Identification of C. elegans ASNA-1 domains and tissue requirements that differentially influence platinum sensitivity and growth control

ASNA1 plays an essential role in cisplatin chemotherapy response, type 2 diabetes, and heart disease. It is also an important biomarker in the treatment response of many diseases. Biochemically, ASNA1 has two mutually exclusive redox-modulated roles: a tail-anchored protein (TAP) targeting function in the reduced state and a holdase/chaperone function in the oxidized state. Assigning biochemical roles of mammalian ASNA1 to biomedical functions is crucial for successful therapy development. Our previous work showed the relevance of the C. elegans ASNA-1 homolog in modeling cisplatin response and insulin secretion. Here we analyzed two-point mutants in highly conserved residues in C. elegans ASNA-1 and determined their importance in separating the cisplatin response function from its roles in insulin secretion. asna-1(ΔHis164) and asna-1(A63V) point mutants, which both preferentially exist in the oxidized state, displayed cisplatin sensitivity phenotype as well as TAP insertion defect but not an insulin secretion defect. Further, using targeted depletion we analyzed the tissue requirements of asna-1 for C. elegans growth and development. Somatic depletion of ASNA-1 as well as simultaneous depletion of ASNA-1 in neurons and intestines resulted in an L1 arrest. We concluded that, targeting single residues in ASNA-1 affecting Switch I/Switch II domain function, in comparison to complete knockdown counteracted cisplatin resistance without jeopardizing other important biological functions. Taken together, our study shows that effects on health caused by ASNA1 mutations can have different biochemical bases.

Introduction C. elegans ASNA-1 /S. cerevisiae GET3 /mammalian ASNA1 are essential cytosolic ATPases that are evolutionarily related to the bacterial arsenite transport factor ArsA [1], however the function in eukaryotes has clearly separated [2]. The mammalian homolog ASNA1 not only plays an essential role in tail-anchored (TA) protein insertion [3] and insulin secretion [4] but has been characterized as an important biomarker in treatment response in schizophrenia [5], predicting disease severity of dengue virus infection [6], differentiation between subjects with active and latent tuberculosis [7], a biomarker for abnormalities associated with ultra-highrisk for psychosis [8], as well as a biomarker for Down's syndrome [9]. Mutations in ASNA1 are associated with pediatric cardiomyopathy [10] and mutations in its pathway protein calcium modulating ligand (CAML) are linked in patients to hypotonia, brain abnormalities, and epilepsy [11]. Further ASNA1 interacts with and likely modulates the function of vesicle-associated protein B (VAPB) which is mutated in cases of amyotrophic lateral sclerosis [12]. Therefore, it is clear that ASNA1 has a role in a variety of human diseases and a deeper understanding of clinically relevant functions of ASNA1 is of general interest.
The diverse roles of ASNA1 in different disease manifestations are reflected in the different biochemical roles of this protein. The best understood canonical and most intensively studied role of ASNA-1/GET3/ASNA1 is in the endoplasmic reticulum (ER) targeting of a special class of membrane proteins called tail-anchored proteins (TAP) [13]. This targeting function strictly requires its ATPase activity [14] and the transmembrane regions of TAP proteins associate with the hydrophobic groove in dimeric ASNA1 before their insertion into the ER membrane. Insertion of TAPs via ASNA1/GET3 requires the ER membrane proteins WRB/CAML (in mammals) or Get1/Get2 (in yeast) [15] to which ASNA1/GET3 hands off the bound TAPs.
The ATPase activity of the ASNA-1/GET3 protein requires the function of the Switch I and Switch II domains. These two domains bind ATP and form so called 'loaded spring' that is released upon ATP hydrolysis [16,17]. The Switch II region collapses after ATP hydrolysis causing disruption in the network of conserved cross-monomer interactions that stabilize the composite groove [14]. In addition, the Switch II domain in yeast Get3 is essential for the transition from an open to a closed state and functions to assure structural changes that are induced by ATP or TAP binding [14]. This structural change ensures TAP binding to GET3 [18][19][20]. Structurally, residues in Switch I and Switch II domains of GET3 coordinate water and Mg 2+ molecules important for the ATPase activity [14]. The interaction between Switch I and Switch II domains is essential because an Asp57Asn mutant in the Switch I domain of GET3 is defective for ATP hydrolysis and unable to rescue the hygromycin-sensitive phenotype of Δget3 yeast [14].
However, it is clear that not all functions of ASNA1 and GET3 are a consequence of their canonical TAP targeting function. There are alternative roles for ASNA1/GET3 where it acts independently of its TAP insertion function and for these functions it thus does not require interaction with Get1/Get2 or WRB/CAML. Mammalian ASNA1 interacts with the tailanchored protein, VAPB, but only via the FFAT motif in the MSP domain and not its hydrophobic groove which is essential for its interaction with tail-anchored proteins. Hence mammalian ASNA1 can display novel interaction patterns even with a tail-anchored protein that does not involve the canonical ASNA1/WRB/CAML pathway. VAPB may also work as an alternative receptor for ASNA1 instead of WRB/CAML [12]. Importantly for our understanding of ASNA1/GET3 functions, elegant biochemical and cell biological studies show upon oxidation-induced structural rearrangements GET3 acts as a holdase chaperone that binds to aggregated proteins under conditions of oxidative stress and ATP depletion [21,22]. Oxidation of conserved cysteines in Get3 converts the dimeric GET3 to a tetramer. This structural change masks the hydrophobic domain required for TA protein binding and substantially reduces its ATPase activity: Both the TA protein binding and ATPase properties of GET3, which are essential for its role in TAP targeting and interaction with WRB/GET1, are completely dispensable for the chaperone function of oxidized Get3. These lines of evidence demonstrate that under conditions of oxidative stress, Get3 promotes functions that do not require WRB/ GET1. Moreover, Voth et al., 2014 show that oxidized GET3 also acts as a general chaperone that can refold denatured citrate synthase in vitro in the absence of WRB/GET1 [22]. Moreover, GET3 directly binds to non-tail anchored proteins such as the chloride transporter Gef1p [23] and to the Gα subunit Gpa1 to act as a guanine nucleotide exchange factor [24]. In both cases, the binding is biologically meaningful since it leads to altered biochemical outcomes that are independent of GET1.
Our previous work on C. elegans ASNA-1 showed that ASNA-1 promotes insulin secretion and led later to the findings that insulin secretion is defective in ASNA1 knockdown mice leading to type 2 diabetes [25,26]. Further, the role of ASNA1 in the cisplatin sensitivity of human tumor cells was also modeled successfully in worms, and importantly the human ASNA1 gene can replace the worm gene for both cisplatin and insulin secretion functions [25,27]. Our previous work in worms demonstrates the TAP targeting function of ASNA-1 and shows that cisplatin exposure negatively impacts only targeting of the ASNA-1-dependent tail-anchored protein SEC-61β to the ER while having no effect on ASNA-1 independent TAPs [28].
Like the yeast ortholog, worm ASNA-1 is also oxidized via the two conserved cysteines. We find that the relevant roles of ASNA-1 in insulin secretion and protection from cisplatin toxicity can be separated based on the oxidative state of the protein. As with GET3, ASNA-1 exists in two redox-sensitive states: reduced (ASNA-1 RED ) and oxidized (ASNA-1 OX ) [28]. The single point mutant asna-1(ΔHis164), in which the protein preferentially exists in the oxidized state, is not only as sensitive to cisplatin treatment as a complete deletion mutant but also has a severe defect in TAP insertion [27,28] showing that separation of function mutations exists and indicating that ASNA-1 function in insulin secretion is most likely independent of its role in TAP targeting and cisplatin response [27,28].
Here, we examined the expression of ASNA-1 and performed selective depletion to identify the tissue requirements for growth promotion. We also expanded our analysis of the asna-1 (ΔHis164) worms that have a lesion in the Switch II domain. Here we showed that this mutant, which has high levels of oxidized ASNA-1, has normal levels of insulin secretion and function. Further, we tested a set of viable and homozygous C. elegans asna-1 point mutants to explore the possibility of completely separating clinically relevant functions of asna-1 with genetic variants. As a result, we characterized the highly conserved alanine in position 63 adjacent to Switch I as a relevant target for function separation. asna-1(A63V) mutants were sensitive to cisplatin but did not display any morphology, growth, lifespan, brood size, or oxidative stress phenotypes that were shown by asna-1 deletion mutants. All these features made asna-1 (A63V) animals an excellent model to study the role of asna-1 in the response to cisplatin without the compounding effects of other phenotypic defects. Indeed, asna-1(A63V) mutants showed cisplatin sensitivity defect and TAP insertion defect that was as severe as that seen in deletion mutants without any detectable defect in insulin secretion or signaling pathways. Moreover, this mutation led to higher levels of the oxidized form of the protein and altered its subcellular distribution. Lastly, we performed in silico modeling to better understand the consequences of the structural changes caused by the single-point mutations. For a multifunctional protein with non-overlapping functions, drugs affecting one function while not perturbing the other functions would reduce the undesirable side effects and might increase the chemotherapeutic response.

