ASNA-1 oxidation induced by cisplatin exposure enhances its cytotoxicity by selectively perturbing tail anchored protein targeting

Cisplatin is a frontline cancer treatment, but intrinsic or acquired resistance is common. We previously showed that ASNA-1/TRC40 inactivation increases cisplatin sensitivity in mammalian cells and a Caenorhabditis elegans asna-1 knockdown model. ASNA-1 has conserved tail-anchored protein (TAP) targeting and insulin secretion functions. Here we examined the mechanism of ASNA-1 action. We show that ASNA-1 exists in two physiologically-responsive redox states with separable TAP-targeting and insulin secretion functions. Cisplatin-generated ROS targeted ASNA-1 oxidation, resulting in a selective targeting defect of an ASNA-1-dependent TAP. Increased ASNA-1 oxidation sensitized worms to cisplatin cytotoxicity. Mutants with a redox balance favoring oxidized ASNA-1 were cisplatin sensitive as null mutants by diverting ASNA-1 away from its TAP-targeting role and instead perturbing endoplasmic reticulum (ER) function. Mutations in the ASNA-1 receptor required for TAP insertion induced equivalent cisplatin sensitivity. We reveal a previously undescribed cellular dysfunction induced by cisplatin, identify a cisplatin target, and show that drug exposure causes TAP targeting-induced ER dysfunction. Therapeutic oxidation of ASNA-1 could be a clinically useful means to increase cisplatin sensitivity, reduce cytotoxic drug doses, and counteract cisplatin resistance. AUTHOR SUMMARY Cisplatin is a very effective anti-cancer drug and is widely used as a frontline treatment. However, tumor resistance limits its use. Tumor re-sensitization would improve cancer treatment. ASNA-1/TRC40 knockdown in Caenorhabditis elegans and mammals results in cisplatin hypersensitivity, but the underlying mechanistic details are largely unknown. We show that in C. elegans ASNA-1 mutants, increased cisplatin killing is coupled with delocalization of a tail-anchored protein, SEC-61β, a membrane protein that should reach the ER and is instead mistargeted. Like its homologs, the reduced form of worm ASNA-1 is needed for targeting activity. Targeting is blocked upon ASNA-1 oxidation after cisplatin treatment, likely via reactive oxygen species (ROS) generated by cisplatin treatment. Nevertheless, the oxidized form of the protein can execute other functions like insulin secretion. We show also that mutants with high oxidized ASNA-1 levels are cisplatin sensitive. Additionally, cisplatin induced mistargeting strictly acts through ASNA-1 inactivation. Thus, we define a pathway from cisplatin exposure that targets protein (ASNA-1) inactivation, consequently leading to mis-targeting of proteins that need ASNA-1 for their maturation. This multi-step process provides vital information about likely proteins that can be targeted by drugs to enhance cisplatin mediated killing and improve chemotherapy. SUMMARY STATEMENT Sensitizing tumors to cisplatin would be of considerable therapeutic benefit. Here we show a novel mechanism of cisplatin sensitization via oxidation of ASNA-1 in a Caenorhabditis elegans model.


