 |
 |
For my sins I am making figures of TLR structures. These are dimeric in activated forms and glycosylated. So I sort of assumed that the sugars would inherit the chain id of the polypeptide they are attached to. But then in 2Z7X (TLR1/2 heterodimer) when I coloured the gycolysation to fit with the chains it did not look right. Looking at the LINK record again I found the following: LINK ND2 ASN A 199 C1 NDG A 901 1555 1555 1.45 LINK ND2 ASN A 414 C1 NAG A 911 1555 1555 1.45 LINK ND2 ASN A 442 C1 NAG A 921 1555 1555 1.45 LINK ND2 ASN B 51 C1 NAG A 801 1555 1555 1.46 LINK ND2 ASN B 163 C1 NAG B 811 1555 1555 1.45 LINK ND2 ASN B 330 C1 NAG A 821 1555 1555 1.46 LINK ND2 ASN B 429 C1 NAG A 831 1555 1555 1.45 So only _one_ of the NAGs bound to chain B ends up getting a B chain id! I'm guessing this is because in the dimers the bulk of the sugar ends up closer to chain A rather than the chain B which it is attached to. And according the the RCSB "chain IDs for all bound moieties and waters are assigned based on their proximity (number of contacts) to the nearest polymer".(http://deposit.rcsb.org/depoinfo/download/dp.pdf section 4.4.1 Chain ID assignment ). So is this why these chains get switched? If so, I have to say that is crazy - surely the covalent bonds, detected when preparing the LINK description, should always take precedence? Maybe more reasonable treatment has been applied in some cases - but I would've hoped that there would be some consistency - and so maybe the rules need re-writing to remove these anomalies. I was only making a figure - but for sure someone will be trying to use the data for something more useful and quantitative. This sort of stuff will make their analysis difficult. All the best Martyn Martyn Symmons Cambridge