ASNA-1 is broadly expressed in C. elegans
Expression from a transgene expressing ASNA-1::GFP (svIs56) was seen in sensory neurons, the intestine, and the hypodermis [25]. Taking advantage of the CRISPR/Cas9 technique we obtained a strain where two tags, mNeonGreen (mNG) and the auxin-inducible degron (AID), were inserted into the genomic locus of ASNA-1 at the C-terminus of the protein (ASNA-1::mNG::AID; syb2249; strain PHX2249). The mNeonGreen tag allowed us to more precisely characterize the ASNA-1 expression pattern in the context of all upstream and downstream elements. The mNeonGreen signal was visible strongly in the pharynx, intestine, muscles, proximal and distal germline cells, oocytes, developing vulva, vulval muscles, spermatheca, sperm, sheath cells, the most proximal pair of gonadal sheath cells (Fig 1), and in head neurons (S1 Fig). This analysis not only confirmed our previous observations [25] but also characterized previously unknown ASNA-1 expression sites in the C. elegans reproductive system. We concluded that the germline and somatic gonad are likely important places for ASNA-1 function given the broad expression of ASNA-1 in these tissues. To more precisely identify the cells expressing ASNA-1 in each tissue we next examined a bi-cistronic ASNA-1^SL2::mNeonGreen::H2B allele syb5730 of the gene (strain PHX5730). From this allele, the ASNA-1 mRNA independently makes untagged ASNA-1 and nuclear-localized mNeonGreen to report on each nucleus that expresses the asna-1 mRNA. Analysis of the mNeonGreen signal in asna-1(syb5730) animals showed that virtually every cell in the body expresses the ASNA-1 mRNA throughout development in both the soma and the germline (S2 Fig).

