Natural SEL1L variants rescue a model of NGLY1 deficiency and modify ERAD function and proteasome sensitivity

TL;DR

This study shows that natural variants in the SEL1L gene can improve survival in a model of NGLY1 deficiency by enhancing ERAD function and proteasome resistance.

Contribution

The study identifies SEL1L variants as novel modifiers of NGLY1 deficiency and provides insights into their role in ERAD and proteasome pathways.

Findings

SEL1L variants S780P and Δ806–809 increase survival in a Drosophila model of NGLY1 deficiency.

These variants improve resistance to ER stress and enhance ERAD function.

The variants also protect against proteasome inhibition defects in NGLY1 deficiency.

Abstract

N-glycanase 1 (NGLY1) deficiency is an ultra-rare disease caused by autosomal recessive loss-of-function mutations in the NGLY1 gene. NGLY1 removes N-linked glycans from glycoproteins in the cytoplasm and is thought to help clear misfolded proteins from the endoplasmic reticulum (ER) through the ER associated degradation (ERAD) pathway. Despite this, the physiological significance of NGLY1 in ERAD is not understood. The best characterized substrate of NGLY1 is NRF1, a transcription factor that upregulates proteasome expression and the proteasome bounce-back response. We previously performed a genetic modifier screen using a Drosophila model of NGLY1 deficiency and identified potential modifiers that alter the lethality of the model. We identified two protein-coding variants in Hrd3/SEL1L: S780P and Δ806–809. Both variants are localized to the SEL1L cytoplasmic tail, an uncharacterized domain. SEL1L is a component of the ERAD complex that retrotranslocates misfolded proteins from the ER to the cytoplasm for degradation. We used CRISPR to generate fly lines carrying these SEL1L variants in a common genetic background and tested them with our model of NGLY1 deficiency. Validating our previous screen, the SEL1LS780P and SEL1LΔ806-809 variants increased the survival of the NGLY1 deficiency model, compared to the SEL1LS780 variant. To determine how these SEL1L variants were modifying lethality in NGLY1 deficiency, we interrogated the ERAD and NRF1 signaling pathways. We found that the SEL1LS780P and SEL1LΔ806-809 variants improve resistance to ER stress, with enhanced ERAD function as a likely contributing mechanism. This effect depends on NGLY1 activity, further implicating NGLY1 in general ERAD function. We also found that, in heterozygous NGLY1 null flies, these variants protect against some defects like increased lethality caused by proteasome inhibition. These results provide new insights into the role of SEL1L in the disease pathogenesis of NGLY1 deficiency. SEL1L is a strong candidate modifier gene in patients, where variability in presentation is common.

NGLY1 deficiency is a debilitating rare genetic disorder. There are currently no treatment options for NGLY1 deficiency and NGLY1 biology remains poorly understood. We previously performed a genetic modifier screen in a Drosophila model of NGLY1 deficiency and identified a number of candidate modifier genes that impacted the survival of our model. Modifier genes can help reveal NGLY1 biology and NGLY1 deficiency disease pathogenesis. In this study, we follow-up on two natural protein-coding variants of Hrd3 (the fly version of the human gene, SEL1L) that increased the survival to adulthood of our NGLY1 deficiency model. SEL1L is a critical component of an important quality control pathway called the endoplasmic reticulum associated degradation (ERAD) pathway. We discovered that these SEL1L variants increased resistance to ER stress and may enhance ERAD function. These SEL1L variants also modify NGLY1 deficiency sensitivity to proteasome inhibition. This study confirms SEL1L as an important modifier gene of NGLY1 deficiency. Further study of the ERAD and proteasome degradation pathways may reveal additional candidate modifier genes of NGLY1 deficiency and potential targets for therapeutic development.

Figures

SEL1L variants increase NGLY1 deficiency eclosion rates and survival to adulthood.(A) Amino acid sequences for the three SEL1L variant alleles. Red boxes highlight differences between alleles. (B) SEL1L(S/P, +/+) and SEL1LΔ806-809/+ genotypes significantly increase eclosion rates and survival to adulthood (~70%) with NGLY1 knockdown when compared to SEL1L+/+ (45%, p < 0.0001). Chi-squared test. (C) Lifespan of NGLY1 KD flies shows no significant difference in survival to adulthood between SEL1L genotypes. Cox proportional hazard regression analysis.

SEL1L variants affect proteasome inhibition sensitivity.(A) NGLY1 + /- larvae are smaller when treated with 5μM BTZ, but there is no effect of SEL1L genotype (Larval size on DMSO: SEL1L+/+ 23.59 ± 4.97, SEL1LS780P/+ 19.37 ± 9.16, and SEL1LΔ806-809/+ 21.32 ± 6.73. Decreased larval size on BTZ: SEL1L+/+ 9.19 ± 1.61, p < 0.0001; SEL1LS780P/+ 8.55 ± 2.04, p < 0.0001; and SEL1LΔ806-809/+ 10.07 ± 1.14, p < 0.01). (B) When treated with 1μM BTZ, the NGLY1 + /- larvae show a SEL1L genotype-dependent decrease in size. While SEL1L+/+ showed no change between DMSO (19.65 ± 5.98) and BTZ (17.88 ± 4.32), SEL1LS780P/+ (DMSO 21.03 ± 3.42, BTZ 16.45 ± 3.65, p < 0.0001) and SEL1LΔ806-809/+ (DMSO 24.92 ± 3.39, BTZ 20.08 ± 2.69, p = 0.001) genotypes were smaller when treated with BTZ compared to DMSO treated larvae. (C) NGLY1 WT larvae show no significant decrease in size with 1μM BTZ treatment, regardless of SEL1L genotype (Larval size on DMSO: SEL1L+/+ 25.75 ± 6.91, SEL1LS780P/+ 23.74 ± 6.94, and SEL1LΔ806-809/+ 26.42 ± 4.29; larval size on BTZ: SEL1L+/+ 24.75 ± 5.63, SEL1LS780P/+ 20.96 ± 7.38, and SEL1LΔ806-809/+ 25.70 ± 5.23).

