5 June 2000 Kevin Karplus T0095 alpha(E)-catenin fragment Associates with the cytoplasmic domain of cadherins. Primary importance in cell adhesion. Strong similarity to vinculins (which are also involved in cell adhesion on the cytoplasmic side of adhesion plaques). no good blast hits no double-blast hits 38 sequences in T2K alignment, in 3 families: 1) vinculin (2 subfamilies) VINC_CAEEL Vinculin (C. elegans) VINC_CHICK Vinculin (vertebrate) 2) alpha catenin (2 subfamilies) pir||T24221 predicted proteins, C. elegans CTN1_HUMAN alpha catenin 3) alpha catenin homolog pir||T08703 alpha-catenin homolog Searching with target HMM: no good hits (best Evalue=6.5) chain E-value FSSP rep 1sch[AB] 6.5 1bgp peanut peroxidase 1tui[ABC] 7.8 1efcA elongation factor tu 1ttt[ABC] 7.8 1efcA elongation factor tu 1b23P 7.8 1efcA elongation factor tu 1eft 7.9 1efcA elongation factor tu 1aip[ABEF] 11 1efcA elongation factor tu Searching template model library: no good hits chain E-value FSSP rep 1bjwA 6.8 1bjwA aspartate aminotransferase 1d2fA 21 1d2fA maly protein 1b8gA 37 1b8gA 1-aminocyclopropane-1-carboxylate synthase FSSP relationships: 1efcA 1d2fA 4.7 4.2 103 361 11 0 0 14 S maly protein 1efcA 1bjwA 2.5 4.3 103 382 8 0 0 14 S aspartate aminotransferase biological_unit 1bjwA 1d2fA 38.1 2.2 344 361 20 0 0 15 S maly protein 1bjwA 1b8gA 39.7 2.5 362 421 18 0 0 23 S 1-aminocyclopropane-1-carboxylate synthase (acc synthas 1d2fA 1b8gA 33.3 2.6 338 421 17 0 0 28 S 1-aminocyclopropane-1-carboxylate synthase (acc synthas It looks like the three template model hits are quite similar (PLP-dependent transferase superfamily in SCOP) and somewhat related to the elongation factor-tu hits. Predicted secondary structure is very helical. 1efcA has three domains: 2 beta barrels and an alpha-beta-alpha The PLP-dependent transferases have a lot of alpha helices, but have a very important beta sheet as part of alpha-beta-alpha. The helix predictions are quite strong, so fitting the target to an alpha-beta-alpha will be very difficult. Tue Jun 6 13:35:30 PDT 2000 Redid 2ry prediction with new neural net. Mon Jun 12 15:20:59 PDT 2000 Kevin Karplus Looked at CAFASP predictions: num SCOP 10 3.31.1 10 1.110.1 8 1.89.1 6 8.1.5 6 2.42.1 6 2.41.3 5 1.110.8 1bgp 1.89.1 1efcA 2.41.3, 2.42.1, 3.31.1 1bjwA 3.62.1 1d2fA 3.62.1 1b8gA 3.62.1 It looks like the 1efcA hits and the 3.62.1 should both be explored. Mon Jun 26 09:52:16 PDT 2000 Remade 2ry predictions Sat Aug 26 15:39:21 PDT 2000 Kevin Karplus Remade 2track predictions New top hits still weak: % Sequence ID Length Simple Reverse E-value SCOP 1dovA 181 -67.31 -9.91 4.2e-01 ? 2a0b 125 -44.48 -9.21 4.2e-01 1.25.9 6prcM 323 -46.02 -9.03 4.2e-01 6.2.1,? 1eyvA 156 -35.70 -8.96 1.1e+00 ? 1abwA 283 -49.53 -8.28 1.1e+00 1.1.1 Mon Sep 4 13:03:04 PDT 2000 Kevin Karplus All the 2-track hits are possible based on secondary structure---I should look at them. 1dovA is several long helices, so a possible fold. 6prcM has some beta strands at the beginning followed by many helices. 1eyvA has several helices. 