Soma-specific asna-1 knockdown leads to L1 larval arrest
Worms lacking maternal and zygotic asna-1 arrest at the 1 st larval (L1) stage even in the presence of food [25]. L1 arrest is a characteristic of wild-type larvae hatched in the absence of food or of strong daf-2/insulin receptor mutants [29]. Taking into consideration how broadly ASNA-1 is expressed in somatic and germline tissues, we asked whether somatic depletion of ASNA-1::mNG::AID would cause a larval arrest phenotype similar to that seen after complete asna-1 genetic depletion. Arabidopsis TIR1 is an F-box protein that interacts with the AID tag in the presence of auxin to assemble an E3 ubiquitin ligase complex and targets AID-containing proteins for rapid degradation by the proteasome [30,31]. Tissue or cell-type-specific expression of TIR1 limits the degradation of AID tagged proteins to TIR1-expressing cells. Taking the advantage of the auxin-inducible tissue-specific knockdown system, we created the strain expressing ASNA-1::mNG::AID (syb2249) and TIR1 pan-somatic driver ieSi57. 4th larval stage (L4) hermaphrodites were exposed to 1mM auxin (AUX) for 48h while they produced progeny and then removed from the plate. Their progeny, which remained on the auxin-containing plates, were analyzed 48 hours after the removal of the mothers (S3A Fig). Similarly handled L4 worms on non-auxin NGM plates served as a control. First, we observed that asna-1::mNG::AID;ieSi57 mothers did not display any morphological defects after 48h exposure to auxin but did produce fewer progeny in comparison to non-treated control animals. On auxin plates all their progeny arrested at the L1 stage (Fig 2A). The auxin-mediated depletion of ASNA-1 was effective because the mNeonGreen signal was not visible in any of the previously characterized somatic tissues (Figs 1 and 2A). However, we saw a strong mNeonGreen signal in the Z2 and Z3 germline progenitor blast cells (Fig 2A), which give rise to the entire germline during development [32]. None of the arrested larvae, even 72 hours after removal of the mothers, showed more than the two cells (Z2 and Z3) in the gonad primordium. The cells could be identified as Z2 and Z3 because they lay in the center of the gonad primordium, were in contact with each other [32], and expressed ASNA-1::mNG::AID while it was depleted in all surrounding somatic cells. To further characterize the effects of ASNA-1 depletion, we sought to establish if asna-1::mNG::AID;ieSi57 worms exposed to auxin only from the 1st larval stage onwards would display further development and possibly reach later larval stages. Staged L1 hermaphrodites were exposed to 1mM auxin (AUX) for 24h, 48h, and 72h and analyzed at these time points. (S3B Fig). Similarly staged unexposed asna-1::mNG::AID;ieSi57 larvae served as controls. We observed that all animals on AUX plates were still arrested at the L1 stage (Figs 2B and S4), whereas control animals reached adulthood. As expected, auxintreated animals were characterized by a lack of ASNA-1::mNG::AID expression in the somatic tissues and we observed a strong ASNA-1::mNG::AID signal in the germline progenitors ( Fig  3C). However, in contrast to worms hatched on AUX-containing plates, ASNA-1::mNG::AID was seen in many more germline cells with a maximum of 14 positive cells (Fig 3C). In total, 13/20 worms analyzed had more than 4 germline cells. This was in contrast to the mNeon-Green signal only being observed in the two primordial germ cells Z2 and Z3 in animals hatched on AUX plates (Fig 2A). However, the larvae in the second experiment did not have the alae that are characteristic of L2 animals indicating that depleting somatic ASNA-1::mNG:: AID after the birth of L1 larvae allowed for further development of the germline but did not allow the larvae to reach the L2 larval stage. The size of the gonad was also characteristic of L1 stage larvae. By contrast, auxin-mediated depletion of ASNA-1 from the L4 stage onwards did not block progression to the adult stage.

Intestine-specific asna-1 knockdown leads to developmental delay
Our previous research has established that asna-1 is required cell non-autonomously for growth. Both neuron and intestine-specific promotors driving asna-1 expression rescue the asna-1(ok938) body size phenotype [25], while ASNA-1 is required in the intestine but not the neurons for the rescue of the cisplatin sensitivity phenotype [27]. There is also a requirement for ASNA-1 in the gut for proper tail-anchored protein targeting [28]. Therefore, we sought to establish a role of the intestinal function of ASNA-1 in growth and development to ask whether the L1 arrest phenotype could be ascribed to the TAP targeting function in the intestine. For that purpose, we constructed a strain expressing an intestine-specific TIR1 driver

PLOS GENETICS
C. elegans ASNA-1 function in insulin secretion and cisplatin ieSi61 in the ASNA-1::mNG::AID (syb2249) background. To compare pan-somatic depletion of ASNA-1 with gut-specific depletion, as before we exposed asna-1::mNG::AID;ieSi61 L4 hermaphrodites to 1mM auxin (AUX) for 48h, removed them from the plate and analyzed their progeny 48h after the removal of the mothers (S3A Fig). Worms on NGM plates not containing auxin served as a control. Microscopy revealed an absence of mNeonGreen fluorescence in the intestine while pharyngeal and germline expression was visible at normal levels (Fig 3A). Intestine-specific ASNA-1 depletion starting in developing embryos led to a delay in development in comparison to control animals (Fig 3A). The larvae hatched on these plates took an additional day to reach adulthood. When L1 larvae were exposed to 1mM AUX (S3B Fig), no delay in growth was seen (Fig 3B). Thus, there was a requirement for ASNA-1 in the intestine for growth but the requirement was not as stringent as that seen with pan-somatic depletion. Based on that we concluded that growth and development dependent on ASNA-1 was a multitissue requirement or that auxin-mediated depletion was not complete.

Simultaneous depletion of ASNA-1 in the intestine and neurons is equivalent to pan-somatic depletion
Having shown that intestinal depletion of ASNA-1 was not sufficient to produce the L1 growth arrest phenotype, we wished to determine whether we could narrow down the tissue requirements that lead to a strong L1 arrest phenotype seen in asna-1 mutants [25] upon pan-somatic depletion. To do this we first depleted by auxin treatment ASNA-1::mNG::AID in neurons using the pan-neuronal driver reSi7 [31]. asna-1::mNG::AID (syb2249); reSi7 worms were exposed to auxin from the time of hatching by putting 4th larval stage mothers on auxin-supplemented NGM plates. The worms produced a normal number of progeny and the progeny displayed no growth defects or developmental delays. To ask whether the 1st larval stage arrest was caused by simultaneously depleting ASNA-1 in both the intestine and neurons, we next made and tested a strain of the genotype asna-1::mNG::AID (syb2249);reSi7;ieSi61. In this strain, the two TIR1 drivers promoted the simultaneous depletion of ASNA-1 in both tissues. In these worms, we observed that when 4th larval stage hermaphrodites were placed on auxinsupplemented NGM plates (S3A Fig), they still produced a large number of progeny as adults but all the progeny (n>250) arrested as 1st larval stage sized animals. This arrest reproduced the effect of depleting ASNA-1 in all somatic tissues (Fig 3C) and indicated that both a neuropeptide secretion defect (such as DAF-28 secretion) and a defect in the trafficking of SEC-61β alone or along with other TAPs is required for the control of growth at the 1st larval stage. To obtain support for this notion, we next examined an asna-1 mutant with a known defect in TAP insertion and determined its insulin/IGF signaling status.

