Murine Leukemia Virus GlycoGag Antagonizes SERINC5 via ER-phagy Receptor RETREG1

Abstract

Serine incorporator 5 (SERINC5) is a host restriction factor that targets certain enveloped viruses, including human immunodeficiency virus type 1 (HIV-1) and murine leukemia virus (MLV). It integrates into the viral envelope from the cell surface, inhibiting viral entry. SERINC5 is transported to the cell surface via polyubiquitination, while a single K130R mutation retains it in the cytoplasm. Both HIV-1 Nef and MLV glycoGag proteins antagonize SERINC5 by reducing its expression in producer cells. Here, we report that MLV glycoGag employs selective autophagy to downregulate SERINC5, demonstrating a more potent mechanism for decreasing its cell surface expression. Although glycoGag is a type II integral membrane protein, it primarily localizes to the cytoplasm and undergoes rapid proteasomal degradation. Employing the K130R mutant, we show that Nef, primarily associated with the plasma membrane, downregulates SERINC5 only after it has trafficked to the cell surface, whereas glycoGag can reduce its expression before reaching the plasma membrane while still in the cytoplasm. Nonetheless, an interaction with SERINC5 stabilizes and recruits glycoGag to the plasma membrane, enabling it to downregulate SERINC5 from the cell surface. Through affinity-purified mass spectrometry analysis combined with CRISPR/Cas9 knockouts, we find that glycoGag’s activity depends on reticulophagy regulator 1 (RETREG1), an ER-phagy receptor. Further knockout experiments of critical autophagy genes demonstrate that glycoGag downregulates cytoplasmic SERINC5 via micro-ER-phagy. These findings provide crucial new insights into the ongoing arms race between retroviruses and SERINC5 during infection.

AUTHOR SUMMARY

HIV-1 Nef and MLV glycoGag are unrelated viral proteins, yet both counteract the same host restriction factor, SERINC5, to facilitate productive infection. In this study, we report a novel pathway through which glycoGag downregulates SERINC5. We demonstrate that while Nef downregulates SERINC5 only after it has trafficked to the cell surface, glycoGag can directly downregulate SERINC5 in the cytoplasm before it reaches the plasma membrane. Furthermore, we show that this pathway is mediated by the ER-phagy receptor RETREG1, which targets SERINC5 for degradation via micro-ER-phagy. This mechanism provides a more effective means of blocking SERINC5 antiviral activity. These findings reveal that retroviruses have evolved different strategies to antagonize SERINC5, highlighting the critical role of SERINC5 in restricting retroviral infections.

INTRODUCTION

Retroviruses, in particular, lentiviruses, express accessory proteins to antagonize host restrictions and establish productive infection. Negative factor (Nef) and glycosylated Gag (glycoGag) are two unrelated accessory proteins expressed from human immunodeficiency virus type 1 (HIV-1) or murine leukemia virus (MLV), but antagonize the same host restriction factor, serine incorporator 5 (SERINC5) that strongly blocks virus entry [1].

SERINC5 (Ser5) was originally discovered in human T cells that restricts HIV-1 infection but is antagonized by Nef [2, 3]. Ser5 also restricts MLV and other enveloped viruses [4-11]. Human Ser5 is a ~45-kDa integral membrane protein with ten transmembrane helices organized into two subdomains [12]. The trans-Golgi network (TGN)-located cullin-3 (Cul3)-Kelch-like protein 20 (KLHL20) E3 ubiquitin (Ub) ligase polyubiquitinates Ser5 on the lysine 130 (K130) via K33/K48-branched Ub chains [13]. The K33-linked Ub chains direct Ser5 from TGN to the plasma membrane [13], where Ser5 is incorporated into the budding virus particles to inhibit the virus entry [14-17].

Nef internalizes Ser5 from the cell surface via receptor-mediated endocytosis and targets Ser5 to the Rab5⁺ early, Rab7⁺ late, and Rab11⁺ recycling endosomes, resulting in degradation in endolysosomes [18]. Nef downregulation of Ser5 is dependent on the Ser5 intracellular loop 4 (ICL4) [19]. Nef interacts with the Cyclin K (CCNK)/Cyclin-dependent kinase 13 (CDK13) complex that phosphorylates Ser5 on the serine 360 (S360) in ICL4 [20]. Phosphorylated Ser5 is endocytosed and targeted to endolysosomes for degradation via the K48-linked polyubiquitination. GlycoGag could also downregulate Ser5 from the cell surface, but the precise mechanism has not been explored as completely as Nef [5, 9].

ER-phagy, often used interchangeably with reticulophagy, is a selective autophagy of the endoplasmic reticulum (ER) that manages ER quality control and size by degrading and recycling portions of the ER. Autophagy is divided into macroautophagy/autophagy, microautophagy, and chaperone-mediated autophagy (CMA) in mammalian cells [21]. Macroautophagy sequesters cargos into autophagosomes, which undergo a series of fusion processes with the late endosomes/lysosomes and mature into functional autolysosomes for bulk degradation. This process is mediated by the autophagosome biosynthesis machinery and the MAP1LC3/LC3 lipidation machinery that includes many autophagy-related (ATG) proteins. Microautophagy directly engulfs and sequesters protein substrates into lysosomes for degradation. CMA selects proteins with a KFERQ-related motif, which binds to the cytosolic chaperone, heat shock protein family A (Hsp70) member 8 (HSPA8), for recruitment to the lysosome surface. The HSPA8-associated cargo then binds to lysosome-associated membrane protein (LAMP) 2a, which is the CMA receptor for lysosomal entry.

Similarly, ER-phagy is also divided into three primary types, including macro-ER-phagy, micro-ER-phagy, and LC3-dependent vesicular transport. ER-phagy is initiated by ER-phagy receptors on the ER membrane, including RETREG1/FAM134B, RETREG1-2 (N-terminally truncated RETREG1), RETREG2/FAM134A, RETREG3/FAM134C, RTN3L, ATL3, SEC62, CCPG1, and TEX264 [22]. These receptors have at least one reticulon-homology domain (RHD) and/or LC3-interacting region (LIR), marking the ER portions for autophagic degradation [23].

We now report that while Nef downregulates Ser5 only at the cell surface via endolysosomes, glycoGag can directly downregulates Ser5 from the ER via ER-phagy. We further identify RETREG1 as the ER-phagy receptor that targets Ser5 to micro-ER-phagy for degradation.

