The Zn²⁺ transporter ZIP7 enhances endoplasmic-reticulum-associated protein degradation and prevents neurodegeneration in Drosophila

Summary:

Proteotoxic stress drives numerous degenerative diseases. Cells initially adapt to misfolded proteins by activating the unfolded protein response, including endoplasmic-reticulum-associated-protein-degradation (ERAD). However, persistent stress triggers apoptosis. Enhancing ERAD is a promising therapeutic approach for protein misfolding diseases. The ER-localized Zn²⁺ transporter ZIP7 is conserved from plants to humans and required for intestinal self-renewal, Notch signaling, cell motility and survival. However, a unifying mechanism underlying these diverse phenotypes was unknown. In studying Drosophila border cell migration, we discovered that ZIP7-mediated Zn²⁺ transport enhances the obligatory deubiquitination of proteins by the Rpn11 Zn²⁺ metalloproteinase in the proteasome lid. In human cells, ZIP7 and Zn²⁺ are limiting for deubiquitination. In a Drosophila model of neurodegeneration caused by misfolded rhodopsin (Rh1), ZIP7 overexpression degrades misfolded Rh1 and rescues photoreceptor viability and fly vision. Thus, ZIP7-mediated Zn²⁺ transport is a previously unknown, rate-limiting step for ERAD in vivo with therapeutic potential in protein misfolding diseases.

eTOC Blurb

The proteasome degrades misfolded proteins, and enhancing proteasome activity may slow age-related degenerative diseases. Guo and Mutch et al. report that ZIP7-mediated Zn2+ transport is limiting for toxic protein degradation in Drosophila and human cells, and ZIP7 overexpression rescues vision in a fly model of the blinding disease retinitis pigmentosa.

Graphical Abstract

Normal development and healthy aging require robust protein quality control. During normal cellular life, up to 30% of protein molecules misfold and are targeted for degradation by the proteasome. ^1,2 Some proteins are naturally more prone to misfolding than others and some cells experience greater protein folding challenges than others. Mutations can increase the susceptibility of proteins to misfolding, as can small protein aggregates that seed larger ones. Long-lived cells like stem cells and neurons, as well as secretory cells are especially dependent on proteasome activity for survival. Unfortunately, proteasome activity declines with age. ^3,4

Healthy cells degrade misfolded proteins. If the load of misfolded proteins increases, inducing endoplasmic reticulum (ER) stress, cells adapt by activating the unfolded protein response (UPR) ⁵, expanding the ER, enhancing ER-associated degradation (ERAD), reducing the rate of new protein synthesis, and altering transcriptional programs. If the stress resolves, the cell recovers. If the stress persists and overwhelms the adaptive response, cells die. There is great interest in identifying mechanisms that enhance ERAD in anticipation that such approaches will prevent or reverse degenerative diseases and promote healthy aging. ^6,7 Conversely, inhibitors of proteasomal degradation of misfolded proteins are in clinical use to treat cancers, especially malignancies of highly secretory cells, such as antibody-secreting B cells.

ZIP7 is an evolutionarily-conserved, ER Zn²⁺ transporter that promotes cell survival and migration in diverse cell types and organisms. ^8–11 ZIP7 promotes intestinal self-renewal, is required for B cell differentiation, and is overexpressed in multiple cancers. ^8,12 Loss of ZIP7 causes ER stress in a variety of cell types within organisms as diverse as plants and yeast to flies and humans. However, the mechanism by which ZIP7 mitigates ER stress and contributes to these diverse biological functions is unknown, ¹³ and the effects of ZIP7 overexpression are largely unexplored.

We previously identified the Drosophila ortholog of ZIP7 (dZIP7, aka Catsup) in a screen for mutations that disrupt border cell migration ¹⁴ and in a border cell gene expression profile. ¹⁵ Border cells in the Drosophila ovary provide an in vivo model of collective cell migration that is amenable to unbiased genetic screening. ¹⁶ Here we show that it is by mitigating ER stress that dZIP7 promotes border cell migration. Induction of ER stress by expressing a misfolded protein (Rh1^G69D) ^17,18 blocks border cell migration. Remarkably, overexpression of dZIP7 is sufficient to degrade misfolded Rh1^G69D, prevent ER stress, and thereby rescue border cell migration. We further show that ER to cytosol Zn²⁺ transport is rate-limiting for ERAD in Drosophila and human cells. dZIP7 overexpression in photoreceptor cells is sufficient to degrade Rh1^G69D in photoreceptor cells and prevent Rh1^G69D-induced retinal degeneration. These results illuminate a previously unappreciated rate-limiting requirement for Zn²⁺ in ERAD and suggest ZIP7 overexpression as a potential gene therapy for autosomal dominant retinitis pigmentosa and other degenerative diseases.

dZIP7 promotes border cell migration and prevents ER stress

Drosophila ovaries are composed of ovarioles, which are strings of egg chambers (Figure 1A) progressing through 14 stages of development, culminating with mature eggs. Each egg chamber is composed of 15 nurse cells and one oocyte (germ cells), surrounded by ~850 epithelial follicle cells. At stage 9 (Figure 1B), 4–8 border cells round up at the anterior end of the egg chamber, delaminate from the follicular epithelium, and migrate posteriorly, reaching the anterior border of the oocyte by stage 10. Border cell clusters are composed of 4–6 migratory cells that surround and carry two non-migratory polar cells. Expression of dZIP7 RNAi in the outer, migratory border cells using fruitlessGal4 19 inhibited migration (Figures 1C and 1G). The defect was rescued by co-expression of UAS-dZIP7::V5 (Figures 1D and 1G). Reduction of dZIP::GFP confirmed the effectiveness of the RNAi (Figures 1E, E’ and 1F, F’). Border cell migration was also impaired when dZIP7 RNAi was driven by the c306Gal4 (Figure 1G), which is expressed in both polar and migratory cells. FruitlessGal4-driven RNAi impaired border cell migration at least as much as c306Gal4, indicating that dZIP7 was primarily required in the outer, migratory cells (Figure 1G). This was further supported by mosaic clone analysis, in which the severity of border cell migration defects were proportional to the number of outer, migratory cells that were homozygous mutant (Figure S1A–C) and mutant cells were mostly excluded from leading positions (Figures S1D–E).

Figure 1: dZIP7 knockdown causes border cell migration defects and ER stress.

(A, B) Developing Drosophila egg chambers expressing dZIP7::GFP. DNA is in blue. F-actin is in magenta. Border cells migrate during stage 9. (B) and complete migration by stage 10 (A). (C, D) Stage 10 egg chambers with fruitlessGal4 driving expression of UAS-dZiP7RNAi and (C) UAS-GFPnls or (D) UAS-dZIP7::V5 in outer, migratory border cells. (E-F’) dZIP7::GFP expression (black/gray) in control border cells (E, E’) or c306Gal4>dZIP7RNAi (F, F’). E’ and F’ are the same clusters as E and F but shown in inverted grayscale. (G) Quantification of stage 10 migration defects in c306Gal4 (magenta) and fruitlessGal4 (purple) driving the indicated transgenes. Each dot represents the average of >24 egg chambers (n=3 independent experiments). Error bars=SEM. (H-H”) A mosaic border cell cluster composed of some control cells (RFP+, which can be dZIP7+/+ or dZIP7−/+, magenta, H’ ) and one homozygous dZIP7 mutant cell (RFP−, outlined). Polar cells (p) express higher levels of RFP compared to outer border cells. Xbp1::EGFP (green, H”) is a marker for ER stress. DNA is in blue. (I-I”) Anti-PDI antibody staining (magenta) reveals that clones of cells expressing dZIP7RNAi and GFPnls exhibit ER expansion. (J-J”) Mosaic clones expressing dZIP7RNAi and dZIP7::V5 and RFP (magenta) show similar PDI staining as neighboring wild type cells. Scale bars=20 μm.