Insulin signaling is normal and germline development defects are less severe in asna-1(ΔHis164) compared to asna-1 deletion mutants
Our previous work shows that asna-1(ΔHis164) mutants are as sensitive to cisplatin as deletion mutants in the gene and have a severe TAP insertion defect [28]. Further, we find that the ASNA-1 ΔHis164 ::GFP protein largely exists in the oxidized state [28]. Transgene expressing asna-1 ΔHis164 ::GFP rescues the asna-1 growth phenotype of the deletion mutant indicating indirectly that the point mutant protein can restore insulin signaling [27]. To examine the growth phenotypes of asna-1(ΔHis164) mutants, we directedly measured the insulin secretion and signaling efficiency by analysis of DAF-16::GFP localization and DAF-28/insulin::GFP secretion. The cytoplasmic localization of DAF-16::GFP indicates high IIS levels since DAF-16::GFP is localized to nuclei in mutants with insulin signaling defects [33]. The uptake of secreted DAF-28/insulin::GFP by coelomocytes, which are found in the open circulatory system, reports on insulin secretion [25]. In asna-1(ΔHis164) animals DAF-16::GFP protein was cytoplasmic and the DAF-28/insulin::GFP protein was secreted and taken up by coelomocytes at normal levels (Figs 4A, 4B, S5A, and S5B). The analysis of both reporters showed that insulin secretion and signaling were unaffected in asna-1(ΔHis164) mutants in comparison to the strong defects in asna-1 deletion mutants. Although both mutants did differ in IIS status, both the asna-1(ΔHis164) and asna-1 deletion mutants were sterile. The bigger size of the asna-1 (ΔHis164) worms (Fig 4C) suggested that germline defects in the two mutants differ significantly. We sought to understand these differences by examining the germline phenotype of asna-1(ok938) and asna-1(ΔHis164) animals. For this purpose, we used the transgenes ltIs37 and ltIs38 to simultaneously monitor germ cell membranes and germline nuclei in vivo. asna-1 (ok938) animals did not produce any oocytes or sperm and showed a gonad migration defect (Fig 4C). asna-1(ΔHis164) also showed a gonad migration defect but the germlines contained oocytes and sperm as well as non-fertilized embryo-type structures in the uterus. These cells contained only one nucleus and displayed no embryonic divisions (S6A Fig). Ovulation, the transit of the ovum through the spermatheca, and entry of oocytes into the uterus require the presence of sperm [34]. This led us to investigate whether asna-1(ΔHis164) mutants might be defective in fertilization, taking into consideration the strong ASNA-1 expression in mature oocytes, sperm, and the spermatheca in wild-type worms (Figs 1 and S2). Indeed, the ltIs37 marker expressing pie-1p::mCherry::his-58 to visualize germline nuclei showed that sperm was present in the spermatheca in asna-1(ΔHis164) animals as in asna-1(+) (S6B Fig). However, since the oocytes that passed through the spermatheca did not contain more than one nucleus or show any embryonic divisions, we concluded that there was a defect either in the oocytes or sperm of asna-1(ΔHis164) animals. The difference in germline defect between point mutants and deletion mutants stands in striking contrast to the similarity in cisplatin sensitivity phenotype of these two mutants [28] indicating that different functions can be independently targeted.

Analysis of ASNA-1 point mutants reveals the importance of a conserved alanine for cisplatin resistance
We have previously shown that C. elegans ASNA-1 exists in two redox states: reduced (ASNA-1 RED ) and oxidized (ASNA-1 OX ), and that important biological functions of the protein are separable based on the redox state [28]. As mentioned, these conclusions were drawn based on an analysis of the asna-1 single point mutant: asna-1(ΔHis164), which also showed the sterility phenotype described above. Therefore, we wondered if this was a unique characteristic of mutating His164 or whether any other single point mutation in the ASNA-1 protein would also be able to completely separate the important biological functions of ASNA-1: insulin secretion, TAP targeting, and cisplatin sensitivity while not causing sterility in the worm. For this reason, we examined a set of seven strains generated by the Million Mutation Project, each carrying a missense point mutation in ASNA-1 along with other mutations in the genome [35] (S1 Table). Of these seven, only in four mutants were the mutated amino acids conserved with human ASNA1 (S1 Table and S7A Fig). None of the single amino acid changes caused reduced steady-state levels of the ASNA-1 protein (S8 Fig). We performed a cisplatin survival analysis of all seven strains to ask if any of those mutants could increase sensitivity to cisplatin treatment. The LD 50 of cisplatin for asna-1(ok938) deletion mutants is 300 μg/mL after 24h treatment [27] and hence we used this concentration for our experiments. We found that only two strains, asna-1(gk687101) and asna-1(gk592672), showed increased cisplatin sensitivity after 24h exposure to the drug (Fig 5A). Since asna-1(gk592672) strain produced a more severe sensitivity phenotype, we focused on this point mutation in our further analysis. The asna-1 (gk592672) mutation is an alanine to valine change, at a highly conserved position 63 which corresponds to alanine 82 in the human homolog and is positioned close to the Switch I domain (S7A Fig). This domain plays an essential role in the activation of ATP hydrolysis during TA protein targeting. We used the GnomAD browser (https://gnomad.broadinstitute.org/ ) to assess the effect of this point mutation in the human homolog of ASNA-1. The browser provides not only a list of variations but also a critical view of the clinical application [36]. In the exome sequence data for 60,706 individuals, ASNA1 has 79 observed single nucleotide variants (SNVs) that resulted in missense mutations. In those, A82 was detected to be mutated to threonine (T) (A82T) with two allele counts in the European (non-Finnish) population. Poy-Phen-2 (http://genetics.bwh.harvard.edu/pph2/) is a tool that predicts the possible impact of amino acid substitution on the structure and function of human proteins. The human ASNA1 (A82T) mutation was classified as 'probably damaging'. This was also true for the alanine to valine substitution. Mindful of other mutations in other genes in the asna-1(gk592672) containing strain we outcrossed the strain 13 times and performed the cisplatin sensitivity assay again. Indeed, the outcrossed strain, hereafter called asna-1(A63V), was as sensitive to cisplatin exposure as the asna-1(ok938) deletion mutant. The cisplatin sensitivity phenotype was rescued by a single copy of wild-type ASNA-1 fused to GFP (knuSi184) expressed from a transgene (Fig 5B). This transgene also rescued the ASNA-1 protein null mutant for the cisplatin sensitivity phenotype (Fig 5B). We concluded that change in a highly conserved alanine at position 63 was the genetic locus causing cisplatin sensitivity as severe as that seen in the deletion mutant and no other mutation in the genetic background contributed to the phenotype.

