Probing the physiological role of the plastid outer-envelope membrane using the oemiR plasmid collection

Abstract

Plastids are the site of complex biochemical pathways, most prominently photosynthesis. The organelle evolved through endosymbiosis with a cyanobacterium, which is exemplified by the outer envelope membrane that harbors more than 40 proteins in Arabidopsis. Their evolutionary conservation indicates high significance for plant cell function. While a few proteins are well-studied as part of the protein translocon complex the majority of outer envelope protein functions is unclear. Gaining a deeper functional understanding has been complicated by the lack of observable loss-of-function mutant phenotypes, which is often rooted in functional genetic redundancy. Therefore, we designed outer envelope-specific artificial micro RNAs (oemiRs) capable of downregulating transcripts from several loci simultaneously. We successfully tested oemiR function by performing a proof-of-concept screen for pale and cold-sensitive mutants. An in-depth analysis of pale mutant alleles deficient in the translocon component TOC75 using proteomics provided new insights into putative compensatory import pathways. The cold stress screen not only recapitulated 3 previously known phenotypes of cold-sensitive mutants but also identified 4 mutants of additional oemiR outer envelope loci. Altogether our study revealed a role of the outer envelope to tolerate cold conditions and showcasts the power of the oemiR collection to research the significance of outer envelope proteins.

Introduction

Plant chloroplasts are the cellular site of photosynthesis and host a number of interwoven pathways central to plant metabolism. Chloroplasts represent one specific type of plastid. However, all plastids originate from a single endosymbiotic event involving a cyanobacterial-like cell and an ancient host cell. As a result, plastids are enclosed by double-membranes consisting of the outer (OE) and inner (IE) envelope (Dyall et al. 2004).

Most studies have primarily focused on elucidating the IE function, as the OE was often regarded as a nonspecific molecular sieve or as a remnant of the food vacuole from its engulfment during endosymbiosis (Day and Theg 2018). Nevertheless, in recent years the significance of the OE and its proteins was demonstrated in various biological processes. The majority of OE proteins were in fact inherited from their prokaryotic ancestors, but key additions of eukaryotic descent required for integration of the engulfed cell are equally essential (Barth et al. 2022; Breuers et al. 2011). The mosaic nature of the membrane was instrumental for successful endosymbiosis and controls organelle biogenesis as well as plastid division.

In Arabidopsis, the OE harbors over 40 proteins (OEPs). OEPs thought to be of prokaryotic origin include porin-type channels, such as translocon of the outer chloroplast membrane (TOC)75, OEP21, OEP24, OEP37, and OEP40. Despite a lack of structural homology, they all function in preprotein and metabolite transport across the OE. Eukaryotic-type OEPs include tail-anchored proteins, such as OEP7 and OEP9, and GTPase receptors TOC34 and TOC159 (Barth et al. 2022). Apart from preprotein import and metabolite shuttling, OE proteins represent important players in other critical cellular functions, such as lipid biosynthesis (e.g. TGD4), or plastid division (PDV1/2) (Miyagishima et al. 2006; Xu et al. 2008). More recently, proteins belonging to the chloroplast-associated protein degradation (CHLORAD) pathway have unraveled an intriguing role of the OE in organellar protein ubiquitination and protein degradation (Ling et al. 2019; Woodson et al. 2015).

A plant's ability to adjust to environmental perturbations depends heavily on changes in chloroplast metabolism. Consequentially, the import of nuclear-encoded proteins as well as the shuttling of metabolites in and out of the chloroplast play a vital role in such acclimation processes (Kleine et al. 2021; Schwenkert et al. 2022). In recent years, proteins located in the chloroplast envelope including a number of OEPs have also been linked to plant acclimation responses in particular toward low temperature. In an extensive proteomics approach, Trentmann et al. (2020) could show that several OEPs are differentially regulated after cold treatment.

Forward genetic screening for plant mutants with altered stress responses would be a powerful approach to identify these and additional OEPs with roles in acclimation in planta. Unfortunately, such screens have limited success rates when multiple genes encode proteins with redundant functions (Cutler and McCourt 2005). This hurdle can be overcome by using an artificial microRNA (amiR) approach with constructs that have the ability to target and downregulate multiple homologs that potentially serve similar functions (Hauser et al. 2013; Jover-Gil et al. 2014). In addition, amiR lines are mostly hypomorphic enabling the study of gene loss effects in loci which cause embryo lethality if lost entirely (Kunz et al. 2014b). Thus far, amiR screens have successfully helped to identify redundant proteins involved in processes such as auxin transport, abscisic acid signaling, arsenite, and cadmium responses (Hauser et al. 2019; Xie et al. 2021; Zhang et al. 2018). Considering the fact that many OEPs are represented by gene families, amiR-based forward genetic screens provide an ideal approach to dissect the molecular fine-tuning capacities of this chloroplast protein subset.

