AlphaLasso interesting cases


AF-A0A140GV22-F1 is an uncharacterized protein from Rhodoplanes sp. Z2-YC6860. It has 163 amino acids with one cysteine bond (between residues 39 and 97) which creates a loop. Loop's secondary structure is composed mainly of beta-strands, turns and bends. C-tail can be divided into two parts based on their secondary structure. First part (residues 98-133) has a similar secondary structure as the loop. It pierces the loop and winds around it two times creating a pretzel-like supercoiled lasso with a tiny beta-barrel fold. Second part of the tail (residues 134-163) starts with a turn, then goes back and structured as an alpha-helix goes through the pretzel-barrel motif, piercing the loop again.

Whole structure constitutes a LS4+++-C supercoiled lasso. N-tail (residues 1-38) is unstructured and has low pLDDT. Whole structure has a high average pLDDT = 81.31 lowered by the N-tail part. Part composed solely by lasso (loop + C-tail) has a very high pLDDT = 90.36.

AF-A0A140GV22-F1



AF-A0A2E5AR49-F1 is a “Type 9 secretion system plug protein N-terminal domain-containing protein” from Flavobacteriales bacterium. It has 436 amino acids with one cysteine bond (between residues 35 and 279). This protein may at first look similar to AF-A0A140GV22-F1 due to plug-like shape, “opposite” lasso type: LS4- - - +C supercoiled lasso, unstructured N-tail, high pLDDT but very high lasso pLDDT. However loop and C-tail create a fold more similar to beta-propeller than to a beta-barrel. Loop creates a majority of the fold. C-tail winds around a loop creating the remaining part of the fold.

Unfortunately, a crucial part of the loop near the cysteine bond (residues 268-278) has low/very low pLDDT, therefore maybe one shouldn't trust the position of the cysteine bond? Indeed, quick inspection of other proteins with the same name show a variety of lasso types (L-1C, L-1N, L-2C, LL+2,-2), which are the results of a different position of a cysteine bond. Moreover in InterPro there are around 3000 proteins with similar names and only 9 of them have lasso. Looks like the lasso is here only by accident.

AF-A0A2E5AR49-F1



AF-C4XEF8-F1 is a “Gamma-glutamylcyclotransferase AIG2-like domain-containing protein" from Mycoplasma fermentans composed of only 141 amino acids, with a very high pLDDT (94.25, lasso part 94.71). Cysteine bond which connects residues 46 and 111 creates a loop. Whole protein resembles an alpha-beta-sandwich fold. Both loop's and N-tail's secondary structure is composed of an alternating sequence: beta-strand, alpha-helix, beta-strand. N-tail winds around a loop two times and creates a supercoiled lasso LS3---N, with beta-strands sticking to loop's beta-strands. C-tail is composed of two short alpha-helices connected by a turn, which together with other alpha-helices create an orthogonal bundle.

AF-C4XEF8-F1