RESULTS

MLV glycoGag antagonizes Ser5 much more efficiently than HIV-1 Nef.

MLV glycoGag is an N-terminally extended form of the Gag polyprotein and is modified by three N-glycosylation sites (Fig. 1A). Although glycoGag was initially described as a type II integral membrane protein [24], it has also been shown to adopt a type I integral membrane topology [25]. To clarify this discrepancy, we utilized the transmembrane hidden Markov model (TMHMM) to analyze glycoGag. The prediction indicates that glycoGag features an N-terminal intracellular domain, a C-terminal extracellular domain, and a transmembrane (TM) domain located at the junction of the leader and MA regions (Fig. 1A). We have previously reported that glycoGag downregulates Ser5 through its leader domain, further supporting its intracellular localization [9]. Therefore, glycoGag is a type II transmembrane protein.

Figure 1. MLV-glycoGag antagonizes Ser5 much more effectively than HIV-1 Nef.

A) The MLV-Gag precursor (p65) is translated from an alternatively spliced mRNA and cleaved into the matrix (MA), p12, capsid (CA), and nucleocapsid (NC). The glycoGag (p80) translation is initiated from a CUG codon upstream of Gag, containing three N-glycosylation sites (N113, N480, N505, highlighted in red). This glycoprotein is predicted to function as a type II transmembrane protein, as suggested by the transmembrane hidden Markov model (TMHMM).

B) Nef-defective HIV-1 were produced from HEK293T cells after transfection with 1 μg HIV-1 proviral pNL-ΔEΔN vector, 500 ng HIV-1 Env expression vector pNLnΔBS, 50 ng pCMV6-Ser5, and 3 μg pcDNA3.1-Nef or pcDNA3.1-glycoMA. After being normalized by p24^Gag ELISA, viral infectivity was measured via infection of TZM-bI cells. Results were shown as relative values, with the infectivity of viruses produced alone set as 100.

C) The expression levels of Ser5, Nef, and glycoMA in HEK293T cells in (A) were determined by western blotting (WB). Levels of Ser5, Nef, and glycoMA protein expression were quantified by ImageJ and are shown as relative values, with the value of Ser5 alone or Nef set as 1.

D) The Ser5 antagonism activity is calculated as the relative levels of Ser5 downregulation normalized to Nef or glycoMA expression.

Error bars represent standard error of measurements (SEMs) calculated from two (A) or three (C, D) independent experiments.

We compared the Ser5 downregulation activity by Nef and glycoGag. Since glycoGag is functionally exchangeable with glycoMA [2, 3], we used glycoMA in this measurement. Nef-deficient (ΔNef) HIV-1 viruses were produced from HEK293T cells after transfection with an HIV-1 proviral vector, an HIV-1 Env expression vector, a Ser5 expression vector, and a vector expressing either Nef or glycoMA. Subsequently, viral infectivity was assessed after infection of the HIV-1 luciferase-reporter cell line TZM-bI. As anticipated, Ser5 substantially reduced ΔNef HIV-1 infectivity by approximately 60-fold, an effect that was antagonized by both Nef and glycoMA (Fig. 1B). However, glycoMA restored viral infectivity to a greater extent than Nef. When we analyzed protein expression in the viral producer cells by Western blotting (WB), we found that, despite glycoMA being expressed at approximately sevenfold lower levels than Nef, it reduced Ser5 expression at least twofold more effectively (Fig. 1C). After normalizing Ser5 downregulation against the expression levels of Nef and glycoMA, it became clear that glycoMA counteracted Ser5 at least tenfold more effectively than Nef (Fig. 1D).

MLV glycoGag is rapidly turned over in proteasomes.

We assessed the stability of glycoGag protein after infection of NIH3T3 cells with MLV for 24 hrs. After treatment with cycloheximide (CHX) for 0.5, 1.0, and 2.0 hrs, viral proteins were detected by WB. Notably, glycoGag exhibited significantly less stability than Gag proteins (p65, p30), with a half-life of approximately 15 min (Fig. 2A, lanes 1-8). We directly expressed Nef, glycoMA, and glycoGag in HEK293T cells to confirm this finding and compared their stability. Both glycoMA and glycoGag had a half-life of approximately 15 min, while Nef had a half-life exceeding 10 hrs (Fig. 2B).

Figure 2. MLV glycoGag is rapidly turned over in proteasomes.

A) NIH3T3 cells were infected with MLV. After 24 hours of infection or transfection, cells were treated with 50 μM CHX or dimethyl sulfoxide (DMSO) as a vehicle control. Cells were collected at the indicated time points and the protein expression was analyzed by WB. Levels of viral proteins (glycoGag, p30, p65) were quantified using ImageJ. Results are shown as relative values, with the values at time zero set as 1.A.

B) Nef, glycoMA, and glycoGag were expressed in HEK293T cells and treated with 50 μM CHX or DMSO. Their expression and stability were determined and quantified similarly.

C) Nef, glycoMA, and glycoGag were expressed in HEK293T cells and treated with 10 mM 3-methyladenine (3-MA), 20 μM LY294002 (Ly), 100 nM bafilomycin A1 (Baf-A1), or 20 μM MG132. Their expression was analyzed by WB and quantified by Image J. Results are shown in relative values, with the values in the absence of CHX treatment set as 1.

D) Nef, glycoMA, and glycoGag were expressed with Cas9 in HEK293T cells in the presence of Cul3- and/or KLHL20-specific sgRNAs. Protein expression was determined by WB.

All experiments were repeated at least twice, and similar results were obtained.

To investigate the mechanism behind the rapid turnover of glycoGag, we treated cells with the proteasomal inhibitor MG132, as well as autophagy and lysosomal inhibitors, 3-methyladenine (3-MA), LY294002 (LY), and bafilomycin A1 (Baf-A1). None of these inhibitors affected Nef expression (Fig. 2C, lanes 1-6). While 3-MA, LY, and Baf-A1 did not yield significant changes, MG132 markedly increased the expression of glycoMA and glycoGag (Fig. 2C, lanes 12, 18, 24). We previously reported that Ser5 is polyubiquitinated by Cul3-KLHL20 E3 ubiquitin ligase [13]. When we silenced the expression of this E3 ligase complex using CRISPR ribonucleoprotein (RNP) complexes containing Cul3- or KLHL20-specific guide RNAs (sgRNA), we found that while Nef expression remained unchanged, the expression levels of both glycoMA and glycoGag increased (Fig. 2D). Thus, our results indicate that glycoGag is targeted to proteasomes for degradation.