SEL1L variants increase survival under proteasome inhibition.(A) When treated with 1μM BTZ, NGLY1 + /- larvae eclose at significantly higher rates when carrying the SEL1LS780P/+ (83%, p < 0.0001) and SEL1LΔ806-809/+ (96%, p < 0.0001) genotypes compared to the SEL1L+/+ genotype (59%). (B) When treated with 1μM BTZ, adult NGLY1 + /- flies with the SEL1LS780P/+ (p < 0.0001) and SEL1LΔ806-809/+ (p < 0.0001) genotypes live longer than the SEL1L+/+. Cox proportional hazard regression analysis. (C) There is no significant difference in survival between adult NGLY1 + /- flies treated with DMSO regardless of SEL1L genotype.

SEL1L variants enhance resistance to ER stress in an NGLY1-dependent manner.To induce ER stress and test ERAD, we expressed Rh1G69D under the control of GMR-GAL4. Rh1G69D is a misfolded rhodopsin protein. The ‘Misfolded protein’ label refers to expression of Rh1G69D. (A) Knockdown of SEL1L reduces eye size in this ER stress model expressing Rh1G69D in the eye. The Control fly is wildtype for SEL1L. (Eye size of Control 12.38 ± 0.93, SEL1L RNAi 10.93 ± 0.65, p = 0.0016). (B) In the ER stress model, with an NGLY1 wildtype background, SEL1LS780P/+ (9.13 ± 0.49, p < 0.0001) and SEL1LΔ806-809/+ (9.29 ± 0.69, p < 0.0001) genotypes significantly increase eye size compared to the SEL1L+/+ genotype (8.36 ± 0.33). (C) Eye-specific knockdown of NGLY1 in the Rh1G69D ER stress model results in no difference in eye size between the SEL1L+/+ and SEL1LS780P/+ genotypes. The SEL1LΔ806-809/+ genotype shows a significantly smaller eye size (p < 0.05). (Eye size: SEL1L+/+ 12.60 ± 0.50, SEL1LS780P/+ 12.34 ± 0.89, and SEL1LΔ806-809/+ 12.16 ± 0.65). (D) Heterozygous loss of NGLY1 in the Rh1G69D ER stress model results in no difference in eye size between the SEL1L+/+ and SEL1LS780P/+ genotypes. The SEL1LΔ806-809/+ genotype shows a significantly smaller eye size (p < 0.0001). (Eye size: SEL1L+/+ 10.58 ± 0.76, SEL1LS780P/+ 10.88 ± 0.61, and SEL1LΔ806-809/+ 8.67 ± 0.54).

Full text

Introduction

N-glycanase 1 (NGLY1) deficiency is an ultra-rare disease and the first identified congenital disorder of deglycosylation (CDDG). The disease is caused by autosomal recessive loss-of-function mutations in the NGLY1 gene [1,2]. NGLY1 is a cytosolic deglycosylating enzyme that removes N-linked glycans from proteins. NGLY1 is thought to be a component of the endoplasmic reticulum (ER) associated degradation (ERAD) pathway, an important cellular quality control mechanism which removes misfolded proteins from the ER to the cytosol for degradation [3,4]. However, the loss of NGLY1 shows little effect on ERAD and does not prevent the degradation of misfolded proteins [5–7]. Therefore, despite its known function as a deglycosylating enzyme, the physiological significance of NGLY1 and disease pathogenesis remains poorly understood.

NGLY1 deficiency is marked by extensive phenotypic heterogeneity, even among patients with identical NGLY1 mutations [2,8], suggesting the presence of genetic modifiers. In a previous genetic screen, we identified 61 potential modifier genes that were associated with changes in survival in our Drosophila model of NGLY1 deficiency [9]. From this screen, our top hit was Ncc69 (Drosophila ortholog for human NKCC1/2), which encodes for a conserved ion transporter and we showed that it is both a substrate of NGLY1 and a modifier of NGLY1 deficiency [9]. Another interesting candidate modifier gene we identified in the screen was Hrd3 (hereon referred to by the human ortholog SEL1L). Through a genome-wide association study (GWAS), we identified a natural missense variant in SEL1L that was associated with increased survival in the Drosophila NGLY1 deficiency model [9]. This SEL1L variant is a substitution of serine 780 for a proline (SEL1L^S780P^). Additionally, we identified a private protein-coding deletion in the strain showing near complete rescue of NGLY1 deficiency lethality. This second variant is a deletion of amino acids 806–809 (SEL1L^Δ806-809^). Both variants are 26 amino acids apart and are located in the cytoplasmic tail of SEL1L.

SEL1L is a single-pass ER membrane protein and a critical, well-established component of ERAD. ERAD functions alongside other quality control mechanisms such as the unfolded protein response (UPR) and autophagy to maintain ER homeostasis and prevent ER stress [10–12]. The SEL1L-Hrd1 ERAD complex is the most conserved branch of ERAD from yeast to humans and translocates misfolded proteins from the ER to the cytosol for proteasomal degradation [13–15]. The luminal domain of SEL1L assists in the recognition of ERAD substrates in the ER lumen and the transmembrane domain helps move proteins through the ER membrane to the cytosol [16]. The cytoplasmic tail of SEL1L is a highly disordered region across species and its function is unknown.

