Loss of the endoplasmic reticulum protein Tmem208 affects cell polarity, development, and viability

Significance

The biological significance of TMEM208 is unknown in multicellular organisms. We show that loss of Tmem208, the fly ortholog of human TMEM208, results in lethality, and escapers exhibit defects in the wing and eye indicative of a disruption in planar cell polarity (PCP). Tmem208 binds the PCP receptor Frizzled and helps maintain its proper levels. An individual who carries biallelic loss-of-function alleles of TMEM208 presents with developmental delay and a multisystem disorder consistent with PCP defects.

Abstract

Nascent proteins destined for the cell membrane and the secretory pathway are targeted to the endoplasmic reticulum (ER) either posttranslationally or cotranslationally. The signal-independent pathway, containing the protein TMEM208, is one of three pathways that facilitates the translocation of nascent proteins into the ER. The in vivo function of this protein is ill characterized in multicellular organisms. Here, we generated a CRISPR-induced null allele of the fruit fly ortholog CG8320/Tmem208 by replacing the gene with the Kozak-GAL4 sequence. We show that Tmem208 is broadly expressed in flies and that its loss causes lethality, although a few short-lived flies eclose. These animals exhibit wing and eye developmental defects consistent with impaired cell polarity and display mild ER stress. Tmem208 physically interacts with Frizzled (Fz), a planar cell polarity (PCP) receptor, and is required to maintain proper levels of Fz. Moreover, we identified a child with compound heterozygous variants in TMEM208 who presents with developmental delay, skeletal abnormalities, multiple hair whorls, cardiac, and neurological issues, symptoms that are associated with PCP defects in mice and humans. Additionally, fibroblasts of the proband display mild ER stress. Expression of the reference human TMEM208 in flies fully rescues the loss of Tmem208, and the two proband-specific variants fail to rescue, suggesting that they are loss-of-function alleles. In summary, our study uncovers a role of TMEM208 in development, shedding light on its significance in ER homeostasis and cell polarity.

TMEM (TransMEMbrane) proteins are present in the cell membrane as well as other membrane-bound organelles including the nucleus, mitochondria, lysosome, and endoplasmic reticulum (ER). In humans, the TMEM protein family consists of over 300 members, many of which have been associated with cancer, neurodegeneration, and a range of genetic disorders (1–10). Many of these TMEM proteins are localized to the ER. Despite their abundance, the functional importance of these ER-localized TMEM proteins in multicellular organisms remains poorly characterized. To gain insight into the role of ER-resident TMEM proteins in development, in this study, we conducted an RNA interference (RNAi) screen targeting conserved ER-resident TMEM proteins in fruit flies. This screen revealed that many genes encoding ER-localized TMEM proteins are essential for fruit fly (Drosophila melanogaster) development. Through the Undiagnosed Diseases Network (UDN), we identified an individual with compound heterozygous variants in TMEM208 who presents with developmental delay, dysmorphism, multiple hair whorls, seizures, and other developmental abnormalities involving multiple organs. These data prompted us to assess the role of this gene in development in more detail using fruit flies to establish a human disease model. Based on experiments in yeast and human cell lines, TMEM208 has been implicated in the transport of nascent polypeptides into the ER via the signal recognition particle (SRP)-independent (SND) pathway (11, 12). Proteins destined for the cell membrane, endomembrane system, and secretory pathway are first targeted to the ER before they are sent to their destination. Proteins with an N-terminal signal peptide and/or transmembrane domain are transported to the ER cotranslationally via another pathway, the SRP pathway (13, 14). However, proteins with a C-terminal transmembrane domain are transported to the ER posttranscriptionally via the Guided Entry of Tail (GET) pathway (13, 14). The Snd2 protein in yeast, the ortholog of human TMEM208/hSnd2, was discovered as a component of the SND pathway which shows a preference toward proteins with an internal transmembrane domain (11). Genetic experiments in yeast suggest that the SND pathway acts as a backup for the other two pathways and that some proteins can use SND and SRP/GET pathways (SI Appendix, Fig. S1A). Based on experiments in human cell lines, the SND pathway was shown to be dependent on TMEM208 (12, 15). However, the phenotypes associated with loss of TMEM208 in multicellular organisms have not yet been established.

