The role of the NMD factor UPF3B in olfactory sensory neurons

Kun Tan; Samantha H Jones; Blue B Lake; Jennifer N Chousal; Eleen Y Shum; Lingjuan Zhang; Song Chen; Abhishek Sohni; Shivam Pandya; Richard L Gallo; Kun Zhang; Heidi Cook-Andersen; Miles F Wilkinson

doi:10.7554/eLife.57525

. 2020 Aug 10;9:e57525. doi: 10.7554/eLife.57525

The role of the NMD factor UPF3B in olfactory sensory neurons

Kun Tan ^1,^†, Samantha H Jones ^1,^†, Blue B Lake ², Jennifer N Chousal ¹, Eleen Y Shum ¹, Lingjuan Zhang ³, Song Chen ², Abhishek Sohni ¹, Shivam Pandya ¹, Richard L Gallo ³, Kun Zhang ², Heidi Cook-Andersen ^1,⁴, Miles F Wilkinson ^1,^5,^✉

Editors: Didier YR Stainier⁶, Didier YR Stainier⁷

PMCID: PMC7452722 PMID: 32773035

Abstract

The UPF3B-dependent branch of the nonsense-mediated RNA decay (NMD) pathway is critical for human cognition. Here, we examined the role of UPF3B in the olfactory system. Single-cell RNA-sequencing (scRNA-seq) analysis demonstrated considerable heterogeneity of olfactory sensory neuron (OSN) cell populations in wild-type (WT) mice, and revealed that UPF3B loss influences specific subsets of these cell populations. UPF3B also regulates the expression of a large cadre of antimicrobial genes in OSNs, and promotes the selection of specific olfactory receptor (Olfr) genes for expression in mature OSNs (mOSNs). RNA-seq and Ribotag analyses identified classes of mRNAs expressed and translated at different levels in WT and Upf3b-null mOSNs. Integrating multiple computational approaches, UPF3B-dependent NMD target transcripts that are candidates to mediate the functions of NMD in mOSNs were identified in vivo. Together, our data provides a valuable resource for the olfactory field and insights into the roles of NMD in vivo.

Research organism: Mouse

Introduction

Nonsense-mediated RNA decay (NMD) is a conserved pathway originally discovered by virtue of its ability to degrade aberrant RNAs harboring premature termination codons (PTCs) and thus protect cells from truncated, potentially toxic, dominant-negative proteins (Chang et al., 2007; Conti and Izaurralde, 2005; Kurosaki et al., 2019; Lykke-Andersen and Jensen, 2015; Palacios, 2013). Subsequently, it was discovered that NMD degrades subsets of normal RNAs, with loss or disruption of NMD leading to dysregulation of 5–20% of the normal transcriptome in species spanning the phylogenetic scale (Chan et al., 2007; He et al., 2003; Mendell et al., 2004). This discovery raised the possibility that the function of NMD extends beyond quality control. This notion has been supported by scores of subsequent studies showing that NMD factors are critical for many fundamental processes, including development, differentiation, cell proliferation, the integrated stress response, the unfolded protein response, and autophagy (Chang et al., 2007; Karam et al., 2015; Kurosaki et al., 2019; Nasif et al., 2018).

NMD is well-studied at the biochemical level, with over 15 proteins known to be involved in this pathway (Chang et al., 2007; Kurosaki et al., 2019). Three of these proteins—UPF1, UPF2, and UPF3—are present in all eukaryotes and considered to be the core NMD factors (Conti and Izaurralde, 2005). UPF1 is an RNA helicase that forms a complex with the adaptor proteins UPF2 and UPF3. In vertebrates, UPF3 is encoded by two paralogs: UPF3A (also called ‘UPF3’) and UPF3B (also called ‘UPF3X’). UPF3A serves as a weak NMD factor and NMD repressor, while UPF3B is a NMD branch-specific factor that stimulates NMD (Chan et al., 2007; Shum et al., 2016). UPF3B directly binds to the exon-junction complex (EJC), a large multi-subunit complex recruited to RNAs just upstream of exon-exon junctions after RNA splicing (Woodward et al., 2017). The EJC triggers NMD when allowed to interact with other NMD factors. Evidence suggests that EJCs are displaced by the ribosome during the pioneer round of translation, and thus only EJCs deposited downstream of the stop codon defining the main open-reading frame (ORF) are able to elicit NMD (Dostie and Dreyfuss, 2002). Ribosomes would also be predicted to displace the last EJC when the termination codon resides ~50 nucleotides or less upstream of the last exon-exon junction, based on the length of the EJC and ribosome footprints. This has led to the ‘−50-nt boundary rule,’ an empirically verified dictum which states that only in-frame stop codons further than ~50 nt upstream of the last exon-exon junction elicit NMD (Nagy and Maquat, 1998). While there are exceptions to this −50-nt boundary rule (Carter et al., 1996), it reliably predicts a large proportion of EJC-dependent NMD target mRNAs (Boehm et al., 2014; Gehring et al., 2005; Hurt et al., 2013). NMD can be triggered by other molecular signals in addition to downstream EJCs. For example, long 3’-untranslated regions (UTRs) and short ORFs upstream of the main ORF (uORFs) can, in some cases, trigger NMD in an EJC-independent manner (Barrett et al., 2012; Bühler et al., 2006; Chang et al., 2007; Hurt et al., 2013; Kebaara and Atkin, 2009; Rebbapragada and Lykke-Andersen, 2009).

While considerable progress has been made in understanding the molecular features that elicit NMD, we are still largely in the dark with regard to which transcripts are targeted for rapid decay. It is critical to define such NMD target mRNAs in order to begin to unravel the molecular mechanisms by which NMD influences biological processes. A particularly large gap in the field is the identity of NMD targets in vivo.

Many lines of evidence suggests that NMD is not a single linear pathway but instead consists of several branches, each of which depends on different factors and promotes the decay of different sets of transcripts (Chan et al., 2007; Gehring et al., 2005; Mabin et al., 2018). In this report, we focus on the UPF3B-dependent branch of NMD, which has been shown to be important for the nervous system. Pedigree analysis of numerous families harboring mutations in the UPF3B gene have demonstrated that both nonsense and missense mutations cause intellectual disability in humans (Nguyen et al., 2014; Tarpey et al., 2007). Humans with UPF3B mutations also commonly have autism, schizophrenia, and/or attention-deficit/hyperactivity disorder (Nguyen et al., 2014; Tarpey et al., 2007). To understand the underlying mechanism for these behavioral defects, we generated Upf3b-deficient mice (Huang et al., 2011; Karam et al., 2015). These Upf3b-null mice suffer from specific learning and memory deficits, including fear-conditioned learning, and thus replicate some aspects of the behavioral defects in UPF3B-deficient humans (Huang et al., 2018). In part, these behavioral defects may stem from abnormal neural connectivity, as cortical pyramidal neurons from Upf3b-null mice undergo impaired dendritic spine maturation in vivo (Huang et al., 2018). Furthermore, cultured UPF3B-depleted neural cells have subtle dendrite outgrowth defects (Jolly et al., 2013), and expression of UPF3B mutants reduces neurite branching (Alrahbeni et al., 2015). The behavioral defects in Upf3b-null mice may also result from neural differentiation and/or maturation defects that were uncovered using loss-of-function approaches in neural precursor cells in vitro, or by forced expression of UPF3B mutants in cell lines in vitro (Alrahbeni et al., 2015; Huang et al., 2018; Jolly et al., 2013).

In this communication, we examine the role of UPF3B in the olfactory system, a useful model for studying neural development and function. There is also considerable clinical interest in the olfactory system, as olfactory defects predict the later onset of numerous CNS disorders, including Parkinson’s and Alzheimer’s disease (Doty, 2012). Olfactory dysfunction also strongly associates with autism (Rozenkrantz et al., 2015). The olfactory epithelium (OE) retains a life-long capacity for neurogenesis and harbors a robust regeneration system that responds to injury (Whitman and Greer, 2009). Importantly, the olfactory system is much simpler than the CNS. Mature olfactory sensory neurons (mOSNs) develop via a relatively simple linear pathway involving horizontal basal cells (HBCs), globose basal cells (GBCs), and immature olfactory sensory neurons (iOSNs). Both HBCs and GBCs are stem cells, but the two types have different roles (Schwob et al., 2017). HBCs are reserve stem cells, as they are normally quiescent and only undergo proliferative expansion in response to OE injury (Peterson et al., 2019). In contrast, GBCs are a heterogeneous cell population that consists of constitutively active stem cells as well as progenitors (Schwob et al., 2017). Lineage-tracing analysis and single-cell RNA sequencing (scRNA-seq) analysis have shown that after proliferative expansion, HBCs and GBCs give rise to iOSNs, which are responsible for undergoing maturation (Fletcher et al., 2017). Of note, iOSNs share markers with another OSN stage called ‘immediate neural precursors (INPs).” Given the ambiguity of the nomenclature, we will refer to cells with either INP or iOSN characteristics as iOSNs. iOSNs ultimately differentiate into mOSNs, which send an axon to neurons in the glomeruli region of the olfactory bulb, relaying olfactory information from the outside world to the CNS. mOSNs recognize odorants through chemosensory receptors, including olfactory receptors (OLFRs), members of the G-protein-coupled receptor super-family, as well as trace amine–associated receptors, guanylate cyclases, and members of the membrane-spanning 4-pass A gene family (Bear et al., 2016; Saraiva et al., 2019).

To gain insight into the nature of the cells in the OE and their developmental relationships, recent studies have performed transcriptome profiling using whole OE, pools of sorted OSNs, single OSNs, or single OE cells (Fletcher et al., 2017; Ibarra-Soria et al., 2014; Saraiva et al., 2015; Saraiva et al., 2019; Tan et al., 2015). These studies have revealed new OE cell subsets, inferred the developmental pathways of both OSN and non-neural OE cells, defined classes of genes exhibiting enriched expression and unique patterns of expression in different OE subsets, and revealed the expression patterns and dynamics of OLFRs during OSN development and in individual mOSNs. These studies have also advanced our understanding of mammalian olfaction evolution.

In this study, we ascertain whether the NMD factor, UPF3B, has roles in the olfactory system. Using scRNA-seq and RNA-seq analyses, we obtained evidence that UPF3B influences the frequency of specific OSN subsets, broadly suppresses the expression of immune genes in OSNs, and shapes the Olfr gene repertoire. We also identified high-confidence NMD target mRNAs in vivo that are candidates to act downstream of UPF3B in mOSNs. As part of our analysis, we also provide new cellular and molecular information on WT OSNs and their development in vivo. Our findings in Upf3b-null NMD-deficient mice introduce a useful biological system to understand the role of RNA metabolism in neurons, and our scRNA-seq, RNA-seq, and RiboTag datasets are new resources that can be used by the olfactory field.

Results

UPF3B-regulated genes in mOSNs

To assess whether NMD-deficient Upf3b-null mice have an olfactory defect, we measured their weight during their growth phase. This follows from the fact that newborn mice are blind and therefore depend on the olfactory system to initiate milk suckling for survival (Logan et al., 2012). We quantified the weight of Upf3b-null and littermate WT mice and found that Upf3b-null mice have a statistically significant postnatal weight deficit (p<0.05; Figure 1A). The weight deficit occurs soon after birth, becomes progressively worse during postnatal development, and is corrected after reaching adulthood. This specific pattern of weight loss is characteristic of mice harboring a partial olfactory defect (Riera et al., 2017). In contrast, newborn mice that completely lack sense of smell are incapable of sensing their source of milk and die soon after birth (Hongo et al., 2000). As further evidence that Upf3b-null mice have an olfactory defect, we found that HBC, iOSN and mOSN marker genes (Krt5, Gap43, and Gnal, respectively) exhibited significantly decreased expression in Upf3b-null as compared to WT OE (Figure 1B). We followed up by testing Upf3b-null mice for evidence for specific olfactory deficits and observed trends but did not observe statistically significant effects (Figure 1—figure supplement 1A), providing further evidence of a partial olfactory defect.

Figure 1. — (A) The weight of *Upf3b*-null vs. WT (wild type) mice at the indicated time points. *Upf3b*-null mice gain weight slowly during postnatal development but then reach the weight of WT mice at the last time point (9 weeks), a pattern indicative of a partial olfactory defect. *, p<0.05. (B) qPCR analysis of olfactory marker genes in *Upf3b*-null and WT OE (n = 6). **, p<0.01; ****, p<0.0001. (C) Heatmap of genes differentially expressed in mOSNs from *Upf3b*-null (KO) vs. WT mice (four biological replicates from each are shown). Row names labeled as green are *Olfr* genes. Right, the most statistically significant GO terms associated with upregulated genes (top) and downregulated genes (bottom) after *Upf3b* loss. (D) A list of most statistically enriched GO terms associated with the 52 high-confidence UPF3B-dependent NMD target mRNAs we identified in mOSNs.

Figure 1—figure supplement 1. — (A) The weight of *Upf3b*-null vs. WT (wild type) mice at the indicated time points. *Upf3b*-null mice gain weight slowly during postnatal development but then reach the weight of WT mice at the last time point (9 weeks), a pattern indicative of a partial olfactory defect. *, p<0.05. (B) qPCR analysis of olfactory marker genes in *Upf3b*-null and WT OE (n = 6). **, p<0.01; ****, p<0.0001. (C) Heatmap of genes differentially expressed in mOSNs from *Upf3b*-null (KO) vs. WT mice (four biological replicates from each are shown). Row names labeled as green are *Olfr* genes. Right, the most statistically significant GO terms associated with upregulated genes (top) and downregulated genes (bottom) after *Upf3b* loss. (D) A list of most statistically enriched GO terms associated with the 52 high-confidence UPF3B-dependent NMD target mRNAs we identified in mOSNs.

Given that mOSNs are the functional units of the OE, we next focused our attention on these cells. We identified UPF3B-regulated genes in mOSNs by performing RNA-seq analysis on FACS-purified mOSNs (YFP+ cells) from R26-eYFP; Omp-Cre mice (Figure 1—figure supplement 1B). Four samples were analyzed from each genotype (Figure 1—figure supplement 2A and Supplementary file 1). The expression of OSN precursor/OSN canonical markers are shown in Figure 1—figure supplement 2D. RNA-seq analysis identified 235 differentially expressed genes between Upf3b-null and WT mOSNs (q < 0.05) (Figure 1C and Supplementary file 2). We validated our RNA-seq analysis by qPCR analysis and immunofluorescence (Figure 1—figure supplement 2B,C).

Among the 127 upregulated genes were several involved in neurogenesis, including Lrp2, Hk2, Notch2, Gdf11, Fos, Ptch1, and Spry2. Gene ontology (GO) analysis revealed enrichment for ‘organ/system development,’ ‘cell-cell adhesion,’ ‘leukocyte activation,’ and ‘cell differentiation/proliferation’ functions (Figure 1C). In contrast, the 108 downregulated genes were most enriched for GO functions associated with olfaction: ‘sensory perception,’ ‘G-proteins,’ ‘detection of chemical stimulus,’ and ‘signal transduction.’ Indeed, we found that the majority (78 out of 108) of these significantly downregulated genes are Olfr genes (marked in green in Figure 1C; the expression of all Olfr genes in Upf3b-null and control mOSNs is shown in Figure 1—figure supplement 2E and Supplementary file 1). We follow-up on this surprising finding below. Other genes downregulated in Upf3b-null mOSNs include those involved in CNS synaptic transmission (Slc17a6), chromatin remodeling (Chd1), and sensory neuronal plasticity (Cwc22) (Supplementary file 2).

Identification of NMD target mRNAs in mOSNs

NMD is thought to influence biological systems by virtue of its ability to promote the decay of specific subsets of mRNAs (Lykke-Andersen and Jensen, 2015). As described in the introduction, there is dearth of knowledge regarding the identity of such NMD target RNAs, particularly in cells in their normal in vivo context. Our RNA-seq analysis of purified mOSNs from Upf3b-null and WT mice provided an opportunity to identify in vivo direct NMD targets. Because NMD is a negative regulatory pathway (it degrades its targets), the 127 RNAs upregulated in Upf3b-null mOSNs are candidates to be direct NMD targets. Among them, we found that 73 had at least one of the well-established molecular features known to elicit NMD, including an exon-exon junction >50 nt downstream of the main ORF (dEJ) (Table 1; see the Introduction for an explanation of NMD-inducing features [NIFs]). Thus, these 73 mRNAs are strong candidates to be UPF3B-dependent NMD target mRNAs in mOSNs.

Table 1. UPF3B-dependent NMD target mRNAs in mOSNs.

Symbol	log2FC (KO/WT)	Padj	dEJ	uORF	3'UTR length	Symbol	log2FC (KO/WT)	Padj	dEJ	uORF	3'UTR length
Prelid3a	1.099967	0.003745	YES	NO	1572	Fmo2	2.02039	0.014815	NO	NO	2411
1700025G04Rik	0.662926	0.012989	NO	YES	8870	Gab2	0.98414	0.003018	NO	NO	3927
6030419C18Rik	0.73232	0.036112	NO	YES	55	Gdf11	1.429234	0.005353	NO	NO	2811
9330159F19Rik	0.542375	0.017617	NO	YES	3408	Gldn	2.115908	0.045841	NO	NO	2970
Adcy6	2.587005	0.002078	NO	YES	2356	Hk2	2.296045	0.033161	NO	NO	2285
Cdh24	1.560901	0.001303	NO	YES	121	Lbh	1.417311	0.024315	NO	NO	2498
Fam84b	0.719841	0.001704	NO	YES	3969	Luc7l	0.492061	8.19E-05	NO	NO	3738
Inpp5f	1.178064	0.043839	NO	YES	949	Map3k9	0.79841	0.021555	NO	NO	1029
Lrp2	2.504276	0.008534	NO	YES	1305	Msrb3	1.668851	0.033388	NO	NO	2972
Mafg	0.577538	0.046713	NO	YES	4167	Neurl3	1.966306	0.00546	NO	NO	1763
Plxnc1	2.322167	0.048567	NO	YES	2320	Notch2	1.68375	0.047733	NO	NO	2917
Prdm4	0.420203	0.027945	NO	YES	1160	Plekha5	0.608634	0.004216	NO	NO	3461
Ptch1	0.768864	0.01088	NO	YES	3205	Rab43	1.033148	0.0151	NO	NO	3737
Ptger2	3.032221	0.035664	NO	YES	1825	Rac2	3.029035	0.038392	NO	NO	2319
Sash3	2.352656	0.033245	NO	YES	1309	Raver2	1.921185	0.027779	NO	NO	1892
Serpinb11	1.991555	0.002719	NO	YES	468	Rflnb	0.755198	0.017617	NO	NO	2716
Snx33	1.512032	0.012417	NO	YES	1258	Sik1	2.027635	1.48E-06	NO	NO	2035
Zfp36	1.802697	0.025165	NO	YES	774	Slc38a6	1.21151	0.025847	NO	NO	1512
Agap2	1.264604	0.00099	NO	NO	1357	Slc5a1	2.563582	0.00527	NO	NO	1868
Aox2	1.36834	0.018035	NO	NO	1640	Swap70	1.863436	0.009993	NO	NO	2169
Atp10d	3.315656	0.017617	NO	NO	2384	Tgm2	2.395934	0.042993	NO	NO	1399
Bhlhe40	1.435423	0.000192	NO	NO	1593	Themis2	3.496025	0.015464	NO	NO	1053
Btg2	1.281148	0.000173	NO	NO	2199	Tmprss2	2.167673	0.005867	NO	NO	1456
Cybrd1	2.372842	0.002733	NO	NO	4269	Tob2	0.667082	0.001453	NO	NO	2459
Cyth4	2.162221	0.045105	NO	NO	1455	Ywhag	0.644673	0.017707	NO	NO	2586
Ermn	1.686519	0.005793	NO	NO	2641	Zcchc6	0.512078	0.003018	NO	NO	1346

Open in a new tab

Given that NMD degrades its target RNAs, this predicts that its targets should be stabilized after inactivation of UPF3B. Thus, we measured the stability of the 127 mRNAs upregulated in Upf3b-null mOSNs using a method that infers RNA stability based on pre-mRNA and steady-state mRNA levels (Alkallas et al., 2017). This method revealed that 82 of 127 upregulated genes encode mRNAs stabilized in Upf3b-null mOSNs as compared to WT mOSNs (Supplementary file 2). Of these 82 stabilized and upregulated mRNAs, 52 have at least 1 of the 3 well-established NIFs (Table 1), and thus we classified these 52 mRNAs as high-confidence mOSN NMD targets. The statistically enriched GO biological functions of the proteins encoded by these 52 mRNAs are listed in Figure 1D.

To determine whether these high-confidence NMD target mRNAs correspond to known NMD targets, we assembled a list of likely mouse NMD substrates defined by previous studies (Supplementary file 3). To qualify to be in this list, the RNA must have at least one known NMD-inducing feature (NIF) (Palacios, 2013) and experimental evidence from at least one assay that it is an NMD substrate (e.g. high UPF1 occupancy or upregulation and/or stabilization in response to NMD-factor depletion). We found that 11 of these previously defined likely mouse NMD target mRNAs overlapped with the 52 high-confidence targets identified in our study: Atp10d, Lbh, Slc38a6, Tgm2, Notch2, Ywhag, Luc7l, Ptch1, 1700025G04Rik, Ptger2, and Msrb3. Of note, it is not surprising that only a proportion of the upregulated mRNAs we identified in NMD-deficient mOSNs are previously known NMD targets, as NMD target mRNAs can be tissue-, cell type-, and NMD factor-specific (Huang et al., 2011). The list of previously defined candidate NMD targets that we compared with were defined in non-neuronal tissues and cell lines made deficient in NMD by knocking down or eliminating factors other than UPF3B (Supplementary file 3).

The mOSN transcriptome and translome

We next determined the translation rate of mRNAs in mOSNs, both as a resource for the field and to address the relationship of NMD with translation in vivo. We assayed the translation rate of mRNAs in mOSNs using RiboTag mice, which express an epitope-tagged ribosomal protein, RPL22^HA, which is incorporated into actively translating ribosomes specifically in cells expressing CRE (Sanz et al., 2009). Immunoprecipitation (IP) of the cell lysates of interest with an HA antibody purifies the ribosome-associated mRNAs (Figure 2A, left) with an efficiency associated with polysome density (Hornstein et al., 2016). To examine ribosome density specifically in WT mOSNs, we isolated RiboTag-labeled mRNA from the OE of RiboTag; Omp-Cre mice and performed RNA-seq analysis. As a validation of cell-type specificity, we found that IP of OE lysates with the HA antisera enriched for the mOSN marker, Omp, whereas these lysates were depleted of the HBC and GBC markers, Krt5 and Lgr5, respectively (Figure 2A, right). We then elucidated inferred translation efficiency (TE) for all expressed mRNAs in mOSNs – the ‘mOSN translome’ – by calculating the ratio of the IP signal from the RiboTag mice OE lysates over mOSN steady-state mRNA level, the latter determined as described above (Supplementary file 2).

Figure 2. — (A) Left, strategy used to define the mOSN translome. Right, RNAseq analysis of the expression of gene markers for HBCs (*Krt5*), GBCs (*Lgr5*), and mOSNs (*Omp*) in the indicated samples. (B) Average inferred translation efficiency (TE) of mOSN mRNAs with the indicated 3’UTR length ranges. *, p<0.05. (C) Average 3’UTR length of mOSN mRNAs with the indicated range of TE values. *, p<0.05. (D) mOSN mRNAs from WT mice stratified by steady-state mRNA level (transcriptome) and TE. The number of genes in each category is indicated. (E) Top enriched GO terms associated with the different categories of genes defined in (D). (F) Analysis of upregulated mRNAs (candidate NMD targets) and downregulated mRNAs (indirect targets) are shown on the left and right, respectively. The average shift in expression in *Upf3b*-null mOSNs relative to WT mOSNs is shown for mOSNs binned by TE (a and c have the highest and lowest TE values, respectively). *, p<0.05. (G) Scatter plot of the 52 high-confidence mOSN NMD targets, showing TE vs. NMD magnitude (upregulation in *Upf3b*-null mOSNs). Both values are log2-transformed. (H) Scatterplot showing the TE of mRNAs in *Upf3b*-null vs. WT mOSNs. (I) mRNAs exhibiting significantly altered TE in response to *Upf3b* loss. (J) mOSN mRNAs from *Upf3b-*null (KO) mice stratified by steady-state mRNA level (transcriptome) and TE. The number of cells in each category is indicated.

Given that 3’UTR length has been shown to influence translation rates in cultured cells (Spies et al., 2013), we examined the relationship of 3’UTR length and TE in mOSNs in vivo. We found that mOSN mRNAs harboring 3’UTRs of >2 kb have much higher average TE than mOSN mRNAs harboring shorter 3’UTRs (Figure 2B). Highly translated mOSN mRNAs have an average 3’UTR length of ~1.8 kb, while lowly translated mOSN mRNAs have an average 3’UTR length of only ~0.9 kb (Figure 2C).

To assess the potential functional relevance of translation, we binned WT mOSN mRNAs into three groups: high (top 30%), medium (middle 40%), and low (bottom 30%) (Supplementary file 2). We also binned WT mOSN mRNAs into three groups based on their steady-state level (Supplementary file 2), allowing us to place mOSN mRNAs into the nine categories shown in Figure 2D. GO analysis revealed that category #1—which is mRNAs expressed at high level that are also highly translated—encode proteins that tend to function in ‘metabolism,’ ‘intercellular transport,’ and ‘catabolism’ (Figure 2E). Categories #2 and #3—which are also highly translated mRNAs but less well expressed at the RNA level than category #1—encode proteins with strikingly different functions: ‘development,’ ‘cell migration,’ and ‘morphogenesis’ (Figure 2E). Category #6—which is lowly expressed and modestly translated mRNAs—encode proteins involved in ‘signal transduction,’ ‘differentiation,’ and ‘development,’ including ‘nervous system development’ (Figure 2E). The categories with most Olfr genes—#4 and #5—are also only moderately translated (Figure 2E). Upf3b-null mOSNs had similar numbers of mRNAs in the nine categories as WT mOSNs (compare Figure 2J with Figure 2D), indicative of UPF3B not altering the mOSN transcriptome and translome globally. Rather, UPF3B influences specific mRNAs, as described above for the mOSN transcriptome, and below for the mOSN translome.

The relationship between NMD and translation in vivo

NMD is a translation-dependent pathway, based on protein-synthesis inhibitor and transfection experiments in immortalized cell lines (Belgrader et al., 1993; Carter et al., 1995; Karousis and Mühlemann, 2019). Our mOSN transcriptome and translome data from Upf3b-null and WT mice provided an opportunity to address the relationship of NMD with translation in vivo. Given that higher translation rates allow for a higher frequency of stop codon recognition, it follows that higher translation rates might drive stronger NMD. This predicts that more highly translated mOSN mRNAs will have a higher NMD response than lowly translated mOSN mRNAs. To test this, we binned mRNAs statistically upregulated in Upf3b-null mOSNs into three groups stratified by TE. The most highly translated group was statistically more upregulated (i.e., had stronger NMD) than the least translated group (Figure 2F, left). As a negative control, we examined downregulated mRNAs (as these would not be direct NMD targets) and found no statistical difference between degree of downregulation and TE (Figure 2F, right).