Ser5 stabilizes glycoGag by targeting it to the plasma membrane.

We began by comparing the subcellular localization of Nef and glycoGag using confocal microscopy. While Nef was predominantly associated with the plasma membrane, both glycoMA and glycoGag were localized in the cytoplasm and enriched in punctate areas (Fig. 3A). To confirm this observation, we isolated cytosolic (Cyto) and membrane (Mem) fractions from cells expressing Ser5, Nef, and glycoMA, detecting their expression in these fractions via WB. Ser5 and Nef were found in both fractions, whereas glycoMA was only detected in the cytosolic fraction (Fig. 3B, lanes 1, 2, 5, 7, 8, 11).

Figure 3. Ser5 stabilizes glycoGag by targeting it to the plasma membrane.

A) Nef with a C-terminal GFP-tag, glycoMA with a C-terminal HA-tag, and glycoGag with a C-terminal Myc-tag were expressed in HeLa cells. After 24 hours, cells were stained with DAPI for the nuclei and with anti-HA or anti-Myc followed by Alexa Fluor 488-conjugated goat anti-mouse antibodies. Nef, glycoMA, and glycoGag expression were detected by confocal microscopy (scale bar 5 μm).

B) Ser5-FLAG was expressed with Nef-HA or glycoMA-HA in HEK293T cells. Cells were separated into cytosolic (Cyto) and plasma membrane (Mem) fractions, where Ser5, Nef, and glycoMA expression were detected by WB.

C) Experiments were repeated as in (B). Proteins were immunoprecipitated (IP) from Mem and Cyto fractions by anti-FLAG and detected by WB.

D) Nef, glycoMA, and glycoGag were expressed alone or with Ser5 in HEK293T cells. After treatment with 50 μM CHX, cells were collected at the indicated time points and protein expression was analyzed by WB. Levels of Nef, glycoMA, and glycoGag expression were quantified using ImageJ. Results are shown as relative values, with the values at time zero in the presence of Ser5 set as 1.

All experiments were repeated at least twice, and similar results were obtained.

Previously, we identified a strong interaction between Ser5 and glycoMA, but not with Nef, through co-immunoprecipitation (co-IP) [5]. Thus, we aimed to determine whether Ser5 affects the subcellular distribution of glycoMA. When co-expressed with Nef and glycoMA, Ser5 did not alter the levels of Nef in the cytosolic and membrane fractions (Fig. 3B, lanes 2, 3, 5, 6); however, it significantly increased the levels of glycoMA in the membrane fraction (Fig. 3B, lanes 11, 12). We further conducted co-IP experiments to examine the interactions of Ser5 with Nef and glycoMA. While Ser5 could not pull down Nef in the cytosolic or membrane fractions, it effectively pulled glycoMA from both fractions (Fig. 3C).

We then tested whether Ser5 affects the stability of glycoGag, glycoMA, and Nef after they were co-expressed in HEK293T cells. Ser5 did not alter Nef stability (Fig. 3D, lanes 1–12); however, it significantly stabilized glycoGag and glycoMA, increasing their half-lives to over 2 hrs (Fig. 3D, lanes 13–24, 25–36).

MLV glycoGag downregulates Ser5 in the cytoplasm.

Ser5 K130 is located at the border between the third transmembrane domain (TMD3) and the first intracellular loop (ICL1), which is polyubiquitinated by Cul3-KLHL20. This residue is also conserved in murine Ser5 (mSer5) as K131 (Fig. 4A). When these lysine residues were mutated to arginine, both the Ser5 K130R and mSer5 K131R mutant exhibited a cytoplasmic distribution, whereas their wild-type (WT) counterparts were associated with the plasma membrane, as shown by confocal microscopy (Fig. 4B). Following cell fractionation into cytosolic and membrane fractions, both K130R and K131R were found in the cytosolic but not membrane fraction (Fig. 4C, lanes 3, 4). In addition, they both lost the antiviral activity against glycoGag-deficient MLV (MLVΔglycoGag) and Nef-deficient HIV-1 (HIV-1ΔNef) (Fig. 4D). Thus, K131 is functionally conserved in mSer5 and is essential for its plasma membrane localization.

Figure 4. MLV glycoGag downregulates Ser5 in the cytoplasm.

A) Ser5 and mSer5 amino acid sequences containing the conserved lysine (K130, K131) are aligned. The ICL1 and TMD3 regions are indicated. Dots indicate identical residues.

B) Ser5 and mSer5 WT proteins and their lysine mutants (K130R, K131R) with a GFP-tag were expressed in HeLa cells and detected by confocal microscopy (scale bar 5 μm).

C) Ser5 and mSer5 proteins and their lysine mutants were expressed in HEK293T cells and their expression in Mem and Cyto fractions were analyzed by WB.

D) MLVΔglycoGag and HIV-1ΔNef luciferase-reporter viruses were produced from HEK293T cells in the presence of indicated Ser5 proteins. After infection of NIH3T3 or TZM-bI cells, their infectivity was compared by measuring intracellular firefly luciferase activity. Results are normalized by the p30^Gag or p24^Gag protein levels of virions and presented as relative values, with the infectivity of viruses produced in a control (Ctrl) vector set as 100. Error bars indicate SEMs calculated from two experiments.

E) Ser5, mSer5, and their lysine mutants were expressed with Nef or glycoMA in HEK293T cells and protein expression was analyzed by WB. 19K/R, Ser5 with all 19 lysine residues replaced with arginine; 23K/R, mSer5 with all 23 lysine residues replaced with arginine.

F) Ser5 and K130R were expressed with Nef or glycoMA in HEK293T cells, where Cas9 and KLHL20- or Cul3-specific sgRNA was also expressed. Protein expression was detected by WB.

G) Ser5-GFP (wild-type, WT) and its mutants (S360A, S249A) were expressed with HA-tagged Nef or glycoMA in HeLa cells. Cells were stained with DAPI (4′,6-diamidino-2-phenylindole) for the nuclei (blue) and with anti-HA followed by Alexa Fluor 594-conjugated goat anti-mouse antibodies for viral proteins and observed by confocal microscopy (scale bar 5 μm).