To study the functional consequences of Tmem208 loss, we generated a fly mutant, assessed the expression pattern of this gene, and characterized the loss-of-function phenotypes. We also generated a GFP-tagged Tmem208 allele to determine the subcellular localization of the Tmem208 protein. Loss of the gene results in lethality with few escapers exhibiting reduced lifespan, neurological issues, some planar cell polarity (PCP) defects, as well as other developmental defects. A human TMEM208 transgene rescues the lethality of the mutants, indicating functional conservation. Loss of Tmem208 also induces a mild ER stress. Furthermore, our data revealed that Tmem208 interacts with Frizzled (Fz), a component of the PCP pathway, and helps to maintain proper levels of Fz. Using the humanized fly models, we demonstrated that the variants identified in a UDN subject are loss-of-function alleles. Altogether, our findings show that Tmem208 and TMEM208 are evolutionarily conserved players in multicellular development, and their loss results in multisystem defects.

Results

ER-Resident TMEM Proteins Are Important for Fruit Fly Development.

In an attempt to identify the possible subcellular localization of ~300 human TMEM proteins (www.uniprot.org), we performed a literature survey. This revealed that a significant fraction (n > 60) of human TMEM proteins are localized to the ER membrane, and ~70% of the human ER-resident TMEM protein-encoding genes have at least one ortholog in flies, and ~60% of these are highly conserved (Drosophila RNAi Screening Center Integrative Ortholog Prediction Tool (DIOPT) score ≥7/15) (16). To evaluate the role of the highly conserved ER proteins in development, we targeted 38 TMEM genes in an RNAi screen. We performed either a ubiquitous knockdown [Tubulin(Tub)-GAL4] or a wing-specific knockdown [nubbin(nub)-GAL4] that drives expression in the pouch region of the larval wing disc using the UAS-GAL4 system (17). Thirty RNAis caused a partial or complete lethality upon ubiquitous gene knockdown and 12 of the 38 genes caused a phenotype upon wing-specific gene knockdown. Eleven of the RNAi lines cause both lethality and a wing phenotype (SI Appendix, Table S1). In some instances, such as the knockdown of CG14646 (ortholog of human TMEM129) and CG30389 (ortholog of human TMEM57), wing-specific knockdown resulted in severely affected wings (SI Appendix, Fig. S1B). In other instances, RNAi-mediated knockdown resulted in a spectrum of phenotypes including wing notching, blistering, wing vein and crossvein malformation, as well as crumpled wings (SI Appendix, Fig. S1B).

One of the candidates identified in the screen was CG8320 an as yet-uncharacterized fly ortholog of human TMEM208. The corresponding human and fly proteins share 43% identity and 65% similarity in amino acid composition (DIOPT score 14/15) as well as a similar topology (SI Appendix, Fig. S1 C and D). Specifically, ubiquitous knockdown of Tmem208 results in lethality while wing knockdown produces blisters, crossvein formation defects, and PCP-like defects. The latter phenotype drew our interest as the UDN identified an individual with compound heterozygous variants in TMEM208. To interrogate the relevance of those TMEM208 variants, we sought to further characterize the function of this gene in flies.

Tmem208 Is Broadly Expressed, and Its Loss Causes Lethality.

To evaluate the role of Tmem208 in development, we generated a knockout allele, Tmem208^KG4, using CRISPR-induced homologous recombination (18). Two guide RNAs targeting the 5′ and 3′Untranslated Region (UTRs) of the endogenous gene along with the donor plasmid with left and right homology arms were used to replace the entire open reading frame of Tmem208 with a Kozak-GAL4 (GAL4 with a Kozak sequence for efficient translation) sequence (Fig. 1A). This allele expresses the GAL4 gene under the endogenous promoter of Tmem208. We used this allele to drive nuclear-localized reporter transgene UAS-mCherry.NLS (nuclear localized fluorescent protein) and examined the expression pattern at different developmental stages (19, 20). The gene is expressed in larval, pupal, and adult stages in fruit flies (Fig. 1B). In the 3rd instar larvae, Tmem208 is broadly expressed in almost all tissues including wing disc, leg disc, eye disc, brain, salivary gland, anterior gut, fat body, and Malpighian tubules (Fig. 1C). We also noted expression of Tmem208 in the adult fly brain (SI Appendix, Fig. S2A). To identify the specific cell populations in the brain that express Tmem208, we used Tmem208^KG4> UAS-mCherry.NLS flies and performed costaining against mCherry and either Elav (Embryonic lethal abnormal vision), a pan-neuronal protein, or Repo (Reversed polarity), a glial protein (21). As shown in SI Appendix, Fig. S2A, Tmem208 is expressed in subsets of neurons and glia of the adult brain.