To further examine whether high translation rate is associated with strong NMD magnitude, we binned the 52 high-confidence NMD substrates we defined above into two groups: those with little or no translation and those with high translation (cut-off: log2TE > 1). We then independently plotted these two sets of mRNAs in terms of TE and NMD magnitude (i.e. the degree of upregulation in Upf3b-null mOSNs relative to WT mOSNs). The results show that the high-translation group exhibited a correlation between their inferred translation rate and NMD magnitude (R² = 0.5; Figure 2G). In contrast, the low-translation group of mRNAs exhibited no correlation between their translation rate and NMD magnitude (R² = 0.06; Figure 2G). Together, these results support that NMD is translation-dependent in vivo and that its magnitude tends to be enhanced for highly translated mRNAs.

Our mOSN translome data also allowed us to assess the reciprocal question: does Upf3b influence translation in vivo? When we plotted the TE of mRNAs when expressed in Upf3b-null mOSNs vs. when expressed in WT mOSNs, we found that the vast majority of mRNAs were similarly translated in both genetic backgrounds, as measured by RiboTag analysis (Figure 2H, Supplementary file 2). Only 16 mOSN mRNAs migrated off the diagonal and thus had a significant change in TE as a result of Upf3b loss (Figure 2H,I).

Identification of OE cell clusters

To determine whether UPF3B influences the cellular composition of the OE, we performed scRNA-seq analysis on dissociated OE cells from 4 Upf3b-null and 4 WT mice. After filtering out poor quality cells, 25,165 cells remained for subsequent analysis. Biological replicates exhibited similar cell distributions (Figure 3A). Using a nonlinear dimensionality-reduction technique—uniform manifold approximation and projection (UMAP)—we identified cell clusters corresponding to 16 known cell types in the OE (Figure 3B). Some of the gene markers used to define these cell clusters are shown in Figure 3C. Genes exhibiting enriched expression in each of the 16 cell types are listed in Supplementary file 4.

Re-clustering of OSN precursors/OSNs (HBCs, GBCs, iOSNs, and mOSNs) revealed several cell sub-clusters within each of these four stages (Figure 3D,E). The identification of these sub-clusters suggested that each of these developmental stages exhibit considerable heterogeneity, at least at the transcriptome level. Genes exhibiting enriched expression in each sub-cluster are shown in Supplementary file 4.

HBC are known to be reserve stem cells, while GBCs consist of active stem cells and progenitors (Schwob et al., 2017). Consistent with this, cell-cycle analysis showed that all four HBC sub-clusters primarily contain quiescent cells, while the GBC sub-clusters have many cells that are proliferating (Figure 3F). All four HBC sub-clusters express similar levels of well-established HBC markers, including Krt5 and Trp63 (Figure 4A). These HBC sub-clusters are each uniquely marked by novel gene markers that we identified (Figure 4A and Supplementary file 4).

GBCs also segregated into four sub-clusters (Figure 3D), which is consistent with past studies demonstrating that GBCs are heterogeneous (Cau et al., 1997; Manglapus et al., 2004; Schwob et al., 2017). Monocle pseudotime trajectory analysis suggested that these four sub-clusters have a linear developmental relationship, with GBC-1 the most immature, GBC-2 more advanced, and GBC-3 and −4 the most advanced (Figure 4B). Consistent with this developmental trajectory, both GBC-1 and GBC-2 express the early GBC marker Ascl1 (Cau et al., 1997; Manglapus et al., 2004). GBC-1 is likely to be more primitive than GBC-2, based on the frequent and high expression of the later GBC markers Neurog1 and Neurod1 (Cau et al., 1997; Manglapus et al., 2004) in the latter, not the former (Figure 4C). While GBC-3 and −4 are clearly GBCs based on the expression of several GBC markers (e.g. Neurog1 and Neurod1), they also express iOSN markers (e.g. Gng8 and Gap43 [Iwema and Schwob, 2003; Figure 4C]), consistent with GBC-3 and -4 being iOSN precursors and hence advanced GBCs. iOSNs also segregated into several cell sub-clusters that each express unique genes (Figure 4D). These iOSN sub-clusters follow a ‘linear’ pattern as depicted by the UMAP algorithm (Figure 3D), consistent with them representing sequential developmental states, each with unique transcriptomes.

Most WT mOSNs segregated into three different cell clusters (Figure 3D), each of which preferentially express different genes (Figure 4E). GO and KEGG signaling pathway analyses indicated that these three mOSN sub-clusters are enriched for different functions and signaling pathways, respectively (Figure 4F; Supplementary file 4).

OSN molecular pathways

Monocle pseudotime analysis of the OSN precursor/OSN cell clusters indicated that they follow a HBC→GBC→iOSN→mOSN trajectory (Figure 5A), consistent with previous studies (Fletcher et al., 2017; Schwob et al., 2017; Tepe et al., 2018). To define candidate molecular events occurring during OSN development, we identified genes whose expression is statistically enriched along this pseudotime trajectory (Supplementary file 4). This analysis identified 4 distinct patterns of gene expression dynamics that we named groups 1 to 4 (Figure 5B). Group-1 genes are dominated by genes expressed transiently in HBCs, including the previously defined HBC-marker genes Trp63, Krt5, and Krt14. Group-1 genes are statistically enriched for ‘signal transduction’ and various ‘development’ categories (Figure 5B). Group-2 genes contain GBC genes; indeed the GBC markers Ascl1, Neurod1, and Lgr5 are enriched in group 2. ‘Cell cycle process’ is statistically enriched (Figure 5B), consistent with the fact that GBCs undergo self-renewal and proliferative expansion. Group-3 genes are mainly expressed in iOSNs, include the well-established iOSN marker genes Lhx2, Ncam1, and Gap43. ‘Neuron development’ is enriched in group 3 (Figure 5B), consistent with the fact that iOSNs are undergoing the final stages of development prior to becoming mature neurons. Group-4 genes are mainly expressed in mOSNs; enriched GO categories include ‘mitochondrion organization,’ ‘metabolism,’ and ‘cellular respiration’.

KEGG signaling pathway analysis revealed that genes involved in different signaling pathways are enriched in each of the 4 groups (Figure 5C). For example, Hippo pathway genes are enriched in group 1 (Figure 5D), raising the possibility this signaling pathway may be important for maintaining HBC stem cells in the quiescent state or eliciting their activation in response to insults.

We also screened for transcription factors preferentially expressed at different stages of OSN development. We identified 209, 178, 169, and 135 transcription factor genes exhibiting enriched expression in groups 1, 2, 3 and 4, respectively (Supplementary file 4). The top 3 transcription factors in each group and their DNA-binding specificity are shown in Figure 5E.

UPF3B impacts HBCs and mOSNs

The array of UPF3B-dependent NMD targets we identified in mOSNs (Figure 2) raised the possibility that UPF3B has roles in mOSNs and possibly OSN precursors. To assess this, we first determined whether loss of UPF3B impacts the frequency of HBCs, GBCs, iOSNs, and mOSNs. scRNA-seq analysis revealed that there was a significant reduction in the frequency of HBCs in Upf3b-null mice relative to WT mice, when compared to either all OSN precursors/OSNs or all OE cells (p<0.05; Figure 6A). As validation, IHC staining with the HBC marker, TRP63, showed that the density of TRP63+ cells was significantly less in Upf3b-null OE than WT OE (Figure 6—figure supplement 1A). This effect appeared to be specific, as we observed no significant difference in the relative proportion of GBCs, iOSNs, and mOSNs between Upf3b-null and WT mice (Figure 6—figure supplement 1B). However, we cannot rule out that the variability among the four samples for each genotype might have obscured a subtle change in the fraction of GBCs, iOSNs, or mOSNs in Upf3b-null mice. This variability might either be the result of biological differences between individual mice or differences in dissection and/or cell dissociation. However, as further evidence that the overall frequency of mOSNs was not affected in Upf3b-null mice, the mOSN marker, OMP, was similarly expressed (at both the RNA and protein levels) in OE from Upf3b-null and WT mice, as assessed in OE preparations obtained from different mice (but of the same genotypes) than those used for scRNA-seq analysis (Figure 1B and Figure 6—figure supplement 1C).

Figure 6. — (A) The fraction of HBCs per all OSN precursors/OSNs (HBCs, GBCs, iOSNs and mOSNs) (left) or all OE cells (right), in *Upf3b*-null (KO) and WT mice, as determined by scRNA-seq analysis. *, p<0.05. (B) The percentage of cells from the indicated cell sub-clusters in *Upf3b*-null (KO) and WT mice, as determined by scRNAseq analysis. (C) Cell number in each cell sub-cluster, as defined in Figure 3D. (D) Most statistically enriched GO terms in the mOSN-2 and −4 sub-clusters. (E) Heatmap depicting the expression pattern of anti-microbial genes in the indicated cell subsets. (F) Left: Western blot analysis of endogenous CAMP protein level in the OE from *Upf3b-*null (KO) and WT mice. Right: quantification of CAMP level normalized against GAPDH (n = 3). *, p<0.05. (G) IF analysis of adult mouse OE sections co-stained with antisera against CAMP (red) and OMP (green). Nuclei were stained with DAPI (blue). (**H, I**) The percentage of mOSNs (H) and iOSNs (I) in our scRNAseq datasets that express *Olfr* genes. Left, all known *Olfr* genes. Right, the 78 *Olfr* genes significantly downregulated in *Upf3b*-null mice, based on RNAseq analysis (Figure 1C). *, p<0.05.

Figure 6—figure supplement 1. — (A) The fraction of HBCs per all OSN precursors/OSNs (HBCs, GBCs, iOSNs and mOSNs) (left) or all OE cells (right), in *Upf3b*-null (KO) and WT mice, as determined by scRNA-seq analysis. *, p<0.05. (B) The percentage of cells from the indicated cell sub-clusters in *Upf3b*-null (KO) and WT mice, as determined by scRNAseq analysis. (C) Cell number in each cell sub-cluster, as defined in Figure 3D. (D) Most statistically enriched GO terms in the mOSN-2 and −4 sub-clusters. (E) Heatmap depicting the expression pattern of anti-microbial genes in the indicated cell subsets. (F) Left: Western blot analysis of endogenous CAMP protein level in the OE from *Upf3b-*null (KO) and WT mice. Right: quantification of CAMP level normalized against GAPDH (n = 3). *, p<0.05. (G) IF analysis of adult mouse OE sections co-stained with antisera against CAMP (red) and OMP (green). Nuclei were stained with DAPI (blue). (**H, I**) The percentage of mOSNs (H) and iOSNs (I) in our scRNAseq datasets that express *Olfr* genes. Left, all known *Olfr* genes. Right, the 78 *Olfr* genes significantly downregulated in *Upf3b*-null mice, based on RNAseq analysis (Figure 1C). *, p<0.05.

Our identification of HBC, GBC, iOSN, and mOSN sub-clusters (Figure 3D) gave us an opportunity to elucidate whether UPF3B has a role in this unexpected heterogeneity. Despite no significant effect on the mOSN stage as a whole (Figure 6—figure supplement 1B), we observed a striking increase in the frequency of 1 of the 4 mOSN sub-clusters—mOSN-4—in Upf3b-null mice (Figure 6B,C). This sub-cluster represented <1% of all OSNs in most WT mice and increased by an average of 25-fold in Upf3b-null mice (p<0.05). Conversely, Upf3b-null mice had an almost complete loss of another mOSN sub-cluster—mOSN-2—a sub-cluster that was populated by many cells in most WT mice (Figure 6B,C). While this reduction failed to reach statistical significance because of variability between samples (p=0.24), it is supported by the independent tSNE plots we generated for Upf3b-null vs. WT OSNs (Figure 3D). Together these results raise the possibility that a ‘mOSN subset switch’ occurs in Upf3b-null mice. Pearson correlation analysis showed that mOSN-2 and −4 sub-clusters are less related to each other in expression profile than they are to the other two mOSN sub-clusters (Figure 6—figure supplement 2A). Indeed, these two mOSN sub-clusters have remarkably distinct molecular characteristics (Figure 6D and Supplementary file 4). Thus, the simultaneous loss and acquisition of these mOSN subsets in Upf3b-null mice has the potential to alter olfaction.

UPF3B shapes the OLFR repertoire and suppresses immune gene activation

To define genes that are candidates to act downstream of NMD in different OSN cell populations, we used our scRNA-seq datasets to identify genes differentially expressed in the Upf3b-null vs. WT cell clusters (Supplementary file 4). This revealed that a major category of Upf3b-regulated genes in OSNs are immune genes, including a large fraction of genes encoding antimicrobial proteins (Supplementary file 5). This was intriguing, as it raised the possibility that OSNs not only normally function in olfaction but also in defense against microbes, a reasonable possibility given that the OE is direct contact with the outside environment. The expression of anti-microbial genes was not confined to the mature neurons in the OE (i.e. mOSNs), as we found that most of these immune-defense genes were also expressed and upregulated in Upf3b-null mice at the HBC, GBC, and iOSNs stages (Figure 6E). More than half (48 out of 88) of these upregulated mRNAs encoding immune-related proteins harbor at least one NIF (Supplementary file 5). This suggests that many of these mRNAs encoding immune system factors are directly targeted for decay by the NMD pathway. This has interesting potential physiological consequences, as described in the Discussion.

Among the antimicrobial genes expressed and upregulated in Upf3b-null OSN precursors and OSNs was Camp (also known as ‘Cramp’), which encodes a member of the cathelicidin family of antimicrobial peptides that has an important role in the defense against microbial infections, and functions in cell chemotaxis, immune mediator induction, and inflammatory response regulation (Zhang and Gallo, 2016). To further assess its regulation, we performed western blot analysis with a validated anti-CAMP antiserum, which showed that CAMP protein is expressed in the OE and is upregulated in Upf3b-null mice (Figure 6F). As further evidence, immunofluorescence analysis detected modest anti-CAMP staining in OE cells (as well as strong staining in the in the lamina propria), both of which were increased in Upf3b-null mice (Figure 6G and Figure 6—figure supplement 2B).

The other major category of genes that we discovered are regulated by Upf3b in the OE is Olfr genes. As described above, our RNA-seq analysis of purified mOSNs from Upf3b-null and WT mice revealed that the majority of genes downregulated in response to Upf3b loss are Olfr genes (Figure 1C). In total, we identified 78 Olfr genes significantly downregulated in Upf3b-null mOSNs (Figure 1C). We considered the possibility that these 78 Olfr genes are regulated by Upf3b through a common local cis-acting regulatory sequence, but against this hypothesis, we found that these 78 Olfr genes are widely chromosomally distributed (Figure 6—figure supplement 2C).

There are more than 1000 olfactory receptor (Olfr) genes in mice (Zhang and Firestein, 2002). Individual OSNs select a single Olfr gene for expression from this large repertoire (Chess et al., 1994; Malnic et al., 1999; Serizawa et al., 2003). This unique mechanism raised the possibility that rather than regulating Olfr expression per se, Upf3b might instead influence the decision whether or not specific Olfr genes are selected to be the dominant receptors in mOSNs. In other words, Upf3b might increase the probability that these 78 Olfr genes that are selected to be expressed in individual mOSNs. If selected less often in Upf3b-null OSNs, these 78 Olfr genes would appear to be downregulated, but would instead be expressed in fewer OSN. To address this model, we made use of our scRNA-seq datasets. We found that 490 of 3887 mOSNs in WT mice (13%) express one of these 78 Olfr genes as the dominant Olfr gene. In contrast, only 328 of 4654 mOSNs in Upf3b-null mice (7%) express one of these 78 Olfr genes as the dominant Olfr gene (Figure 6H). In contrast, when all known Olfr genes were considered as a group, there was no significant difference in the percentage of Olfr genes selected in Upf3b-null vs. WT mOSNs (Figure 6H). Likewise, we found that these 78 Olfr genes were under-represented in Upf3b-null iOSNs as compared with WT iOSNs (5.6% vs 8.0%, p<0.05) (Figure 6I). Together, these results support a model in which UPF3B promotes the selection of these 78 Olfr genes to be the dominant Olfr gene expressed in individual mOSNs, a model we elaborate on in the Discussion.

Discussion

NMD factors have been shown to have numerous roles in the development and function of neurons (Jaffrey and Wilkinson, 2018). As described in the Introduction, the NMD factor examined in our study – UPF3B – has been shown to be necessary for normal cognition in humans and its loss is associated with several neuro-developmental disorders (Nguyen et al., 2014; Tarpey et al., 2007). While the precise roles of UPF3B in behaviors is not known, it has been shown that UPF3B is critical for both neural differentiation and mature neuronal functions (Alrahbeni et al., 2015; Huang et al., 2018; Jolly et al., 2013). In addition to UPF3B, other NMD genes are likely to have roles in the nervous system (Jaffrey and Wilkinson, 2018). For example, copy-number gain and loss of several genes encoding proteins involved in NMD—including UPF2, UPF3A, SMG6, RBM8A, EIF4A3, and RNPS1—are statistically significantly associated with neural-developmental disorders in humans (Nguyen et al., 2013). Mutations in RBM8A have been shown to cause TAR syndrome, which can cause cognitive dysfunction (Jaffrey and Wilkinson, 2018). In mice, loss of a single copy of Rbm8a or other EJC genes (Magoh or Eif4e) causes microcephaly and severe neural defects (Mao et al., 2017). In worms, flies and mice, genetic perturbation of other NMD genes causes neural defects, including synaptic and axon guidance defects (Colak et al., 2013; Giorgi et al., 2007; Long et al., 2010; Zheng et al., 2012). Two recent studies revealed that conditional loss of the NMD gene, Upf2, in specific neural populations in mice causes a variety of intriguing defects, including aberrant behavior, spine density, and synaptic plasticity (Johnson et al., 2019; Notaras et al., 2019). Together, these studies make a strong case that NMD has roles in the CNS.

Here, we report the first investigation of the role of NMD in the olfactory system. One of our major findings was that Upf3b loss causes shifts in gene expression in OSNs. One major class of genes impacted by Upf3b is the Olfr genes. These genes have evolved to allow recognition of the large array of odors encountered by higher organisms. In mice, there are >1000 Olfr genes, each of which encode a G-coupled receptor that binds to a restricted set of odorants (Godfrey et al., 2004; Zhang and Firestein, 2002). In order to interpret the information from a given odor, it is critical that only a single OLFR be expressed in each mOSN. This is accomplished by a novel gene regulatory mechanism that selects only a single Olfr gene to be expressed in any given mOSN (Chess et al., 1994; Malnic et al., 1999; Serizawa et al., 2003). While the underlying mechanism for this ‘one‐neuron‐one‐receptor’ rule is not fully understood, a prevailing model is that a stochastic mechanism drives a single Olfr to become dominate transcriptionally, a decision that is reinforced by feedback mechanisms (Dalton et al., 2013; Lewcock and Reed, 2004; Serizawa et al., 2004; Serizawa et al., 2005).

The first indication that UPF3B might have a role in the selection of Olfr genes came from our RNA-seq analysis, which revealed that the majority of genes expressed at lower level in Upf3b-null mOSNs are Olfr genes. In total, we found that 78 Olfr genes are statistically downregulated in Upf3b-null mOSN. To address mechanism, we performed scRNA-seq analysis and found that these 78 Olfr genes are rarely represented as the dominant genes in individual mOSNs in Upf3b-null mice. This defect was also present at the iOSN stage, suggesting that Upf3b is involved directly or indirectly in determining which Olfr gene are selected for dominant expression during OSN development.

A caveat is the OE contains zones enriched for mOSNs expressing particular sets of OLFRs (Miyamichi et al., 2005; Ressler et al., 1994), and thus even though we made an effort to dissect the entire OE for RNA-seq analysis, it is possible that there is zonal heterogeneity in the samples we analyzed. To reduce this potential bias, we pooled dissociated OE cells from 3 mice for FACS sorting. Confidence that the 78 Olfr genes are regulated by Upf3b comes from the reproducibility of the regulation in independent samples (Figure 1C) and validation by qPCR (Figure 1—figure supplement 2B). Furthermore, our single-cell RNA-seq analysis (which analyzed samples different from those analyzed by RNA-seq) verified the regulation of these 78 Olfr genes (Figure 6H).

How might Upf3b influence the selection of this particular set of Olfr genes? Given that UPF3B is a NMD factor, it could promote the decay of an mRNA encoding a repressor that acts to regulate the selection of these 78 Olfr genes for dominant expression. To test this model, we screened genes exhibiting significantly upregulated expression in Upf3b-null OSNs for those that encode factors known to regulate Olfr gene expression or have binding sites in Olfr promoters (Clowney et al., 2011; Dalton et al., 2013; Hirota and Mombaerts, 2004; Markenscoff-Papadimitriou et al., 2014; McIntyre et al., 2008; Michaloski et al., 2006; Wang et al., 1997). This screen identified two genes—Mafg and Irf8—that fulfilled this criteria. Both encode transcriptional repressors (Igarashi et al., 1994; Salem et al., 2014) that bind O/E consensus sites found in Olfr gene promoters (Michaloski et al., 2006). Thus, Mafg and Irf8 are candidates to act directly downstream of NMD in a regulatory circuit that suppresses the transcription of these 78 Olfr genes. Mafg is a member of the Maf subfamily of basic leucine-zipper transcription factor genes that encode small proteins containing a B-ZAP DNA-binding domain, but lack a transactivation domain, and thus members of this family dimerize to form transcriptional repressors (Igarashi et al., 1994). MAFG is best known for its ability to regulate globin transcription in erythroid cells; our results raise the possibility that MAFG also functions in OSNs to regulates Olfr genes. IRF8 regulates the development hematopoietic cells; its expression in OSNs raises the possibility that this transcription factor also functions in OSNs.

Our findings support a model in which IRF8 and MAFG normally subtly repress the transcription of a subset of Olfr genes in OSNs to fine-tune their expression. Our evidence suggests that IRF8 and MAFG are encoded by NMD target mRNAs, so when NMD is disrupted, these transcriptional repressors are overexpressed, leading to reduced expression of their Olfr gene targets in developing OSNs. Thus, NMD deficiency would be expected to reduce the probability that these particular Olfr genes will be chosen to be the ‘dominant Olfr gene’ in individual mOSNs, which is precisely what we observed in Upf3b-null mice.

A non-mutually exclusive possibility is that Upf3b dictates the selection of Olfr genes by influencing OSN development. In support, several of the genes we found were regulated by Upf3b have been reported to play essential roles in neurogenesis, including Lrp2, Hk2, Notch2, Gdf11, Fos, Ptch1, Spry2, and Cwc22. Upf3b could also indirectly influence the Olfr repertoire by differentially affecting the survival of OSNs harboring different OLFRs. In support, we found that Upf3b loss upregulates Fos, which is associated with OSN apoptosis (Michel et al., 1994).

The other major class of genes regulated by Upf3b in OSNs is antimicrobial genes. This finding, coupled with our finding that OSNs constitutively express these anti-microbial genes (albeit at low levels), suggests that OSNs function not only in olfaction but also in defense against microbes in the bronchial airways. In support, a recent study showed that inflammation causes OSNs to switch from a role in olfaction to immune defense (Chen et al., 2019). This raises the interesting possibility that loss of Upf3b triggers OE inflammation, which, in turn, diverts OSNs from functioning in olfaction to immune defense, thereby causing deficient olfaction. In support, another recent study reported that NMD disruption causes neuro-inflammation in the central nervous system (Johnson et al., 2019). In particular, this study found that Upf2 conditional knockout in the murine forebrain leads to immune infiltration, coupled with deficits in memory, synaptic plasticity, social, and vocal communication (Johnson et al., 2019). Importantly, they found that anti-inflammatory agents partially rescued many of these deficits, indicating that the inflammation is at least partially responsible for the neural defects in these Upf2-conditional knockout mice. It will be intriguing to determine whether humans with UPF3B mutations also suffer from neuro-inflammation and whether this is responsible for their intellectual disability.

Our finding that loss of UPF3B upregulates a very large number of immune-related genes in OSNs, over half of which encode mRNAs that have NIFs and thus may be direct NMD targets (Supplementary file 5), raises the possibility that this ‘immune induction’ response to NMD inhibition is physiologically important. In this regard, it is notable that some viruses have been shown to inhibit NMD, and, in turn, NMD can inhibit viral infection (Wachter and Hartmann, 2014; Wada et al., 2018). Coupled with our data, these findings raise the intriguing possibility that the reason that OSNs express high levels of antimicrobial genes in response to NMD inhibition is because this provides a means to cope with infectious agents, particularly those that inhibit NMD as a means to avoid the antiviral actions of NMD.

In addition to NMD inhibition directly upregulating mRNAs encoding immune factors in OSNs, we identified candidate intermediary factors that may act in a circuit to achieve the same aim. In particular, we identified three mRNAs—Notch2, Bhlhe40, and Rac2—which are high-confidence NMD targets in mOSNs (Table 1) that encode factors previously shown to regulate the expression of many genes encoding inflammatory mediators and antimicrobial proteins (Dooley et al., 2009; Jarjour et al., 2019; Shang et al., 2016).

Our scRNA-seq analysis indicated that Upf3b impacts the steady-state frequency of specific OSN precursor and OSN cell subsets. We found that Upf3b-null mice have decreased numbers of HBCs, suggesting that UPF3B promotes the maintenance of these reserve stem cells. This effect appeared to be specific, as we observed no significant effects on GBCs, which also serve as olfactory stem cells, but unlike HBCs, function to generate new mOSNs constitutively (Schwob et al., 2017). We also observed that Upf3b-null mice acquired a specific group of mOSNs harboring a unique transcriptome that are hardly present in WT mice. This mOSN-4 sub-cluster is enriched for many genes, such as Tuba1a, Nsg1, Chchd10, Eml2, Ubb, and Gldc, which suggests that Upf3b normally represses these genes. It remains to be determined whether the aberrant over-expression of these genes causes aberrant mOSN function. We also found that Upf3b-null mice largely lack a mOSN sub-cluster—mOSN-2—that we found contained large numbers of cells in most WT mice. The mOSN-2 sub-cluster is likely to be functional, as genes enriched in this sub-cluster include Pten, App, Cnga2, Nrp2, Ncam1, Adcy3, Gnal, Atf5, and Gfy (Supplementary file 4), all of which are known to be essential for olfactory epithelium development and/or olfaction. This reciprocal shift in these two mOSN sub-clusters in Upf3b-null mice raises the possibility that UPF3B loss converts the mOSN-2 sub-cluster into the mOSN-4 sub-cluster. This remains to be determined, as does the physiological consequences of these shifts in mOSN sub-populations. Another important area for future investigation is to determine whether these cell-subset alterations in Upf3b-null mice are cell autonomous or non-cell autonomous.

