18 May 2000 Kevin Karplus FtsA from Thermotoga maritima Closest BLAST hits 1dcn[ABCD], 1auw[ABCD] Closest double-blast hit 1dkgD, 1dcn[ACD] Hit FSSP representative 1dkgD 1ba1 1dcn* 1auwA 1auw* 1auwA 1ba1 and 1auwA are distantly related in FSSP: NR. STRID1 STRID2 Z RMSD LALI LSEQ2 %IDE REVERS PERMUT NFRAG TOPO PROTEIN 158: 1ba1 1auwA 2.0 3.1 46 447 9 0 0 5 S delta 2 crystallin Mutant biological_unit The t2k alignment has 99 sequences, but none of them are in PDB, so this is not going to be a trivial alignment. The top target model hits are all very strong hits to homologs of 1ba1. The top template model hits are 1dgkD, 1kay, 1ba1, 1bupA (all homologs of 1ba1). I think that the BLAST hits are misleading and we should go with one of the 1ba1 homologs (1ba1 itself, 1dgkD, or one of the top target hits: 1ngj, 3hsc, 1ngi, 1hpm). The 1dgkD template works well with the template model, but not as well with the target model (though still quite well). Tue Jun 6 13:35:30 PDT 2000 Redid 2ry prediction with new neural net. Mon Jun 26 09:49:04 PDT 2000 Remade 2ry predictions Mon Jun 26 09:52:02 PDT 2000 Remade 2ry predictions Sat Aug 26 15:34:06 PDT 2000 Kevin Karplus Remade 2track predictions Top hits now % Sequence ID Length Simple Reverse E-value SCOP 1bupA 386 -184.17 -155.45 1.7e-64 3.50.1.1.1 1ba1 386 -183.69 -154.97 4.5e-64 3.50.1.1.1 1kay 381 -181.88 -153.15 1.2e-63 3.50.1.1.1 1dkgD 383 -155.42 -123.06 1.3e-50 3.50.1.1.3 30 Aug 2000 Rachel Karchin ftsA contains an ATP-binding site . . . and is a prokaryotic member of a protein family which includes actin, an eukaryotic protein also involved in cytokinesis. The ability of Fts A to bind ATP is accompanied by its chemical modification by phosphorylation in one of its residues (T205). M. Vicente web page http://web.bham.ac.uk/bcm4ght6/vicente/vic.html The actin fold is well studied. It has "extensive structural similarity to the NH2-terminal domain of heat-shock cognate protein Hsc70 . . . Despite the absence of significant similarity of their amino acid sequences, both prtoeins share the same core architecture and a common nucleotide binding site resembling the structure of hexokinase . . . The common fold consists of two alpha/beta domains, which are connected by a putative hinge with an ATP-binding site situated between the domains. Each domain contains a five-stranded beta sheet of identical topology . . . prokaryotic cell cycle proteins FtsA, MreB, StbA . . . are predicted to posess a structure similar to actin in the common core region . . . the diversity of biological function is provided by the polymorphism of the loops joining the beta- strands and helices in the core region and by inserted domains that show high variability. The actin fold. Kabsch, et. al. FASEB J 1995 Feb;9(2):167-74 Also well studied is the ATPase domain common to actin, hexokinase and hsp70 protein families. I've placed the article "An ATPase domain common to prokaryotic cell cycle proteins, sugar kinases, actin and hsp70 heat shock proteins" Bork et. al. Proc Natl Acad Sci U S A 1992 Aug 15;89(16):7290-4 in t89/papers This paper contains a nice picture of the structure of the ATPase domain. Hits we are considering: 1ba1 is a heat-shock cognate 70 protein. 1dkg is a hexokinase 1ngj is a heat-shock cognate 70 protein. 3hsc " " " " " 1ngi (not in PDB sum) [HEAT-SHOCK COGNATE 70KD PROTEIN] 1hpm is an n-terminal atpase fragment from of 7okda heat-shock cognate protein Wed Aug 30 10:58:56 PDT 2000 Kevin Karplus OK, it looks like we have the right hit (based on web page Rachel found). Is there any info in the paper that will help us choose the alignment? Wed Aug 30 13:14:02 PDT 2000 Rachel Karchin According to the "ATPase domain" paper, important conserved residues in the domain are found (with respect to a human actin) at: D11 G13 D154 Q137 G156 G342 A144 G301 G302 which in the target appear to map to (using multiple alignment shown in the paper) D14 G16 N212 S190 G214 G386 S193 G336 G337 1ba1-T0089-fssp-global.pw.a2m some conservaton on helices, but too many unaligned sheets and pieces of helices to be a good match 1ba1-T0089-global.pw.a2m similar conservation level to fssp-global a few unaligned sheets and helices 1ba1-T0089-local.pw.a2m some conservation but deletions are badly placed on sheets. 