H) Ser5 and the two mutants were expressed with Nef or glycoMA in HEK293T cells and protein expression was analyzed by WB.

I) Ser5 was expressed with a CCNK-shRNA expression vector and a Nef or glycoMA expression vector in HEK293T cells and protein expression was analyzed by WB.

J) Ser5 was expressed with a CDK13-shRNA expression vector and a Nef or glycoMA expression vector in HEK293T cells and protein expression was analyzed by WB.

All experiments were at least repeated twice, and similar results were obtained.

Next, we tested how glycoGag downregulates these lysine mutants. Ser5 and mSer5 contain a total of 19 or 23 lysine residues, respectively. All these lysine residues were mutated to arginine, generating two other lysine mutants, Ser5 19K/R and mSer5 23K/R. When Ser5 (WT, K130R, 19K/R) and mSer5 (WT, K131R, 23K/R) proteins were co-expressed with HIV-1 Nef or MLV glycoMA, Nef downregulated only the WT proteins but not any lysine mutants (Fig. 4E, lanes 2, 5, 8, 11, 14, 17). In contrast, glycoMA downregulated all these Ser5 proteins, including all these lysine mutants (Fig. 4E, lanes 3, 6, 9, 12, 15, 18). Additionally, we found that glycoMA still effectively downregulated Ser5 and K130R when Cul3 and KLHL20 were knocked down by their CRISPR RNP complexes, although the Nef downregulation of Ser5 was effectively blocked (Fig. 4F). Collectively, these results demonstrate that although Nef strictly downregulates Ser5 at the cell surface, glycoGag can downregulate Ser5 within the cytoplasm as well.

Previously, we reported that Nef recruits the CCNK/CDK13 kinase complex to phosphorylate Ser5 on S360, which is required for Nef downregulation of the cell surface Ser5 [20]. Thus, we determined whether this S360 phosphorylation is required for glycoGag downregulation of Ser5. We expressed Ser5 and its two serine-to-alanine mutants (S360A, S249A) with Nef or glycoMA in HeLa cells, and their expression was observed by confocal microscopy. As we reported already [20], Nef internalized Ser5 and S249A, but not S360A (Fig. 4G, left panels). Nevertheless, glycoMA internalized all these Ser5 proteins (Fig. 4G, right panels). We then expressed Ser5 WT, 249A, and S360A with Nef or glycoMA in HEK293T cells and compared the Ser5 downregulation by WB. Nef downregulated Ser5 WT and S249A, but not S360A, whereas glycoMA downregulated all these Ser5 proteins (Fig. 4H). Lastly, we expressed Ser5 with these viral proteins in HEK293T cells where CCNK was silenced by short-hairpin RNAs (shRNAs). When their expression was determined by WB, Nef could no longer downregulate Ser5 in the absence of CCNK (Fig. 4I, lanes 1 to 4). Nonetheless, glycoMA still downregulated Ser5 in these cells (Fig.4I, lanes 5 to 8). We further silenced CDK13 and confirmed that CDK13 is required for Nef, but not glycoMA downregulation of Ser5 (Fig. 4J). These results further confirm that Ser5 downregulation by glycoGag does not necessarily require Ser5 on the cell surface.

MLV glycoGag targets cytoplasmic Ser5 to lysosomes for degradation.

To understand how the cytosolic Ser5 is downregulated, we co-expressed Ser5 and K130R with Nef or glycoMA in HEK293T cells and treated cells with MG132 and Baf-A1. Baf-A1 completely blocked Ser5 downregulation by Nef or glycoMA, as well as K130R by glycoMA, while MG132 showed no effect (Fig. 5A), indicating that both Nef and glycoMA target these Ser5 proteins to lysosomes for degradation. We next expressed eGFP-tagged Ser5 or K130R with mCherry-tagged LAMP1 in the presence of Nef and glycoMA and detected their co-localization by confocal microscopy. Ser5-LAMP1 colocalization was detected by the presence of Nef and glycoMA, and K130R-LAMP1 colocalization was detected only by the presence of glycoMA (Fig. 5B), corroborating the WB results. We further expressed Ser5, Nef, glycoMA, and glycoGag with the ER marker calreticulin (CALR). We found that Nef showed minimal colocalization with CALR, compared to its primary localization to the plasma membrane (Fig. 5C). Similarly, Ser5 mainly localized to the plasma membrane but also showed some intracellular colocalization with CALR. Notably, glycoGag and glycoMA strongly colocalized with CALR.

Figure 5. MLV glycoGag targets cytoplasmic Ser5 to lysosomes for degradation.

A) Ser5 and K130R were expressed with Nef or glycoMA in HEK293T cells and treated with 20 μM MG132 or 100 nM Baf-A1. Protein expression was analyzed by WB.

B) The co-localization of Ser5 or K130R with LAMP1 was determined in HeLa cells by confocal microscopy (scale bar 5 μm).

C) Nef and Ser5 with a C-terminal GFP-tag were expressed with calreticulin (CALR) with a mCherry-tag in HeLa cells. Alternatively, glycoMA and glycoGag with an C-terminal HA-tag were expressed with CALR with a GFP-tag in HeLa cells. GlycoMA and glycoGag were stained with mouse anti-HA followed by Alexa Fluor 594-conjugated goat anti-mouse antibodies. Co-localization of Nef, Ser5, glycoMA, and glycoGag with CALR was determined by confocal microscopy (scale bar 5 μm).

D) The colocalization of Nef, Ser5, glycoGag, and glycoMA with the ER marker calreticulin (CALR) were determined by confocal microscopy (scale bar 5 μm). Write down method: .....

E) Indicated proteins with a FLAG or HA-tag were expressed in HEK293T cells and immunoprecipitated with anti-FLAG or anti-HA antibodies. Affinity purified (AP) samples were analyzed by mass spectrometry (MS). Identified ER-phagy receptors are listed.

Our findings thus far indicate that Ser5 and glycoGag can be detected in the ER, with glycoGag downregulating cytoplasmic Ser5 through lysosomal degradation. We hypothesized that glycoGag downregulation of Ser5 is mediated by ER-phagy. Thus, we employed affinity-purified (AP) mass spectrometry (MS) to identify its specific ER-phagy receptor. Ser5 with a FLAG-tag was co-expressed with glycoGag and Nef, both with a HA tag, in HEK293T cells, and proteins were purified by anti-FLAG or anti-HA beads. In a total of 5 samples analyzed, ATL3, SEC62, and TEX264 were detected in all samples (Fig. 5D). RETREG2 and RETREG3 were identified in samples from cells expressing Ser5 alone, Ser5 plus Nef, and Ser5 plus glycoGag. RETREG1 and CCPG1 were detected only in samples from cells expressing Ser5 plus glycoGag. Thus, we decided to investigate how these ER-phagy receptors get involved in the Ser5 degradation.