INTRODUCTION
Solid tumors consist of both dividing and non-dividing post-mitotic cells (Komlodi-Pasztor et al., 2012), and both populations must be eliminated to reduce tumor bulk. The frontline cytotoxic chemotherapy cisplatin rapidly kills both dividing and non-dividing cells but via different mechanisms (Stewart, 2007). In dividing cells, cisplatin directly binds to DNA to cause DNA damage, cell cycle arrest, and apoptosis (Kelland, 2007;Legin et al., 2014). In nondividing cells, it kills by binding to antioxidants like glutathione, elevating reactive oxygen 3 species (ROS) levels that have downstream effects on a variety of cellular processes (Dasari and Tchounwou, 2014). Identification of proteins that are rapidly oxidized after cisplatin exposure would provide important information on the mechanism of cisplatin-induced death in post-mitotic cells. If cytotoxicity requires the oxidation of specific proteins, their targeted oxidation could be combined with lower doses of cisplatin to achieve equivalent anti-tumor activity as high-dose single agent regimens. The longstanding limitation of cisplatin use is the eventual development of tumor resistance and significant side-effects of nephrotoxicity and ototoxicity. The use of lower but effective doses would address these important limitations.
ASNA-1/TRC40/GET3 is a redox-sensitive protein whose knockdown increases the sensitivity of tumor cells to cisplatin-induced death (Hemmingsson et al., 2009). We previously established Caenorhabditis elegans as a relevant system to model the cisplatin killing effect in mammalian post-mitotic cells, since C. elegans asna-1 mutants are sensitive to cisplatininduced death (Hemmingsson et al., 2010). Human ASNA-1/TRC40 can act as a functional substitute in worm mutants, further underlining the model's similarity to human cells (Kao et al., 2007). We have also shown that ASNA-1 in mouse and C. elegans promotes insulin secretion and that targeted mouse knockouts develop type 2 diabetes (T2D) (Kao et al., 2007;Norlin et al., 2016). This work led us to propose that the cisplatin response and insulin secretion functions are separable (Hemmingsson et al., 2010).
A possible molecular basis for the separation of functions emerges from extensive cellular and structural biology studies on ASNA-1 homologs in yeast, mammals, and plants (Mateja et al., 2009;Mateja et al., 2015;Srivastava et al., 2016;Voth et al., 2014). The yeast homolog GET3 can adopt two alternative redox-sensitive forms with non-overlapping functions (Voth et al., 2014). The best understood function is of the reduced dimeric form, which acts as part of a targeting complex to guide tail-anchored membrane proteins (TAPs) for insertion into the endoplasmic reticulum (ER) membrane (Favaloro et al., 2008;Schuldiner et al., 2008;Stefanovic and Hegde, 2007). However, GET3 can also act as a general chaperone and a holdase by adopting an oxidized tetrameric structure with internal disulfide bonds via oxidation of critical redox-sensitive cysteines (Powis et al., 2013;Voth et al., 2014). In vitro, GET3 adopts a zinc-stabilized conformation needed for TAP targeting when conserved cysteines are reduced but acts as a holdase and general chaperone when these residues are oxidized (Voth et al., 2014).
Thus, a better understanding of how ASNA-1 works and which form of the protein should be targeted in different pathologies requires matching the disease with the molecular state before rational drug regimens can be devised. 4 Here we investigated links between cisplatin-induced oxidative stress, the redox-responsive nature of ASNA-1, and the role of ASNA-1 in cisplatin cytotoxicity. ASNA-1 is needed for ER targeting of a model TAP via its ER-localized receptor WRB-1. Specifically, we asked whether ASNA-1 activity in worms can be targeted and modified by cisplatin exposure and whether this would explain its role as a sensitivity factor. We show that reduced monomeric ASNA-1 (ASNA-1 RED ) protects the organism from cisplatin cytotoxicity. Cisplatin generates ROS, and drug exposure rapidly leads to higher levels of oxidized ASNA-1 (ASNA-1 OX ) at the expense of ASNA-1 RED and cisplatin treatment in wild-type worms perturbs model TAP targeting in an ASNA-1 dependent manner. Cisplatin-induced oxidation of ASNA-1 is not a bystander effect.
Rather, mutants with increased levels of oxidized ASNA-1 are sensitive to cisplatin death. We find that increased oxidation of ASNA-1 is equivalent to complete depletion of ASNA-1 with respect to cisplatin response without having any effect on its insulin secretion function.
Increased levels of oxidized ASNA-1 also re-distributes the protein away from its juxtamembrane localization. Failure of TAP targeting is important for cisplatin-induced death, because wrb-1 -/mutants are also cisplatin sensitive. Thus, we delineate a pathway in which cisplatin-induced ROS generation inactivates the cisplatin response function of ASNA-1, which in turn perturbs the targeting of a TAP protein to the ER membrane. Taken together, our studies identify ASNA-1 as a biologically relevant cisplatin target, reveal which molecular form of ASNA-1 is modulated, and demonstrates that one strategy to increase cisplatin sensitivity would be biasing ASNA-1 to the oxidized state.

ASNA-1 is present in physiologically redox-sensitive states
The oxidation state of the S. cerevisiae ASNA-1 homolog, GET3, modulates its activities. We therefore asked whether C. elegans ASNA-1 could also act as a redox-regulated switch and whether this switch responds to internal physiological changes. We performed non-reducing SDS-PAGE on proteins isolated from wild-type worms (Fig. 1A) and control worms expressing ASNA-1::GFP ( Fig. 1B-H), and found that C. elegans ASNA-1 was present in both its oxidized and reduced states ( Fig. 1A-H). GFP alone was not detected in the oxidized state (Fig. 1H).
Consistent with the findings in yeast (Voth et al., 2014), there was no oxidation of ASNA-1::GFP when two conserved cysteines at positions 285 and 288 were substituted for serines ( Fig. 1G). Moreover, the balance between ASNA-1:GFP OX and ASNA-1:GFP RED could be 5 altered by exposure of worms to different pro-oxidants: 10 mM H2O2/1mM CuSO4 (Fig. 1B) or 5 mM sodium arsenite (Fig. 1C). ASNA-1:GFP OX was diminished by exposing animals to the antioxidant mitoTempo (Fig. 1D). The balance shifted towards more ASNA-1:GFP OX at the expense of ASNA-1:GFP RED in sod-2(gk257) and mev-1(kn1) mutants, which have high endogenous ROS levels (Fig. 1E,F) (Labuschagne et al., 2013). We concluded that ASNA-1 exists in two alternative molecular states and that their balance can be altered by environmental factors and internal physiological changes.

