Structural Insights into the Broad-Spectrum Antiviral Target Endoplasmic Reticulum Alpha-Glucosidase II

Abstract

Targeting the host-cell endoplasmic reticulum quality control (ERQC) pathway is an effective broad-spectrum antiviral strategy. The two ER resident α-glucosidases whose sequential action permits entry in this pathway are the targets of glucomimetic inhibitors. Knowledge of the molecular details of the ER α-glucosidase II (α-Glu II) structure was limited. We determined crystal structures of a trypsinolytic fragment of murine α-Glu II, alone and in complex with key catalytic cycle ligands, and four different broad-spectrum antiviral iminosugar inhibitors, two of which are currently in clinical trials against dengue fever. The structures highlight novel portions of the enzyme outside its catalytic pocket which contribute to its activity and substrate specificity. These crystal structures and hydrogen-deuterium exchange mass spectrometry of the murine ER alpha glucosidase II heterodimer uncover the quaternary arrangement of the enzyme’s α- and β-subunits, and suggest a conformational rearrangement of ER α-Glu II upon association of the enzyme with client glycoproteins.

Discussion of Chapter 19 in Dengue and Zika: Control and Antiviral Treatment Strategies

This discussion was held at the 2nd Advanced Study Week on Emerging Viral Diseases at Praia do Tofo, Mozambique.Transcribed by Hilgenfeld R and Vasudevan SG (Eds); approved by Dr. Nicole Zitzmann.

• Aravinda de Silva: I guess this is more of a basic question about how the quality control system works. So you said that you have the sugar with one glucose and then, when it is strained, it is released and calnexin binds. Or am not i getting it?“.

• Nicole Zitzmann: Slightly. It is not possible [to happen at the same time; ie for calnexin and gluII to bind to the monoglucosylated substrate]. The ER alpha-glucosidase II has two substrates, the diglucosylated and the monoglucosylated [N-linked glycan]. And then [the question is] does calnexin get there first [i.e. to the monoglucosylated glycan], and calnexin is obviously also going on-off, on- off. And so eventually, you can’t have alpha-glucosidase cleaving the last glucose off with calnexin still bound to it, it’s physically impossible. So the glucosidase II will have to come in there at some point when calnexin is not bound.

• Aravinda de Silva: So the question is: How does the quality control work? Is it the glucosidase that actually does that?

• Nicole Zitzmann: No. The real player here is UGGT, the UDP-glucose-glucosyl transferase, the one on the other end of the calnexin cycle. That is the real quality controller that scans the protein that is trying to fold and sees wether it is folded or not. Nobody knows exactly how that works, so we wanted the structure. Now that we have the structure we can address some of these questions.

• Aravinda de Silva: So it mysteriously scans proteins in the act of folding...

• Nicole Zitzmann: Well, it possibly samples for disulfide bridges and you know that the textbooks always say that exposed hydrophobic patches can’t be there. It has a lot of thioredoxin-like domains doing all sorts of jobs.

• Aravinda de Silva: Do you know the fate of the viral proteins? What happens to them?

• Nicole Zitzmann: That entirely depends on the virus and that’s sadly, although it is broad spectrum, or at least in vitro it is antiviral against all of these viruses, why it totally depends on the virus. Say, in HIV, you’ll get a slight misfolding of the gp120 and the virus can still get secreted, but you have a slight misfolding in the V1 and V2 loop. The virus can still bind to the next cell and it can bind to CD4, but it cannot make the conformational change to bind to the coreceptors. So in HIV you get virus out, so if you just measure the RNA, you wouldn’t know any better, but you have a totally non-infectious virus. Other viruses see gross misfolding. For dengue, I think Joanna [Miller] will show some data. It looks like we actually retain them. And if you retain things inside the cell, they eventually will be degraded by the ERAD pathway. And for most other viruses, you’ll get grossly misfolded [viral envelope glyco] proteins inside the cell and they get degraded inside the cell.

• George Gao: Are there any evidences for quality control factors or putative factors such as your enzyme being upregulated in the ER during virus infection like HIV or Ebola?

• Nicole Zitzmann: The enzymes of the calnexin cycle? Yes they are. So when you have a viral infection, there are thousands of viral proteins trying to fold in addition to the normal host proteins that are trying to fold. So yes, the cell senses this extra burden and triggers the unfolded protein response, the UPR response. This can lead to the upregulation of the main players I mentioned.

• George Gao: So what happens for a chronic infection like HIV? Is it also upregulated?

• Nicole Zitzmann: It depends on wether you have active viral replication in the cell. So if you have active viral replication it will be upregulated. But HIV is often latent.

• Laura Rivino: So you are inhibiting an enzyme that is a host enzyme. And so what is the effect on the host?

• Nicole Zitzmann: It is actually not very dramatic. We have a large therapeutic window with alpha glucosidases and we hit the viruses much earlier, before we hit the host proteins. The reason for this is so far only speculative, but we have a paper hopefully coming out soon where we have proven that for HIV. Most viruses need to oligomerize their envelope glycoproteins. For instance, gp120 in HIV is a trimer. So even if you only slightly misfold one of them, the whole trimer cannot work anymore and for HIV, we know that actually for viral fusion, you’ll need several, about seven, of these to come together. So you can actually have a massive amplification effect with only a little bit of misfolding and that is why we hit the viruses before we hit the host. Most host enzymes can also use endomonosidases in the Golgi to get around these problems but viruses can’t. So you could speculate if it has something to do with the fact that viral proteins have not necessarily, at least initially, co-evolved with our cells. They really need a lot of help to fold their proteins in the ER.

• Siew Pheng Lim: So you mentioned the alpha-glucosidase inhibitors have been repeatedly looked at for HIV, HBV, HCV etc. What are your thoughts in terms of what makes a better inhibitor and which angle are you are aiming at?

• Nicole Zitzmann: So personally, my group is after an allosteric inhibitor at the moment for this enzyme, but there are a lot of approaches in this field and we are not the only players in this. There is a lot of medicinal chemistry going on and having determined the structure you know there are some very obvious things that one would be able to do.