asna-1(A63V) mutants have a normal brood size and insulin/IIS signaling and secretion
To obtain more evidence for the separable functions of ASNA-1, we characterized the asna-1 (A63V) strain. As in the case of the asna-1(ΔHis164) protein [28], the A63V mutation did not affect the steady-state levels of the protein (Figs 5C and S9). ASNA-1 A63V ::GFP expressed from a multi-copy transgene was detected in the ER (S7B Fig) in the same manner as the wildtype protein ASNA-1::GFP. asna-1(A63V) animals did not display the enhanced ER stress phenotypes (S7C and S7D Fig) shown by the asna-1(ok938) deletion mutant [28]. While asna-1 (A63V) animals were as sensitive to cisplatin as the deletion mutant (Fig 5B), they were fertile with a normal brood size and life span (Fig 5D and 5E), and had properly developed germlines (Fig 5F). This characteristic opened up the possibility of separating the cisplatin, TAP insertion, and insulin functions of ASNA-1 without compromising on the sterility phenotypes seen in asna-1(ok938) or asna-1(ΔHis164) mutants [28,37] Importantly, asna-1(A63V) mutants had neither the insulin signaling nor the insulin secretion defect (Figs 5, 5H, S5A, and S5B), which was measured respectively by cytosolic/nuclear localization of DAF-16::GFP and uptake of DAF-28::GFP by the coelomocytes. This evidence led to the conclusion that indeed functions of ASNA-1 in cisplatin sensitivity and insulin signaling/secretion were completely separable in animals without the compromised function in the germline and further that the cisplatin sensitivity of asna-1 mutants was purely a soma-specific defect since the germline is normal in asna-1(A63V) mutants.

asna-1(A63V) mutants separate the TAP insertion function from insulin secretion
We have previously shown that asna-1(ok938) and asna-1(ΔHis164) mutants displayed a strong defect in the targeting of the TAP SEC-61β [28]. This conclusion was supported by two independent tests. First, Pearson colocalization analysis following confocal microscopy of worms co-expressing GFP::SEC-61β with the ER marker, mCherry::SP12. Second, N-linked glycosylation analysis of SEC-61β to monitor the insertion of SEC-61β into the ER membrane [28]. Since asna-1(ok938) and asna-1(ΔHis164) mutants share the cisplatin sensitivity phenotype of asna-1(A63V) mutants, we therefore wondered if asna-1(A63V) animals could support ASNA-1-dependent SEC-61β insertion. Pearson correlation analysis of confocal microscope images of the two co-expressed transgenes (GFP::SEC-61β and mCherry::SP12) in asna-1 (A63V) mutants revealed a significantly defective TAP targeting phenotype (Fig 6A and 6B). We next measured the direct insertion of the SEC-61β protein into the ER membrane [28]. We used a transgene (rawEx64) expressing a tagged SEC-61β in which the beta-opsin tag is inserted just before the stop codon. The opsin tag contains a reactive arginine which is glycosylated by ER luminal enzymes when the C-terminus is exposed to the ER lumen. This reports on the proper insertion of the SEC-61β protein into the ER membrane. The glycosylation event is revealed by the change in electrophoretic mobility of the glycosylated tagged protein. This analysis revealed that ER insertion of SEC-61β was decreased in the asna-1(A63V) mutant background (S10 Fig) consistent with the conclusions drawn from confocal microscopy analysis.

The A63V mutation alters the subcellular localization of ASNA-1
ASNA-1 physically interacts with the ER membrane resident receptor WRB-1 [28]. We next investigated whether the defect in GFP::SEC-61β targeting might be also associated with the reduced association of the ASNA-1(A63V) protein with the membrane. Separation of cytoplasmic from membrane fractions by ultracentrifugation revealed significantly reduced membrane association of the ASNA-1 A63V ::GFP protein compared to that seen in ASNA-1 + ::GFP (Figs 6C and S11). We concluded that asna-1(A63V) mutants shared the TAP insertion defect seen in the asna-1 deletion mutants likely via a reduced ability to associate with ER membranes. This defect was manifested without affecting the insulin signaling and indeed those two vital functions of ASNA-1 were likely independent of each other.