dZIP7 is the ortholog of ZIP7, a Zn²⁺ transporter that moves Zn²⁺ from the ER to the cytoplasm and suppresses ER stress in many cell types including mammalian intestinal stem cells ^8,20, cancer cells ²⁰, Drosophila imaginal discs ⁹ and even in plants ¹³. We confirmed that dZIP7 localizes to the ER in border cells where it co-localized with the ER chaperone PDI much more significantly than with F-actin or the nucleus (Figure S2). Additionally, homozygous mutant dZIP7 border cells expressed the ER stress reporter Xbp1s::EGFP ²¹ whereas dZIP7^+/− and dZIP7^+/+ cells did not (Figures 1H–H”). We also observed increased expression of the ER chaperone PDI in homozygous dZIP7 RNAi-expressing follicle cell clones (Figures 1I–I”), a phenotype that we rescued with a wild type, V5-tagged dZIP7 transgene (Figures 1J–J”). Accumulation of XBP1 and PDI are indicative of induction of an adaptive unfolded protein response (UPR). ²² We conclude that cells lacking dZIP7 experience ER stress and impaired migration, raising the question as to whether ER stress, per se, inhibits motility.

dZIP7 is limiting for ERAD

To address whether ER stress inhibits migration, we expressed a misfolded rhodopsin protein, Rh1^G69D, known to induce ER stress. ^23,24 Misfolded rhodopsin accumulates in the ER and causes retinal degeneration in flies and humans. ^24–27 We found that Rh1^G69D accumulated to high levels intracellularly in border cells, induced ER stress, and blocked their migration (Figures 2A–2E). Similarly, RNAi against the protein chaperone HSC70–3, which is known to cause ER stress, inhibited border cell migration in 92% of egg chambers (n=3 independent experiments). Thus, ER stress caused by unrelated mechanisms - either loss of dZIP7, expression of a misfolded protein, or inhibition of a chaperone – correlates with impaired motility.

Figure 2: dZIP7 enhances proteasomal degradation of Rh1^G69D.

(A) Stage 10 egg chamber expressing Rh1^G69D (magenta) and the ER stress sensor Xbp1::EGFP (green) in border cells using c306Gal4. Rh1^G69D accumulation induced ER stress and blocked migration. (B-D) Co-expressing dZip7 rescued migration (B-C) and reduced Xbp1::EGFP (B-D) and Rh1^G69D (B-E). (C) Each dot represents one experiment (n>17 clusters). Error bars=SEM. (D) Each dot represents an individual border cell. Error bars=95% confidence intervals. (E) Each dot represents one experiment (n>17 clusters). Error bars=95% confidence intervals. (F-H) Stage 10 egg chambers expressing misfolded Rh1^G69D in border cells (insets) treated with 10μM of the MG132 proteasome inhibitor for 5 hours and stained with an antibody against Rh1. dZip7 co-expression with Rh1^G69D did not reduce Rh1 protein levels in the presence of MG132 showing that the dZIP7-mediated degradation of Rh1^G69D is mediated by the proteasome and that dZIP7 functions upstream of the 20S core enzyme blocked by MG132. (I) Knocking down the ERAD component Ufd1 in border cells causes migration defects. Each dot represents one experiment (n>30 clusters). Error bars=95% confidence intervals. *P≤0.05, ** P≤0.01, *** P≤0.001. Scale bars=20 μm.

Since loss of dZIP7 caused ER stress, we wondered if dZIP7 overexpression might suppress ER stress, so we co-expressed Rh1^G69D and dZIP7::V5 in border cells. Interestingly, dZIP7::V5 expression restored normal border cell migration (Figures 2B and 2C), reduced ER stress (Figures 2B and 2D) and virtually eliminated Rh1^G69D protein (Figures 2B and 2E). Together, these results suggest that dZIP7 is a limiting factor for degrading misfolded Rh1, which is known to be degraded by ERAD ^23–25, though a role for ZIP7 in ERAD has not previously been described.

If ZIP7 enhances ERAD, the proteasome inhibitor MG132 should block the ability of dZIP7 overexpression to promote Rh1^G69D degradation. In the absence of MG132, ZIP7 overexpression virtually eliminated Rh1^G69D protein (Figures 2B and 2E). In contrast, MG132 largely prevented the ability of ZIP7 to promote Rh1^G69D degradation (Figures 2F–2H). MG132 also prevented dZIP7-mediated rescue of the Rh1^G69D border cell migration defect (Figure 2G compared to Figure 2B). If ZIP7 normally promotes migration by promoting ERAD, then inhibiting ERAD should block migration. So we inhibited expression of the ERAD component known as Ubiquitin fusion-degradation 1-like (Ufd1), which is known to be required for degradation of Rh1^G69D. Ufd RNAi also inhibited border cell migration (Figure 2I). We conclude that ZIP7 likely promotes ERAD.

Misfolded rhodopsin is ubiquitinated by ER membrane-localized ubiquitin ligases (HRD1 and SORDD1/2), extracted from the ER membrane via the p97 chaperone TER94, and degraded by the proteasome (Figure 3A). ^35,37 Our data described above implicate ZIP7 and cytosolic Zn²⁺ as limiting for this process. To investigate which step of the ERAD process requires dZIP7, we used an antibody that recognizes ubiquitinated proteins to compare wild type, dZIP7-overexpressing, and dZIP7 RNAi cells. One hypothesis was that dZIP7 might provide Zn²⁺ to ubiquitin ligases such as Hrd1 and SORDD1/2, which are Zn²⁺-binding proteins that reside in the ER membrane (Figure 3A). If this were true, we might expect dZIP7 overexpression to increase - and dZIP7 RNAi to decrease - the abundance of polyubiquitinated proteins (PUBs). However, we observed the opposite effect. dZIP7 RNAi increased polyubiquitinated protein abundance (Figures 3B, 3B’) compared to the lacZ control (Figures 3C, #C’ and S3A–S3B‘). Conversely, dZIP7 overexpression essentially eliminated detectable polyubiquitinated proteins (Figures 3D and 3E). We also found that Rh1^G69D expression increased the accumulation of PUBs compared to the control (Figures S3A–B‘ and D), and ZIP7 overexpression suppressed the effect (Figures S3C, S3C‘, and S3D). Since PUBs accumulated in the absence of dZIP7, we conclude that ZIP7 functions downstream of ubiquitination to enhance ERAD.

Figure 3: dZIP7 enhances ERAD and promotes Rpn11-mediated deubiquitination of proteins required for proteasome entry.

(A) Schematic of proteasomal processing of misfolded proteins. (B-D’) Representative images of stage 8 egg chambers stained with an antibody against ubiquitinated proteins (PUB, green in B-D, FIRE LUT in B’-D’). (E) Quantification of fluorescence intensity of PUB staining in control (lacZ), dZIP7 overexpression and dZIP7RNAi expressing border cells. (F-G’) dZip7 overexpression reduced ubiquitinated protein levels in egg chambers treated with MG132 (10μM, 5 hours). Relative fluorescence intensity is quantified in (H). (I-J’) dZip7 overexpression does not prevent ubiquitinated protein buildup in egg chambers treated with Rpn11 inhibitor Capzimin (Czm, 20μM, 5 hours); this effect is quantified in (K). Dots represent individual border cell clusters. Error bars=95% confidence intervals. *** P≤0.001, **** P≤0.0001. Scale bars=20 μm.

dZIP7 reduced the abundance of polyubiquitinated proteins suggesting that it might be required instead for deubiquitination. There are many deubiquitinating enzymes (DUBs), but Rpn11 stood out as a top candidate. Whereas deubiquitination by some DUBs can rescue proteins from degradation, deubiquitination by Rpn11 is an essential prerequisite for entry of client proteins into the proteasome core and thus, like dZIP7, is essential for misfolded protein degradation (Figure 3A). ²⁸ Furthermore, unlike most DUBs, Rpn11 requires Zn²⁺ for catalysis ^29–31. In fact, Rpn11 is the only Zn²⁺-requiring DUB implicated in ERAD. ³² In vitro, when ubiquitinated substrates and the 26S proteasome assemble into complexes, chelating Zn²⁺ inactivates Rpn11. Addition of Zn²⁺ is sufficient to reactivate Rpn11 and stimulate deubiquitination ²⁹. Since dZIP7 transports Zn²⁺ to the cytosol ²⁰ and promotes loss of ubiquitinated proteins and degradation of Rh1^G69D, we hypothesized that dZIP7 might be limiting for Rpn11 activity.

