Conserved residues of the immunosuppressive domain of MLV are essential for regulating the fusion-critical SU-TM disulfide bond

Abstract

The Env protein of murine leukemia virus (MLV) is the prototype of a large clade of retroviral fusogens, collectively known as gamma-type Envs. Gamma-type Envs are found in retroviruses and endogenous retroviruses (ERVs) representing a broad range of vertebrate hosts. All gamma-type Envs contain a highly conserved stretch of 26-residues in the transmembrane subunit (TM) comprising two motifs, a putative immunosuppressive domain (ISD) and a CX₆CC motif. Extraordinary conservation of the ISD and its invariant association with the CX₆CC suggests a fundamental contribution to Env function. To investigate ISD function, we characterized several mutants with single amino acid substitutions at conserved positions in the MLV ISD. A majority abolished infectivity, although we did not observe a corresponding loss in intrinsic ability to mediate membrane fusion. Ratios of the surface subunit (SU) to capsid protein (CA) in virions were diminished for a majority of the ISD mutants, while TM:CA ratios were similar to wild type. Specific loss of SU reflected premature isomerization of the labile disulfide bond that links SU and TM prior to fusion. Indeed, all non-infectious mutants displayed significantly lower disulfide stability than wild-type Env. These results reveal a role for ISD positions 2, 3, 4, 7, and 10 in regulating a late step in entry after fusion peptide insertion but prior to creation of the fusion pore. This implies that the ISD is part of a larger domain, comprising the ISD and CX₆CC motifs, that is critical for the formation and regulation of the metastable, intersubunit disulfide bond.IMPORTANCEThe gamma-type Env is a prevalent viral fusogen, found within retroviruses and endogenous retroviruses across vertebrate species and in filoviruses such as Ebolavirus. The fusion mechanism of gamma-type Envs is unique from other Class I fusogens such as those of influenza A virus and HIV-1. Gamma-type Envs contain a hallmark feature known as the immunosuppressive domain (ISD) that has been the subject of some controversy in the literature surrounding its putative immunosuppressive effects. Despite the distinctive conservation of the ISD, little has been done to investigate the role of this region for the function of this widespread fusogen. Our work demonstrates the importance of the ISD for the function of gamma-type Envs in infection, particularly in regulating the intermediate steps of membrane fusion. Understanding the fusion mechanism of gamma-type Envs has broad implications for understanding the entry of extant viruses and aspects of host biology connected to co-opted endogenous gamma-type Envs.

Keywords: envelope glycoprotein; fusion; gamma-type Env; immunosuppressive domain; infectivity; murine leukemia virus.

Fig 1

Gamma-type Envs are shared across multiple genera in the retroviridae and beyond and contain a distinctive disulfide bond between SU and TM. (A) Retroviral Env sequence showcasing common features of murine leukemia virus (MLV) and related gamma-type Envs. The location of the receptor binding domain is based on MLV; however, the location of this region can be highly variable and is not completely defined for many Envs. Of note are two cysteine motifs that are joined via a disulfide bond (shown in pink). HR1 and HR2 correspond to the heptad repeats. ISD corresponds to the immunosuppressive domain. (B) Geneious Prime (v2022.2.2) generated neighbor-joining tree of envelope proteins using an alignment of TM only, colored by genus except for ebolavirus and syncytin-1. Bootstrap values are shown at branch points. The asterisk corresponds to the gammaretrovirus used for the experiments in this study.

Fig 2

Conservation of the ISD and CX₆CC motifs across highly divergent Env sequences. (A) Alignment of 37 whole or partial gamma-type entry protein sequences. Percent identity at each position is shown in the histogram (top). Black boxes indicate aligned residues, while horizontal lines indicate gaps in the alignment. Low overall identity is observed for SU and higher for TM. Alignment was generated using clustal omega as implemented in Geneious Prime (v2022.2.2). (B) The region with the highest conservation and longest unbroken span is shown zoomed in and contains the 26 amino acid stretch that comprises the ISD and CX₆CC. The diversity of sequences used within the alignment represents hundreds of millions of years of evolution for both viruses and hosts.

Fig 3

MLV ISD mutants are correctly processed but largely incapable of infection. (A) Mutagenic strategy to create eight single mutants and two double mutants to the most conserved regions. *Mutant E14R + A20F is not within the most conserved residues but instead replicates a mutant used previously in the literature (34, 36) believed to have no immunosuppression. Logo plot was based on alignment from Fig. 2 showing the conserved residues of the ISD. Leucine 1 of the ISD corresponds to L538 in Moloney MLV (accession NC_001501.1). (B) Expression of Env in whole cell lysates via reducing western blot of all ISD mutants. All Envs were C-terminally tagged with an Avi epitope tag for visualization. To specifically observe Env expression and processing, constructs encoding Env were transfected without Gag-Pol. Blots were stained with a rabbit anti-Avi polyclonal antibody to visualize TM and precursor. Env processing is demonstrated by the presence of cleaved TM at ~17 kD. A furin deletion mutant was used as a negative control. (C) Infectivity of each Envelope mutant pseudotyped with a Mo-MLV core and co-transfected with a packageable GFP plasmid. Virus was produced in 293T cells, and murine NIH-3T3 fibroblasts were used as target cells. GFP expression in target 3T3 cells, as measured by flow cytometry, was used as a measure of infection. Three infections, each performed in triplicate, are shown per condition. Data points are colored by infection round (**** =P < 0.0001, ns = not significant, ordinary one-way ANOVA with multiple comparisons).

