Lyme disease is no fun, and in some people it’s a lot worse than no fun. We know a fair amount about it – the main event is that deer ticks spread the bacterium Borrelia burgdorferi when they bite humans. For most people, it’s an acute infection that goes away, and which can be treated with antibiotics. Famously, though, there are people who show longer-term consequences, which are likely to be due (at least partly) to some sort of autoimmune response set off during the initial infection. There are plenty of examples of this sort of thing with both bacterial and viral infections. But what is it about that bacterium that causes trouble?
We are once again in the thickets of human immunology with these questions, of course. Here’s a new paper with some details that no one up until now has appreciated. Over the years, there have been several candidates for the bacterial “immunogen” that we are reacting to. The front-runner has been outer surface protein A (OspA), a lipid-conjugated bacterial protein that has been the subject of several vaccine development efforts. There are another compounds, though, that also seems to set off an immune response, in particular a glycoglycerolipid (BbGl-II) that was first characterized in detail about twenty years ago. As the new paper under discussion mentions, though, the follow-up to that work has been confusing. People have synthesized the compound but found it to be not very active as an immunogen, leading to various hypotheses about the disconnect in the various data sets.
This latest work provides a concise answer for that problem: the structure has been wrongly assigned all these years. It’s a diacylglycerol attached to a galactose, all right, but it’s a galactofuranose rather than the expected galactopyranose. For those of you who aren’t into carbohydrate chemistry, that basically means that the sugar is in a five-membered-ring form rather than a six-membered one. Carbohydrates are fun that way, for some definitions of “fun” – I mean, I did my PhD work using them as chiral starting materials for synthesis, so I can take them just fine, but they can give other chemists headaches trying to keep track of all the variations. For example, the five-membered and six-membered forms of a given sugar each have two other possible isomers of their own (alpha and beta) at their C1-carbons, but I think I’ll save that for a separate more carbohydrate-focused post, if anyone can stand it.
At any rate, the furanose form is almost certainly not the thermodynamically favored one in this molecule, which is why no one assumed that it was the real structure, and the usual proton NMRs of the two forms are going to be rather hard to distinguish. Carbon-13 NMR is more diagnostic – this work provides all the characterization data, but as the authors note, you’d still want to have authentic synthesized samples of all the possible forms to be really sure, as they went to the trouble of making.
And although that might not be much of a difference in the NMR or in LC/MS, it can be a big one to the immune system. Indeed, comparing the alpha-galactrofuranose, the beta-galactofuranose, and the previously assigned alpha-galactopyranose in cell assays clearly shows that it’s the first one that sets off a response. Further experiments shows that this is through the TLR2 system, which is what you’d expect, since that one is known to be sensitive to all sorts of foreign lipid-containing molecules. Lyme disease researchers in this area have indeed been working on the wrong structure for twenty years now. And now that this has been cleared up, further vaccine development and the search for the human target that gets mis-identified by the immune system should both be that much more clear. This isn’t the only case where foreign galacto-lipid molecules are known to be trouble, either, and it makes you wonder how many of those have been mischaracterized. Good to see!