The ASNA-1(A63V) mutant protein preferentially exists in the oxidized state
We have previously shown that C. elegans ASNA-1 is a redox-modulated protein and its functions are separable based on the redox state of the protein [28]. We obtained evidence for this functional separation based on the analysis of asna-1(ΔHis164) point mutants in which the protein preferentially exists in an oxidized state. This mutant was defective for TAP insertion and cisplatin resistance [28] and as shown in Fig 4, ΔHis164 mutants were normal for insulin secretion/signaling function. We therefore wondered, based on the analysis of asna-1(A63V) mutants, if in asna-1(A63V) animals the separation of ASNA-1 functions is reflected in changes in the oxidative state of the protein. To address this, we performed non-reducing SDS-PAGE to evaluate the oxidative state of the ASNA-1 A63V ::GFP protein. Indeed, the analysis revealed that more of the ASNA-1 A63V protein exists in the oxidized state (Figs 6D and  S11). Strikingly, the shift in ASNA-1 A63V ::GFP to the oxidized state occurred in the absence of an increased expression of markers of ER stress (hsp-4) (S7C and S7D Fig), or mitochondrial stress markers (hsp-6 and hsp-60) (S7E Fig), whereas oxidative stress markers (gst-4, gst-30, and gst-38) were slightly increased (S7F Fig). Cisplatin treatment was not able to increase even further the amount of oxidized ASNA-1 A63V ::GFP (Fig 6E). We concluded that the A63V mutants in ASNA-1 had inherently high ASNA-1 OX levels, which resulted in sensitivity to cisplatin. They also had a TAP insertion defect but maintained not only normal insulin secretion function but also did not have a compromised germline phenotype. This conclusion is consistent with our previous analysis which showed that lower levels of reduced ASNA-1 had no impact on growth and DAF-28 secretion.

in silico modeling of ASNA-1(A63V) and ASNA-1(ΔHis164) proteins predicts the importance of Switch I and Switch II domains in function separation
The model of the wild-type ASNA-1 protein assembled as a dimer shows that the dimerization interface presents three main interaction regions: the zinc atom coordination, the ATP binding site, and the Histidine164 contacts (S12 Fig). The model obtained by the single point mutation A63V showed that the substitution of alanine by valine reduced the distance between the helix α3 and the helix α2 by the addition of the isopropyl group at the position 63 (helix α3) pointing directly towards the Thr32 (helix α2) in the P-loop motif (Fig 7A). The model proposed a movement correction of the threonine hydroxyl group induced by this close contact with the Valine. The Thr32 plays a key role in the coordination of the Mg atom (Figs 7A and S12), its hydration state, and thereby the binding and catalysis of the ATP. The binding of ATP is a key event in the formation of the closed dimer and the formation of the closed interface by its direct participation in it. We conclude that A63V change might influence the P-loop structure and affect ATP binding.
The wild-type model showed the close interaction between the His164 and His164' of each monomer to perform the dimer interface (Fig 7B). The model of ASNA-1 presenting the deletion of the His164 showed that the loss of histidine implies a loss of one of the key amino acids that create the dimeric interface and does not suggest the formation of compensatory Representative western blot after reducing and non-reducing SDS-PAGE to detect levels of oxidized and reduced ASNA-1 A63V ::GFP. Control worms expressed the ASNA-1(+)::GFP transgene. Blots were probed with anti:GFP antibody and tubulin was the loading control. For full uncropped blot source images including replicates used for quantification, see S11 Fig. Graph represents the band intensity quantification of oxidized/reduced ASNA-1 A63V ::GFP. Statistical significance was determined by the independent two-sample t-test. Experiments were performed in triplicate. Bars represent ± SD. (E) Representative western blot after reducing and non-reducing SDS-PAGE to detect levels of oxidized and reduced ASNA-1 A63V ::GFP upon cisplatin (+CP) treatment (300 μg/mL for 6h). Control worms expressed the ASNA-1 A63V ::GFP transgene. Blots were probed with anti:GFP antibody and tubulin was the loading control. For full uncropped blot source images including replicates used for quantification, see S13 Fig

PLOS GENETICS
interactions. (Fig 7B). Therefore, the deletion of this amino acid would affect the stability of the closed conformation and probably its formation in a negative fashion.

