Bop1 is required to establish precursor domains of craniofacial tissues

Abstract

Bop1 can promote cell proliferation and is a component of the Pes1-Bop1-WDR12 (PeBoW) complex that regulates ribosomal RNA processing and biogenesis. In embryos, however, bop1 mRNA is highly enriched in the neural plate, cranial neural crest and placodes, and potentially may interact with Six1, which also is expressed in these tissues. Recent work demonstrated that during development, Bop1 is required for establishing the size of the tadpole brain, retina and cranial cartilages, as well as controlling neural tissue gene expression levels. Herein, we extend this work by assessing the effects of Bop1 knockdown at neural plate and larval stages. Loss of Bop1 expanded neural plate gene expression domains (sox2, sox11, irx1) and reduced neural crest (foxd3, sox9), placode (six1, sox11, irx1, sox9) and epidermal (dlx5) expression domains. At larval stages, Bop1 knockdown reduced the expression of several otic vesicle genes (six1, pax2, irx1, sox9, dlx5, otx2, tbx1) and branchial arch genes that are required for chondrogenesis (sox9, tbx1, dlx5). The latter was not the result of impaired neural crest migration. Together these observations indicate that Bop1 is a multifunctional protein that in addition to its well-known role in ribosomal biogenesis functions during early development to establish the craniofacial precursor domains.

RESULTS AND DISCUSSION

The Six1 transcription factor plays an important role in several developmental processes, and is a key regulator in establishing the cranial placodes that give rise to many of the sensory organs of the head, including the olfactory sensory epithelium, the cranial sensory ganglia and the inner ear (reviewed in: Saint-Jeannet and Moody, 2014; Schlosser, 2021). In humans, mutations in SIX1 and an interacting protein, EYA1, can result in Branchio-oto-renal (BOR) syndrome, which is characterized by hearing loss, branchial fistulas and kidney defects (reviewed in Moody et al., 2015; Neal et al., 2023). We previously conducted a screen for vertebrate homologues of proteins shown to interact with the Drosophila version of Six1 (Sine oculis) to identify other potential Six1-interacting proteins (Neilson et al., 2010). One of the candidates was Block of Proliferation 1 (Bop1), which in cell lines promotes cell proliferation and is a well-characterized component of the Pes1-Bop1-WDR12 (PeBoW) complex that is required for ribosomal RNA processing and biogenesis (Strezoska et al., 2000; Lapik et al., 2004; Rohrmoser et al., 2007). Although Bop1 is expressed ubiquitously in human adult tissues (www.proteinatlas.org/ENSG00000261236-BOP1/tissue), as expected for a protein that performs these essential cellular roles, during development of mouse, frog and fish embryos Bop1 mRNA is highly enriched in the neural plate, cranial placodes, cranial neural crest and pronephros, tissues that are affected in BOR (Kudoh et al., 2001; Gray et al., 2004; Neilson et al., 2010; Gärtner et al., 2022). Furthermore, RNA-seq and proteomic datasets demonstrate that Xenopus Bop1 has peak mRNA levels and high protein levels at neural plate and larval stages (Xenbase.org, based on Session et al., 2016; Peshkin et al., 2019). Thus, the embryonic expression data suggest an additional, developmental function for Bop1, consistent with reports that other proteins involved in ribosomal biogenesis, such as Pes1 and Ppan, have additional developmental functions including regulating gene expression in neural and pronephric tissues (Gessert et al., 2007; Bugner et al., 2011; Tecza et al., 2011). Recently, Gärtner et al. (2022) demonstrated that Bop1 also is required for establishing the size of the tadpole brain, retina and cranial cartilages, as well as controlling the levels of gene expression in embryonic brain and retina. They also provided evidence that these effects are likely independent of a ribosomal biogenesis or cell proliferation function, which is consistent with the idea that Bop1 performs multiple functions, sometimes referred to as protein “moonlighting” (Jeffery, 2018).