RETREG1 enhances Ser5 downregulation by glycoGag.

We reasoned that if an ER-phagy receptor is involved in Ser5 degradation, an increase in its expression should enhance this degradation. We co-expressed Ser5 with Nef and glycoGag in the presence of ectopic expression of all known ER-phagy receptors, including ATL3, RTN3L, SEC62, TEX264, CCPG1, RETREG1 (R1), RETREG1-2 (R1-2), RETREG2 (R2), and RETREG3 (R3). Our results showed that none of these ER-phagy receptors enhanced Nef downregulation of Ser5 (Fig. 6A). However, RETREG1 did increase Ser5 downregulation by glycoGag, while the other receptors did not exhibit this effect (Fig. 6B, lanes 23, 24).

Figure 6. RETREG1 enhances glycoGag downregulation of Ser5.

A) Ser5 was expressed with Nef in the presence of indicated ER-phagy receptors in HEK293T cells, and their expression was determined by WB.

B) Ser5 was expressed with glycoGag in the presence of indicated ER-phagy receptors in HEK293T cells, and their expression was determined by WB.

RETREG1 is required for glycoGag downregulation of cytoplasmic Ser5.

To confirm the RETREG1 activity, we knocked out RETREG1, RETREG3, TEX264, CCPG1, RTN3L, ATL3, and SEC62 using CRISPR in HEK293T cells and assessed whether glycoGag could still downregulate K130R in these knockout cells. None of these knockouts (KOs) disrupted the glycoGag activity, except for RETREG1-KO (Fig. 7A, lanes 3, 4). Additionally, we silenced RETREG2 via two small interfering RNA (siRNA#1 and #2) (Fig. S1) and found that this gene is not required for the glycoGag activity (Fig. 7B). To further confirm the requirement for RETREG1, we ectopically expressed RETREG1 and its N-terminally truncated isoform (RETREG1-2), as well as RETREG2 and RETREG3, in the RETREG1-KO cells. Only RETREG1 restored the glycoGag activity, while the other proteins did not (Fig. 7C, lanes 3, 4).

Figure 7. RETREG1 is required for glycoGag downregulation of cytoplasmic Ser5.

A) Ser5 K130R was expressed with glycoGag in indicated ER-phagy receptor-knockout (KO) cells, and their expression was determined by WB.

B) K130R was expressed with glycoGag in HEK293T cells in the presence of small interfering RNA (siRNA) #1 and #2 that specifically silence RETREG2. Ser5 and glycoGag expression were determined by WB.

C) K130R was expressed with glycoGag in HEK293T RETREG1 (R1)-KO cells in the presence of indicated ectopic ER-phagy receptors, and their expression was determined by WB.

D) Ser5 and K130R were expressed with glycoGag in HEK293T WT and R1-KO cells and their expression was determined by WB.

E) Ser5 was expressed with Nef or glycoGag in HEK293T cells and immunoprecipitated. Alternatively, Nef and glycoGag were expressed in HEK293T cells and immunoprecipitated. The endogenous RETREG1 in these samples was detected by WB.

Next, we investigated how RETREG1 affects the downregulation of wild-type Ser5 by glycoGag. We co-expressed Ser5 WT and the K130R mutant with glycoGag in both HEK293T WT and RETREG1-KO cells. The downregulation of the K130R mutant was completely blocked in the KO cells (Fig. 7D, lanes 7, 8), whereas the downregulation of Ser5 WT was only partially inhibited (Fig. 7D, lanes 3, 4). These results suggest that when the RETREG1-dependent degradation pathway is disrupted, glycoGag can still downregulate Ser5, likely by trafficking to the cell surface and undergoing degradation in endolysosomes.

Finally, we investigated the interaction between Ser5, glycoGag, and RETREG1 using co-IP. Ser5-FLAG, Nef-HA, and glycoGag-HA were expressed in HEK293T cells, and endogenous RETREG1 was pulled down using anti-FLAG or anti-HA antibodies. Neither Ser5, Nef, nor glycoGag alone, nor the combination of Ser5 and Nef, successfully pulled down RETREG1 (Fig. 7E, lanes 4, 5, 9, 10). However, Ser5 co-expressed with glycoGag did pull down RETREG1 (Fig. 7E, lane 6). These results collectively demonstrate that glycoGag targets Ser5 for ER-phagy by recruiting RETREG1.

MLV glycoGag targets Ser5 to micro-ER-phagy for degradation.

To investigate the precise ER-phagy mechanism, we first determined whether it depends on autophagosomes. Thus, we tested the glycoGag activity in HEK293T cell lines where the key components of the autophagosome biosynthesis machinery, including PIK3C3/VPS34 and Beclin 1 (BECN1), and the MAP1LC3/LC3 lipidation machinery, including ATG3, ATG5, and ATG7, were knocked out by CRISPR/Cas9 [26]. We also did this experiment in a HEK293T cell line where the autophagy receptor sequestosome 1 (SQSTM1/p62) was knocked out [27]. Ser5 WT and K130R were co-expressed with Nef or glycoMA, and their downregulation was assessed by WB. As expected, Nef did not downregulate K130R in either WT or any knockout cells (Fig. 8A, lanes 9–12, 19–24, 31–36). Nef downregulated Ser5 in most cell lines except for SQSTM1- and PIK3C3-knockout cells (Fig. 8A, lanes 1–6, 13–18, 25–30). GlycoMA downregulated both Ser5 and K130R in all knockout cells (Fig. 8A, lanes 37–72). We further knocked down the CMA receptor LAMP2a using siRNAs and observed that glycoMA continued to downregulate Ser5 and K130R (Fig. S2). These findings suggest that Ser5 downregulation by Nef via the endolysosomal pathway requires SQSTM1/p62 and PIK3C3. Importantly, glycoGag downregulates Ser5 via micro-ER-phagy, rather than macro-ER-phagy, LC3-dependent vesicular transport, or CMA.