Fig. 1.

CG8320/Tmem208 is a broadly expressed gene essential for survival. (A) Generation strategy of Tmem208^KG4 allele. (B) Expression of Tmem208^KG4-driven mCherry.NLS at different stages of the fly lifecycle. (C) Expression of Tmem208^KG4-driven mCherry.NLS in different larval tissues. Upper panel (from Left to Right): wing disc, leg disc, eye disc, and larval brain; Lower panel (from Left to Right): salivary gland, anterior gut, fat body, and Malpighian tubule. Scale bar: 100 µm. (D) Tmem208 transcripts are undetectable in Tmem208^KG4 mutants. (E and F) Tmem208^KG4is a strong loss-of-function allele that is complemented by introduction of one copy of a P[acman] genomic rescue (GR) construct (22). To determine the statistical significance between two genotypes, a two-tailed Student’s t test was used. To determine the statistical significance between three genotypes, one-way ANOVA followed by a Tukey’s post hoc test was used. Error bars represent SEM (****P < 0.0001).

The Tmem208^KG4 allele results in semilethality with ~10% escapers when tested in heterozygous flies carrying one of two molecularly defined deficiencies (Df  ), (Df(2R)Exel7138/CyO and Df(2R)BSC308/CyO). Quantitative RT-PCR data revealed no transcript in Tmem208^KG4/(Df(2R)Exel7138 animals (Fig. 1D). Escapers display a significantly reduced life span when compared to control flies (SI Appendix, Fig. S2B). The lethality associated with the Tmem208 mutants is fully rescued by introducing a genomic rescue (GR) construct (22) that contains a copy of the Tmem208 gene (Fig. 1 E and F). In summary, Tmem208 is broadly expressed, important for survival, and escapers have a reduced lifespan.

Tmem208 Mutants Display Cell Polarity Defects.

In addition to the reduced lifespan, Tmem208^KG4/Df mutant escapers displayed a spectrum of morphological defects. For example, in 29.6% of the escapers (n = 140), we observed a defect in tarsal development (kinked segment) (SI Appendix, Fig. S2C). We also noted morphological anomalies in eye and wing tissues. Retinal sections revealed a variety of subtle defects. The fly eye consists of ~800 units called ommatidium. Each ommatidium typically consists of eight photoreceptor (PR) neurons (R1-R8) with each having a prominent rhabdomere that plays a critical role in phototransduction. Six of the PRs rhabdomeres (R1 to R6) are peripheral, whereas two PRs rhabdomeres (R7 and R8) are central (23). The R7 and R8 PR are on top of each other, and only seven PR/rhabdomeres are visible in each section (Fig. 2A). In some mutant ommatidia, we observe rhabdomere loss, whereas in other ommatidia, there are more than seven rhabdomeres as well as elongated rhabdomeres (Fig. 2 B and B’ ). Additionally, in several instances, we observe a “supernumerary PR”, a cell that looks like another R3 PR leading to an M shape of the PR organization rather than a trapezoidal shape (Fig. 2B). We also noticed a mild PCP defect-like phenotype with miss-rotated ommatidia (Fig. 2 A’ and B’ ). Transmission electron microscopy (TEM) demonstrates that the extra-rhabdomere phenotype is due to the presence of two rhabdomeres in the same PR (Fig. 2 A” and B” ). This defect appears to be an apical–basal polarity defect (24–27). During development, apicobasal polarity is established in the PR cells, and the apical membrane then reorganizes to form the stalk and rhabdomeres (28). Loss of proteins that establish apicobasal polarity during rhabdomere formation can cause similar phenotypes as those shown in Fig. 2 B and B”. In sum, these data suggest that loss of Tmem208 causes mild cell polarity defects in the fly eye.

Fig. 2.

Loss of Tmem208 causes developmental defects and impairs cell polarity. (A and B) Retinal images of control and escaper mutant flies. (A’ and B’) Arrow diagram showing the polarity of each ommatidium in the retina. “o” indicates missing rhabdomere and “+” indicates extra rhabdomere. Black and Red arrows indicate Dorsal and Ventral chirality. (A” and B”) TEM images of single ommatidia from the control (Upper) and mutant (Lower) animals. Scale bar: 2 µm. (C–E) Wing images from control and two Tmem208 escaper mutant flies. (C’ and D’) Higher magnification images from wing from control (Left) and mutant (Right) showing the orientation of wing bristles. (F) The area in the red box indicates the region used for the images shown in C’ and D’. The wing image in F was drawn using Biorender.