1efcA has no long runs of helices, so seems like an improbable match (unless our 2ry structure is all wrong). The 3.62.1 hits in CAFASP have no more than 3 helices together, so also seem unlikely. There is some consensus in CAFASP on 1.110 folds. Here are the 22 FSSP chains with those domains: 3bct 001.110.001.001.001 1qgrA 001.110.001.001.003 1qbkB 001.110.001.001.004 1bk5A 001.110.001.001.005 1b3uA 001.110.001.002.001 1b89A 001.110.001.003.001 1lrv 001.110.001.004.001 1qmmA 001.110.001.005.001 1ycsB 001.110.002.001.001 1bd8 001.110.002.001.003 1iknD 001.110.002.001.007 2myo 001.110.002.001.008 1sw6A 001.110.002.001.009 1dcqA 001.110.002.001.010 1pbv 001.110.003.001.001 1qsaA 001.110.005.001.001 1ft1A 001.110.006.001.001 1qjaA 001.110.007.001.002 1a17 001.110.008.001.001 1qqeA 001.110.008.001.002 1bpoA 001.110.009.001.001 2occE 001.110.010.001.001 Tue Sep 5 09:26:16 PDT 2000 Kevin Karplus Top alignments are now SCOP 1dovA/T0095-1dovA-2track-local 1dovA 181 -67.31 -9.91 3.7e-04 ? Helix lengths match fairly well, but the first two predicted helices are missing. Note: 1dovA is an alpha catenin fragment, so it SHOULD be the best match! Top fssp matches: 1dovA 26.5 0.0 181 181 100 alpha-catenin fragment 1dowA 21.4 1.6 179 205 97 alpha-catenin fragment beta-catenin fragment 1nfn 9.9 3.1 107 132 11 apolipoprotein e3 fragment 1a36A 9.0 11.1 102 544 14 topoisomerase i fragment Mutant biological_unit DNA Slightly moving the helices (still keeping the turns in the same places) increases the residue identities slightly: 1dovA/T0095-1dovA-karplus.a2m Let's go with this prediction. 2a0b/T0095-2a0b-2track-local 2a0b 125 -44.48 -9.21 3.7e-04 1.25.9 6prcM/T0095-6prcM-2track-global 6prcM 323 -16.38 -9.59 3.7e-04 6.2.1,? 6prcM/T0095-6prcM-2track-local 6prcM 323 -46.02 -9.03 3.7e-04 6.2.1,? 1abwA/T0095-1abwA-2track-global 1abwA 283 -39.78 -8.84 1.0e-03 1.1.1 1abwA/T0095-1abwA-2track-local 1abwA 283 -49.53 -8.28 1.0e-03 1.1.1 1b8gA/1b8gA-T0095-vit T0095 244 -12.63 -8.55 1.0e-03 3.62.1 1dovA/T0095-1dovA-2track-global 1dovA 181 -53.34 -8.05 1.0e-03 ? 1eyvA/T0095-1eyvA-2track-global 1eyvA 156 -19.99 -8.14 1.0e-03 ? 1eyvA/T0095-1eyvA-2track-local 1eyvA 156 -35.70 -8.96 1.0e-03 ? 2a0b/T0095-2a0b-2track-local 2a0b 118 -43.97 -8.94 1.0e-03 1.25.9 1b8gA/1b8gA-T0095-local T0095 244 -18.84 -7.45 2.7e-03 3.62.1 1d2fA/1d2fA-T0095-local T0095 244 -20.22 -7.98 2.7e-03 3.62.1 1eyvA/T0095-1eyvA-2track-local 1eyvA 131 -34.21 -7.98 2.7e-03 ? 1iknD/1iknD-T0095-global T0095 244 -7.31 -7.17 2.7e-03 1.110.2 1ycsB/1ycsB-T0095-global T0095 244 -9.58 -7.60 2.7e-03 1.110.2 1b8gA/1b8gA-T0095-fssp-global T0095 244 -7.67 -6.61 7.4e-03 3.62.1 1bd8/1bd8-T0095-global T0095 244 -9.39 -6.22 7.4e-03 1.110.2 1bjwA/1bjwA-T0095-local T0095 244 -22.47 -6.51 7.4e-03 3.62.1 1d2fA/1d2fA-T0095-vit T0095 244 -11.87 -6.86 7.4e-03 3.62.1 1eyvA/T0095-1eyvA-2track-global 1eyvA 131 -18.50 -6.33 7.4e-03 ? 2a0b/T0095-2a0b-local 2a0b 118 -17.02 -6.18 7.4e-03 1.25.9 2a0b/T0095-2a0b-vit 2a0b 118 -11.11 -6.55 7.4e-03 1.25.9 1bjwA/1bjwA-T0095-vit T0095 244 -13.33 -5.30 2.0e-02 3.62.1 1qbkB/T0095-1qbkB-2track-local 1qbkB 880 -54.56 -5.47 2.0e-02 1.110.1 1qqeA/T0095-1qqeA-2track-local 1qqeA 292 -53.84 -5.24 2.0e-02 1.110.8 1sw6A/1sw6A-T0095-global T0095 244 -6.58 -5.97 2.0e-02 1.110.2 2myo/2myo-T0095-fssp-global T0095 244 -5.46 -5.99 2.0e-02 2myo/2myo-T0095-global T0095 244 -5.93 -5.45 2.0e-02 For T0095, let's use 1dovA/T0095-1dovA-karplus.a2m 1dovA was our top hit with the 2-track HMM, and it is the "dimerization and beta-catenin-binding region of alpha catenin". Our target is alpha-catenin. The residue identity is low, but the 2ry match is great, with helices of exactly the right lengths. When combined with the functional match, I think this is our best bet. The structure wasn't released until 12 Jul 00, which explains why "no known homolog" was reported for the target. Tue Sep 5 10:35:28 PDT 2000 remaking 2track