ASNA-1 promotion of TAP insertion is independent of its role in insulin secretion
We next examined whether redox-induced structural changes in ASNA-1 impacted its wellstudied functions and whether these functions were affected by alterations of the cellular ASNA-1 redox balance. While the ASNA-1/TRC40 pathway guides TAPs to the ER membrane, other pathways (EMC, HSP70/HSC40, SND) can potentially compensate for the absence of ASNA-1 to produce little overall defect in TAP targeting (Aviram and Schuldiner, 2017). To investigate whether the TAP targeting function was affected by the absence of worm ASNA-1, we established an in vivo model for TAP targeting in C. elegans. In vertebrates and yeast, SEC61β is a model TAP that requires ASNA1/TRC40 for correct targeting to the ER membrane (Favaloro et al., 2008;Hegde and Keenan, 2011;Schuldiner et al., 2008;Stefanovic and Hegde, 2007) via binding of ASNA-1 homologs to the transmembrane domain (TMD) of SEC61β. This domain is highly conserved in the C. elegans homolog sec-61.B (hereafter called SEC61β). Indeed, Co-IP/MS/MS analysis detected SEC61β as an ASNA-1::GFP-interacting partner ( Table 1). Cell fractionation using GFP-tagged SEC61β confirmed that the protein was found only in the membrane fraction, even when expressed under a strong intestine-specific promoter (Fig. 2F). Co-expression of GFP::SEC-61β in intestinal cells of wild-type C. elegans with the mCherry-tagged rough ER (RER)-specific protein, SP12 (spcs-1), resulted in near complete colocalization in a pattern characteristic of ER ( Fig. 2A,B and Fig. S1). This colocalization pattern was disrupted in animals expressing GFP::SEC61β ΔTMD with its TMD deleted. Instead, large aggregates of GFP::SEC61β ΔTMD accumulated in cytoplasmic regions distinct from the RER ( Fig. 2A,B). Taken together, this established that worm SEC-61β was an ASNA-1-interacting partner that localized to the ER membranes via its TMD and thus was a good model protein to further study the contribution of ASNA-1 to TAP targeting. This localization required ASNA-1 since, in the two worm asna-1 protein null mutants (ok938 and sv42) (Kao et al., 2007), the ER localization of GFP::SEC-61β was significantly decreased and the protein was instead detected in cytoplasmic foci ( Fig. 2A,B). The targeting defect was solely 6 due to lack of ASNA-1, since wild-type ASNA-1 expressed from a transgene completely rescued the TAP targeting defect ( Fig. 2A,B and Fig. S1). Detection of the same phenotype in two independent deletion mutants and rescue with wild-type ASNA-1 ruled out any contribution from background genetic effects.
asna-1 mutants have multiple phenotypes including elevated autophagy levels ( Fig. S2D), high ER stress (Natarajan et al., 2013), and low insulin signaling levels (Hemmingsson et al., 2010;Kao et al., 2007). We asked whether these phenotypes contribute to the TAP targeting defect in asna-1(ok938) mutants. Inducing ER stress in wild-type animals to levels equivalent to asna-1(ok938) mutants ( Fig. S2E) and did not affect GFP::SEC-61β targeting (Fig. 2C). Starvation, which reduces insulin signaling and increases autophagy levels in C. elegans (Henderson and Johnson, 2001) and disrupts insulin secretion in unc-31/CAPS (Speese et al., 2007) mutants, did not affect GFP::SEC-61β targeting to the ER (Fig. 2C). Conversely, forcing high levels of DAF-28/insulin secretion in the tom-1/tomosyn mutant (Gracheva et al., 2007) failed to suppress the TAP targeting defect of asna-1(ok938) animals (Fig. 2C). Therefore, changes in ER morphology, insulin secretion activity, autophagy, or ER stress levels do not contribute to TAP targeting defects in asna-1 mutants. These findings support the hypothesis that the insulin secretion function of ASNA-1 and high ER stress levels do not contribute to defective TAP targeting in asna-1 mutants, thereby providing the first evidence of independent ASNA-1 functions in C. elegans.
A receptor for TAP insertion interacts with ASNA-1 and has a role in cisplatin detoxification 7 ASNA-1-mediated insertion of TAPs requires a receptor at the ER membrane. WRB is a component of the heterodimeric receptor for TAP targeting (Vilardi et al., 2011). Its C. elegans homolog WRB-1 is the likely receptor for ASNA-1. Using co-IP/MS/MS analysis, we detected WRB-1 as an interaction partner of ASNA-1::GFP ( Table 1). Western blot analysis showed a significant decrease in ASNA-1 protein levels in wrb-1(tm5938) mutants (Fig. 2D) indicative of their intracellular association. WRB-1::GFP was detected in the ER (Fig. 2E) and was required for GFP::SEC-61β targeting to the ER membrane ( Fig. 2A,B) to roughly the same extent as ASNA-1. Furthermore, wrb-1(tm5938) produced a phenotype similar to asna-1 mutants for proteinaceous aggregates (Fig. S2A), swollen ER (Fig. S2B), and autophagosomes ( Fig. S2C). To determine whether WRB-1 is required for other ASNA-1 functions such as cisplatin responses, we tested the survival of wrb-1(tm5938) mutants after exposure to cisplatin; wrb-1 mutants displayed cisplatin hypersensitivity (Fig. 3D). Thus, mutations in two different genes that similarly affect TAP targeting also demonstrated increased cisplatin sensitivity, suggesting that TAP targeting is associated with cisplatin sensitivity.