We also observed wing defects in escapers. Some of the escapers (31.4%, n = 86) exhibit anterior crossvein formation defects (Fig. 2 C and D). In addition, 13.2% (n = 53) of the escapers display improper wing folding (Fig. 2E), and ~10% of escapers display locally misaligned wing hairs, a phenotype reminiscent of PCP or PCP effector defects (Fig. 2 C’–F). Hence, the data again suggests that loss of Tmem208 causes a PCP-like phenotype in eye and wing with incomplete penetrance.

Tmem208 Encodes an ER-Resident Protein, and Its Loss Induces Mild ER Stress.

To assess the subcellular localization of Tmem208 protein in vivo, we generated a Tmem208-GFP allele. A Green Fluorescent Protein (GFP) sequence, along with linker sequences, was inserted in Tmem208 (between amino acids R99 and E100) using CRISPR-mediated homologous recombination (Fig. 3A) (18). The resulting allele produced an internally GFP-tagged Tmem208 protein. Animals carrying the Tmem208-GFP allele in homozygous condition or in trans with the Tmem208^KG4 allele are lethal. However, introducing one copy of GR rescued the lethality associated with this allele (SI Appendix, Fig. S3A), indicating that it is a loss-of-function allele. We used the salivary gland and fat body from heterozygous Tmem208-GFP larva and performed immunostaining for GFP to assess the protein localization and costained for Calnexin, an ER marker. As shown in Fig. 3B and SI Appendix, Fig. S3B, GFP and Calnexin colocalize in the ER indicating that Tmem208-GFP is properly localized.

Fig. 3.

Tmem208 encodes an endoplasmic-reticulum resident protein, and its loss causes mild ER stress. (A) Nature of the Tmem208-GFP allele. (B) Colocalization of Tmem208-GFP and calnexin proteins in the salivary gland. Scale bar: 5 µm. (C) Control and mutant wing discs with Bip/GRP78 staining and its quantification. (D) Western blot and quantification showing relative Bip/GRP78 protein levels in Tmem208^KG4 mutants. (E) Western blot and quantification showing relative p-Eif2α protein levels in Tmem208^KG4 mutants. (F) Anti-GFP (for detecting Xbp1-GFP protein) staining in the pouch area of the wing disc from the animals of the indicated genotype and related quantification. Scale bar: 50 µm. The fold changes in the experimental samples are relative to the control samples. To determine the statistical significance between two genotypes, two-tailed Student’s t test was used. Error bars represent SEM (*P < 0.05; **P < 0.01).

Next, we explored the cellular consequences of loss of Tmem208. An earlier study reported increased ER stress upon knockdown of TMEM208 in human U2OS epithelial-like cell lines (29). We therefore evaluated the levels of different ER stress markers in Tmem208 mutants by performing western blot and/or immunocytochemistry for the ER stress markers. When unfolded proteins accumulate in the ER lumen, Bip (encoded by Hsc70-3, the fly ortholog of human HSPA5/GRP78) senses these proteins and initiates an ER stress response (30–32). We measured the Bip protein level in Tmem208^KG4 mutants using two different approaches: First, we performed immunostaining using the anti-Bip antibody in the wing discs of Tmem208^KG4/Df larvae; second, we performed western blot analyses to detect the Bip protein levels in these mutants. As shown in Fig. 3 C and D, a ~1.5-fold increase in the Bip protein levels was observed in both wing discs as well as in mutants. Subsequent signaling cascades lead to the phosphorylation of Eif2α and repression of the translational machinery (30, 33). Consistent with this, the p-Eif2α protein levels were elevated in Tmem208 mutants when compared to controls (Fig. 3E). Finally, upon initiation of ER stress, alternative splicing allows for the expression of Xbp1 (32, 34, 35). We measured the levels of Xbp1 upon RNAi-mediated knockdown of Tmem208 using a transgene in which Xbp1 is located upstream of a GFP construct. In this reporter line, ER stress induces Xbp1 mRNA splicing leading to the expression of Xbp1-GFP (32, 35). As shown in Fig. 3F, an increase in the GFP levels is observed in the pouch region of 3rd instar larval wing discs upon knockdown of Tmem208. Taken together, these data indicate that loss of Tmem208 causes a modest but consistent ER stress response.

