Mon Aug 24 19:26:07 PDT 1998 Kevin Karplus t85 is C554_NITEU in swissprot (though that has 24 more residues on the beginning). Extra info: contains 4 hemes, related to known fold, "no sequence identity", is a monomer. Wu-blast finds some modest hits: 2cdv -5.2590967 1a2i -3.912023 2cthA -3.912023 2cthB -3.912023 2cym -3.912023 Double-blast finds 2cdv (same score, no intermediate) Note: 2cdv is a cytochrome (FSSP rep 3cyr) as is t85. This is a very probable hit. t85.t98_6 includes 2cym as one of the homologs---2cdv and 2cym are very similar (so 2cdv is probably included under another name). t85.t98_6 finds several: 1a2i -62.680 ? 2cth[AB] -62.680 3cyr Cytochrome c3 2cym -62.680 3cyr 2cdv -60.290 3cyr 1wad -35.280 3cyr 3cyr -32.350 3cyr 1czj -17.850 1aqe Cytochrome c3 1aqe -17.780 1aqe 2cy3 -12.890 3cyr 1fgj[AB] -10.600 1fgja hydroxylamine oxidoreductase The first several hits are all cytochrome c3 (3cyr or 1aqe as FSSP reps) The best hits with the library models are also cytochromes: 2cy3 -12.170 3cyr Cytochrome c3 2mtaC -7.760 2mtaC Cytochrome c5511 1dvh -6.770 1c53 (cytochrome c553) 1cyi -6.760 1cyj 100% cytochrome c6 (cytochrome c553) 1fcdC_1 -6.220 Flavocytochrome c sulfide dehydrogenase (fcsd) 5cytR -5.920 1ycc 61% Cytochrome c (reduced) 1onrA -5.630 1onrA transaldolase b 351c -5.360 ? cytochrome c551 oxidized Summing both ways improves a few of the cytochrome scores, so that all the top scores are cytochromes: 1a2i -62.680 3cyr (1a2i=2cth[AB]=2cym) 2cth[AB] -62.680 3cyr This has best resolution of the group. 2cym -62.680 3cyr 2cdv -60.290 3cyr 1wad -35.280 3cyr 3cyr -32.350 3cyr 2cy3 -25.06 3cyr 1czj -17.850 1aqe 1aqe -17.780 1aqe 1dvh -12.38 1c53 (79%) cytochrome c553 reduced 5cytR -11.21 1ycc 61% 1fgj[AB] -10.600 1fgjA 351c -10.58 ? 2mtaC -9.89 2mtaC 1cyi -9.44 1cyj 100% cytochrome c6 (cytochrome c553) The top fssp scores are 1aqe -8.530 cytochrome c3 mutant 1xbrA -7.910 t protein fragment DNA 1ak4C -7.110 cyclophilin 1ycc -6.200 cytochrome c (isozyme 1) 1fkj -5.970 fk506 binding protein 1wit -5.960 twitchin 18th igsf module Top scores with target98-mixed (using w0.5) are 1czj -15.980 1aqe 99% 1aqe -14.180 1aqe 1fcdC -10.720 1fcdC 1etpA -10.200 1etpA cytochrome c4 1fgjA -9.190 1fgja hydroxylamine oxidoreductase 1ccr -5.780 1ycc 57% Cythochrome c 1ctj -5.630 1cyj 65% cytochrome c6 1c53 -5.280 1c53 cytochrome c553 1cry -4.600 1ycc 47% cytochrome c2 1cyj -4.280 1cyj cytochrome c6 (cytochrome 553) 1kdu -4.040 5hpgA plasminogen activator 1krn -3.850 5hpgA plasminogen fragment Top scores with target98 library using viterbi scoring: 2cy3 -9.430 3cyr 1onrA -8.840 1onrA transaldolase b 351c -6.980 ? cytochrome c551 oxidized 2mtaC -6.470 2mtaC 1fcdC_1 -5.200 1fcdC 1dvh -5.160 1c53 5cytR -5.150 1ycc 1cyi -5.010 1cyj 1stu -4.640 1stu maternal effect protein staufen 1atx -4.630 1apf 64% ATX ia 1edt -4.600 2ebn 33% Endo-beta-n-acetylglucosaminidase h, endo h 1pk4 -4.540 5hpgA Plasminogen kringle 1ctm_2 -4.320 1hcz 100% Cytochrome f (reduced) 1sbp -3.780 1sbp sulfate binding protein 1prcC -3.360 1prcC photsynthetic reaction center 1cc5 -3.270 1cc5 1fcdC_2 -3.060 1fcdC t85.remote_4.varh50 includes a LOT of cytochromes, starting with all the 1ycc homologs. THen it has 1mey[CFG] -79.260 zinc-finger? 