Altered redox balance that favors ASNA-1 OX causes cisplatin sensitivity
To obtain direct evidence for the separable functions of ASNA-1, we tested a set of asna-1 point mutant strains for cisplatin sensitivity (Thompson et al., 2013). asna-1(gk592672) substitutes a highly conserved alanine at position 63 to a valine, hereafter called asna-1(A63V).
This mutation did not affect the ASNA-1 protein stability (Fig. 4A). Transgene-expressed ASNA-1 A63V ::GFP protein was detected in the ER (Fig. 4B). The 13-times outcrossed asna-1(A63V) mutant was as sensitive to cisplatin exposure as the deletion mutant, and this phenotype was rescued by wild type transgene-expressed ASNA-1::GFP (Fig. 4C). This transgene also rescued the ASNA-1 protein null mutant for the cisplatin sensitivity phenotype The antioxidant epigallocatechin gallate (EGCG) reduces cisplatin effectiveness (Pan et al., 2015) and has been used as an antioxidant in several C. elegans studies (Xiong et al., 2018;Zhang et al., 2009). Notably, pre-treatment with EGCG reduced asna-1(A63V) cisplatin sensitivity but had no protective effect in the asna-1 deletion mutant (Fig. 4C). Modulation of cisplatin sensitivity by EGCG suggested that the effect of the A63V change might alter the balance between oxidized and reduced ASNA-1 and that ASNA-1 was a direct target for the antioxidant effect of EGCG. Comparison of oxidized and reduced ASNA-1 A63V ::GFP to ASNA-1::GFP revealed that the point mutation shifted the balance from the reduced to the oxidized form (Fig. 4L,M), indicating that a shift to more oxidized ASNA-1 caused increased cisplatin sensitivity. To examine this further, we tested the redox balance in worms expressing ASNA-1 ΔHis164 ::GFP, which is inactive for rescue of the cisplatin response phenotype (Hemmingsson et al., 2010). Again, substantially more oxidized ASNA-1 was detected ( Fig.   4L,M). To examine this further, we tested an asna-1(ΔHis164) mutant strain generated by 9 Crispr/CAS9 technology. This strain was as sensitive to cisplatin as the asna-1(ok938) deletion mutant (Fig. 4C). Thus, two ASNA-1 mutants that displayed increased oxidation were as sensitive to cisplatin as the deletion mutants. Moreover, sod-2(gk257) and mev-1(kn1) mutants, which harbor high levels of ASNA-1 OX without changes in overall ASNA-1 levels (Fig. 1C,D), were also cisplatin sensitive (Fig. 6C). Taken together, a redox balance that favors oxidized ASNA-1 results in the same cisplatin sensitivity phenotype as complete depletion of ASNA-1, leading to the conclusion that reduced ASNA-1 protects against cisplatin toxicity. Notably, cisplatin sensitivity becomes a redox-sensitive event in the asna-1 A63V genetic background, indicating a strategy to modulate its killing effect. Strikingly, the shift in ASNA-1 A63V to the oxidized state occurred even in the absence of an increase in the expression of markers of ER stress (hsp-4) ( Fig. S4A-C), mitochondrial stress (hsp-6 and hsp-60) (Fig. S4F), or oxidative stress (gst-4, gst-30 and gst-38) ( Fig. S4G) in contrast to high ROS levels in sod-2(gk257) and mev-1(kn1) mutants. We conclude that the A63V and ΔHis164 mutants in ASNA-1, which have inherently high ASNA-1 OX levels, are sensitive to cisplatin but have normal insulin secretion function. This analysis proves that the redox balance of ASNA-1 is necessary for maintaining its essential functions and that reduced ASNA-1 protects against cisplatin toxicity, most likely through its TAP targeting activity.

