30 July 1998 wu-blast finds one sequence weakly 1c6s -1.17 double-blast finds one sequence weakly 1ffh -5.32 t77.t98_6 finds a few sequences weakly: 1uag -5.880 1uag 1erv -4.440 1erv thoredoxin mutant 1spbP -4.260 1spbP 1cq[gh]A -3.980 1erv 1md[ijk]A -3.980 1erv 1tr[suvw] -3.970 1erv [34]trx -3.880 1erv 1gotB -3.510 1tbgA 7-bladed beta propellor 2trcB -3.510 1tbgA 1g[gp]2B -3.500 1tbgA 1tbg[ABCD] -3.500 1tbgA 1hmp[AB] -3.080 1dbrA (all the 1erv homologs are extremely similar--deliberate mutation experiments) The template library finds a few sequences weakly: 1scuA_1 -4.590 1scuA 1scuA -4.220 1scuA (similar to 1ido Z=5.4, id=7%) 1ido -4.030 1ido 1bbpA -3.680 1bbpA 1lfaA -2.730 1ido (only 32% identity with 1ido) 1fcdA_2 -2.620 1fcdA (only 16% identity with 1ido) 1uag similar to 1ido (Z=3.2, id=7%) 1erv similar to 1ido (Z=2.5, id=14%) It looks like many of our predictions are at least weakly consistent! Note: summing both ways only improved 1spbP (to -4.44), which is not similar to the otehr predictions. t77.remote_4 moves 1erv to the top, then 1uag, but is otherwise quite similar to target98_6. 31 July 1998 I can get a hand-alignment of 1uag to t77 that respects the secondary structure, and conserves 18 residues with 4 gaps, but the resulting structure is not very compact. (t77-1uag-hand1.a2m) The 1erv alignment t77-1erv-hand1.a2m looks a bit more promising, with a compact structure and 21 conserved residues (3 gaps, one an unpleasant one omitting a central strand of a beta sheet). The secondary structure is mostly preserved. The 1ido alignmen 1ido-t77-hand1.a2m also looks promising: compact, 19 conserved residues, two small gaps. 5 August 1998 Christian T0077 is RL3E_YEAST From PMID:9121443 [Vilardell J, et al. Ribosomal protein L32 of Saccharomyces cerevisiae influences both the splicing of its own transcript and the processing of rRNA. Mol Cell Biol. 1997 Apr; 17(4): 1959-1965. PMID: 9121443; UI: 97219985. ] Ribosomal protein L32 of Saccharomyces cerevisiae binds to and regulates the splicing and the translation of the transcript of its own gene. Selecting for mutants deficient in the regulation of splicing, we have identified a mutant form of L32 that no longer binds to the transcript of RPL32 and therefore does not regulate its splicing. The mutation is the deletion of an isoleucine residue from a highly conserved hydrophobic domain near the middle of L32. The mutant protein supports growth, at a reduced rate, and is found at normal levels in mature ribosomes. However, in cells homozygous for the mutant gene, the rate of processing of the ribosomal RNA component of the 60S ribosomal subunit is severely reduced, leading to an insufficiency of 60S subunits. L32 must be considered a remarkable protein. Composed of only 104 amino acids, it appears to interact with three distinct RNA molecules to influence three different elements of RNA processing and function in three different locations of the cell: the processing of pre-rRNA in the nucleolus, the splicing of the RPL32 transcript in the nucleus, and the translation of the spliced RPL32 mRNA in the cytoplasm. Fri Aug 21 18:29:04 PDT 1998 Kevin Karplus Starting over with new Makefile. Picked up rl3e_yeast.txt. Sat 22 Aug Wu-blast best hit 1c6s -1.17 double-blast still nothing. No PDB files in t77.t98_6.a2m t77.t98_6 finds 1uag -6.630 1uag 1erv -5.210 1erv 1pbk -4.740 1fkj 42% 1cq[gh]A -4.520 1erv 1md[ijk]A -4.520 1erv 1tr[suvw] -4.520 1erv [34]trx -4.520 1erv 1gotB -4.320 1tbgA 99% 2trcB -4.320 1tbgA 98% 1g[gp]2B -4.290 1tbgA 99% 1tbg[ABCD] -4.290 1tbgA 1tif -4.100 1tif Target98 library models find 1scuA_1 -5.520 1scuA 1scuA -4.640 1scuA 1mysA_2 -4.500 1lvk 47% ? 1ido -4.440 1ido 1bbpA -3.650 1bbpA 1lfaA -3.490 1ido 32% Summing both ways gets 1uag -6.630 1uag 1tif -5.9 1tif 1spbP -5.84 1spbP 1scuA_1 -5.520 1scuA 1erv -5.210 1erv 1pbk -4.740 1fkj 42% 1scuA -4.640 1scuA 1mdiA -4.55 1erv 95% FSSP library models with viterbi get 1bovA -5.950 1xyzA -5.490 1a2kC -5.260 1a04A -4.980 1bdb -4.410 1al3 -4.340 t98-mixed library with w0.5 finds 1atzA -6.140 1atzA 1atzB -4.760 1atzA 100% 1fcdA -2.810 1fcdA 1hdcA -2.760 1bdb 30% target98 with viterbi finds 1hdcA -7.630 1bdb 30% 1scuA_1 -7.460 1scuA 1mysA_2 -5.930 1lvk 47% 1bbpA -5.280 1bbpA 1scuA -5.230 1scuA 1ido -5.130 1ido 2mev2 -4.770 1aym2 35% 1tpg_2 -4.730 1tpg 23 August 1998 Top hits are 1hdcA, 1scuA_1, 1uag, 1atzA, 1bovA, 1mysA_2, 1tif, 1spbP, 1xyzA, 1bbpA, 1a2kC, 1scuA, 1erv, 1ido, (Note: need to use 2mysA---1mysA is obsolete.) Best alignment scores are: 1uag/t77-1uag-global 1uag 428 -10.32 -8.83 1erv/t77-1erv-global 1erv 105 -7.01 -7.95 1hdcA/1hdcA-t77-vit T0077 105 -8.23 -6.97 