Tmem208 Helps to Maintain Fz Protein Levels.

Since Tmem208 null mutants displayed a PCP-like phenotype and Fz-signaling regulates establishing PCP, we subsequently tested whether Tmem208 interacts with Fz. PCP orients cells to create a polarity within tissue. Several of the proteins in this pathway are transmembrane membrane proteins (36) and are imported into the ER prior to being transported to the cell membrane. In the classical Fz-signaling pathway, six proteins are involved (36, 37). Three of these proteins (Prickle, Diego, and Dishevelled) do not have a transmembrane (TM) domain, whereas the other three proteins (Fz, Van Gogh, and Starry night/Flamingo) contain C-terminal or internal TM domains, suggesting that Tmem208 may play a role in facilitating the transport of these latter proteins. To assess this possibility, we tested whether Fz interacts directly with Tmem208 using immunoprecipitation (IP). We ubiquitously overexpressed HA-tagged Tmem208 and performed IP using anti-HA antibody, followed by western blot analysis targeting against Fz. Our results showed that endogenous Fz indeed immunoprecipitated with Tmem208 (Fig. 4A), providing evidence that these two proteins interact.

Fig. 4.

Tmem208 interacts with Fz and helps maintain its levels. (A) Western blot from co-IPs showing an interaction between Fz and Tmem208 proteins. (B and C) Western blot and quantification showing the relative level of Fz protein in Tmem208^KG4 escaper mutant flies. (D) Images of wing discs showing the Fz protein levels of the indicated genotypes. Scale bar: 50 µm (lower magnification images) and 2 µm (higher magnification images). The Lower panel shows the merged images for Fz and DAPI. (E) Relative fold change of Fz protein in the wing discs of the indicated genotypes. To determine the statistical significance between two genotypes, two-tailed Student’s t test was used. Error bars represent SEM (*P < 0.05; **P < 0.01).

Next, we measured the effects of Tmem208 loss on Fz protein levels using western blotting. As shown in Fig. 4 B and C, Fz protein levels are reduced in Tmem208^KG4/Df escaper flies. In addition, we performed immunostaining in the larval wing disc to detect the levels of Fz protein. However, we were not able to detect endogenous Fz protein in the 3rd instar larval wing discs. We therefore overexpressed Fz either in a wild-type background or in a Tmem208-GFP mutant heterozygous background and consistently observed decreased Fz protein levels in the wing discs of Tmem208-GFP heterozygous larva (Fig. 4 D and E). Taken together, our data indicate that Tmem208 interacts with Fz and plays a role in maintaining proper Fz protein levels.

Biallelic Variants in TMEM208 Are Associated with Developmental Defects in Humans.

Through the UDN, we identified a 5-y-old child with global developmental delay and a multisystemic disorder (SI Appendix, Fig. S4A). The individual was born with neonatal respiratory distress, gut malrotation, and dysmorphic features. Additionally, the subject has abnormal skeletal and ocular features, lymphopenia, and heart defects. Other features included failure to thrive, perioral cyanosis with feeding, seizures, idiopathic intracranial hypertension with papilledema, hypoglycemia, idiopathic dilatation of the main pulmonary artery and aortic root, two posterior parietal hair whorls, frontal prominence, bilateral epicanthal folds, micrognathia, short neck, and a mild 5th finger clinodactyly (see SI Appendix, Supplementary Clinical Information, for more detail). Genome sequencing revealed compound heterozygous variants in TMEM208 (NM_014187.3) (SI Appendix, Fig. S4B). The first variant (NM_014187.3: c.80T>C, p.L27P) is a point mutation (Fig. 5A) and the CADD score is 25.1 (38). This allele is present in trans to a frameshift variant (NM_014187.3: c.177delT, p.F59fs*13) (Fig. 5A) that likely leads to the production of a truncated TMEM208 protein (predicted to be truncated in the middle of the second transmembrane domain) or the absence of protein because of nonsense-mediated decay.

Fig. 5.