Discussion
An intensively studied role of ASNA-1 homologs is in the targeting of a subset of tail-anchored proteins to the ER membrane and beyond. However, it has also been clear that ASNA-1 and its homologs have roles that are not connected to its tail-anchored targeting function. These include modulation of chloride channel function, a holdase function for aggregated proteins, and the binding to the ALS-related protein VAPB protein among others [12,21,38]. Earlier work from our lab has also addressed this issue and separated insulin secretion from cisplatin response [28]. Since ASNA-1 promotes the ER targeting of SEC-61β in the intestine [28] and is also required for the neuronal secretion of DAF-28/insulin [25], one objective of this study was to examine further the relative contributions of different tissues for growth and reproduction. A second major goal was to discover whether single point mutations in C. elegans ASNA-1 will fulfill the requirements for ASNA-1 function for cisplatin response separate from its roles in growth and reproduction.
First, we characterized the expression pattern of ASNA-1 in somatic tissues as well as in the germline. We showed that general somatic tissue depletion led to the strict L1 arrest phenotype identical to that seen in asna-1 mutants depleted from both maternal and zygotic contributions. We also found that intestine-specific depletion of ASNA-1 led to a growth defect but that it was not as severe as pan-somatic depletion, raising the possibility that the contribution of ASNA-1 to the L1 arrest likely came from both neurons and the intestine. This was further confirmed by the simultaneous depletion of ASNA-1 in the intestine and neurons. Pansomatic depletion after embryogenesis showed that asna-1 was required for the development of worms from the L1 to the L2 stage but not from the L4 larval stage to adulthood. This indicated that ASNA-1 cannot be regarded as a housekeeping gene whose function is needed in a constitutive manner.
The million mutation project (MMP) is a powerful resource for exploring mutations in almost every C. elegans gene [35]. It led to a collection of a large number of mutations in over 2000 whole genome sequenced strains. We tested seven single-point mutants in ASNA-1 generated by this project for cisplatin sensitivity (S1 Table). Since all these seven strains were homozygous for the point mutants and were fertile, it was very likely that none of the seven point mutants affected reproduction. Given that the mutants were fertile we expected only mild changes in the overall effect on the function of the mutant ASNA-1. However, we cannot rule out the possibility that a compensatory second site mutant might act in some cases. Only two of the seven strains were cisplatin sensitive. We concluded that heavy mutation load per se, which is present in all the strains will not result in cisplatin sensitivity, and worms that were unhealthy were not inherently prone to cisplatin sensitivity due to some non-specific reason. In the screen of viable ASNA-1 mutants, which would be able to separate the clinically relevant functions of ASNA-1 in insulin secretion and cisplatin resistance, we identified asna-1(A63V) as important in function separation.
The asna-1(A63V) mutant produces a protein that preferentially exists in the oxidized state, has a strong cisplatin sensitivity phenotype and tail-anchored protein insertion defect, and has a defective subcellular localization. Importantly, despite these defects, the mutation did not affect IIS activity nor caused any growth or developmental problems. This allowed us to genetically separate asna-1 functions even further without compromising the viability of the worms. Evidence for functional separation emerged from the earlier study of another point mutant: asna-1(ΔHis164). We had previously shown that this mutant too is cisplatin sensitive and the mutant protein preferentially accumulates in the oxidized state [28]. Here we present evidence that the IIS functions of ASNA-1 remain largely intact in the ΔHis164 mutant and the germline development is substantially better than that seen in the asna-1 deletion mutant since oocyte and sperm are formed and ovulation (the exit of oocytes from the ovary) also occurs. Taken together our work shows that a marked shift towards the accumulation of higher levels of oxidized ASNA-1 does not perturb insulin secretion and growth but does have an impact on TAP targeting and the cisplatin sensitivity phenotype (Table 1 and S14 Fig).
In silico modeling showed that the A63V mutation, which is close to the Switch I motif, affects the coordination of a conserved Threonine in the P-loop motif with possible effects on ATP binding and hydrolysis. Further, another function separating mutation in highly conserved Histidine 164 lies in the Switch II motif and had no effect on insulin secretion and signaling. This indicates that Switch I and Switch II and P-loop motifs are excellent druggable targets that increase cisplatin sensitivity while sparing the insulin secretion role of ASNA-1. Indeed, ATPases have been an attractive target for drug discovery for many years considering that their dysregulation is at the origin of many human diseases. Already several ATPase inhibitors have been described [39][40][41][42] and have been shown to be beneficial in the treatment of various diseases. One of the possibilities to target the Switch I/II domains in ASNA1 in order to separate the ASNA1 functions might be the use of small molecule, competitive ATP inhibitors. In particular, it would be beneficial for the treatment of cancer types that are resistant to chemotherapy. One of the alternatives would be a synthesis of a specific ATP inhibitor produg, specific for Switch I and Switch II domains present in ASNA1, which would also contain a small peptide carrier. The activation of the drug would require the cleavage of the peptide by specific proteases which would allow for the design of tissue-specific ATP inhibitor. Full validation of this strategy is still needed, but it seems clear that targeting the domains important for ATPase activity in ASNA1 is an appealing approach.
To be able to develop targeted cancer therapies, there is a need to have a better understanding of which cellular pathways are involved in the cisplatin response and drug resistance. Based on the results presented here we believe that separating ASNA1 functions in cisplatin response in insulin secretion by targeting the ATPase domain of the protein by small molecule drugs that act as competitive ATP inhibitors will allow us to re-sensitize tumors to the cisplatin treatment leaving insulin secretion undisturbed.

C. elegans strains maintenance and synchronization
The Bristol strain (N2) was the wild-type. The asna-1(gk592672) containing strain VC40357 was outcrossed 13 times using unc-32(e189) and oxTi719 as balancers and kept in trans to the Table 1. Summary of phenotypes of asna-1 deletion and point mutants discussed in the paper to highlight the different phenotypes displayed with differing levels of ASNA-1 activity. Cisplatin sensitivity is independent of germline development ability and insulin secretion ability but linked to the TAP targeting phenotype. Animals with a large increase in the ratio of oxidized v/s reduced ASNA-1 will still display the cisplatin sensitivity phenotype indicating that increased ASNA-1 levels do not protect these mutants from cisplatin induced death. hT2(qIs48) balancer. ASNA-1 A63V ::GFP was generated by mutagenesis of pVB222GK [25] to introduce the A63V change using the QuickChange II Kit (Agilent Technologies). The resulting plasmid (pGK200) was used to generate the extrachromosomal array rawEx8. Genomic integration of rawEx8 yielded rawIs16. rawIs16 worms were outcrossed 4 times before analysis. The single-copy asna-1::gfp (knuSi184) contained 1.4 kb upstream promoter sequence driving the genomic asna-1 coding region fused to GFP just before the stop codon followed by the tbb-2 3' UTR. This construct was inserted on chromosome II at the ttTi5605 locus using MosSCI technology [43] by Knudra Transgenics. rawEx64 expressed vha-6p-3xFlag::SEC-61β::opsin as an extrachromosomal array and was outcrossed 2 times before use [28]. All C. elegans strains used in this study are listed in S2 Table. Worms were cultured under standard conditions at 20˚C on nematode growth media (NGM) plates unless stated otherwise and the E. coli strain OP50 was used as a food source. Synchronization: Synchronous larval populations were obtained by gravity separation as described previously [44].