Cisplatin inactivates the protective function of ASNA-1 by converting it to the oxidized state
Having shown that reduced ASNA-1 is protective, we sought to examine whether part of cisplatin's killing mechanism was by directly targeting and inactivating ASNA-1. Several groups have shown in mammalian cancer cell models that a central feature of cisplatin treatment is oxidative stress induction and ROS generation (Brozovic et al., 2010;Choi et al., 2015;Itoh et al., 2011;Lu et al., 2016). Since ASNA-1 is required for cisplatin resistance and ASNA-1 function is modulated by oxidation, we next asked if cisplatin exposure induced oxidative stress in C. elegans. A short cisplatin exposure regimen was chosen at 300 µg/mL or 600 µg/mL, since these concentrations do not decrease the viability of wild-type animals, thereby allowing meaningful cell biology analysis. These experimental conditions were indeed sufficient to induce significant elevations in ROS (Fig. 5D), increased protein carbonylation (Fig. 5C), and robustly activate the oxidative stress response genes gcs-1 and gst-4 (Fig. 5A,B). We next tested whether cisplatin exposure influences the oxidation state of ASNA-1. ASNA-1::GFPexpressing worms were exposed to cisplatin at a concentration that significantly reduced the survival of asna-1(ok938) mutants without affecting wild-type animals (Fig. 4C). Under these 10 conditions, the balance of ASNA-1::GFP shifted towards the oxidized form (Fig. 5E,F). Thus, cisplatin exposure directly increases asna-1 oxidation, allowing us to conclude that ASNA-1 is a molecular target of cisplatin.

Cisplatin selectively delocalizes an ASNA-1-dependent TAP from the ER
Having established that ASNA-1 RED protects against cisplatin toxicity and mindful of the role of the reduced form in TAP targeting, we next determined whether cisplatin treatment and TAP targeting were directly related. To this end, we tested whether the increased oxidized ASNA-1 observed in cisplatin-treated worms resulted in a biologically meaningful outcome. Exposure of asna-1(+) animals expressing GFP::SEC61β to cisplatin significantly delocalized this TAP away from the ER membrane at levels similar to those seen in untreated asna-1(ok938) animals ( Fig. 6A,B). Remarkably, cisplatin treatment had no effect on the ER localization of the two ASNA-1-independent TAPs CytB5.1 (cytb-5.1) and SERP-1.1 (Fig. 6A,B). Quantitative confocal microscopy revealed that conditions that substantially delocalized GFP::SEC-61β had no effect on the ER localization of GFP::CYTB-5.1 and GFP::SERP-1.1. Therefore, delocalization of GFP::SEC-61β in cisplatin-treated worms was not due to more a more generalized effect on membrane integrity. We next tested the effect of cisplatin exposure on DAF-28/insulin::GFP secretion and observed no secretion defect (Fig. 5G); the membrane events associated with insulin maturation, packaging into dense core vesicles, and release were unaffected. Therefore, the effects on GFP::SEC-61β localization were not caused by general cisplatin effects on cellular function and animal health. Since cisplatin oxidizes ASNA-1 and shifts the protein to its non-TAP associated role, we next asked whether GFP::SEC-61β delocalization in cisplatin-treated animals was solely via its modulation of ASNA-1 activity or due to another role in this process. ER targeting of GFP::SEC61β was compared in asna-1(ok938) mutants with and without cisplatin treatment; GFP::SEC61β was delocalized to the same extent in both conditions (Fig. 6A,B). Since there was no additive effect, cisplatin exposure requires ASNA-1 activity to exert its TAP delocalization effect.
Our studies support the hypothesis that cisplatin directly targets ASNA-1, leading to its rapid oxidation without affecting insulin secretion. We conclude that a shift in redox balance that lowers ASNA-1 RED levels blocks insertion of its client TAP into the ER membrane.
Furthermore, cisplatin treatment of wild-type worms phenocopied the asna-1 deletion mutant phenotype for delocalization of GFP::SEC61β but not for the insulin secretion function. The ER is an organelle whose function is affected by cisplatin exposure via its effect on ASNA-1 oxidation.

