Tue Jul 11 10:58:57 PDT 2000 Kevin Karplus T0110 Protein HI1288, H. influenzae wu-blast: very weak hits 1bal, 1bbl double-blast: no hits t2k: 31 sequences, none in PDB secondary structure: alternating alpha/beta? 3 helices, and either 2 long strands or 2 hairpins? No target model hits with T2K No template model hits with T2K CAFASP SAM-T99 target hit: CHAIN E-Value FSSP SCOP 1hra 20.19 2nllB 7.37.1 This hit is to a zinc-binding protein, but none of the cysteines that coordinate the zinc are present. What remains is a helix, that matches OK with the end of a helix in the predicted secondary structure. Wed Jul 26 11:26:33 PDT 2000 CAFASP hits fold #server #found #top #within10% 4.47 4 4 3 1 5.5 4 6 2 1 2.43 5 5 1 0 1.4 5 5 0 0 4.109 4 4 1 1 3.16 4 4 0 0 4.48 2 2 2 0 4.44 3 3 1 0 4.59 3 3 0 0 1.65 3 3 0 0 2.80 2 3 1 0 4.54 2 2 1 0 4.145 2 2 1 0 We could examine some of these hits, but there doesn't seem to be much consensus. Sat Aug 26 14:34:56 PDT 2000 Kevin Karplus Remade 2track predictions % Sequence ID Length Simple Reverse E-value SCOP 1lehA 364 -28.13 -9.82 4.2e-01 3.2.1,3.53. 1mml 265 -25.79 -9.50 4.2e-01 5.8.1 1cf9A 753 -24.77 -9.33 4.2e-01 3.16.1,5.5.1 1egaA 301 -29.04 -9.13 4.2e-01 3.31.1,4.44.2 1otgA 125 -28.95 -9.05 4.2e-01 4.65.1 1ek0A 170 -24.03 -8.64 1.1e+00 ? 1tsm 316 -25.02 -8.27 1.1e+00 4.96.1 We have one 3.16+5.5 hit (a cafasp match) also a 4.44 hit---none are impressive scores. Still we should check them. Tue Sep 5 10:35:22 PDT 2000 remaking 2track 5 September 2000 Rachel Karchin Functional information: http://www.tigr.org/tigr-scripts/hmm/hmm_report.spl?acc=TIGR00082 "ribosome-binding factor A" or rbfA From: Bylund et. al. "RimM and RbfA Are Essential for Efficient Processing of 16S rRNA in Escherichia coli" J Bacteriol, January 1998, p. 73-82, Vol. 180, No. 1 found at http://jb.asm.org/cgi/content/full/180/1/73?view=full&pmid=9422595 RbfA acts as a cold-shock protein. "RbfA has been shown to act as a high-copy-number suppressor of a cold-sensitive mutation in the 5'-terminal helix of 16S rRNA, and the growth of a strain lacking RbfA is cold sensitive. . . . RbfA seems to be important for the maturation of 16S rRNA. Evidently, increased expression of rbfA suppresses both a cold-sensitive mutation in 16S rRNA and the slow growth and translational deficiency of a rimM mutant" The article above cites: [15] Jones, P. G., and M. Inouye. 1996. RbfA, a 30S ribosomal binding factor, is a cold-shock protein whose absence triggers the cold-shock response. Mol. Microbiol. 21:1207-1218 cold shock RNA-binding domains listed in SCOP? These might be related to our rRNA binding domain ??? 1sro S1 RNA-binding domain of polyribonucleotide phosphorylase, PNP from Escherichia coli 2.38.4.5.3 Class: All beta proteins Fold: OB-fold barrel, closed or partly opened n=5, S=10 or S=8; greek-key Superfamily: Nucleic acid-binding proteins Family: Cold shock DNA-binding domain-like barrel, closed; n=5, S=8 also 2.38.4.5.5 1a62 1a8v[ab] 2a8v[abc] 1a63 Tue Sep 5 18:07:49 PDT 2000 Kevin Karplus good signal in short target model---may be suitable for fragfinder and undertaker. The 2.38.4 SCOP superfamily has the following FSSP reps: 1a0i 1a1d 1a62 1ah9 1b8aA 1bkb 1c0aA 1ckmA 1gpc 1gvp 1hjp 1jmcA 1krs 1lylA 1mjc 1otcA 1otcB 1pfsA 1pysB 1quqA 1quqB 1rip 1rl2A 1sro 3ullA 6 Sept 2000 Rachel Karchin top scoring alignments T0110-1otgA-2track-global.pw.a2m 125 -13.38 -13.07 6.8e-06 T0110-1ek0A-2track-global.pw.a2m 170 -0.06 -12.02 1.8e-05 T0110-1cf9A-2track-local.pw.a2m 727 -24.80 -9.33 3.7e-04 T0110-1egaA-2track-local.pw.a2m 301 -29.04 -9.13 3.7e-04 T0110-1lehA-2track-local.pw.a2m 364 -28.13 -9.82 3.7e-04 