As described in the Introduction, few direct NMD target RNAs have previously been defined in vivo. Our study fills this gap by identifying high-confidence NMD target mRNAs in mOSNs in vivo. Many of the NMD targets we identified in mOSNs have long 3’UTRs, raising the possibility that mOSNs have a predilection for degrading mRNAs with this particular NIF. By analogy, evidence suggests that mRNAs harboring long 3’UTRs are also preferentially targeted for destruction by NMD in male germ cells (Bao et al., 2016). Several of the NMD target mRNAs that we identified in mOSNs are good candidates to have roles in OSN development. For example, Gdf11 functions in negative-feedback control of OE neurogenesis; Lrp2 promotes the proliferation of neural precursor cells in the subependymal zone of the olfactory bulb; and Notch2 is required for maintaining sustentacular cell function in the OE (Gajera et al., 2010; Kawauchi et al., 2009; Rodriguez et al., 2008). Other NMD target mRNAs that we identified, including Ptch1 and Hk2, encode proteins known to be important for the development of neurons outside of the olfactory system (Iulianella and Stanton-Turcotte, 2019; Zheng et al., 2016).

Our study provides a useful resource for the olfactory field. For example, our scRNA-seq analysis identified putative new OSN precursor and OSN cell subsets. While we do not know the significance of this heterogeneity, the genes differentially expressed by the sub-clusters we identified suggests functional relevance. For example, the genes differentially expressed by the 4 cell-sub clusters we identified for both GBCs and iOSNs suggested that these sub-clusters represent distinct developmental stages. Our results are consistent with Fletcher et al., who demonstrated that that the 1 GBC and 4 INP/iOSN sub-clusters they identified follow a linear developmental pattern (Fletcher et al., 2017). Our genome-wide determination of mOSN mRNA expression levels and ribosome occupancy (i.e. translation rates) will be useful for future studies to determine how transcription, translation, and other post-transcriptional processes coordinate to regulate the expression of large sets of genes in mature neurons in vivo. We divided mOSN-expressed mRNAs into nine categories based on steady-state mRNA level and ribosome occupancy, allowing dissection of common functions encoded by similarly regulated mRNAs. Given that translation is a highly energy-consuming process (Lynch and Marinov, 2015), it is likely that there has been strong selection pressure for many mRNAs to be translated inefficiently. Indeed, we found that modestly translated mRNAs encode many key mOSN proteins, including receptors, signaling factors, and developmental regulators.

In conclusion, our study provides an invaluable set of resources for the olfactory field and identifies a post-transcriptional regulatory pathway that impacts OSNs.

Materials and methods

Key resources table.

Reagent type (species) or resource	Designation	Source or reference	Identifiers	Additional information
Gene (Mus musculus)	Upf3b	GenBank	Gene ID: 68134
Genetic reagent (Mus. musculus)	C57BL/6J	Jackson Laboratory	Stock #: 000664 RRID:MGI:3028467
Genetic reagent (Mus. musculus)	Upf3b-null mice	PMID:21925383	RRID:MGI:6110148	Miles Wilkinson lab
Genetic reagent (Mus. musculus)	R26-eYFP mice	PMID:11299042		Obtained from Dr. Maike Sander (UCSD)
Genetic reagent (Mus. musculus)	Omp-Cre mice	PMID:22057188		Obtained from Dr. Haiqing Zhao (Johns Hopkins University)
Genetic reagent (Mus. musculus)	RiboTag mice	PMID:19666516		Obtained from Dr. Paul Ameiux (University of Washington)
Antibody	Rabbit monoclonal anti-OMP (EPR19190)	Abcam	Cat# ab183947 RRID:AB_2858281	IF (1:400), WB (1:2000)
Antibody	Goat polyclonal anti-OMP	FUJIFILM Wako Chemicals	Cat# 544–10001-WAKO RRID:AB_2315007	IF (1:200)
Antibody	Rabbit polyclonal anti-CAMP	Generated by Richard L. Gallo laboratory	PMID:11442754	IF (1:200)
Antibody	Rabbit polyclonal anti-FUT10	Proteintech	Cat#: 18660–1-AP RRID:AB_10641997	IF (1:200)
Antibody	Donkey anti-Goat IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor 488	Thermo Fisher Scientific	Cat#: A-11055 RRID:AB_2534102	IF (1:1000)
Antibody	Donkey anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 555	Thermo Fisher Scientific	Cat#: A-31572 RRID:AB_162543	IF (1:1000)
Sequence-based reagent	Fosl2_F	This paper	PCR primers	CCGCAGAAGGAGAGATGAG (from IDT)
Sequenced-based reagent	Fosl2_R	This paper	PCR primers	GCAGCTTCTCTGTCAGCTC (from IDT)
Sequence-based reagent	Ptger2_F	This paper	PCR primers	TGCTCCTTGCCTTTCACAATC (from IDT)
Sequenced-based reagent	Ptger2_R	This paper	PCR primers	CCTAAGTATGGCAAAGACCCAAG (from IDT)
Sequence-based reagent	Adcy6_F	This paper	PCR primers	TTCCTGACCGTGCCTTCTC (from IDT)
Sequenced-based reagent	Adcy6_R	This paper	PCR primers	CACCCCGGTTGTCTTTGC (from IDT)
Sequence-based reagent	Ptch1_F	This paper	PCR primers	ACCTCCTAGGTAAGCCTCC (from IDT)
Sequenced-based reagent	Ptch1_R	This paper	PCR primers	CACCCACAATCAACTCCTCC (from IDT)
Sequence-based reagent	Cwc22_F	This paper	PCR primers	CAGAAGACAGATACACAGAGCAAG (from IDT)
Sequenced-based reagent	Cwc22_R	This paper	PCR primers	CTCTCTCTCTCTCTCTGCGTTT (from IDT)
Sequence-based reagent	Fut10_F	This paper	PCR primers	CCAGGGCCTTCCTATTCTACG (from IDT)
Sequenced-based reagent	Fut10_R	This paper	PCR primers	CTGAATGTGGCCGTATGGTTG (from IDT)
Sequence-based reagent	Gdpd3_F	This paper	PCR primers	TGATCCGACACTTGCAGGAC (from IDT)
Sequenced-based reagent	Gdpd3_R	This paper	PCR primers	GCTGTGGGGTAATCGGTCAT (from IDT)
Sequence-based reagent	Olfr827_F	This paper	PCR primers	TGGGATGGTTCTTCTGGGAA (from IDT)
Sequenced-based reagent	Olfr827_R	This paper	PCR primers	ACCGTGGAGTAGGAGAGGTC (from IDT)
Sequence-based reagent	Rpl19_F	This paper	PCR primers	CCTGAAGGTCAAAGGGAATGTG (from IDT)
Sequenced-based reagent	Rpl19_R	This paper	PCR primers	CTTTCGTGCTTCCTTGGTCTT (from IDT)
Commercial assay or kit	Chromium Single Cell 3' Library and Gel Bead Kit	10X Genomics	Cat# 120237
Commercial assay or kit	iScript cDNA synthesis Kit	BioRad	Cat# 170–8891
Commercial assay or kit	SsoAdvanceD Universal SYBR Green Supermix	BioRad	Cat# 172–5274
Commercial assay or kit	RNeasy Mini Kit	Qiagen	Cat# 74104
Software, algorithm	Cell Ranger Version 2.1.1	10x genomics	Cell Ranger Version 2.1.1
Software, algorithm	Seurat (v3.1.5)	Designed by Rahul Satija laboratory	PMID:31178118
Software, algorithm	Monocle (v2.16.0)	Designed by Cole Trapnell laboratory	PMID:28114287
Software, algorithm	NIH ImageJ (v1.8.0)	NIH	Version 1.8.0

Open in a new tab

Mice

This study was carried out in strict accordance with the Guidelines of the Institutional Animal Care and Use Committee (IACUC) at the University of California, San Diego. The protocol was approved by the IACUC at the University of California, San Diego (permit number: S09160). All studies were conducted on adult male mice housed under a 12 hr light:12 hr dark cycle and provided with food and water ad libitum. Of note, we only performed analyses on male mice. Since Upf3b is X-linked gene, we analyzed Upf3b^+/y (WT) and Upf3^-/y (KO) mice. All mouse strains used for analysis were backcrossed to C57BL/6J for at least eight passages.

Behavioral and weight analyses

To assess the effect of UPF3B loss on mouse weight, 19male pups (nine Upf3b-null and ten WT mice) from Upf3b^+/- × WT breeders (6 litters) were assessed, performed as described previously (Tan et al., 2016). For pre-weaning pups, to reduce stress, forceps and gloves were changed frequently between cages.

For the coyote/bobcat urine experiment, 10 male mice (10- to 16 weeks of age) from each genotype were analyzed. Each mouse was placed into a cage for 10 min to acclimatize, a strip of filter paper soaked with coyote urine (Snow Joe) or bobcat urine (Predator Pee) was placed into the cage for 5 min, and the amount of time the mouse was in the vicinity of the filter paper was determined by video recording. Each mouse was tested separately in the absence of humans or other mice in the room.

RNA-seq analysis

For each mOSN sample analyzed, 3 C57BL/6J male mice (8- to 9-weeks old) were pooled. Four replicate samples were analyzed per genotype (Upf3b^+/y; Omp-Cre; R26-eYFP and Upf3b^-/y; Omp-Cre; R26-eYFP). Cell sorting experiments were performed on two separate days, with two samples sorted per day. The OE was dissected as described (Gong, 2012) and dissociated using the Papain Dissociation System (Worthington) at 37°C for 15 min, followed by extensive trituration. Cells were filtered using a 40‐μm strainer (Falcon). After spinning at 200 g for 5 min, cells were resuspended in Hanks’ balanced salt solution (HBSS) containing 3% FBS (Gibco) but without Ca²⁺ and Mg²⁺. The cell suspension was mixed with propidium iodide (final concentration of 1 μg/ml) and the OMP-eYFP⁺ cells were sorted by flow-cytometry. RNA was isolated from the OMP-eYFP⁺ cells using TriZOL (Life Technologies), followed by a secondary purification step using a RNeasy column (Qiagen). Total RNA was assessed for quality using an Agilent Bioanalyzer, and samples determined to have an RNA Integrity Number (RIN) of at least 8 or greater were used to generate RNA libraries using Illumina's TruSeq RNA Sample Prep Kit, following the manufacturer's specifications, with the RNA fragmentation time adjusted to 5 min. RNA-seq was performed at the Institute of Genomic Medicine at UCSD. RNA libraries were multiplexed and sequenced with 100 base pair (bp) pair end reads on an Illumina HiSeq4000. The average number of reads per sample ranged from approximately 15 to 22 million reads. Reads were filtered for quality and aligned with STAR (2.5.2b) against Mus musculus release-90, Ensembl genome (GRCm38). The exon counts were aggregated for each gene to build a read count table using SubRead function featureCounts (Liao et al., 2014). Using the exon start/end positions, we extracted the exon sequences from the mm10 mouse genome, and ligated them together in silico for each transcript. For each entry, the entire transcript sequence was subtracted from the known CDS sequence (obtained as above) to identify 3’UTR length. DEGs were defined using DESeq2 (Love et al., 2014) using a threshold q-val of <0.05. The R package program ‘pheatmap’ was used for clustering and to generate heatmap plots. GO analysis was done using database for annotation, visualization and integrated discovery (DAVID), version v6.8. To infer relative RNA stability, we used the REMBRANDTS program (Alkallas et al., 2017) following the tutorial (https://github.com/csglab/REMBRANDTS).

RiboTag analysis

For each mOSN sample analyzed, three C57BL/6J male mice (8 to 9-weeks old) were pooled. Three replicate samples were analyzed per genotype (Upf3b^+/y; Omp-Cre; RiboTag and Upf3b^-/y; Omp-Cre; RiboTag). The OEs was dissected as described (Gong, 2012), homogenized, washed with HBSS, centrifuged at 16,000 g at 4°C for 10 min, the supernatant was transferred into a new tube and incubated with HA antisera (#16B12; BioLegend, CA) at 4°C for 2.5 hr. Ribosome-bound RNAs were captured on anti-HA agarose beads (Pierce) for 1 hr at 4°C on a tube rotator. RNA libraries were multiplexed and sequenced with 50 bp single-end reads on an Illumina HiSeq4000. RNA sequencing, alignment, and downstream analyses were done as described above for RNA-seq analysis. TE was determined by dividing RiboTag reads by RNA-seq reads. Log2-transformed transcripts per million (TPM) values were used to segregate mRNAs into different categories.

scRNA-seq analysis

Four C57BL/6J male mice (7 to 8-weeks old) per genotype (Upf3b^+/y and Upf3b^-/y) were used to obtain OE for scRNA-seq analysis. After dissecting the OE as described (Gong, 2012), the cells were dissociated following the 10X Genomics Chromium sample preparation protocol. Briefly, tissue was cut into 1 mm³ pieces and digested in HBSS without Ca²⁺ and Mg²⁺ and supplemented with 44 U/ml Dispase (Invitrogen), 1000 U/ml Collagenase type II (Invitrogen) and 10 mg/ml DNaseI (Sigma), for 20 min at 37°C with gentle agitation. The digested tissue was centrifuged at 300 rcf for 5 min and washed in HBSS without Ca²⁺ and Mg²⁺. Dissociated cells were resuspended in 3% FBS in PBS. Dead cells were removed using the ClioCell Dead Cell Removal kit (Amsbio) following the manufacturer's instructions. Single cells were resuspended in 0.04% BSA in PBS (w/v) and loaded on the 10x Chromium chip. Cell capturing, and library preparation was carried as per kit instructions (Chromium Single Cell Kit [v2 chemistry]). The resultant libraries were size selected, pooled, and sequenced using 2 × 100 paired-end sequencing protocol on an Illumina HiSeq 4000 instrument. The libraries initially underwent shallow sequencing to access quality and to adjust subsequent sequencing depth based on the capture rate and unique molecular indices (UMI) detected. All sequencing analyses were performed at the Institute of Genomic Medicine at UCSD.

As described previously (Sohni et al., 2019; Tan et al., 2020b; Tan et al., 2020a), de-multiplexed raw sequencing reads were processed and mapped to the mouse genome (mm10) using Cell Ranger software (v2.0) with default parameters. We filtered raw count matrices by excluding cells expressing less than 200 detectably expressed genes and genes expressed in less than 3 cells. Each library was tagged with a library batch ID and combined across independent experiments using the Seurat package (Butler et al., 2018) in R. To check the quality of the single-cell data and to remove multiplets, we performed Seurat-based filtering of cells based on three criteria: number of detected features (nFeature_RNA) per cell, number of UMIs expressed per cell (nCount_RNA) and mitochondrial content, using the following threshold parameters: nFeature_RNA (>500), nCount_RNA (>1,500), and percentage of mitochondrial genes expressed (<0.2%). We used known lineage marker profiles to exclude cell multiplets (cells expressing different lineage markers) and cell-free droplets. Gene expression values were log normalized and regressed by mitochondrial expression (‘percent.mt’) and cell cycle gene expession (‘S.Score’ and ‘G2M.Score’) using the SCTransform function. Batch correction was performed using the JackStraw functions in the Seurat package.

To identify cell clusters, we employed the UMAP algorithm (Becht et al., 2019). The FindMarkers function (a Wilcoxon rank sum test) was used to determine differential gene expression between clusters (set at minimum expression in 25% of cells). The DoHeatmap function was used to generate an expression heatmap for specific cells and features. GO analysis (DAVID v6.8) was done using the top differentially (positively) expressed genes, with a p-adjusted cut off of 0.01.

Single-cell pseudotime trajectories were constructed with the Monocle two package (v2.10.1) (Qiu et al., 2017) according to the provided documentation (http://cole-trapnell-lab.github.io/monocle-release/). UMI counts were modeled as a negative binomial distribution. The ordering genes were identified as having high dispersion across cells (mean_expression >= 0.01; dispersion_empirical >= 1). The discriminative dimensionality reduction with trees (DDRTree) method was used to reduce data to two dimensions. Differentially expressed genes were identified and used for dynamic trajectory analysis (NO discovery rate [FDR]<0.01) to order cells in pseudotime. The plot_pseudotime_heatmap function was used to generate heatmaps.

NIF analysis

To define NIFs, Refseq-defined transcripts were first converted into Ensemble transcript IDs and their sequences were obtained using the UCSC Table Browser. NIFs were identified in these transcripts using an algorithm written in Python 2.7, Zenith.py, created by the Wilkinson laboratory. Only transcripts with a detectable 5’UTR and 3’UTR were considered. A transcript was defined as harboring a dEJ if it contained at least one exon-exon junction ≥50 nt downstream of the stop codon terminating the main ORF. A transcript was defined as harboring an uORF if the following criteria were met: (i) the ORF is in the 5’ UTR, (ii) the start codon and surrounding nts are in a context known to initiate translation (a purine at the −3 position or a guanine at the +4 position, relative to the A in the AUG initiation codon [+1]) (Kozak, 1986), (iii) the ORF is ≥30 nt long, and (iv) the ORF does not overlap with the main ORF (to reduce the probability that translation could be re-initiated, thereby allowing the transcript to escape NMD).

Immunofluorescence analysis

Adult mice were anesthetized and perfused with 4% paraformaldehyde (PFA; Sigma). OE was dissected and fixed in 4% PFA at 4°C for 24 hr, then transferred to 70% ethanol. After embedding in paraffin, 5 µM sections were prepared, deparaffinized 2 times in xylene, followed by serial dilutions of ethanol. Unmasking was performed with IHC-TekTM epitope retrieval solution using a steamer (IHCWORLD) for 40 min. Blocking was performed by incubating with 5% serum (from the species that the secondary antibody was raised in) for 1 hr at room temperature. The sections were then incubated overnight with the primary antibody (goat polyclonal OMP, rabbit anti-CAMP [Gallo et al., 1997]) at 4°C and incubated with secondary antibody (Donkey anti-Goat IgG [H+L] conjugated with Alexa Fluor 488 or Donkey anti-Rabbit IgG [H+L] conjugated with Alexa Fluor 555) for 1 hr at room temperature. The nuclei were counterstained with DAPI, and a coverslip was placed over the sections with mounting medium.

Western blot analysis

OEs were incubated in radioimmunoprecipitation assay (RIPA) buffer (Bio-Rad) supplemented with protease inhibitor cocktail (Sigma) on ice for 30 min, followed by centrifugation at 16,000 g for 15 min at 4°C. The lysates were then transferred to new tubes, and protein level was quantified using the DC Protein Assay kit (Bio‐Rad). Twenty micrograms of the protein samples were separated on an 15% polyacrylamide gel, and Western blot analysis was performed as previously described (Ramaiah et al., 2019). Quantification of the blots was performed using NIH ImageJ (1.8.0).

Acknowledgements

We thank the UCSD Institute for Genomic Medicine for technical support and the San Diego Supercomputer Center for providing data analysis resources. We are grateful to Drs. Maike Sander (UCSD), Haiqing Zhao (Johns Hopkins University), and Paul Ameiux (University of Washington) for providing R26-eYFP, Omp-Cre, and RiboTag mice, respectively.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Miles F Wilkinson, Email: mfwilkinson@health.ucsd.edu.

Didier YR Stainier, Max Planck Institute for Heart and Lung Research, Germany.

Funding Information

This paper was supported by the following grant:

Eunice Kennedy Shriver National Institute of Child Health and Human Development R01 HD093846 to Miles F Wilkinson.

Additional information

Competing interests

No competing interests declared.

Author contributions

Resources, Data curation, Formal analysis, Validation, Writing - original draft.

Conceptualization, Data curation.

Formal analysis.

Conceptualization.

Data curation.

Supervision.

Conceptualization, Supervision, Funding acquisition, Project administration, Writing - review and editing.

Ethics

Animal experimentation: This study was carried out in strict accordance with the Guidelines of the Institutional Animal Care and Use Committee (IACUC) at the University of California, San Diego. The protocol was approved by the IACUC at the University of California, San Diego (permit number: S09160).

Additional files

Supplementary file 1. Quality control matrices of Upf3b-null and control mOSNs RNA-seq datasets.

(1) QC metrics, (2) reads count, and (3) TPM values from Upf3b-null (KO) and control (WT) mOSN RNA-seq and RiboTag datasets. (4) The expression of all annotated Olfr genes in Upf3b-null (KO) and control (WT) mOSN samples, based on RNA-seq analysis.

elife-57525-supp1.xlsx^{(26.4MB, xlsx)}

Supplementary file 2. mOSN transcriptome, translome, and NMD targets.

(1) Genes differentially expressed between Upf3b-null and WT mOSNs (as defined by RNA-seq analysis), including their known NIFs. (2) High-confidence NMD targets: transcripts both upregulated and stabilized in Upf3b-null mOSNs. (3) The transcriptome and translome of Upf3b-null and WT mOSNs, based on RNA-seq and RiboTag analyses. (4) Translational efficiency (TE) of RNAs expressed in Upf3b-null and WT mOSNs, as defined by RNA-seq and RiboTag analyses.

elife-57525-supp2.xlsx^{(4MB, xlsx)}

Supplementary file 3. Putative mouse NMD target RNAs defined by previous studies.

elife-57525-supp3.xlsx^{(177.2KB, xlsx)}

Supplementary file 4. Genes exhibiting enriched expression in scRNA-seq-defined OE cell subsets.

(1) Genes exhibiting enriched expression in all OE cell types identified by scRNA-seq analysis. (2) Genes exhibiting enriched expression in HBC, GBC, iOSN, and mOSN sub-clusters identfied by scRNA-seq analysis. (3) Genes exhibiting enriched expression in the 4 temporally distinct gene groups defined by pseudotime trajectory analysis of HBCs, GBCs, iOSNs, and mOSNs. (4) Transcription factor (TF) genes exhibiting enriched expression in the 4 temporally distinct gene groups defined by pseudotime trajectory analysis of HBCs, GBCs, iOSNs, and mOSNs.

elife-57525-supp4.xlsx^{(2.2MB, xlsx)}

Supplementary file 5. Immune genes expressed in HBCs, GBCs, iOSNs, and mOSNs.

(1) Immune genes enriched in Upf3b-null HBC, GBC, iOSN, and mOSN sub-clusters identified by scRNA-seq analysis. (2) NIFs present in immune genes expressed in Upf3b-null OSNs.

elife-57525-supp5.xlsx^{(34.8KB, xlsx)}

Transparent reporting form

elife-57525-transrepform.docx^{(245.8KB, docx)}

Data availability

Sequencing data have been deposited in GEO under accession code GSE146043.

The following dataset was generated:

Tan K, Jones SH, Wilkinson MF. 2020. The role of the NMD factor UPF3B in olfactory sensory neurons. NCBI Gene Expression Omnibus. GSE146043

The following previously published datasets were used:

Hurt JA, Robertson AD, Burge CB. 2013. Global Analyses of UPF1 Binding and Function Reveal Expanded Scope of Nonsense-Mediated mRNA Decay. NCBI Gene Expression Omnibus. GSE41785

Mao H, McMahon JJ, Tsai YH, Wang Z, Silver DL. 2016. Haploinsufficiency for Core Exon Junction Complex Components Disrupts Embryonic Neurogenesis and Causes p53-Mediated Microcephaly. NCBI Gene Expression Omnibus. GSE85576

Mooney CM, Jimenez-Mateos EM, Engel T, Mooney C, Diviney M, Venø MT, Kjems J, Farrell MA, O'Brien DF, Delanty N, Henshall DC. 2017. RNA sequencing of synaptic and cytoplasmic Upf1-bound transcripts supports contribution of nonsense-mediated decay to epileptogenesis. RNA sequencing of synaptic and cytoplasmic Upf1-bound transcripts supports contribution of nonsense-mediated decay to epileptogenesis.

Bao j, Vitting-Seerup K, Waage J, Tang C, Ge Y, Porse BT, Yan W. 2016. UPF2-Dependent Nonsense-Mediated mRNA Decay Pathway Is Essential for Spermatogenesis by Selectively Eliminating Longer 3'UTR Transcripts. NCBI Gene Expression Omnibus. GSE55180

McIlwain DR, Pan Q, Reilly PT, Elia AJ, McCracken S, Wakeham AC, Itie-Youten A, Blencowe BJ, Mak TW. 2010. Smg1 is required for embryogenesis and regulates diverse genes via alternative splicing coupled to nonsense-mediated mRNA decay. Smg1 is required for embryogenesis and regulates diverse genes via alternative splicing coupled to nonsense-mediated mRNA decay.

Weischenfeldt J, Damgaard I, Bryder D, Theilgaard-Mönch K. Thoren LA, Nielsen FC, Jacobsen SEW, Nerlov C, Porse BT. 2008. NMD Is Essential for Hematopoietic Stem and Progenitor Cells and for Eliminating By-Products of Programmed DNA Rearrangements. NMD Is Essential for Hematopoietic Stem and Progenitor Cells and for Eliminating By-Products of Programmed DNA Rearrangements.

Thoren LA, Nørgaard GA, Weischenfeldt J, Waage J, Jakobsen JS, Damgaard I, Bergström FC, Blom AM, Borup R, Bisgaard HC, Porse BT. 2010. UPF2 Is a Critical Regulator of Liver Development, Function and Regeneration. UPF2 Is a Critical Regulator of Liver Development, Function and Regeneration.