In vitro, when Rpn11 is activated by addition of Zn²⁺ but the 20S core proteases are blocked, deubiquitinated substrates stall and accumulate. ²⁹ So we tested the effect of ZIP7 overexpression in the presence of MG132. We observed that ZIP7 still stimulated deubiquitination (Figures 3F–3H). Together with the observation that ZIP7 overexpression failed to degrade Rh1^G69D in the presence of MG132 (Figures 2F–2H), this result supports the idea that ZIP7 enhances deubiquitination of misfolded proteins upstream of 20S proteasomal degradation, possibly by providing Zn²⁺ to Rpn11.

If ZIP7 promotes Rpn11-mediated deubiquitination, an Rpn11 inhibitor should block the ability of dZIP7 to enhance deubiquitination of misfolded proteins. To test this prediction, we used the potent and selective Rpn11 inhibitor capzimin, which blocks the Zn²⁺ within the Rpn11 catalytic site. ³³ Capzimin largely prevented the effects of ZIP7 on deubiquitination (Figures 3I–3K’). Rpn11 RNAi also caused a severe border cell migration defect (Figure S3E). We conclude that dZIP7 enhances the obligatory Zn²⁺- and Rpn11-dependent deubiquitination of misfolded proteins prior to proteasomal degradation.

ZIP7-mediated Zn²⁺ transport is limiting for ERAD in Drosophila

ZIP7 resides in the ER membrane and transports Zn²⁺ from the ER to the cytosol ^20,34. To test whether the Zn²⁺ transporter activity of dZIP7 is important for border cell migration, we introduced point mutations, H315A and H344A, which replace histidine residues that are required for Zn²⁺ transport ³⁵ (Figure 4A, purple) and are conserved between ZIP7 and a more distant family member from Arabidopsis IRT1 (Figure S4A). As controls, we engineered dZIP7^H187A and dZIP7^H183A mutants (Figure 4A, green), which are not located in the ZIP7 core transmembrane domains involved in Zn²⁺ transport and are predicted not to affect transport. ³⁵ We generated transgenic flies expressing the mutants under Gal4/UAS control and included a V5 tag so that we could monitor protein abundance and localization. We then co-expressed each of these RNAi-resistant transgenes with dZIP7 RNAi and evaluated protein expression and border cell migration. All the proteins were stably expressed and correctly localized to the ER (Figures 4B and 4C and S4H and S4I). The point mutations predicted not to disrupt Zn²⁺ transport, dZIP7^H187A and dZIP7^H183A, rescued border cell migration to nearly wild type levels whereas neither dZIP7^H344A nor dZIP7^H315A provided significant rescue (Figures 4B and 4C and S4H and S4I), as quantified in Figure 4D.

Figure 4: Cytosolic Zn²⁺ is limiting for ERAD.

(A) Schematic of transmembrane domains and topology of dZIP7. Point mutations H183A and H187A res within the second transmembrane domain while H315A and H344A are within the highly conserved HELP domain and CHEXPHEXGD motif on the fourth and fifth transmembrane domains required for Zn2+ transport. (B-C) Co-localization of V5-tagged, RNAi-resistant dZIP7 mutants with the ER marker PDI (green) in border cells. (D) Quantification of incomplete migration at stage 10 in egg chambers expressing dZIP7RNAi with the indicated mutant forms. N=3 independent experiments. Error bars=SEM. (E-H’) Mosaic expression of dZIP7RNAi together with the indicated mutant forms of dZIP7 marked by RFPnls (magenta) and stained for Notch or EGFR (green). Scale bars=20 μm. Quantification of the fold change of Notch (I) and EGFR (J) expression in dZIP7 mutants compared to control cells.(K) Representative western blot on HeLa cell protein extract probed for polyubiquitinated protein and GAPDH. Cells were treated with/without proteasome inhibitor MG132 and zinc pyrithione (ZnPyr), a zinc ionophore. Treatments from left to right: DMSO, 500nM MG132, 500nM MG132 + 1μM ZnPyr, 500nM ZnPyr. (L) Adding ZnPyr to MG132-treated cells reduces polyubiquitinated protein levels. N= 4 experiments. (M) Western blot on HeLa cells with DOX-inducible ZIP7 that were treated with/without doxycycline and with/without MG132 and probed for polyubiquitinated protein and GAPDH. (N). ZIP7 overexpression reduces polyubiquitinated protein in HeLa cells treated with MG132. N= 4 experiments. Error bars= standard deviation. *P≤0.05, **P≤0.01, ****P≤0.0001.

In ZIP7 knockdown cells, membrane proteins, such as Notch and EGFR, accumulate abnormally in the ER ⁹. We also observed this in clones of border cells expressing dZIP7 RNAi (Figures S5A and S5B‘). In Drosophila follicle cells, as in imaginal disc cells, the effect is specific to some but not all membrane proteins. For example, E-cadherin does not accumulate abnormally in dZIP7 RNAi-expressing cells (Figures S5C and S5C‘ and ⁹).

To test whether Zn²⁺ transport is essential to prevent this accumulation, we assessed Notch and EGFR staining in clones of ZIP7 RNAi cells co-expressing the mutant forms of ZIP7, compared to neighboring wild type cells within the same cluster as an internal control (Figures 4E–4J and S4J–S4M‘). the Zn²⁺-transport-proficient proteins (dZIP7^H183A and dZIP7^H187A) exhibited similar levels of Notch and EGFR as wildtype cells in the same cluster (Figures 4E–F’, 4I and 4J, and S4J–S4K‘). By contrast, the Zn²⁺-transport-deficient proteins (dZIP7^H344A and dZIP7^H315A) failed to rescue and accumulated more Notch and EGFR than wild type cells of the same cluster (Figures 4G–J and S4L–M‘). From these experiments, we conclude that Zn²⁺ transport is an essential function of dZIP7 in promoting ERAD.

Zn²⁺ and ZIP7 are limiting for deubiquitination of proteasome client proteins in human cells

ZIP7 is nearly ubiquitously expressed in cells from organisms as diverse as plants and animals ¹³. Our results, combined with the observations that the proteasome is highly abundant ²⁸ whereas cytosolic [Zn²⁺]_free is extremely low (~100pM - 1 nM) ²⁰, suggested that Zn²⁺ might be rate limiting for proteasome activity, particularly when cells are stressed. So, we tested the ability of a Zn²⁺ ionophore, pyrithione (an organic salt of zinc capable of permeating cell membranes) ²⁰ to enhance deubiquitination of proteins in human cells.

This ionophore has previously been shown to rescue ER stress in ZIP7-deficient HeLa cells ²⁰, indicating that an important function of ZIP7 is to increase cytosolic Zn²⁺ (rather than decreasing it in the ER). So, we incubated HeLa cells with MG132 to induce ER stress and increase the abundance of polyubiquitinated proteins (Figures 4K and 4L). We then tested the effect of the Zn²⁺ ionophore. The added Zn²⁺ reduced the accumulation of polyubiquitinated proteins in the MG132-treated cells (Figures 4K and 4L). We conclude that cytosolic Zn²⁺ is limiting for proteasomal degradation of misfolded proteins in cells with ER stress.

To test whether overexpression of human ZIP7 could enhance protein deubiquitination in HeLa cells as it does in border cells, we generated HeLa cells expressing doxycycline-inducible ZIP7 (Figure S4N). Inhibition of the proteasome with MG132 increased the abundance of ubiquitinated proteins and dox-induced ZIP7 overexpression significantly reduced that effect (Figures 4M and 4N). We conclude that in Drosophila and human cells, ZIP7 provides limiting Zn²⁺ for deubiquitination of proteasome client proteins.

ZIP7 is required to degrade misfolded Notch and EGFR whereas ER stress independently inhibits Notch transcriptional responses

Notch and EGFR accumulate abnormally in the ER in ZIP7-deficient fly wing disc cells ⁹ and human cancer cells. ³⁶ We also observed abnormal accumulation of Notch (Figures 5A, 5A’ and S5A, A‘) and EGFR (Figures 5B, B’ and S5B, B‘) but not E-cadherin (Figures S5C and S5C‘) in dZIP7 knockdown follicle cell clones, supporting the generality of the phenomenon. Accumulation of Notch and EGFR following ZIP7 knockdown or inhibition in flies and mammalian cells has previously been attributed to a defect in protein trafficking ^9,28. However, our results implicate defective ERAD as the likely cause of abnormal intracellular Notch and EGFR in ZIP7 k.d. cells.

