(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



A structurally conserved site in AUP1 binds the E2 enzyme UBE2G2 and is essential for ER-associated degradation

['Christopher E. Smith', 'Laboratory Of Protein Dynamics', 'Signaling', 'Center For Cancer Research', 'Nci', 'National Institutes Of Health', 'Frederick', 'Maryland', 'United States Of America', 'Yien Che Tsai']

Date: 2022-01 

Endoplasmic reticulum–associated degradation (ERAD) is a protein quality control pathway of fundamental importance to cellular homeostasis. Although multiple ERAD pathways exist for targeting topologically distinct substrates, all pathways require substrate ubiquitination. Here, we characterize a key role for the UBE2G2 Binding Region (G2BR) of the ERAD accessory protein ancient ubiquitous protein 1 (AUP1) in ERAD pathways. This 27-amino acid (aa) region of AUP1 binds with high specificity and low nanomolar affinity to the backside of the ERAD ubiquitin-conjugating enzyme (E2) UBE2G2. The structure of the AUP1 G2BR (G2BR AUP1 ) in complex with UBE2G2 reveals an interface that includes a network of salt bridges, hydrogen bonds, and hydrophobic interactions essential for AUP1 function in cells. The G2BR AUP1 shares significant structural conservation with the G2BR found in the E3 ubiquitin ligase gp78 and in vitro can similarly allosterically activate ubiquitination in conjunction with ERAD E3s. In cells, AUP1 is uniquely required to maintain normal levels of UBE2G2; this is due to G2BR AUP1 binding to the E2 and preventing its rapid degradation. In addition, the G2BR AUP1 is required for both ER membrane recruitment of UBE2G2 and for its activation at the ER membrane. Thus, by binding to the backside of a critical ERAD E2, G2BR AUP1 plays multiple critical roles in ERAD.

Funding: This Research was supported in whole by the Center for Cancer Research, National Cancer Institute, National Institutes of Health Intramural Research Program project numbers ZIA BC010292 (Allan M. Weissman), ZIA BC010326 (Xinhua Ji), and ZIA BC011131 (R. Andrew Byrd) and by federal funds from the National Cancer Institute, National Institutes of Health, under contract HHSN261200800001E (Leidos Biomedical Research). The funders had no role in study design, data collection and analysis, or preparation of the manuscript.

This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.

We now provide evidence demonstrating that AUP1 is required for UBE2G2-mediated substrate ubiquitination. By contrast, both AUP1 and UBE2G2 are dispensable for the constitutive ubiquitination of HRD1 that we observe in cells. Furthermore, while ERAD of both HRD1 (ERAD-L) and TRC8 (ERAD-C) substrates are seemingly unaffected by the loss of the acyltransferase and CUE domains, the G2BR is absolutely required for degradation of these substrates. We have similarly found requirements for G2BR AUP1 and for G2BR gp78 in the degradation of ERAD-M substrates for which HRD1 and gp78, respectively, are implicated. We also report the structure of the G2BR AUP1 in complex with UBE2G2, confirm intermolecular contacts critical for this high-affinity interaction, and assess the role of G2BR AUP1 in UBE2G2:RING affinity. Additionally, we have determined that the G2BR AUP1 facilitates the critical role of UBE2G2 in ERAD through multiple mechanisms including increasing its levels, recruiting it to the ER membrane, and markedly enhancing its ubiquitin-conjugating activity.

Although it is evident that AUP1 plays a role in ERAD, how it functions and whether it is involved in substrate ubiquitination have not been assessed. Lipid droplets have been postulated as an alternative to proteinaceous channels as a mechanism for dislocating proteins from the ER during ERAD [ 48 ], and mutations in the AUP1 acyltransferase domain reduce lipid droplet formation in cells loaded with oleic acid [ 35 ]. However, at least in yeast, deficiencies in lipid droplet formation do not correlate with defective ERAD [ 49 ]. Until now, the significance of the AUP1 acyltransferase domain in ERAD has not been evaluated. There is evidence that the AUP1 CUE domain provides a means of interaction with components in the HRD1 ERAD pathway as well as with misfolded proteins [ 35 ]. This interaction likely occurs by binding ubiquitinated proteins. The CUE domain has also been shown to play a role in lipid droplet clustering [ 36 ] and in ubiquitination of AUP1 itself [ 35 ]. Again, however, there has not been a direct assessment of the significance of this domain in ERAD. Along the same lines, while AUP1 is found to associate with UBE2G2 in a G2BR AUP1 -dependent manner [ 35 , 37 , 50 ], it has not been determined whether G2BR AUP1 or its interaction with UBE2G2 is of significance in ERAD.