References

Alkallas R, Fish L, Goodarzi H, Najafabadi HS. Inference of RNA decay rate from transcriptional profiling highlights the regulatory programs of Alzheimer's disease. Nature Communications. 2017;8:909. doi: 10.1038/s41467-017-00867-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
Alrahbeni T, Sartor F, Anderson J, Miedzybrodzka Z, McCaig C, Müller B. Full UPF3B function is critical for neuronal differentiation of neural stem cells. Molecular Brain. 2015;8:33. doi: 10.1186/s13041-015-0122-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bao J, Vitting-Seerup K, Waage J, Tang C, Ge Y, Porse BT, Yan W. UPF2-Dependent Nonsense-Mediated mRNA decay pathway is essential for spermatogenesis by selectively eliminating longer 3'UTR Transcripts. PLOS Genetics. 2016;12:e1005863. doi: 10.1371/journal.pgen.1005863. [DOI] [PMC free article] [PubMed] [Google Scholar]
Barrett LW, Fletcher S, Wilton SD. Regulation of eukaryotic gene expression by the untranslated gene regions and other non-coding elements. Cellular and Molecular Life Sciences. 2012;69:3613–3634. doi: 10.1007/s00018-012-0990-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bear DM, Lassance JM, Hoekstra HE, Datta SR. The evolving neural and genetic architecture of Vertebrate Olfaction. Current Biology. 2016;26:R1039–R1049. doi: 10.1016/j.cub.2016.09.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
Becht E, McInnes L, Healy J, Dutertre C-A, Kwok IWH, Ng LG, Ginhoux F, Newell EW. Dimensionality reduction for visualizing single-cell data using UMAP. Nature Biotechnology. 2019;37:38–44. doi: 10.1038/nbt.4314. [DOI] [PubMed] [Google Scholar]
Belgrader P, Cheng J, Maquat LE. Evidence to implicate translation by ribosomes in the mechanism by which nonsense codons reduce the nuclear level of human triosephosphate isomerase mRNA. PNAS. 1993;90:482–486. doi: 10.1073/pnas.90.2.482. [DOI] [PMC free article] [PubMed] [Google Scholar]
Boehm V, Haberman N, Ottens F, Ule J, Gehring NH. 3' UTR length and messenger ribonucleoprotein composition determine endocleavage efficiencies at termination codons. Cell Reports. 2014;9:555–568. doi: 10.1016/j.celrep.2014.09.012. [DOI] [PubMed] [Google Scholar]
Bühler M, Steiner S, Mohn F, Paillusson A, Mühlemann O. EJC-independent degradation of nonsense immunoglobulin-mu mRNA depends on 3' UTR length. Nature Structural & Molecular Biology. 2006;13:462–464. doi: 10.1038/nsmb1081. [DOI] [PubMed] [Google Scholar]
Butler A, Hoffman P, Smibert P, Papalexi E, Satija R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nature Biotechnology. 2018;36:411–420. doi: 10.1038/nbt.4096. [DOI] [PMC free article] [PubMed] [Google Scholar]
Carter MS, Doskow J, Morris P, Li S, Nhim RP, Sandstedt S, Wilkinson MF. A regulatory mechanism that detects premature nonsense codons in T-cell receptor transcripts in vivo is reversed by protein synthesis inhibitors in vitro. Journal of Biological Chemistry. 1995;270:28995–29003. doi: 10.1074/jbc.270.48.28995. [DOI] [PubMed] [Google Scholar]
Carter MS, Li S, Wilkinson MF. A splicing-dependent regulatory mechanism that detects translation signals. The EMBO Journal. 1996;15:5965–5975. doi: 10.1002/j.1460-2075.1996.tb00983.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cau E, Gradwohl G, Fode C, Guillemot F. Mash1 activates a cascade of bHLH regulators in olfactory neuron progenitors. Development. 1997;124:1611–1621. doi: 10.1242/dev.124.8.1611. [DOI] [PubMed] [Google Scholar]
Chan WK, Huang L, Gudikote JP, Chang YF, Imam JS, MacLean JA, Wilkinson MF. An alternative branch of the nonsense-mediated decay pathway. The EMBO Journal. 2007;26:1820–1830. doi: 10.1038/sj.emboj.7601628. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chang YF, Imam JS, Wilkinson MF. The nonsense-mediated decay RNA surveillance pathway. Annual Review of Biochemistry. 2007;76:51–74. doi: 10.1146/annurev.biochem.76.050106.093909. [DOI] [PubMed] [Google Scholar]
Chen M, Reed RR, Lane AP. Chronic inflammation directs an olfactory stem cell functional switch from neuroregeneration to immune defense. Cell Stem Cell. 2019;25:501–513. doi: 10.1016/j.stem.2019.08.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chess A, Simon I, Cedar H, Axel R. Allelic inactivation regulates olfactory receptor gene expression. Cell. 1994;78:823–834. doi: 10.1016/S0092-8674(94)90562-2. [DOI] [PubMed] [Google Scholar]
Clowney EJ, Magklara A, Colquitt BM, Pathak N, Lane RP, Lomvardas S. High-throughput mapping of the promoters of the mouse olfactory receptor genes reveals a new type of mammalian promoter and provides insight into olfactory receptor gene regulation. Genome Research. 2011;21:1249–1259. doi: 10.1101/gr.120162.110. [DOI] [PMC free article] [PubMed] [Google Scholar]
Colak D, Ji SJ, Porse BT, Jaffrey SR. Regulation of axon guidance by compartmentalized nonsense-mediated mRNA decay. Cell. 2013;153:1252–1265. doi: 10.1016/j.cell.2013.04.056. [DOI] [PMC free article] [PubMed] [Google Scholar]
Conti E, Izaurralde E. Nonsense-mediated mRNA decay: molecular insights and mechanistic variations across species. Current Opinion in Cell Biology. 2005;17:316–325. doi: 10.1016/j.ceb.2005.04.005. [DOI] [PubMed] [Google Scholar]
Dalton RP, Lyons DB, Lomvardas S. Co-opting the unfolded protein response to elicit olfactory receptor feedback. Cell. 2013;155:321–332. doi: 10.1016/j.cell.2013.09.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dooley JL, Abdel-Latif D, St Laurent CD, Puttagunta L, Befus D, Lacy P. Regulation of inflammation by Rac2 in immune complex-mediated acute lung injury. American Journal of Physiology-Lung Cellular and Molecular Physiology. 2009;297:L1091–L1102. doi: 10.1152/ajplung.90471.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dostie J, Dreyfuss G. Translation is required to remove Y14 from mRNAs in the cytoplasm. Current Biology. 2002;12:1060–1067. doi: 10.1016/S0960-9822(02)00902-8. [DOI] [PubMed] [Google Scholar]
Doty RL. Olfactory dysfunction in parkinson disease. Nature Reviews Neurology. 2012;8:329–339. doi: 10.1038/nrneurol.2012.80. [DOI] [PubMed] [Google Scholar]
Fletcher RB, Das D, Gadye L, Street KN, Baudhuin A, Wagner A, Cole MB, Flores Q, Choi YG, Yosef N, Purdom E, Dudoit S, Risso D, Ngai J. Deconstructing olfactory stem cell trajectories at Single-Cell resolution. Cell Stem Cell. 2017;20:817–830. doi: 10.1016/j.stem.2017.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gajera CR, Emich H, Lioubinski O, Christ A, Beckervordersandforth-Bonk R, Yoshikawa K, Bachmann S, Christensen EI, Götz M, Kempermann G, Peterson AS, Willnow TE, Hammes A. LRP2 in ependymal cells regulates BMP signaling in the adult neurogenic niche. Journal of Cell Science. 2010;123:1922–1930. doi: 10.1242/jcs.065912. [DOI] [PubMed] [Google Scholar]
Gallo RL, Kim KJ, Bernfield M, Kozak CA, Zanetti M, Merluzzi L, Gennaro R. Identification of CRAMP, a cathelin-related antimicrobial peptide expressed in the embryonic and adult mouse. Journal of Biological Chemistry. 1997;272:13088–13093. doi: 10.1074/jbc.272.20.13088. [DOI] [PubMed] [Google Scholar]
Gehring NH, Kunz JB, Neu-Yilik G, Breit S, Viegas MH, Hentze MW, Kulozik AE. Exon-junction complex components specify distinct routes of nonsense-mediated mRNA decay with differential cofactor requirements. Molecular Cell. 2005;20:65–75. doi: 10.1016/j.molcel.2005.08.012. [DOI] [PubMed] [Google Scholar]
Giorgi C, Yeo GW, Stone ME, Katz DB, Burge C, Turrigiano G, Moore MJ. The EJC factor eIF4AIII modulates synaptic strength and neuronal protein expression. Cell. 2007;130:179–191. doi: 10.1016/j.cell.2007.05.028. [DOI] [PubMed] [Google Scholar]
Godfrey PA, Malnic B, Buck LB. The mouse olfactory receptor gene family. PNAS. 2004;101:2156–2161. doi: 10.1073/pnas.0308051100. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gong Q. Culture of mouse olfactory sensory neurons. Current Protocols in Neuroscience. 2012;3:24. doi: 10.1002/0471142301.ns0324s58. [DOI] [PMC free article] [PubMed] [Google Scholar]
He F, Li X, Spatrick P, Casillo R, Dong S, Jacobson A. Genome-wide analysis of mRNAs regulated by the nonsense-mediated and 5' to 3' mRNA decay pathways in yeast. Molecular Cell. 2003;12:1439–1452. doi: 10.1016/S1097-2765(03)00446-5. [DOI] [PubMed] [Google Scholar]
Hirota J, Mombaerts P. The LIM-homeodomain protein Lhx2 is required for complete development of mouse olfactory sensory neurons. PNAS. 2004;101:8751–8755. doi: 10.1073/pnas.0400940101. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hongo T, Hakuba A, Shiota K, Naruse I. Suckling dysfunction caused by defects in the olfactory system in genetic arhinencephaly mice. Neonatology. 2000;78:293–299. doi: 10.1159/000014282. [DOI] [PubMed] [Google Scholar]
Hornstein N, Torres D, Das Sharma S, Tang G, Canoll P, Sims PA. Ligation-free ribosome profiling of cell type-specific translation in the brain. Genome Biology. 2016;17:149. doi: 10.1186/s13059-016-1005-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
Huang L, Lou CH, Chan W, Shum EY, Shao A, Stone E, Karam R, Song HW, Wilkinson MF. RNA homeostasis governed by cell type-specific and branched feedback loops acting on NMD. Molecular Cell. 2011;43:950–961. doi: 10.1016/j.molcel.2011.06.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
Huang L, Shum EY, Jones SH, Lou CH, Dumdie J, Kim H, Roberts AJ, Jolly LA, Espinoza JL, Skarbrevik DM, Phan MH, Cook-Andersen H, Swerdlow NR, Gecz J, Wilkinson MF. A Upf3b-mutant mouse model with behavioral and neurogenesis defects. Molecular Psychiatry. 2018;23:1773–1786. doi: 10.1038/mp.2017.173. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hurt JA, Robertson AD, Burge CB. Global analyses of UPF1 binding and function reveal expanded scope of nonsense-mediated mRNA decay. Genome Research. 2013;23:1636–1650. doi: 10.1101/gr.157354.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ibarra-Soria X, Levitin MO, Saraiva LR, Logan DW. The olfactory transcriptomes of mice. PLOS Genetics. 2014;10:e1004593. doi: 10.1371/journal.pgen.1004593. [DOI] [PMC free article] [PubMed] [Google Scholar]
Igarashi K, Kataoka K, Itoh K, Hayashi N, Nishizawa M, Yamamoto M. Regulation of transcription by dimerization of erythroid factor NF-E2 p45 with small maf proteins. Nature. 1994;367:568–572. doi: 10.1038/367568a0. [DOI] [PubMed] [Google Scholar]
Iulianella A, Stanton-Turcotte D. The hedgehog receptor Patched1 regulates proliferation, neurogenesis, and axon guidance in the embryonic spinal cord. Mechanisms of Development. 2019;160:103577. doi: 10.1016/j.mod.2019.103577. [DOI] [PubMed] [Google Scholar]
Iwema CL, Schwob JE. Odorant receptor expression as a function of neuronal maturity in the adult rodent olfactory system. The Journal of Comparative Neurology. 2003;459:209–222. doi: 10.1002/cne.10583. [DOI] [PubMed] [Google Scholar]
Jaffrey SR, Wilkinson MF. Nonsense-mediated RNA decay in the brain: emerging modulator of neural development and disease. Nature Reviews Neuroscience. 2018;19:715–728. doi: 10.1038/s41583-018-0079-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jarjour NN, Schwarzkopf EA, Bradstreet TR, Shchukina I, Lin CC, Huang SC, Lai CW, Cook ME, Taneja R, Stappenbeck TS, Randolph GJ, Artyomov MN, Urban JF, Edelson BT. Bhlhe40 mediates tissue-specific control of macrophage proliferation in homeostasis and type 2 immunity. Nature Immunology. 2019;20:687–700. doi: 10.1038/s41590-019-0382-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Johnson JL, Stoica L, Liu Y, Zhu PJ, Bhattacharya A, Buffington SA, Huq R, Eissa NT, Larsson O, Porse BT, Domingo D, Nawaz U, Carroll R, Jolly L, Scerri TS, Kim HG, Brignell A, Coleman MJ, Braden R, Kini U, Jackson V, Baxter A, Bahlo M, Scheffer IE, Amor DJ, Hildebrand MS, Bonnen PE, Beeton C, Gecz J, Morgan AT, Costa-Mattioli M. Inhibition of Upf2-Dependent Nonsense-Mediated decay leads to behavioral and neurophysiological abnormalities by activating the immune response. Neuron. 2019;104:665–679. doi: 10.1016/j.neuron.2019.08.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jolly LA, Homan CC, Jacob R, Barry S, Gecz J. The UPF3B gene, implicated in intellectual disability, autism, ADHD and childhood onset schizophrenia regulates neural progenitor cell behaviour and neuronal outgrowth. Human Molecular Genetics. 2013;22:4673–4687. doi: 10.1093/hmg/ddt315. [DOI] [PubMed] [Google Scholar]
Karam R, Lou CH, Kroeger H, Huang L, Lin JH, Wilkinson MF. The unfolded protein response is shaped by the NMD pathway. EMBO Reports. 2015;16:599–609. doi: 10.15252/embr.201439696. [DOI] [PMC free article] [PubMed] [Google Scholar]
Karousis ED, Mühlemann O. Nonsense-Mediated mRNA decay begins where translation ends. Cold Spring Harbor Perspectives in Biology. 2019;11:a032862. doi: 10.1101/cshperspect.a032862. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kawauchi S, Kim J, Santos R, Wu HH, Lander AD, Calof AL. Foxg1 promotes olfactory neurogenesis by antagonizing Gdf11. Development. 2009;136:1453–1464. doi: 10.1242/dev.034967. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kebaara BW, Atkin AL. Long 3'-UTRs target wild-type mRNAs for nonsense-mediated mRNA decay in Saccharomyces cerevisiae. Nucleic Acids Research. 2009;37:2771–2778. doi: 10.1093/nar/gkp146. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kowalczyk MS, Tirosh I, Heckl D, Rao TN, Dixit A, Haas BJ, Schneider RK, Wagers AJ, Ebert BL, Regev A. Single-cell RNA-seq reveals changes in cell cycle and differentiation programs upon aging of hematopoietic stem cells. Genome Research. 2015;25:1860–1872. doi: 10.1101/gr.192237.115. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kozak M. Point mutations define a sequence flanking the AUG initiator Codon that modulates translation by eukaryotic ribosomes. Cell. 1986;44:283–292. doi: 10.1016/0092-8674(86)90762-2. [DOI] [PubMed] [Google Scholar]
Kurosaki T, Popp MW, Maquat LE. Quality and quantity control of gene expression by nonsense-mediated mRNA decay. Nature Reviews Molecular Cell Biology. 2019;20:406–420. doi: 10.1038/s41580-019-0126-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lewcock JW, Reed RR. A feedback mechanism regulates monoallelic odorant receptor expression. PNAS. 2004;101:1069–1074. doi: 10.1073/pnas.0307986100. [DOI] [PMC free article] [PubMed] [Google Scholar]
Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics. 2014;30:923–930. doi: 10.1093/bioinformatics/btt656. [DOI] [PubMed] [Google Scholar]
Logan DW, Brunet LJ, Webb WR, Cutforth T, Ngai J, Stowers L. Learned recognition of maternal signature odors mediates the first suckling episode in mice. Current Biology. 2012;22:1998–2007. doi: 10.1016/j.cub.2012.08.041. [DOI] [PMC free article] [PubMed] [Google Scholar]
Long AA, Mahapatra CT, Woodruff EA, Rohrbough J, Leung HT, Shino S, An L, Doerge RW, Metzstein MM, Pak WL, Broadie K. The nonsense-mediated decay pathway maintains synapse architecture and synaptic vesicle cycle efficacy. Journal of Cell Science. 2010;123:3303–3315. doi: 10.1242/jcs.069468. [DOI] [PMC free article] [PubMed] [Google Scholar]
Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology. 2014;15:550. doi: 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lykke-Andersen S, Jensen TH. Nonsense-mediated mRNA decay: an intricate machinery that shapes transcriptomes. Nature Reviews Molecular Cell Biology. 2015;16:665–677. doi: 10.1038/nrm4063. [DOI] [PubMed] [Google Scholar]
Lynch M, Marinov GK. The bioenergetic costs of a gene. PNAS. 2015;112:15690–15695. doi: 10.1073/pnas.1514974112. [DOI] [PMC free article] [PubMed] [Google Scholar]
Mabin JW, Woodward LA, Patton RD, Yi Z, Jia M, Wysocki VH, Bundschuh R, Singh G. The exon junction complex undergoes a compositional switch that alters mRNP structure and Nonsense-Mediated mRNA decay activity. Cell Reports. 2018;25:2431–2446. doi: 10.1016/j.celrep.2018.11.046. [DOI] [PMC free article] [PubMed] [Google Scholar]
Malnic B, Hirono J, Sato T, Buck LB. Combinatorial receptor codes for odors. Cell. 1999;96:713–723. doi: 10.1016/S0092-8674(00)80581-4. [DOI] [PubMed] [Google Scholar]
Manglapus GL, Youngentob SL, Schwob JE. Expression patterns of basic helix-loop-helix transcription factors define subsets of olfactory progenitor cells. The Journal of Comparative Neurology. 2004;479:216–233. doi: 10.1002/cne.20316. [DOI] [PubMed] [Google Scholar]
Mao H, Brown HE, Silver DL. Mouse models of Casc3 reveal developmental functions distinct from other components of the exon junction complex. RNA. 2017;23:23–31. doi: 10.1261/rna.058826.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
Markenscoff-Papadimitriou E, Allen WE, Colquitt BM, Goh T, Murphy KK, Monahan K, Mosley CP, Ahituv N, Lomvardas S. Enhancer interaction networks as a means for singular olfactory receptor expression. Cell. 2014;159:543–557. doi: 10.1016/j.cell.2014.09.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
McIntyre JC, Bose SC, Stromberg AJ, McClintock TS. Emx2 stimulates odorant receptor gene expression. Chemical Senses. 2008;33:825–837. doi: 10.1093/chemse/bjn061. [DOI] [PMC free article] [PubMed] [Google Scholar]
Mendell JT, Sharifi NA, Meyers JL, Martinez-Murillo F, Dietz HC. Nonsense surveillance regulates expression of diverse classes of mammalian transcripts and mutes genomic noise. Nature Genetics. 2004;36:1073–1078. doi: 10.1038/ng1429. [DOI] [PubMed] [Google Scholar]
Michaloski JS, Galante PA, Malnic B. Identification of potential regulatory motifs in odorant receptor genes by analysis of promoter sequences. Genome Research. 2006;16:1091–1098. doi: 10.1101/gr.5185406. [DOI] [PMC free article] [PubMed] [Google Scholar]
Michel D, Moyse E, Brun G, Jourdan F. Induction of apoptosis in mouse [correction of rat] olfactory neuroepithelium by synaptic target ablation. Neuroreport. 1994;5:1329–1332. [PubMed] [Google Scholar]
Miyamichi K, Serizawa S, Kimura HM, Sakano H. Continuous and overlapping expression domains of odorant receptor genes in the olfactory epithelium determine the dorsal/ventral positioning of glomeruli in the olfactory bulb. Journal of Neuroscience. 2005;25:3586–3592. doi: 10.1523/JNEUROSCI.0324-05.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Nagy E, Maquat LE. A rule for termination-codon position within intron-containing genes: when nonsense affects RNA abundance. Trends in Biochemical Sciences. 1998;23:198–199. doi: 10.1016/S0968-0004(98)01208-0. [DOI] [PubMed] [Google Scholar]
Nasif S, Contu L, Mühlemann O. Beyond quality control: the role of nonsense-mediated mRNA decay (NMD) in regulating gene expression. Seminars in Cell & Developmental Biology. 2018;75:78–87. doi: 10.1016/j.semcdb.2017.08.053. [DOI] [PubMed] [Google Scholar]
Nguyen LS, Kim HG, Rosenfeld JA, Shen Y, Gusella JF, Lacassie Y, Layman LC, Shaffer LG, Gécz J. Contribution of copy number variants involving nonsense-mediated mRNA decay pathway genes to neuro-developmental disorders. Human Molecular Genetics. 2013;22:1816–1825. doi: 10.1093/hmg/ddt035. [DOI] [PubMed] [Google Scholar]
Nguyen LS, Wilkinson MF, Gecz J. Nonsense-mediated mRNA decay: inter-individual variability and human disease. Neuroscience & Biobehavioral Reviews. 2014;46 Pt 2:175–186. doi: 10.1016/j.neubiorev.2013.10.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
Notaras M, Allen M, Longo F, Volk N, Toth M, Li Jeon N, Colak D. UPF2 leads to degradation of dendritically targeted mRNAs to regulate synaptic plasticity and cognitive function. Molecular Psychiatry. 2019;21:5. doi: 10.1038/s41380-019-0547-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Palacios IM. Nonsense-mediated mRNA decay: from mechanistic insights to impacts on human health. Briefings in Functional Genomics. 2013;12:25–36. doi: 10.1093/bfgp/els051. [DOI] [PubMed] [Google Scholar]
Peterson J, Lin B, Barrios-Camacho CM, Herrick DB, Holbrook EH, Jang W, Coleman JH, Schwob JE. Activating a reserve neural stem cell population in Vitro Enables Engraftment and Multipotency after Transplantation. Stem Cell Reports. 2019;12:680–695. doi: 10.1016/j.stemcr.2019.02.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
Qiu X, Mao Q, Tang Y, Wang L, Chawla R, Pliner HA, Trapnell C. Reversed graph embedding resolves complex single-cell trajectories. Nature Methods. 2017;14:979–982. doi: 10.1038/nmeth.4402. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ramaiah M, Tan K, Plank TM, Song HW, Dumdie JN, Jones S, Shum EY, Sheridan SD, Peterson KJ, Gromoll J, Haggarty SJ, Cook-Andersen H, Wilkinson MF. A microRNA cluster in the Fragile-X region expressed during spermatogenesis targets FMR1. EMBO Reports. 2019;20:e46566. doi: 10.15252/embr.201846566. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rebbapragada I, Lykke-Andersen J. Execution of nonsense-mediated mRNA decay: what defines a substrate? Current Opinion in Cell Biology. 2009;21:394–402. doi: 10.1016/j.ceb.2009.02.007. [DOI] [PubMed] [Google Scholar]
Ressler KJ, Sullivan SL, Buck LB. A molecular dissection of spatial patterning in the olfactory system. Current Opinion in Neurobiology. 1994;4:588–596. doi: 10.1016/0959-4388(94)90061-2. [DOI] [PubMed] [Google Scholar]
Riera CE, Tsaousidou E, Halloran J, Follett P, Hahn O, Pereira MMA, Ruud LE, Alber J, Tharp K, Anderson CM, Brönneke H, Hampel B, Filho CDM, Stahl A, Brüning JC, Dillin A. The sense of smell impacts metabolic health and obesity. Cell Metabolism. 2017;26:198–211. doi: 10.1016/j.cmet.2017.06.015. [DOI] [PubMed] [Google Scholar]
Rodriguez S, Sickles HM, Deleonardis C, Alcaraz A, Gridley T, Lin DM. Notch2 is required for maintaining sustentacular cell function in the adult mouse main olfactory epithelium. Developmental Biology. 2008;314:40–58. doi: 10.1016/j.ydbio.2007.10.056. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rozenkrantz L, Zachor D, Heller I, Plotkin A, Weissbrod A, Snitz K, Secundo L, Sobel N. A mechanistic link between olfaction and autism spectrum disorder. Current Biology. 2015;25:1904–1910. doi: 10.1016/j.cub.2015.05.048. [DOI] [PMC free article] [PubMed] [Google Scholar]
Salem S, Langlais D, Lefebvre F, Bourque G, Bigley V, Haniffa M, Casanova JL, Burk D, Berghuis A, Butler KM, Leahy TR, Hambleton S, Gros P. Functional characterization of the human dendritic cell immunodeficiency associated with the IRF8(K108E) mutation. Blood. 2014;124:1894–1904. doi: 10.1182/blood-2014-04-570879. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sanz E, Yang L, Su T, Morris DR, McKnight GS, Amieux PS. Cell-type-specific isolation of ribosome-associated mRNA from complex tissues. PNAS. 2009;106:13939–13944. doi: 10.1073/pnas.0907143106. [DOI] [PMC free article] [PubMed] [Google Scholar]
Saraiva LR, Ibarra-Soria X, Khan M, Omura M, Scialdone A, Mombaerts P, Marioni JC, Logan DW. Hierarchical deconstruction of mouse olfactory sensory neurons: from whole mucosa to single-cell RNA-seq. Scientific Reports. 2015;5:18178. doi: 10.1038/srep18178. [DOI] [PMC free article] [PubMed] [Google Scholar]
Saraiva LR, Riveros-McKay F, Mezzavilla M, Abou-Moussa EH, Arayata CJ, Makhlouf M, Trimmer C, Ibarra-Soria X, Khan M, Van Gerven L, Jorissen M, Gibbs M, O'Flynn C, McGrane S, Mombaerts P, Marioni JC, Mainland JD, Logan DW. A transcriptomic atlas of mammalian olfactory mucosae reveals an evolutionary influence on food odor detection in humans. Science Advances. 2019;5:eaax0396. doi: 10.1126/sciadv.aax0396. [DOI] [PMC free article] [PubMed] [Google Scholar]
Schwob JE, Jang W, Holbrook EH, Lin B, Herrick DB, Peterson JN, Hewitt Coleman J. Stem and progenitor cells of the mammalian olfactory epithelium: taking poietic license. Journal of Comparative Neurology. 2017;525:1034–1054. doi: 10.1002/cne.24105. [DOI] [PMC free article] [PubMed] [Google Scholar]
Serizawa S, Miyamichi K, Nakatani H, Suzuki M, Saito M, Yoshihara Y, Sakano H. Negative feedback regulation ensures the one receptor-one olfactory neuron rule in mouse. Science. 2003;302:2088–2094. doi: 10.1126/science.1089122. [DOI] [PubMed] [Google Scholar]
Serizawa S, Miyamichi K, Sakano H. One neuron-one receptor rule in the mouse olfactory system. Trends in Genetics. 2004;20:648–653. doi: 10.1016/j.tig.2004.09.006. [DOI] [PubMed] [Google Scholar]
Serizawa S, Miyamichi K, Sakano H. Negative feedback regulation ensures the one neuron-one receptor rule in the mouse olfactory system. Chemical Senses. 2005;30:i99–i100. doi: 10.1093/chemse/bjh133. [DOI] [PubMed] [Google Scholar]
Shang Y, Smith S, Hu X. Role of notch signaling in regulating innate immunity and inflammation in health and disease. Protein & Cell. 2016;7:159–174. doi: 10.1007/s13238-016-0250-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shum EY, Jones SH, Shao A, Dumdie J, Krause MD, Chan WK, Lou CH, Espinoza JL, Song HW, Phan MH, Ramaiah M, Huang L, McCarrey JR, Peterson KJ, De Rooij DG, Cook-Andersen H, Wilkinson MF. The antagonistic gene paralogs Upf3a and Upf3b govern Nonsense-Mediated RNA decay. Cell. 2016;165:382–395. doi: 10.1016/j.cell.2016.02.046. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sohni A, Tan K, Song HW, Burow D, de Rooij DG, Laurent L, Hsieh TC, Rabah R, Hammoud SS, Vicini E, Wilkinson MF. The neonatal and adult human testis defined at the Single-Cell level. Cell Reports. 2019;26:1501–1517. doi: 10.1016/j.celrep.2019.01.045. [DOI] [PMC free article] [PubMed] [Google Scholar]
Spies N, Burge CB, Bartel DP. 3' UTR-isoform choice has limited influence on the stability and translational efficiency of most mRNAs in mouse fibroblasts. Genome Research. 2013;23:2078–2090. doi: 10.1101/gr.156919.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tan L, Li Q, Xie XS. Olfactory sensory neurons transiently express multiple olfactory receptors during development. Molecular Systems Biology. 2015;11:844. doi: 10.15252/msb.20156639. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tan K, Wang Z, Zhang Z, An L, Tian J. IVF affects embryonic development in a sex-biased manner in mice. Reproduction. 2016;151:443–453. doi: 10.1530/REP-15-0588. [DOI] [PubMed] [Google Scholar]
Tan K, Song HW, Wilkinson MF. Single-cell RNAseq analysis of testicular germ and somatic cell development during the perinatal period. Development. 2020a;147:dev183251. doi: 10.1242/dev.183251. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tan K, Song HW, Thompson M, Munyoki S, Sukhwani M, Hsieh TC, Orwig KE, Wilkinson MF. Transcriptome profiling reveals signaling conditions dictating human spermatogonia fate in vitro. PNAS. 2020b;117:17832–17841. doi: 10.1073/pnas.2000362117. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tarpey PS, Raymond FL, Nguyen LS, Rodriguez J, Hackett A, Vandeleur L, Smith R, Shoubridge C, Edkins S, Stevens C, O'Meara S, Tofts C, Barthorpe S, Buck G, Cole J, Halliday K, Hills K, Jones D, Mironenko T, Perry J, Varian J, West S, Widaa S, Teague J, Dicks E, Butler A, Menzies A, Richardson D, Jenkinson A, Shepherd R, Raine K, Moon J, Luo Y, Parnau J, Bhat SS, Gardner A, Corbett M, Brooks D, Thomas P, Parkinson-Lawrence E, Porteous ME, Warner JP, Sanderson T, Pearson P, Simensen RJ, Skinner C, Hoganson G, Superneau D, Wooster R, Bobrow M, Turner G, Stevenson RE, Schwartz CE, Futreal PA, Srivastava AK, Stratton MR, Gécz J. Mutations in UPF3B, a member of the nonsense-mediated mRNA decay complex, cause syndromic and nonsyndromic mental retardation. Nature Genetics. 2007;39:1127–1133. doi: 10.1038/ng2100. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tepe B, Hill MC, Pekarek BT, Hunt PJ, Martin TJ, Martin JF, Arenkiel BR. Single-Cell RNA-Seq of mouse olfactory bulb reveals cellular heterogeneity and Activity-Dependent molecular census of Adult-Born neurons. Cell Reports. 2018;25:2689–2703. doi: 10.1016/j.celrep.2018.11.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wachter A, Hartmann L. NMD: nonsense-mediated defense. Cell Host & Microbe. 2014;16:273–275. doi: 10.1016/j.chom.2014.08.015. [DOI] [PubMed] [Google Scholar]
Wada M, Lokugamage KG, Nakagawa K, Narayanan K, Makino S. Interplay between coronavirus, a cytoplasmic RNA virus, and nonsense-mediated mRNA decay pathway. PNAS. 2018;115:E10157–E10166. doi: 10.1073/pnas.1811675115. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wang SS, Tsai RY, Reed RR. The characterization of the Olf-1/EBF-like HLH transcription factor family: implications in olfactory gene regulation and neuronal development. The Journal of Neuroscience. 1997;17:4149–4158. doi: 10.1523/JNEUROSCI.17-11-04149.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
Whitman MC, Greer CA. Adult neurogenesis and the olfactory system. Progress in Neurobiology. 2009;89:162–175. doi: 10.1016/j.pneurobio.2009.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
Woodward LA, Mabin JW, Gangras P, Singh G. The exon junction complex: a lifelong guardian of mRNA fate. Wiley Interdisciplinary Reviews: RNA. 2017;8:e1411. doi: 10.1002/wrna.1411. [DOI] [PubMed] [Google Scholar]
Zhang X, Firestein S. The olfactory receptor gene superfamily of the mouse. Nature Neuroscience. 2002;5:124–133. doi: 10.1038/nn800. [DOI] [PubMed] [Google Scholar]
Zhang LJ, Gallo RL. Antimicrobial peptides. Current Biology : CB. 2016;26:R14–R19. doi: 10.1016/j.cub.2015.11.017. [DOI] [PubMed] [Google Scholar]
Zheng S, Gray EE, Chawla G, Porse BT, O'Dell TJ, Black DL. PSD-95 is post-transcriptionally repressed during early neural development by PTBP1 and PTBP2. Nature Neuroscience. 2012;15:381–388. doi: 10.1038/nn.3026. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zheng X, Boyer L, Jin M, Mertens J, Kim Y, Ma L, Ma L, Hamm M, Gage FH, Hunter T. Metabolic reprogramming during neuronal differentiation from aerobic glycolysis to neuronal oxidative phosphorylation. eLife. 2016;5:e13374. doi: 10.7554/eLife.13374. [DOI] [PMC free article] [PubMed] [Google Scholar]

