Viral immune evasion strategies: the human cytomegalovirus US2 immunoevasin and protein degradation from the endoplasmic reticulum

Relatively small genomes and high replication rates allow viruses and bacteria to accumulate mutations, continuously presenting the host immune system with new challenges. On the other side of the trenches, an increasingly well-adjusted host immune response, shaped by co-evolutionary history, makes a pathogen’s life a rather complicated endeavor. It is, therefore, no surprise that pathogens either escape detection or modulate the host immune response, often by redirecting normal cellular pathways to their advantage 1 . Persistent viruses of the Herpesviridae family such as human cytomegalovirus (HCMV) can cause a life-long latent infection, with controlled sporadic reactivation in the face of a fully primed host immune system. This viral survival strategy is critically dependent on immune evasion proteins or “immunoevasins”, viral gene products committed to subvert the host immune system to escape elimination 2-4. The class I Major Histocompatibility Complex (MHC) antigen presentation pathway is utilized by the immune system for sampling of the intracellular environment 5 . Peptide antigens derived from cytosolic proteolysis are loaded on newly synthesized class I MHC molecules in the endoplasmic reticulum (ER), after which the trimeric class I MHC complex - composed of the type I membrane glycoprotein class I MHC heavy chain (HC), the soluble light chain, and a short antigenic peptide of 8 to 10 aminoacids – travels through the secretory pathway to the cell surface, where it is subjected to scrutiny by the immune system. When a cell is virally infected, cytosolic proteins of viral origin, like endogenous cellular proteins, are subjected to proteasomal degradation, and thus virus-derived peptides access the class I MHC antigen presentation pathway. Cytotoxic T lymphocytes (CTLs) inspect surface class I MHC products for the presence of virus-derived peptide antigens, whereas natural killer (NK) cells score for perturbations in expression of class I MHC molecules at the surface of the virus-infected cell. Detection of either – both of which can be caused by viruses - can lead to activation of the CTL and NK cell cytolytic program, ultimately resulting in destruction of the cell harboring the virus 6 . Because it is involved in sampling of the intracellular environment inhabited by viruses, the class I MHC antigen presentation pathway presents a formidable challenge for viral pathogens and is, not surprisingly, the target of many viral immune evasion strategies, particularly by herpesviruses 4,7-11. HCMV alone encodes several immunoevasins that interfere with the class I MHC antigen presentation pathway 3 . The HCMV unique short region (US) genes US2 and US11 encode two ER-resident type I membrane glycoproteins capable of catalyzing the transport of newly synthesized class I MHC HC molecules through the ER membrane to the cytosol, leading to HC degradation by the cytosolic proteasome. This ER-to-cytosol transport, as it constitutes movement of proteins in the reverse direction of translocation of nascent polypeptide chains into the ER 12,13, is termed retrotranslocation or dislocation. HCMV US2 and US11-induced HC dislocation from the ER membrane and subsequent HC degradation allows HCMV to avoid cell surface display of viral peptides and thus evade the host immune surveillance afforded by the class I MHC pathway. The virus-induced dislocation process bears many similarities to cellular ER quality control mechanisms, which ensure the fidelity and regulation of protein expression during cell life and differentiation 14,15. After insertion into the ER membrane or lumen, proteins that fail to fold or assemble correctly are destroyed. Misfolded or misassembled ER proteins, or ER proteins that are subject to degradation in response to environmental cues, are dislocated or retro-translocated into the cytosol, where deglycosylation, ubiquitination, and ultimately proteasomal proteolysis dispense with the defective polypeptides 15. The machinery involved in the extraction of misfolded proteins from the ER is poorly defined. The degradation of class I MHC molecules not only endows HCMV with the capacity to evade the immune system 12,13 but also constitutes a model system for the study of dislocation as a cellular process of general importance, for example, in the context of diseases caused by accumulation of misfolded proteins in the secretory pathway or premature degradation of proteins with slow folding kinetics, or intoxication by some bacterial toxins and viruses 16-19. US2- and US11-mediated dislocation of class I MHC heavy chains is unique in terms of the speed and specificity of degradation 4,12,13,20 and because HCs are neither misfolded nor degraded as a result of normal cellular signals like most known substrates, but as a result of the expression of either viral protein. Notwithstanding the unique nature of this virus-mediated process, the US2 and US11 systems have allowed the characterization of important mechanistic details of the dislocation reaction: the study of the US11 pathway, in particular, led to the discovery of previously unknown components of the cellular dislocation machinery, the DERLIN proteins, that have since been shown to be involved in ER dislocation more generally 21-26. Although targeting the same substrate (HCs) and inducing the same outcome (HC degradation), the