A candidate to play a Cue1p-like role in mammals is ancient ubiquitous protein 1 (AUP1). This 410-amino acid (aa) protein is inserted into both the ER membrane and lipid droplets through an N-terminal hydrophobic “hairpin” sequence. The cytoplasmic region of AUP1 includes both an acyltransferase and a CUE domain, which is followed by a carboxyl-terminal G2BR (G2BR AUP1 ) [ 35 – 38 ]. AUP1 has been suggested to play a role in the retrotranslocation step for ERAD-L substrates of HRD1, including the Null Hong Kong (NHK) variant of alpha1-antitrypsin (α1ΑΤ) and a truncation mutant of Ribophorin I (RI 332 ) [ 35 ]. Although an RNA interference (RNAi) screen did not recapitulate a requirement for AUP1 in HRD1-mediated ERAD-L [ 39 ], a more recent CRISPR/Cas-9 screen did implicate AUP1 in this pathway [ 40 ]. This CRISPR/Cas-9 screen, as well as a fluorescence insertional mutagenesis screen [ 41 ], also found AUP1 to be required for degradation of fluorescent cytosolic proteins that can associate with the ER membrane. These fluorescent substrates engage the ERAD-C machinery as a consequence of being fused to the well-described hydrophobic CL1 degron [ 42 – 46 ]. In both of these screens, TRC8/RNF139 was found to be an E3 for these ERAD-C substrates [ 40 , 41 ] and, in the insertional mutagenesis screen, another E3 MARCH6/TEB4 was also implicated [ 41 ]. In contrast to findings for ERAD-L and ERAD-C substrates, AUP1 is not required for degradation of INSIG-1, an ERAD-M substrate targeted by gp78 [ 40 , 47 ].

In mammals, there may be up to two dozen E3s that are resident to the ER and potentially involved in ERAD. Approximately half of these have been implicated in the degradation of either naturally occurring or model substrates [ 1 , 20 – 22 ]. We and others have characterized the requirements for the function of the E3 gp78 (aka AMFR or RNF45) in ERAD [ 17 , 23 – 28 ]. This polytopic ER protein has an extended carboxyl-terminal cytosolic region that includes three domains critical for its activity [ 23 ]. These include a RING domain, a CUE domain that robustly binds ubiquitin, and a binding site for the E2 UBE2G2, which is the mammalian ortholog of Ubc7p [ 29 ]. This E2 binding site in gp78 is referred to as the UBE2G2 Binding Region (G2BR) [ 23 , 30 ]. The gp78 G2BR (G2BR gp78 ), analogous to the U7BR of yeast Cue1p, binds to the backside of UBE2G2, increases E2:RING affinity, and stimulates ubiquitination in vitro [ 17 ]. However, the functions that it performs in cells in facilitating ERAD have not been directly assessed. Unlike gp78, neither the ortholog of yeast Hrd1p, HRD1/Synoviolin [ 31 – 33 ], nor that of Doa10p, MARCH6/TEB4 [ 32 , 34 ], encode a G2BR-like region or CUE domain. Similarly, such domains have not been described for other ERAD E3s. The question then arises as to whether there is a mammalian equivalent of yeast Cue1p.

In the yeast Saccharomyces cerevisiae, where this process has been most extensively studied, there are two primary ERAD E3s, Hrd1p and Doa10p, both of which are polytopic RING-type E3s that are resident to the ER [ 5 – 7 ]. These two E3s are central to ERAD complexes that recognize substrates depending on the topology of their degradation-targeting signals (degrons). Hrd1p is generally responsible for targeting proteins for ERAD that have luminal or ER membrane degrons (ERAD-L and ERAD-M, respectively), while Doa10p has been implicated primarily in the targeting of proteins with degrons in their cytosolic domains (ERAD-C) [ 8 – 11 ]. Hrd1p and Doa10p function in ERAD primarily with two E2s: Ubc6p, which is carboxyl-terminally ER membrane anchored, and Ubc7p, which is not membrane bound [ 12 , 13 ]. Ubc7p associates with the ER membrane via interactions with the ERAD accessory protein Cue1p [ 14 ], which is anchored to the ER by a single N-terminal transmembrane domain. Cue1p also contains a cytoplasmic ubiquitin-binding CUE domain that plays a role in degradation of some substrates and has been shown to facilitate ubiquitin chain elongation [ 15 , 16 ]. The carboxyl-terminal Ubc7-Binding Region (U7BR) of Cue1p binds to the “backside” of Ubc7p—an area that is distinct from both the RING domain–interacting region of the E2 and its catalytic Cys and surrounding residues [ 17 – 19 ]. Importantly, the U7BR can activate ubiquitination in vitro by allosterically increasing the affinity of Ubc7p for the RING domains of both Hrd1p and Doa10p [ 19 ].