In previous studies, both SEL1L and NGLY1 were identified as genetic modifiers of NRF1, a transcription factor responsible for the proteasome bounce-back response [17,18]. NRF1 is co-translated and glycosylated in the ER before being retrotranslocated to the cytosol, a process that requires ERAD machinery, including SEL1L, HRD1, and p97 [13,17,19]. Once in the cytosol, NRF1 is deglycosylated by NGLY1, which edits asparagine residues to aspartate and is required for NRF1 activation, localization, and transcriptional function [18,20]. The protease DDI1/2 also cleaves NRF1 to promote its activation and nuclear translocation [20]. However, some studies suggest that full-length NRF1 can reach the nucleus in the absence of DDI1/2 activity, indicating that its role may be context or species-specific [17,18]. Although SEL1L has been genetically linked to NRF1 activity and is a critical component of the Hrd1 ERAD complex, its specific mechanistic role in NRF1 retrotranslocation has not been fully delineated [13,17,21]. Nevertheless, SEL1L is strongly implicated in mediating NRF1 dislocation from the ER and may modulate the efficiency of NRF1 activation under proteasome stress conditions. Under healthy, homeostatic conditions, activated NRF1 is constitutively degraded by the proteasome. Under conditions of proteasomal stress, however, NRF1 is not degraded, accumulates in the cytosol, and is transported to the nucleus where it acts as a transcription factor and upregulates genes that increase proteasome function, including proteasome subunit genes [17,18,22]. This activation of NRF1 is known as the proteasome bounce-back response. NGLY1 and ERAD machinery are necessary for NRF1 activation and the loss of either prevents the proteasome bounce-back response [17,18].

In this study, we characterized the functional consequences of the two new SEL1L variants identified in our NGLY1 deficiency genetic screen. We placed the SEL1L variants on an isogenic background to test the effects each variant has on both SEL1L and NGLY1. The SEL1L variants increased eclosion rates and survival to adulthood in our NGLY1 deficiency model, validating the observations from the modifier screen. The SEL1L variants also enhance ERAD in an NGLY1-dependent manner and provide a protective fitness advantage during proteasome inhibition. Our results suggest that SEL1L is a modifier of NGLY1 and that interactions between SEL1L, NGLY1, and NRF1 may underlie these observed changes in fitness. These genetic interactions are potential targets for NGLY1 deficiency treatment.

Results

SEL1LS780P and SEL1LΔ806-809 variants increase eclosion rates and survival to adulthood of NGLY1 deficiency model

In a previous study, we crossed our NGLY1 deficiency Drosophila model, which uses the GAL4/UAS system to ubiquitously express RNAi against NGLY1, with nearly 200 strains of the Drosophila Genetic Reference Panel (DGRP) [9,23]. On a standard laboratory background, the NGLY1 deficiency model has ~ 30% survival to adulthood [7,9]. In the DGRP strains, eclosion rates and survival to adulthood of the NGLY1 deficiency model ranged from 0 to 100%, indicating that lethality is highly modifiable by genetic background. We performed a GWAS to identify candidate modifier genes associated with increased survival to adulthood. One of the top associated variants was the S780P missense variant in SEL1L. The SEL1L^S780P^ minor allele was associated with increased eclosion rates and survival to adulthood of the NGLY1 deficiency model, compared to the common SEL1L^S780^ allele. We also discovered that the DGRP strain (DGRP strain 379) with a nearly 100% eclosion rate and survival to adulthood in the screen harbored a private SEL1L variant, resulting in the deletion of amino acids 806–809 (SEL1L^Δ806-809^). This strain also carries the more common S780 allele. SEL1L^Δ806-809^ was not formally identified through the GWAS because it is a private variant in a single DGRP strain. Because these two variants were both in the cytoplasmic tail of SEL1L, a functionally uncharacterized region of the protein, we sought to understand how these variants were affecting eclosion rates and survival to adulthood of the NGLY1 deficiency model. We used CRISPR to place each of the SEL1L variants (Fig 1A) onto the same isogenetic background, creating three strains that are homozygous for SEL1L^S780^, SEL1L^S780P^, or SEL1L^Δ806-809^. This allowed us to test for phenotypic differences specific to each SEL1L variant.

SEL1L variants increase NGLY1 deficiency eclosion rates and survival to adulthood.(A) Amino acid sequences for the three SEL1L variant alleles. Red boxes highlight differences between alleles. (B) SEL1L(S/P, +/+) and SEL1LΔ806-809/+ genotypes significantly increase eclosion rates and survival to adulthood (~70%) with NGLY1 knockdown when compared to SEL1L+/+ (45%, p < 0.0001). Chi-squared test. (C) Lifespan of NGLY1 KD flies shows no significant difference in survival to adulthood between SEL1L genotypes. Cox proportional hazard regression analysis.

In our original screen, we crossed a strain carrying both a GAL4 and NGLY1 RNAi transgene with strains of the DGRP. Lethality was scored based on eclosion rates and survival to adulthood of the F1 flies, which had half of their genomes from the NGLY1 RNAi strain and half from the different DGRP strains. Because the NGLY1 RNAi strain is homozygous for the common, wildtype SEL1L^S780^ allele, any new SEL1L variant introduced by the DGRP strain in the F1 generation was heterozygous with the SEL1L^S780^ allele. For clarity, we will refer to the more common major S780 allele as “wildtype” or SEL1L^+/+^. Therefore, the relevant SEL1L genotypes from our screen are: SEL1L^+/+^, SEL1L^S780P/+^, or SEL1L ^Δ806-809/+^. We focus our analyses on these three SEL1L genotypes throughout this study.

Based on the original DGRP screen, we expected that the SEL1L^S780P/+^ and SEL1L^Δ806-809/+^ genotypes would increase eclosion rates and survival to adulthood of the NGLY1 model, compared to the wildtype SEL1L genotype. To validate the results of the screen, we crossed the same NGLY1 RNAi strain used in the modifier screen with each of our new SEL1L variant CRISPR strains to generate flies that have the exact SEL1L genotypes from the screen. The eclosion rates and survival to adulthood was determined in the same manner as the original screen, by dividing the number of NGLY1 knockdown flies by the largest balancer class in its cross [9]. There were significantly increased eclosion rates and survival to adulthood of NGLY1 knockdown flies with the SEL1L^S780P/+^ and SEL1L^Δ806-809/+^ genotypes compared to the SEL1L^+/+^ genotype (Fig 1B and S1 Data). The SEL1L^+/+^ genotype had a 45% eclosion rate and survival to adulthood compared to ~70% in the SEL1L^S780P/+^ (p < 0.001) and SEL1L^Δ806-809/+^ (p < 0.001) genotypes. This result nicely replicates our previous genetic screen that showed increased eclosion rates and survival to adulthood in the DGRP lines with these particular SEL1L variants and suggests that the SEL1L^S780P^ and SEL1L^Δ806-809^ alleles are protective against NGLY1 deficiency. We also evaluated the lifespan of these surviving NGLY1 knockdown flies, but found no significant differences between the SEL1L^S780P/+^ and SEL1L^Δ806-809/+^ and the SEL1L^+/+^ genotype (Fig 1C and S1 Data), suggesting that the interaction occurs during development.

