Last time I talked about immune evasion, I said how disconcerting it was to learn that immune evasion genes of mouse cytomegalovirus really don’t seem to have much of an impact on viral pathogenesis (except for the ability to infect tonsils and therefore, perhaps, spread between mice). I was equally disconcerted by a relatively old paper that I only recently found.
Adenoviruses are medium-sized DNA viruses that often establish long-term persistent infections.1 These persistent infections aren’t truly latent (like herpesviruses), nor are they “chronic”, like hepatitis C or HIV, which continue to replicate robustly for a long time; instead, adenoviruses seem to be able to linger around in what seems to be a low-level active infection, without doing any particular harm. As far as I know, which isn’t all that far, the mechanism by which adenoviruses are able to do this isn’t well understood. There are many different strains of adenoviruses, over 40 for humans alone, and they tend to have different lifestyles and infect different tissues in different ways. The various types of human (and some other) adenoviruses are grouped into subgroups A through F, based on sequence similarity. Research-wise, the most popular subgroup by far is Subgroup C, including types 2 and 5; these are the types that most commonly are used as vectors for vaccines and gene therapy.
The first example of viral T cell immune evasion to de described, as far as I know, was in adenovirus: 2 A glycoprotein in the E3 genomic region (“E3gp19k”) binds to MHC class I and prevents it from reaching the cell surface, thus presumably reducing the ability of CTL to recognize the infected cell. (It does other things as well, but we won’t go into that now.) About ten years ago, Linda Gooding’s lab looked at the effects of deleting this gene on mouse infection;3 their answer was that it doesn’t really do much of anything (“the E3 protein gp19K alters neither afferent nor efferent immune responses”). However, human adenoviruses don’t replicate well, if at all, in mice, so it’s not really a very good model, and so I wasn’t very distressed by that observation. I comforted myself with the knowledge that E3gp19k is quite conserved in its sequence, so that it must be strongly selected for.
Oddly, though, while E3gp19k sequences are conserved within adenovirus subgroups, between the different subgroups they’re really quite different. (Phylogenetic tree on the left shows the clustering of E3gp19k among a number of adenovirus subgroups.) So something keeps the protein sequence similar among similar viruses — it’s strongly selected — but between subgroups it can diverge quite widely. I’m not sure what that means. Presumably this reflects something about the different virus lifestyles, but I don’t know enough about the different lifestyles to make any sweeping generalizations; nor is there anything know as yet about any functional differences (if any) between the various subgroups’ E3gp19k.
In fact, some human adenovirus subgroups don’t even seem to have any E3gp19k at all. I assumed that these viruses either have a very divergent variant of E3gp19k, that I wasn’t picking up in my Blast searches, or else that they have evolved some other form of T cell immune evasion molecule. However, that assumption has been deeply shaken by the paper I mentioned:
Lack of effect of mouse adenovirus type 1 infection on cell surface expression of major histocompatibility complex class I antigens.
Kring SC, Spindler KR.
J Virol. 1996 Aug;70(8):5495-502.
Here, they’re looking at mouse adenovirus rather than human. This virus, like some of the human subgroups, apparently has no E3gp19k protein in its genome. Spindler — probably guessing, like me, that the virus must have some other protein with a similar function — put the virus through a pretty exhaustive battery of experiments, and quite conclusively show that at least in vitro the virus doesn’t do a damn thing to MHC class I. So, either the virus does something in vivo that it doesn’t do in vitro:
One model is that a unique cell type exists in which MAV-1 infection down regulates class I MHC surface expression. … While endogenous expression of MAV-1 E1A activity did not induce a decrease of class I MHC antigen levels in MAV-1-infected cells in our studies, it remains possible that we did not analyze the unique cell type in which this mechanism could operate.
Or else the virus doesn’t care about avoiding CTL, and if that’s the case, it strengthens the case against immune evasion molecules (even in viruses that express them) being all that important for pathogenesis:
An alternative model is that decreased class I MHC surface expression is not important to the ability of MAV-1 to persist and that another mechanism is active … the demonstrated ability of the E3 gp19K protein to decrease the surface expression of class I MHC antigens in vitro may not be indicative of an ability to significantly alter the cell-mediated immune response to an infection in vivo.
My bias, for a long time, has been that viruses like herpesviruses and adenoviruses must care deeply about evading CTL. I’m going to have to rethink that bias.
- They’re kind of like the poor man’s herpesvirus.[↩]
- An adenovirus type 2 glycoprotein blocks cell surface expression of human histocompatibility class I antigens. Burgert HG, Kvist S. Cell. 1985 Jul;41(3):987-97.[↩]
- The role of human adenovirus early region 3 proteins (gp19K, 10.4K, 14.5K, and 14.7K) in a murine pneumonia model. Sparer TE, Tripp RA, Dillehay DL, Hermiston TW, Wold WS, Gooding LR. J Virol. 1996 Apr;70(4):2431-9.[↩]
I am everything but an expert in viral immune evasion, but I have always naively assumed that this mechanism (using dedicated evasion proteins) is restricted to big-genome viruses, such as herpes- or poxviridae. This (probably wrong) idea had been fueled by detecing several big-virus genes that have been obviously hijacked from the host genome and typically now fulfil a decoy role.