1aayA -51.380 zinc-finger peptide It seems clear that some cytochrome is the correct model, and several score very highly. The best-scoring alignments of the several cytochromes checked are 2cthA/t85-2cthA-vit 2cthA 107 -68.08 -66.71 2cdv/t85-2cdv-vit 2cdv 107 -66.64 -64.95 2cthA/t85-2cthA-global 2cthA 107 -57.07 -54.23 2cdv/t85-2cdv-global 2cdv 107 -55.58 -51.93 3cyr/t85-3cyr-vit 3cyr 107 -33.92 -31.91 3cyr/t85-3cyr-global 3cyr 107 -26.09 -25.49 2cthA/2cthA-t85-global T0085 211 -29.31 -18.51 2cdv/2cdv-t85-global T0085 211 -27.91 -17.69 1aqe/1aqe-t85-global T0085 211 -26.21 -17.19 3cyr/3cyr-t85-fssp-global T0085 211 -21.83 -17.08 1fcdC/1fcdC-t85-global T0085 211 -18.86 -15.66 2cdv/2cdv-t85-vit T0085 211 -17.20 -15.35 2cthA/2cthA-t85-vit T0085 211 -18.27 -14.69 3cyr/3cyr-t85-global T0085 211 -23.32 -14.60 3cyr/3cyr-t85-fsspt98-global T0085 211 -23.54 -14.54 1aqe/t85-1aqe-vit 1aqe 110 -16.10 -14.30 1aqe/t85-1aqe-global 1aqe 110 -14.27 -13.94 1etpA/1etpA-t85-global T0085 211 -18.19 -13.63 3cyr/3cyr-t85-const-global T0085 211 -19.34 -12.84 1aqe/1aqe-t85-vit T0085 211 -15.62 -12.51 3cyr/3cyr-t85-vit T0085 211 -12.44 -10.36 1fcdC/1fcdC-t85-fssp-global T0085 211 -13.44 -9.94 1ycc/1ycc-t85-fssp-global T0085 211 -12.52 -9.12 1aqe/1aqe-t85-fssp-global T0085 211 -14.16 -8.98 1ycc/1ycc-t85-glob T0085 211 -11.93 -8.38 1c53/t85-1c53-vit 1c53 79 -7.84 -7.52 In both directions, the best is 2cthA 2cthA/t85-2cthA-vit 2cthA 107 -68.08 -66.71 2cthA/2cthA-t85-global T0085 211 -29.31 -18.51 So it looks like 2cthA is likely to be our best bet. 2cthA-t85-global: T0085 -ADAPFEGRK------------------KCSSCHKA--QAQSWKDTAHAKAMESLKPNVKKEAK AP G K KC CH A S K 2cthA APKAPADGLKMEATKQPVVFNHSTHKSVKCGDCHHPVNGKEDYRKCGTAGCHDSMDKKDKSAKG T0085 QKAKLDPAKDytQDKDCVGCHVDGFGQKggytiESPKPMLTGVGCESCHG-p118k. K CVGCHV G K LTG CH 2cthA YYHVMHDKNT..KFKSCVGCHVEVAGAD.....AAKKKDLTGCKKSKCHE-...... t85-2cthA-vit: T0085 .....ADAPFEGRKKCSSCHKAQA..QSWKDTAHAKAMESLKPNVKKEAKQKAKLDPAKDYTQD KC CH A S K AK K T 2cthA a28sv---------KCGDCHHPVNgkEDYRKCGTAGCHDSMD-KKDKSAKGYYHVMHDKN-TKF T0085 KDCVGCHVDGFGQKGGYTIESPKPMLTGVGCESCHGPGRN... K CVGCHV G K LTG CH 2cthA KSCVGCHVEVAG-----ADAAKKKDLTGCKKSKCH-------- Note: 2cthA only aligns to the first 92 residues of t85, leaving 119 unaccounted for! We may need to look for the second part of t85 separately. The 1fcdC alignments may have the best hope of covering the whole thing, though their alignments are not as good as the 2cthA one on the first 92 residues. Constructed t85-2nd from the tail of t85 (after the CH at the end of the 2cthA match), and started a search using it in subdirectory t85-2nd. Top wu-blast hit: 1yeb -0.357 Nothing found by double-blast. t85-2nd.t98_6 finds [123]dmr -5.090 1dmr dmso reductase 1ce[ab][AB] -4.760 5hpgA 58% plasminogen fragment 1hp[jk] -4.740 5hpgA 52% plasminogen fragment 1qli -4.740 1qli cysteine and glycine-rich protein fragment 1pkr -4.700 5hpgA 58% plasminogen fragment 1cor -4.600 451c cytochrome c551 1mil -4.310 1mil shc adaptor protein fragment 1pbk -4.240 1fkj 42% fkbp25 c-terminal domain 1dms -4.030 ? dmso reductase It looks like plasminogen or dmso reductase are the best bets here. The 5hpgA homologs started to appear even for the whole chain, so the plasminogen lead may be more promising. I'll have to look up what characterizes "kringle" domains---are the cystine bridges important? 