Figure 5: dZIP7 knockdown results in Notch & EGFR accumulation and reduced Notch transcriptional activity.

(A-B’) Intracellular Notch (A, A’) and EGFR (B, B’) accumulated in epithelial follicle cell clones expressing dZIP7RNAi (GFP+, green) relative to neighboring wild type cells. (C-D’) Mosaic clones of follicle cells expressing Rh1^G69D and GFP. Rh1^G69D expression does not cause accumulation of Notch (C, C’) or EGFR (D, D’) relative to wild type cells. (E-H) Notch transcriptional activity visualized with a Notch responsive element reporter (white). (E-F) Notch in dZIP7RNAi expressing cells (GFP+) compared to neighboring wild type (GFP−) cells. (G-H) Notch in Rh1^G69D- expressing cells (GFP+) compared to wild type cells (GFP−). **** P≤0.0001. Scale bars=20 μm.

Interestingly, neither Notch (Figures 5C and 5C’) nor EGFR (Figures 5D and 5D’) accumulated abnormally in Rh1^G69D-expressing follicle cells, including border cells. Yet, both ZIP7 knockdown cells (Figures 5E and 5 E’) and Rh1^G69D-expressing cells (Figures 5G and 5G’) exhibited reduced Notch transcriptional responses (Figures 5F and 5H). We conclude that dZIP7 knockdown causes two independent effects on Notch: accumulation of misfolded protein in the ER due to reduced ERAD and inhibition of Notch transcriptional activity, presumably as a consequence of the ER stress response. ^37–39

dZIP7 overexpression prevents retinal degeneration caused by Rh1^G69D

The observations that dZIP7 overexpression is sufficient to degrade Rh1^G69D, reduce ER stress, and rescue border cell migration and the ubiquity of ZIP7 and proteasomes suggested that dZIP7 overexpression might also be effective at suppressing retinal degeneration due to folding-defective rhodopsin. To test this hypothesis, we co-expressed UAS-dZIP7::V5 with UAS-Rh1^G69D in fly photoreceptor cells using GMR-Gal4. Eye morphology was normal in flies expressing dZIP7 alone (Figures 6A, 6B and 6G) whereas Rh1^G69D causes severe disruption of eye morphology compared to controls ^{17,23–25,27} (Figures 6C, 6D and 6G). Co-expression of Rh1^G69D and dZIP7 fully rescued eye morphology in the majority of flies examined (Figures 6E, 6F and 6G).

Figure 6: dZIP7 overexpression prevents Rh1^G69D retinal degeneration.

(A, C, E) Representative light photomicrographs of retinal morphology. (B, D, F) DIC images of retinal imprint morphology. (G) Quantification of number of flies with rough eye. dZIP7 overexpression allows ~70% of flies to develop with normal eyes. Each dot represents an average of >10 flies observed in each experiment. (H-I’) Larval eye discs expressing Rh1^G69D stained with antibodies against photoreceptors (Elav, magenta) and Rh1 (green). (J) dZIP7 expression significantly reduces Rh1^G69D protein accumulation. (K-N) Representative ERG recordings of one-week-old flies. (O) Quantification of ERG recordings. dZIP7 co-expression with Rh1^G69D returns ERG amplitude to control levels. Error bars represent 95% confidence intervals. *P≤0.05, ** P≤0.01, **** P≤0.0001.

To determine whether the mechanism of dZIP7 overexpression was the same in photoreceptor cells as border cells, we stained eye discs from third instar larvae with an antibody that labels all neuronal nuclei (Elav) and an antibody against Rh1. Rhodopsin is normally not expressed in larval photoreceptor cells so any Rh1 detected Rh1^G69D expressed from GMR-Gal4. Rh1 was abundantly expressed in the cytoplasm of photoreceptor cells in GMR-Gal4;UAS-Rh1G^69D;UAS-lacZ discs (Figures 6H and 6H’), whereas it was nearly undetectable in the majority of GMR-Gal4;UAS-Rh1^G69D;UAS-dZIP7 discs (Figures 6I–6J). We conclude that dZIP7 overexpression promotes the degradation of Rh1^G69D in photoreceptors as it does in border cells, demonstrating the generality of the effect in disparate cell types.

To test whether dZIP7 overexpression could restore visual function, we carried out electroretinogram (ERG) recordings, which measures the summed responses of all retinal cells to light. Control flies display a corneal negative receptor potential upon turning on a light stimulus, which quickly decays to baseline upon termination of the light (Figure 6K). The large maintained component of the ERG results principally from activation of the phototransduction cascade. The on- and off-transient responses, which are nearly coincident with the initiation and cessation of the light stimulus (Figure 6K), depends on synaptic transmission from the photoreceptor cells to postsynaptic cells in the optic lobes. Expression of Rh1^G69D in photoreceptor cells greatly diminished the amplitude of the ERG, and eliminated the on- and off-transients (Figures 6L and S6). Of note, overexpression of dZIP7 in Rh1^G69D photoreceptor cells restored a normal ERG, including a full receptor potential and synaptic transmission, as evidenced by the on- and off-transients (Figures 6M and S6). As a control, we found that expression of dZIP7 alone in photoreceptor cells had no measurable adverse effects on the ERG (Figure 6N), as quantified in Figures 6O and S6. These data demonstrate that over-expression of dZIP7 prevents the deleterious impact of Rh1^G69D on the response of retinal cells to light without side effects.

Discussion

A model for dZIP7 function: Zn²⁺ transport from the ER to the cytosol is limiting for ERAD and mitigation of ER stress

dZIP7 is a conserved protein that goes by names including ZRT1 in yeast, IRT1 in plants, dZIP7 or Catsup in Drosophila, and SLC39a7/Zip7/Ke4 in mammals. While many studies come to a common conclusion that loss or inhibition of ZIP7 disrupts ER homeostasis in cells from plants to flies and humans ^{8,13,20,36,40,41}, the mechanism has been unclear. ¹³ Moreover it has been unclear whether the disparate phenotypes caused by ZIP7 knockdown, which include defects in intestinal self-renewal, B cell differentiation, cell motility and survival amongst others, have a common underlying cause or represent pleiotropic activities of the ZIP7 protein. The data presented here provide evidence for an unanticipated and possibly unifying mechanism. Our data support a model in which dZIP7 promotes ERAD and prevents ER stress by providing free Zn²⁺ to enhance the catalytic activity of the Rpn11 DUB in the proteasome lid (see graphical abstract). This role for dZIP7 is critical since free Zn²⁺ is present at exceedingly low intracellular levels ²⁰, and therefore could be limiting. In the absence of ZIP7, misfolded/unfolded proteins accumulate and cause ER stress. Although Rpn11 is a top candidate for ZIP7-provided Zn²⁺, our experiments cannot rule out that other Zn²⁺ binding proteins may contribute to the beneficial effects of ZIP7. For example, the chaperone p97 extracts misfolded and ubiquitinated proteins from the ER and transfers them to the 26S proteasome, and one subunit of p97, Npl4, is a Zn²⁺ finger protein. Our experiments do not rule out the possibility that ZIP7 promotes deubiquitination by enhancing p97, in addition to, or instead of Rpn11 activity.

It is striking that dZIP7 overexpression is sufficient to enhance proteasomal degradation of misfolded proteins, including Rh1^G69D, preventing the harmful effects of ER stress including blindness. In contrast to earlier work that suggested that ZIP7 primarily promotes trafficking of membrane proteins such as Notch and EGFR ^9,36, our results show that release of Zn²⁺ from the ER to the cytosol via ZIP7 is limiting for ERAD.