Figure 8. MLV glycoGag targets Ser5 to micro-ER-phagy for degradation.

A) Ser5 and K130R were expressed with Nef or glycoMA in HEK293T WT and indicated knockout cells that do not express ATG3, ATG5, ATG7, BECN1, SQSTM1/p62, or PIK3C3. Protein expression was determined by WB.

B) Ser5 and K130R were expressed with Nef or glycoMA in HEK293T cells and treated with 10 mM 3-MA, 20 μM Ly, or 100 nM Baf-A1. Protein expression was analyzed by WB.

C) A proposed model for MLV glycoGag downregulation of Ser5. GlycoGag is expressed as a type II integral membrane protein in the ER, where it interacts with Ser5. This interaction leads to the recruitment of RETREG1, resulting in Ser5 degradation via micro-ER-phagy (1). If Ser5 is not cleared within the ER, it proceeds to the Golgi apparatus and is subsequently delivered to the cell surface after undergoing polyubiquitination. Throughout this process, glycoGag is also recruited to the cell surface by Ser5. At the cell surface, both glycoGag and HIV-1 Nef can internalize Ser5 through early endosomes (EE) and late endosomes (LE), culminating in degradation within endolysosomes (2).

To further investigate this micro-ER-phagy mechanism, we tested the glycoGag activity again after treatment with three autophagy inhibitors, Baf-A1, 3-MA, and Ly. Both Nef- and glycoMA-mediated downregulation of Ser5 were blocked by these inhibitors (Fig. 8B, lanes 1–5, 11–15). However, glycoMA-mediated downregulation of K130R was blocked by Baf-A1 and 3-MA, but not Ly (Fig. 8B, lanes 16–20). These results indicate that while PIK3C3 is not involved, an unidentified 3-MA-sensitive but Ly-insensitive factor is necessary for this micro-ER-phagy pathway.

DISCUSSION

We reported that glycoGag employs a mechanism similar to that of Nef for downregulating Ser5 from the cell surface [9]. It was therefore surprising to find that glycoGag does not require S360 phosphorylation for Ser5 downregulation, a modification that is critical for Nef’s ability to downregulate Ser5. We also discovered that glycoGag effectively downregulates the Ser5 K130R mutant, indicating that it can target Ser5 in the cytoplasm. Consistently, we observed that although glycoGag is classified as a type II integral membrane protein, it primarily localizes to the cytoplasm rather than the plasma membrane. Additionally, glycoGag recruits RETREG1, an ER-phagy receptor and targets Ser5 to the ER-phagy pathway. This suggests that glycoGag can clear Ser5 before it reaches the plasma membrane, offering a more effective means to antagonize its antiviral activity. However, due to its interaction with Ser5, glycoGag is also recruited to the plasma membrane, which enables it to downregulate Ser5 from the cell surface. Thus, glycoGag employs two complementary mechanisms to degrade Ser5-one via endolysosomal degradation and another through ER-phagy-demonstrating a more potent antagonism of Ser5 than Nef (Fig. 8C).

We find that, unlike Nef, glycoGag is rather short-lived, with a half-life of ~15 min, but still antagonizes Ser5 very effectively. Unlike long-lived proteins and damaged cell organelles that are cleared by autophagy, short-lived proteins are often degraded by proteasomes. Consistently, glycoGag is degraded in proteasomes but not lysosomes. These results are reminiscent of the HIV-1 virion infectivity factor (Vif) protein, which also has a very short half-life due to proteasomal degradation [28]. Vif antagonizes the apolipoprotein B mRNA editing enzyme, catalytic subunit 3 (APOBEC3) host restriction factors by targeting them to proteasomes for degradation [29]. It was reported that higher levels of Vif expression block the HIV-1 Gag processing, so the rapid Vif degradation of Vif could be evolutionarily beneficial to HIV-1 [30]. We speculate that glycoGag may also block the MLV Gag maturation, so its rapid degradation is evolutionarily beneficial to MLV.

The endolysosomal system selects extracellular cargo via endocytosis and phagocytosis for degradation. We reported that Nef internalizes Ser5 from the cell surface via receptor-mediated endocytosis, which is subsequently recruited to lysosomes via endosomes [18]. Unlike the autophagy pathways, the endolysosmal machinery remains to be fully defined. Here, we found that this Nef activity is blocked in SQSTM1-KO and PIK3C3-KO cells, highlighting a new role of SQSTM1/p62 and PIK3C3/VPS34 in endolysosomal degradation. SQSTM1 is a macroautophagy receptor that recruits polyubiquitinated cargo to the autophagosome membrane for degradation. Thus, SQSTM1 should play a similar role in this endolysosomal pathway by recruiting polyubiquitinated Ser5 proteins during this degradation process. In addition, PIK3C3/VPS34 is responsible for producing lipid phosphatidylinositol-3-phosphate (PI3P) that is required for macroautophagy and endocytosis. PIK3C3 assembles two protein complexes with different cellular proteins, and both complexes contain BECN1. Complex I produces PI3P at the phagophore, promoting the autophagosome formation, and complex II produces PI3P at early endosomes, promoting the endocytic sorting [31]. Because BECN1 is not required for Ser5 downregulation by Nef, it is unclear how PIK3C3 gets involved in endolysosomal degradation.

RETREG1/FAM134B was the first protein identified as an ER-phagy receptor [32], which belongs to the protein family with sequence similarity 134 (FAM134), including two other conserved ER-phagy receptors FAM134A/RETREG2 and FAM134C/RETREG3 [33]. They all have an N-terminal RHD with two hydrophobic domains connected by a hydrophilic loop. Each hydrophobic domain forms a hairpin within the membrane bilayer and induces ER membrane curvature. They also have an identical LIR in the C-terminal disordered region, which binds to autophagosomes. RETREG1 is also expressed as a truncated isoform RETREG1-2/FAM134B-2, which lacks the first 141 N-terminal residues and RHD [34, 35]. All these proteins are involved in starvation-induced macro-ER-phagy. Now, we show that RETREG1 is also involved in micro-ER-phagy. Because RETREG1-2 is inactive, RHD should be critical for RETREG1 to initiate this micro-ER-phagy.