25 August 1998 Kevin Karplus All the approx 400 sequences in 5hpgA.target98-pdb conserve all 6 cysteines, so I'm reasonably certain that a prediction of a kringle domain that doesn't have the cystine bridges is VERY unlikely. With the target98 library models the bets hits are 2mtaC -8.240 2mtaC 1cyi -7.500 1cyj 100% 5cytR -6.560 1ycc 61% 1onrA -6.310 transaldolase B 351c -5.120 ? cytochrime c551 oxidized 1fcdC_1 -4.390 1fcdC 1dvh -3.720 1c53 1cc5 -3.400 1cc5 Summing both ways gives 2mtaC -8.240 2mtaC 5cytR -7.98 1ycc 1onrA -7.59 1onrA transaldolase B 1cyi -7.500 1cyj 1pk4 -5.42 5hpgA Plasminogen 2ebn -5.12 2ebn Endo-beta-n-acetylglucosaminidase h, endo h 351c -5.120 ? cytochrome c551 oxidized [123]dmr -5.090 1dmr dmso reductase 1ce[ab][AB} -4.760 5hpgA 58% plasminogen fragment 1hp[jk] -4.740 5hpgA 52% plasminogen fragment 1qli -4.740 1qli cysteine and glycine-rich protein fragment 1pkr -4.700 5hpgA 58% plasminogen fragment With viterbi on the FSSP models, the best are 1ycc -6.200 1ycc 1fkj -5.970 1fkj fk506 binding protein 1wit -5.960 twitchin 18th igsf module 1avc -4.900 annexin vi (p68, protein iii, 67-kda-calcimedin, lipoco 1cc5 -4.720 1cc5 1occM -3.510 1occM cytochrome c oxidase (ferrocytochrome c:oxygen oxidored... 1pysA -3.360 1pysA phenylalanyl-trna synthetase biological_unit With the target98-mixed models the best are 1ctj -6.810 1cyj 65% 1ccr -6.250 1ycc 57% 1cyj -5.570 1cyj 1cry -5.460 1ycc 47% 1fcdC -5.190 1fcdC 1czj -4.850 1aqe 99% 1cot -4.610 1ycc 34% 1c2rA -4.370 1ycc 33% 1etpA -4.270 1etpA 1cc5 -3.940 1cc5 1fgjA -3.910 1fgjA hydroxylamine oxidoreductase With Viterbi scoring on the target98 models the best are 1onrA -8.840 1onrA transaldolase B 2mtaC -7.200 2mtaC 351c -6.180 ? cytochrome c551 oxidized 1hsq -5.250 1ckaA 35% Phospholipase c-gamma (sh3 domain) 5cytR -5.150 1ycc 61% 1cyi -5.010 1cyj 100% 1semA -4.980 1ckaA 36% sem-5 (grb2 (mammalian homologue), drk (drosophila 1edt -4.600 2ebn 33% Endo-beta-n-acetylglucosaminidase h, endo h 2hnp -4.010 1pty 99% Protein-tyrosine phosphatase 1b (human) 1fcdC_1 -3.850 1fcdC 1sbp -3.790 1sbp sulfate-binding protein 1chl -3.720 1sis 79% chlorotoxin 1dvh -3.490 1c53 79% 1fd2 -3.450 6fd1 100% ferredoxin 1cc5 -3.270 1cc5 With the t85-2nd.remote_4 target model, the best are 1ycc homologs, which are included in the alignment. The large numbers of cytochrome hits suggests to me that the target has two domains, BOTH of which are good cytochromes. For the first domain, I almost certainly want to use 2cthA, but the second domain needs more thought. Note: 3cyr is a tetraheme cytochrome, though 2cthA seems to have only 2 hemes (double-check this---I may be misreading the PDB files). I'll build constrained and fssp-t98 alignments for several of the cytochromes that came up as hits, and see if they help any with the alignment. Here are the top non-self alignments for the second half 1ycc/1ycc-t85-2nd-fssp-global t85-2nd 119 -12.51 -11.86 2mtaC/2mtaC-t85-2nd-global t85-2nd 119 -11.09 -10.98 1ycc/1ycc-t85-2nd-global t85-2nd 119 -11.93 -10.42 1ycc/1ycc-t85-2nd-const-global t85-2nd 119 -11.37 -10.31 1c53/1c53-t85-2nd-fsspt98-global t85-2nd 119 -10.56 -9.94 1ycc/1ycc-t85-2nd-fsspt98-global t85-2nd 119 -12.45 -9.44 