DISCUSSION
Cisplatin-generated ROS is a central event in post-mitotic cells. The precise targets of these oxygen species are poorly understood, and their identification remains critical to understanding cisplatin cytotoxicity. Here we identified ASNA-1 as an oxidative stress-induced molecular target of cisplatin and provide a molecular basis for the cisplatin hypersensitivity caused by ASNA-1 knockdown.
This study was underpinned by the observations that cisplatin induces oxidative stress (Itoh et al., 2011), ASNA-1 knockdown increases sensitivity to cisplatin (Hemmingsson et al., 2010), that the yeast homolog, GET3, exists in both oxidized and reduced states, and that GET3 must be reduced to target some TAPs to the ER membrane (Voth et al., 2014). Using live cell imaging, we have shown that worm ASNA-1 also has TAP targeting activity and that WRB-1 is its likely receptor for this function. wrb-1 mutants were also cisplatin sensitive, thereby linking TAP targeting to cisplatin sensitivity and uncovering a new cisplatin sensitivity factor.
ASNA-1 existed in both oxidized and reduced states in a manner requiring redox-reactive cysteines. We provide several lines of evidence showing that elevated ASNA-1 OX levels increases cisplatin sensitivity. Strikingly, a redox-imbalanced mutant that favored accumulation of ASNA-1 OX was as sensitive as a mutant lacking ASNA-1. Cisplatin exposure rapidly increased ROS and ASNA-1 OX levels. Moreover, an ASNA-1-dependent TAP was specifically delocalized as a result of cisplatin treatment.

ASNA-1 is required for a TAP-targeting protein
We characterized the TAP-targeting activity of C. elegans ASNA-1 and found that it possesses robust TAP targeting activity shared with its binding partner, the ER-localized protein WRB- Consequently, it was important to establish the in vivo contribution of ASNA-1 to TAP targeting in intact animals. We assayed this activity in large intestinal cells, because ASNA-1 12 expressed by these cells was amenable to quantitative confocal microscopy. We showed that the other pathways do not bypass the need for ASNA-1 activity, since SEC-61b, which binds to ASNA-1 ( Table 1), was significantly delocalized in asna-1 mutants ( Fig. 2A,B).
Importantly, using this assay, we also found two ASNA-1-independent ER-localized TAPs (Fig. 6A), in agreement with findings in other systems (Colombo et al., 2009;Favaloro et al., 2008;Stefanovic and Hegde, 2007). The contribution of other pathways to the targeting of ASNA-1-dependent TAPs in worms is unclear. It is possible that if TAP targeting is studied in C. elegans tissues other than the intestines, the EMC or the HSP70/HSC40 pathways may play a greater role in the process.

WRB-1 participates in TAP targeting and is the likely ER receptor for ASNA-1
C. elegans WRB-1 was identified as the closest homolog of mammalian WRB, containing two predicted trans-membrane domains and the important coiled-coil domain. Evidence that WRB-1 is the receptor for the ASNA-1/TAP complex comes from the fact that it was ER-localized ( Fig. 2E) and was an ASNA-1-binding protein ( Table 1), as with its homologs in other species (Carvalho et al., 2019;Vilardi et al., 2011). Moreover, ASNA-1 protein levels were lower in wrb-1 mutants (Fig. 2D), a property also seen upon mammalian WRB knockdown (Colombo et al., 2016;Rivera-Monroy et al., 2016). Taken together, these three properties of WRB-1 satisfy the requirement to regard this protein as the ER-based receptor. Crucially, TAP localization was defective in two wrb-1 mutants to a similar extent seen in asna-1 mutants (Fig.   2B).

Cisplatin cytotoxicity is linked to defects in TAP targeting but not the insulin secretion function of ASNA-1
The observation that TAP targeting was defective in wrb-1 mutants allowed us to address two important questions about TAP protein function. First, we showed that wrb-1 mutants were also hypersensitive to cisplatin (Fig. 3D), thus identifying mutants in two separate genes with TAP targeting and cisplatin hypersensitivity function. Second, we found that wrb-1 mutants were not defective in insulin/IIS signaling and secretion function (Fig. 3), as observed in asna -1 mutants (Kao et al., 2007). We therefore concluded that insulin/IIS signaling pathway defects play no role in the asna-1(-) cisplatin hypersensitivity phenotype. Moreover, this also meant that ASNA-1-dependent TAP targeting has little role in ASNA-1-induced DAF-28/insulin secretion. This was consistent with our previous finding that daf-2/insulin receptor mutants without insulin signaling capacity are completely resistant to cisplatin (Hemmingsson et al., 13 2010). Further evidence that the cisplatin and insulin functions of ASNA-1 are separable emerged from the analysis of the asna-1(A63V) mutant, which had the striking phenotype of completely separating the cisplatin and insulin secretion functions (Fig. 4): "null" for the cisplatin phenotype and "wild-type" for the growth and insulin secretion function of ASNA-1.
Since ASNA-1 protein levels were similar in wild-type and the asna-1(A63V) mutant (Fig. 4A), this functional separability (cisplatin versus insulin secretion) was not due to a threshold effect, in which a differential phenotypic effect would be seen if one function required less total ASNA-1 protein.