TMEM208 variants are loss of function in nature. (A) Schematic of the TMEM208 gene and protein with the relative position of the variants. (B) Graph showing the extent of rescue upon expression of human TMEM208 transgenes. (C) Wing images showing the bristle defects and their rescue with human TMEM208 transgene expression. The area in the red box indicates the region used for the images. The wing image in C was drawn using Biorender. (D) Duration of bang sensitivity in adult escapers with loss of Tmem208. (E and F) Percentage of flies with the indicated genotypes showing bang sensitivity at two different time points, including young flies at 3 to 5 d old (E) and aged flies at 21 to 23 d old (F). Error bars represent SEM (*P < 0.05; **P < 0.01; ****P < 0.0001). (G) Graph showing the XBP1 exon 4 usage in fibroblasts from controls and the affected individual. The rate of splicing out part of exon 4 was quantified in each sample (n = 3), normalized against the exon 4 to 5 junction that is present in all isoforms. Differences in exon 4 usage were tested using a 2-sample t test. (H) Relative CHOP protein levels at different time points upon a 12-h thapsigargin treatment of fibroblasts from unaffected control and affected individual.

We assessed the nature of the TMEM208 variants identified in the UDN subject using a “humanized” fruit fly model. As shown in Fig. 5B, loss of Tmem208 causes lethality in fruit flies, and expression of the human TMEM208 reference transgene rescues the lethality of these mutants. In contrast, expression of the transgene encoding the p.L27P variant only partially rescued the lethality, whereas the p.F59Lfs*13 did not significantly rescue the lethality (Fig. 5B). These data suggest that both variants are loss-of-function alleles: The p.L27P behaves as a mild loss-of-function allele while p.F59Lfs*13 is a strong loss-of-function allele. In addition, the expression of reference TMEM208 rescues the PCP-like defects, whereas expression of the variants only partially rescues the PCP phenotypes (Fig. 5C).

Since the patient developed seizures, we tested whether loss of Tmem208 causes a seizure-like phenotype in tmem208^KG4/Df escaper flies. A mechanical stress, caused by a brief vortex stimulation, causes immobility as well as involuntary movements similar to the seizure-like behavior in bang-sensitive flies, whereas the wild-type flies are relatively unaffected by this stimulus (39). We noted minor bang sensitivity in the young mutant flies when compared to control flies (Fig. 5D). The bang sensitivity in Tmem208^KG4/Df escapers is significantly rescued by the reference transgene, whereas the variants are less effective at rescuing the bang sensitivity (Fig. 5E). Although there is no significant difference in rescue ability of the bang sensitivity phenotypes between the reference TMEM208 and p.L27P variant in young flies, when the flies are aged, we observed an increase in bang sensitivity for the flies expressing TMEM208 p.L27P variant (Fig. 5F). These data support our observations that the p.L27P acts as a partial loss-of-function variant.

We next focused on determining the mechanism by which the p.L27P variant impacts TMEM208 function. We first assessed the subcellular localization of the reference and variant protein in the wing discs. We did not observe any difference in localization between the reference and p.L27P variant TMEM208 proteins (SI Appendix, Fig. S4 C and D). We also assessed the ER stress levels in the patient-derived fibroblasts. First, we performed RNA-seq on skin fibroblasts from the patient and control and assessed usage of the alternative isoform of XBP1 with partial exon 4 exclusion, a known marker of ER stress (34). Normalization of the exclusion against the exon 4 to 5 junction present in all XBP1 isoforms showed that there were a significantly higher proportion of XBP1 transcripts with partial exon 4 in the patient-derived fibroblasts (UDP11522) compared to control cells (Fig. 5G). Additionally, during exposure to thapsigargin, an ER stressor, levels of the protein CHOP (encoded by the human DDIT3 gene), a transcription factor that is up-regulated with ER stress (40, 41), increased continuously in patient cells during the12-h treatment, unlike control cells (Fig. 5H). These data indicate that the patient cells have a decreased ability to respond to ER stress.

Discussion

Tmem208 is broadly expressed during fruit fly development, and its loss causes lethality with few escapers. The latter show a reduced lifespan, seizure-induction-sensitivity, as well as defective leg, eye, and wing development. Our data also revealed cell polarity defects including PCP-like defects in the eyes and wings of flies when Tmem208 is lost. Consistent with a role of this gene in multicellular development, an individual with compound heterozygous variants in TMEM208 showed developmental anomalies, dysmorphism, and seizures. We also observe an ER stress response upon loss of Tmem208/TMEM208 in fruit flies and patient cells. Hence, we argue that loss of fly Tmem208 and its human ortholog impairs development, at least in part, by causing cell polarity and ER defects.