Because Drosophila interactome data indicate that its Bop1 ortholog (CG5033) is likely to interact with Sine oculis, the homologue of vertebrate Six1 (Neilson et al., 2010), we investigated whether Bop1 is required for the establishment of the progenitors of the cranial neural crest and sensory placodes at neural plate stages when Six1 is highly expressed (Pandur and Moody, 2000). To accomplish this, we depleted endogenous Bop1 protein on one side of the embryo by injecting an equimolar mixture of two antisense morpholino oligonucleotides (MOs) into the dorsal animal and ventral animal blastomeres of the 8-cell stage embryo, which are the major precursors of the cranial neural crest and placodes (Moody and Kline, 1990), and assessed gene expression at neural plate and larval stages. Each translation blocking MO recognizes both the L chromosome and S chromosome homoeologs of Bop1, was verified to be specific and effective, and in combination phenocopied Crispant mutants (Gärtner et al., 2022).

We first examined the effects of Bop1 knockdown on genes expressed at neural plate stages in the neural plate, neural crest and preplacodal ectoderm. The domains of three genes expressed in the neural plate (sox2, sox11, irx1) were each broader on the MO-injected side in a high percentage of embryos (Figure 1A-C). This is consistent with the expansion of another neural plate marker, sox3, at earlier neural ectodermal stages, as reported by Gärtner et al. (2022); interestingly, these authors did not observe expansion of sox2 at this earlier stage, perhaps because it is not upregulated by sox3 until neural plate stages (Rogers et al., 2009). Conversely, the gene expression domains of the epidermal border (dlx5), neural crest (foxd3, sox9), and preplacodal ectoderm (six1, sox11, irx1, sox9) were each reduced on the MO-injected side compared to the control side (Figure 1D-J). The effect on foxd3 is consistent with results reported by Gärtner et al. (2022). We next analyzed embryos in which the neural plate domains were expanded to determine if that phenotype occurred at the expense of neural crest or placode. In those foxd3 samples in which the neural crest expression was affected (n=45), 80.0% had a visibly expanded neural plate, as measured from the midline neural groove (Figure 1E). In those irx1 samples with expanded neural plate expression (n=28), 71.4.6% had diminished placode expression (Figure 1C). In those sox11 samples with expanded neural plate expression (n=36), 77.8% had diminished placode expression (Figure 1H’). Thus, in the majority of cases when loss of Bop1 expands the neural plate domain it is at the expense of neural crest and placodes. This result was also found with knockdown of other Six1 cofactors (Neilson et al., 2017, Neilson et al., 2020; Tavares et al., 2021; Jourdeuil et al., 2023), and is consistent with the observation that ectopic expression of neural plate genes represses neural crest, placode and epidermal genes (Klein et al., 2022). We therefore conclude that Bop1 is among those genes required for apportioning the early embryonic ectoderm into the domains required for proper development of the cranial central (neural plate) and peripheral (neural crest, placodes) nervous systems. These results support the findings that loss of Bop1 results in smaller brains and retinas at tadpole stages (Gärtner et al., 2022).

Figure 1: Changes in gene expression at neural plate stages after Bop1 knock-down.

(A-C) Knockdown of endogenous Bop1 caused an expansion of the neural plate (np) domains of sox2 (A), sox11 (B) and irx1 (C). Black bars indicate the width of the neural plate on the control side and red bars indicate the width of the neural plate on the Bop1-depleted side (asterisk). In (C), the black arrow indicates the irx1 domain on the control side and the red arrow indicates it on the Bop1-depleted side (asterisk).

(D) Bop1 knock-down reduced the epidermal expression of dlx5 along the border on the neural plate. Black arrow indicates the dlx5 domain on the control side and red arrow indicates it on the Bop1-depleted side (asterisk).

(E-F) Knockdown of endogenous Bop1 caused a reduction of the neural crest domains of foxd3 (E) and sox9 (F). Black arrows indicate expression domains on the control sides and red arrows indicate them on the Bop1-depleted sides (asterisk) of each embryo. In (E), the black bar indicates the width of the neural plate on the control side and the red bar indicates it on the Bop1-depleted side (asterisk).

(G-J) Knockdown of endogenous Bop1 caused a reduction of the preplacodal ectoderm domains of six1 (G), sox11 (H, H’), irx1 (I) and sox9 (J). Black arrows indicate placode expression domains on the control sides and red arrows indicate them on the Bop1-depleted sides (asterisk) of each embryo. In (H’), the black bar indicates the width of the neural plate on the control side and the red bar indicates an expanded domain on the Bop1-depleted side (asterisk).

Percentages indicate the frequency of each phenotype and numbers in parentheses indicate the number of embryos examined. Images in A-H’ are frontal views with dorsal to the top; images in I-J are anterolateral views with dorsal to the top.