TAP-targeting analysis
The assay was performed as described previously [28]. In short, live 1-day-old adult animals were sedated in 1 mM Levamisole/M9 and mounted onto 2% agarose pads. The int8 and int9 cells in the posterior intestine were imaged. The fluorescence signals were analyzed at 488 nm and 555 nm by the LSM700 Confocal Laser Scanning Microscope (Carl Zeiss) with LD C-Apochromat 40x/1.1 W Corr. objective. Image processing of Z-stacks was performed with the ZEN Lite program (Zeiss). Correlation quantification was done using Volocity software (PerkinElmer). Correlation quantification was done using the Automatic Thresholding [45] method to set thresholds objectively.

Glycosylation analysis of tagged SEC61β
Young adult worms expressing 3xFlag::SEC-61β::opsin from the rawEx64 transgene were homogenized and taken for protein concentration determination. All samples were solubilized in Laemmli buffer, separated by SDS-PAGE followed by western transfer and detection by immunoblotting with anti-Flag (F1804, Sigma) antibody.

Western blot analysis
Reducing SDS-PAGE: Synchronized young adult worms were homogenized and protein concentration was determined using the BCA assay (Thermo Scientific). Samples were boiled for 10 min in reducing loading buffer (SDS/ β-mercaptoethanol). Proteins were separated by SDS-PAGE and blotted onto PVDF membranes. Non-reducing SDS-PAGE: Lysates were boiled for 10 min in non-reducing (without β-mercaptoethanol) loading buffer and cooled to room temperature for 10 min. To protect free cysteine thiols from post-lysis oxidation, iodoacetamide was added to the samples at a final concentration of 25 mM followed by a 30 min incubation in darkness at room temperature. Proteins were separated by SDS-PAGE and blotted onto nitrocellulose membranes. Antibodies: anti-ASNA-1 antibody [25], anti-GFP antibody (3H9, Chromotek), and anti-Flag M2 antibody (F1804, Sigma) were used. To assess equal loading, membranes were stripped and probed with anti-alpha tubulin (T5168, Sigma). Band quantification was performed using ImageJ software [46].

Subcellular fractionation
Young adult animals grown at 20˚C were harvested, washed, and lysed in extraction buffer (50 mM Tris, pH7.2, 250 mM sucrose, 2 mM EDTA). Supernatants were centrifuged for 60 min at 100,000xg at 4˚C. The supernatant fraction was concentrated using Vivaspin Concentrators (Sigma). The pellet fraction was resuspended in 1x Laemmli buffer (Biorad). Proteins in both fractions were separated by SDS-PAGE and blotted onto PVDF membranes. Proteins were detected using an anti-GFP antibody (3H9, Chromotek) or anti-alpha tubulin (T5168, Sigma).

Auxin treatment
Animals were transferred to NGM plates containing indicated concentration of the water-soluble auxin derivative naphthaleneacetic acid (N610; Phyto-Tech Labs). The auxin-containing agar plates were prepared on the day of use from a freshly made 800mM stock solution in water and stored in the dark.

Cisplatin sensitivity assay
Cisplatin plates were prepared as described previously [28]. L4 larvae were isolated and grown for 24h before exposure to cisplatin. After 24h cisplatin exposure, death was determined by the absence of touch-provoked movement when stimulated by harsh touch using a platinum wire.

ASNA-1 model prediction
Three models for the C. elegans ASNA-1 (CELE_Zk637.5 #1-342) were obtained using the 3D modeling prediction program MODELLER [47], having as a template the structure of the Get3 protein complex of Chaeromium thermophilum binding AMPPNP a non-hydrolyzable ATP agonist (PDB:3IQW) [48]. Two of the models were obtained by loading the sequence with one of the following mutations: A63V and ΔH164. The third structure does not include any mutation and mimics the wild-type protein. The model was assembled as a dimer and in complex with AMPPNP, Mg and Zn mimicking the structure obtained for Chaeromium thermophilum (PDB:3IQW). (A) 4th larval stage (L4) hermaphrodites were exposed to 1mM auxin (AUX) for 48h while they produced progeny and then removed from the plate. Their progeny remaining on the auxin-containing plates were analyzed 48 hours after removal of the mothers. Similarly handled L4 worms on non-auxin NGM plates served as a control. (B) Staged L1 hermaphrodites were exposed to 1mM auxin (AUX) for 24h, 48h, and 72h and analyzed at these time points. Similarly staged unexposed larvae on non-auxin NGM plates served as a control. asna-1(A63V) animals (n = 5). Statistical significance was determined by the independent twosample t-test. Bars represent mean ± SD. (D) Relative mRNA analysis of ER stress reporter hsp-4 in 1-day old adult asna-1(A63V) animals. Statistical significance was determined by the independent two-sample t-test. Experiments were performed in triplicate. F44B9.5 was used as a normalizing control. Bars represent mean ± SEM. Relative mRNA analysis of (E) the mitochondrial stress reporters (hsp-6 and hsp-60) and (F) oxidative stress reporters (gst-4, gst-30, and gst-38) in 1-day old adult wild-type and asna-1(A63V) animals. Statistical significance was determined by the independent two-sample t-test. Experiments were performed in triplicate.  (1), cisplatin resistance/sensitivity (2) and DAF-28/insulin secretion (3). A tail anchored protein (TAP) is shown in green with a cylindrical C-terminal membrane spanning domain. Normal TAP insertion is depicted with three TAP protein molecules in the ER membrane bilayer, while lower TAP insertion is shown with one such molecule. ASNA-1 protein in loss-of-function (lf) scenario is shown in red to indicate the low ASNA-1 levels that are present only because of maternal inheritance in the homozygous mutants from asna-1/+ mothers. The altered redox balance in the point mutants (A63V and ΔHis164) resulting in lower levels of reduced ASNA-1 leads to engagement of fewer TAP protein molecules and consequently fewer TAP protein molecules inserted into the ER membrane bilayer. Created with BioRender. (TIFF) S1 Table. List of million mutation project strains carrying a point mutation in ASNA-1.