Fig 4

MLV ISD mutants are still fusogenic. (A) Cell-to-cell fusion ability of all Env mutants as measured via flow cytometry. To facilitate fusion, the R-peptide was removed from all Envs. HEK293T cells were transfected with either MLV Env in an IRES-GFP expression construct (ecoMLV-EnvAvi-pCGCG), the receptor MCAT1 in an IRES-mCherry expression construct (pMSCV-MCAT1-IRES-mCherry), or an empty mCherry vector (pMSCV-IRES-mCherry). Two days post-transfection, cell populations were resuspended and combined in flow cytometry tubes such that each Env was combined with MCAT1 and an empty vector. After 20 minutes, cells were pelleted and fixed in 2% paraformaldehyde, then sent for flow cytometry analysis. All events that were dual positive for both mCherry and GFP were considered recently fused cells. An empty GFP vector was used as a no-Env control. A line is drawn at this no-Env control for visual comparison. (B) Cell-to-cell fusion as performed in A but with the R-peptide intact. An ordinary one-way ANOVA with multiple comparisons was performed for A and B (**** =P < 0.0001, *** =P < 0.001, ** =P < 0.01, * =P < 0.05, ns = not significant). (C) Representative fluorescent images of syncytia formation between GFP and mCherry expressing cell populations, corresponding to wild-type MLV Env and a no Env control.

Fig 5

Incorporation of MLV Env ISD mutants into MLV particles. (A) Incorporation of Envelope mutants as seen on a western blot of viral pellets. Samples were treated with 2-mercaptoethanol to reduce all disulfide bonds for clarity in resolving TM. MLV SU was visualized with an anti-MLV-SU antibody, TM was visualized with the same anti-Avi stain used previously, and capsid was imaged using an anti-p30 antibody. (B) Western blot of viral pellets to measure incorporation of envelope mutants treated with a protease inhibitor, amprenavir, at 0.8 mM. MLV SU was visualized with an anti-MLV-SU antibody, TM was visualized with the same anti-Avi stain used previously, and capsid was imaged using an anti-p30 antibody. The blots in A and B are representative and correspond to the blue-colored data points in each corresponding graph, panels C, D, E, and F. (C) Ratio of TM to Capsid across three experiments (points colored by experiment). Densitometry was used to quantify levels of CA and TM for three sets of viral pellets. No significant difference was measured between wild-type and all mutant conditions. (D) Ratio of SU to capsid across three experiments (points colored by experiment). Densitometry was used to quantify levels of CA and SU for three sets of viral pellets. (E) Ratio of TM to capsid across three experiments treating virus with a protease inhibitor (points colored by experiment). Densitometry was used to quantify levels of CA and TM for three sets of viral pellets, using only the 30 kD capsid band for quantification. No significant difference was measured between wild-type and all mutant conditions using an ordinary one-way ANOVA with multiple comparisons. (F) Ratio of SU to capsid across the same three experiments as in E with points colored by experiment (** =P < 0.01, * =P < 0.05, ns = not significant. Ordinary one-way ANOVA with multiple comparisons).

Fig 6

ISD mutations do not prevent the formation of the disulfide bond between SU and TM. (A) Representative blot of NEM-treated whole cell lysates of MLV ISD mutants, stained with anti-Avi to detect TM either in precursor or in disulfide-bonded Env. Disulfide-bonded Env migrates to around 100 kD, precursor protein to around 80 kD, SU alone would migrate to 70 kD but cannot be visualized with the Avi antibody. (B) The amount of disulfide-bonded Env for each mutant was quantified relative to the Env protein present in the lysate for four independent experiments (ns = not significant, one-way ANOVA with multiple comparisons).

Fig 7

MLV ISD mutants have a less stable disulfide bond. (A) Schematic of conserved cysteine containing motifs and formation/isomerization of the disulfide bond between SU and TM subunits. (B) Non-reduced viral pellets were separated with SDS PAGE and visualized with western blot. NEM treatment was used to block isomerization of the disulfide bond, essentially fixing any bonded Env present in the samples. Non-treated samples were maintained in non-reducing conditions to preserve intact disulfide bonds. Mutant AWLC no longer has the free cysteine needed to isomerize and acts as a size control for the disulfide-bonded form of Env. Blots were stained with an anti MLV-SU antibody to enable visualization of disulfide-bonded Env and SU alone. (C) Densitometry of the disulfide bonded band compared to total protein present for untreated supernatant samples across five experiments (representative blot shown in panel B). Points are colored by individual experiments. (D) Densitometry quantification of three western blots of MLV Env ISD mutants combined with an AWLC mutation to block isomerization. Samples were harvested in non-reducing conditions and separated with SDS PAGE. Precursor and disulfide-bonded Env were detected with an anti-Avi antibody. Ratios of disulfide-bonded Env to precursor are shown for a selection of ISD mutants. Points are colored by individual experiements (ns = not significant, one-way ANOVA with multiple comparisons).

Fig 8

Structure of the pre-fusion Ebola GP trimer highlighting highly conserved residues of the ISD. (A) Top-down view of the central axis of the pre-fusion crystal structure of Ebola GP (6QD8). Side chains are shown for residues 1, 2, 3, 4, 7, 8, and 10 of the ISD. (B) Zoomed-out view of panel A. (C) Single heterodimer (GP1-s-s-GP2) with side chains shown as in A. The intersubunit bond is indicated with an arrow to label the linkage between GP1 and GP2, while the intrasubunit bond that forms between the first two cysteines of the CX₆CC is called out by a star. (D) Zoomed-out view of panel C showing the full Ebola GP trimer. (E) Top-down view of the alpha helix formed by the ISD. Side chains are shown for residues 1, 2, 3, 4, 7, 8, and 10. For clarity, only one heterodimer is shown.