C. elegans ASNA-1 is a redox-sensitive protein
C. elegans ASNA-1 exists in both reduced (ASNA-1 RED ) and oxidized (ASNA-1 OX ) forms ( Fig.   1, Fig. 7A). This is consistent with Voth et al. (2014), who showed that yeast GET3 changes from a reduced, ATPase-dependent TAP targeting state to an oxidized ATPase-independent general chaperone/holdase state. The ASNA-1 RED /ASNA-1 OX balance was biologically relevant, since it was sensitive to changes in animal physiology. Worm mutants with high ROS levels or cultivation conditions that increase ROS shifted the balance towards more oxidized ASNA-1, while exposure to antioxidants shifted the balance in the opposite direction (Fig. 1).
The ASNA-1 RED /ASNA-1 OX balance had biological relevance, since cisplatin sensitivity was increased in the asna-1(A63V) mutant that produced a preferentially oxidized protein variant ( Fig. 4L, Fig. 7B). Notably, the cisplatin hypersensitivity of asna-1(A63V) animals was reduced by pre-treatment with the antioxidant EGCG, indicating that the redox state of ASNA-1(A63V) is relevant for cisplatin cytotoxicity (Fig. 4C). Reversal of the cisplatin sensitivity phenotype with antioxidants strongly supports this conclusion. The shift in the ASNA-1 RED /ASNA-1 OX balance towards higher ASNA-1 OX levels is not a peculiarity of the alanine 63 mutation; indeed, ASNA-1 ΔHis164 displayed even higher ASNA-1 OX levels (Fig. 4L). Transgenic expression of ASNA-1 ΔHis164 does not rescue cisplatin hypersensitivity of the null mutant but maintains largely intact insulin/IGF signaling (Hemmingsson et al., 2010). We speculate that ASNA-1 point mutants that are TAP-targeting defective exist at higher levels in the oxidized state.
ASNA-1 is member of a group of C. elegans proteins, along with PRDX-2, IRE-1, and GLB-12, whose activity changes with oxidation (Henau et al., 2015;Hourihan et al., 2016;Oláhová et al., 2008). In vivo redox changes are dynamic in C. elegans with respect to both tissue type and age (Back et al., 2012;David et al., 2010;Walther et al., 2015). Oxidized proteins are likely 14 to be widespread in the proteome, since quantitative redox proteomics has identified many proteins with redox-sensitive cysteines in H2O2-treated animals (Kumsta et al., 2011) and in day 2 adults (Knoefler et al., 2012).

Increased levels of oxidized ASNA-1 sensitize cells to cisplatin killing
Only 5-10% of cisplatin is found in the nuclei of exposed cells, the remainder being cytoplasmic and membrane associated (Brozovic et al., 2010). Non-nuclear cisplatin has been shown to modify kinase signaling, Ca 2 + pump activity, and membrane dynamics (Brozovic et al., 2010;Dasari and Tchounwou, 2014). In light of our findings that cisplatin modifies ASNA-1 function (Fig. 7B), it will be instructive to study which -if any -of these processes are caused by changes in ASNA-1 oxidation levels and ASNA-1-dependent TAP targeting. Our study shows that cisplatin has an effect on the ER in worm intestinal cells, which was not due to a general effect on cellular membranes at the concentrations of cisplatin used, since only ASNA-1-dependent TAPs were delocalized and ASNA-1-independent TAPs were targeted normally (Fig. 6).
Further, this treatment had no effect on DAF-28/insulin secretion, a process that also requires several membrane budding and fusion events at various cellular locations (Fig. 5G) the killing effect cannot only be attributed to induced DNA damage but is also likely to be partially be due to effects on the ER. We have shown that increased levels of oxidized ASNA-1 beyond a certain threshold will sensitize cells to cisplatin. We propose that small molecule drugs that increase oxidized ASNA-1 at the expense of the reduced form will enhance cisplatin cytotoxicity. Conversely, the antioxidant EGCG reduces cisplatin effectiveness. This work also reveals for the first time that cisplatin impacts TAP targeting. The cellular biology of the TAP pathway has been studied in detail and the roles of several other proteins besides ASNA1/TRC40 delineated. It is likely that small molecule drugs that affect other components of the TAP targeting pathway would also increase cisplatin sensitivity; indeed, wrb-1 mutants were as sensitive as asna-1 mutants and also represent a new cisplatin sensitivity factor. This analysis of ASNA-1 demonstrates that cisplatin perturbs ER function, which might also explain other effects of cisplatin on signaling pathways, Ca 2+ pump function, and membrane properties.
We previously showed that while cisplatin does not induce ER stress, combinatorial use with an ER stress inducer sensitizes resistant worms to cisplatin (Natarajan et al., 2013). This is 15 consistent with the idea that cisplatin treatment sensitizes the ER in a metastable state due to TAP targeting defects and if ER function is further compromised then the cells will become hypersensitive to cisplatin cytotoxicity. Further, increased endogenous levels of oxidized ASNA-1 increases cisplatin sensitivity. Drugs that increase ASNA-1 oxidation, target other components of the TAP biogenesis pathway, or induce mild effects on ER function will enhance cisplatin sensitivity, address the problem of cisplatin resistance, and permit lower combinatorial doses. Modeling drug action in non-mammalian genetic model systems is extremely useful for identifying potential protein targets to improve cancer therapy.
svIs143(Pnhx-2::mCherry::lgg-1) was generated by genomic integration (as above) of Insulin assays. Larval arrest phenotypes were scored in the F1 generation, grown at 20°C, from wrb-1 dsRNA-injected mothers. Worms harboring integrated daf-16::gfp and daf-28::gfp arrays were grown at 20°C and imaged using a Nikon Ni-E microscope, equipped with Hammamatsu Orca flash4.0 camera. daf-16::gfp animals were analysed within 10 minutes to 18 avoid artifacts due to stressful conditions. DAF-28::GFP uptake by coelomocytes was scored in adult svIs69; ok938, in svIs69; wrb-1(RNAi) worms and in svIs69;asna-1(A63V).  Cisplatin sensitivity assay. Cisplatin plates were prepared using MYOB media with 2 % agar in which the drug was added at a concentration of 300 μg/mL or 500 μg/mL. Cisplatin solution from a stock at 1 mg/mL (Accord Healthcare AB) was added to autoclaved medium after cooling it to 52°C. L4 larvae were picked to a new plate and allowed to grow for 24h before exposure to cisplatin. Cisplatin exposure was for 24 ± 1h at 20°C and death was determined by absence of touch-provoked movement when worms were stimulated by a platinum wire. Immunoprecipitation assay. Nematodes were grown on NGM plates with OP50 for 4 days.