Unlike macroautophagy and CMA, the mechanism for microautophagy still remains elusive. Based on the morphological changes of the lysosome, three types of microautophagy have been proposed: lysosomal protrusion, lysosomal invagination, and endosomal invagination [36]. Like CMA, the endosomal microautophagy selects proteins with a KFERQ-like motif and depends on HSPA8. However, unlike CMA, it uptakes cytosolic proteins in the late endosomes via the endosomal sorting complex required for transport (ESCRT) machinery. Notably, we find that the K130R degradation by glycoGag was blocked by 3-MA but not Ly. These two autophagy inhibitors inhibit class I PI3Ks and they also have some other cellular targets [37, 38]. We suggest that there is a 3-MA-sensitive but Ly-insensitive host factor that plays a critical role in the microautophagy. Further characterization of this unknown factor will collect critical new insights into the molecular machineries of microautophagy.

Collectively, we have uncovered a novel retrovirus antagonism of host restriction via ER-phagy, which offers critical new insights into the arms race between retroviruses and Ser5 during infection.

MATERIALS AND METHODS

Ethics statement.

This study does not include any live participants such as human subjects and animals.

Cell lines.

Human embryonic kidney (HEK) 293 cell line transformed with SV40 large T antigen (HEK293T), mouse embryonic fibroblasts cells (NIH3T3), and human cervical carcinoma cell line HeLa were purchased from American Type Culture Collection (ATCC) (CRL-3216, CRL-1658, CRM-CCL-2). The HIV-1 luciferase reporter TZM-bI cell line (ARP-8129) was obtained from the National Institutes of Health (NIH) HIV AIDS Reagent Program. HEK293T ATG3-, ATG5-, ATG7-, BECN1-, and PIK3C3-knockout cell lines were provided by Xiaojun Wang [26]. HEK293T SQSTM1-knockout cell line was reported [27]. HEK293T RETREG1-, RETREG3-, and TEX264-knockout cell lines were generated by CRISPR-Cas9 and confirmed by Sanger DNA sequencing, as we did previously [39, 40]. The oligo sequences for single-guide RNA (sgRNAs) targeting RETREG1, RETREG3 and TEX264 are 5’-CGCGGTAACCTGGCTGCTCG-3’, 5’-TTGTCTAATGCGTCGGGTCT-3’, and 5’-GATAAGTGCCGATGTGCCGT-3’, respectively. The generation of HEK293T CCPG1-, RTN3L-, ATL3-, and SEC62-knockout cell lines will be reported somewhere else [41].

All these cells were maintained in Dulbecco’s modified Eagle medium (DMEM; Thermo Fisher Scientific, 11965092) supplemented with 10% fetal bovine serum (Inner Mongolia Opcel Biotechnology) and 1% penicillin-streptomycin (pen-strep; Thermo Fisher Scientific, 10378016) and cultivated at 37°C in humidified atmosphere in a 5% CO₂ incubator.

Antibodies and inhibitors.

The mouse anti-HA (H3663), rabbit anti-FAM134B (HPA012077), and horseradish peroxidase (HRP)-conjugated anti-FLAG (A8592), anti-HA (H6533), and anti-actin (A3854) antibodies were purchased from Sigma. The mouse anti-Myc (2776S) and rabbit anti-Cullin3 (2759S) were purchased from Cell Signaling. Rabbit anti-CycK (Ab85854) was purchased from Abcam; rabbit anti-CDC2L5 (NB100-68268) and KLHL20 (NBP1-79570-20) were from Novus; rabbit anti-LAMP2a (AF1036) was from Beyotime; Alexa Fluor 594-conjugated goat anti-mouse (A11032) and Alexa Fluor 488-conjugated goat anti-mouse (A11029) were from Invitrogen. 3-Methyleadenine (M9281), LY-294002 (L9908), MG132 (C2211), and cyclohexamide (C7698) were from Sigma; bafilomycin A1 (sc-201550) was from Santa Cruz Biotechnology.

Bacterial strains.

Escherichia coli (E. coli) HB101 cells (Promega, L2011) were used to produce HIV-1 proviral vectors. These bacteria were cultured in Luria-Bertani (LB) broth (Hopebio, HB0128) on a shaking incubator at 30 °C. All other plasmids were produced from E. coli DH5α (Vazyme, C502) cells and were cultured in LB broth at 37°C on a shaking incubator. Penicillin 100 μg/mL or Kanamycin 50 μg/mL (Solarbio, A8180, K8020) were added to the medium for selective growth of transformed bacteria.

Expression vectors.

The Env and Nef-deficient HIV-1 proviral vector pNLΔEΔN (pNLenCAT-Xh) and HIV-1 Env expression vector pNLnΔBS were reported [42]. The MLV provirus vector (pNCA) was a gift from Stephen Goff (Addgene plasmid # 17363 ; http://n2t.net/addgene:17363 ; RRID:Addgene_17363) [43]. GlycoGag-deficient MLV proviral vector pNCAΔgG and pLNCX2-Luc were reported [9]. A codon-optimized glycoGag gene (accession no. J02255) was cloned into pcDNA3.1 vector via NheI/HindIII digestion, which contained a Myc- or HA-tag in the p12-coding region. pCMV-Ser5-EGFP, pCMV-Ser5-K130R-EGFP, pCMV6-Ser5-FLAG, pCMV6-Ser5-K130R-FLAG, pCMV6-Ser5-1-19K/R-FLAG, pCMV6-Ser5-S360A-FLAG, pCMV6-Ser5-S249A-FLAG, pCMV6-mSer5-FLAG, pcDNA3.1-SF2Nef-HA, and pcDNA3.1-glycoMA-HA were reported [5, 9, 20]. pEGFP-N1 vector expressing SF2Nef was constructed by PCR via KpnI/AgeI digestion. pCMV-LAMP1-mCherry was created by directly cloning the LAMP1 from pCMV-LAMP1-mGFP into pCMV vector with a mCherry tag. S360A and S249A mutations were also introduced into pCMV-Ser5-EGFP by site-directed mutagenesis. mSer5 K131R mutation in pCMV6-mSer5-FLAG and pCMV-mSer5-EGFP were also created by site-directed mutagenesis. Primers and cloning methods are available upon request.