two viral immunoevasins have been shown to differ in many of the mechanistic details of the dislocation reaction, such as which class I MHC alleles are targeted by each viral product or what the HC cytoplasmic segment (cytoplasmic tail) length and sequence requirements each displays. Here we show that US2 and US11 also differ in their ability to target distinct folding stages of class I MHC products. We generated an unfolded HC molecule through site-directed mutagenesis of the cysteine residues that form the membrane-proximal alpha3 domain disulphide bond, and show that only US11 is able to catalyze its degradation 27. The fact that US11 targets HC molecules independently of their folding/assembly status while US2 targets only folded class I MHC structures suggests that US11 and US2 act at different stages of the class I MHC biosynthetic pathway. The differences between the US2- and US11-mediated dislocation processes extend to the cellular machinery that each utilizes: US11 uses its transmembrane domain to recruit HCs to a human homologue of yeast Der1p, Der1-like protein 1 or DERLIN-1, a protein essential for the degradation of a subset of misfolded ER proteins. Interestingly, DERLIN-1 is essential for the degradation of HC molecules induced by US11, but not by US2 21. In fact, proteins essential for US2-dependent HC dislocation have remained elusive. We conducted a screen for such proteins by comparing interacting partners of wild type (active) US2 with interacting partners of a dislocationincompetent US2 mutant that lacks the US2 cytosolic tail. We identified signal peptide peptidase (SPP) as a specific partner for the active form of US2. We showed that reduction of SPP levels by RNA interference (RNAi) inhibits HC dislocation by US2, whereas the same knockdown of SPP has no effect on US11-mediated dislocation of HC. Moreover, we showed that while US2 binds SPP, it does not bind DERLIN-1; conversely, US11 binds DERLIN-1 but not SPP. Taken together, our data suggests that the two viral immunoevasins utilize distinct cellular machinery and possibly subvert two distinct pathways of dislocation from the mammalian ER. SPP is an ER-resident presenilin-type intramembrane-cleaving aspartic protease which, as the name suggests, cleaves signal peptide remnants resulting from prior signal peptidase processing from the ER membrane 28. The first role assigned to SPP was the generation of HLA-E epitopes, i.e., peptide fragments derived from intramembrane cleavage of the signal sequence of polymorphic class I MHC molecules such as HLA-A and HLA-B. HLA-E presentation is important for NK cell surveillance as it allows NK cells to detect perturbations in synthesis of polymorphic class I MHC products 29. The fact that reduction of SPP levels by RNAi inhibits HC dislocation by US2 implicates SPP in the US2 dislocation pathway and suggests a novel role for this intramembrane-cleaving aspartic protease in ER dislocation. We also present evidence for a role of the p97 protein in HC dislocation by US2. The trimeric Cdc48p(p97)/NPL4/UFD1 complex, composed of the AAATPase p97 (Cdc48p in yeast) and its co-factors NPL4 and UFD1, is involved in recognition and/or extraction of polyubiquitinated ERAD substrates from the ER membrane, in a step that precedes their proteasomal degradation 30,31. Although the involvement of p97 in US11- mediated dislocation had been shown previously in semi-permeabilized cells 32, we confirmed the p97 requirement in intact US11 cells: reduction of p97 levels by RNAi leads to impaired HC degradation in US11 cells. The same p97 knockdown also impairs US2-mediated HC dislocation, constituting the first evidence for a role of the p97/NPL4/UFD1 complex in the US2 pathway. This convergence of the US2 and US11 pathways, presumably at a pre-proteasomal step catalyzed by the p97/NPL4/UFD1 complex, occurs notwithstanding the seemingly non-overlapping SPP/DERLIN-1 routes of substrate recognition/targeting to dislocation prior to extraction from the ER membrane, reinforcing the notion of the existence of multiple ER dislocation pathways. Furthermore, we show that the US2 cytosolic tail is not only necessary but also sufficient for recruitment of SPP. We generated mutant US2 molecules lacking one or more US2 sequence/structural elements to analyze the contribution of the US2 lumen, the US2 transmembrane domain (TMD) and the US2 tail to association with SPP; a “mini-molecule” targeted to the ER membrane and containing only the tail aminoacids of the US2 molecule is still capable of binding to SPP. More importantly, while investigating the contribution of the US2 TMD we found that the US2 TMD is dispensable for association with SPP, yet crucial for US2 function in HC dislocation: a chimeric US2 molecule in which the US2 TMD has been replaced with that of an unrelated type I membrane glycoprotein is inactive in HC dislocation, although retaining the ability to bind the HC substrate and SPP. This has allowed us to refine the model for US2’s role in dislocation from the ER membrane: HC dislocation is dependent not just on US2 tail-mediated recruitment of SPP but also on additional US2 TMD-mediated interactions within the plane of the membrane. We hypothesize that the US2 TMD is involved in further engagement of SPP or in the recruitment or engagement of other protein(s) involved in dislocation. Finally, we discuss the implications of our findings in viral protein-induced dislocation to cellular ER quality control.