eLife. doi: 10.7554/eLife.57525.sa1

Decision letter

Editor: Didier YR Stainier¹

Reviewed by: David M Bedwell², Luis R Saraiva³

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Thank you for submitting your article "The NMD factor UPF3B shapes olfactory neurogenesis" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by Didier Stainier as the Senior and Reviewing Editor. The following individuals involved in review of your submission have agreed to reveal their identity: David M Bedwell (Reviewer #1); Luis R Saraiva (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Reviewer #1:

In this study, the authors examined the role of the NMD factor UPF3B in olfactory neurogenesis. They identified stage-specific olfactory sensory neuron (OSN) cell clusters in WT mice and found that UPF3B loss resulted in changes in the proportion of these cell clusters. They also identified a number of interesting UPF3B-dependent NMD target transcripts in vivo using single-cell RNA-sequencing (scRNAseq) analysis, including strong candidates with roles in olfactory neurogenesis, and neural development. Finally, they showed that UPF3B dampens the expression of a large cadre of anti-microbial genes and promotes the selection of specific olfactory receptor (Olfr) genes. This study provides important new insights into the olfactory field and identifies an important new post-transcriptional regulator that shapes OSN gene expression and OSN development.

Overall, this is an interesting manuscript with exiting new insights into the role of UPF3B in regulating gene expression in OSN cell clusters. However, I suggest that the authors consider the following concerns:

1) I am concerned about the lack of validation of your RNA-seq data in Figure 1. The only secondary validation seemed to be the experiment using the method of Alkallas et al. to measure the stability of 127 mRNAs upregulated in Upf3b-null mOSNs. You state that this method "infers RNA stability based on pre-mRNA and steady-state mRNA levels". Going to the Alkallas paper, they say that it is "a method for unbiased estimation of differential mRNA decay rate from RNA-sequencing data by modeling the kinetics of mRNA metabolism." So, it seems to uphold the changes in abundance as predicted by the RNA-seq, but it seems that you are using RNA-seq to validate RNA-Seq data, which seems to be a circular argument. I didn't see any other data examining the validity of your RNA-seq data. Is there not any other way to confirm these results from the FACS-purified mOSNs?

2) The labels of many panels (F-J) in Figure 2 do not match the text or figure legend. This made it difficult to try to match the right panel with the description in the text (and legend). Please correct the labeling.

Reviewer #2:

The authors apply a mixed bulk and single-cell RNA-seq based approach to address an interesting and important question regarding the role of NMD in the neurogenesis and development of OSNs in the mouse nose. As it stands, the structure and language used in many sections of this manuscript made it confusing and hard to follow. Critically, the lack of rigor in the statistical and methodological reporting impact my ability to properly assess the validity of the results and conclusions of this study.

Thus, there are several points that I would like to see the authors address:

1) Based the body of olfaction-related literature cited, study design and naïveté of the analysis, it is clear that the authors are not well familiarized with the olfactory system. In the last 6 years, multiple studies have performed RNA-seq in the mouse whole OE, pools of sorted OSNs and also single OSNs and other cell types (PMIDs: 25187969, 31392275, 26670777, 26646940, 26541607, 28506465). I was surprised to note that with the exception of the above underlined reference, the authors failed to mention and discuss this work in their submitted manuscript. Why is this important? Some key results from those studies would have helped guide the experimental design and data analysis presented in this submitted manuscript (see point 2).

2) Some of the key results in this manuscript arise from FACS experiments done using dissociated cells from the nose of OMP-YFP mice, but details are missing regarding the number of cells sorted from each mouse, and also the number of cells (and animals) that incorporate each sample sequenced. Why is this important? Each mouse has on average 10 million mOSNs, each expressing 1 out of ~1100 intact Olfr genes. Since Olfr genes are expressed in the OE in a zonal fashion, dissection, dissociation and FACS experimental procedures can all lead to subsampling and subsequently severely bias the results. Moreover, it is known from previous FACS+RNA-seq experiments using OMP-GFP animals, that 2 sub-populations exist among the mOSNs expressing Omp, one expressing lower Omp levels than the other, and representing mOSNs that are not yet in the latest maturation stage possible (PMID: 26670777). Thus, in order to critically assess the validity of the conclusions from the analyses relative to Figure 1 and Figure 2, the authors should provide the following information:

2.1) The exact number of sorted cells for each of their sequenced samples, and FACS plots (and associated gating cutoffs used) in supplementary data. This could also help explain the puzzling result of why half of the KO and WT samples share the same space in the PCA from Figure 1—figure supplement 1B.

2.2) How many Olfr genes (list pseudogenes and intact genes) can the authors detect in each of the sequenced samples and how do they compare between all samples? A sample capturing the whole diversity of the OE should show evidence for expression for 98.9% of all annotated Olfr genes (PMID: 26670777). If this is not the case, the downstream analysis is compromised and the conclusions not well supported.

3) The Materials and methods section severely lacks rigor and structure. Some methodological sections need more detail and structure. For example, for every experiment described, the age, number of replicates, genotype and sex of the animals used should be stated. Also, explicit details are lacking throughout the manuscript regarding which exact mouse crossings (for the genetically modified lines) were used in different experiments. More importantly, methods describing some experiments shown in the manuscript are entirely missing from the manuscript. For example, how were the OE samples from WT and KO mice collected and processed for whole tissue RNA-seq and for immunofluorescence? how were the weighting experiments from (Figure 1A) performed? How were the behavioral tests in Figure 1—figure supplement 1A performed? How were the samples prepared ahead of FACS and how many cells, and from how many animals, are contained in each biological replicate of FACS-purified OSNs? How were Western blots performed?

4) The authors mention they analyzed several neuronal markers, which they show in Figure 1B. Not all markers tested are for neurons. In reality, they probed the gene expression for 3 canonical markers of iOSNs (Gap43), HBCs (Krt5) and mOSNs (Gnal). Please correct the main text accordingly. Also, since the authors performed RNA-seq on the whole OE, the RNA-seq gene expression estimates for the canonical markers of the main cell types populating the OE should be shown (I suggest plotting something similar to similar to Figure 1C in PMID: 31392275).

4.1) In this context, since Omp is the canonical marker for all mOSNs subtypes, I am puzzled as to why the authors used Gnal instead. This is especially relevant because they show the RQ values for Omp in Figure 1—figure supplement 1C, but no significant difference was observed between WT and KO, which does not fit with the results from the other markers used in Figure 1B. Please explain.

4.2) Also, it is clear from the immunochemistry experiments in Figure 6G that the layer containing OMP+ cells is thicker in WT vs. KO mice, suggesting that a higher number of OMP+ cells exist in WT mice (in line with the hypothesis that KO mice have reduced olfactory abilities, possibly caused to lower numbers of mOSNs). The authors should check if OMP+ cells are more highly prevalent throughout the whole OE in WT vs. KO, or if this decrease is zonally restricted. A counting of OMP+ per mm² would help convince me and clarify this topic.

5) The authors mention: "Despite the reduction in GBCs and iOSNs, Upf3b-null mice had normal numbers of mOSNs (53% vs. 47% in the control) (Figure 6A). Despite the reduction in GBCs and iOSNs, Upf3b-null mice had normal numbers of mOSNs (53% vs. 47% in the control) (Figure 6A). Consistent with this, the mOSN marker, OMP, was similarly expressed (at both the RNA and protein levels) in OE from Upf3b-null and control mice (Figure 1—figure supplement 1C, D)". How can the authors exclude the hypothesis that potential biases caused by dissection and dissociation (see point 2 above), are the cause for these differences? Do the authors have statistics like shown in Figure 6A for individual mice, or were all the WT and KO samples pulled together in a single experiment for each genotype?

6) The authors mention that in Figure 6G "immunofluorescence analysis showed that an anti-CAMP antiserum showed a strong signal in the OE, as well as the lamina propria". While the presence of CAMP staining in the lamina propria is clear, I cannot distinguish the signal in the OE. Please provide evidence of this or rephrase.

7) It is critical the authors should validate (at least) a subset of the key differentially expressed genes (namely some Olfrs) identified in their RNA-seq analyses. This is important to confirm the validity of the RNA-seq data and refute the possibility that the differentially expressed genes identified are due to biases in dissection and/or dissociation. This validation could be done with immunohistochemistry or with in-situ hybridization, followed by counting of the positive cells across the whole OE in multiple animals and the two genotypes.

8) In its current form, the manuscript is not easy to follow. In my opinion, this is partly caused by the structure and language used in several sections of this manuscript, and partly because of the way data is presented in the main figures. In the revised version, the authors should consider simplifying the language to make it more accessible and improving the structure and the way the data is presented, as means to improve clarity (I give several examples on this topic below).

Reviewer #3:

The manuscript "The NMD factor UPF3B shapes olfactory neurogenesis" by Tan, Jones and Wilkinson studies the role of UPF3B-dependent NMD in olfactory epithelium (OE). This branch of the nonsense-mediated RNA decay (NMD) pathway is critical for human cognition. Mutations in Upf3b have been described as a potential cause of mental retardation, autism, childhood-onset schizophrenia, bipolar disorder and attention-deficit hyperactivity disorder in several families. Using Upf3b-null mice, the authors perform the powerful single-cell RNAseq analysis in OE and delineate the roles of UPF3B-dependent NMD in various stages of OE neurogenesis. First, by showing that several olfactory neural markers exhibited significantly decreased expression in Upf3b-null as compared to control OE, the authors confirm an olfactory defect upon loss of UPF3B protein. Their RNAseq data on sorted mOSNs identifies more than 200 differentially expressed genes between Upf3b-null and control mOSNs. Several genes involved in neurogenesis were among these genes suggesting a role of UPF3B in OE neurogenesis. They then screened the statistically upregulated genes for features known to trigger NMD. By combining this screen with a recently developed method that allows unbiased estimation of mRNA decay, the authors determined likely NMD targets within mOSNs. The authors also provide new cellular and molecular information on wild type OSNs and their development in vivo. Although this part of the study is not directly related to NMD, it provides a unique set of resource for the olfactory field. The scRNAseq analysis was employed to identify novel mOSN cell subsets, and thereby determination of the specific cell subsets of mature olfactory sensory neurons that are regulated by UPF3B-dependent NMD. Their experiments are in-depth and their findings are novel. While I have much enthusiasm for this study, there are some issues that need to be attended to prior to consideration for publication:

Specific comments:

1) The authors should consider changing the title, maybe using a more specific term such as “olfactory epithelium neurogenesis”, as olfactory neurogenesis might be confused with the neurogenesis ongoing in olfactory bulb in adulthood (constant neuron production for olfactory bulb from SVZ). Same rewording should be applied to throughout the text as “olfactory neurogenesis” is broadly used in the manuscript.

2) The authors found that 15 previously defined likely mouse NMD target mRNAs (based on upregulation and/or stabilization in response to NMD factor depletion or high UPF1 occupancy) overlapped with mRNAs upregulated in Upf3b-null mOSNs: Atp10d, Lbh, Slc38a6, Tgm2, Rgl3, Notch2, Ywhag, Luc7l, Ptch1, 1700025G04Rik, Tle3, Ptprn, Ptger2, Dhps, and Msrb3. To assess the direct NMD targets, the authors measured the stability of the 127 mRNAs upregulated in Upf3b-null mOSNs using a method that infers RNA stability based on pre-mRNA and steady-state mRNA levels. This method revealed that 82 of 127 upregulated genes encode mRNAs stabilized in Upf3b-null mOSNs as compared to control mOSNs. Of these 82 stabilized and upregulated mRNAs, 52 had at least 1 of the 3 NMD-inducing features (NIFs) (1- an exon-exon junction >50 nt downstream of the stop codon, 2- an uORF, or 3- a long 3'UTR) that they examined. Thus, the authors classified these 52 mRNAs as high confidence mOSN NMD targets.

4 of the 15 previously defined NMD targets (Tle3, Ptprn, Rgl3, and Dhps) that initially overlapped with the list of 127 upregulated mRNAs, never made to the list of 52 upregulated, stabilized, and with NIFs. The authors should provide an explanation for this discrepancy. Is it because those 4 genes were identified as NMD targets solely based on their upregulation in response to NMD factor depletion in previous studies? If so, that list or at least those 4 genes should be omitted from the text. I can see that the authors were trying to do a thorough analysis by combining stability, upregulation and NMD-inducing features to determine the direct targets of NMD. However, those 15 previously defined likely mouse NMD target mRNAs are confusing.

3) Presence of uORF and an exon-exon junction >50 nt downstream of the stop codon are well-established NIFs. Is a long 3' UTR alone a well-established NIF? Although the authors used a stability analysis, they should be careful with their interpretation of direct NMD targets, especially with the exon-length criteria. The majority of the direct-target list (34 mRNAs of 52 mRNAs) did not have either a uORF or an exon-exon junction >50 nt downstream of the stop codon, but just a long 3'UTR. They should either provide additional validation for these 34 mRNAs or rephrase the relevant part. Intriguingly, 8 of the previously "defined" 15 NMD-target mRNAs mentioned above were among the 34 mRNAs with long 3'UTRs, while only 3 mRNAs with uORF were common in both lists. Were those 8 mRNAs found to have high UPF1 occupancy or stability upon depletion of a different NMD factor than UPF3B? Because this could provide additional support that they could be direct NMD targets.

4) The four subtypes in olfactory epithelium, and their trajectory (HBC → GBC → iOSN → mOSN) are well established. However, the authors spent more than 2 pages in describing OE cell subsets and molecular pathways active during olfactory neurogenesis. Although Upf3b-null mice have reduced numbers of HBCs, more GBCs, and more iOSNs, Upf3b-null mice still had normal numbers of mOSNs (53% vs. 47% in the control). Because the authors focused only in OSNs when identifying direct targets of NMD (due to the OSNs being the functional unit in OE), the part describing other OE cell subsets and molecular pathways active during olfactory neurogenesis should be more concise.

5) Table S1 has overwhelming data (5 excel sheets), and it is hard to pull out the antimicrobial genes from 12,000 genes. How many are they? Are they more than what is shown in Figure 6E? If so, it should be shown in a table (similar format to Table 1).

6) The authors suggest a model that UPF3B protein promotes the selection of the downregulated 78 Olfr genes to be the dominant Olfr gene expressed in individual mOSNs. Is there a transcription factor or other kind of transcriptional regulator that positively regulates Olfr genes and is downregulated in KO mOSNs? Did any known repressor of Olfr gene expression show an increased expression in KO mOSNs?

7) The scRNAseq data shows that the Upf3b-null OEs lack mOSN-2 sub-cluster. Could the downregulation of 78 Olfr genes in KO mOSNs be a reflection of this? Maybe, those 78 genes are predominantly expressed in the mOSN-2 sub-cluster. The authors should discuss this.

8) The Western blot of CAMP looks convincing. However, it is not specific to the OSNs. The CAMP signal is not clear in mOSNs in immunohistochemistry images. It looks like that CAMP was upregulated in the entire OE, signal being weak in mOSNs compared to other cell types. Showing CAMP staining only with DAPI in another panel (no merging with OMP staining) would work.

9) In addition to Olfr genes, the authors discover that UPF3B in OSNs also regulates antimicrobial genes. Unlike the Olfrs, the antimicrobial genes are upregulated in KO mOSNs. Similar to comment #6, is any of the direct NMD targets identified in Figure 2 known to trigger inflammation or antimicrobial genes? The authors should offer a mechanism/s as to how NMD indirectly regulates these major sets of genes. Their finding of another major class of genes is regulated by UPF3B suggests that the UPF3B prortein has independent functions within OSNs, one being suppression of OE inflammation. Or, can loss of Olfr genes and an increase in immune response be connected? Can one be responsible for the other?

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Thank you for submitting your article "The role of the NMD factor UPF3B in olfactory sensory neurons" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by Didier Stainier as the Senior and Reviewing Editor. The following individuals involved in review of your submission have agreed to reveal their identity: David M Bedwell (Reviewer #1); Luis R Saraiva (Reviewer #2).

While the revised manuscript was found to be much improved, there are several remaining issues that should be addressed.

Reviewer #1:

In this study, the authors examined the role of the NMD factor UPF3B in olfactory neurogenesis. They identified stage-specific olfactory sensory neuron (OSN) cell clusters in WT mice and found that UPF3B loss resulted in changes in the proportion of these cell clusters. They also identified a number of interesting UPF3B-dependent NMD target transcripts in vivo using single-cell RNA-sequencing analysis, including strong candidates with roles in olfactory neurogenesis, and neural development. Finally, they showed that UPF3B dampens the expression of a large cadre of anti-microbial genes and promotes the selection of specific olfactory receptor (Olfr) genes. This study provides important new insights into the olfactory field and identifies an important new post-transcriptional regulator that shapes OSN gene expression and OSN development.

Overall, this is an interesting manuscript with exciting new insights into the role of UPF3B in regulating gene expression in OSN cell clusters. The authors adequately addressed each of my previous concerns.

Reviewer #2:

The authors have made improvements in the manuscript, but unfortunately, they have not convincingly addressed some of my major concerns. Below I am detailing my responses to the rebuttal letter point by point:

1) No more comments. This concern is addressed.

2.1) The authors then present "two lines of evidence against a significant subsampling bias in their samples: First, each sample was pooled from 3 mice. Second, most samples had a similar percentage of OMP-YFP+ cells (between 5-7%)".

Unfortunately, without knowing the numbers of cells sorted in each sample, I am failing to see how any of these two reasons argue against it. The olfactory epithelium of young adult mice (8-10 weeks) is populated on average by ~10 million mature OSNs (mOSNs). If each sample contains on average 10,000 sorted cells, that would capture only 0.1% of all mOSNs. If each sample contains 100,000 sorted cells, that would still only make up 1% of all mOSNs, and so on. Since the authors "did not make note of the output cell number" in each sample, how could pooling animals or sorting similar percentages of cells between animals be used as an argument against subsampling? Also, only 4/8 of their samples had between 5-7% of sorted OMP-YFP+ cells, which is not "most samples". I am assuming the authors kept the FACS files, and thus should be able to retrieve those values. Is this the case?

Finally, what is the logic underlying the following statement: "We note that our manuscript is currently ~30% over the word limit."?

2.2) The authors here claim that they asked, "what genes are regulated by a particular factor in mOSNs" and that "the depth of RNA-seq analysis we used was sufficient to draw several conclusions". One of the unexpected major conclusions of the authors is that "over half of the downregulated genes in Upf3-null mOSNs are Olfr genes" and state that they "regard this finding as a robust result, as the RNA-seq coverage for Upf3b-null and control OSNs was 95.8% and 96.0%, respectively, of all annotated Olfr gene".

Since Olfr and many other genes are expressed zonally in the olfactory epithelium (PMID: 7812149, 15814789, 12709059, 32209480, etc), differences in dissection between animals could easily account for the differences observed in the RNA-seq data from sorted cells. Thus, showing the gene expression distribution of all Olfr genes for each of the sorted samples is a key QC step for this study and it should feature early on in the main figures. The authors should include that in the main text and present a plot containing the mean gene expression values (in log10) for all annotated mouse Olfr in either descending or ascending order for the WT and KO, include similar plots in supplementary for each individual sample and include all the gene names (and not just Ensembl IDs) in the “TPM” sheet of Supplementary file 1 to make this accessible to the community. It would also be very useful to present in supplementary a Venn diagram comparing all the Olfrs expressed in WT vs KO mice.

3) The Materials and methods are now much improved. Only one minor comment here: the authors should include the commercial source of the coyote and bobcat urines.

4) As I mentioned in my original comment, as Gap43 and Krt5 are not expressed in mature OSNs, they should not be called “olfactory neuronal markers”. Instead, Gap43 is a canonical marker for immature neurons and Krt5 a canonical marker for horizontal basal cells. In line with this, the authors do mention that "mOSN marker gene, Omp, is enriched in these mOSN samples, and that the non-mOSN markers genes, Krt5 and Lgr5, are de-enriched in these samples". Finally, the authors mention that their RNA-seq analysis was done "purified mOSNs, not whole OE". I was indeed aware of this, as it is was already explicit in the manuscript. I suggested plotting something similar to Figure 1C in PMID: 31392275, for 2 main reasons: i) it summarizes in one plot the gene expression levels for the canonical markers for all major cell types populating the olfactory epithelium, and ii) it visually allows the reader assess the levels of “contamination/hitchhiking” by other cell types during the sorting procedure.

Please correct this by specifically stating the cell types that those genes are indicative of (also, by “de-enriched” I assume the authors mean “depleted”?).

4.1) The Omp expression data should feature in Figure 1B, as it is the best canonical marker for mOSNs (also see previous response). It is well established that Omp and Gnal are expressed ubiquitously in the majority of mOSNs across all the olfactory epithelium, so I do not see how that could be an explanation.

4.2) No more comments. This concern is addressed.

5) Regarding Figure 6—figure supplement 1B in the revised manuscript: HBCs and GBCs are not OSNs, and it is incorrect to refer them as such. Only immature and mature OSNs should be called OSNs. Please correct this in the figures and throughout the text. Also, on the right-hand panel of Figure 6—figure supplement 1B it would be useful to list all the other cell types that make up 100%.

6) No more comments. This concern is addressed.

7) Olfr genes are expressed zonally in the olfactory epithelium (PMID: 7812149, 15814789, 12709059, 32209480, etc) and differences in dissection between animals could easily translate into differential expression artifacts. Moreover, different mouse strains express different Olfrs at different levels (PMID: 28438259). Since the authors i) have not yet convincingly addressed my previous concerns about a potential bias in dissection (see above), ii) the mice used in this studies were on mixed or different genetic backgrounds, and iii) one of the major claims of the authors is that half of the differentially expressed genes are Olfrs; the authors should validate at least their top 2 differentially expressed Olfrs across the two genotypes.

8) This comment is addressed.

Reviewer #3:

The authors have done an excellent job revising this manuscript. The revised submission is significantly improved. I have no further comments. The manuscript should be accepted as it is.

eLife. 2020 Aug 10;9:e57525. doi: 10.7554/eLife.57525.sa2

Author response

Reviewer #1:

[…]

1) I am concerned about the lack of validation of your RNA-seq data in Figure 1. The only secondary validation seemed to be the experiment using the method of Alkallas et al. to measure the stability of 127 mRNAs upregulated in Upf3b-null mOSNs. You state that this method "infers RNA stability based on pre-mRNA and steady-state mRNA levels". Going to the Alkallas paper, they say that it is "a method for unbiased estimation of differential mRNA decay rate from RNA-sequencing data by modeling the kinetics of mRNA metabolism." So, it seems to uphold the changes in abundance as predicted by the RNA-seq, but it seems that you are using RNA-seq to validate RNA-Seq data, which seems to be a circular argument. I didn't see any other data examining the validity of your RNA-seq data. Is there not any other way to confirm these results from the FACS-purified mOSNs?