SEL1L variants impact sensitivity to proteasome inhibition in NGLY1+/- flies

We next interrogated pathways that involve both SEL1L and NGLY1 to understand how these SEL1L variants are protecting against NGLY1 deficiency. Both NGLY1 and SEL1L were previously identified as modifier genes of NRF1, which encodes for an important transcription factor that upregulates proteasome genes in response to proteasomal stress [17,18]. The importance of NGLY1 in NRF1 function is well established. NGLY1 mutants have reduced proteasome function and are exquisitely sensitive to proteasome stress because NRF1 is not processed [17,18,24,25]. Previous studies have shown that heterozygous NGLY1 null larvae, which are otherwise normal, are sensitive to proteasome inhibition, leading to larval size defects [24,25]. Although SEL1L was identified as a genetic modifier of NRF1, its role in NRF1 signaling has not been determined. We hypothesized that these SEL1L variants would affect NRF1 signaling and modify phenotypes in an NGLY1 deficiency model.

We tested whether the SEL1L variants affect proteasome sensitivity in NGLY deficient Drosophila using the proteasome inhibitor bortezomib (BTZ). Homozygous NGLY1 null Drosophila are embryonic lethal, but heterozygous NGLY1 null flies are phenotypically normal when unchallenged. We used heterozygous NGLY1 null flies as a model of NGLY1 deficiency because of their known increased sensitivity to proteasome inhibition [24,25]. Because the NGLY1 null strain also carries the common wildtype S780 allele, when we cross this strain with our CRISPR generated SEL1L variant strains, we generate heterozygous NGLY1 null flies with the same SEL1L genotypes to what we tested in the NGLY1 knockdown model: SEL1L^+/+^, SEL1L^S780P/+^, or SEL1L^Δ806-809/+^.

Heterozygous NGLY1 null Drosophila larvae develop smaller when exposed to proteasome inhibition, compared to NGLY1 wildtype and DMSO-treated heterozygous NGLY1 null controls [24,25]. In previous studies, heterozygous NGLY1 null larvae, when exposed to 5μM BTZ, are significantly smaller than DMSO-treated heterozygous NGLY1 null larvae [25]. We observed an equally strong decrease in larval size with the treatment of 5μM bortezomib in all heterozygous NGLY1 null larvae compared to DMSO controls, regardless of SEL1L genotype (Fig 2A and S2 Data). SEL1L genotype does not impact the size defects induced by 5μM BTZ in NGLY1 heterozygous null larvae.

SEL1L variants affect proteasome inhibition sensitivity.(A) NGLY1 + /- larvae are smaller when treated with 5μM BTZ, but there is no effect of SEL1L genotype (Larval size on DMSO: SEL1L+/+ 23.59 ± 4.97, SEL1LS780P/+ 19.37 ± 9.16, and SEL1LΔ806-809/+ 21.32 ± 6.73. Decreased larval size on BTZ: SEL1L+/+ 9.19 ± 1.61, p < 0.0001; SEL1LS780P/+ 8.55 ± 2.04, p < 0.0001; and SEL1LΔ806-809/+ 10.07 ± 1.14, p < 0.01). (B) When treated with 1μM BTZ, the NGLY1 + /- larvae show a SEL1L genotype-dependent decrease in size. While SEL1L+/+ showed no change between DMSO (19.65 ± 5.98) and BTZ (17.88 ± 4.32), SEL1LS780P/+ (DMSO 21.03 ± 3.42, BTZ 16.45 ± 3.65, p < 0.0001) and SEL1LΔ806-809/+ (DMSO 24.92 ± 3.39, BTZ 20.08 ± 2.69, p = 0.001) genotypes were smaller when treated with BTZ compared to DMSO treated larvae. (C) NGLY1 WT larvae show no significant decrease in size with 1μM BTZ treatment, regardless of SEL1L genotype (Larval size on DMSO: SEL1L+/+ 25.75 ± 6.91, SEL1LS780P/+ 23.74 ± 6.94, and SEL1LΔ806-809/+ 26.42 ± 4.29; larval size on BTZ: SEL1L+/+ 24.75 ± 5.63, SEL1LS780P/+ 20.96 ± 7.38, and SEL1LΔ806-809/+ 25.70 ± 5.23).

To determine whether there might be more subtle effects, we treated the heterozygous NGLY1 null larvae with a lower concentration of 1μM BTZ. The SEL1L^+/+^ larvae showed no significant size differences between BTZ and DMSO treatments. However, SEL1L^S780P/+^ (p < 0.0001) and SEL1L^Δ806-809/+^ (p = 0.001) larvae had significant decreases in larval size with 1μM BTZ treatment compared to DMSO (Fig 2B and S2 Data). This indicates that larval size in heterozygous NGLY1 null larvae carrying the SEL1L^S780P/+^ and SEL1L^Δ806-809/+^ genotypes are particularly sensitive to proteasome inhibition compared to SEL1L^+/+^. On an NGLY1 wildtype background, the different SEL1L genotypes showed no larval size changes with the 1μM BTZ treatment (Fig 2C and S2 Data).