In this study, we designed a collection of 36 binary pGreen-based vectors outfitted with OE-specific amiRNAs (oemiRs) targeting all to date verified OEPs. The tool was used to generate an initial oemiR plant mutant pool. As a proof-of-concept, this pool was screened for pale/photosynthesis-related as well as cold acclimation phenotypes. Since several OEPs are involved in preprotein import photosynthesis-related phenotypes pale plants were expected. To further investigate the aforementioned link between OEPs and cold acclimation we chose cold treatment as a screening condition. One of the isolated pale plants was identified as a toc75 loss-of-function mutant, a component of the protein import complex. This mutant was analyzed in more detail to understand the molecular consequences of impaired preprotein import on the cellular plant proteome.

Methods and material

Genetic redundancy predictions

Genetic redundancy was predicted among gene pairs using the model described by Cusack et al. (2021). Homologous gene pairs that clustered together in the phylogenetic tree were paired for analysis to determine the likelihood of their functional redundancy. Several genes that were related but did not cluster together in the phylogenetic analysis were also paired for model validation (Supplementary Data 1). Features were generated as described in the previous publication with one modification: coding sequences were here aligned in RAxML-NG using the Jones-Taylor-Thornto (JTT) model (Cusack et al. 2021). The model implemented Random Forest and was trained on the “extreme redundancy” dataset with 200 features.

oemiR plasmid collection generation and agrobacterium transformation

All target-specific antisense and sense amiRs (amiR* and amiR, respectively) sequences were designed using the Web MicroRNA Designer, WMD3 (https://www.weigelworld.org/). Each amiR constructs used in this study was cloned individually. The vector backbone was PCR amplified using the Platinum SuperFi II DNA Polymerase and the primers vec_fwd and vec_rev (Supplementary Data 1) (ThermoFisher Scientific), digested with DpnI and gel purified (Macherey&Nagel). The template for the vector backbone amplification was a fully functional and binary amiR expression clone in the pGREEN-based vector backbone vector called pG20_MCS_Hyg (Pratt et al. 2020). For each amiR construct, primers were designed with the amiR* and amiR being flanked by 5′ and 3′ sequences binding to the template vector (Supplementary Data 1). All amiR fragments were PCR amplified with the respective individual primer pairs using the Phusion polymerase (New England Biolabs) and subsequently gel purified (Macherey & Nagel). The vector backbone and the amiR fragments were assembled using Gibson seamless cloning (New England Biolabs) according to the manufacturer's instructions resulting on a functional binary expression vector. Maps of all vectors are provided in Supplementary Data 2.

PCR analysis and sequencing

Genomic DNA was isolated according to Kotchoni and Gachomo (2009). Subsequently, the amiR construct was amplified from the genomic DNA with Taq-Polymerase PCR using the primers mir319_for and HSP18_rev. The PCR program used consisted of an initial denaturation step of 30 seconds at 95°C. This was followed by a continued denaturation step of 30 seconds at 95°C, an annealing step for 50 seconds at 49°C, and extension step for 1 minute at 68°C. This was repeated for a total of 36 cycles before a final extension step for 5 minutes at 68°C. PCR products were sequenced (Sanger sequencing, LMU Faculty of Biology, Genetics Sequencing Service).

Plant growth and arabidopsis transformation

Arabidopsis plants were grown either on soil or on sterile solid ½ Murashige–Skoog–Medium (MS) medium. Plants were grown under long-day conditions (day: 16 h 100 μmol photons m⁻² s_-1, 21°C; night: 8 h dark, 16°C) in a climate chamber (cold treatment and growth on plates) or the greenhouse. Treatment at 4°C was performed under the same light/dark regime. Arabidopsis thaliana Columbia (Col-0) accession was used as wild type strain. Initially, individual Agrobacteria cultures carrying one of the 36 plasmids each were grown as overnight cultures. The next morning, all cultures were normalized to the same OD₆₀₀ and used to inoculate one joined 350 ml culture. Stably transformed Arabidopsis plants were generated using the floral dip method (Clough and Bent 1998). The dipping procedure was repeated after one week. For the selection of transformed plants, 1% agar containing ½ MS medium was supplemented with hygromycin. An ½MS mix with vitamins was used without the addition of sucrose. The medium was then adjusted to a pH of 5.7 with KOH before autoclaving and subsequent supplementation with 15 ug*ml⁻¹ hygromycin. In addition to nonselected Col-0 a transgenic line carrying a hygromycin resistance [control, (Kunz et al. 2014a)], was employed as a control. No apparent changes from the Col-0 wild-type were observed.

Chlorophyll (Chl) a fluorescence measurements

Chl a fluorescence of intact plants was measured using imaging pulse-amplitude-modulation fluorometry (Imaging PAM, Walz, Effeltrich, Germany) as described previously (Schneider et al. 2019). In brief, plants were dark-adapted for 20 min and exposed to a pulsed measuring light (intensity 1, gain 2, damping 1) and a saturating light flash (intensity 10) to calculate F_v/F_m, Photosystem II (PSII) quantum yield (ΦII = (Fm′ − F)/Fm′] and PS II regulated nonphotochemical energy loss [ΦNPQ = (Fm − Fm′)/Fm′] and PSII quantum yield of nonregulated nonphotochemical energy loss [ΦNO = F/Fm].