1cyj/1cyj-t85-2nd-fsspt98-global t85-2nd 119 -11.58 -8.93 2mtaC/2mtaC-t85-2nd-const-global t85-2nd 119 -12.93 -8.87 1cc5/1cc5-t85-2nd-const-global t85-2nd 119 -12.34 -8.70 1cc5/1cc5-t85-2nd-fsspt98-global t85-2nd 119 -11.41 -8.32 2mtaC/2mtaC-t85-2nd-fssp-global t85-2nd 119 -7.92 -8.14 2mtaC/2mtaC-t85-2nd-fsspt98-global t85-2nd 119 -12.56 -7.76 1onrA/1onrA-t85-2nd-vit t85-2nd 119 -9.67 -7.74 1fcdC/1fcdC-t85-2nd-fsspt98-global t85-2nd 119 -10.51 -7.70 5hpgA/5hpgA-t85-2nd-fssp-global t85-2nd 119 -5.57 -7.56 1cyj/1cyj-t85-2nd-global t85-2nd 119 -10.67 -7.43 1cyj/1cyj-t85-2nd-const-global t85-2nd 119 -11.93 -7.30 1etpA/1etpA-t85-2nd-fsspt98-global t85-2nd 119 -8.25 -7.18 1wit/1wit-t85-2nd-fssp-global t85-2nd 119 -3.29 -6.86 1cc5/1cc5-t85-2nd-fssp-global t85-2nd 119 -7.99 -6.71 3cyr/3cyr-t85-2nd-vit t85-2nd 119 -7.88 -6.58 5hpgA/t85-2nd-5hpgA-global 5hpgA 84 -7.12 -6.53 1cyj/1cyj-t85-2nd-fssp-global t85-2nd 119 -7.25 -6.46 1fkj/1fkj-t85-2nd-fssp-global t85-2nd 119 -6.85 -6.42 2mtaC/2mtaC-t85-2nd-vit t85-2nd 119 -6.87 -5.86 1pk4/t85-2nd-1pk4-global 1pk4 79 -5.79 -5.78 5hpgA/5hpgA-t85-2nd-const-global t85-2nd 119 -3.70 -5.77 1fkj/1fkj-t85-2nd-global t85-2nd 119 -3.82 -5.69 1cc5/1cc5-t85-2nd-global t85-2nd 119 -8.63 -5.51 1c53/1c53-t85-2nd-vit t85-2nd 119 -7.06 -5.41 1fcdC/1fcdC-t85-2nd-vit t85-2nd 119 -7.37 -5.41 5hpgA/5hpgA-t85-2nd-fsspt98-global t85-2nd 119 -3.49 -5.27 2ebn/t85-2nd-2ebn-global 2ebn 285 -7.98 -5.26 5hpgA/5hpgA-t85-2nd-global t85-2nd 119 -3.47 -5.22 1ycc/1ycc-t85-2nd-vit t85-2nd 119 -8.04 -5.06 The 1ycc-t85-2nd-fssp-global alignment leaves 23 residues between the domains unaligned and is not too great in the middle: t85-2nd g23kkTPRKDL-AKKGQDFHFEERCSACHLNYEGSPWKGAKAPYTPFTPEVDAKYTFKFDEMV T K AKKG F RC CH G P K F 1ycc .....TEFKAGSAKKGATL-FKTRCLQCHTVEKGGPHKVGPNLHGIFGRHSGQAEGYSYTDAN t85-2nd KEVKAMHEHYKL--------------EGVFEGEPKFKFHDEFQASAKPAKK-gk. F G K K K 1ycc IKKNVLWDENNMSEYLTNPKKYIPGTKMAFGGLKKEKDRNDLITYLKKACE-... 2mtaC-t85-2nd-global offers some possibility for filling that between-domain gap: t85-2nd gpgrn-------FRGDHRKSGQAFEKSGKKTPRKDLAKKGQDFH-------------FEERC G A E K G C 2mtaC .....APQFFNIIDGSPLNFDDAMEEGRDTEAVKHFLETGENVYNEDPEILPEAEELYAGMC t85-2nd SACHlnyeGSPWKGAKAP-------------------------------------------- S CH G G P 2mtaC SGCH....GHYAEGKIGPGLNDAYWTYPGNETDVGLFSTLYGGATGQMGPMWGSLTLDEMLR t85-2nd --------------------------YTPFTP-e53gk. TPF P 2mtaC TMAWVRHLYTGDPKDASWLTDEQKAGFTPFQP-...... Wed Aug 26 09:56:24 PDT 1998 Going through the top hits in order (see "see-all"), then regrouping by similarity of the predictions: The first group are all very similar, getting 4 heme groups in the first half of the sequence (probably too many)---3cyr, 2cthA, 2cdv, 1aqe. Of these, I like the 2cthA alignments best. 2cthA-t85-global 25 residues conserved, including helix CVGCHV, missing initial beta sheet. 2cthA-t85-vit 21 residues conserved, including helix CVGCHV, missing initial beta sheet, doesn't try topick up beginning of sequence. The t85-2cthA-global alignment looks pretty good in 3D, but is missing the initial beta hairpin that holds 2 or 3 of the 4 heme groups against the protein. This COULD be replaced by a beta sheet after the end of the aligned region. All the gaps and insertions occur naturally at the most exposed portions of loops. HEM2 has no coordinating residues conserved, HEM 3 and HEM4 have excellent conservation on one side, but the beta sheet on the other side is missing. HEM 1 has one histidine, but the other is not present (replaced by a proline). 