To validate, we performed qPCR analysis on 8 genes with known functions, such as in neural development, from the list of genes shown by our RNA-seq analysis to be differentially expressed between Upf3b-null mOSNs and control mOSNs. This qPCR analysis confirmed the regulation of 7 out of 8 of these genes (Figure 1—figure supplement 2B in the revised manuscript). The one exception – Ptger2 – was not detectably expressed using the qPCR primers we used (data not shown). Genes shown to be upregulated in Upf3-null mOSNs by RNA-seq analysis were found to also be upregulated in Upf3b-null mOSNs as judged by qPCR; the same concordance between the 2 assays was shown for downregulated genes.

2) The labels of many panels (F-J) in Figure 2 do not match the text or figure legend. This made it difficult to try to match the right panel with the description in the text (and legend). Please correct the labeling.

We thank reviewer 1 for noticing this. We corrected the labeling in the revised manuscript.

Reviewer #2:

[…]

Thus, there are several points that I would like to see the authors address:

1) Based the body of olfaction-related literature cited, study design and naïveté of the analysis, it is clear that the authors are not well familiarized with the olfactory system. In the last 6 years, multiple studies have performed RNA-seq in the mouse whole OE, pools of sorted OSNs and also single OSNs and other cell types (PMIDs: 25187969, 31392275, 26670777, 26646940, 26541607, 28506465). I was surprised to note that with the exception of the above underlined reference, the authors failed to mention and discuss this work in their submitted manuscript. Why is this important? Some key results from those studies would have helped guide the experimental design and data analysis presented in this submitted manuscript (see point 2).

We thank reviewer 2 for providing the PubMed numbers for these genome-wide studies on the topic of OE cells. We had previously read most of these papers, but not all of them. In our revised manuscript, we cite all 6 of these papers in the Introduction as follows: “To gain insight into the nature of the cells in the OE and their developmental relationships, recent studies have performed transcriptome profiling using whole OE, pools of sorted OSNs, single OSNs, or single OE cells (Fletcher et al., 2017; Ibarra-Soria, Levitin, Saraiva, and Logan, 2014; Saraiva et al., 2015; Saraiva et al., 2019; Tan, Li, and Xie, 2015). […] These studies have also advanced our understanding of the evolution of mammalian olfaction”.

2) Some of the key results in this manuscript arise from FACS experiments done using dissociated cells from the nose of OMP-YFP mice, but details are missing regarding the number of cells sorted from each mouse, and also the number of cells (and animals) that incorporate each sample sequenced. Why is this important? Each mouse has on average 10 million mOSNs, each expressing 1 out of ~1100 intact Olfr genes. Since Olfr genes are expressed in the OE in a zonal fashion, dissection, dissociation and FACS experimental procedures can all lead to subsampling and subsequently severely bias the results. Moreover, it is known from previous FACS+RNA-seq experiments using OMP-GFP animals, that 2 sub-populations exist among the mOSNs expressing Omp, one expressing lower Omp levels than the other, and representing mOSNs that are not yet in the latest maturation stage possible (PMID: 26670777). Thus, in order to critically assess the validity of the conclusions from the analyses relative to Figure 1 and Figure 2, the authors should provide the following information:

We thank reviewer 2 for bringing up this issue. While we cannot completely rule out a subtle subsampling bias, we provide evidence below in our response to point 2.1, against this possibility.

2.1) The exact number of sorted cells for each of their sequenced samples, and FACS plots (and associated gating cutoffs used) in supplementary data. This could also help explain the puzzling result of why half of the KO and WT samples share the same space in the PCA from Figure 1—figure supplement 1B.

The FACS plots of all the dissociated OE samples and associated gating cut-offs used for our RNA-seq analysis are in Figure 1—figure supplement 1B. Unfortunately, we did not make note of the output cell number. However, two lines of evidence against a significant subsampling bias are the following: First, each sample was pooled from 3 mice. Second, most samples had a similar percentage of OMP-YFP+ cells (between 5-7%). With regard to reviewer 2’s comment above about OMP-high and -low cell populations, we did not detect these two populations, perhaps because we used a different reporter mouse than reviewer 2 did in their analysis (Saraiva et al., 2015). We have not indicated this in the revised manuscript, but can do so if deemed important. We note that our manuscript is currently ~30% over the word limit.

2.2) How many Olfr genes (list pseudogenes and intact genes) can the authors detect in each of the sequenced samples and how do they compare between all samples? A sample capturing the whole diversity of the OE should show evidence for expression for 98.9% of all annotated Olfr genes (PMID: 26670777). If this is not the case, the downstream analysis is compromised and the conclusions not well supported.

The reference brought up by reviewer 2 required “deep analysis” to answer a key question it addresses – what is the nature of the expressed Olfr repertoire? This is a fascinating question, as mice harbor >1000 Olfr genes and thus it is important to know whether all of these genes are actually expressed. Using RNA-seq analysis, the paper shows that at least 98.9% of all Olfr genes in mice are detectably expressed in the OE. Thus, the vast majority of the >1000 Olfr genes in the mouse genome are expressed and, thus, have the potential to be functional. To draw this conclusion, it was critical that the authors do extremely deep RNA-seq analysis on a very large number of mOSNs.

In our case, we were instead asking what genes are regulated by a particular factor in mOSNs. The depth of RNA-seq analysis we used was sufficient to draw several conclusions. For example, we were able to define numerous in vivo NMD RNA targets from our RNA-seq analysis. Unexpectedly, our RNA-seq analysis revealed that over half of the downregulated genes in Upf3-null mOSNs are Olfr genes (Figure 1C in the revised manuscript). We regard this finding as a robust result, as the RNA-seq coverage for Upf3b-null and control OSNs was 95.8% and 96.0%, respectively, of all annotated Olfr genes. Given that this coverage is only slightly less that of the study described above (98.9%), it is unlikely that we missed many Olfr genes significantly misregulated in Upf3b-null OSNs. But even if this case (a caveat we can bring up in our manuscript if deemed important), our data still strongly supports the concept that a surprisingly large fraction of Olfrs are regulated by NMD in mOSNs. Furthermore, our data documents 78 specific Olfrs subject to this regulation.

3) The Materials and methods section severely lacks rigor and structure. Some methodological sections need more detail and structure. For example, for every experiment described, the age, number of replicates, genotype and sex of the animals used should be stated. Also, explicit details are lacking throughout the manuscript regarding which exact mouse crossings (for the genetically modified lines) were used in different experiments. More importantly, methods describing some experiments shown in the manuscript are entirely missing from the manuscript. For example, how were the OE samples from WT and KO mice collected and processed for whole tissue RNA-seq and for immunofluorescence? how were the weighting experiments from (Figure 1A) performed? How were the behavioral tests in Figure 1—figure supplement 1A performed? How were the samples prepared ahead of FACS and how many cells, and from how many animals, are contained in each biological replicate of FACS-purified OSNs? How were Western blots performed?

We apologize for this previously missing information, which has been added to the revised Materials and methods section.

4) The authors mention they analyzed several neuronal markers, which they show in Figure 1B. Not all markers tested are for neurons. In reality, they probed the gene expression for 3 canonical markers of iOSNs (Gap43), HBCs (Krt5) and mOSNs (Gnal). Please correct the main text accordingly. Also, since the authors performed RNA-seq on the whole OE, the RNA-seq gene expression estimates for the canonical markers of the main cell types populating the OE should be shown (I suggest plotting something similar to similar to Figure 1C in PMID: 31392275).

With regard to the qPCR analysis of whole OE, we corrected the text to say: “we found that olfactory neural markers (Gap43, Krt5, and Gnal) exhibited significantly decreased expression in Upf3b-null as compared to control OE (Figure 1B)”. With regard to the RNA-seq analysis, this was done on purified mOSNs, not whole OE – we have revised the manuscript to make this clearer. We showed in Figure 2A in the revised manuscript that the expression of the mOSN marker gene, Omp, is enriched in these mOSN samples, and that the non-mOSN markers genes, Krt5 and Lgr5, are de-enriched in these samples.

4.1) In this context, since Omp is the canonical marker for all mOSNs subtypes, I am puzzled as to why the authors used Gnal instead. This is especially relevant because they show the RQ values for Omp in Figure 1—figure supplement 1C, but no significant difference was observed between WT and KO, which does not fit with the results from the other markers used in Figure 1B. Please explain.

In the originally submitted manuscript, we showed Omp mRNA and OMP protein expression in the same figure (Figure 6—figure supplement 1C, D in the revised manuscript) to convey – in one place – the evidence that OMP expression was not significantly changed in Upf3b-null mice. If deemed important, we could move this data to a main figure. A likely explanation for why Omp and Gnal respond differently in Upf3b-null OE is these mOSN markers label different mOSN sub-populations.

4.2) Also, it is clear from the immunochemistry experiments in Figure 6G that the layer containing OMP+ cells is thicker in WT vs. KO mice, suggesting that a higher number of OMP+ cells exist in WT mice (in line with the hypothesis that KO mice have reduced olfactory abilities, possibly caused to lower numbers of mOSNs). The authors should check if OMP+ cells are more highly prevalent throughout the whole OE in WT vs. KO, or if this decrease is zonally restricted. A counting of OMP+ per mm² would help convince me and clarify this topic.

We thank reviewer 2 for noticing this. The image in Figure 6G in our originally submitted manuscript is not representative with respect to the typical width of the mOSN layer we observed in Upf3b-null OE. We provide representative images from three individual mice from each genotype in Figure 6—figure supplement 2B in the revised manuscript. Quantification showed that OMP+ cell density is not significantly between the two genotypes.

5) The authors mention: "Despite the reduction in GBCs and iOSNs, Upf3b-null mice had normal numbers of mOSNs (53% vs. 47% in the control) (Figure 6A). Despite the reduction in GBCs and iOSNs, Upf3b-null mice had normal numbers of mOSNs (53% vs. 47% in the control) (Figure 6A). Consistent with this, the mOSN marker, OMP, was similarly expressed (at both the RNA and protein levels) in OE from Upf3b-null and control mice (Figure 1—figure supplement 1C, D)". How can the authors exclude the hypothesis that potential biases caused by dissection and dissociation (see point 2 above), are the cause for these differences? Do the authors have statistics like shown in Figure 6A for individual mice, or were all the WT and KO samples pulled together in a single experiment for each genotype?

Figure 6—figure supplement 1B shows the fraction of HBCs, GBCs, iOSNs, and mOSN in the individual samples we analyzed by scRNA-seq analysis. With regard to mOSNs, we found their representation differs to some extent between each sample and there is neither an obvious nor a statistical difference (p = 0.91) between the 4 Upf3b-null and 4 control samples when compared to all OSNs (Figure 6—figure supplement 1A). The same was the case when we determined the fraction of mOSNs per all OE cells (p = 0.44, Figure 6—figure supplement 1B). These results are in agreement with our finding that there was no statistical difference in OMP protein expression as determined by Western, which was analyzed in OE from different mice than those used for scRNA-seq, thereby increasing confidence in the result (Figure 6—figure supplement 1D in our revised manuscript). That said, we cannot rule out that small differences in the dissection and/or cell dissociation used to generate the samples obscured a subtle change in the fraction of mOSNs in the 2 genotypes. Given that reviewer 2 has considerable experience with scRNA-seq analysis of the OE, they can perhaps appreciate the challenges in this experiment, including that cell isolation must be done extremely quickly, thereby precluding pooling OE from several mice to reduce heterogeneity (by contrast, we pooled OE from 3 mice per sample for RNA-seq analysis). In the revised manuscript, we now explicitly state this as a caveat in our revised manuscript: “However, we cannot rule out that the variability among the 4 samples for each genotype might have obscured a subtle change in the fraction of these other OSN stages in Upf3b-null mice. This variability might either be the result of biological differences between individual mice or differences in dissection and/or cell dissociation”. We also did not observe a statistically significant difference in either GBCs or iOSNs: analysis of the individual samples showed no statistical difference between Upf3b-null and control mice, when compared to all OSNs (p = 0.49 and 0.90,respectively, Figure 6—figure supplement 1A) or compared to all OE cells (p = 0.59 and 0.94, respectively, Figure 6—figure supplement 1B). We previously suggested an apparent shift in the frequency of both GBCs and iOSNs in the KO, based on a striking shift in their average frequency, but this did not hold up to statistical scrutiny.

With regard to HBCs, statistical analysis showed a significant decrease in their frequency in Upf3b-null mice, either when compared to all OSNs (p = 0.04, Figure 6—figure supplement 1A) or compared to all OE cells (p = 0.01, Figure 6—figure supplement 1B). To test the validity of this, we performed IHC staining with the HBC marker, TRP63, and found that the density of TRP63+ cells was significantly less in Upf3b-null OE than WT OE (Figure 6—figure supplement 1A ). We have revised the manuscript in light of these new results and analyses.

6) The authors mention that in Figure 6G "immunofluorescence analysis showed that an anti-CAMP antiserum showed a strong signal in the OE, as well as the lamina propria". While the presence of CAMP staining in the lamina propria is clear, I cannot distinguish the signal in the OE. Please provide evidence of this or rephrase.

We agree that the anti-CAMP staining is modest in mOSNs. In the revised manuscript, we edited the description of this data as follows: “As further evidence, immunofluorescence analysis detected modest staining in cells in the OE, as well as strong staining in the in the lamina propria, both of which were increased in Upf3b-null mice (Figure 6G and Figure 6—figure supplement 2B)”.

7) It is critical the authors should validate (at least) a subset of the key differentially expressed genes (namely some Olfrs) identified in their RNA-seq analyses. This is important to confirm the validity of the RNA-seq data and refute the possibility that the differentially expressed genes identified are due to biases in dissection and/or dissociation. This validation could be done with immunohistochemistry or with in-situ hybridization, followed by counting of the positive cells across the whole OE in multiple animals and the two genotypes.

We validated our RNA-seq analysis by two independent approaches. We validated using qPCR analysis, as described in answer to reviewer 1, critique 1 (Figure 1—figure supplement 2). For validation at the protein level, we elected to test FUT10, as commercial antibodies were available against this protein and it is an interesting factor that is known to promote maintenance of neural stem cells in an undifferentiated state (Kumar et al., 2013). We performed immunofluorescence analysis on FUT10 expression in OE from 3 mice from both genotypes. We co-stained with the mOSN marker, OMP, to label mOSNs. The results showed that the anti-FUT10 antibody signal is broadly decreased in Upf3b-null OE, including in Upf3b-null mOSNs (Figure 1—figure supplement 2C). This result is consistent with decreased Fut10 RNA expression in Upf3b-null mOSNs, as determined by RNA-seq (log2FC=-0.87).

8) In its current form, the manuscript is not easy to follow. In my opinion, this is partly caused by the structure and language used in several sections of this manuscript, and partly because of the way data is presented in the main figures. In the revised version, the authors should consider simplifying the language to make it more accessible and improving the structure and the way the data is presented, as means to improve clarity (I give several examples on this topic below).

We thank reviewer 2 for bringing up this concern, which we have addressed by making a considerable effort to revise wording throughout the entire manuscript (note: alterations that do not impact science content are not marked with the editing program). As part of this, we revised the introductory sentences for some sections to make the new topic more clear, and when necessary, to transition from the previous topic. Finally, we simplified the description of the OE cell clusters and associated molecular pathways identified by our scRNA-seq analysis, as suggested by reviewer 3.

Reviewer #3:

[…]

Specific comments:

1) The authors should consider changing the title, maybe using a more specific term such as “olfactory epithelium neurogenesis”, as olfactory neurogenesis might be confused with the neurogenesis ongoing in olfactory bulb in adulthood (constant neuron production for olfactory bulb from SVZ). Same rewording should be applied to throughout the text as “olfactory neurogenesis” is broadly used in the manuscript.

We elected to broaden the title of our manuscript to: “The role of the NMD factor UPF3B in olfactory sensory neurons.” In addition, we no longer use the term “olfactory neurogenesis” in the revised manuscript.

2) The authors found that 15 previously defined likely mouse NMD target mRNAs (based on upregulation and/or stabilization in response to NMD factor depletion or high UPF1 occupancy) overlapped with mRNAs upregulated in Upf3b-null mOSNs: Atp10d, Lbh, Slc38a6, Tgm2, Rgl3, Notch2, Ywhag, Luc7l, Ptch1, 1700025G04Rik, Tle3, Ptprn, Ptger2, Dhps, and Msrb3. To assess the direct NMD targets, the authors measured the stability of the 127 mRNAs upregulated in Upf3b-null mOSNs using a method that infers RNA stability based on pre-mRNA and steady-state mRNA levels. This method revealed that 82 of 127 upregulated genes encode mRNAs stabilized in Upf3b-null mOSNs as compared to control mOSNs. Of these 82 stabilized and upregulated mRNAs, 52 had at least 1 of the 3 NMD-inducing features (NIFs) (1- an exon-exon junction >50 nt downstream of the stop codon, 2- an uORF, or 3- a long 3'UTR) that they examined. Thus, the authors classified these 52 mRNAs as high confidence mOSN NMD targets.

4 of the 15 previously defined NMD targets (Tle3, Ptprn, Rgl3, and Dhps) that initially overlapped with the list of 127 upregulated mRNAs, never made to the list of 52 upregulated, stabilized, and with NIFs. The authors should provide an explanation for this discrepancy. Is it because those 4 genes were identified as NMD targets solely based on their upregulation in response to NMD factor depletion in previous studies? If so, that list or at least those 4 genes should be omitted from the text. I can see that the authors were trying to do a thorough analysis by combining stability, upregulation and NMD-inducing features to determine the direct targets of NMD. However, those 15 previously defined likely mouse NMD target mRNAs are confusing.

There are many possible explanations for why 4 previously defined putative NMD target mRNAs did not make our list of high-confidence NMD targets in mOSNs. The most obvious possibility is that some or all of these 4 RNAs are not actually NMD targets. In support, none of these 4 RNAs have well-established NMD-inducing features (NIFs) established by the field. The Ptprn and Rgl3 mRNAs we found were upregulated have 3’ UTRs of only 522 and 364 nt, respectively; neither has a short upstream open reading frame (ORF); and neither has an exon-exon junction downstream of the main ORF (based on their ENSEMBL transcript IDs of ENSMUST00000027404 and ENSMUST00000045726, respectively). The other 2 mRNAs we found were upregulated are non-coding isoforms of Tle3 and Dhps (ENSMUST00000159140 and ENSMUST00000129826, respectively) and thus, by definition, lack these 3 NIFs. Furthermore, 3 of these 4 RNAs was defined as a putative NMD target mRNA based only on being upregulated in NMD-deficient mice (Author response table 1). Thus, these 3 may be indirectly regulated by NMD. The one exception, Tle3, was defined as a likely NMD target based on (i) upregulation in mouse embryonic stem cells (mESCs) depleted of the NMD factor UPF1, (ii) upregulation in mESCs treated with a protein synthesis inhibitor to block NMD, and (iii) high occupancy of the NMD factor UPF1, based on CLIPseq experiments (Hurt, Robertson, and Burge, 2013). Thus, Tle3 mRNA may be a bona fide NMD target in mESCs. One explanation for why we did not find that Tle3 mRNA is significantly stabilized in Upf3b-null mOSNs is that it is targeted by NMD in a cell type- or NMD-factor-specific manner. In support of this, we previously showed that some RNAs downregulated by NMD in some tissues are not downregulated by NMD in other tissues (Huang et al., 2011). In addition, we found that NMD targeting also depends on the NMD branch involved, based on knockdown or knockout of different NMD factors (Huang et al., 2011).

To address the above issue, we now report overlap analysis of our high-confidence NMD target mRNAs (rather than all Upf3b-null-upregulated mRNA) with previously identified candidate NMD targets: “We found that 11 of these previously defined likely mouse NMD target mRNAs overlapped with the 52 high-confidence targets identified in our study: Atp10d, Lbh, Slc38a6, Tgm2, Notch2, Ywhag, Luc7l, Ptch1, 1700025G04Rik, Ptger2, and Msrb3.”.

Author response table 1. Previously identified putative NMD target mRNAs not in our mOSN high-confidence NMD target mRNA list.

3) Presence of uORF and an exon-exon junction >50 nt downstream of the stop codon are well-established NIFs. Is a long 3' UTR alone a well-established NIF? Although the authors used a stability analysis, they should be careful with their interpretation of direct NMD targets, especially with the exon-length criteria. The majority of the direct-target list (34 mRNAs of 52 mRNAs) did not have either a uORF or an exon-exon junction >50 nt downstream of the stop codon, but just a long 3'UTR. They should either provide additional validation for these 34 mRNAs or rephrase the relevant part. Intriguingly, 8 of the previously "defined" 15 NMD-target mRNAs mentioned above were among the 34 mRNAs with long 3'UTRs, while only 3 mRNAs with uORF were common in both lists. Were those 8 mRNAs found to have high UPF1 occupancy or stability upon depletion of a different NMD factor than UPF3B? Because this could provide additional support that they could be direct NMD targets.

There is extensive literature demonstrating that a long 3’UTR is sufficient to elicit NMD, including the following papers (Ge, Quek, Beemon, and Hogg, 2016; Hogg and Goff, 2010; Kishor, Ge, and Hogg, 2019). Therefore, we consider the 34 genes encoding mRNAs with long 3’UTR that are upregulated in Upf3b-null mOSNs as strong candidates to be NMD targets. As reviewer 3 mentioned, a large number of the NMD targets we identified in mOSNs have long 3’UTRs, which raises the possibility that mOSNs have a predilection for degrading mRNAs with this particular NIF, as we now mention in the manuscript. We also note that “By analogy, evidence suggests that male germ cells preferentially target mRNAs harboring long 3’UTRs for destruction by NMD (Bao et al., 2016).”

As reviewer 3 noted, 8 of the 15 genes overlapping between our high-confidence mouse mOSN NMD target list and the list we established of previously defined putative mouse NMD targets encode mRNAs with long 3’UTRs. As shown in Author response table 2, all of these 8 mRNA were defined on the basis of being upregulated in NMD-deficient tissues or cell lines. In most cases, the core NMD factor gene, Upf2, was conditionally ablated in the tissue or cell type under study. In two cases, the NMD gene, Smg1, was mutated. Thus, all were defined on the basis of inactivation of a NMD factor gene different than the one we studied – Upf3b – thereby increasing confidence that these are bona fide NMD target mRNAs. We have not brought up the above discussion of these 8 genes in the revised manuscript, but we can add this if deemed important. Our revised manuscript is currently ~30% over the length limit.

Author response table 2. High-confidence NMD targets expressed in mOSNs harboring long 3’UTRs that overlap with putative NMD target mRNAs identified in other cell types by previous studies.

4) The four subtypes in olfactory epithelium, and their trajectory (HBC → GBC → iOSN → mOSN) are well established. However, the authors spent more than 2 pages in describing OE cell subsets and molecular pathways active during olfactory neurogenesis. Although Upf3b-null mice have reduced numbers of HBCs, more GBCs, and more iOSNs, Upf3b-null mice still had normal numbers of mOSNs (53% vs. 47% in the control). Because the authors focused only in OSNs when identifying direct targets of NMD (due to the OSNs being the functional unit in OE), the part describing other OE cell subsets and molecular pathways active during olfactory neurogenesis should be more concise.

We agree we were too verbose and thus we have substantially reduced the length of this section.

5) Table S1 has overwhelming data (5 excel sheets), and it is hard to pull out the antimicrobial genes from 12,000 genes. How many are they? Are they more than what is shown in Figure 6E? If so, it should be shown in a table (similar format to table 1).

We thank reviewer 3 for this suggestion. We separated original Table S1 into three supplementary tables, with the RNA-seq and translome data as one supplementary table, scRNA-seq data as another table, and regulated immune genes as another table. We also added a table (Supplementary file 5 in the revised manuscript) listing all enriched immune-associated genes in different Upf3b-null OSN sub-clusters.

6) The authors suggest a model that UPF3B protein promotes the selection of the downregulated 78 Olfr genes to be the dominant Olfr gene expressed in individual mOSNs. Is there a transcription factor or other kind of transcriptional regulator that positively regulates Olfr genes and is downregulated in KO mOSNs? Did any known repressor of Olfr gene expression show an increased expression in KO mOSNs?

To determine whether there are either direct or indirect NMD target mRNAs encoding transcriptional regulators that might act in such a circuit, we screened genes exhibiting significantly altered expression in Upf3b-null OSNs for those that encode factors known to regulate Olfr gene expression or have binding sites in Olfr promoters (Clowney et al., 2011; Dalton, Lyons, and Lomvardas, 2013; Hirota and Mombaerts, 2004; Markenscoff-Papadimitriou et al., 2014; McIntyre, Bose, Stromberg, and McClintock, 2008; Michaloski, Galante, and Malnic, 2006; Wang, Tsai, and Reed, 1997). While our screen did not identify any downregulated genes that fulfilled the above criteria, our screen did identify 2 upregulated genes – Mafg and Irf8 – that fulfilled our criteria. Because these 2 genes are upregulated in Upf3b-null mOSNs, they may be direct NMD target mRNAs. Intriguingly, Mafg and Irf8 both encode transcriptional repressors (Igarashi et al., 1994; Salem et al., 2014) that bind O/E consensus sites found in most Olfr gene promoters (Michaloski et al., 2006). Thus, Mafg and Irf8 are candidates to act directly downstream of NMD in a regulatory circuit that suppresses the transcription of these 78 Olfr genes. Mafg is a member of the Maf subfamily of basic leucine-zipper transcription factor genes that encode small proteins containing a b-zip DNA-binding domain, but lack a transactivation domain, and thus members of this family dimerize to form transcriptional repressors (Igarashi et al., 1994). MAFG is best known for its ability to regulate globin transcription in erythroid cells; our results raise the possibility that MAFG also functions in OSNs to regulates Olfr genes. IRF8 regulates the development hematopoietic cells; its expression in OSNs raises the possibility that this transcription factor also functions in OSNs.

Together, these findings support a model in which (i) IRF8 and MAFG normally subtly repress the transcription of a subset of Olfr genes in OSNs to fine tune their expression and (ii) IRF8 and MAFG are encoded by NMD target mRNAs, so when NMD is disrupted, these transcriptional repressors are overexpressed, leading to reduced expression of their Olfr gene targets in developing OSNs. Thus, NMD deficiency would be expected to reduce the probability that these particular ^genes will be chosen to be the “dominant Olfr gene” in individual mOSNs, which is precisely what we observed in Upf3b-null mice. In the revised manuscript, we have posited this model and provided the supporting evidence (a brief version of what is described above) in the Discussion.

We note it is also possible that Upf3b dictates the selection of Olfr genes by influencing olfactory epithelial neurogenesis. In support, several genes we found to be UPF3B-regulated have been reported to play essential roles in neurogenesis, including Lrp2, Hk2, Notch2, Gdf11, Fos, Ptch1, Spry2, and Cwc22, which we mention in the revised manuscript. Finally, it is possible that Upf3b differentially influences the survival of OSNs bearing different OLFRs. Consistent with this, we found the Fos is up-regulated in Upf3b-null mOSNs. It has been reported that the induction of Fos in OSNs cause cell apoptosis (Michel, Moyse, Brun, and Jourdan, 1994).

7) The scRNAseq data shows that the Upf3b-null OEs lack mOSN-2 sub-cluster. Could the downregulation of 78 Olfr genes in KO mOSNs be a reflection of this? Maybe, those 78 genes are predominantly expressed in the mOSN-2 sub-cluster. The authors should discuss this.