Blue native (BN)-PAGE

Thylakoid membranes were isolated and solubilized with 1% ß-dodecylmaltoside as described previously (Schwenkert et al. 2006). Samples were normalized according to fresh weight and separated on a NativePAGE 3 bis 12% Bis-Tris gel (Invitrogen).

Proteome analysis

Proteome analysis was performed using total protein extracts isolated from 3-week old leaf material (4 biological replicates). Protein preparation, trypsin digestion, and liquid chromatography-tandem mass spectrometry (LC-MS/MS) were performed as described previously (Marino et al. 2019). Raw files were processed using the MaxQuant software version 2.1.3.0 (Cox and Mann 2008). Peak lists were searched against the Arabidopsis reference proteome (Uniprot, www.uniprot.org, version April 2021) using the built-in Andromeda search engine (Cox et al. 2011) with default settings and “mach-between-runs' was enabled. Proteins were quantified using the label-free quantification algorithm (LFQ) (Cox et al. 2014). For hierarchical clustering, the ANOVA significant differentially expressed proteins based on the log₂ of the z-score normalized LFQ intensities. Envelope proteins were identified according to annotated envelope proteins in Bouchnak et al. (2019), omitting ribosomal proteins.

Computational analysis

CLC Main Workbench 20.4 (QIAGEN) was used to generate the phylogenetic tree. Downstream proteomics statistical analysis was performed using Perseus version 2.0.6.0 (Tyanova et al. 2016), R, and RStudio. Enrichment analysis was performed with ShinyGO (http://bioinformatics.sdstate.edu/go) (Ge et al. 2020) and subcellular localization analysis was performed with SUBA5 (https://suba.live/) (Hooper et al. 2017).

Results and discussion

Design of the oemiR plasmid collection

To design the oemiR plasmid collection we selected candidates with a high confidence in respect of their subcellular and OE localization, respectively. Almost all selected protein targets were confirmed in previous proteomics studies and/or by immunoblots or fluorescence-based methods. Table 1 provides either a representative reference for a proteomics dataset or a functional study investigating the localization. Since IE and OE cannot be properly separated in Arabidopsis, studies on Pisum sativum OE membranes identifying the respective paralogues were included (Simm et al. 2013). Moreover, key publications analyzing the functions of the individual proteins are provided in Table 1. For several OEPs distinct functions have been assigned in the last decades. These assignments range from the formation of membrane channels enabling protein and metabolite trafficking to plastid division, lipid biosynthesis, and signaling. Since most functions are tied to their varying structural prerequisites we summarized structural classifications of all OE proteins in Table 1 (Fish et al. 2022). Interestingly, a number of OE proteins form ß-barrels, which are common in Gram-negative bacteria and mitochondrial membranes where they generally serve as porins. Among these, OEP23 and JASSY stand out, as their structure prediction suggest that they are incomplete ß-barrel proteins. JASSY belongs to the START/RHO_alpha_C/PITP/Bet_v1/CoxG/CalC superfamily, which contains a conserved ligand-binding domain forming a large hydrophobic binding cavity, which gives rise to its function as an 12-oxo Phytodienoic Acid (OPDA) transporter (Radauer et al. 2008; Guan et al. 2019). OEP23 has been shown to possess ion permeability, however, the exact transport substrates remain to be identified (Goetze et al. 2015). Further studies are also required to analyze the structural details of these potentially incomplete and unusual ß-barrel proteins.

Table 1.

Targets of the oemiR plasmid collection.

Group ID	Protein name	(potential) Function	Locus	amiR (sense)	Ref Function	Ref Loc	Structure
1	OEP24A OEP24B	Solute/ion channel Solute/ion channel	At1g45170 At5g42960	TTGAATGTAAGTCAGATTCAC	Pohlmeyer et al. (1998)	Simm et al. (2013)	β-barrel
2	OEP21A OEP21B	Solute/ion channel Solute/ion channel	At1g20816 At1g76405	TTCATCGCACAGAAGTAACTT	Bolter et al. (1999)	Simm et al. (2013)	β-barrel
3	OEP40	Glucose channel	At3g57990	TAATTCAGCGCTCATGCGCAT	Harsman et al. (2016)	Harsman et al. (2016)	β-barrel
4	OEP37	Solute/ion channel	At2g43950	TAATCTATCGCAAAGTCCCGA	Goetze et al. (2006)	Simm et al. (2013)	β-barrel
5	OEP16-1	Amino Acid channel	At2g28900	TGATAGTTGCTAAATACACGT	(Philippar et al. 2007; Pohlmeyer et al. 1997; Pudelski et al. 2012)	Simm et al. (2013)	α-helical
6	OEP16-2	Amino Acid channel	At4g16160	TTCTTAGTAGACCTTTAGCGC	(Philippar et al. 2007; Pohlmeyer et al. 1997; Pudelski et al. 2012)	Philippar et al. (2007)	α-helical
7	OEP16-4	Amino Acid channel	At3g62880	TATTCGTGAATAAACTGGCCT	(Philippar et al. 2007; Pohlmeyer et al. 1997; Pudelski et al. 2012)	Philippar et al. (2007)	α-helical
8	OEP23	Solute/ion channel	At2g17695	TAAGACGTTATCTCATACCAA	Goetze et al. (2015)	Trentmann et al. (2020)	Other
9	WBC7	ABC transporter	At2g01320	TTCAAATTTAGCGTACAGCGA	Sanchez-Fernandez et al. (2001)	Simm et al. (2013)	α-helical
10	JASSY-1	OPDA exporter	At1g70480	TTTTCATAGAGTGATCTGCGC	Guan et al. (2019)	Guan et al. (2019)	Other
11	JASSY-2	Unknown/OPDA exporter	At1g23560	TCAATTACATTGACCTACCAG	Guan et al. (2019)		Other
12	MGDG2 MGDG3	Monogalactolipid synthesis	At5g20410 At2g11810	TAACATACGGCACGTTGCCTC	Kobayashi et al. (2009)	Awai et al. (2001)	α-helical
13	DGD1 DGD2	Digalactolipid synthesis	At3g11670 At4g00550	TCTTCTGCGGTTGTTGTGCAA	Kobayashi et al. (2009)	Kelly et al. (2003)	Other
14	TGD4	Lipid import	At3g06960	TATAAATGGTAACTTGGGCCA	Xu et al. (2008)	Bouchnak et al. (2019)	Other
15	SFR2	Diacylglycerol acyltransferase	At3g06510	TAATTTGAGACCTAATAGCAG	(Fourrier et al. 2008; Thorlby et al. 2004)	Simm et al. (2013)	Other
16	LACS9	Long-chainacyl-CoAsynthetase	At1g77590	TCATATTACGGTTGTGACCTA	Shockey et al. (2002)	Simm et al. (2013)	α-helical
17	OEP9-1	Unknown	At1g16000	TTAACAGTGTGCAAATGACAC		Dhanoa et al. (2010)	α-helical
18	OEP9-2	Unknown	At1g80890	TATGTAGTTGACTAGAGTCTA			α-helical
19	OEP7	Unknown	At3g52420	TCAAATAAACGATCATGACGC		Lee et al. (2001)	Tail anchored
20	CHUP1	Actin binding	At3g25690	TACTTTACAGAACTATGTCCT	Oikawa et al. (2003)	Oikawa et al. (2003)	Other
21	PDV1	Plastid division	At5g53280	TTGCTACTAAAGAATAGCCGC	Okazaki et al. (2015)	Miyagishima et al. (2006)	Other
22	PDV2	Plastid division	At2g16070	TTCTGGCTAAAATTGACCCGA	Okazaki et al. (2015)	Miyagishima et al. (2006)	Tail anchored
23	PTM	Retrograde signaling	At5g35210	TGATTATACGGCAGGAGGCAG	Sun et al. (2011)	Froehlich et al. (2003)	Other
24	HXK1	Glucose-responsive sensor hexokinase	At4g29130	TATTACCGAAAAATGGCGCTG	Jang et al. (1997)	Simm et al. (2013)	Signal anchored
25	CRL	OEP80 insertion	At5g51020	TTATAGTCGTCAAATGCGCTC	Asano et al. (2004), Yoshimura et al (2023)	(Simm et al. 2013)	Signal anchored
26	THF1	Sugar signaling; thylakoid formation	At2g20890	TTTATATAGAGTATCTCCCAT	Huang et al. (2006); (Keren et al. 2005)	Trentmann et al. (2020)	α-helical
27	TOC75-III TOC75-IV TOC75-I	Protein import	At3g46740 At4g09080 At1g35860	TACCGAGTTTCACACCCGCAC	Baldwin et al. (2005)	Simm et al. (2013)	β-barrel
28	TOC159	Protein import receptor	At4g02510	TAGAATTGCGAGTAAAGGCAG	Kubis et al. (2004)	Simm et al. (2013)	Other
29	TOC90	Protein import receptor	At5g20300	TATATTATTCTGTGACTCCCC	Hiltbrunner et al. (2004)	Bouchnak et al. (2019)	Other
30	TOC120 TOC132	Protein import receptor	At3g16620 At2g16640	TATGTTTAACCGAGCTGTCCT	Kubis et al. (2004)	Simm et al. (2013)	Other
31	TOC33 TOC34	Protein import receptor	At1g02280 At5g05000	TGTACACATCCAAACGGGCAA	Weibel et al. (2003)	Bouchnak et al. (2019)	Tail anchored
32	TOC64	Protein import	At3g17970	TTTTATCGATAAAAGCGCCGG	Qbadou et al. (2007)	Simm et al. (2013)	Signal anchored
33	OEP80	Protein insertion	At5g19620	TAACACGCACCCCTAAAGCAT	Patel et al. (2008)	Trentmann et al. (2020)	β-barrel
34	SP2 P36	Degradation of TOC complex, CHLORAD	At3g44160 At3g48620	TTTAATCGGACGCACATGCAA	Ling et al. (2019)	Ling et al. (2019)	β-barrel
35	KOC1	Tyrosine kinase	At4g32250	TTAACACCAGTAATGACGCGG	Zufferey et al. (2017)	Trentmann et al. (2020)	Tail anchored
36	PAP2 PAP9	Phosphatase	At1g13900 At2g03450	TGTACATTGGTCTATGCCCTT	Sun et al. (2012)	Sun et al. (2012)	Tail anchored

PTM, PHD type transcription factor with transmembrane domains; CRL, crumpled leaf.

To gain insight into the issue of functional redundancy among OE proteins, we aligned all cDNA sequences and compared these in a radial phylogenetic tree to identify gene clusters (Fig. 1a). Several gene pairs or highly similar gene family members became apparent by clustering. Next up, the OE gene pairs were evaluated to determine their relative likelihood of functional redundancy, i.e. the risk that no phenotype would emerge in a single loss-of-function mutant. Indeed, the redundancy model suggested that there was a high risk of functional redundancy (Redundancy score ≥ 0.5) for all except one of the homologous gene pairs analyzed (Fig. 1, a and b).

Fig. 1.

Design of the oemiR plasmid collection. a) Circular phylogram based on all 46 outer envelope protein (OEP) cDNA sequences. Gene pairs targeted by one amiR are indicated in brackets and scores adjacent to gene pairs indicate the predicted likelihood of redundancy between the 2 genes, with 0 indicating that the genes were likely nonredundant and 1 indicating that the genes were likely redundant. b) Predicted redundancy scores for 25 pairs of genes among the OEPs. Genes that cluster together tend to have higher redundancy scores (≥0.5) than those that do not cluster together, consistent with the closer genetic relationships among clustered genes. c) One-Step amiR cloning workflow. Each construct was cloned individually by PCR amplifying the vector backbone and the amiR insert followed by In-Fusion assembly, resulting in a fully functional binary vector. pUBQ10: Ubiquitin10 promoter; amiR* and amiR: Specific amiR antisense and sense sequence, respectively; HygR: Hygromycin resistance cassette; LB and RB: Left and right T-DNA border, respectively. d) Plant mutant screening. T₁ seeds obtained after transformation with the oemiR plasmid collection were selected for hygromycin-resistant seedlings. At the age of 7 days, resistant seedlings (80 in total) were transferred to soil and grown for 7 days at 21°C before being transferred to 4°C for 7 days. At stages 1 and 2 phenotypes were monitored, Chl a fluorescence measurements were taken and plants differing from the control were subjected to PCR analysis and sequencing. AmiR group IDs were identified as indicated.

The main goal of the oemiR design was to find the minimal number of amiRs capable of downregulating as many gene targets as possible to bypass potential functional redundancy. Encouragingly, amiRs could be designed for all predicted functional redundant gene pairs, except JASSY and OEP9. Overall, 36 amiR constructs were sufficient to target all 46 genes (as indicated with colors in Fig. 1a). All amiR group IDs targeting one or more gene loci transcripts along with the used amiR sequences are listed in Table 1. To streamline the molecular cloning, we established a new amiR one-step cloning process (Fig. 1c). Initially, the original MIR319a employed for the design of amiRs (Schwab et al. 2006), was inserted into an updated binary pGREEN-based vector called pG20_MCS_Hyg (Pratt et al. 2020). The high-copy vector then served as a PCR template to (1) incorporate the target-specific amiR* and amiR sequences into a new DNA fragment also containing the stem loop and (2) amplify the vector backbone. Subsequently, both fragments were assembled using Gibson seamless cloning (New England Biolabs).

Workflow of the oemiR proof-of-concept screen

All 36 oemiR constructs were transformed into Col-0 plants (T₀) using standard Agrobacteria-mediated floral-dip. The resulting hygromycin-resistant T₁ progeny was subjected to a 2-step screening procedure for mutants impaired in growth, leaf paleness, photosynthesis, and/or sensitivity to cold treatment (Fig. 1d). In total, 80 T₁ plants were analyzed. At an age of 14 days (long day 16 h light/8 h dark period conditions), phenotypes were visually inspected. Additionally, the plants' photosynthetic capacity was evaluated by PAM chlorophyll fluorometry (timepoint 1). F_v/F_m, the maximum quantum efficiency of photosystem II (PSII), was evaluated as a general parameter reflecting plant fitness. Subsequently, plants were transferred to long day 16 h/8 h conditions at 4°C. After 7 days, F_v/F_m was measured again (timepoint 2). Generally, most plants did not show any morphological or other noticeable abnormalities. Nevertheless, some plants were smaller and paler than the control, already at timepoint 1. These plants were therefore subjected to PCR and sequencing to identify the causative amiR.

Analysis of amiR-toc75 plants

Upon sequencing analysis of the pale plants described above, amiR ID 27, targeting the TOC75 gene family, was identified. Since TOC75 is responsible for the import of most nuclear-encoded chloroplast proteins, this phenotype matched our expectations. Moreover, reduced import rates leading to a similar phenotype in a toc75-III RNAi mutant allele were observed previously (Huang et al. 2011). TOC75 proteins are part of the outer membrane protein of 85 kDa (Omp85) superfamily also found in the outer membranes of gram-negative bacteria and mitochondria (Hsu and Inoue 2009; Day et al. 2019). Initially, 3 genes were assigned to the TOC75 family. Apart from the ubiquitously expressed TOC75-III, TOC75-IV might play a role in etioplasts is part of the family, as well as TOC75-I, which was classified as a pseudogene (Jackson-Constan and Keegstra 2001; Baldwin et al. 2005). Later, TOC75-V/OEP80 was identified via proteomics studies as a TOC75 paralogue (Eckart et al. 2002) involved in the insertion of other OE β-barrel proteins (Patel et al. 2008; Gross et al. 2021). Complete loss of OEP80 renders mutants embryolethal (Patel et al. 2008). Since also TOC75-III null mutants are embryolethal (Baldwin et al. 2005), we chose the amiR-toc75 lines to (a) verify the specificity of our amiRNA constructs and (b) to analyze how TOC75 downregulation affects the plant proteome in more detail.

To confirm the observed phenotype as caused by amiR ID 27, we performed an independent transformation with Agrobacteria only containing this plasmid. AmiR ID 27 was designed to target only TOC75-I, TOC75-III, and TOC75-IV. Due to its distinct function in chloroplast biogenesis, a separate amiR was designed for OEP80 (Table 1). Nevertheless, since the amiR ID 27 target site displayed some similarly to OEP80, this amiR represented an ideal tool to verify targeting specificity of our designed oemiRs (Fig. 2a). Two independent amiR-toc75 mutant lines were obtained and the progenitors from this transformation both yielded pale plant individuals in the T₁ and T₂ generations (Fig. 2b). As a first step, a BN-PAGE was performed to investigate the overall integrity of thylakoid membrane complexes. All photosynthetic complexes were found to be reduced, especially PSII-LHCII supercomplexes, pointing toward a pleiotropic phenotype caused by lack of nuclear-encoded chloroplast proteins (Fig. 2c).

Fig. 2.

Analysis of amiR-toc75 mutants. a) Alignment of the amiR ID 27 target region. b) Three-week-old amiR-toc75 plants. c) Large-pore BN gel of protein complexes from Col-0 and amiR-toc75 chloroplasts. (d–g) Chl a fluorescence measurements of Col-0 and amiR-toc75 plants. The asterisks indicate statistically significant differences according to one-way analysis of variance comparing Col-0 to each mutant (*P < 0.05; **P < 0.01; ***P < 0.001; n = 8).

To further analyze the photosynthetic performance of the mutants we measured chl a fluorescence parameters at increasing light intensities. F_v/F_m was found slightly, but significantly reduced as compared to Col-0 wild-type plants (Fig. 2d). Moreover, both amiR-toc75 alleles displayed reduced PSII yield (ΦII), probably caused by a lower abundance of PSII complexes, which harbor a number of nuclear-encoded subunits. This was accompanied by higher regulated nonphotochemical quenching (ΦNPQ) and slightly elevated nonregulated nonphotochemical quenching (ΦNO) indicative of oxidative stress caused by plastid malfunction in amiR-toc75 plants (Fig. 2e–g).

Next up, we were interested to analyze the extent of TOC75 protein downregulation by the amiR. Additionally, we also expected to gain a global understanding of the molecular consequences and potential compensatory mechanisms in response to the loss of a protein import translocon unit. Therefore, we performed label-free protein quantification on leaf extracts using mass spectrometry (Fig. 3a–f). Of the 3 predicted amiR targets (TOC75-I, TOC75-III, TOC75-IV), only TOC75-III protein was identified. This can be explained by the low TOC75-IV abundance in leaf chloroplasts (Baldwin et al. 2005). TOC75-I is a pseudogene as mentioned above. Nevertheless, TOC75-III was shown to be substantially downregulated by 3-fold in amiR-toc75-1 and 3.6-fold in amiR-toc75-2, respectively. All other significantly differentially regulated proteins in both amiR-toc75 vs Col-0 were identified by performing a Student t-test. The -log₂ fold changes of all up- and downregulated proteins were plotted against the P-value and depicted as a volcano plot (Fig. 3a and Supplementary Data 3). Notably, OEP80 was unchanged in both mutant lines indicating a high target specificity toward TOC75-III by the amiR construct. In order to compare the 2 independently obtained mutant lines we matched the number of overlapping up- and downregulated proteins in both lines and found most proteins (between 62 and 75%) to be regulated in the same manner (Fig. 3b).

Fig. 3.

Proteome analysis of amiR-toc75 mutants. a) Volcano plots showing differentially regulated proteins in amiR-toc75-1/amiR-toc75-2 vs Col-0. Each dot represents one protein, plotted according to P-value and relative abundance ratio amiR-toc75-1/amiR-toc75-2 and Col-0, upper and lower panel, respectively. Toc75-III and OEP80 are highlighted. The dashed line indicates the -log₁₀P-value of 1.5. b) Venn diagrams showing the overlapping up- and downregulated proteins in both amiR-toc75 lines. c) Hierarchical cluster analysis of the differentially expressed proteins. Bar represents the relative z-score. d) Subcellular enrichment analysis of the clusters identified in (c). e) KEGG pathway enrichment analysis of the clusters identified in (c). f) More than 2-fold up- and downregulated OE and IE proteins extracted from the clusters identified in (c). g) GO biological process enrichment analysis of upregulated plastid proteins identified in (d).

