One more human glycosyltransferase

We’ve found a gene, encoded for by most animals that we think is another mannosyltransferase. For a protein glycosylation pathway that we were pretty sure involved two enzymes, and one major substrate only a few years ago, it’s now taking up a fair amount of space on our pathway figures. The name of the gene is TMEM260. Unlike the other new enzymes that we found, it doesn’t look like there’s an opportunity to backronym² meaning into this name, so there’s likely a renaming due for this at some point. You can read about it at our preprint ³ that we dropped at Research Square paper ⁴ in PNAS. I’ll give some background information here that didn’t sit properly in the manuscript.

Lots of guilt by association for heart defects

Soon after we decided to try and figure out if TMEM260 encoded for a glycosyltransferase, a publication came out reporting on mutations in this gene in patients with developmental defects ⁵ in the kidney, heart and brain. In our paper, we’ve got a lot of new data that sheds some more light onto these phenotypes, but there’s also many individual data points that also point us in that direction. We know from our glycoproteomics that this mannosyltransferase targets Plexin D1, and there’s a match between one of the phenotypes (“Truncus arteriosus”) and phenotypes arising from a deficiency in PLXND1 ⁶ . If you dig into the phenotypes for CDG-Ie, and PMM2-CDG, you’ll find descriptions of heart defects (patent ductus arteriosus and truncus arteriosus) ^{7 8 9} (via ¹⁰ ). Looking upstream from the plexins,

GATA6 has been implied ¹¹ to regulate semaphorin/plexin signalling via regulation of Plexin A2 and SEMA3C.

Beyond these more well supported phenotypes, there’s very sparse information from when TMEM260 was known by it’s street name of C14orf101, that links it to non-Hodgkin lymphoma survival ¹² . Apart from some reports that semaphorin signalling is implicated in adhesive and metastatic potential, there’s little to go on.

A strange C-terminal domain

One confusing thing when we first started to look at this protein was that we didn’t understand what was happening at the C-terminal. There weren’t any domains annotated for it on InterPro, and it seemed pretty unlikely that there was a good 300 amino acids of protein just sitting there without any structure.

It was a stroke of luck that I ran across the preprint for Max Bileschi’s new PFAM-N method ¹³ that makes use of deep learning to perform better PFAM annotations. When I dug through the data, I saw that this method finally classified the human TMTCs as part of the right PFAM family. I reached out to Max over Twitter, and we began to discuss the mannosyltransferases. I hoped that the new method would be able to identify any domains on this C-terminal. In the end, there weren’t any matches for the majority of the C-terminal domain, but he did confirm that there were bits of this region that look like TPRs. The Alphafold2 model, backed up my hunch that was something there. It’s likely that this protein will get classified into a new CAZy family, so there’ll be two brand new human families to our name!

Mostly conserved in metazoa

We can find this gene pretty broadly in eukaryotes, but interestingly, there doesn’t look to be an orthologue in the invertebrates. The conserved domain in TMEM260 is also found in prokaryotes, although it’s unclear what this enzyme would be doing in these organisms.

The future

There’s a lot of function to unpack with this new enzyme, and the fact that it glycosylates many important molecules apart from plexins (including Hepatocyte growth factor receptor and Macrophage-stimulating protein receptor) means that it’s probably going to be pretty interesting.

Footnotes