1uag/t77-1uag-vit 1uag 428 -8.57 -6.56 1ido/1ido-t77-vit T0077 105 -6.03 -6.52 1ido/1ido-t77-global T0077 105 -0.75 -6.47 1erv/t77-1erv-vit 1erv 105 -6.65 -6.24 1spbP/1spbP-t77-fssp-global T0077 105 -6.82 -6.18 1tif/t77-1tif-global 1tif 76 -5.73 -5.62 1ido/1ido-t77-const-global T0077 105 -0.48 -5.48 1scuA/1scuA-t77-const-global T0077 105 -4.30 -5.36 1ido/1ido-t77-fsspt98-global T0077 105 -0.49 -5.10 1atzA/1atzA-t77-vit T0077 105 -5.20 -4.82 1scuA/1scuA-t77-vit T0077 105 -5.31 -4.79 1tif/1tif-t77-global T0077 105 -4.70 -4.35 1spbP/t77-1spbP-vit 1spbP 71 -8.23 -4.32 1atzA/t77-1atzA-global 1atzA 184 -4.23 -4.19 1ido/t77-1ido-global 1ido 184 -5.56 -3.96 1spbP/1spbP-t77-vit T0077 105 -5.04 -3.66 1spbP/t77-1spbP-global 1spbP 71 -4.54 -3.65 1spbP/1spbP-t77-const-global T0077 105 -5.87 -3.60 1spbP/1spbP-t77-global T0077 105 -4.87 -3.37 1tif/1tif-t77-fssp-global T0077 105 -3.40 -3.17 1bovA/1bovA-t77-fssp-global T0077 105 -1.52 -2.21 1a2kC/t77-1a2kC-vit 1a2kC 196 -0.94 -2.13 1tif/t77-1tif-vit 1tif 76 -3.09 -2.13 Target 77 is known to bind RNA (both ribosomal and its own mRNA). Are any of our templates RNA binding? 1hdcA is structurally somewhat similar to some sugar-binding proteins. 1atzA is structurally similar to several sugar-binding proteins, including 2dri and 1a04A. Note: 1ido is similar to the same proteins. 1bovA is structurally similar to 3ullA, a DNA-binding protein. 1tif IS a ribosome binding factor! This looks promising! 1spbP is structurally similar to 1ris, a ribosomal protein, and to 1ha1) 1xyzA is structurally similar to 1pud which recognizes trna. 1a2kC is a nuclear transport protein. Mon Aug 24 09:31:43 PDT 1998 Kevin Karplus t77-1uag-global has 14 residue identities, and 2 inserts, but is not very compact. Secondary structure has moderate agreement. 1ido/1ido-t77-vit gets 11 conserved residues with 2 small gaps. 1ido/1ido-t77-global gets 10 conserved residues with 2 small gaps. 1ido/1ido-t77-const-global gets 12 residue ids with 2 small gaps. The alignments are basically the same, with the gaps shifted over by 1 or 2. All have pretty good secondary structure agreement---the const-global alignment probably being the best. t77-1erv has 9 residue identities with 2 inserts, and is fairly compact. The first strand in the middle of the beta sheet is missing, but perhaps the edge strand could move over. Secondary structure agreement is moderately good, and the inserts are on outside turns. This is an alpha-beta-alpha structure. 1hdcA-t77-vit has 5 residues conserved on short alignment (24 residues). The alignment can be extended a little bit, but not far, since it is at the beginning of 1hdcA. 1spbP/1spbP-t77-fssp-global has 13 residue identities with 1 gap, but only aligns 48 residues. Secondary structure agreement is poor. 1tif/t77-1tif-global has 15 residue identities, but 2 short gaps and a long insert. With a little hand-editing (t77-1tif-hand1), I can get excellent agreement on the beta sheet and the helix on the active side (14 conserved residues with 2 small inserts), but the back remains unaligned. This is the perhaps the most convincing alignment I've seen, and the similarity in function increases my confidence. All the alignments I've seen are to alpha-beta-alpha folds, so I'm pretty sure that this is an alpha-beta-alpha. 25 August 1998 Christian I've made t77last as a target directory for the part of t77 that does not match 1tif. Interestingly enough, it seems as though a structure has been determined for the prokaryotic counterpart of this protein, but all that exists in PDB is its header. Structure is 0RPL. Fri Aug 28 11:06:28 PDT 1998 Kevin Karplus I think the 0rpl hit is bogus---a different protein with a similar name---since it did not come up in the homology search of pdb, but is in all-protein. There are three conflicting predictions coming in from t77last: 1bdb/1bdb-t77-vit or 1bdb/t77last-1bdb-hand1 1ihp/t77last-1ihp-hand1 1mit/t77last-1mit-hand1 The (t77last) 1bdb-t77-vit alignment does not conflict with the 1tif alignment, but could be in almost any ortientation. The beta sheet could fit onto the exposed face of the beta sheet of 1tif, but the extra matching part of t77last-1bdb-hand1 runs out into space with nothing to stabilize it. Th 1ihp alignment results in matching up a helix to the final helix of 1tif, but then running off with another long helix way out away from everything else. This seems unlikely somehow. By unaligning the helix at the beginning of t77last-1mit-hand1 (t77last-1mit-hand2), we get a domain that can fit fairly comfortably on the exposed face of the beta sheet of 1tif. Current best guess: t77-1tif-hand1 t77last-1mit-hand2