The ubiquitin and proteasome-dependent degradation of proteins from the endoplasmic reticulum (ER) via endoplasmic reticulum–associated degradation (ERAD) is a critical mechanism for both protein quality control and regulation of protein levels. Dysfunction of this homeostatic mechanism is associated with a wide range of pathologies. ERAD involves the tightly coupled processes of protein recognition, ubiquitination, retrotranslocation or dislocation, and proteasomal degradation. Central to this process are specific pairs of ubiquitin protein ligases (E3s) and ubiquitin-conjugating enzymes (E2s) [ 1 – 4 ].

Results

The G2BRAUP1 allosterically increases the affinity of UBE2G2 for the gp78 RING domain Our previous studies of UBE2G2:G2BRgp78 revealed a significant change in nuclear magnetic resonance (NMR) chemical shifts for the backbone 15N-1HN resonances of UBE2G2 upon binding of G2BRgp78 [17,24]. The recognition that the backbone structure of UBE2G2 changes very little between the apo and G2BRgp78 bound states suggested that the chemical shifts were due to variations in the hydrogen bond strengths that stabilize the core structure. Nevertheless, subtle changes in populations of dynamic conformers at the gp78 RING:UBE2G2 interface take place to promote the significant allosteric effect observed for binding of gp78 RING to UBE2G2 when G2BRgp78 is bound [69]. We explored whether the same effects are present in the interaction with G2BRAUP1. The 15N-1HN chemical shifts of UBE2G2 exhibit the same dramatic shifts upon binding G2BRAUP1 as upon binding G2BRgp78, suggesting similar allosteric effects (Fig 5A–5C, Biological Magnetic Resonance Data Bank deposition https://doi.org/10.13018/bmrbig33). Furthermore, consistent with the very similar crystal structures of UBE2G2:G2BRgp78 and UBE2G2:G2BRAUP1 that we report here, the UBE2G2 chemical shifts in the 15N-1HN HSQC spectrum showed a very similar shift from the apo UBE2G2 (Fig 5C). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Titration of UBE2G2:G2BR with gp78 RING. Superposition of 15N-1HN HSQC NMR spectra for apo-UBE2G2 and UBE2G2 bound to G2BRAUP1 (A) or G2BRgp78 (B). Spectra for UBE2G2 (150 μM) bound to G2BR were acquired at a 1:1.1 ratio. (C) Superposition of the spectra for apo-UBE2G2, UBE2G2:G2BRAUP1, and UBE2G2:G2BRgp78, illustrating the equivalence of perturbations between G2BRAUP1 and G2BRgp78. (D) Expansion of overlaid resonances corresponding to UBE2G2 residues showing fast exchange binding kinetics for gp78 RING binding to both UBE2G2:G2BRAUP1 (left panels) and UBE2G2:G2BRgp78 (right panels). Spectra were acquired at ratios of RING to UBE2G2:G2BR of 0, 0.33, 0.67, 1.0, 2.0, and 3.0. (E) Binding curves as a function of [RING]/[UBE2G2:G2BR] ratio and analyses for representative residues E12, Q15, L66, and V113 of UBE2G2:G2BRgp78 (orange) and UBE2G2:G2BRAUP1 (magenta) UBE2G2:G2BR binding to gp78 RING. Errors are reported from the regression analysis of the data in S1 Table (B, D). G2BR, UBE2G2 Binding Region; NMR, nuclear magnetic resonance. https://doi.org/10.1371/journal.pbio.3001474.g005 We have previously found that the addition of G2BRgp78 or U7BR to their cognate E2s resulted in a significant increase in E2:RING affinity as assessed by NMR, which correlated with enhanced ubiquitination in vitro [17,19]. Taking advantage of the well-characterized gp78 RING domain, we assessed the relative effects of G2BRgp78 and G2BRAUP1 on the affinity of UBE2G2 for this domain. Titration of UBE2G2 and UBE2G2:G2BRAUP1 with the gp78 RING confirmed that the gp78 RING exhibited mostly fast exchange kinetics and that the increased affinity of gp78 RING was enhanced to a similar degree by G2BRAUP1 as by G2BRgp78. Through chemical shift perturbation (CSP) mapping, backbone amide resonances of G2BR-bound UBE2G2 were examined upon titration of the RING domain (Fig 5D and 5E, Biological Magnetic Resonance Data Bank deposition https://doi.org/10.13018/bmrbig33). Titration of UBE2G2:G2BRgp78 and of UBE2G2:G2BRAUP1 resulted in measured affinities of 20.3 ± 1.5 μM and 22.5 ± 1.8 μM, respectively. These results indicate equivalent enhancements of approximately 10-fold for G2BRgp78 and G2BRAUP1 on the affinity of UBE2G2 for the gp78 RING, which was measured as 207.4 ± 6.4 μM. Consistent with this increase in affinity, the G2BRAUP1, like the G2BRgp78, increased ubiquitination in an in vitro autoubiquitination assay employing UBE2G2 with either the gp78 RING domain or those of HRD1 and TRC8 (S4A and S4B Fig). Thus, as assessed both biophysically and functionally, G2BRAUP1 has similar effects on UBE2G2 activity to what is observed with G2BRgp78.