During multicellular growth and development, cell polarity acts at two levels: apicobasal polarity at the cellular level and PCP at the tissue level (37, 42–44). Somatic mutations in apicobasal polarity-related genes typically cause cancers. On the other hand, germline mutations in PCP-related genes cause developmental defects (37, 42, 44, 45). Mutations in PCP pathway genes in humans cause Robinow syndrome and Joubert syndrome, both characterized by a spectrum of symptoms including developmental delay as well as neurological, skeletal, cardiac, and kidney problems (46–51). In mice, loss of PCP leads to developmental defects including abnormal hair pattern formation, as well as cardiac, skeletal, and central nervous system defects (37, 52–57). The transmembrane proteins that help in establishing cell polarity are processed via the ER. Hence, impaired ER function may affect the polarity pathway. The proband with autosomal recessive variants in TMEM208 displays symptoms that can be related to the phenotypes observed in mice and humans with impaired PCP. These include developmental delay, skeletal, cardiac, and neurological issues, as well as hair patterning defects. Interestingly, the proband also has intercranial hypertension. The CSF inside the brain ventricles moves by synchronized activity of cilia, and the ciliary pattern formation as well as the directional flow of the fluid is regulated by PCP (58–62). Hence, improper beating of the cilia due to abnormal PCP may cause this hypertension.

The proband also presented with seizures. Mutations in PCP-signaling genes, PRICKLE1 and PRICKLE2, have been identified to cause seizures in humans (63). Consistent with these observations, loss of Prickle in mice or flies also causes seizure phenotypes (63) suggesting a connection between PCP and seizure. In summary, the observed PCP defects in mice and humans are consistent with some of the phenotypes observed in the proband. However, the symptoms are relatively mild in the proband probably because of the loss of TMEM208 function is partial and TMEM208^L27P likely retains quite a bit of function. It should also be noted that the proband has other symptoms including metabolic issues, microphthalmia, and lymphopenia. Impaired ER homeostasis and up-regulated ER stress cause pathological conditions including metabolic, ocular, and immunological issues (64, 65). Hence, these additional symptoms in the proband may be due to ER stress that may affect other proteins unrelated to PCP.

Our study also reveals that Tmem208 directly interacts with the PCP pathway protein Fz and helps maintain its levels. Based on previous studies, TMEM208 may target proteins with an internal transmembrane domain (11, 12). Fz harbors multiple internal transmembrane domains (66, 67). It also has an N’-terminal signal peptide (67), which should facilitate its transport via the SRP pathway. Hence, Fz may require both SRP and SND pathways, and loss of Tmem208 may impair the proper dosage of Fz by reducing its translocation to the ER. The mild change of Fz levels and the fact that only mild rotation but no chirality defects are found in Tmem208 mutant eyes suggest that Tmem208 does not directly affect core PCP functions, but rather may affect PCP effectors including Rho-GTPase family members and Rok or tricornered, fuzzy (61, 68–70). This is consistent with the notion that the albeit stronger wing hair orientation defects do not show the typical core PCP global effects, but rather local ones. Therefore, it is possible that the levels of some of these effector proteins are impaired due to loss of Tmem208, and the partial loss of Fz synergize to cause a PCP-like phenotype. Alternatively, the potential PCP-like phenotypes in the proband can also be linked to ciliary defects.

How does the loss of Tmem208 activate the ER stress response? TMEM208 helps in the import of proteins into the ER, and it also shares substrates with the two other ER protein import pathways, the SRP and GET pathways (11–13, 15, 71). Thus, loss of Tmem208 may affect the import of specific proteins into the ER and activate the Unfolded Protein Response (UPR). Indeed, we noted an elevation of Bip, which is up-regulated upon a UPR (30–32). It is also possible that Tmem208 not only facilitates the translocation of proteins into the ER lumen but also helps in chaperoning nascent proteins in the ER lumen. Finally, an abnormal ER membrane lipid composition, including altered chain length and/or saturation of the fatty acids (72–74), induces a mild ER stress via inositol-requiring enzyme 1 and PKR-like ER Kinase pathways (73, 75, 76), both of which are activated upon loss of Tmem208. Further studies are required to investigate whether loss of Tmem208 affects the composition of ER membrane lipids.

Materials and Methods

A detailed description of the materials and methods used in this study can be found in SI Appendix, Supplementary Materials and Methods and Table S1.

Fly Stock Maintenance and Longevity Assay.