We favor the model that the step in ERAD that is most sensitive to Zn²⁺ is deubiquitination of client proteins by Rpn11 and/or p97-mediated transfer of misfolded proteins to the proteasome. ^29,30 Our in vivo genetic and pharmacological studies are concordant with in vitro biochemistry. Rpn11 requires Zn²⁺ to deubiquitinate client proteins. ^29–31,33 This is an essential step so that the client protein can enter into the 20S proteasome for degradation by trypsin, chymotrypsin, and caspase-like endoproteases. Rpn11 also enhances proteasomal degradation by allowing ubiquitin to be recycled. Worden et al ²⁹ were able to assemble a complex in vitro composed of a ubiquitinated substrate and the 26S proteasome, including Rpn11. In the presence of a Zn²⁺ chelator, the complex assembles but Rpn11 is catalytically inactive, so ubiquitinated substrates accumulate. Upon addition of Zn²⁺, Rpn11 deubiquitinates the client protein. In the presence of the 20S protease inhibitor epoxomicin, Rpn11 deubiquitinates the client but it is not degraded, so deubiquitinated protein accumulates.

We observe remarkably similar effects by manipulating dZIP7 in vivo as Worden et al observed by manipulating Zn²⁺ in vitro. 29 In the absence of dZIP7, ubiquitinated proteins accumulate, whereas upon overexpression of dZIP7 in the presence of MG132, deubiquitinated substrate proteins (e.g. Rh1^G69D) accumulate. It is reasonable to propose that cytosolic free Zn²⁺ could be rate-limiting because proteasomes are abundant, whereas cytosolic free Zn²⁺ is vanishingly rare at ~1 nM ²⁰, which is ~100-fold less than the typical free cytosolic [Ca²⁺]. We propose that ZIP7 provides rate-limiting Zn²⁺ to p97 and/or Rpn11, and thus that the level of ZIP7 determines a cell’s capacity to degrade misfolded proteins. In support of this idea and the generality of the mechanism proposed here, we found that increasing intracellular Zn²⁺ enhanced deubiquitination of proteins in a human cell line in the presence of MG132.

Why is ZIP7-mediated Zn²⁺ transport dispensable for ubiquitinating misfolded proteins even though the E3 ubiquitin ligases that catalyze that reaction are Zn²⁺-binding proteins? In contrast to Rpn11, which can fold and assemble into proteasome/substrate complexes even in the presence of a Zn²⁺ chelator, Zn²⁺ is a structural element of the RING finger domains of the ubiquitin ligases. We suspect that the ligases would fail to fold in its absence and therefore likely have a high affinity for Zn²⁺.

Despite the fact that Zn²⁺ is the second most abundant divalent cation in cells, free Zn²⁺ is exceptionally rare in the cytosol because nearly all Zn²⁺ is bound to proteins. While an essential trace element, excess cytosolic Zn²⁺ can be toxic ⁴², yet ZIP7 overexpression did not cause detectable harm either to follicle cells in the ovary or to photoreceptor cells in the eye. This suggests that the Zn²⁺ transported to the cytosol via ZIP7 might predominantly exert its effects locally near the ER. ZIP7 may not directly bind to the ERAD machinery though because it was not detected in an extensive proteomic analysis. ⁴³ The human genome encodes 24 Zn²⁺ transporters, 14 of which belong to the ZIP family which move Zn²⁺ into the cytoplasm from outside the cell or from inside an organelle while 10 are members of the ZnT family which transport Zn²⁺ out of the cell or into organelles. ^44,45 The large sizes of these families are consistent with the idea that local Zn²⁺ sources may be important for promoting necessary Zn²⁺-dependent processes without increasing global levels, which would be toxic.

Biomedical implications of the role of ZIP7 in ERAD

Our finding that dZIP7 overexpression alleviates ER stress and cellular death due to Rh1^G69D expression has some general biomedical implications. Dominant mutations in rhodopsin that impair folding and cause accumulation in the ER cause retinal degeneration in human patients ²⁶, for which there is no effective prevention or therapy. Over-expression of proteins that enhance ERAD is a promising therapeutic strategy. Additionally, toxic protein aggregates have been proposed to kill neurons by inhibiting ERAD in numerous neurodegenerative diseases including Huntington’s, Altzheimer’s, Parkinson’s, frontotemporal dementia, and others, even when the toxic protein is not localized in the ER. ^46,47 Thus, strategies to enhance ERAD may be useful in treating multiple degenerative diseases.

The suppression of ER stress and border cell migration by dZIP7 overexpression is consistent with the observation that ZIP7 is over-expressed in numerous cancers where it promotes survival, proliferation and migration and correlates with disease progression, invasion, and metastasis. ^10–12,48 A ZIP7 inhibitor was identified in a screen for drugs to treat Notch-dependent cancers, based on the model that ZIP7 is important for Notch trafficking ³⁶. We show that ER stress impairs Notch transcriptional activity independent of any trafficking defect because Rh1^G69D inhibits Notch signaling without abnormal Notch or EGFR protein accumulation. Precisely how ER stress or the UPR inhibits Notch signaling is not yet clear, but the observation that a pharmacological inhibitor of ZIP7 was identified as a suppressor of Notch signaling by the Notch intracellular domain (NICD) in cultured U2OS osteosarcoma cells ³⁶ suggests that there is a deeply conserved requirement for ZIP7 for Notch transcriptional activity. Nolin et al ³⁶ showed that ZIP7 inhibition causes accumulation of full length Notch and a decrease in the NICD, and concluded that Notch activation by proteolysis was likely perturbed upon inhibition of ZIP7. An alternative interpretation is that full-length Notch accumulates in the ER lumen due to inhibition of ERAD, and that the NICD activity is inhibited by the global ER stress response.

Our results suggest that ZIP7 inhibitors might be effective against cancers that rely especially heavily on proteasomes. Proteasome inhibitors such as bortezomib are approved for the treatment of B cell malignancies including multiple myeloma and mantle cell lymphoma ⁴⁹. Our results suggest that ZIP7 inhibitors might be repurposed to treat those cancers as well, especially considering that resistance typically develops against a single therapeutic agent. Moreover, by inhibiting ERAD specifically rather than all proteasomes generally, ZIP7 inhibition may be less toxic than bortezomib. Interestingly, hypomorphic mutations in ZIP7 cause a B cell deficiency due to defects in B cell differentiation in human patients ⁵⁰. Although the mechanism underlying this phenotype is unknown, our results implicate ZIP7 in the UPR, and mutations that compromise the UPR also cause B cell deficiency due to defective B cell differentiation. So, the results presented here suggest a possible link between these otherwise disparate observations. B cell development appears to depend upon a functional UPR and ER stress response, perhaps to ensure resilience to the natural ER stress B cells experience when they secrete large quantities of antibody.

Finally, the similarities in dZIP7 functions and phenotypes across disparate cells, tissues, and organisms suggests that the border cell system offers an excellent model for deciphering the fundamental and conserved effects of this protein in vivo.

Limitations of the Study

The data do not currently distinguish whether Zn²⁺ is limiting only for Rpn11 activity or also for Npl4, which is a zinc finger domain protein and subunit of the p97 chaperone complex. The data also do not currently exclude the possibility that ZIP7 promotes border cell migration through a mechanism other than promoting ERAD, such as transit through the ER or protein folding within the ER, though the data are consistent with the model that a primary and general function of ZIP7 is to stimulate ERAD by providing Zn²⁺ to ER-associated proteasomes. Further work will be required to determine if the ZIP7-mediated Zn²⁺ ions remain in the local vicinity of the ER, or diffuse to activate proteasomes throughout the cell. Whether ZIP7 overexpression can prevent retinitis pigmentosa in mammals including humans remains to be tested. Whether ZIP7 overexpression can prevent neurodegenerative diseases other than retinitis pigmentosa also remains to be investigated.

STAR METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Denise Montell (dmontell@ucsb.edu).

Materials availability

Drosophila lines and generated reagents are available upon request.

Data and code availability

• All data are available in the main text or the supplementary materials. Raw data will be available upon request from the lead contact.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this work paper is available from the Lead Contact upon request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Drosophila husbandry & genetics

Flies were raised in vials containing 5mL of standard cornmeal-yeast food. See recipe: https://bdsc.indiana.edu/information/recipes/molassesfood.html. Flies were kept in incubators at 18, 25, or 29°C with 12 hour light/dark cycles and 80% humidity. Female flies were used for experiments on egg chambers. For experiments on the eye, both female and male flies were used as sex did not appear to influence results.