pCMV6-mSer5-23K/R-FLAG, pCMV6-RETREG1-HA (Accession no. NM_001034850.3), pCMV6-RETREG2-HA (Accession no. NM_001321109.2), pCMV6-RETREG3-HA (Accession no. NM_178126.4), pCMV6-RETREG1-2-HA (Accession no. NM_019000.5), pCMV6-SEC62-HA (Accession no. NM_003262.4), and pCMV6-TEX264-FLAG (Accession no. NM_015926.6) were purchased from Gensoul Technology. pCMV6-mSer5-23K/R-FLAG, pCMV6-RTN3L-FLAG (Accession no. NM_001265590), pCMV6-ATL3-FLAG (Accession no. NM_015459), and pCMV6-CCPG1-FLAG (Accession no. NM_004748) were purchased from Comate Biosciences. pEGFP-N1 vector expressing CALR was constructed by PCR via XhoI/BspEI digestion. pCAGGS-CALR-mCherry was constructed by cloning CALR into the pCAGGS-mCherry vector via EcoRI digestion. pMJ920 Cas9 was a gift from Jennifer Doudna (Addgene plasmid # 42234) [44]. pGEM-T vectors expressing KLHL20 and Cul3 sgRNAs were reported previously [13]. pLKO.1-CCNKshRNA and pLKO.1-CDK13shRNA vectors were reported previously [20].

Synthesis of siRNAs.

LAMP2a (5’-GCACCCAUGCUGGAUAUTT-3’/5’-AUAUCCAGCAUGAUGGUGCTT-3’), RETREG2 (5’-ACCUCUGCCUCCCAGGUUCTT-3’/5’-GTUGGAGACGGAGGGUCCAAG-3’), and control siRNAs (5’-UUCUCCGAACGUGUCACGUdTdT-3’/5’-ACGUGACACGUUCGGAGAAdTdT-3’) were synthesized by Genesoul Technology.

Viral infectivity determination.

Nef-deficient HIV-1 for a single round of replication was produced by transfecting HEK293T cells with pNLΔEΔN and pNLnΔBS in the presence of Ser5 expression vectors using polyethylenimine (PEI) as transfection reagent. Viruses were collected after 48 hours and quantified with p24^Gag ELISA. Viral infectivity was determined by infecting TZM-b1 cells in triplicate in 96-well plates for 48 hours. Cells were lysed and luciferase activities were determined using the Firefly Luciferase Assay Kit (UElandy, F6024XL). Infectivity was calculated after being normalized by p24^Gag [45].

GlycoGag-deficient MLV was produced from HEK293T cells after transfection with pNCAΔgG and pLNCX2-Luc in the presence of Ser5 expressing vectors. After 48 hours, virions were purified by ultracentrifugation and quantified by p30^Gag via WB. Viral infectivity was determined by intracellular luciferase activity after infection of NIH3T3 cells, which was normalized by p30^Gag.

Protein stability analysis.

MLV was produced from HEK293T cells after transfection with pNCA for 48 hours and used to infect NIH3T3 cells cultured in 6-well plates. Alternatively, NIH3T3 cells were transfected with a glycoGag expression vector using Lipofectamine^® 3000 (Invitrogen, L3000-015), or HEK293T cells were transfected with a glycoGag, glycoMA, and Nef expression vector using PEI as transfection reagents. After 24 hours of infection or transfection, cells were treated with 50 μM CHX or DMSO for different times and analyzed by WB.

Western blotting (WB).

HEK293T cells were seeded either in 6-well plates or 10-cm dishes and transfected with expression vectors according to experimental design. After 24 hours of transfection cells were either lysed with RIPA buffer containing protease inhibitors (Target Mol, USA) for protein extraction or collected for extraction of membrane proteins. Membrane and Cytosol Protein Extraction Kit (Beyotime, P0033-1) were used for membrane protein extraction. Total cell lysate and membrane proteins were then applied to sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and then blotted on polyvinylidene difluoride (PVDF) membranes (Sigma, ISEQ00010). Membranes were then blocked with 5% skimmed milk powder dissolved in TBST (Tris-buffered saline) followed by incubation with primary and secondary antibodies. The blotted proteins were detected by enhanced chemiluminescence substrate (Applygen, P1010) using X-ray films (FUJI). Adobe Photoshop and Adobe Illustrator were used to generate the figures.

Immunoprecipitation.

To detect RETREG1 interactions with Ser5, Nef, and glycoGag, FLAG-tagged Ser5 or HA-tagged Nef, or glycoGag were expressed or co-expressed in HEK293T cells. After 24 hours of transfection, cells were lysed with RIPA buffer and proteins were precipitated with anti-FLAG M2 magnetic beads (Sigma, M8823) or anti-HA Affinity Gel (Beyotime P2287). Cell lysate (Input) and immunoprecipitated (IP) samples were analyzed by WB [40].

Confocal microscopy.

HeLa cells with an initial density of 1.5~2.0 ×10⁵/dish were seeded on poly-L-lysine coated coverslips and transfected with indicated vectors using Lipofectamine^® 3000. After 24 hours, cells were washed with phosphate buffer saline (PBS) and fixed with 4% paraformaldehyde for 5 minutes. Cells were then permeabilized with 0.1 % Triton X-100 for 10 min followed by 2 hours’ blocking with 10% FBS. Cells were then incubated with a mouse anti-HA antibody overnight at 4 °C, washed and coated with secondary antibodies Alexa Fluor 594-conjugated goat anti-mouse or with Alexa Fluor 488-conjugated goat anti-mouse for 1 h. Cells were washed, and the nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma, D9542) for 30 minutes followed by washing. The fluorescence signals were then visualized by confocal microscope (ZEISS, LSM880).

Mass Spectrometry.

HEK293T cells were transfected with the Ser5-FLAG, Nef-HA, and glycoGag-HA expression vectors. After 24 hours of transfection, cells were lysed with RIPA buffer. The cell lysate was incubated with anti-FLAG M2 magnetic beads or anti-HA Affinity Gel for 24 hours at 4 °C. The beads and the gel were then washed with PBS five times and mixed with protein loading buffer (5X) and electrophoresed in SDS-PAGE. The gel was then stained with Coomassie blue and sent to the Laboratory of Proteomics, Institute of Biophysics, Chinese Academy of Sciences, for Nano LC-MS/MS and database search analysis. The raw data can be downloaded from iProX (project ID: IPX0010260000).

Statistical Analysis.

All experiments were performed independently at least three times unless indicated. SPSS Statistics Software (Version 23; IBM, Inc., New York, USA) was used for the data analysis. Quantitative values of data were expressed as mean ± standard error of measurements (SEMs) and represented by error bars.