2cdv-t85-global 24 residues conserved, missing initial beta sheet, very similar to 2cthA. 2cdv-t85-vit 21 residues conserved (doesn't try to pick up little piece at beginning of sequence). The t85-2cvd-global alignment is essentially the same as the t85-2cthA one. t85-3cyr-vit 18 residues conserved, 1gap, two inserts, missing initial beta sheet. Aligns up to CESCH. Gaps and inserts naturally on most exposed parts of loops. t85-3cyr-global same as t85-3cyr-vit 3cyr-t85-fssp-global 21 conserved residues, very similar to t85-3cyr-vit 3cyr-t85-global 20 conserved residues, very similar to t85-3cyr-vit 3cyr-t85-fsspt98-global 20 conserved residues, very similar to t85-3cyr-vit. 3cyr-t85-const-global, 21 conserved residues, similar to 3cyr-t85-fssp-global. 3cyr-t85-vit 20 conserved residues, similar to other 3cyr alignments. 1aqe-t85-fsspt98-global 19 residues conserved, still missing initial beta sheet, and a bit gappier than 3cyr and 2cthA alignments. 1aqe-t85-global 20 residues conserved, still missing initial beta sheet, and a bit gappier than 3cyr and 2cthA alignments. Very similar to fsspt98 alignment. t85-1aqe-vit 10 conserved residues, ends early at CVGCHVD helix. t85-1aqe-global 14 conserved residues, moves final histidine match way too far into the t0085 sequence. 1aqe-t85-const-global 19 residues conserved, very similar to 1aqe-t85-fsspt98-global. 1aqe-t85-vit 19 residues conserved, similar to other 1aqe alignments. 1aqe-t85-fssp-global is similar to other 1aqe alignments. -------------------- The 1fdc and 1etpA alignments pick up two heme groups, but not very convincing: The 1fcdC/1fcdC-t85-global alignment gets 2 heme groups nicely covered, but there are some large, uncloseable gaps, immediately next to the area most conserved in the alignments (DKDCVGCHVDG). This alignment extends further than the previous ones (out to DFHFEE). The 1fcdC/1fcdC-t85-fsspt98-global alignment also gets 2 heme groups nicely, but still has problems--the helices covering the backs of the heme groups are missing. The 1fcdC/1fcdC-t85-const-global is very similar to 1fcdC-t85-fsspt98-global. The 1fcdC/1fcdC-t85-fssp-global alignment also gets 2 heme groups nicely, but still has problems--the helices covering the back of one of the heme groups are missing, and gaps are still a bit too large. The 1etpA/1etpA-t85-global alignment again gets 2 heme groups, nicely coordinated by their histidines, but again has a large uncloseable gap. 1epta-t85-fsspt98 is similar. 1eptA-t85-const-global is similar. 1epta-t85-fssp-global is similar. ------------------------------ There is another match to the END of t85: The 1ycc/1ycc-t85-fssp-global alignment picks up one heme group at the very end of t85, though it needs a little hand editing. (t85-1ycc-hand1.a2m) 1ycc-t85-global is the same. 1ycc-t85-fsspt98-global shifts part by 1 residue to get a better initial match (should be incorporated into a hand alignment). 1ycc-t85-const-global is the same as 1ycc-t85-fsspt98-global. 