The dZIP7 mutant was generated by ethyl methanesulfonate (EMS) mutagenesis ¹⁵. The mutation results in a glycine (G) to aspartic acid (D) substitution at amino acid 178. The FLP/FRT system was used to generate dZIP7^G178D homozygous mutant clones by combining FRT40A-dZIP7^G178D with hsFLP12,yw;ubi:GFPnls, FRT40A or hsFLP12,yw;ubi:RFPnls, FRT40A/(CyO). hsFLP;AyGal4,UAS-Redstinger.nls6 (RFPnls in figure legends) was given as a gift by Dr. Jocelyn McDonald and it was used to generate FLP out clones driving UAS-dZIP7RNAi;UAS-GFPnls (Fig.1 I–I”), UAS-dZIP7RNAi;UAS-dZIP7.V5 (Fig.1 J–J”), UAS-dZIP7RNAi;UAS-dZIP7H183A, UAS-dZIP7RNAi;UAS-dZIP7H187A (Fig.4 E–F, Sup Fig.4 J–K’), UAS-dZIP7RNAi;UAS-dZIP7H344A, UAS-dZIP7RNAi;UAS-dZIP7H315A (Fig.4 G–H, Sup Fig.4 L–M’). hsFLP;;Aygal4,UAS-mCD8GFP (lab stock) induced UAS-dZIP7RNAi and UAS-dZIP7RNAi;Notch-Responsive-Element-RFP FLP out clones (Fig.5 A–B’, E–E’, Sup Fig.5 A–B’). hsFLP;AyGal4,UAS-GFP (lab stock) induced Rh1^G69D and Rh1^G69D;Notch-Responsive-Element-RFP FLP out clones (Fig.5 C–D). Heat shocks were performed at 37°C for 1 hour and followed by 29°C incubation before dissection.

dZIP7::GFP expression pattern was visualized using the line from VDRC 318542 in the fTRG stocks library. The UAS-dZIP7RNAi transgenic line is from VDRC 100095 P{KK103630}VIE-260B. UAS-HSC70–3RNAi (VDRC 14882 w1118; P{GD6484}). The following stocks were from the Bloomington Drosophila Stock Center: dZIP7.V5 (# 63229, w[*]; sna[Sco]/CyO; P{w[+mC]=UAS-dZIP7.V5}6 ER stress marker UAS-Xbp1-EGFP.HG (#60731, w[*]; P{w[+mC]=UAS-Xbp1.EGFP.HG}3), UAS-Ufd1 RNAi (#41823 y[1] v[1]; P{y[+t7.7] v[+t1.8]=TRiP.GL01251}attP2), UAS-Rpn11 RNAi (#33662 y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP.HMS00071}attP2).

Additional transgenic Drosophila stocks used: UAS-white RNAi/Cyo is a lab stock ⁵¹, UAS-Rh1^G69D was a gift from Dr. Hyung Don Ryoo lab ²⁵, UAS-lacZ is a lab stock, C306-Gal4 is a lab stock, UAS-Notch.Intracellular.Domain on the third chromosome was a gift from Artavanis-Tsakonas lab ⁵².

Cell Culture Conditions & Maintenance

HeLa cells derived from a human (female) cervical cancer (ATCC^® CCL-2^™) were grown in high glucose Dulbecco’s Modified Eagle Medium (DMEM) containing GlutaMAX^™ and pyruvate (Gibco #10–569-044), plus 10% heat-inactivated fetal bovine serum (Sigma-Aldrich #F4135). For HeLa-ZIP7 cells, the media was supplemented with 1 μg/mL puromycin (Gibco 414 #A1113803). Media was sterile filtered using Nalgene^™ Rapid-Flow^™ Sterile Disposable Filter Units with PES, CN, SFCA or Nylon Membranes. Cells were maintained in incubators at 37°C with 5% CO2 and 90% humidity. Cells were not tested for mycoplasma contamination and were not authenticated.

METHOD DETAILS

Design of UAS-RNAi-resistant dZIP7 point mutations

dZIP7 point mutations were cloned into vector pUASt-attb with forward primer ctctgaatagggaattgggATGGCCAAACAAGTGGCTGA and reverse primer ccgcagatctgttaacgtcaCGTAGAATCGAGACCGAGGAGAG. The vector was injected into attp2 flies y1 w67c23; P{CaryP}attP2 by BestGene Inc.

When generating UAS-dZIP7-point-mutations, we designed the construct so it cannot be targeted by the dZIP7RNAi sequences by substituting redundant codons for the same amino acids within the region targeted by the RNAi. The RNAi resistant sequence is reported below. Nulceotide substitutions are shown in lower case: ACAcGGcCAttcCCAtGAcATGtcCATcGGctTGTGGGTgCTgGGcGGcATtATcGCgTTtCTgagcGTcGAaAAgtTGGTgCGtATcCTgAAaGGaGGcCAcGGcGGcCAtGGaCAttcCCAcGGcGCcCCcAAaCCcAAgCCcGTcCCcGCcAAaAAgAAaagCagcGAtAAgGAgGAttcCGGcGAcGGcGAtAAgCCcGCcAAaCCcGCgAAaATtAAaagCAAaAAgCCcGAgGCcGAaCCcGAgGGaGAgGTcGAaATcagCGGaTAtcTGAAccTGGCcGCcGAtTTcGCcCAtAAtTTtACgGAcGGatTGGCgATtGGaGCgagCTAccTGGCcGGaAAttcCATcGGaATcGTcACgACcATtACcATctTGtTGCAtGAgGTcCCgCAcGAaATcGGcGAtTTcGCgATcCTgATcAAaagcGGaTGcagCcGcCGcAAaGCcATGCTgtTGCAaCTgGTgACcGCcCTgGGcGCccTGGCcGGaACcGCcCTgGCcCTgtTGGGcGCcGGcGGaGGcGAtGGcagcGCgCCcTGGGTgcTGCCgTTtACcGCgGGaGGcTTcATcTAtATtGCcACcGTcAGcGTgtTGCCcGAatTGCTgGAaGAaagcACcAAgtTGAAgCAaagctTGAAaGAgATtTTcGCctTGCTgACCGGCGTAGCCCTAATGATCGTTATCGCCAAGTTCGAGGg.

Point mutations were designed by changing codons at the following site dZIP7H183A (CAC to GCC), dZIP7H187A (CAT to GCT), dZIP7H315A (CAT to GCT), dZIP7H344A (GCT to CAT).

Immunofluorescence staining of Drosophila tissues

For immunostaining of egg chambers, 10 female flies were fattened with yeast for 2–3 days at 29°C. Egg chambers were dissected from ovaries of female flies in Schneider’s medium with 10% Fetal Bovine Serum (FBS) (pH=6.85–6.95) as described previously ⁵³.

For immunostaining of larval eye imaginal discs, 10 3rd instar larva were dissected in Phosphate Buffered Saline (PBS).

Freshly dissected tissues were fixed in 4% paraformaldehyde in PBS for 15 minutes, then washed three times with PBS+0.4%TritonX-100 (PBST). They were then incubated overnight at 4°C in PBST with primary antibodies. Tissues were washed three times with PBST. Secondary antibodies were incubated for 1 hour at room temperature, together with Hoechst to stain for nuclear DNA, and phalloidin to stain for F-actin. After secondary antibody incubation, samples were washed three times with PBST and mounted in VECTASHIELD mounting medium from Vector Laboratories.

For treatment with proteasome inhibitors, egg chambers were dissected and incubated in Schneider’s medium with 10% FBS containing the inhibitor for 5 hours at room temperature. MG132 was purchased from Millipore Sigma (474791, CalBiochem) and Capzimin from Aobious (AOB8855). Post-treatment, samples were fixed and immunostained as described above.