1ba1-T0089-vit.pw.a2m not compact T0089-1ba1-2track-global.pw.a2m Good 2ry structure match and conservation. Several insertions badly placed on sheets. Best so far. Conserves D14, G16, G214, G336, G337 asparagine replaces aspartate for "N212" Kevin -- you might want to look at this one. T0089-1ba1-2track-local.pw.a2m Similar to global. Check with measure-shift T0089-1ba1-global.pw.a2m Also good 2ry structure match. Aligns a region in the target from around pos 40-60 (seq. positions) which doesn't seem to have much importance. T0089-1ba1-local.pw.a2m Good conservation, not as much alignment as global. T0089-1ba1-vit.pw.a2m " " " " The last five all look to be of similar quality. 1dkgD-T0089-global.pw.a2m Larger unaligned regions than best 1ba1 alignments and some deletions on strands T0089-1dkgD-2track-global.pw.a2m Better than 1dkgD-T0089-global but more unaligned regions than best 1ba1 and many insertions on strands. Wed Aug 30 16:43:06 PDT 2000 I edited T0089-1ba1-2track-global to get 1ba1 T0089-1ba1-karplus.a2m, moving gaps to the "backside" of the protein. All the conserved sites mentioned in the article were the same in the various alignments we looked at, and were not moved in the hand editing. Need to look at 1bupA and 1kay (particularly 2track-global and global). Wed Aug 30 17:08:50 PDT 2000 Rachel Karchin T0089-1bupA-2track-global.pw.a2m Good 2ry structure match and strong conservation. Unaligned regions look reasonable. Conserves D14, G16, G214, G336, G337, G386 Does not conserve these residues: S193 aligned with an A, N212 aligned with D, S190 aligned with an E T0089-1kay-2track-global.pw.a2m Good 2ry structure match and good conservation. Unaligned regions are plausible. Same as above on the key residues from the article. Thu Aug 31 10:40:15 PDT 2000 Kevin Karplus 1ba1 and 1bupA are almost identical sequences. I edited T0089-1bupA-2track-global to match the edits I did to 1ba1 and called the result T0089-1bupA-karplus.a2m I did a blast search of PDB with 1ba1 to find out how identical the alternatives are: resolution 1ba1 100% 1.9 3hsc=1ngj=1ngi=1hpm 96% 1.7 1hpm (2.45 1ngi,2.35 1ngj, 1.90 3hsc) 1ngb 96% 2.38 1bupA 96% 1.7 1ngd 96% 2.44 1ngf 96% 2.38 1nga 96% 2.4 1atr 96% 2.34 1ngh 96% 2.50 1ngg 96% 2.43 1ats 96% 2.43 1ngc 96% 2.48 1nge 96% 2.43 1ba0 96% 1.90 1kaz 96% 1kax 96% 1kay 96% 2bupA 96% 1.70 1qqoA 96% 1.90 1qqmA 95% 1.90 1qqnA 94% 1hjoA 85% 2.3 1dkgD 48% 2.8 Since the differences between 1bupA and 1ba1 are small, we can either choose 1bupA for the higher resolution, or 1ba1 for the slightly better-looking pattern of conservation on the first beta hairpin. (Both have the same critical active-site residues.) T0089-1dkgD-karplus2.a2m The 1dkgD structure, though only 48% identical, can be aligned about as well as the 1ba1 and 1bupA alignments. The central residues are slightly more conserved, and one of the short inserts in 1ba1/1bupA is present in 1dkgD, but the residue conservation is slightly worse in the first and last beta sheets. I looked at the 1ba1-T0089-fssp-global alignment, but was not impressed with it. There are a couple of places where it may be better than the hand-edited alignment---most notably, I can't decide about the placement of the predicted helix akklsviiharlreimskskkffreveakiveeg I looked at the placement of this region in 1dgkD and found a better alignment there, saving it as T0089-1dkgD-karplus2.a2m I suggest we use T0089-1dkgD-karplus2.a2m as model 1 T0089-1ba1-karplus.a2m as model 2 Hmm---I moved one of the active sites (EPTA in 1ba1) which we had aligned to 190 SLVS before and has now been moved to 271 EMFY. Better check the paper, which I don't have here. Looked at paper---don't see any reason to prefer their alignment of the EPTA to ours.