This is an interesting hypothesis. It predicts that some of these 78 Olfr genes will be amongst the genes significantly differentially expressed in the mOSN-2 sub-cluster as compared to the other mOSN sub-clusters. However, refuting this hypothesis, no Olfr genes are in this list, as shown in Supplementary file 4 in the revised manuscript. Instead genes enriched in the mOSN-2 sub-cluster include Pten, App, Cnga2, Nrp2, Ncam1, Adcy3, Gnal, Atf5, and Gfy, all of which are known to be essential for olfactory epithelium development and/or olfaction.

8) The Western blot of CAMP looks convincing. However, it is not specific to the OSNs. The CAMP signal is not clear in mOSNs in immunohistochemistry images. It looks like that CAMP was upregulated in the entire OE, signal being weak in mOSNs compared to other cell types. Showing CAMP staining only with DAPI in another panel (no merging with OMP staining) would work.

We agree that the anti-CAMP staining in mOSNs is modest. In the revised manuscript, we edited our description of this data to say: “As further evidence, immunofluorescence analysis detected modest staining in cells in the OE, as well as strong staining in the in the lamina propria, both of which were increased in Upf3b-null mice (Figure 6G and Figure 6—figure supplement 2B).”. Per reviewer 2 request, we have also shown the same image with only the CAMP and DAPI signal (no OMP) (Author response image 1).

Author response image 1. — Nuclei were stained with DAPI (blue).

9) In addition to Olfr genes, the authors discover that UPF3B in OSNs also regulates antimicrobial genes. Unlike the Olfrs, the antimicrobial genes are upregulated in KO mOSNs. Similar to comment #6, is any of the direct NMD targets identified in Figure 2 known to trigger inflammation or antimicrobial genes? The authors should offer a mechanism/s as to how NMD indirectly regulates these major sets of genes. Their finding of another major class of genes is regulated by UPF3B suggests that the UPF3B prortein has independent functions within OSNs, one being suppression of OE inflammation. Or, can loss of Olfr genes and an increase in immune response be connected? Can one be responsible for the other?

We thank reviewer 3 for bringing this up, as we neglected to point out that a remarkably large number of the upregulated mRNAs encoding anti-microbial and other immune proteins have the potential to be direct NMD targets. We found that 48 out of 88 mRNAs encoding immune-related proteins that are upregulated in Upf3b-null OSNs harbor at least one NIF (Supplementary file 5 in the revised manuscript). Thus, these 48 mRNAs are good candidates to be direct targets of the NMD pathway, as we point out in the revised manuscript.

With regard to indirect regulation, the most promising candidate we found to mediate such regulation is NOTCH2, which we identified as being encoded by a high-confidence NMD target mRNA (Figure 2 in the revised manuscript). NOTCH2 has been shown to regulate the expression of many genes encoding inflammatory mediators and antimicrobial proteins, as reviewed in Shang et al. (Shang, Smith, and Hu, 2016). In addition, two other high-confidence NMD targets – Bhlhe40 and Rac2 – encode immune system regulators. These candidate regulatory circuit are discussed in the revised manuscript.

Reviewer 3 also asks whether the increased expression of anti-microbial genes in Upf3b-null OSNs might somehow be causally connected with the reduced numbers of mOSNs expressing specific OLFRs in Upf3b-null mice. One possibility is that some of the 78 OLFRs less represented in Upf3b-null mice bind to odorants that normally suppress immune activation. It is also possible is that these 78 OLFRs are underrepresented in Upf3b-null OSNs because loss of Upf3b increases the likelihood of apoptosis of these cells, which, in turn induces anti-microbial genes in surviving OSNs. We have not mentioned either of these scenarios in the revised manuscript, as we do not consider them very likely. However, if deemed important, we would be happy to discuss these or other models in a subsequent revision.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Reviewer #2:

[…]

1) No more comments. This concern is addressed.

2.1) The authors then present "two lines of evidence against a significant subsampling bias in their samples: First, each sample was pooled from 3 mice. Second, most samples had a similar percentage of OMP-YFP+ cells (between 5-7%)".

Unfortunately, without knowing the numbers of cells sorted in each sample, I am failing to see how any of these two reasons argue against it. The olfactory epithelium of young adult mice (8-10 weeks) is populated on average by ~10 million mature OSNs (mOSNs). If each sample contains on average 10,000 sorted cells, that would capture only 0.1% of all mOSNs. If each sample contains 100,000 sorted cells, that would still only make up 1% of all mOSNs, and so on. Since the authors "did not make note of the output cell number" in each sample, how could pooling animals or sorting similar percentages of cells between animals be used as an argument against subsampling? Also, only 4/8 of their samples had between 5-7% of sorted OMP-YFP+ cells, which is not "most samples". I am assuming the authors kept the FACS files, and thus should be able to retrieve those values. Is this the case?

The FACS purification of mOSNs (YFP+ cells) was done by a graduate student, Samantha Jones, who left the laboratory two years ago. Her FACS files only recorded the number of cells used to make the plots we provided in our manuscript (Figure 1—figure supplement 1B). Unfortunately, these FACS files did not indicate the total number of cells sorted and, thus, as we indicated in the previous rebuttal letter, we are sorry but we cannot provide this information. That said, the graduate remembers that she sorted approximately one-half of the dissociated cells that she obtained from dissected olfactory epithelium. While this means that there was some subsampling (as only ~½ of the cells was sorted), nonetheless, a substantial number of mOSNs (YFP+ cells) must have been purified for RNA-seq analysis.

Despite this subsampling, there are several reasons we believe our findings from these FACS-purified cells is valid. First, dissociated OE cells from 3 mice were pooled together to decrease sampling bias (also indicated in our previous rebuttal letter as noted by reviewer 2). This was done to reduce variation caused by differences in individual mice, as well as differences in dissection that might lead to differential recovery of cells in the different “zones” of the olfactory epithelium. In total, we analyzed mOSNs from 24 mice (3 mice x 4 samples x 2 genotyes = 24) for RNA-seq analysis. Second, given that single-cell suspensions were used for FACS sorting (a method that randomly sorts cells), it is unlikely that subsampling would lead to a consistent bias between different samples. Third, confidence in our results comes from the finding that our RNA-seq analysis showed consistent regulation of most of the 78 Olfr genes (in several independent Upf2-null and control mOSN samples; Figure 1C in the revised manuscript). Fourth, to test the veracity of our RNA-seq analysis, we performed qPCR analysis on 3 of these Olfr genes, which verified the regulation of all 3 (Figure 1—figure supplement 2B in the revised manuscript; of note, this figure also shows data from non-Olfr genes whose regulation we also verified; this was presented in the previous rebuttal letter). Finally, our single-cell RNA-seq analysis also verified the regulation of these 78 Olfr genes (Figure 6H in the revised manuscript). Of note, these OE samples were different than those analyzed by RNA-seq.

In the revised manuscript, we briefly bring up these issues as follows: “A caveat is the OE contains zones enriched for mOSNs expressing particular sets of OLFRs (Miyamichi, Serizawa, Kimura, and Sakano, 2005; Ressler, Sullivan, and Buck, 1994), and thus even though we made an effort to dissect the entire OE for RNA-seq analysis, it is possible that there is zonal heterogeneity in the samples we analyzed. […] Furthermore, our single-cell RNA-seq analysis (which analyzed samples different from those analyzed by RNA-seq) also verified the regulation of these 78 Olfr genes (Figure 6H).”.

Finally, what is the logic underlying the following statement: "We note that our manuscript is currently ~30% over the word limit."?

The eLife author guidelines states that “…we suggest that authors try not to exceed 5,000 words in the main text…” Our manuscript has >7,000 words.

2.2) The authors here claim that they asked, "what genes are regulated by a particular factor in mOSNs" and that "the depth of RNA-seq analysis we used was sufficient to draw several conclusions". One of the unexpected major conclusions of the authors is that "over half of the downregulated genes in Upf3-null mOSNs are Olfr genes" and state that they "regard this finding as a robust result, as the RNA-seq coverage for Upf3b-null and control OSNs was 95.8% and 96.0%, respectively, of all annotated Olfr gene".

Since Olfr and many other genes are expressed zonally in the olfactory epithelium (PMID: 7812149, 15814789, 12709059, 32209480, etc), differences in dissection between animals could easily account for the differences observed in the RNA-seq data from sorted cells. Thus, showing the gene expression distribution of all Olfr genes for each of the sorted samples is a key QC step for this study and it should feature early on in the main figures. The authors should include that in the main text and present a plot containing the mean gene expression values (in log10) for all annotated mouse Olfr in either descending or ascending order for the WT and KO, include similar plots in supplementary for each individual sample and include all the gene names (and not just Ensembl IDs) in the “TPM” sheet of Supplementary file 1 to make this accessible to the community. It would also be very useful to present in supplementary a Venn diagram comparing all the Olfrs expressed in WT vs KO mice.

To address this issue, we examined the expression of all Olfr genes in each of our sorted samples (Figure 1—figure supplement 2E in the revised manuscript). This heatmap shows that the individual samples do not show an obvious bias in their expression pattern.

As requested, we also generated a plot showing the Log10-transformed TPM values (mean values) for all Olfr genes (Supplementary file 1 in the revised manuscript). A gene list with all annotated Olfr genes (in ascending order) was added to the Supplementary file 1. As requested, we also provide a Venn diagram showing the expression pattern of all Olfrs, including the few uniquely expressed in Upf3b-null or control mOSNs (Author response image 2). Author response table 3 lists the expression pattern of all these uniquely expressed Olfrs in all samples. This table shows that Olfrs expressed only in Upf3b-null or control mOSNs tend to be lowly expressed; none of these are among the 78 Olfr genes we found are statistically downregulated in Upf3b-null mOSNs (they were filtered out because of low read count).

Author response table 3. Olfr genes uniquely expressed in Upf3b-null and control mOSNs, as determined by our RNA-seq analysis. Reads count are shown.

3) The Materials and methods are now much improved. Only one minor comment here: the authors should include the commercial source of the coyote and bobcat urines.

In the revised manuscript, we provided the commercial source of the coyote urine (Snow Joe) and bobcat urines (Predator Pee).

4) As I mentioned in my original comment, as Gap43 and Krt5 are not expressed in mature OSNs, they should not be called “olfactory neuronal markers”. Instead, Gap43 is a canonical marker for immature neurons and Krt5 a canonical marker for horizontal basal cells.

We have made these corrections.

In line with this, the authors do mention that "mOSN marker gene, Omp, is enriched in these mOSN samples, and that the non-mOSN markers genes, Krt5 and Lgr5, are de-enriched in these samples". Finally, the authors mention that their RNA-seq analysis was done "purified mOSNs, not whole OE". I was indeed aware of this, as it is was already explicit in the manuscript. I suggested plotting something similar to Figure 1C in PMID: 31392275, for 2 main reasons: i) it summarizes in one plot the gene expression levels for the canonical markers for all major cell types populating the olfactory epithelium, and ii) it visually allows the reader assess the levels of “contamination/hitchhiking” by other cell types during the sorting procedure.

In Figure 1—figure supplement 2D we provide a heatmap showing the expression of canonical markers for all major cell types. Most mOSN markers are highly expressed in all 8 samples, and the other markers are more lowly expressed in all 8 samples.

Please correct this by specifically stating the cell types that those genes are indicative of (also, by “de-enriched” I assume the authors mean “depleted”?).

In the revised manuscript, we have corrected these descriptions. We have also changed “de-enriched” to “depleted.”

4.1) The Omp expression data should feature in Figure 1B, as it is the best canonical marker for mOSNs (also see previous response). It is well established that Omp and Gnal are expressed ubiquitously in the majority of mOSNs across all the olfactory epithelium, so I do not see how that could be an explanation.

We moved the Omp expression data to Figure 1B. While we agree that it is well established that Omp and Gnal are ubiquitously expressed in the majority of mOSNs, these two markers may be differentially affected by loss of Upf3b, thereby explaining our results. As support, we found that Upf3b loss leads to almost complete loss of one mOSN sub-cluster and acquisition of another mOSN sub-cluster, based on our scRNA-seq analysis. The differential expression of Omp and Gnal could also be because Upf3b influences the expression of one of these genes.

4.2) No more comments. This concern is addressed.

5) Regarding Figure 6—figure supplement 1B in the revised manuscript: HBCs and GBCs are not OSNs, and it is incorrect to refer them as such. Only immature and mature OSNs should be called OSNs. Please correct this in the figures and throughout the text. Also, on the right-hand panel of Figure 6—figure supplement 1B it would be useful to list all the other cell types that make up 100%.

We thank reviewer 2 for noting this and have made the requisite corrections. We certainly appreciate that HBCs and GBCs are OSN precursors and not yet functional OSNs. With regard to Figure 6—figure supplement 1B, we have made the requested alteration.

6) No more comments. This concern is addressed.

7) Olfr genes are expressed zonally in the olfactory epithelium (PMID: 7812149, 15814789, 12709059, 32209480, etc) and differences in dissection between animals could easily translate into differential expression artifacts. Moreover, different mouse strains express different Olfrs at different levels (PMID: 28438259). Since the authors i) have not yet convincingly addressed my previous concerns about a potential bias in dissection (see above), ii) the mice used in this studies were on mixed or different genetic backgrounds, and iii) one of the major claims of the authors is that half of the differentially expressed genes are Olfrs; the authors should validate at least their top 2 differentially expressed Olfrs across the two genotypes.

We apologize for neglecting to indicate in the previous submission that all mice used for the experiments described herein were backcrossed to C57BL/6 for at least 8 passages. We have added this information to the Materials and methods. With regard to the validation requested, qPCR analysis validated the regulation of 3 of 3 Olfr genes that we tested, as described above (Figure 1—figure supplement 2B).

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Citations

Tan K, Jones SH, Wilkinson MF. 2020. The role of the NMD factor UPF3B in olfactory sensory neurons. NCBI Gene Expression Omnibus. GSE146043 [DOI] [PMC free article] [PubMed]
Hurt JA, Robertson AD, Burge CB. 2013. Global Analyses of UPF1 Binding and Function Reveal Expanded Scope of Nonsense-Mediated mRNA Decay. NCBI Gene Expression Omnibus. GSE41785 [DOI] [PMC free article] [PubMed]
Mao H, McMahon JJ, Tsai YH, Wang Z, Silver DL. 2016. Haploinsufficiency for Core Exon Junction Complex Components Disrupts Embryonic Neurogenesis and Causes p53-Mediated Microcephaly. NCBI Gene Expression Omnibus. GSE85576 [DOI] [PMC free article] [PubMed]
Mooney CM, Jimenez-Mateos EM, Engel T, Mooney C, Diviney M, Venø MT, Kjems J, Farrell MA, O'Brien DF, Delanty N, Henshall DC. 2017. RNA sequencing of synaptic and cytoplasmic Upf1-bound transcripts supports contribution of nonsense-mediated decay to epileptogenesis. RNA sequencing of synaptic and cytoplasmic Upf1-bound transcripts supports contribution of nonsense-mediated decay to epileptogenesis. [DOI] [PMC free article] [PubMed]
Bao j, Vitting-Seerup K, Waage J, Tang C, Ge Y, Porse BT, Yan W. 2016. UPF2-Dependent Nonsense-Mediated mRNA Decay Pathway Is Essential for Spermatogenesis by Selectively Eliminating Longer 3'UTR Transcripts. NCBI Gene Expression Omnibus. GSE55180 [DOI] [PMC free article] [PubMed]
McIlwain DR, Pan Q, Reilly PT, Elia AJ, McCracken S, Wakeham AC, Itie-Youten A, Blencowe BJ, Mak TW. 2010. Smg1 is required for embryogenesis and regulates diverse genes via alternative splicing coupled to nonsense-mediated mRNA decay. Smg1 is required for embryogenesis and regulates diverse genes via alternative splicing coupled to nonsense-mediated mRNA decay. [DOI] [PMC free article] [PubMed]
Weischenfeldt J, Damgaard I, Bryder D, Theilgaard-Mönch K. Thoren LA, Nielsen FC, Jacobsen SEW, Nerlov C, Porse BT. 2008. NMD Is Essential for Hematopoietic Stem and Progenitor Cells and for Eliminating By-Products of Programmed DNA Rearrangements. NMD Is Essential for Hematopoietic Stem and Progenitor Cells and for Eliminating By-Products of Programmed DNA Rearrangements. [DOI] [PMC free article] [PubMed]
Thoren LA, Nørgaard GA, Weischenfeldt J, Waage J, Jakobsen JS, Damgaard I, Bergström FC, Blom AM, Borup R, Bisgaard HC, Porse BT. 2010. UPF2 Is a Critical Regulator of Liver Development, Function and Regeneration. UPF2 Is a Critical Regulator of Liver Development, Function and Regeneration. [DOI] [PMC free article] [PubMed]

Supplementary Materials

Supplementary file 1. Quality control matrices of Upf3b-null and control mOSNs RNA-seq datasets.

elife-57525-supp1.xlsx^{(26.4MB, xlsx)}

Supplementary file 2. mOSN transcriptome, translome, and NMD targets.

elife-57525-supp2.xlsx^{(4MB, xlsx)}

Supplementary file 3. Putative mouse NMD target RNAs defined by previous studies.

elife-57525-supp3.xlsx^{(177.2KB, xlsx)}

Supplementary file 4. Genes exhibiting enriched expression in scRNA-seq-defined OE cell subsets.

elife-57525-supp4.xlsx^{(2.2MB, xlsx)}

Supplementary file 5. Immune genes expressed in HBCs, GBCs, iOSNs, and mOSNs.

(1) Immune genes enriched in Upf3b-null HBC, GBC, iOSN, and mOSN sub-clusters identified by scRNA-seq analysis. (2) NIFs present in immune genes expressed in Upf3b-null OSNs.

elife-57525-supp5.xlsx^{(34.8KB, xlsx)}

Transparent reporting form

elife-57525-transrepform.docx^{(245.8KB, docx)}

Data Availability Statement

Sequencing data have been deposited in GEO under accession code GSE146043.

The following dataset was generated:

Tan K, Jones SH, Wilkinson MF. 2020. The role of the NMD factor UPF3B in olfactory sensory neurons. NCBI Gene Expression Omnibus. GSE146043

The following previously published datasets were used:

Hurt JA, Robertson AD, Burge CB. 2013. Global Analyses of UPF1 Binding and Function Reveal Expanded Scope of Nonsense-Mediated mRNA Decay. NCBI Gene Expression Omnibus. GSE41785

[bib1] Alkallas R, Fish L, Goodarzi H, Najafabadi HS. Inference of RNA decay rate from transcriptional profiling highlights the regulatory programs of Alzheimer's disease. Nature Communications. 2017;8:909. doi: 10.1038/s41467-017-00867-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] Alrahbeni T, Sartor F, Anderson J, Miedzybrodzka Z, McCaig C, Müller B. Full UPF3B function is critical for neuronal differentiation of neural stem cells. Molecular Brain. 2015;8:33. doi: 10.1186/s13041-015-0122-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib3] Bao J, Vitting-Seerup K, Waage J, Tang C, Ge Y, Porse BT, Yan W. UPF2-Dependent Nonsense-Mediated mRNA decay pathway is essential for spermatogenesis by selectively eliminating longer 3'UTR Transcripts. PLOS Genetics. 2016;12:e1005863. doi: 10.1371/journal.pgen.1005863. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] Barrett LW, Fletcher S, Wilton SD. Regulation of eukaryotic gene expression by the untranslated gene regions and other non-coding elements. Cellular and Molecular Life Sciences. 2012;69:3613–3634. doi: 10.1007/s00018-012-0990-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib5] Bear DM, Lassance JM, Hoekstra HE, Datta SR. The evolving neural and genetic architecture of Vertebrate Olfaction. Current Biology. 2016;26:R1039–R1049. doi: 10.1016/j.cub.2016.09.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] Becht E, McInnes L, Healy J, Dutertre C-A, Kwok IWH, Ng LG, Ginhoux F, Newell EW. Dimensionality reduction for visualizing single-cell data using UMAP. Nature Biotechnology. 2019;37:38–44. doi: 10.1038/nbt.4314. [DOI] [PubMed] [Google Scholar]

[bib7] Belgrader P, Cheng J, Maquat LE. Evidence to implicate translation by ribosomes in the mechanism by which nonsense codons reduce the nuclear level of human triosephosphate isomerase mRNA. PNAS. 1993;90:482–486. doi: 10.1073/pnas.90.2.482. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] Boehm V, Haberman N, Ottens F, Ule J, Gehring NH. 3' UTR length and messenger ribonucleoprotein composition determine endocleavage efficiencies at termination codons. Cell Reports. 2014;9:555–568. doi: 10.1016/j.celrep.2014.09.012. [DOI] [PubMed] [Google Scholar]

[bib9] Bühler M, Steiner S, Mohn F, Paillusson A, Mühlemann O. EJC-independent degradation of nonsense immunoglobulin-mu mRNA depends on 3' UTR length. Nature Structural & Molecular Biology. 2006;13:462–464. doi: 10.1038/nsmb1081. [DOI] [PubMed] [Google Scholar]

[bib10] Butler A, Hoffman P, Smibert P, Papalexi E, Satija R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nature Biotechnology. 2018;36:411–420. doi: 10.1038/nbt.4096. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] Carter MS, Doskow J, Morris P, Li S, Nhim RP, Sandstedt S, Wilkinson MF. A regulatory mechanism that detects premature nonsense codons in T-cell receptor transcripts in vivo is reversed by protein synthesis inhibitors in vitro. Journal of Biological Chemistry. 1995;270:28995–29003. doi: 10.1074/jbc.270.48.28995. [DOI] [PubMed] [Google Scholar]

[bib12] Carter MS, Li S, Wilkinson MF. A splicing-dependent regulatory mechanism that detects translation signals. The EMBO Journal. 1996;15:5965–5975. doi: 10.1002/j.1460-2075.1996.tb00983.x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] Cau E, Gradwohl G, Fode C, Guillemot F. Mash1 activates a cascade of bHLH regulators in olfactory neuron progenitors. Development. 1997;124:1611–1621. doi: 10.1242/dev.124.8.1611. [DOI] [PubMed] [Google Scholar]

[bib14] Chan WK, Huang L, Gudikote JP, Chang YF, Imam JS, MacLean JA, Wilkinson MF. An alternative branch of the nonsense-mediated decay pathway. The EMBO Journal. 2007;26:1820–1830. doi: 10.1038/sj.emboj.7601628. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib15] Chang YF, Imam JS, Wilkinson MF. The nonsense-mediated decay RNA surveillance pathway. Annual Review of Biochemistry. 2007;76:51–74. doi: 10.1146/annurev.biochem.76.050106.093909. [DOI] [PubMed] [Google Scholar]

[bib16] Chen M, Reed RR, Lane AP. Chronic inflammation directs an olfactory stem cell functional switch from neuroregeneration to immune defense. Cell Stem Cell. 2019;25:501–513. doi: 10.1016/j.stem.2019.08.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] Chess A, Simon I, Cedar H, Axel R. Allelic inactivation regulates olfactory receptor gene expression. Cell. 1994;78:823–834. doi: 10.1016/S0092-8674(94)90562-2. [DOI] [PubMed] [Google Scholar]

[bib18] Clowney EJ, Magklara A, Colquitt BM, Pathak N, Lane RP, Lomvardas S. High-throughput mapping of the promoters of the mouse olfactory receptor genes reveals a new type of mammalian promoter and provides insight into olfactory receptor gene regulation. Genome Research. 2011;21:1249–1259. doi: 10.1101/gr.120162.110. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] Colak D, Ji SJ, Porse BT, Jaffrey SR. Regulation of axon guidance by compartmentalized nonsense-mediated mRNA decay. Cell. 2013;153:1252–1265. doi: 10.1016/j.cell.2013.04.056. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib20] Conti E, Izaurralde E. Nonsense-mediated mRNA decay: molecular insights and mechanistic variations across species. Current Opinion in Cell Biology. 2005;17:316–325. doi: 10.1016/j.ceb.2005.04.005. [DOI] [PubMed] [Google Scholar]

[bib21] Dalton RP, Lyons DB, Lomvardas S. Co-opting the unfolded protein response to elicit olfactory receptor feedback. Cell. 2013;155:321–332. doi: 10.1016/j.cell.2013.09.033. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib22] Dooley JL, Abdel-Latif D, St Laurent CD, Puttagunta L, Befus D, Lacy P. Regulation of inflammation by Rac2 in immune complex-mediated acute lung injury. American Journal of Physiology-Lung Cellular and Molecular Physiology. 2009;297:L1091–L1102. doi: 10.1152/ajplung.90471.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] Dostie J, Dreyfuss G. Translation is required to remove Y14 from mRNAs in the cytoplasm. Current Biology. 2002;12:1060–1067. doi: 10.1016/S0960-9822(02)00902-8. [DOI] [PubMed] [Google Scholar]

[bib24] Doty RL. Olfactory dysfunction in parkinson disease. Nature Reviews Neurology. 2012;8:329–339. doi: 10.1038/nrneurol.2012.80. [DOI] [PubMed] [Google Scholar]

[bib25] Fletcher RB, Das D, Gadye L, Street KN, Baudhuin A, Wagner A, Cole MB, Flores Q, Choi YG, Yosef N, Purdom E, Dudoit S, Risso D, Ngai J. Deconstructing olfactory stem cell trajectories at Single-Cell resolution. Cell Stem Cell. 2017;20:817–830. doi: 10.1016/j.stem.2017.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] Gajera CR, Emich H, Lioubinski O, Christ A, Beckervordersandforth-Bonk R, Yoshikawa K, Bachmann S, Christensen EI, Götz M, Kempermann G, Peterson AS, Willnow TE, Hammes A. LRP2 in ependymal cells regulates BMP signaling in the adult neurogenic niche. Journal of Cell Science. 2010;123:1922–1930. doi: 10.1242/jcs.065912. [DOI] [PubMed] [Google Scholar]

[bib27] Gallo RL, Kim KJ, Bernfield M, Kozak CA, Zanetti M, Merluzzi L, Gennaro R. Identification of CRAMP, a cathelin-related antimicrobial peptide expressed in the embryonic and adult mouse. Journal of Biological Chemistry. 1997;272:13088–13093. doi: 10.1074/jbc.272.20.13088. [DOI] [PubMed] [Google Scholar]

[bib28] Gehring NH, Kunz JB, Neu-Yilik G, Breit S, Viegas MH, Hentze MW, Kulozik AE. Exon-junction complex components specify distinct routes of nonsense-mediated mRNA decay with differential cofactor requirements. Molecular Cell. 2005;20:65–75. doi: 10.1016/j.molcel.2005.08.012. [DOI] [PubMed] [Google Scholar]

[bib29] Giorgi C, Yeo GW, Stone ME, Katz DB, Burge C, Turrigiano G, Moore MJ. The EJC factor eIF4AIII modulates synaptic strength and neuronal protein expression. Cell. 2007;130:179–191. doi: 10.1016/j.cell.2007.05.028. [DOI] [PubMed] [Google Scholar]