Because a consistent feature of BOR is hearing loss, we next examined the effects of Bop1 knockdown on gene expression in the larval otic vesicle, which is the primordium of the inner ear. On the control side of larvae, six1, pax2, irx1, sox9, and dlx5 are broadly expressed in the otic vesicle (Figure 2A, B, C, E, G), otx2 in the ventral compartment (Figure 2D) and tbx1 in the posteroventral region (Figure 2F). Each of these genes has been shown to play important roles in the formation and regionalization of the inner ear (Depew et al., 1999; Burton et al., 2004; Saint-Germain et al., 2004; Moraes et al., 2005; Morsli et al., 1999; Raft et al., 2004; Torres et al., 1996). Knockdown of Bop1 reduced the size of the expression domains of all seven of these otic vesicle transcription factors in a high percentage of embryos (Figure 2). The pattern of dlx5 otic expression appeared different because on the control side the vesicle had involuted into a cup whereas it remained a spherical vesicle on the knockdown side. These results are consistent with a previous report that otx2 expression in the midbrain and retina was reduced in Bop1 morphants (Gärtner et al., 2022). In a subset of the larvae stained for otic vesicle genes, the diameter of the vesicle on the knockdown side was compared to the diameter of the vesicle on the control side of the same embryo. The otic vesicles on the knockdown side were larger in 1.31%, smaller in 37.6%, and the same in 61.09% (n=383), indicating that both the size of otic vesicle (e.g., Fig. 2B, E, F, G) and the level of expression (e.g., Fig 2A, C, D) can be affected.

Figure 2: Changes in gene expression at larval stages after Bop1 knock-down.

Knockdown of endogenous Bop1 caused a reduction of expression in the otic vesicle (ov) domains of six1 (A), pax2 (B), irx1 (C), otx2 (D), sox9 (E), tbx1 (F) and dlx5 (G). The pattern of dlx5 otic expression appeared different because on the control side the vesicle had involuted into a cup whereas it remained a spherical vesicle on the knockdown side. In addition, Bop1 knockdown caused a reduction in expression in the branchial arch (ba) domains of sox9 (E), tbx1 (F) and dlx5 (G). Black arrows indicate expression domains on the control sides and red arrows indicate them on the Bop1-depleted sides (asterisk) of each embryo. Percentages indicate the frequency of each phenotype and numbers in parentheses indicate the number of embryos examined. Images are side views with dorsal to the top.

A subset of the cranial neural crest migrates ventrally to form the mesenchymal tissue of the branchial arches and eventually differentiates into cranial cartilages (Noden, 1983). It was notable, therefore, that the branchial arch expression domains of sox9 and tbx1 were reduced in a high proportion of larvae (Figure 2E, F) because these genes are required for neural crest migration into the branchial arches (Spokony et al., 2002; Moraes et al., 2005) as well as later chondrogenesis including regulating the expression of a cartilage matrix gene, col2a1 (Mori-Akiyama et al., 2003; Tazumi et al., 2010). The expression levels of dlx5, which is required for patterning the branchial arches as well as chondrogenesis (Depew et al., 1999; Kraus and Lufkin, 2006), also were reduced in the majority of embryos (Figure 2G). However, dlx5 continued to be expressed in only the distal regions of the branchial arches, as in controls (Depew et al., 1999; Kraus and Lufkin, 2006), indicating that Bop1 does not regulate the patterning of its expression. It is likely that the disrupted expression of these genes contributed to the defects in ventral cranial cartilages reported by Gärtner et al. (2022). These authors also showed that co-expression of foxd3 mRNA in Bop1 morphants did not rescue the cranial cartilage defects, suggesting that loss of Bop1 does not perturb the specification of neural crest cells, but rather interferes with some aspect of their migration into the branchial arches or their differentiation into chondrogenic progenitors. Our data suggest that migration is not affected because sox9-, tbx1- and dlx5-expressing neural crest cells entered the branchial arches in Bop1 morphants. To test the ability of cranial neural crest to migrate in Bop1 morphants, we performed two tests of neural crest migration. First, the progenitor blastomeres of the cranial neural crest were labeled with membrane GFP that allowed us to follow the movement of these cells into the branchial migratory streams (Cousin et al., 2012). Control (n=279 in 4 independent trials) and Bop1 morphant (n=256 in 4 independent trials) embryos were scored for success of cells entering the migratory streams. There was no significant difference in the percentage of embryos showing successful movement into them (Figure 3A; p=0.077). Second, premigratory cranial neural crest were dissected from embryos, placed into explant culture and allowed to migrate in vitro for 10 hours according to the methods of Cousin and Alfandari (2018). The ratio of the 2-dimensional area of each explant between 0 and 10 hours (end area/beginning area) was calculated for controls (n=15 from 4 independent trials) and normalized to 1 ± 0.0 SD; the ratio of the area of Bop1 morphant explants (n=11 from 4 independent trials) was not significantly different (Figure 3B; p=0.122). These results, in combination with the previous demonstration that Bop1 knockdown neither decreased proliferation nor increased cell death (Gärtner et al., 2022), which frequently occurs when ribosomal proteins or biogenesis factors are depleted (Turi et al., 2019), lead us to propose that knockdown of Bop1 affects the chondrogenesis pathway. However, future analyses will be necessary to uncover the molecular pathway(s) by which this is achieved.