RNA isolation and quantitative
Worms were washed 3 times with M9, lysed using Next Advance cell disruptor with 0.2mm stainless steel beads for 3 minutes at 4°C in lysis buffer (10mM Tris/Cl, 150mM NaCl, 0.5mM EDTA, 0,5% NP-40). Bradford protein assay was performed to measure protein concentration.
Lysate was added to 35µl of GFP-Trap®_MA magnetic beads (Chromotek) and tumbled endover-end for 1 hour at 4°C. Beads were magnetically separated and washed 3 times with 500µl of wash buffer I (0,01% Tween-20 in 50mM TEAB) for 10 min, followed by 3 times wash II (50mM TEAB) for 10 min. Elution of protein from the beads was performed by adding 50 µl 0.2M glycine (pH 2.5) for 30 sec, followed by magnetic separation. 5 µl 1M Tris base (pH10 .4) was added for neutralization. GFP expressing strain (unc-119(ed3);oxTi880) was used as a negative control for GFP-interacting partners. Experiment was performed three times for identification of WRB-1 as an interacting partner.

Carbonylated protein detection. Protein carbonylation was determined by OxyBlot Protein
Oxidation Detection Kit (Millipore). Wild-type animals were synchronized to the young adult stage. Worms were exposed to cisplatin at concentration of 300 μg/mL for 3 or 6 hours. Next worms were washed off the plate using M9 solution. Worms were homogenized as described   Error bars represent ± SD. (B) Quantification of relative GFP intensity for gcs-1::GFP (n ≥ 15) and gst-4::GFP ( n ≥15) transgenic strains. Statistical significance was determined by the independent two-sample t-test (*p < 0.05, **p < 0.01, ***p < 0.001). Bars represent mean ± SD.
(C) Influence of cisplatin exposure on protein oxidation evaluated using the OxyBlot kit. Proteins recovered from lysed worms exposed to cisplatin for 3 h at a concentration of 300 μg/mL or 600 μg/mL were labeled with DNP solution (Oxyblot) to visualize the presence of protein carbonyls by immunoblotting.
(D) ROS production levels as estimated by the H2DCFDA assay. Wild-type strain was exposed to 600 μg/mL cisplatin for 1 h. Statistical significance was determined by the independent two-sample t-test (n ≥ 1000). Experiments were performed in triplicate (*p < 0.05). Bars represent mean ± SD. and oxidized (ASNA-1 OX ). ASNA-1 RED participates in the insertion of TAP into the ER membrane, while ASNA-1 OX assures proper insulin secretion. (B) When the cells are exposed to cisplatin, elevated ROS levels perturb the ASNA-1 redox balance.
ASNA-1 shifts towards the ASNA-1 OX state at the expense of ASNA-1RED, which impairs TAP targeting to the ER without affecting insulin secretion. Because less ASNA-1 RED is available, ER homeostasis is affected and cells are more sensitive to cisplatin.
Bars represent mean ± SD.