T0110-1mml-2track-local.pw.a2m 65 -25.79 -9.50 3.7e-04 T0110-1otgA-2track-local.pw.a2m 125 -28.95 -9.05 3.7e-04 T0110-1quqB-global.pw.a2m 114 -1.40 -9.76 3.7e-04 --first 2.38.4 T0110-1ek0A-2track-local.pw.a2m 170 -24.03 -8.64 1.0e-03 T0110-1tsm-2track-local.pw.a2m 316 -25.02 -8.27 1.0e-03 T0110-1hra-local.pw.a2m 80 -14.29 -6.08 6.9e-03 T0110-1hra-vit.pw.a2m 80 -8.98 -6.03 7.2e-03 1quqA-T0110-fssp-global.pw.a2m 128 -6.51 -5.92 2.0e-02 T0110-1ah9-vit.pw.a2m 71 -5.46 -3.46 1.4e-01 T0110-1otgA-2track-global.pw Strong alignment, conservation and 2ry match. Only unaligned pieces are a strand, helix turn, end of a strand and piece of a loop. T0110-1ek0A-2track-global.pw Some good conservation on helices, loops but not on strands. Two strands, a helix and some loop regions not alinged. 2ry match is good. T0110-1cf9A-2track-local.pw Not compact. T0110-1egaA-2track-local.pw Dimer connected by a loop. One domain of the dimer is well aligned to the target, with a few deletions reasonably placed at ends of strands and small pieces of loops. Also small piece of one helix is deleted. Aligned region has good conservation and excellent 2ry match. Sun Sep 10 11:07:38 PDT 2000 Kevin Karplus Re-examining the folds Rachel looked at: 1otgA/T0110-1otgA-2track-global missing initial strand (interior to sheet). Not much conservation in sheet---seems unlikely. 1ek0A/T0110-1ek0A-2track-global missing two interior strands. 1egaA/T0110-1egaA-2track-local ok conservation of helix, and good 2ry match. Not much residue identity on sheet, and final predicted helix past end of alignment. Possible, but not great. Now let's look at some more: T0110-1lehA-2track-local good 2ry match, low residue identity, missing first (edge) strand of sheet. possible, but not great. Can improve residue ID slightly: 1lehA/T0110-1lehA-karplus T0110-1mml-2track-local crosses domain boundaries, rather than putting all strands in one sheet. T0110-1otgA-2track-local very similar to 1otgA/T0110-1otgA-2track-global T0110-1quqB-global fold 2.38.4 poor 2ry match and poor residue identity Mon Sep 11 12:50:45 PDT 2000 Kevin Karplus T0110-1ek0A-2track-local Poor conservation, missing 3 of 6 strands of sheet. T0110-1tsm-2track-local Poor conservation, missing 2 of 5 strands of sheet. No reason to expect sheet to bend where this one does (no residue identity near there). T0110-1hra-local A one-helix match. No conservation in Zn-binding turn of template. 1quqA-T0110-fssp-global poor 2ry match and poor residue identity T0110-1ah9-vit 6 identical residues in 18-residue fragment. predicted helix, but is turns in template. Poor 2ry match, but can be extended to cover beta barrel. 1ah9/T0110-1ah9-karplus.a2m  Sor far the possible, but not strong predictions are 1egaA/T0110-1egaA-2track-local.pw.a2m.gz 10 identical residues good 2ry prediction HEEHE flat sheet 1lehA/T0110-1lehA-karplus.a2m 12 identical residues good 2ry prediction EEHEH flat sheet 1ah9/T0110-1ah9-karplus.a2m 15 identical residues adequate 2ry prediction 5-strand barrel Note: psi-pred and jpred2 pretty much agree with us on the 2ry prediction, lowering the confidence in the 1ah9 prediction Trying a TOPS search on http://www3.ebi.ac.uk/tops/topsQ.html for H (5,15) E (2,8) E (2,10) H (2,5) E (3,10) H OPS search version $Revision: 1.181 $ database=pdb sses=6 sses=[h,e,e,h,e,h] inserts=[[1=(5,15),2=(2,8),3=(2,10),4=(2,5),5=(3,10)],0,5] Number of Hbonds=0 hbonds=[] par_anti=[parmin=0..268435455,parmax=0..268435455,antimin=0..268435455,antimax=0..268435455] neghbonds=[] Number of chiralities=0 chiralities=[] general,[sum,name,nodes,cath-number,other],name=$empty,sorted=y hbond_match=weak,chirality_match=weak