Antibodies used: chicken anti-GFP (1:200) (ab13970 Abcam plc.), mouse anti-Notch intracellular domain (1:100) (C17.9C6 DSHB), rat E-cadherin antibody (1:50) (DCAD2 DSHB), V5 Tag Monoclonal Antibody-Alexa Fluor 555 (1:100) (2F11F7 Invitrogen), mouse anti-PDI (1:200) (ADI-SPA-891-D, Enzo Life Sciences), mouse anti-dEGFR (1:2000) (E2906 Sigma Aldrich), mouse anti-Rhodopsin 1 (1:100) (4C5 DSHB), mouse anti-Ubiquitinated proteins clone FK2 (1:400) (04–263 Sigma Aldrich). Secondary antibodies were all Invitrogen AlexaFluor conjugated goat antibodies. Mouse anti-PDI and mouse anti-V5–555 co-staining was done by first staining with PDI primary and secondary antibodies, followed by a washout, and application of anti-V5–555 overnight.

Doxycycline-inducible ZIP7 overexpression

ZIP7 was cloned from cDNAs of HeLa cell line into pcDNA3.1 V5-His B vector (Invitrogen V81020) by the ApaI restriction site using primers: forward (5’-CCGCTCGAGTCTAGAGGGCCATGGCCAGAGGCCTGGGG-3’) and reverse (5’-TTACCTTCGAACCGCGGGCCTCACTCAAGGTGGGCAATCAGCACCATC-3’). An HA-tag was added to Zip7 3’ by amplification with the following primers: forward (5’-cctccatagaagattctagagccaccATGGCCAGAGGCCTGGGG-3’) and reverse (5’-tatatagcggccgcttaAGCGTAATCTGGAACATCGTATGGGTACTCAAGGTGGGCAATC-3’). Zip7-HA was transiently cloned in AddGene plasmid 46970 using XbaI and NotI restriction sites (New England Biolabs, R0145 and R3189). Zip7-HA was then amplified using the following primers: forward (5’-tatatagctagcgccaccATGGCCAGAGGC-3’) and reverse (5’-atatataccggtAGCGTAATCTGGAACATCGtatggg-3’). mCherry (AddGene, 124428) was amplified using the following primers: forward (5’-tatataACGCGTatggtgagcaagggcgagg-3’) and reverse (5’-atatatGGATCCttacttgtacagctcgtccatgcc-3’). Zip7-HA and mCherry were cloned in a doxycycline-inducible lentiviral vector (AddGene, 71782) upstream and downstream of a P2A sequence, using NheI-AgeI and MluI-BamHI respectively (New England Biolabs, R3131, R3552, R3198, R3136). PCRs were carried out with Phusion^® High-Fidelity PCR kit (New England Biolabs). The transfer plasmid was delivered to HeLa cells (ATCC^® CCL-2^™) by lentiviral transduction. Packaging was achieved in Lenti-X 293T cells (Takara, 632180) by transfection with the transfer plasmid, pCMV-VSV-G (AddGene, 8454) and pCMV-dvpr-dR8.2 (AddGene, 8455), using X-tremeGENE^™ HP DNA Transfection Reagent (Sigma, 6366236001). The media was changed 24h after transfection and lentiviral particles were collected 24h after media change, then centrifuged 2min at 2000rpm (RT) and filtered (0.45 μm Acrodisc. Syringe Filters PALL 4654; VWR, 28143–352). Recipient cells were transduced by incubation and centrifugation with filtered lentiviral media in the presence of 2 μg/mL polybrene (Sigma, 1003) (30 min, 2000 rpm, RT). The resulting cell line was sorted twice for mCherry positivity on a Sony MA900 Multi-Application Cell Sorter with a 561nm excitation laser (Biological Nanostructures Laboratory, CNSI, UCSB). mCherry expression was induced by treatment with 1 μg/mL doxycycline added to culturing media for 24 hours prior to sorting. After the first round of sorting, the cell line was expanded before being treated and sorted again.

Western blotting

HeLa cells (ATCC^® CCL-2^™) were plated at a density of 200.000 cells/well in 6-well cell culture plates (Genesee Scientific, #25–105). The following day, cells were treated with MG-132 (Calbiochem, #474791–1MG), 1-Hydroxypyridine-2-thione zinc salt (Millipore Sigma, #H6377–10G) and/or DMSO (Fisher Scientific, #BP231–100) as indicated in the figure legend. The next day, cells were rinsed with PBS (Gibco^™, #20–012-050) and proteins were extracted in 200μl of Laemmli buffer (Sigma-Aldrich, #S3401–10VL). Samples were heated for 5 minutes at 95°C. Proteins were separated on 4–20% Mini-PROTEAN^® TGX^™ Precast Protein Gels (Bio-Rad, #4561096) in 10% Tris-Glycine (Bio-Rad, #1610771)/0.1% SDS (Bio-Rad, #1610418) running buffer. Molecular weights were defined with Precision Plus Protein^™ All Blue Prestained Protein Standards (Bio-Rad, #1610373). Transfer on nitrocellulose membrane (Bio-Rad, #1620115) was carried out at 40V for 3 h in aqueous buffer 10% Tris-Glycine (Bio-Rad, #1610771)/20% methanol (Fisher Scientific, #A412P-4). Transfer quality was assessed by Ponceau staining (Santa Cruz Biotechnology, #sc-301558). Membranes were blocked 30 min at RT in agitation in Intercept^® (TBS) Blocking Buffer (LI-COR Biosciences, #927–60003). Antibodies were diluted in 50% Intercept^® (TBS) Blocking Buffer/45% H₂0/5% TBS 10X (Bio-Rad, #1706435)/0.1% Tween-20 (Sigma-Aldrich, #P1379). Primary antibodies were incubated overnight at 4°C in agitation and washed 2×10 min in TBS 1X/0.1% Tween-20. Secondary antibodies were incubated 1 h at RT in agitation and washed 4×10 min in TBS 1X/0.1% Tween-20. Membranes were rinsed in TBS 1X and imaged using an Odyssey imaging system (LI-COR Biosciences). Antibodies and concentrations used: 1:750 anti-ubiquitinated proteins antibody, clone FK2 (Sigma-Aldrich, #04–263), 1:10.000 GAPDH antibody (Thermo Fisher Scientific, #PA5–85074), 1:500 anti-Actin (DSHB, #JLA20), 1:15.000 IRDye^® 680LT Donkey anti-Mouse IgG Secondary Antibody (LI-COR Biosciences #926–68022) and 1:15.000 IRDye^® 800CW Donkey anti-Rabbit IgG Secondary Antibody (LI-COR Biosciences #926–32213). Raw images were quantified in Fiji designing rectangular regions of interest (ROIs) of constant size for a given protein within a given blot. The ROI for anti-ubiquitinated proteins was designed to encompass the increase in signal observed in the positive control (MG-132). In each lane, the mean fluorescence intensity of the band of interest (anti-ubiquitinated proteins signal) was normalized over the loading control (GAPDH). Representative blots shown in Figure 4 after increasing image size with interpolation.

Electroretinogram (ERG) recordings

ERG recordings were performed as described by Wes et al. with slight modifications ⁵⁵. After dark adaptation for ≥90 sec, individual flies 3–7 days posteclosion were stimulated with a 5-second pulse of bright orange light (~30 mW/cm²) at a frequency of 3.4 pulses per minute. The light-induced retinal field potential signals were amplified using an IE-210 amplifier (Warner Instruments), digitalized with a PowerLab 4/30 (ADInstruments), and stored as raw data on a computer using LabChart 6 software (ADInstruments). For each genotype, 4–6 individual flies were recorded. The peak amplitudes of the on- and off-transients as well as the sustained responses were analyzed from the raw ERG data.

Sequence alignment

dZIP7 and ZIP7 amino acid sequences were acquired from NCBI in a FASTA format. The files were input into T-coffee http://tcoffee.crg.cat/apps/tcoffee/do:regular to generate multiple sequence alignment. The output was fed into Boxshade http://www.ch.embnet.org/software/BOX_form.html to generate the sequence alignment with black and gray shades to show the conserved sequence region. dZIP7 transmembrane domain annotation is predicted by Kambe et. al, 2015 ³⁵ and Alphafold Protein Structure Database ^{56 57}

Experimental design

All fly crosses were made at least 3 times. Dissection and immunostaining experiments were performed at least 3 times. Drosophila retinal morphology was examined 5 times. Independent ERG recordings were performed at least 4 times. For cell culture experiments, cells were plated, treated, and analyzed via western blotting 4 times. Sample size was determined by using prior knowledge of minimum sample size. Experiments were not randomized and investigators were not blinded. No data was excluded from the results shown.