[bib30] Godfrey PA, Malnic B, Buck LB. The mouse olfactory receptor gene family. PNAS. 2004;101:2156–2161. doi: 10.1073/pnas.0308051100. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib31] Gong Q. Culture of mouse olfactory sensory neurons. Current Protocols in Neuroscience. 2012;3:24. doi: 10.1002/0471142301.ns0324s58. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib32] He F, Li X, Spatrick P, Casillo R, Dong S, Jacobson A. Genome-wide analysis of mRNAs regulated by the nonsense-mediated and 5' to 3' mRNA decay pathways in yeast. Molecular Cell. 2003;12:1439–1452. doi: 10.1016/S1097-2765(03)00446-5. [DOI] [PubMed] [Google Scholar]

[bib33] Hirota J, Mombaerts P. The LIM-homeodomain protein Lhx2 is required for complete development of mouse olfactory sensory neurons. PNAS. 2004;101:8751–8755. doi: 10.1073/pnas.0400940101. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib34] Hongo T, Hakuba A, Shiota K, Naruse I. Suckling dysfunction caused by defects in the olfactory system in genetic arhinencephaly mice. Neonatology. 2000;78:293–299. doi: 10.1159/000014282. [DOI] [PubMed] [Google Scholar]

[bib35] Hornstein N, Torres D, Das Sharma S, Tang G, Canoll P, Sims PA. Ligation-free ribosome profiling of cell type-specific translation in the brain. Genome Biology. 2016;17:149. doi: 10.1186/s13059-016-1005-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib36] Huang L, Lou CH, Chan W, Shum EY, Shao A, Stone E, Karam R, Song HW, Wilkinson MF. RNA homeostasis governed by cell type-specific and branched feedback loops acting on NMD. Molecular Cell. 2011;43:950–961. doi: 10.1016/j.molcel.2011.06.031. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib37] Huang L, Shum EY, Jones SH, Lou CH, Dumdie J, Kim H, Roberts AJ, Jolly LA, Espinoza JL, Skarbrevik DM, Phan MH, Cook-Andersen H, Swerdlow NR, Gecz J, Wilkinson MF. A Upf3b-mutant mouse model with behavioral and neurogenesis defects. Molecular Psychiatry. 2018;23:1773–1786. doi: 10.1038/mp.2017.173. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib38] Hurt JA, Robertson AD, Burge CB. Global analyses of UPF1 binding and function reveal expanded scope of nonsense-mediated mRNA decay. Genome Research. 2013;23:1636–1650. doi: 10.1101/gr.157354.113. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib39] Ibarra-Soria X, Levitin MO, Saraiva LR, Logan DW. The olfactory transcriptomes of mice. PLOS Genetics. 2014;10:e1004593. doi: 10.1371/journal.pgen.1004593. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib40] Igarashi K, Kataoka K, Itoh K, Hayashi N, Nishizawa M, Yamamoto M. Regulation of transcription by dimerization of erythroid factor NF-E2 p45 with small maf proteins. Nature. 1994;367:568–572. doi: 10.1038/367568a0. [DOI] [PubMed] [Google Scholar]

[bib41] Iulianella A, Stanton-Turcotte D. The hedgehog receptor Patched1 regulates proliferation, neurogenesis, and axon guidance in the embryonic spinal cord. Mechanisms of Development. 2019;160:103577. doi: 10.1016/j.mod.2019.103577. [DOI] [PubMed] [Google Scholar]

[bib42] Iwema CL, Schwob JE. Odorant receptor expression as a function of neuronal maturity in the adult rodent olfactory system. The Journal of Comparative Neurology. 2003;459:209–222. doi: 10.1002/cne.10583. [DOI] [PubMed] [Google Scholar]

[bib43] Jaffrey SR, Wilkinson MF. Nonsense-mediated RNA decay in the brain: emerging modulator of neural development and disease. Nature Reviews Neuroscience. 2018;19:715–728. doi: 10.1038/s41583-018-0079-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib44] Jarjour NN, Schwarzkopf EA, Bradstreet TR, Shchukina I, Lin CC, Huang SC, Lai CW, Cook ME, Taneja R, Stappenbeck TS, Randolph GJ, Artyomov MN, Urban JF, Edelson BT. Bhlhe40 mediates tissue-specific control of macrophage proliferation in homeostasis and type 2 immunity. Nature Immunology. 2019;20:687–700. doi: 10.1038/s41590-019-0382-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib45] Johnson JL, Stoica L, Liu Y, Zhu PJ, Bhattacharya A, Buffington SA, Huq R, Eissa NT, Larsson O, Porse BT, Domingo D, Nawaz U, Carroll R, Jolly L, Scerri TS, Kim HG, Brignell A, Coleman MJ, Braden R, Kini U, Jackson V, Baxter A, Bahlo M, Scheffer IE, Amor DJ, Hildebrand MS, Bonnen PE, Beeton C, Gecz J, Morgan AT, Costa-Mattioli M. Inhibition of Upf2-Dependent Nonsense-Mediated decay leads to behavioral and neurophysiological abnormalities by activating the immune response. Neuron. 2019;104:665–679. doi: 10.1016/j.neuron.2019.08.027. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib46] Jolly LA, Homan CC, Jacob R, Barry S, Gecz J. The UPF3B gene, implicated in intellectual disability, autism, ADHD and childhood onset schizophrenia regulates neural progenitor cell behaviour and neuronal outgrowth. Human Molecular Genetics. 2013;22:4673–4687. doi: 10.1093/hmg/ddt315. [DOI] [PubMed] [Google Scholar]

[bib47] Karam R, Lou CH, Kroeger H, Huang L, Lin JH, Wilkinson MF. The unfolded protein response is shaped by the NMD pathway. EMBO Reports. 2015;16:599–609. doi: 10.15252/embr.201439696. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib48] Karousis ED, Mühlemann O. Nonsense-Mediated mRNA decay begins where translation ends. Cold Spring Harbor Perspectives in Biology. 2019;11:a032862. doi: 10.1101/cshperspect.a032862. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib49] Kawauchi S, Kim J, Santos R, Wu HH, Lander AD, Calof AL. Foxg1 promotes olfactory neurogenesis by antagonizing Gdf11. Development. 2009;136:1453–1464. doi: 10.1242/dev.034967. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib50] Kebaara BW, Atkin AL. Long 3'-UTRs target wild-type mRNAs for nonsense-mediated mRNA decay in Saccharomyces cerevisiae. Nucleic Acids Research. 2009;37:2771–2778. doi: 10.1093/nar/gkp146. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib51] Kowalczyk MS, Tirosh I, Heckl D, Rao TN, Dixit A, Haas BJ, Schneider RK, Wagers AJ, Ebert BL, Regev A. Single-cell RNA-seq reveals changes in cell cycle and differentiation programs upon aging of hematopoietic stem cells. Genome Research. 2015;25:1860–1872. doi: 10.1101/gr.192237.115. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib52] Kozak M. Point mutations define a sequence flanking the AUG initiator Codon that modulates translation by eukaryotic ribosomes. Cell. 1986;44:283–292. doi: 10.1016/0092-8674(86)90762-2. [DOI] [PubMed] [Google Scholar]

[bib53] Kurosaki T, Popp MW, Maquat LE. Quality and quantity control of gene expression by nonsense-mediated mRNA decay. Nature Reviews Molecular Cell Biology. 2019;20:406–420. doi: 10.1038/s41580-019-0126-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib54] Lewcock JW, Reed RR. A feedback mechanism regulates monoallelic odorant receptor expression. PNAS. 2004;101:1069–1074. doi: 10.1073/pnas.0307986100. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib55] Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics. 2014;30:923–930. doi: 10.1093/bioinformatics/btt656. [DOI] [PubMed] [Google Scholar]

[bib56] Logan DW, Brunet LJ, Webb WR, Cutforth T, Ngai J, Stowers L. Learned recognition of maternal signature odors mediates the first suckling episode in mice. Current Biology. 2012;22:1998–2007. doi: 10.1016/j.cub.2012.08.041. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib57] Long AA, Mahapatra CT, Woodruff EA, Rohrbough J, Leung HT, Shino S, An L, Doerge RW, Metzstein MM, Pak WL, Broadie K. The nonsense-mediated decay pathway maintains synapse architecture and synaptic vesicle cycle efficacy. Journal of Cell Science. 2010;123:3303–3315. doi: 10.1242/jcs.069468. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib58] Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology. 2014;15:550. doi: 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib59] Lykke-Andersen S, Jensen TH. Nonsense-mediated mRNA decay: an intricate machinery that shapes transcriptomes. Nature Reviews Molecular Cell Biology. 2015;16:665–677. doi: 10.1038/nrm4063. [DOI] [PubMed] [Google Scholar]

[bib60] Lynch M, Marinov GK. The bioenergetic costs of a gene. PNAS. 2015;112:15690–15695. doi: 10.1073/pnas.1514974112. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib61] Mabin JW, Woodward LA, Patton RD, Yi Z, Jia M, Wysocki VH, Bundschuh R, Singh G. The exon junction complex undergoes a compositional switch that alters mRNP structure and Nonsense-Mediated mRNA decay activity. Cell Reports. 2018;25:2431–2446. doi: 10.1016/j.celrep.2018.11.046. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib62] Malnic B, Hirono J, Sato T, Buck LB. Combinatorial receptor codes for odors. Cell. 1999;96:713–723. doi: 10.1016/S0092-8674(00)80581-4. [DOI] [PubMed] [Google Scholar]

[bib63] Manglapus GL, Youngentob SL, Schwob JE. Expression patterns of basic helix-loop-helix transcription factors define subsets of olfactory progenitor cells. The Journal of Comparative Neurology. 2004;479:216–233. doi: 10.1002/cne.20316. [DOI] [PubMed] [Google Scholar]

[bib64] Mao H, Brown HE, Silver DL. Mouse models of Casc3 reveal developmental functions distinct from other components of the exon junction complex. RNA. 2017;23:23–31. doi: 10.1261/rna.058826.116. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib65] Markenscoff-Papadimitriou E, Allen WE, Colquitt BM, Goh T, Murphy KK, Monahan K, Mosley CP, Ahituv N, Lomvardas S. Enhancer interaction networks as a means for singular olfactory receptor expression. Cell. 2014;159:543–557. doi: 10.1016/j.cell.2014.09.033. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib66] McIntyre JC, Bose SC, Stromberg AJ, McClintock TS. Emx2 stimulates odorant receptor gene expression. Chemical Senses. 2008;33:825–837. doi: 10.1093/chemse/bjn061. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib67] Mendell JT, Sharifi NA, Meyers JL, Martinez-Murillo F, Dietz HC. Nonsense surveillance regulates expression of diverse classes of mammalian transcripts and mutes genomic noise. Nature Genetics. 2004;36:1073–1078. doi: 10.1038/ng1429. [DOI] [PubMed] [Google Scholar]

[bib68] Michaloski JS, Galante PA, Malnic B. Identification of potential regulatory motifs in odorant receptor genes by analysis of promoter sequences. Genome Research. 2006;16:1091–1098. doi: 10.1101/gr.5185406. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib69] Michel D, Moyse E, Brun G, Jourdan F. Induction of apoptosis in mouse [correction of rat] olfactory neuroepithelium by synaptic target ablation. Neuroreport. 1994;5:1329–1332. [PubMed] [Google Scholar]

[bib70] Miyamichi K, Serizawa S, Kimura HM, Sakano H. Continuous and overlapping expression domains of odorant receptor genes in the olfactory epithelium determine the dorsal/ventral positioning of glomeruli in the olfactory bulb. Journal of Neuroscience. 2005;25:3586–3592. doi: 10.1523/JNEUROSCI.0324-05.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib71] Nagy E, Maquat LE. A rule for termination-codon position within intron-containing genes: when nonsense affects RNA abundance. Trends in Biochemical Sciences. 1998;23:198–199. doi: 10.1016/S0968-0004(98)01208-0. [DOI] [PubMed] [Google Scholar]

[bib72] Nasif S, Contu L, Mühlemann O. Beyond quality control: the role of nonsense-mediated mRNA decay (NMD) in regulating gene expression. Seminars in Cell & Developmental Biology. 2018;75:78–87. doi: 10.1016/j.semcdb.2017.08.053. [DOI] [PubMed] [Google Scholar]

[bib73] Nguyen LS, Kim HG, Rosenfeld JA, Shen Y, Gusella JF, Lacassie Y, Layman LC, Shaffer LG, Gécz J. Contribution of copy number variants involving nonsense-mediated mRNA decay pathway genes to neuro-developmental disorders. Human Molecular Genetics. 2013;22:1816–1825. doi: 10.1093/hmg/ddt035. [DOI] [PubMed] [Google Scholar]

[bib74] Nguyen LS, Wilkinson MF, Gecz J. Nonsense-mediated mRNA decay: inter-individual variability and human disease. Neuroscience & Biobehavioral Reviews. 2014;46 Pt 2:175–186. doi: 10.1016/j.neubiorev.2013.10.016. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib75] Notaras M, Allen M, Longo F, Volk N, Toth M, Li Jeon N, Colak D. UPF2 leads to degradation of dendritically targeted mRNAs to regulate synaptic plasticity and cognitive function. Molecular Psychiatry. 2019;21:5. doi: 10.1038/s41380-019-0547-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib76] Palacios IM. Nonsense-mediated mRNA decay: from mechanistic insights to impacts on human health. Briefings in Functional Genomics. 2013;12:25–36. doi: 10.1093/bfgp/els051. [DOI] [PubMed] [Google Scholar]

[bib77] Peterson J, Lin B, Barrios-Camacho CM, Herrick DB, Holbrook EH, Jang W, Coleman JH, Schwob JE. Activating a reserve neural stem cell population in Vitro Enables Engraftment and Multipotency after Transplantation. Stem Cell Reports. 2019;12:680–695. doi: 10.1016/j.stemcr.2019.02.014. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib78] Qiu X, Mao Q, Tang Y, Wang L, Chawla R, Pliner HA, Trapnell C. Reversed graph embedding resolves complex single-cell trajectories. Nature Methods. 2017;14:979–982. doi: 10.1038/nmeth.4402. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib79] Ramaiah M, Tan K, Plank TM, Song HW, Dumdie JN, Jones S, Shum EY, Sheridan SD, Peterson KJ, Gromoll J, Haggarty SJ, Cook-Andersen H, Wilkinson MF. A microRNA cluster in the Fragile-X region expressed during spermatogenesis targets FMR1. EMBO Reports. 2019;20:e46566. doi: 10.15252/embr.201846566. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib80] Rebbapragada I, Lykke-Andersen J. Execution of nonsense-mediated mRNA decay: what defines a substrate? Current Opinion in Cell Biology. 2009;21:394–402. doi: 10.1016/j.ceb.2009.02.007. [DOI] [PubMed] [Google Scholar]

[bib81] Ressler KJ, Sullivan SL, Buck LB. A molecular dissection of spatial patterning in the olfactory system. Current Opinion in Neurobiology. 1994;4:588–596. doi: 10.1016/0959-4388(94)90061-2. [DOI] [PubMed] [Google Scholar]

[bib82] Riera CE, Tsaousidou E, Halloran J, Follett P, Hahn O, Pereira MMA, Ruud LE, Alber J, Tharp K, Anderson CM, Brönneke H, Hampel B, Filho CDM, Stahl A, Brüning JC, Dillin A. The sense of smell impacts metabolic health and obesity. Cell Metabolism. 2017;26:198–211. doi: 10.1016/j.cmet.2017.06.015. [DOI] [PubMed] [Google Scholar]

[bib83] Rodriguez S, Sickles HM, Deleonardis C, Alcaraz A, Gridley T, Lin DM. Notch2 is required for maintaining sustentacular cell function in the adult mouse main olfactory epithelium. Developmental Biology. 2008;314:40–58. doi: 10.1016/j.ydbio.2007.10.056. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib84] Rozenkrantz L, Zachor D, Heller I, Plotkin A, Weissbrod A, Snitz K, Secundo L, Sobel N. A mechanistic link between olfaction and autism spectrum disorder. Current Biology. 2015;25:1904–1910. doi: 10.1016/j.cub.2015.05.048. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib85] Salem S, Langlais D, Lefebvre F, Bourque G, Bigley V, Haniffa M, Casanova JL, Burk D, Berghuis A, Butler KM, Leahy TR, Hambleton S, Gros P. Functional characterization of the human dendritic cell immunodeficiency associated with the IRF8(K108E) mutation. Blood. 2014;124:1894–1904. doi: 10.1182/blood-2014-04-570879. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib86] Sanz E, Yang L, Su T, Morris DR, McKnight GS, Amieux PS. Cell-type-specific isolation of ribosome-associated mRNA from complex tissues. PNAS. 2009;106:13939–13944. doi: 10.1073/pnas.0907143106. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib87] Saraiva LR, Ibarra-Soria X, Khan M, Omura M, Scialdone A, Mombaerts P, Marioni JC, Logan DW. Hierarchical deconstruction of mouse olfactory sensory neurons: from whole mucosa to single-cell RNA-seq. Scientific Reports. 2015;5:18178. doi: 10.1038/srep18178. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib88] Saraiva LR, Riveros-McKay F, Mezzavilla M, Abou-Moussa EH, Arayata CJ, Makhlouf M, Trimmer C, Ibarra-Soria X, Khan M, Van Gerven L, Jorissen M, Gibbs M, O'Flynn C, McGrane S, Mombaerts P, Marioni JC, Mainland JD, Logan DW. A transcriptomic atlas of mammalian olfactory mucosae reveals an evolutionary influence on food odor detection in humans. Science Advances. 2019;5:eaax0396. doi: 10.1126/sciadv.aax0396. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib89] Schwob JE, Jang W, Holbrook EH, Lin B, Herrick DB, Peterson JN, Hewitt Coleman J. Stem and progenitor cells of the mammalian olfactory epithelium: taking poietic license. Journal of Comparative Neurology. 2017;525:1034–1054. doi: 10.1002/cne.24105. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib90] Serizawa S, Miyamichi K, Nakatani H, Suzuki M, Saito M, Yoshihara Y, Sakano H. Negative feedback regulation ensures the one receptor-one olfactory neuron rule in mouse. Science. 2003;302:2088–2094. doi: 10.1126/science.1089122. [DOI] [PubMed] [Google Scholar]

[bib91] Serizawa S, Miyamichi K, Sakano H. One neuron-one receptor rule in the mouse olfactory system. Trends in Genetics. 2004;20:648–653. doi: 10.1016/j.tig.2004.09.006. [DOI] [PubMed] [Google Scholar]

[bib92] Serizawa S, Miyamichi K, Sakano H. Negative feedback regulation ensures the one neuron-one receptor rule in the mouse olfactory system. Chemical Senses. 2005;30:i99–i100. doi: 10.1093/chemse/bjh133. [DOI] [PubMed] [Google Scholar]

[bib93] Shang Y, Smith S, Hu X. Role of notch signaling in regulating innate immunity and inflammation in health and disease. Protein & Cell. 2016;7:159–174. doi: 10.1007/s13238-016-0250-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib94] Shum EY, Jones SH, Shao A, Dumdie J, Krause MD, Chan WK, Lou CH, Espinoza JL, Song HW, Phan MH, Ramaiah M, Huang L, McCarrey JR, Peterson KJ, De Rooij DG, Cook-Andersen H, Wilkinson MF. The antagonistic gene paralogs Upf3a and Upf3b govern Nonsense-Mediated RNA decay. Cell. 2016;165:382–395. doi: 10.1016/j.cell.2016.02.046. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib95] Sohni A, Tan K, Song HW, Burow D, de Rooij DG, Laurent L, Hsieh TC, Rabah R, Hammoud SS, Vicini E, Wilkinson MF. The neonatal and adult human testis defined at the Single-Cell level. Cell Reports. 2019;26:1501–1517. doi: 10.1016/j.celrep.2019.01.045. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib96] Spies N, Burge CB, Bartel DP. 3' UTR-isoform choice has limited influence on the stability and translational efficiency of most mRNAs in mouse fibroblasts. Genome Research. 2013;23:2078–2090. doi: 10.1101/gr.156919.113. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib97] Tan L, Li Q, Xie XS. Olfactory sensory neurons transiently express multiple olfactory receptors during development. Molecular Systems Biology. 2015;11:844. doi: 10.15252/msb.20156639. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib98] Tan K, Wang Z, Zhang Z, An L, Tian J. IVF affects embryonic development in a sex-biased manner in mice. Reproduction. 2016;151:443–453. doi: 10.1530/REP-15-0588. [DOI] [PubMed] [Google Scholar]

[bib99] Tan K, Song HW, Wilkinson MF. Single-cell RNAseq analysis of testicular germ and somatic cell development during the perinatal period. Development. 2020a;147:dev183251. doi: 10.1242/dev.183251. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib100] Tan K, Song HW, Thompson M, Munyoki S, Sukhwani M, Hsieh TC, Orwig KE, Wilkinson MF. Transcriptome profiling reveals signaling conditions dictating human spermatogonia fate in vitro. PNAS. 2020b;117:17832–17841. doi: 10.1073/pnas.2000362117. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib101] Tarpey PS, Raymond FL, Nguyen LS, Rodriguez J, Hackett A, Vandeleur L, Smith R, Shoubridge C, Edkins S, Stevens C, O'Meara S, Tofts C, Barthorpe S, Buck G, Cole J, Halliday K, Hills K, Jones D, Mironenko T, Perry J, Varian J, West S, Widaa S, Teague J, Dicks E, Butler A, Menzies A, Richardson D, Jenkinson A, Shepherd R, Raine K, Moon J, Luo Y, Parnau J, Bhat SS, Gardner A, Corbett M, Brooks D, Thomas P, Parkinson-Lawrence E, Porteous ME, Warner JP, Sanderson T, Pearson P, Simensen RJ, Skinner C, Hoganson G, Superneau D, Wooster R, Bobrow M, Turner G, Stevenson RE, Schwartz CE, Futreal PA, Srivastava AK, Stratton MR, Gécz J. Mutations in UPF3B, a member of the nonsense-mediated mRNA decay complex, cause syndromic and nonsyndromic mental retardation. Nature Genetics. 2007;39:1127–1133. doi: 10.1038/ng2100. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib102] Tepe B, Hill MC, Pekarek BT, Hunt PJ, Martin TJ, Martin JF, Arenkiel BR. Single-Cell RNA-Seq of mouse olfactory bulb reveals cellular heterogeneity and Activity-Dependent molecular census of Adult-Born neurons. Cell Reports. 2018;25:2689–2703. doi: 10.1016/j.celrep.2018.11.034. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib103] Wachter A, Hartmann L. NMD: nonsense-mediated defense. Cell Host & Microbe. 2014;16:273–275. doi: 10.1016/j.chom.2014.08.015. [DOI] [PubMed] [Google Scholar]

[bib104] Wada M, Lokugamage KG, Nakagawa K, Narayanan K, Makino S. Interplay between coronavirus, a cytoplasmic RNA virus, and nonsense-mediated mRNA decay pathway. PNAS. 2018;115:E10157–E10166. doi: 10.1073/pnas.1811675115. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib105] Wang SS, Tsai RY, Reed RR. The characterization of the Olf-1/EBF-like HLH transcription factor family: implications in olfactory gene regulation and neuronal development. The Journal of Neuroscience. 1997;17:4149–4158. doi: 10.1523/JNEUROSCI.17-11-04149.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib106] Whitman MC, Greer CA. Adult neurogenesis and the olfactory system. Progress in Neurobiology. 2009;89:162–175. doi: 10.1016/j.pneurobio.2009.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib107] Woodward LA, Mabin JW, Gangras P, Singh G. The exon junction complex: a lifelong guardian of mRNA fate. Wiley Interdisciplinary Reviews: RNA. 2017;8:e1411. doi: 10.1002/wrna.1411. [DOI] [PubMed] [Google Scholar]

[bib108] Zhang X, Firestein S. The olfactory receptor gene superfamily of the mouse. Nature Neuroscience. 2002;5:124–133. doi: 10.1038/nn800. [DOI] [PubMed] [Google Scholar]

[bib109] Zhang LJ, Gallo RL. Antimicrobial peptides. Current Biology : CB. 2016;26:R14–R19. doi: 10.1016/j.cub.2015.11.017. [DOI] [PubMed] [Google Scholar]

[bib110] Zheng S, Gray EE, Chawla G, Porse BT, O'Dell TJ, Black DL. PSD-95 is post-transcriptionally repressed during early neural development by PTBP1 and PTBP2. Nature Neuroscience. 2012;15:381–388. doi: 10.1038/nn.3026. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib111] Zheng X, Boyer L, Jin M, Mertens J, Kim Y, Ma L, Ma L, Hamm M, Gage FH, Hunter T. Metabolic reprogramming during neuronal differentiation from aerobic glycolysis to neuronal oxidative phosphorylation. eLife. 2016;5:e13374. doi: 10.7554/eLife.13374. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

The role of the NMD factor UPF3B in olfactory sensory neurons

Kun Tan

Samantha H Jones

Blue B Lake

Jennifer N Chousal

Eleen Y Shum

Lingjuan Zhang

Song Chen

Abhishek Sohni

Shivam Pandya

Richard L Gallo

Kun Zhang

Heidi Cook-Andersen

Miles F Wilkinson

Roles

Abstract

Introduction

Results

UPF3B-regulated genes in mOSNs

Figure 1. Identification of UPF3B-regulated genes and NMD target genes in the olfactory system.

Figure 1—figure supplement 1. Upf3b-null mice behavior and purified mOSNs.

Figure 1—figure supplement 2. UPF3B-regulated genes in mOSNs.

Identification of NMD target mRNAs in mOSNs

Table 1. UPF3B-dependent NMD target mRNAs in mOSNs.

The mOSN transcriptome and translome

Figure 2. The mOSN translome and NMD.

The relationship between NMD and translation in vivo

Identification of OE cell clusters

Figure 3. Identification of OE cell subsets using scRNAseq analysis.

Figure 4. HBC, GBC, iOSN and mOSN heterogeneity.

OSN molecular pathways

Figure 5. Gene groups exhibiting distinct expression dynamics during OSN development.

UPF3B impacts HBCs and mOSNs

Figure 6. UPF3B shapes olfactory neurogenesis.

Figure 6—figure supplement 1. Impact of UPF3B loss on OE cell subsets.

Figure 6—figure supplement 2. mOSN subsets and UPF3B-regulated genes.

UPF3B shapes the OLFR repertoire and suppresses immune gene activation

Discussion

Materials and methods

Key resources table.

Mice

Behavioral and weight analyses

RNA-seq analysis

RiboTag analysis

scRNA-seq analysis

NIF analysis

Immunofluorescence analysis

Western blot analysis

Acknowledgements

Funding Statement

Contributor Information

Funding Information

Additional information

Competing interests

Author contributions

Ethics

Additional files

Data availability

References

Decision letter

Roles

Author response

Author response image 1. IF analysis of adult mouse OE sections co-stained with antisera against CAMP (red) and OMP (green).

Author response image 2. Venn diagram showing the expression of Olfr genes in Upf3b-null and control mOSNs.

Associated Data

Data Citations

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases