Genetic characterization of C. elegans TMED genes

Abstract

Background

p24/Transmembrane Emp24 Domain (TMED) proteins are a set of evolutionarily conserved, single pass transmembrane proteins that have been shown to facilitate protein secretion and selection of cargo proteins to transport vesicles in the cellular secretion pathway. However, their functions in animal development are incompletely understood.

Results

The C. elegans genome encodes eight identified TMED genes, with at least one member from each defined subfamily (α, β, γ, δ). TMED gene mutants exhibit a shared set of defects in embryonic viability, animal movement, and vulval morphology. Two γ subfamily genes, tmed-1 and tmed-3, exhibit the ability to compensate for each other, as defects in movement and vulva morphology are only apparent in double mutants. TMED mutants also exhibit a delay in breakdown of basement membrane during vulva development.

Conclusions

The results establish a genetic and experimental framework for the study of TMED gene function in C. elegans, and argue that a functional protein from each subfamily is important for a shared set of developmental processes. A specific function for TMED genes is to facilitate breakdown of the basement membrane between the somatic gonad and vulval epithelial cells, suggesting a role for TMED proteins in tissue reorganization during animal development.

Introduction

Cells interact with their environment and each other using a complex set of membrane-associated and secreted proteins, lipids and other molecules. In eukaryotes, this highly dynamic and regulated process is mediated by the secretory pathway (reviewed in Brandizzi and Barlowe¹). Appropriate processing and delivery of secretory cargo proteins relies on their sorting and selection into distinct transport vesicles. This selection is dependent on interaction of cargo proteins with coat protein complexes (COPI and COPII) which have multiple binding sites to recruit cargo during formation of the vesicles.² In addition, some cargo proteins rely on transmembrane receptors that interact with the coat proteins to facilitate interaction.³ One type of receptor are the p24/Transmembrane Emp24 Domain (TMED) proteins, a set of evolutionarily conserved, single pass transmembrane proteins that have been shown to facilitate selection or secretion of a variety of cargo proteins in yeast, Drosophila, humans, and other organisms.^4–8 TMED proteins have a short cytoplasmic tail that can interact with both COPI and COPII subunits, allowing the proteins to cycle between the ER and the Golgi.⁹ The luminal portion of the protein includes a coiled-coil domain and a globular GOLD domain that participate in multimerization and cargo recognition^10–12 (Figure 1A). Based on amino acid sequence and other features, TMED proteins are classified into four subfamilies (α, β, δ, γ), and can function as tetramers composed of one member from each group.^13,14,15,16

Figure 1. C. elegans TMED proteins include conserved transmembrane, coiled-coil, and GOLD domains.

A. Membrane organization of TMED proteins, using C. elegans TMED-1 as an example. TMED proteins include a single transmembrane domain, and a globular GOLD (“Golgi Dynamics”¹¹) domain separated from the membrane by a coiled-coil region, resulting in a “fishing rod” with a “protein hook” type structure.²³ TMED-1 structure obtained from AlphaFold,⁴⁸ and annotated using EzMol.⁴⁷ B. Predicted structure for each C. elegans TMED protein, obtained from AlphaFold,⁴⁸ and annotated using EzMol.⁴⁷ Structures are grouped according to subfamily, defined by sequence similarity (Figure 2). All are predicted to include a globular GOLD domain. All but TMED-12 are predicted to include a single transmembrane domain.

The functions for TMED proteins in animal development are incompletely understood. RNAi and genetic mutant studies in mouse and Drosophila demonstrate that at least some genes are essential, and loss of function results in lethality.^17–19 Genetic experiments in budding yeast (S. cerevisiae) argue that TMED proteins function as a tetramer that includes one member from each subfamily.¹⁶ Consequently, single mutants of different genes (or all genes of a subfamily, for subfamilies with multiple members) exhibit a set of shared phenotypes since disruption of any complex member eliminates function of the complex. A similar set of shared functions is observed for some Drosophila TMED genes, such that baiser (δ), logjam (γ), and eclair (α) mutants each exhibit a shared defect in depositing eggs and wing development, plus eclair and CHOp24 (β) both facilitate secretion of the signaling ligand Wnt.⁷ However, knockdown of the single δ subfamily member, baiser, does not impact Wnt secretion, suggesting that the critical components of TMED complexes in animal development may be varied, depending on the process. Furthermore, the potential for cell- and tissue-specific gene expression regulation in a multicellular organism suggests a diverse array of functions or complex compositions for TMED proteins in development that are not apparent in single-celled organisms.

Here we report a systematic evaluation of the TMED gene family in C. elegans. We describe eight TMED genes, including members from each of the four subfamilies, and select five for further study based on orthology and their broad expression in the animals. Homozygous mutants for any of the five genes are viable and fertile, but exhibit a shared profile of defects in locomotion, embryonic development and vulva morphology. Two γ subfamily genes (tmed-1 and tmed-3) exhibit functional redundancy, with certain defects only observed in double mutants. We identify a specific function for TMED genes in facilitating breakdown of the basement membrane between the somatic gonad and epithelial cells during development of the vulva, suggesting a role for TMED proteins in developmental tissue reorganization. Taken together, our results indicate that TMED proteins of each subfamily function in a shared set of biological processes, and establish a genetic and experimental framework for the study of TMED gene function in a multicellular animal.

Results

TMED genes in C. elegans encode representatives from all four protein subfamilies

We identified TMED genes in the C. elegans genome that have been classified by previous researchers in publications or as annotated in WormBase.^20,21 This yielded seven previously named and annotated genes (tmed-4, tmed-12, tmed-10, sel-9, tmed-13, tmed-3, and tmed-1), as well as C26C6.9, which was reported in WormBase as a paralog to the other genes, using the Panther classification system.²² Amino acid sequences of the predicted product for each gene include the characteristic GOLD domain, a globular structure composed of two four-strand antiparallel β sheets.²³ All proteins also include a predicted signal sequence and transmembrane domain, except for TMED-12 which is a non-standard TMED protein as it lacks both (Figure 1B). TMED proteins are classified into four subfamilies (α, β, γ, δ) based on sequence similarity.^13,14 To confirm the classification of the C. elegans proteins, we completed a comparison among TMED proteins from C. elegans, Drosophila melanogaster, Mus musculus and Saccharomyces cerevisiae (Figure 2). This analysis and that of others argues that the C. elegans genome includes at least one representative from each subfamily.²⁰

Figure 2. The C. elegans genome encodes eight TMED proteins with representatives from each of four subfamilies.

A phylogenetic tree built from TMED proteins sequences from C. elegans (CEL), Drosophila melanogaster (DRO), Mus musculus (MUS) and Saccharomyces cerevisiae (SCER). C. elegans proteins group with those of other species into four subfamilies (α, β, δ, γ) that are previously defined based on sequence clustering.^13,14 The C. elegans genome codes for one representative from the δ and β groups (TMED-10 and SEL-9, respectively), two representatives from the α group (TMED-4 and TMED-12), and four representatives from the γ group (TMED-13, TMED-3, TMED-1, and C26C6.9). The protein product of tmed-12 is an incomplete TMED protein, as it does not include a signal sequence or a transmembrane domain (Figure 1). The sequences used in the analysis and their associated accession numbers are provided in Supplemental Table 1.

To better understand how the TMED genes may be functioning in development, we cross-referenced each gene with the single cell RNAseq analysis from L2 larvae reported in Cao et al.²⁴ (Supplemental Table 2). These data show that tmed-4, tmed-10 and sel-9 transcripts are present across all cell types at this stage. tmed-1 and tmed-3 transcripts are present broadly across many cell types, but exhibit more varied cell type regulation²⁴. tmed-4, tmed-13, and C26C6.9 transcripts accumulate in highly restricted patterns, and are present in a limited number of cells. This analysis shows that all cell types express a member from the α, β, and δ subfamilies. In contrast, some cells (specifically, GABAergic neurons) lack any member of the γ subfamily, and many cell types express more than one γ gene. As we are interested in identifying broad roles for TMED proteins in development, we focused our subsequent analyses on the genes from each of the α, β, and δ subfamilies that encode canonical TMED proteins and are present in all cell types (tmed-4, tmed-10, sel-9), and two broadly expressed γ group genes (tmed-1, tmed-3).

TMED gene mutants exhibit behavioral and embryonic defects

We obtained, or generated, alleles that delete all or a portion of each of the five genes, and are predicted to reduce or eliminate gene function as each deletes important functional domains and/or shifts the gene reading frame (allele details in Experimental Procedures). Animals homozygous for each mutation are viable and fertile, but we observed some behavioral and developmental defects. tmed-4, tmed-10, and sel-9(tm6822) mutants exhibit uncoordinated movement (illustrated as body bends per minute, Figure 3A) whereas sel-9(tm5697), tmed-3 and tmed-1 mutants exhibit more normal behavior. Based on the nature of the sel-9(tm5697) allele (Experimental Procedures) and the impact on other traits (below), we hypothesize that this allele reduces sel-9 gene activity, but retains more activity compared to sel-9(tm6822). In contrast, alleles of each γ subfamily gene, tmed-3 and tmed-1, are predicted to be more disruptive to gene function. Since previous studies with yeast TMED genes demonstrated that there is the potential for functional compensation among genes from the same subfamily,^15,16 we constructed tmed-3; tmed-1 double mutants. We found that these animals exhibit uncoordinated movement, arguing that these two genes can compensate for each other in this case. In addition, while all of the mutations can be maintained as homozygous strains, we observed that all mutants except for tmed-3 exhibit a reduced frequency of viable offspring (Figure 3B), and the tmed-1 defect is not enhanced by coincident disruption of tmed-3. While future work will be required to identify the cellular reasons for these defects, the results indicate that TMED mutants exhibit similar properties with respect to movement and embryonic survival. Taken together, these data identify distinct relationships between the two γ subfamily genes, tmed-3 and tmed-1, in their ability to compensate for each other. For movement (and vulva morphology, described below), the two genes are functionally redundant and both must be mutated to observe a phenotype. In contrast, normal embryonic development is reliant on tmed-1, but not tmed-3, despite the fact that both genes are expressed broadly in the embryo.²⁵ Overall, based on these shared phenotypes among genes of all four subfamilies, we conclude that the TMED genes function in a shared set of biological processes.

Figure 3. TMED gene mutants exhibit defects in movement and embryonic survival.

A. TMED mutants exhibit movement defects, and decreased body bends per minute. The movement of tmed-1 and tmed-3 single mutants is similar to wild type, but double mutants exhibit movement defects, with data represented with a box plot. X identifies the mean for each sample, and outliers are represented by a circle. Asterisks indicate statistically different from wild type control (Tukey-Kramer test, P<0.01). Sample size for each condition is >=40 animals. B. TMED mutants produce a proportion of offspring that die embryonically, and do not hatch. This phenotype is absent in tmed-3(gu257) animals, and introduction of this mutation into the tmed-1(tm8108) genetic background does not enhance the phenotype. Error bars correspond to standard error. Asterisks indicate statistically different from control (proportional Z-test with Bonferroni correction, P<0.01). ns = not significant. Sample size for each condition is >=200.

TMED gene mutants exhibit defects in adult vulva morphology

An additional notable defect shared among TMED mutants is in the morphology of the adult hermaphrodite vulva, a structure that connects the gonad to the outside, and is important for the laying of eggs and mating with males.²⁶ The vulva structure is formed from the progeny of three vulval precursor cells (VPCs) that divide to produce a total of 22 cells that undergo a complex set of cell fusions and morphological steps to produce the adult structure.²⁷ In wild-type adult animals, the opening of this structure lies relatively flat with respect to the ventral surface of the animal. In single tmed-4, tmed-10 and sel-9 mutants, and tmed-3; tmed-1 double mutants, however, adult hermaphrodites exhibit a high frequency of a protruding vulva (Pvul) phenotype (Figure 4). To confirm that the protruding vulva defect is caused by the genetic disruption of the mutant TMED gene within the strains, we performed genetic rescue experiments (Figure 5A). We introduced genomic DNA corresponding to each gene and evaluated whether the transgene is capable of rescuing the mutant phenotype (see Experimental Procedures). We indeed found that each mutant (or double mutant in the case of tmed-3; tmed-1) can be rescued by introduction of the wild-type gene. Thus we conclude that the vulval defect in each case results from disruption of TMED gene function, rather than a background mutation in each strain.

Figure 4. TMED gene mutants exhibit a protruding vulva phenotype.

A.-J. Representative images from wild type and TMED mutants illustrating the vulva structure at L4 larval stage, and in young adults. A.-E. The cells and morphology of vulval cells in TMED mutants are similar to wild type at the L4 larval stage (n=10 for each genotype). G.-K. Adult TMED mutants exhibit a protruding vulva (Pvul) phenotype, where the vulva tissue protrudes noticeably out from the ventral cuticle (indicated with a dotted white line). This phenotype is absent in tmed-1 and tmed-3 single mutants, but is prevalent in double mutants. K. Error bars correspond to standard error. Asterisks indicate statistically different from control (proportional Z-test with Bonferroni correction, P<0.01). Sample size for each condition is >130. Full genotypes for images are wild type (N2), tmed-4(tm3823), tmed-10(tm6503), sel-9(tm6822), and tmed-3(gu257); tmed-1(tm8108).

Figure 5. The Pvul phenotype is rescued by introduction of wild-type transgenes, and in genetically chimeric animals bearing wild-type DNA in cells derived from the embryonic AB blastomere.

A. Rescue of the mutant protruding vulva (Pvul) phenotype with a transgene containing wild-type DNA for each gene. Control data Pvul are distinct from the data in Figure 4K as they were collected in parallel to the transgene-bearing strains. Error bars correspond to standard error. Asterisks indicate statistically different from control (proportional Z-test with Bonferroni correction, p<0.01). Sample size for each condition is >30. B. Schematic representation of the genetic chimera method²⁹. GPR-1 is overexpressed (GPR-1(oe)) in maternal oocytes, causing male and female pronuclei to segregate to daughter cells without fusing in the zygote. For this experiment, we evaluated the specific chimera class where replicated maternal chromosomes segregate to the AB blastomere (precursor to the VPCs), while replicated paternal chromosomes segregate to the P1 blastomere (precursor to the somatic gonad). A myo-2::mCherry reporter (hjSi20) expressed in pharyngeal cells is included in the maternal strain, and serves as a marker to identify chimeric animals, as it expresses in anterior pharyngeal cells when retained in AB, and posterior pharyngeal cells when retained in P1²⁹. C. Genetic chimeras for tmed-10 and sel-9 indicate that presence of wild-type gene product in AB-derived cells is sufficient for normal vulval morphology. In each case, animals homozygous for the relevant maternal genotype (including the transgene ccTi1594 that causes GPR-1(oe)) were crossed with wild-type males, or males bearing the mutant genotype. Data were collected from animals with maternal DNA in AB-derived cells (including the VPCs), and paternal DNA in P1-derived cells (including the somatic gonad and anchor cell), as indicated by the restriction of myo-2::mCherry expression to anterior pharyngeal cells. Animals receiving the mutant allele from each parent are genetically mutant in all cells, and similar to standard homozygous mutants (Figure 4). Animals receiving the wild-type allele from the mother that is segregated in the AB blastomere are phenotypically wild type, even if the TMED gene is mutant in P1-derived cells. In contrast, presence of the paternally-provided wild-type allele in P1-derived cells is not typically sufficient to compensate for a maternally-provided mutant genotype in AB-derived cells. Error bars correspond to standard error. Asterisks indicate statistically different from control animals with wild-type genotype derived from both parents (proportional Z-test with Bonferroni correction, P<0.01). Sample size for each condition is >=25.

We wanted to characterize the developmental defect that underlies the vulva morphology defect, to better understand how TMED proteins participate in production of this important reproductive structure. One cause of the protruding vulva phenotype can be abnormal cell division of the VPCs.²⁸ Since there is a stereotypic pattern of vulval cell number, placement and morphology in L4 larval animals that results from the reproducible pattern of developmental cell interactions and divisions,²⁶ we inspected this stage in TMED mutants. We found that tmed-4, tmed-10 and sel-9 mutants, and tmed-3; tmed-1 double mutant L4 animals exhibit no obvious defects (Figure 4), suggesting that the protruding vulva defect results from abnormalities in morphogenesis or cell fate defects not anatomically apparent, rather than a VPC cell lineage defect.

Normal development of the C. elegans vulva involves communication and coordinated development between cells of the somatic gonad and the vulva itself, to produce a functional structure.²⁶ To understand which tissue(s) require normal TMED gene function, we evaluated tmed-10 and sel-9 using the genetic chimera method of Artiles et al.²⁹ (Figure 5B). This method produces animals with a distinct genotype in cells derived from the embryonic AB blastomere (which include the VPCs) from cells derived from the P1 blastomere (which includes all cells of the somatic gonad). Using this method, we identified animals with the mutation in AB (VPCs) but the wild-type allele in P1 (somatic gonad) and vice versa, and found that presence of the wild-type allele in AB-derived cells is sufficient for a normal vulva structure, but presence of the wild-type allele in P1-derived cells generally yields the Pvul phenotype (Figure 5C). We conclude that while the genes may have functions in both tissues required for normal vulva morphogenesis, certain functions provided by AB-derived cells (presumably the VPCs) cannot be compensated for by other cells in the larva that may participate in the process.

TMED gene mutants exhibit defects in basement membrane clearing during vulva development

Finally, we aimed to understand the cellular defects that underlie the vulva development abnormality. Previous work has shown that to facilitate communication and connection between cells of the somatic gonad and the vulva, the anchor cell (AC) of the somatic gonad invades and mediates breakdown of the basement membrane between the two tissues.³⁰ This process includes production and activation of matrix metalloproteinases to hasten the breakdown, as well as active biophysical breaching by the AC.³¹ To ask whether TMED genes participate in basement membrane breakdown and remodeling, we crossed transgenes with fluorescently labeled laminin (LAM-1) into tmed-4, tmed-10, and sel-9 mutants, and found that LAM-1 breakdown is delayed (but not eliminated) in these mutants (Figure 6). We interpret that at least one function for TMED genes in vulva development is in promoting efficient basement membrane reorganization between the somatic gonad and epithelial tissue.

Figure 6. TMED gene mutants exhibit defects in breakdown of the basement membrane between the Anchor Cell and the developing vulval cells.

DIC (A.-D.) and fluorescent (E.-H.) images of wild type and mutant animals bearing tagged LAM-1 (LAM-1::mCherry or LAM-1::Dendra) to highlight basement membrane between the anchor cell and the vulval cells at the point where all VPCs have undergone two rounds of cell division but prior to morphological changes that are initiated by the cells derived from P6.p. In wild-type animals, a clear area of basement membrane removal (A.) and LAM-1 clearing (E.) is apparent, whereas this area is diminished or absent in many TMED mutant animals of a similar stage. The average LAM-1 gap is reduced in TMED mutants compared to wild type (I.). Asterisks indicate statistically different from wild type control (Tukey-Kramer test, P<0.01). Sample size for each condition is >=19 animals.

Discussion

In this study we systematically evaluate TMED genes in C. elegans. The C. elegans genome encodes eight TMED genes, including representatives from all four subfamilies. We used genetic mutants to evaluate the function of five broadly expressed TMED genes, including one gene from each α, β, δ subfamily (tmed-4, tmed-10, sel-9), and two γ subfamily genes (tmed-1 and tmed-3). We show that although the mutants exhibit a shared set of phenotypes, including defects in embryonic development, movement, and vulva morphology. In addition, we show that tmed-1 and tmed-3 exhibit functional redundancy, as both movement and vulva morphology defects are apparent only in double mutants. This ability of genes within a subfamily to compensate for each other is similarly observed in yeast cells, where a strong secretion defect for the Kar2 ER-resident chaperone protein is only observed when all subfamily members are deleted for groups with multiple members (such as the γ group members erp2, erp3, erp4).¹⁶ Similarly, in yeast and animals, elimination of genes representing any of the subfamilies yields a common phenotype and can influence the stability of proteins from other subfamilies.^8,15–17,32 These results, combined with co-immunoprecipitation data from yeast, support the conclusion that TMED proteins function in heteromeric complexes such as tetramers.¹⁶ Our results argue that C. elegans TMED proteins likewise function as tetramers in the biological functions we have evaluated, since α, β, δ, and γ double mutants exhibit a shared set of defects. However, we note that some C. elegans cells express only members from subfamily α, β, and δ, suggesting that although some key functions require a protein member from each subfamily, we do not rule out the possibility of alternative combinations or the possibility that proteins from different subfamilies can partially compensate for each other.

TMED proteins have a dual role in facilitating recruitment of specific cargo proteins to transport vesicles for efficient exit from the endoplasm reticulum (ER), but also removing defective molecules from the Golgi to return them to the ER. For example, the yeast TMED complex facilitates glycosylphosphatidylinositol (GPI)-anchored protein anterograde transport, but also retrieves from the Golgi and targets to the ER proteins that are soluble or have an un-remodeled GPI anchor, ensuring both efficient and accurate presentation of proteins on the membrane.⁵ Another quality control function for TMED proteins is in preventing misfolded proteins from being loaded onto secretory vesicles, and disruption of their activity will activate the unfolded protein response in cells.^33,34 Previous studies have demonstrated this role in protein quality control for C. elegans TMED genes. For example, sel-9 and tmed-10 were identified as modulators of specific alleles of Notch genes lin-12 and glp-1, and reduced TMED gene activity increases the presentation of mutant GLP-1 proteins on the cell membrane, when normally the mutant protein is retained in the cell.³⁵ Similarly, an allele of sel-9 was identified in a screen for mutations that ensure proper removal of inappropriate proteins by mediating the response to proteasome disruption,³⁶ and knockdown of tmed-3 results in increased accumulation of misfolded human α-synuclein in C. elegans cells.³⁷ In contrast to these earlier studies, we aimed to uncover functions for TMED genes in otherwise wild-type animals, to identify functions important for normal developmental processes, and have described specific developmental functions that are not dependent on a genetically sensitized background. While the specific cellular processes affected in the mutants remain to be discovered, we speculate that these mutants uncover processes where specific cargo proteins rely on a TMED receptor complex for efficient recruitment and sorting into secretion vesicles.

We show that TMED gene function is necessary for the normal timing for breakdown of the basement membrane between the somatic gonad and the vulva epithelium during anchor cell (AC) invasion. AC invasion is the initiating step in coordinating the vulva-uterine attachment, and involves a number of important gonad and AC-centered processes, including matrix metalloproteinase protein secretion, AC cell polarization, localized mitochondria and ATP production, focused glucose import and establishment of unique actin scaffolds.^31,38–40 Known roles for the vulval cells are more limited, although they are a likely source for LIN-3/EGF that mediates coordination between the AC and invaginating vulval cells, and they participate in the attachment to the uterine cells.^26,39 While failure to break down basement membrane can result in failure of attachment between uterine and vulval cells, the breakdown delay as we observe here only rarely results in a protruding vulva phenotype,^30,40 arguing that there are additional cellular roles for TMED genes. We anticipate that these functions reflect vulval cell functions, since we observe that presence of wild-type TMED gene function in AB-derived cells (including the vulval cells) is necessary and sufficient for normal vulva morphology in genetically chimeric animals. Future experiments will be necessary to identify what these specific functions are.

Experimental Procedures

Worm maintenance and genetics

Caenorhabditis elegans strains were grown on NGM plates seeded with Escherichia coli OP50 as a food source.⁴¹ All experiments were performed at 20C, unless otherwise noted. The wild type C. elegans is N2 Bristol. Specific strains and genotypes used are listed in Supplemental Table 3.

Production of transgenic lines

Genomic rescue experiments were performed using standard C. elegans methods.⁴² DNA fragments were generated by PCR amplification of each gene, plus at least 500 bp flanking DNA, from N2 (wild-type C. elegans) genomic DNA. PCR primer sequences are provided in Supplemental Table 4. PCR fragments were confirmed based on size using gel electrophoresis, incorporated into injection mixes (12 ng/ul pDP#MM016B (unc-119(+)), 12 ng/ul pCFJ90 (myo-2::mCherry), 25 ng/ul 1kb ladder (NEB), 25 ng/ul PCR product), and micro-injected into the gonad of homozygous mutant animals. Positive transformants were selected on the basis of mCherry fluorescence, and data in Figure 5A were collected from animals selected as mCherry+ in the L4 larval stage, and then allowed to mature 24 hours for data collection in the adult stage.

Deletion alleles of TMED genes

All mutant TMED gene alleles used in this study except for tmed-3(gu257) were generated and the precise deletion breakpoints were described by the National Bioresource Project for the Nematode (https://shigen.nig.ac.jp/c.elegans/). We confirmed the presence of the deletions using PCR, and repeat aspects of the description here, with added interpretation about the predicted alteration to the protein product.

tmed-4(tm3823) is a 327 bp deletion spanning from exon 2 to intron 3 of the gene. If this mutant allele produces a transcript that fails to remove intron 3, then the transcript is predicted to code for the signal sequence and about 1/3 of the GOLD domain before a frameshift and premature stop codon. If, instead, the smaller intron 3 is removed (using a potential cryptic splice site) in a manner to retain the frame, it will yield a protein lacking about 2/3 of the GOLD domain, but retaining the rest of the protein, including the signal sequence, coiled-coil and transmembrane domains. The FX03823 strain description (tmed-4(tm3823)/+) originally identifies tmed-4(tm3823) as a homozygous sterile or lethal allele, as PCR of genomic DNA from the strain consistently yields two products corresponding to the wild-type and deleted products. However, we were able to determine that the apparent wild-type product associated with the strain results from cross-reaction of the PCR primers with the tmed-12 locus. Redesign of primers (Supplemental Table 4) permitted specific tracking of the tmed-4 locus and genotype.

tmed-10(tm6503) is a 169 bp deletion spanning from exon 2 to exon 3 of the gene. This deletion removes the 50 bp intron 2 sequence, and therefore shifts the frame, with one altered amino acid before introduction of an in-frame stop codon. The transcript produced from this allele is predicted to code for the signal sequence and about 1/4 of the GOLD domain before the frameshift and premature stop codon. The FX16564 strain description (tmed-10(tm6503)/nT1[qIs51]) originally identifies tmed-10(tm6503) as a homozygous sterile or lethal allele, but in outcrossing we were able to recover the allele as a homozygote, likely by removal of a linked second mutation.

sel-9(tm5697) is a 487 bp deletion plus 7 bp insertion spanning from exon 1 to intron 2 of the gene. If this mutant allele produces a transcript that fails to remove intron 2, then the transcript is predicted to code for the signal sequence and about 1/3 of the GOLD domain before a frameshift and premature stop codon. If, instead, the smaller intron 2 is removed (using a potential cryptic splice site) in such a way that frame of the predicted product is retained, it will yield a protein lacking about 2/3 of the GOLD domain, but retaining the rest of the protein, including the signal sequence, coiled-coil and transmembrane domains. Based on the phenotype associated with homozygotes bearing this mutant allele, we speculate that it retains some gene activity compared to the sel-9(tm6822) allele.

sel-9(tm6822) is a 548 bp plus 1 bp insertion deletion removing sequences upstream of the presumed transcriptional start site, as well as the first two exons. While the deletion may impact production of a transcript (due to loss of upstream sequences and the presumed promoter), if a transcript is produced the first start codon is in exon three, and out of frame with the wild-type sequence. If somehow a transcript is produced that is translated in-frame, the encoded protein would lack a signal sequence, the GOLD domain, and about 1/4 of the coiled-coil domain. Based on the nature of the mutation, we anticipate that it impacts sel-9 gene function to a greater extent than does sel-9(tm5697).

tmed-1(tm8108) is a 132 bp deletion removing sequences upstream of the presumed transcriptional start site into the first exon, removing the first 24 codons, including the start codon. While the deletion may impact production of transcript (due to loss of upstream sequences and the presumed promoter), if a transcript is produced the first start codon is present in exon 2. While in-frame with the wild-type sequence, utilization of this internal start codon would yield a protein lacking the signal sequence and about 1/3 of the GOLD domain.

Generation of tmed-3(gu257)

tmed-3(gu257) was generated using the method and plasmids of Dickinson et al.,⁴³ with sgRNAs in pDD162 (one targeting the 5’ end of the gene, and one targeting the 3’ end of the gene). A repair template was assembled in pDD282 using the NEBuilder HiFi assembly kit, with flanking homology arms corresponding to sequences upstream and sequences downstream of the tmed-3 coding exons. This design results in a mutant allele in which the coding exons are replaced with the plasmid intervening sequences (GFP > SEC > FLAG). These plasmids were co-injected into the gonad of wild-type C. elegans animals together with a negative selection fluorescent marker (myo-2::mCherry). Candidate repair events were selected (hygromycin^R mCherry−), subjected to SEC excision via heat shock, and validated with PCR and Sanger sequencing.

Sequence analysis

The NCBI accession number and amino acid sequences used in this study are listed in Supplemental Table 1. TMED protein domains were identified using Quick2D to identify beta sheet areas (GOLD domain), TMHMM to identify transmembrane sequences,⁴⁴ and DeepCoil to identify coiled-coil sequences.⁴⁵ The phylogenetic tree in Figure 2 was generated using the Maximum Likelihood, JTT matrix-based model in MEGA.⁴⁶ We utilized the TMED subfamily classifications (α, β, δ, γ) of Carney and Bowen¹³ and Strating et al.¹⁴ We used the protein modeling database AlphaFold to identify existing protein models for each of the eight C. elegans TMED proteins, and used EzMol to annotate protein domains on the structures.^47,48

Phenotypic analyses

Animal movement (Figure 3A) was evaluated as in Sawin et al.⁴⁹ Briefly, well-fed L4 animals were selected and plated onto plates seeded with bacteria (1-2 per plate). For each animal, the number of body bends was counted for 2 minutes in 20 second intervals and recorded. Plates were coded so that data were collected blind with respect to genotype. Embryonic lethality (Figure 3B) was evaluated by selecting eggs from well-fed parents, and placing them together on fresh seeded plates (10 per plate). After 48 hours, the plates were counted for the total number of L4 or adult animals.

Microscopy

Vulval cell morphology was evaluated in L4 and adult animals using DIC optics at 100x magnification (Figure 4). Fluorescently-tagged LAM-1 protein localization was observed in L3 animals under both DIC and fluorescent imaging at 100x magnification (Figure 6). Stage of development was determined first based on the (inferred) number of divisions completed by the P6.p cell. For LAM-1 localization, data were collected from animals in which P5.p-P7.p had all completed two rounds of division, but P6.p had not yet initiated morphological changes or movement toward the anchor cell. The length of the cleared area was measured using a straight line and the measure tool in imageJ.⁵⁰ LAM-1::mCherry (qyIs127) was evaluated in tmed-4(tm3823) and LAM-1::Dendra (qyIs108) in tmed-10(tm6503) and sel-9(tm6822) as the qyIs127 insertion is close to tmed-10 and sel-9 on Chromosome V. Wild type control images were collected using each transgene.

Genetic chimeras

Genetic chimeras were generated using the method of Artiles et al.²⁹ tmed-10(tm6503) or sel-9(tm6822) was crossed into strains bearing transgene ccTi1594 (which causes over-expression of GPR-1 in the germline, yielding segregation of maternal and paternal DNA into separate blastomeres in the zygote, Figure 5B) to produce maternal strains CM2906 (ccTi1594; hjSi20; tmed-10(tm6503)), and CM2909 (ccTi1594; hjSi20; sel-9(tm6822)). The strain PD2217 (ccTi1594 unc-119(ed3); hjSi20) served as the wild-type maternal strain. Hermaphrodites from these strains were crossed with control (N2) males or heterozygous genetic mutant males (tmed-10(tm6503)/nT1(qIs51) or sel-9(tm6822)/nT1(qIs51)). Chimeric offspring in which the maternal genome segregates into AB (and paternal into P1) exhibit a characteristic pattern of fluorescence (mCherry from hjSi20) in the anterior pharynx that is easy to identify under a dissecting microscope²⁹. Chimeric offspring with mCherry+ localized to the anterior pharynx, that also lacked GFP (indicating animals lacking the paternal balancer chromosome nT1(qIs51)) were selected as L4 larvae, and allowed to mature 24 hours to evaluate the adult vulva phenotype.

Bullet points:

• The C. elegans genome includes representatives from all four TMED gene subclasses

• C. elegans TMED mutants exhibit a shared set of phenotypic defects

• C. elegans tmed-1 and tmed-3 exhibit functional redundancy

• C. elegans TMED mutants exhibit a delayed basement membrane breakdown during vulva development, and a protruding vulva phenotype