In your post, you raise the interesting question if some adenoviruses can do without a gp19k or if it is too divergent to be recognized. I have briefly looked into that issue using our sequence comparison methods, which are more sensitive than BLAST. I got mixed results: there are very divergent gp19k versions in class F adenoviruses (human ADV41 and ADV52, simian ADV1) but I was not able to find anything in murine adenoviruses. The class-F gp19k homologs (typcially called RL2 or CR1-beta1) are not only divergent but also have an additional Ig-like domain at the N-terminus. The similarity is restricted to the ectodomain of gp19k, which probably also resembles an Ig-fold.
This does not answer the question if immune evasion is generally important for viruses. I would guess that it is, but probably using a variety of mechanisms, not all of which can be recapitulated in vitro.
The large-genome viruses have more immune evasion genes than the small ones, but that’s sort of a tautology. Some forms of immune evasion are nearly universal. HIV, for example, which I’d call a medium-sized virus, uses nef to downregulate a bunch of immune-related molecules including MHC class I. Polioviruses, which are really tiny, encode a protein that blocks ER-surface transport, and it’s been suggested that this is at least partly an immune evasion gene since it reduced MHC class I surface expression. Outside cell-mediated immunity, things like cytokine evasion are really common, with fairly small viruses like orthomyxoviruses getting all medieval on interferon’s ass.
I did find the AdF CR1 as weakly homologous to gp19k. CR1 is supposedly a complement receptor, and other ad subgroups have CR1s as well. At the time I tried comparing the putative AdF CR1 to other CR1s and to gp19ks and at the time I convinced myself that it was a more plausible CR1 than gp19k; I didn’t keep notes on how I did the comparisons or what the similarities were, and quickly re-running those comparisons I’m not so convinced any more; the similarity to other CR1 is not great either. I don’t know the basis on which the AdF gene was termed CR1 — whether there was a functional basis for it. In any case, it’s certainly way divergent.
To clarify a point, I think it’s well established that immune evasion in general is important to viruses; I’m restricting my question here to the specific subset of immune evasion that supposedly reduces MHC class I-restricted recognition. There are some very impressive experiments showing that some immune evasion genes, especially cytokine inhibitors and decoys, are very important for virus pathogenesis, and that’s true for a wide range of viruses.
I have observed something similar, but after having a closer look I think that these similarities are based on different parts of the sequence. The AdF CR1 proteins have two domains, a V-set like Ig domain (which has similarity to CR1 proteins from other adenovirus classes but not to gp19k) and a 2nd extracellular part that is related to gp19k but not to other CR1 proteins. I would interpret this as a kind of fusion of CR1 and gp19k into one protein for the AdF viruses. By the way, are you sure that the CR1 proteins are supposedly complement receptors? From reading the database entries and a little literature, the “CR1″ probably just means “conserved region 1″.
Interestingly, the Ig domain of the adenovirus CR1s is related to several CMV genes that are involved in cell tropism – maybe they have a similar role in the adenoviruses. It would seem a strange idea to have a cell tropism domain and an MHC evasion domain on the same polypeptide. Without having any real evidence, I would guess that the gp19k proteins mainly do something different from messing with MHCs.
Without having any real evidence, I would guess that the gp19k proteins mainly do something different from messing with MHCs.
As far as I can tell from the literature, no one has really looked at E3gp19k function in most adenovirus subgroups other than subgroup C. (I won’t swear there’s nothing in the literature, though. Searching for “A” or “F” in publemd doesn’t really help narrow the field much.) I have a student who is running through representatives of the various subgroups (fairly superficially, though) and we’ll see if we can figure out something.
It would seem a strange idea to have a cell tropism domain and an MHC evasion domain on the same polypeptide
After a cup of coffee, it occurs to me that more likely than a fused polypeptide is that the genome annotators missed a splice site, and that these are two distinct proteins. (As I’m sure you know, adenoviruses are splicing nightmares.)
[...] Immune evasion: Who needs it? [...]
[...] reticulum. Immunity 3, 207-214.[↩]At least, that’s the accepted wisdom — but see my previous discussions of that.[↩]King, N. J., Maxwell, L. E., and Kesson, A. M. (1989). Induction of class [...]
[...] protein that’s long been shown to block MHC class I antigen processing. In particular, mouse adenovirus 1 does NOT block antigen presentation, at least as determined in one careful study. And while human adenoviruses will just about infect [...]
[...] MHC class I didn’t seem to do all that much. I’ve summarized some of those experiments here and here. For example, the MHC class I immune evasion genes in adenoviruses and in mouse [...]
[...] MHC class I didn’t seem to do all that much. I’ve summarized some of those experiments here and here. For example, the MHC class I immune evasion genes in adenoviruses and in mouse [...]