1cc5-t85-const-global has a rather gappy match to the part that 1ycc matches better. 1cc5-t85-fsspt98-global has an even shorter match to the end of t85. ------------------------------ t85-5cytR-global has a very nice match to the beginning of t85. There 15 conserved residues and one heme group, but a rather large insert to pick up stuff from the end of t85. We can slide this in to close, but one of the CH pairs that probably coordinates a heme group will get covered (t85-5cytR-hand1). ------------------------------ Poor cytochrome matches: t85-1c53-vit matches only KDCVGCH, though clearly to the right place on the cytochrome. t85-1c53-global gets the same helix, but fills in the rest by using huge insertions. 1cyj-t85-fsspt98-global has a weak match to the beginning of t85, but has some missing helices and rather awkward gaps. 2mtaC-t85-global gets a heme group coordinated on FEERCSCH, but has some huge uncloseable gaps. 2mtaC-t85-const-global gets the same coordination, but fills in with different huge gaps. ------------------------------ Non-cytochrome matches (probalby bogus): 1onrA-t85-vit matches the end of a beta strand and a helix---this is not a significant match. 5hpgA-t85-fssp-global gets 8 conserved residues, but does not preserve the cystine bridges that hold the kringle domain together. 5hpgA-t85-const-global gets 11 conserved residues, but does not preserve the cystine bridges that hold the kringle domain together. 1pk4/1pk4-t85-global has 11 conserved residues with unbridgeable gaps. ------------------------------ Note: Prodom doesn't recognize C554_NITEU as having separate domains. 27 August 1998 Kevin Karplus We probably need to build alignments for ALL the cytochrome C structures in the database. The coordinating residues for the 4 heme groups are easy to find, as there are exactly 4 CH pairs in the sequence, and they all have the CxxCHx pattern of the heme-binding site, though not the C-{CPWHF}-{CPWR}-C-H-{CFYW} that prosite uses to define the site signature. The four are CSSCHK (compare with 1bccD/3bccD CSSCHS and 1gks CSSCHD, 1new CKTCHK) CVGCHV (compare with 1a2i/2cthA/2cym CVGCHV) CESCHG (compare with 1fcdC CEKCHV, 1fgjA CDNCHT, 1new CDACHE, 1wad CDDCHH) CSACHL (compare with 2mtaC CSGCHG. 1lmt CNACHL[no heme]) According to prosite, there are 98 cytochrome-C-containing PDB files (cytochrome-c-prosite.ids). I reduced the set to 29 (cytochrome-rep.ids), by choosing fssp representatives where they were close homologs. I included a few extra 1ycc homologs, since that is such a large and diverse set. I'll build alignments for all of these, and add them to the Makefile as possible joint alignments. I made a lot of new alignments, and got a few new ones that may be worth looking at: 2cy3/2cy3-t85-global T0085 211 -23.83 -13.25 1hroA/t85-1hroA-global 1hroA 105 -3.00 -11.69 2c2c/t85-2c2c-global 2c2c 112 -6.44 -11.30 155c/t85-155c-global 155c 134 -6.39 -9.97 1fgjA/t85-1fgjA-global 1fgjA 499 -20.69 -8.51 1ocd/1ocd-t85-global T0085 211 -11.98 -8.47 2frc/2frc-t85-global T0085 211 -11.98 -8.47 451c/t85-451c-global 451c 82 0.47 -8.31 1c52/1c52-t85-global T0085 211 -11.30 -7.86 1prcC/1prcC-t85-global T0085 211 -5.87 -7.85 1onrA/1onrA-t85-vit T0085 211 -9.10 -7.74 Fri Aug 28 17:44:52 PDT 1998 Christian and I looked at some alignments