Figure 3: Neural crest migration is not significantly affected after Bop1 knock-down.

(A) The percentage of embryos that contained labeled neural crest cells in the branchial migratory streams in controls and sibling Bop1 morphants. Bar represents standard deviation. There is no significant difference between the two conditions (ns, p=0.077).

(B) The 2-dimensional area of neural crest explants was measured at the beginning and end of the culture period (10 hours) and expressed as the ratio of end area/beginning area. The ratios for explants from control embryos were set as 1.00±0.00 SD. The ratios for explants from Bop1 morphants did not significantly differ from controls (ns, p=0.122). Bar represents standard deviation.

The results reported herein extend the previous detailed analysis of the early embryonic effects of reducing the levels of endogenous Bop1 (Gärtner et al., 2022). By adding additional markers, we confirm their observations that neural plate gene expression domains are broader and neural crest gene domains are reduced. In addition, as we found by reducing the expression of other putative Six1-interacting proteins (Neilson et al., 2017, Neilson et al., 2020; Tavares et al., 2021; Jourdeuil et al., 2023), we show that placode and epidermal gene domains also were reduced, ultimately leading to otic vesicle defects at larval stages. We also demonstrate that the cranial cartilage defects shown by Gärtner et al. (2022) are likely due to defects in chondrogenesis gene expression rather than neural crest migration into the branchial arches. In future studies it will be important to: 1) biochemically establish Xenopus Bop1 as a Six1-interacting protein by co-immunoprecipitation assays; and 2) determine whether manipulating the levels of Six1 and Bop1 in embryos, by combined knockdown and overexpression assays as performed for Mscr1 and Six1 (Neilson et al., 2020), rescue the early gene expression phenotypes.

It is well established that when mutated, factors that are involved in ribosome biogenesis can lead to significant craniofacial dysmorphologies in experimental models and human patients (e.g., Zhou et al., 2014; Griffin et al., 2015; Falcon et al., 2022; Schreiner et al., 2022). However, there is evidence that some of the proteins known to be necessary for ribosomal synthesis also have other cellular functions, particularly during early development (Gessert et al., 2007; Bugner et al., 2011; Tecza et al., 2011). While Bop1 is a critical component the PeBoW complex, which regulates ribosomal RNA processing, it also apparently has other functions during development (Gärtner et al., 2022). For example, unlike some other ribosomal factors, loss of Bop1 does not result in changes in proliferation or apoptosis. Furthermore, expression of additional c-Myc in Bop1-depleted embryos, which would enhance ribosomal biogenesis via its interactions with RNA polymerases (Schreiner et al., 2022), did not rescue the morphant phenotypes (Gärtner et al., 2022). Thus, Bop1 appears to be a multifunctional protein whose molecular interactions are likely time- and tissue-dependent. The next important step will be to unravel Bop1’s developmental functions at the cellular and molecular levels that appear to be independent of ribosome biogenesis.