query=general([(weak,weak),[1..81,7..87,10..90,13..93,16..96,20..100],[(1..81,h,_41687,_41690,_41693,_41694),(7..87,e,_41724,_41727,_41730,_41731),(10..90,e,_41802,_41805,_41808,_41809),(13..93,h,_41880,_41883,_41886,_41887),(16..96,e,_41958,_41961,_41964,_41965),(20..100,h,_42036,_42039,_42042,_42043)],0..5,[],[parmin=0..268435455,parmax=0..268435455,antimin=0..268435455,antimax=0..268435455],[],[sheet=n,barrel=n,curved_sheet=n,vcurved_sheet=n,sandwich=n],[]],sum,name,nodes,cath,other) Total matched = 0 Search time : 102472 ms Even chopping off the final helix finds nothing---perhaps the search is just a global, and not a local, alignment search. Mon Sep 11 14:14:44 PDT 2000 Kevin Karplus It seems that Rachel has already done an undertaker run with fragments from fragfinder. Unfortunately, the optimum on her run was an all-helical structure. There were several structures found with some beta structure, but nothing compact. Perhaps if we combine the fragfinder fragments with the alignments, we can get something. The undertaker run with alignments does ok with the t110-1lehA.pdb structure, but the t110-1egaA.pdb starting point, which tries to superimpose 1egaA after 1lehA, ends up with terrible clashes, and the optimization just unfolds the proteins. The next try (doing 1egaA first then 1lehA) still has way too many clashes: T0110 36.0582 bits/residue, 266 clashes t110-1egaA at pool[1] 92.539 bits/residue, 281 clashes 14.7156 breaks t110-1lehA at pool[2] 90.7436 bits/residue, 322 clashes 43.0288 breaks and unfolds. Maybe we should try again without all the fragfinder fragments, so that the "InsertAlignment" operator keeps coming back to one of the alignment predictions? On try4, most of the structures still unfolded, but try4/t110-1egaA.7.7.pdb looks pretty good, though the final helix is not packed with the rest, and seems to be rotated a bit. Tue Sep 12 09:10:42 PDT 2000 On try5, the structures stayed folded, but a lot peeled the thrid strand a little away in an attempt to pack the helix. Ones that retained all three strands without breaking final helix are: t110-opt.7.10.pdb 35.6059 bits/residue, 32 clashes has space between helix and sheet t110-opt.2.10.pdb 35.6299 bits/residue, 36 clashes small space between helix and sheet t110-opt.10.10.pdb 35.6317 bits/residue, 33 clashes turns helix so not packed on sheet. t110-opt.18.10.pdb 35.7626 bits/residue, 32 clashes has space between helix and sheet t110-opt.8.10.pdb 35.7889 bits/residue, 36 clashes turns helix slightly, has small space t110-opt.19.10.pdb 35.8973 bits/residue, 31 clashes turns helix slightly, has space t110-opt.12.10.pdb 35.9152 bits/residue, 31 clashes has large space t110-opt.9.10.pdb 35.9735 bits/residue, 33 clashes turns helix slightly, has large space Of these, I think I like try5/t110-opt.2.10.pdb best Tue Sep 12 10:36:32 PDT 2000 On try6, the secondary structure is preserved. Things are still not quite as tightly packed as I'd like, probably because the turn between the first and second strands is bent the wrong way and is interfering with the packing of the final helix. Undertaker does not currently have much chance of bending the turn the other way without destroying the beta sheet, so this is probably about as good as I'll get. t110-try5 30.3729 bits/residue, 36 clashes t110-try5-scwrl at pool[1] 31.5447 bits/residue, 49 clashes t110-rot at pool[2] 30.2814 bits/residue, 34 clashes t110-opt at pool[3] 29.9724 bits/residue, 34 clashes t110-opt-rot at pool[4] 29.7755 bits/residue, 30 clashes t110-opt-scwrl at pool[5] 30.4031 bits/residue, 50 clashes t110-opt-scwrl-rot at pool[6] 29.5521 bits/residue, 26 clashes