QUANTIFICATIONS AND STATISTICAL ANALYSES

Image acquisition, analysis and quantification

Zeiss LSM780 and LSM800 confocal microscopes were used to acquire images. Images were processed using FIJI, rotated and cropped for presentation.

Incomplete migration is when border cell clusters failed to reach the border of oocyte by oogenesis stage 10.

For quantification of RH1 and ubiquitinated proteins: FIJI was used to define the region of interest (ROI) optimal for identifying Gal4 expression and quantifying mean fluorescence intensity of the target channel (either RH1 or ubiquitinated proteins). Image quantifications were done using stage 8, stage 9 and stage 10 egg chambers. Experiment quantifications were normalized to the average of the mean of the control genotype/treatment (C306>LacZ, media treated).

Notch and EGFR protein intensity fold change (Figure 4I, J) was done by measuring the Notch Intracellular Domain and EGFR antibody staining intensity in mutant and control cell cytoplasm at the same imaging plane in the same border cell cluster. Cytoplasmic regions were manually identified and circled by F-actin and nuclei staining. Fold changes of the mutant over control mean intensity were calculated and plotted. More than 5 pairs of cells are quantified across more than 3 independent experiments.

Notch Responsive Element quantification (Figure 5) is done by measuring the Notch responsive element (NRE) fluorescent reporter intensity in the mutant cells and the control cells in the same plane of image. Cell boundaries are manually circled according to F-actin, and the NRE mean intensity in arbitrary units are compared in the mutant and control cell pairs. More than 10 pairs of cells are quantified across more than 4 independent experiments.

Statistical analysis

Statistics were done in Graphpad Prism. Significance was determined through either a Student’s T-Test, Ordinary one-way Anova, or Ordinary two-way Anova with Tukey’s test for multiple comparisons. Error bars are reported in figure legends.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Chicken anti-GFP (1:200)	Abcam	ab13970
Mouse anti-Notch (intracellular domain) (1:100)	DSHB	C17.9C6
Rat anti-E-cadherin (1:50)	DSHB	DCAD2
V5 Tag Monoclonal Antibody Alexa Fluor 555 (1:100)	Invitrogen	2F11F7
Mouse anti-PDI (1:200)	Enzo Life Sciences	ADI-SPA-891-D
Mouse anti-EGFR (1:2000)	Sigma Aldrich	E2906
Mouse anti-Rhodopsin1 (1:100)	DSHB	4C5
mouse anti-Ubiquitinylated proteins clone FK2 (1:400 IF, 1:750 WB)	Sigma Aldrich	04–263
Rabbit anti-GAPDH (1:10000)	ThermoFisher	PA5–85074
Rabbit anti-ZIP7 (1:1000)	Proteintech	19429–1-AP
Mouse anti-Actin (1:500)	DSHB	JLA20
IRDye® 680LT Donkey anti-Mouse IgG Secondary Antibody	LI-COR Biosciences	926–68022
IRDye® 800CW Donkey anti-Rabbit IgG Secondary Antibody	LI-COR Biosciences	926–32213
Rat anti-Elav (1:500)	DSHB	9F8A9
Bacterial and virus strains
Doxycycline-inducible lentiviral vector	AddGene	71782
NEB® 5-alpha Competent E. coli	New England BioLabs	C2987H
Chemicals, peptides, and recombinant proteins
Schneider’s Drosophila Medium	Gibco	21720
Hoechst 33342	Sigma Aldrich	14533
Phalloidin Atto647N	Sigma Aldrich	65906
Triton X-100	Sigma Aldrich	T8787
Paraformaldehyde, 16%	Electron Microscopy Sciences	15710
Phosphate Buffered Saline (10X, pH 7.4)	Quality Biological	119–069–131
Vectasheild anti-fade mounting medium	Vector Laboratories	H-1000
MG-132	Calbiochem	474791–1MG
Capzimin	Aobious	AOB8855
1-Hydroxypyridine-2-thione zinc salt	Millipore Sigma	H6377–10G
DMSO	Fisher Scientific	BP231–100
Laemmli buffer	Sigma-Aldrich	S3401–10VL
Tris-Glycine	Bio-Rad	1610771
SDS	Bio-Rad	1610418
Intercept® (TBS) Blocking Buffer	LI-COR Biosciences	927–60003
Tween-20	Sigma-Aldrich	P1379
Experimental models: Cell lines
HeLa cells	ATCC	CCL-2
Lenti-X 293T cells	Takara	632180
HeLa with DOX-induced ZIP7-HA-P2A-mCherry	This paper	n/a
Experimental models: Organisms/strains
c306Gal4;UAS-LifeActGFP;Gal80ts	Denise Montell Lab Stock, University of California Santa Barbara	n/a
UAS-dZIP7RNAi P{KK103630}VIE-260B	VDRC	100095
UAS-dZIP7 w[*]; sna[Sco]/CyO; P{w[+mC]=UAS-dZIP7.V5}6	BDSC	63229
UAS-Xbp1 -EGFP.HG w[*]; P{w[+mC]=UAS-Xbp1.EGFP.HG}3	BDSC	60731
UAS-HSC70–3RNAi w1118; P{GD6484}	VDRC	14882
UAS-Ufd1RNAi y[1] v[1]; P{y[+t7.7] v[+t1.8]=TRiP.GL01251}attP2	BDSC	41823
UAS-Rpn11RNAi y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP.HMS00071}attP2	BDSC	33662
UAS-white RNAi/Cyo	Denise Montell Lab Stock, University of California Santa Barbara	n/a
UAS-lacZ	Denise Montell Lab Stock, University of California Santa Barbara	n/a
C306-Gal4	Denise Montell Lab Stock, University of California Santa Barbara	n/a
UAS-Notch.Intracellular.Domain	Go et al., 1998 52	n/a
UAS-Rh1G69D	Ryoo et al., 2007 25	n/a
GMR-Gal4 (2nd chromosome)	Denise Montell Lab Stock, University of California Santa Barbara	n/a
dZIP7G178D	Wang et al., 2006 15	n/a
hsFLP12,yw;ubi:GFPnls	Denise Montell Lab Stock, University of California Santa Barbara	n/a
hsFLP12,yw;ubi:RFPnls, FRT40A/(CyO)	Denise Montell Lab Stock, University of California Santa Barbara	n/a
dZIP7::GFP	VDRC	318542
UAS-dZIP7H183A	This paper	n/a
UAS-dZIP7H344A	This paper	n/a
UAS-dZIP7H187A	This paper	n/a
UAS-dZIP7H315A	This paper	n/a
Oligonucleotides
dZIP7 Forward Primer ctctgaatagggaattgggATGGCCAAACAAGTGGCTGA	This paper	n/a
dZIP7 Reverse Primer ccgcagatctgttaacgtcaCGTAGAATCGAGACCGAGGAG AG	This paper	n/a
hZIP7 Forward Primer tatatagctagcgccaccATGGCCAGAGGC	This paper	n/a
hZIP7 Reverse Primer atatataccggtAGCGTAATCTGGAACATCGtatggg	This paper	n/a
Recombinant DNA
pUASt-attb	Denise Montell Lab Stock, University of California Santa Barbara	n/a
pCMV-VSV-G	AddGene	8454
pCMV-dvpr-dR8.2	AddGene	8455
mCherry	AddGene	124428
pcDNA3.1 V5-His B	Denise Montell Lab Stock, University of California Santa Barbara	n/a
Software and algorithms
Adobe Illustrator 2023	Adobe	adobe.com
Prism 9	Graph Pad	graphpad.com
Zen	Zeiss	zeiss.com
ImageJ2 (FIJI)	Schindelin et al., 2012	fiji.sc

Highlights.

The ER-localized Zn²⁺ transporter ZIP7 promotes ER-associated degradation (ERAD)

ZIP7 provides rate-limiting Zn²⁺ to degrade proteins that misfold in the ER

ZIP7 enhances protein deubiquitination by the Zn²⁺ metalloproteinase Rpn11

ZIP7 overexpression prevents blindness in a Drosophila model of neurodegeneration