METHODS

Obtaining embryos and microinjections

Fertilized Xenopus laevis eggs were obtained by gonadotropin-induced natural mating of wild type, outbred adults as previously described (Moody, 2000; Moody, 2018). Embryos were selected at the 2-cell stage if the first cleavage furrow bisected the lightly pigmented region of the animal hemisphere to accurately identify the dorsal-ventral axis (Klein, 1987). When these selected embryos reached the 8-cell stage, both the dorsal animal and ventral animal blastomeres, which are the major contributors to the cranial neural crest and pre-placodal ectoderm (Moody and Kline, 1990), were microinjected on one side of the embryo with an equimolar mixture of MO1 and MO2 (9 ng total), according to standard methods (Moody, 2018). Injections were done on only one side of the embryo and the uninjected side used as an internal control of gene expression. Embryos were cultured in a diluted series of Steinberg’s solution until fixation.

Antisense morpholino oligonucleotides

To knockdown the endogenous level of Bop1 protein, we purchased two lissamine-labeled translation-blocking antisense morpholino oligonucleotides (Gene Tools, LLC) that bind to both the L chromosome and S chromosome homoeologs of Xenopus laevis bop1 mRNA, as verified by Gärtner et al. (2022). One binds to the ATG start site region and one to the upstream 5’UTR. An equimolar combination of both MOs was used in all injections.

In vitro synthesis of antisense RNA probes

Dig-labeled, antisense RNA probes for in situ hybridization (ISH) were synthesized in vitro (MEGAscript kit; Thermo Fisher), as previously described (Sive et al., 2000).

Fixation, histochemistry and in situ hybridization (ISH)

Embryos were cultured to neural plate (stages 16-18) or otic vesicle (stages 28-32) stages (Nieuwkoop and Faber, 1994), fixed in 4% paraformaldehyde (in 0.1M MOPS, 2mM EGTA Magnesium, 1mM MgSO₄, pH 7.4), and processed for ISH, as previously described (Sive et al., 2000). Only embryos in which the lissamine lineage tracer was located in the appropriate tissue domains were analyzed.

Gene expression analysis

To assess whether knockdown of endogenous Bop1 affects the size of the expression domains of the genes of interest, we compared the expression intensity and domain size between the injected, lineage-labeled side and the control, uninjected side of the same embryo. This direct intra-embryo comparison minimizes inter-embryo variation that might occur due to differences in developmental stages or experimental batches. Embryos were scored independently by at least two authors, and the values reported are means of their independent scores. Embryos that showed a more intense and/or larger expression domain on the injected side compared to control side of the same embryo were scored as “broader”. Embryos that showed reduced staining intensity or smaller expression domain on the injected side compared to control side of the same embryo were scored as “reduced”. The percentages of embryos that showed “broader” expression, “reduced” expression, or “no change” between injected and control sides of the same embryo were calculated for each gene assayed. In a subset of the larvae stained for otic vesicle genes, the diameter of the vesicle on the knockdown side was scored as larger, smaller or the same as the diameter of the vesicle on the control side of the same embryo. Embryos assessed for each gene were derived from a minimum of three independent experiments from different sets of outbred parents to attain genetic diversity.

Neural crest migration assays

The extent of cranial neural crest migration was assessed using two approaches. In vivo studies were performed as previously described (Cousin et al., 2012). Briefly, the dorsal animal blastomere of the 8-cell embryo, which is the major neural crest progenitor (Moody and Kline, 1990), was microinjected with 100 pg of membrane-associated GFP (mbGFP) mRNA alone or in combination with Bop1 MOs, as described above. When embryos reached neural crest migratory stages (stages 22-23), those expressing mbGFP in the dorsal cranial region were scored for the presence or absence of fluorescently labeled neural crest cells within the branchial arch migration pathways; partial migration was scored as presence. For each experiment, the results from embryos in which only lineage tracer was injected (control) were used as a normalizer and set to 100%. Control and morphant numbers were compared using a two-tailed, paired Student’s t-test. In vitro studies were performed as described in Cousin and Alfandari (2018). Briefly, one cell of the embryo was injected at the 2-cell stage with either 200pg of mbGFP mRNA alone or in combination with Bop1 MOs. At stage 15, explants of fluorescently labeled cranial neural were dissected, placed singly into a 96-well coated with 20μg/ml of fibronectin, and cultured at 18°C in Danilchik medium containing 50μg/ml gentamicin for about 10-11 hours, as described in Jourdeuil et al. (2023). The surface area occupied by the explant at the beginning and at the end of the culture were determined and the ratio of these areas (end/beginning) were calculated and normalized to the ratio obtained from control explants and compared using a two-tailed, paired Student’s t-test.

Data Availability:

All data for this study are presented in the body of the manuscript. Please contact authors for additional details.