SEL1L variants increase survival of NGLY1+/- flies in response to proteasome inhibition

To further examine the impact of the SEL1L variants on NRF1 signaling, we tested other phenotypes affected by NGLY1 deficiency and proteasome inhibition. When heterozygous NGLY1 null larvae were treated with 1μM BTZ, we observed a SEL1L variant specific effect on survival through eclosion to adulthood. The SEL1L^S780P/+^ and the SEL1L^Δ806-809/+^ larvae showed higher eclosion rates and survival to adulthood at 83% (p < 0.0001) and 96% (p < 0.0001), respectively, compared to the SEL1L^+/+^ flies at 59% (Fig 3A and S3 Data). SEL1L^+/+^ larvae treated with bortezomib showed high rates of non-eclosed and partially eclosed flies. The improved eclosion rates of the SEL1L^S780P/+^ and the SEL1L^Δ806-809/+^ genotypes suggests a protective effect of the SEL1L^S780P^ and SEL1L^Δ806-809^ variants against NGLY1 deficiency and proteasome inhibition during larval development. Importantly, when NGLY1 wildtype larvae were raised on 1μM BTZ, we observed no lethality and nearly 100% eclosion of flies, regardless of SEL1L genotype, demonstrating that lethality to BTZ is NGLY1-dependent (S1 Fig and S3 Data). SEL1L^S780P/+^ and SEL1L^Δ806-809/+^ genotypes provide near complete rescue of the heterozygous NGLY1 null larvae lethality induced by 1μM BTZ.

SEL1L variants increase survival under proteasome inhibition.(A) When treated with 1μM BTZ, NGLY1 + /- larvae eclose at significantly higher rates when carrying the SEL1LS780P/+ (83%, p < 0.0001) and SEL1LΔ806-809/+ (96%, p < 0.0001) genotypes compared to the SEL1L+/+ genotype (59%). (B) When treated with 1μM BTZ, adult NGLY1 + /- flies with the SEL1LS780P/+ (p < 0.0001) and SEL1LΔ806-809/+ (p < 0.0001) genotypes live longer than the SEL1L+/+. Cox proportional hazard regression analysis. (C) There is no significant difference in survival between adult NGLY1 + /- flies treated with DMSO regardless of SEL1L genotype.

To understand if the SEL1L variants impact NRF1 signaling in adult flies, we treated heterozygous NGLY1 null adult flies with bortezomib or DMSO and observed their long-term, adult survival. The adult flies were raised on DMSO during their larval stages and were either maintained on DMSO or BTZ food. SEL1L^+/+^ adult flies showed significantly decreased survival time on BTZ compared to the SEL1L^S780P/+^ (p < 0.0001) and the SEL1L^Δ806-809/+^ (p < 0.0001) genotypes (Fig 3B and S3 Data). The increased survival times of the bortezomib-treated SEL1L^S780P/+^ and SEL1L^Δ806-809/+^ genotypes indicate that these variants are protective against proteasome inhibition during adulthood. When treated with DMSO, heterozygous NGLY1 null adult flies displayed no significant differences in long-term survival, regardless of SEL1L variant genotypes (Fig 3C and S3 Data). These experiments demonstrate that the sensitivity to proteasome inhibition in NGLY1 deficiency is modified by these SEL1L variants, both during larval development and in adulthood. The effects of these SEL1L variants on proteasome sensitivity are only observed when NGLY1 function is reduced, indicating that the interaction between SEL1L and proteasome stress responses is contingent on the loss of NGLY1 activity.

SEL1L variants do not impact the expression of NRF1 target genes

In response to proteasome inhibition, NRF1 upregulates proteasome genes [22]. To test for changes in NRF1 signaling in our SEL1L variant flies, we examined the expression of several proteasomal subunit genes: prosalpha6, prosalpha3, prosbeta2, prosbeta4, and prosbeta5. These five genes were among other proteasome genes that we previously identified as downregulated in our NGLY1 deficiency fly model [7]. We treated flies with BTZ to induce NRF1 signaling. When we compared expression of the proteasome genes, in whole flies, between NGLY1 wildtype and heterozygous NGLY1 null adults, we mostly observed no differences in gene expression across each condition and SEL1L genotype (S2A Fig and S4 Data). As expected, we observed increases in proteasome gene expression with the treatment of BTZ compared to DMSO; however, we found no SEL1L variant specific differences in gene expression in either the NGLY1 wildtype or heterozygous NGLY1 null flies (S2B Fig and S4 Data). While this is unexpected in light of the previous phenotypic data we present, it is possible that there are specific effects in different tissues that are missed when we examine expression in whole flies. To date, it is unknown which tissues contribute to the lethality in NGLY1 deficiency flies and more work is needed to determine which tissues are most impacted by proteasome inhibition.

SEL1LS780P and SEL1LΔ806-809 variants enhance ERAD in an ER stress model

SEL1L is an integral component of the Hrd1 ERAD complex, which retrotranslocates misfolded proteins from the ER lumen and ubiquitinates them for degradation by the proteasome [10,13,15]. Although its role in ERAD remains unclear, NGLY1 has been shown to physically interact with proteins in the ERAD complex, including Derlin-1 and VCP [26,27]. The deglycosylation of ERAD substrates by NGLY1 is thought to prepare misfolded glycoproteins for proteasomal degradation [5,28,29].

We crossed a Drosophila eye model of ER stress with the SEL1L strains to determine the variant-specific effects on ERAD. In this model, a transgene carries an eye-specific GAL4 (GMR-GAL4) that drives the overexpression of a mutant misfolded rhodopsin protein (encoded by UAS-Rh1^G69D^) that constitutively misfolds and induces degeneration in the developing larval eye disc [30–33]. The misfolded rhodopsin protein leads to chronic ER stress in the eye disc, cell death, and a rough eye phenotype in adult flies. These eyes are also significantly smaller than wildtype eyes. This model is sensitive to changes in ERAD function and increased ERAD function is protective against eye degeneration [30]. In larval eye discs, overexpressing Hrd1, the ERAD protein essential for transporting misfolded proteins out of the ER, nearly completely restored the normal appearance of the eye [30]. Increasing Hrd1 levels enhances ERAD, helping to clear misfolded rhodopsin, which in turn prevents ER stress and protects against degeneration. Conversely, knockdown of ERAD components in this model leads to increased eye degeneration [30]. Because SEL1L is a component of the Hrd1 ERAD complex, modifying SEL1L should similarly affect phenotypes in the ER stress eye model.