Since not much is known about the effect of TOC75 downregulation on the composition of the proteome and in order to perform a functional enrichment analysis we subjected the data to hierarchical clustering analysis. Differentially regulated proteins are grouped into 2 clusters. Cluster 1 contained 321 proteins, which were mainly downregulated in both amiR-toc75 lines vs Col-0. Cluster 2 held 619 mainly upregulated proteins in both amiR-toc75 lines vs Col-0 (Fig. 3c). To estimate the compartmental protein abundance of down- vs upregulated proteins we utilized the SUBA5 database. The results shown in Fig. 3d reveal that the majority of downregulated proteins are localized in chloroplasts, which underpins the role of TOC75 in plastid preprotein import. Moreover, other components of the TOC complex, such as TOC159, a major preprotein receptor, were also downregulated in response to low TOC75 abundance (Fig. 3e). Most of the upregulated proteins, however, were cytosolic. For further functional analysis of these proteins, we performed Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. While cluster 1 was heavily enriched in proteins related to photosynthesis and chloroplast functions, the upregulated proteins (cluster 2) were enriched in proteins related to cytosolic ribosome assembly, translation, and protein degradation (Fig. 3f). One explanation is that compensatory mechanisms to level out the effects of reduced plastid protein import become initiated. This is achieved through higher translation rates and faster degradation of unimported, aggregated preproteins.

Interestingly, also a number of chloroplasts upregulated proteins in amiR-toc75 vs Col-0 were identified. GO term enrichment analysis revealed that these comprised proteins with biological functions related to plastid transcription, but also proteins involved in preprotein import (Fig. 3f). Among those were prominent proteins of the inner preprotein translocon (TIC) complex, i.e. TIC40 and TIC56. Also, TIC110 was upregulated, albeit it has been controversially discussed whether TIC110 is a permanent component of the TIC complex (Fig. 3g) (Nakai 2015; Bolter 2018; Jin et al. 2022). In addition, Chaperonin CPN60, 93-kD heat shock protein HSP93-III/CLPC2 and chloroplast heat shock protein HSP90C are upregulated, which are also suggested to play a role in preprotein import (Bolter 2018) (Supplementary Data 3). At first glance, upregulation of these components seems to be a logic consequence if the cell tries to compensate for reduced import rates across the OE. Nevertheless, it also bares the question of how these proteins enter the chloroplast in the absence of a fully assembled functional TOC complex in the first place? Moreover, ten additional stromal proteins were upregulated more than 2-fold. One possibility is that these proteins are preferentially recognized and transported to keep the mutant chloroplasts functional to some degree. Alternatively, some of these proteins might utilize noncanonical import pathways which are yet to be unraveled but have been posited in the past (Armbruster et al. 2009). In summary, this experiment confirmed that amiRs provide a reliable method to study loss-of-function effects in the OE circumventing functional genetic redundancy.

Identification of mutants with lower photosynthetic performance in the cold

A recent study has shown a significant adjustment of the OE proteome in response to cold treatments (Trentmann et al. 2020). To investigate the impact of OEP reduction on cold acclimation, we screened our initial oemiR mutant pool for phenotypic changes in response to low temperature. All 80 T₁ plants were shifted into long-day growth conditions at 4°C. After one week at 4°C F_v/F_m was determined. While the control and most oemiR mutants exhibited an F_v/F_m value around 0.70, ten oemiR lines revealed average F_v/F_m values of 0.55 and below (P < 0.05 Student t-test). Subsequently, mutant plants with a cold-treatment-induced drop in average F_v/F_m were PCR genotyped and sequenced to identify the respective causative amiRs (Fig. 4, a and b). Out of these, 7 mutant plants carried amiR constructs targeting SENSITIVE TO FREEZING2 (SFR2), CHLOROPLAST UNUSUAL POSITIONING PROTEIN 1(CHUP1), OEP40, TOC159, OEP16-2, PLASTID DIVISION2 (PDV2), and the putative ATP binding cassette (ABC)-type transporter WBC7, respectively. Three oemiR lines gave inconclusive sequencing results, potentially due to the presence of multiple insertions. In general, mutant lines carrying multiple amiRs can be cleaned up and turned into oemiR single mutant by backcrossing into Col-0 wild-type plants. The resulting F1 progeny needs to be tested for the heritability of the desired phenotype and genotyped.

Fig. 4.

Screening of oemiR mutants with affected cold acclimation based on F_v/F_m changes from wild-type control plants. a) All mutant plants were treated at 4°C, with Chl a fluorescence measurements taken before and after treatment. Mutants that had an average F_v/F_m value of 0.55 after cold treatment were identified and a Student t-test was performed after cold treatment with each mutant against the control (*P < 0.05; **P < 0.01; ***P < 0.001; n = 4). Genotyping was performed on mutants that showed a statistically significant decrease in average F_v/F_m after cold treatment (P < 0.05). Seven of these mutants were sequenced and identified to harbor amiR constructs targeting SFR2, CHUP1, OEP40, TOC159, OEP16.2, PDV2, and WBC7, respectively. b) Chlorophyll fluorescence imaging before and after cold treatment.