Fly stocks were maintained in standard cornmeal food medium. The following stocks were used: UAS-mCherry.NLS w[1118]; Df(2R)Exel7138/CyO (BL#7883), w[1118]; Df(2R)BSC308/CyO (BL#23691), w[1118]; Dp(2;3)GV-CH321-39L15, PBac{y[+mDint2] w[+mC]=GV-CH321-39L15}VK00031 (BL#89921); dpp-GAL4:UAS-Fz; UAS-Tmem208-HA; UAS-Xbp1-EGFP; UAS-luci-RNAi; UAS-CG8320-RNAi; Actin-GAL4; Nub-GAL4. W¹¹¹⁸ fly strain was used as the Control. UAS-luci-RNAi was used as the Control RNAi. Tmem208^KG4 in all figures represent the genotype: Tmem208^KG4/Df(2R)Exel7138. The crosses were performed at 25 °C. RNAi stocks were procured from Vienna Drosophila Stock Center (please see SI Appendix, Supplementary Material for details) and the RNAi experiments were performed at 29 °C. Lifespan measurement was carried out at 25 °C as described previously (21). A group of 5 to 6 newly emerged flies were kept together and flipped into a fresh vial every 2 to 3 d until all of the flies were dead. At least 50 flies were tested in each group. The log-rank (Mantel–Cox) test was used to determine the statistical significance.

Generation of Tmem208^KozakGal4 and Tmem208-GFP Alleles.

Tmem208^KozakGal4/Tmem208^KG4 and Tmem208-GFP alleles were generated using CRISPR/Cas9 technology as described previously (18). For the Tmem208^KozakGal4 line, the coding sequence of Tmem208 was replaced with KozakGAL4-3XP3EGFP cassette using homologous recombination. In Brief, the gRNAs targeting 5′UTR (AGACGTCGCCACGTAAATAGTGG) and the end of the coding region (CGAGTTGGCCTGGTGGTTAAGGG) of Tmem208 were cloned in pCDF5 vector (Addgene Accession number #73914). A Tmem208-specific plasmid with a restriction cassette, which is flanked on either side by 200-nucleotide homology arms and a gRNA1 target sequence, was synthesized in a pUC57_Kan_gw_OK vector backbone (GENEWIZ, Azenta Life Sciences), and it was used as the homology donor intermediate vector. Subcloning was performed to replace the restriction cassette with an integration cassette that contains KozakGAL4-polyA-FRT-3XP3- EGFP-FRT. Subsequently, yw; iso; attP2(y+){nos-Cas9(v+)} embryos were injected with 250 ng/μL of the completed plasmid together with pCFD5 vector (100 ng/µL) encoding for the gene-specific sgRNAs. The resulting G0 progenies were crossed to yw flies, and 3XP3-EGFP-positive flies were selected to establish the stock.

For Tmem208-GFP allele, the same homology donor intermediate vector and sgRNAs to generate KozakGAL4 alleles were used to replace the coding region of Tmem208 with the codign region tagged with sfGFP. 3XGGS flexible linker was used in either side of the sfGFP to minimize disruption of the protein structure. Linker-sfGFP-linker is inserted between amino acids R99 and E100. The dominant marker Scarless-DsRed (a gift from O’Conner-Giles lab DGRC# 1364) is inserted at the 3′UTR. Fragments for gene coding region, linker-sfGFP-linker, and Scarless-DsRed were PCR amplified with primers containing overlaps, and the fragments are assembled using the NEB-HiFi DNA Assembly kit (New England Biolabs #E2621) using the manufacturer’s instructions, in the homology donor intermediate linearized by BsaI-HF (NEB #R3535). Homology donor vector 250 ng/μL was injected into y w; iso; attP2(y+){nos-Cas9(v+)} embryos, together with pCFD5 vector encoding for the gene-specific sgRNAs (100 ng/µL). The resulting G0 males and females were crossed to yw flies to screen for the presence of 3XP3-DsRed.

Statistical Analyses.

All the statistical analyses were performed using MS Excel and GraphPad Prism (GraphPad Software Inc., CA, USA). For comparing two groups, two-tailed unpaired t tests were carried out, and for comparing more than two groups, ANOVA with an appropriate post hoc test were carried out. The results were presented as bar plots where the error bars represent ± SEM. P-values of more than 0.05 were considered not significant, whereas P-values less than 0.05 were considered significant (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (DOCX)

Data, Materials, and Software Availability

Some study data are available. An MTA would be required to share human fibroblasts. All other data are included in the article and/or supporting information.