We hypothesized that losing SEL1L would decrease ERAD and enhance the eye degeneration phenotype, leading to a smaller eye size. We expressed SEL1L RNAi in the eye discs to knockdown SEL1L in the ER stress model (and NGLY1 wildtype) and observed that the eyes were significantly smaller than in the flies without knockdown of SEL1L (p = 0.0016) (Fig 4A and S5 Data). The reduction in eye size observed in this model following SEL1L knockdown is expected, given SEL1L’s established role in ERAD function. We next crossed the different SEL1L variants onto this model (and NGLY1 wildtype) to investigate their effects on eye size. Because the ER stress model strain carries the common SEL1L wildtype allele, crossing our SEL1L variants onto this model provided the same SEL1L genotypes as previously described. We observed an increase in eye size of the SEL1L^S780P/+^ (p < 0.0001) and SEL1L^Δ806-809/+^ (p < 0.0001) genotypes compared to the SEL1L^+/+^ flies (Fig 4B and S5 Data), opposite of what we observed with the RNAi knockdown experiment. Given that previous studies show that improvement in ERAD function increases eye size, this result suggests that the SEL1L^S780P^ and SEL1L^Δ806-809^ alleles improve resistance to ER stress, likely through enhanced ERAD function, in an NGLY1 wildtype background.

SEL1L variants enhance resistance to ER stress in an NGLY1-dependent manner.To induce ER stress and test ERAD, we expressed Rh1G69D under the control of GMR-GAL4. Rh1G69D is a misfolded rhodopsin protein. The ‘Misfolded protein’ label refers to expression of Rh1G69D. (A) Knockdown of SEL1L reduces eye size in this ER stress model expressing Rh1G69D in the eye. The Control fly is wildtype for SEL1L. (Eye size of Control 12.38 ± 0.93, SEL1L RNAi 10.93 ± 0.65, p = 0.0016). (B) In the ER stress model, with an NGLY1 wildtype background, SEL1LS780P/+ (9.13 ± 0.49, p < 0.0001) and SEL1LΔ806-809/+ (9.29 ± 0.69, p < 0.0001) genotypes significantly increase eye size compared to the SEL1L+/+ genotype (8.36 ± 0.33). (C) Eye-specific knockdown of NGLY1 in the Rh1G69D ER stress model results in no difference in eye size between the SEL1L+/+ and SEL1LS780P/+ genotypes. The SEL1LΔ806-809/+ genotype shows a significantly smaller eye size (p < 0.05). (Eye size: SEL1L+/+ 12.60 ± 0.50, SEL1LS780P/+ 12.34 ± 0.89, and SEL1LΔ806-809/+ 12.16 ± 0.65). (D) Heterozygous loss of NGLY1 in the Rh1G69D ER stress model results in no difference in eye size between the SEL1L+/+ and SEL1LS780P/+ genotypes. The SEL1LΔ806-809/+ genotype shows a significantly smaller eye size (p < 0.0001). (Eye size: SEL1L+/+ 10.58 ± 0.76, SEL1LS780P/+ 10.88 ± 0.61, and SEL1LΔ806-809/+ 8.67 ± 0.54).

We next tested whether ERAD is still improved by the SEL1L^S780P^ and SEL1L^Δ806-809^ alleles when NGLY1 activity is reduced. When we knockdown NGLY1 in the eye of this ER stress model, eye size is no longer increased in the SEL1L^S780P/+^ and SEL1L^Δ806-809/+^ genotypes compared to the SEL1L^+/+^ (Fig 4C and S5 Data). This result suggests that the improvement in ERAD associated with the SEL1L^S780P/+^ and SEL1L^Δ806-809/+^ genotypes is dependent on NGLY1 activity. Heterozygous NGLY1 null flies also showed no improvement in ERAD in the SEL1L^S780P/+^ and SEL1L^Δ806-809/+^ genotypes compared to the SEL1L^+/+^ genotype (Fig 4D and S5 Data). Interestingly, with the loss of NGLY1, the SEL1L^Δ806-809/+^ genotype had smaller eyes than the SEL1L^+/+^ genotype in both the NGLY1 knockdown (p < 0.05) and heterozygous NGLY1 null (p < 0.0001). Notably, our previous work has shown that NGLY1 knockdown alone does not alter eye size in this ER stress model [7]. We conclude that the SEL1L variants only improve resistance to ER stress when wildtype NGLY1 activity is present, suggesting that NGLY1 is required to mediate ERAD-dependent or related protective mechanisms. In the absence or reduction of NGLY1, the SEL1L variants do not increase ERAD in this model. While these variants improved ER stress responses in the eye model, their effect on whole-organism survival under NGLY1 deficiency may reflect additional mechanisms beyond ERAD, such as altered handling of specific substrates like NRF1.

Discussion

In this study, we sought to understand how natural SEL1L protein-coding variants identified in our Drosophila genetic screen influence NGLY1 deficiency lethality and other phenotypes. Our results demonstrate that these SEL1L variants increase survival in our NGLY1 deficiency models and protect against proteasome inhibition during both larval development and adulthood, although they do not mitigate larval size reductions induced by BTZ treatment. While these phenotypes are consistent with enhanced NRF1 activity, direct evidence for NRF1 involvement remains limited. Proteasome subunit gene expression measured from whole-body RNA suggests that the canonical NRF1-mediated bounce-back response is largely intact in heterozygous NGLY1 null animals and is not influenced by the SEL1L variants. However, proteasome gene expression changes in specific tissues and at different developmental windows were not evaluated. Additionally, the increased sensitivity to low-dose bortezomib in our SEL1L variants (Fig 2B) could reflect altered responses to proteotoxic or ER stress beyond NRF1 signaling. Alternative pathways such as the unfolded protein response (UPR) [34], integrated stress response (ISR) [35], or autophagy [36] may contribute to the observed phenotypes. Given the central role of SEL1L in ERAD, it is also possible that enhanced clearance of specific ERAD substrates indirectly affects stress sensitivity or protein homeostasis [10]. While NRF1 remains a compelling candidate, our data do not exclude the contribution of other stress response mechanisms. Further studies will be needed to define the specific downstream effects of SEL1L variants on NGLY deficiency.