A literature research confirmed that loss-of-functions in sfr2, oep40, and chup1 genes have been previously linked to cold-sensitivity. SFR2, a member of the Family-1 β-glycosidases, is partially responsible for removing monogalactolipids from the OE, a lipid-remodeling process vital for freezing tolerance (Thorlby et al. 2004; Moellering et al. 2010). Consequently, sfr2 loss-of-function mutants suffer chloroplast damage after exposure to freezing conditions (Fourrier et al. 2008). Extended cold-treatments, i.e. above 0°C, have not been reported thus far. OEP40 encodes for a regulated β-barrel OE channel permeable to glucose and its phosphorylated derivatives. Loss of OEP40 triggers early flowering under cold temperature conditions, which is indicative of carbohydrate imbalance, particularly in the floral meristem (Harsman et al. 2016). Finally, CHUP1 is an OEP featuring an actin-binding domain responsible for chloroplast positioning and movement within leaf cells (Oikawa et al. 2003; Oikawa et al. 2008). Under ambient light/low-temperature conditions chup1 loss-of-function mutants exhibit lower F_v/F_m values compared to control plants (Kitashova et al. 2021). This phenotype was recapitulated in the amiR-chup1 line isolated in this study.

While the identification of SFR2, OEP40, and CUP1 confirmed the functionality of our screening procedure, the 3 other cold-affected mutants identified are valuable for further studies directed toward the OE role in cold acclimation. The isolated mutants harbored individual amiR constructs targeting TOC159, OEP16-2, PDV2, and WBC7, respectively. TOC159, also known as PLASTID PROTEIN IMPORT 2 (PPI2), encodes for a membrane GTPase that functions as a transit-sequence receptor and represents an integral part required of the translocase complex. Previous work on ppi2 mutants has shown that null alleles are seedling lethal but exhibit an albino phenotype if grown on sucrose (Bauer et al. 2000). The amiR-toc159 line isolated and used in this study did not show such a strong phenotype under normal growth conditions. Nevertheless, the mutant did exhibit cold-sensitivity. This finding underpins the value of milder mutant versions to study complex physiological stress responses. We did also find strongly comprised amiR-toc159 lines reminiscent of the ppi2 allele which we decided to not use for the cold sensitivity experiment.

OEP16-2 was shown to affect metabolic fluxes during abscisic acid (ABA)-controlled seed development and germination (Pudelski et al. 2012). It forms a potential amino-acid selective channel that possesses moderate cation-selectivity but high-conductance. While OEP16-1, which according to the design algorithm is not targeted by the OEP16-2 amiR, is highly abundant in leaf tissue, the OEP16-2 isoform is preferably expressed during germination and late seed development (Pudelski et al. 2012). Nevertheless, the observed leaf phenotype in the amiR-oep16-2 allele may indicate either unpredicted off-targeting toward OEP16-1 or a specialized role of OEP16-2 during cold acclimation which has previously not been reported. The same is true for PDV2, a tail-anchored protein that, along with its homolog PDV1, is a vital component of the division machinery in plastids (Chang et al. 2017; Miyagishima et al. 2006). Finally, WBC7 is a putative ABC transporter located in the OE (Sanchez-Fernandez et al. 2001). The exact molecular function of this protein remains elusive. Nevertheless, confirming our amiR-wpc7 mutant phenotype results, WBC7 has been found to be downregulated under cold conditions (Trentmann et al. 2020).

With this pilot screen, we were able to emphasize a global function of OEPs in cold acclimation. However, to pursue the molecular functionalities causing the observed cold-sensitive phenotypes in amiR-toc159, amiR-oep16-2, amiR-pdv2, and amiR-wbc7 further verification steps are required. As exemplified for amiR-toc75 individual oemiR constructs can be utilized to generate independent, yet comparable mutant lines with high specificity toward the gene(s) of interest. Homozygous T-DNA or CrispR mutants, if viable, can be employed to verify observed phenotype(s) and to unravel the biological role of the respective proteins in cold acclimation and beyond. Future studies for instance exploiting the oemiR collection will dissect the OE's critical role in acclimation under other abiotic and biotic stress scenarios.

Overall, this OE-specific amiR-based screening tool exhibits great potential in overcoming common issues faced with other screening methods, such as seedling lethality and functional redundancy. The oemiR collection demonstrates the power of amiR-based strategies for new studies not limited to the plastid OE, but also other biological subsystems such as whole organelles.

Data availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD041299. Sequence information on the oemiR plasmid collection is provided in Supplementary Data 2. Plasmids are deposited at Belgian Coordinated Collections of Microorganisms, see Supplementary Data 1 for accession numbers (https://bccm.belspo.be/deposit/public/plasmids).

Supplemental material available at G3 online.

Funding

S.S., H.H.K., and D.L. received funding from the DFG, SFB-TR 175, project B06, B09, and B07 respectively. A.I.P. was funded by a National Science Foundation (NSF) Career Award IOS-1553506 to H.H.K.

Author contributions

S.S. and H.H.K.: conceptualization and supervision; S.S., W.T.L., B.S., C.K.Y., S.A.C., A.I.P., and B.B.: investigation; D.L.: review and editing; S.S., H.H.K., and W.T.L.: writing.