together and chose to use 3 parents: 3cyr/t85-crys-hand1 1ycc/1ycc-t85-hand2 1avc/t85-1avc-hand1 Christian may look for a better second half, but I don't expect he'll find one. 29 August 1998 Christian I think I did find a better one. It is in t85-2nd (which is not the previous t85-2nd that I did not have write permissions in; that is t85-2nd.no-permissions). Alignment is 1c52/1c52-t85.cbarrett1.a2m. The reason that I like 1c52 is that it is a cytochrome that associates with the phospholipid membrane electrostatically, probably/possibly through its basic tail. T0085 is also a cytochrome that associates electrostatically with lipid membranes. What I don't like is that structural deletion is necessary for T0085 to superimpose. There was another hit to a 1c52 FSSP family member, 1cor: 1cor -4.600 451c cytochrome c551 This alignment is probably more promising since only a few short deletions are necessary to get some quite promising residue conservation. See the alignment 1cor-t85.cbarrett1.a2m. This structure is similar to the ycc FSSP family members that I like below. Some other structures; 155c is not as promising and 1fgjA is a piece of a very large protein. 2frc, 1ocd, 155c, 2c2c, and 1hroA are from the 1ycc FSSP family: 2frc-t85-2nd.cbarrett1.a2m and 1ocd-t85-2nd-global are very similar, except that they show structural differences in the same region where T85 makes a deletion. This structure is a possibility. 2c2c looks nice, but may require a bit too much structural change. 1hroA allows probably the least structural deletion, but %IDE isn't as high. Currently, I prefer 1cor and 1c52 because they are single-heme binding cytochromes and the latter membrane associates electrostatically. The suspected membrane associating piece (last 8 or so residues) of T85 is unaligned at the end of 1cor, so could form the (helical?) membrane-binding piece. But I don't know one way or another whether 1cor is a membrane-bound protein. 2frc and 1ocd are second on my list. The others were not promising. Mon Aug 31 09:40:36 PDT 1998 Kevin Karplus I merged t85-2nd.no-permissions with t85-2nd and eliminated t85-2nd.no-permissions. Sorry about not getting the directory perrmissions right! 1cor is 68% identical with 451C (both are cytochrome c551). In the area right around the heme-binding site, t85 is more similar to 1ycc than to 1cor or 451c. Christian's alignment to 1cor does not work quite as well with 451c, and some of the insertions he has are mid-helix. I don't like the 1cor alignment much, though I have a hard time articulating why. I'd rather use 1ycc to GSPWK, then switch to 1c52, though I don't like the long gap between DEMV and KEVK. I would unalign a little of that region, so we have enough loose residues to bridge the gap somehow. (t85-2nd/1c52/1c52-t85-hand1.a2m) I think I like this a little better than 1cor, because the transition from the well-conserved part of 1ycc is better, thugh I am still a bit dubious about eith alignment at the end. Looking at the annexin hit again---I think I still prefer 1ycc+1avc, with 1ycc continuing to GSPWK, then switching. 1ycc+1avc looks like it will wrap around the heme just as well as 1cor or 1c52, and with fewer insertions and deletions. I could go with either 1c52 or 1avc though.