The results from our ER stress model indicate that the ability of these SEL1L variants to confer protection against ER stress, may involve enhanced ERAD function and requires NGLY1 activity. In our ER stress eye model, SEL1L variant effects on eye size were only observed in the presence of wildtype NGLY1. When NGLY1 was reduced via knockdown or in a heterozygous NGLY1 null, the improvement in eye size associated with SEL1L^S780P^ and SEL1L^Δ806–809^ alleles was lost, suggesting that ERAD enhancement by these variants requires NGLY1 activity. While ERAD is a separate pathway from NRF1 signaling, the pathways are inextricably linked [17–19,21]. We propose that our SEL1L variants may enhance ERAD efficiency, thereby increasing the amount of NRF1 that is retrotranslocated to the cytosol and available for activation. In heterozygous NGLY1 null animals, where some NGLY1 activity remains, this may permit more efficient NRF1 activation and improved proteasome function. However, recent findings suggest that in the complete absence of NGLY1, NRF1 is not only inactive but may be diverted to a non-canonical ubiquitination pathway that exacerbates proteasome dysfunction [37,38]. Thus, the protective effects of the SEL1L variants we observe are likely contingent on residual NGLY1 activity and may not be generalizable to models of complete NGLY1 loss. While our data are consistent with enhanced NRF1 activation in the presence of the SEL1L variants, we did not directly measure NRF1 processing and nuclear localization. Therefore, the proposed involvement of NRF1 is a hypothesis based on prior literature, rather than a conclusion directly demonstrated in this study.

In addition to NRF1, recent work has shown that NGLY1 is required to deglycosylate misfolded BMP4, facilitating its removal from the ER and allowing properly folded BMP4 to dimerize and engage in canonical signaling [39]. BMP4 is required for proper developmental timing, so impairments in its signaling could affect larval size and growth [40]. Our previous studies have also demonstrated that proper folding and function of Ncc69/NKCC1 requires NGLY1 deglycosylation [9]. In these contexts, impaired clearance of misfolded NGLY1 clients could explain some of the phenotypes, and enhanced ERAD may restore proteostasis by clearing misfolded, misglycosylated proteins. Exploring these alternative mechanisms will be important for fully understanding the genetic interaction of SEL1L with NGLY1.

The SEL1L variants identified in our genetic screen are in the C-terminal cytoplasmic tail of the protein. While this region is not highly conserved across species, it is predicted, in all species, to be an intrinsically disordered domain [41]. Intrinsically disordered domains often mediate transient or multivalent protein–protein interactions and may contribute to dynamic complex formations [42–44]. Given SEL1L’s conserved role in ERAD, it is possible that these variants impact protein-protein interactions important for ERAD function and it will be important to determine whether analogous alterations in the cytoplasmic tail of mammalian SEL1L influence ER homeostasis or modulate phenotypes associated with NGLY1 deficiency. Functional studies in mammals could provide a deeper understanding of this poorly characterized domain of SEL1L and inform potential therapeutic targeting.

This study demonstrates that these naturally occurring SEL1L variants can enhance ERAD and mitigate specific phenotypes associated with NGLY1 deficiency. This raises the intriguing possibility that modulating SEL1L activity or ERAD efficiency could offer a therapeutic angle for this disease. However, further work is needed to determine whether these effects translate to mammalian systems and whether targeting ERAD components would improve proteostasis. It would also be interesting to test how modifications to mammalian SEL1L alter NRF1 processing or influence other stress response pathways. Ultimately, insights from cross-species modeling would help reveal whether modifying ERAD function can be leveraged to ameliorate disease phenotypes in NGLY1 deficiency.

Our findings implicate ERAD and proteasome function as potential therapeutic targets for NGLY1 deficiency patients. While there are still no therapeutic treatments for NGLY1 deficiency patients, there are two ongoing clinical trials, including an intracerebroventricular (ICV) NGLY1 gene replacement therapy [45] and a GlcNAc supplementation trial to treat alacrima [46]. Further study is needed to investigate how enhancing ERAD affects the proteasome bounce-back response in NGLY1 deficiency patients who already suffer from proteasome dysfunction [37]. Most current therapeutics that target ERAD aim to decrease its function [47–49]; however, enhancing ERAD is possible in cells and animal models through increased expression of ERAD complex proteins [49,50]. In addition to SEL1L, our previous modifier screen has also identified other genes involved in ERAD or ER function, including TMEM259, TMTC2, and ERMP1 [9]. Interestingly, for both TMTC2 and ERMP1, our GWAS independently hit two separate fly orthologs of these genes [9]. ERAD components are strong candidates for modifier genes of NGLY1 deficiency patients and may be good drug targets for further development.

Methods

Fly stocks and maintenance

Stocks were maintained on standard agar-dextrose-yeast medium and standard Archon Scientific glucose fly food at 25˚C on a 12-h light/dark cycle. SEL1L CRISPR strains were created by WellGenetics, Inc (www.wellgenetics.com) and alleles were verified by Sanger sequencing (S3 Fig). The following stocks were obtained from the Bloomington Drosophila Stock Center (Bloomington, IN): Tubulin-GAL4 (BDSC #5138), UAS-pngl-RNAi (BDSC #54853), attp2 (BDSC #36303), and attp40 (BDSC #36304). The Tubulin-GAL80 strain was provided by Dr. Carl Thummel (University of Utah). The SEL1L-RNAi strain (v1161) was obtained from the Vienna Drosophila Resource Center [51]. The NGLY1 null allele carries an early stop codon in the NGLY1 gene, and was previously characterized and generously provided by Perlara, PBC [24,25]. The NGLY1 null allele is homozygous lethal and the stock is maintained with the CyO balancer.