The AUP1–UBE2G2 interaction is required for ERAD Inspection of the UBE2G2:G2BRAUP1 interface reveals several contact residues that are likely responsible for their high-affinity binding. Specifically, five positively charged or polar residues of the G2BRAUP1—R382, Q383, K390, R398, and R400—form salt bridges or hydrogen bonds with negatively charged or polar groups on UBE2G2 (Fig 4B and 4C). We first generated alanine mutations on these G2BRAUP1 residues to neutralize positively charged side chains and assessed their effects on UBE2G2 binding and activity in cells by GST pulldown and CHX chase, respectively. Beginning at the N-terminal end of the G2BRAUP1 helix, the R382A/Q383A mutation (2A) was unable to disrupt UBE2G2 binding (Fig 6A). Charge reversal of these residues (R382E/R383E, or 2E) also failed to disrupt UBE2G2 binding. Mutation of all five of the positively charged or polar G2BRAUP1 residues to glutamate (5E), however, led to a loss of detectable UBE2G2 binding. Consistent with the importance of binding to UBE2G2, expression of a full-length AUP1 bearing the 2E mutation restored degradation of NHK in the AUP1 KO cells, to a degree comparable to WT AUP1 (Fig 6B). By contrast, expression of the AUP1 5E mutant failed to restore NHK degradation in these cells. These results support a requirement for UBE2G2 binding in AUP1 function in cells. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. Disruption of UBE2G2:G2BRAUP1 interaction abrogates AUP1 function in cells. (A) GST or indicated GST-AUP1 fusions (aa 292 to 410, expression shown in Coomassie-stained gel) were immobilized on glutathione beads and binding to 35S-UBE2G2 was assessed as in Fig 3A. Mutations include 2A (R382A/Q383A), 2E (R382E/Q383E), and 5E (R382E/Q383E/K390E/R398E/R400E). (B) AUP1 KO cells were transfected with NHK-HA and forms of full-length AUP1-FLAG with the indicated point mutations from (A) and NHK degradation assessed by CHX chase. (C) Binding assay performed as in (A) with AUP1 mutants predicted to disrupt hydrophobic interactions with UBE2G2. (D) Cellular assay performed as in (B) with AUP1-FLAG mutants that disrupt interactions with UBE2G2 in vitro in (C). The data underlying this figure can be found in S1 Data. aa, amino acid; AUP1, ancient ubiquitous protein 1; CHX, cycloheximide; G2BR, UBE2G2 Binding Region; KO, knockout; WT, wild type. https://doi.org/10.1371/journal.pbio.3001474.g006 As mentioned, the pattern of hydrophobic interactions in the UBE2G2:G2BRAUP1 and UBE2G2:G2BRgp78 complexes are highly similar (Fig 4C), suggesting that the van der Waals forces resulting from a perfect landscape match of the 2 molecules is also required for the stabilization of the complex. To test this hypothesis, we mutated the G2BRAUP1 residues L386, Y394, and A397 to aspartate to disrupt this hydrophobic landscape and assessed effects on in vitro binding of UBE2G2. While the Y394D mutation alone reduced binding, mutation of L386 and Y394 together eliminated detectable binding to UBE2G2 (Fig 6C). Binding was similarly lost with a single point mutation, A397D, which makes multiple hydrophobic contacts with UBE2G2 (Fig 4C). Consistent with the in vitro binding results, reexpression of AUP1 harboring either the L386D/Y394D or A397D mutations failed to restore NHK degradation in cells lacking AUP1, confirming the functional importance of these residues (Fig 6D). Taken together, our mutational analysis demonstrates that the high-affinity binding between G2BRAUP1 and UBE2G2, driven by both electrostatic and hydrophobic interactions, is critical for the function of AUP1 in ERAD.

[END]

[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001474

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/