The “ER stress model” contains GMR-GAL4 and UAS-Rh1G69D on the second chromosome and has been previously described [31–33,52]. The endogenous SEL1L in this strain is homozygous for the SEL1L^S780P^ allele. We backcrossed the SEL1L^S780^ allele from our CRISPR generated SEL1L^S780^ strain into the ER stress model for 20 generations to create the desired homozygous SEL1L^S780^ genotype and verified genotype through sequencing.

Proteasome sensitivity larval size assay

Proteasome sensitivity larval size assays were performed as previously described [25]. Standard Drosophila food was melted and cooled to 60˚C prior to the addition of DMSO or bortezomib. NGLY1^+/-^ females were pre-mated with males from a CRISPR SEL1L variant strain for 24 hours. Mated females were placed on food containing DMSO and allowed to lay eggs for approximately 8 hours. Adult flies were removed and after four days of development, 3^rd^ instar larvae were transferred to vials containing food with DMSO or bortezomib. After two days on bortezomib containing food, larvae were genotyped using the presence of GFP. The CyO balancer also carries a GFP marker. Larvae were confirmed to lack the balancer chromosome by checking that they were also GFP negative. Larvae were imaged at 2.5X magnification using a Leica EC3 camera. Larval size was quantified using ImageJ as previously described [24,25].

Survival assays

NGLY1 knockdown (KD) eclosion survival: Virgin females from the CRISPR SEL1L variant strains were fed yeast overnight and then crossed with males from the donor strain UAS-NGLY1RNAi/Cyo,Tubulin-GAL80; Tubulin-GAL4/TM3,Sb. Progeny were collected and scored for the four balancer classes: CyO, Sb, double balanced, or no balancers, with the no balancer flies being the NGLY1 KD. This cross should produce the expected 1:1:1:1 ratio of the four genotypes. Given that there is always a very low level of lethality associated with each balancer, the largest balancer class was considered the closest to the expected number. We scored at least 200 flies per cross. Males and females were combined for a single count. To calculate the proportion of NGLY1 KD flies by generating a ratio of NGLY1 knockdown/largest balancer class as previously described [9]. For long-term survival, flies were collected, placed in a vial and flipped into fresh food every 2–3 days. Lifespan was measured as days post-eclosion. Vials were checked daily for dead flies and recorded.

Eye imaging and quantification

Adult female flies aged 3–5 days were collected under CO2 anesthesia then frozen at -80°C for later imaging. Eyes were imaged at 3x magnification using the Leica EC3 Camera. Eye area was measured as previously described [31–33,53].

Proteasome gene RT-qPCR

Changes in proteasome subunit gene expression were measured using RT-qPCR. Flies with appropriate SEL1L and NGLY1 genotypes were crossed as previously described. 24 hours after eclosion, male flies were collected and placed on either 0.2% DMSO or 3μM BTZ for 12 hours or 24 hours. Immediately after drug treatments, RNA was extracted from 8-10 whole-body flies using a Direct-zol RNA Miniprep (Zymo Research R2061) using TRIzol Reagent (ThermoFisher Cat # 15596026) and including the DNAse step. RNA was converted to cDNA using a ProtoScript II First Strand cDNA Synthesis Kit (NEB Cat # E6560L). RT-qPCR was performed using a QuantStudio 3 96-well 0.2 ml block instrument and PowerUp SYBR Green Master Mix (ThermoFisher Cat #A25741). If available, we used primers from the FlyPrimerBank [54] located at http://www.flyrnai.org/flyprimerbank. Other primers were designed using Primer3Plus [55] located at www.primer3plus.com. All primer sequences listed in S4 Data. When comparing proteasome gene expression between the NGLY1 wildtype and NGLY1 null flies, gene expression was calculated relative to the NGLY1 wildtype of each drug treatment and SEL1L genotype. In the comparison of proteasome gene expression changes with BTZ treatment, gene expression for all SEL1L genotypes was calculated relative to the DMSO treated SEL1L^+/+^ genotype.

Supporting information

S1 FigWildtype NGLY1 larvae are unaffected by proteasome inhibition.NGLY1 WT larvae treated with 1μM BTZ eclose as adults at similar rates regardless of SEL1L genotypes.(TIF)

S2 FigProteasome gene expression changes in response to BTZ.(A) Gene expressions are shown relative to NGLY1 wildtype for each drug treatment. No differences were observed in relative proteasome gene expressions between NGLY1 WT and NGLY1 + /- flies. Top: prosalpha6; bottom: prosbeta5. (B) There is an increase in proteasome gene expression with BTZ treatment compared DMSO treatment. All values are shown relative to the “SEL1L^+/+^ DMSO” proteasome gene expression levels. Top: prosalpha6; bottom: prosbeta5. (**) p < 0.01, (*) p < 0.05.(TIF)

S3 FigSequencing of CRISPR fly lines confirm appropriate SEL1L variants.Sanger sequencing of the three SEL1L CRISPR strains was used to verify the SEL1L alleles. (A) The top chromatograph is the SEL1L wildtype allele and the bottom is the SEL1L^S780P^ allele. The blue and red boxes show the base that is changed in the SEL1L^S780P^ variant allele. (B) The top chromatograph shows the unperturbed SEL1L gene. The box denotes the bases that are deleted, leading to the disruption of amino acids 806–809. The triangle on the bottom chromatograph shows where the deletion occurs in the SEL1L^Δ806-809^ allele.(TIF)

S1 DataNGLY1 ubiquitous knockdown fly counts and long-term survival.(XLSX)

S2 DataLarval size assay measurements.(XLSX)

S3 DataFly counts for eclosion on BTZ and survival data.(XLSX)

S4 DataqPCR results and primers used.(XLSX)

S5 DataFly eye measurements with ER stress model.(XLSX)