PFRMAT AL 
TARGET T0063 
AUTHOR 5529-3140-9255 
REMARK Prediction team BENNER-COHEN has two group leaders, 
REMARK but we will consistently use Steven Benner's predictor 
REMARK number to avoid confusion concerning a second team 
REMARK connected with Fred Cohen (called Cohen, Fred). 
REMARK (Fred Cohen's number as a group leader is 6140-7890-6093). 
REMARK Prediction team members: D.L.Gerloff, G.Cannarozzi, 
REMARK M.Joachimiak, F.E.Cohen & S.A.Benner. 
METHOD 
METHOD This threading alignment prediction follows the secondary 
METHOD structure prediction submitted and accepted in SS format 
METHOD for the same target protein (Acc.code T0063SS156_1). 
METHOD 
METHOD The alignment (33 sequences total), predictions for 
METHOD the three subalignment reflecting the three primary 
METHOD kingdoms of life (eubacteria, eukaryotes, and archae), 
METHOD structural divergence within the superfamily in the C- 
METHOD terminal part, and the impact of the increasing sequence  
METHOD data bases for a secondary prediction like this one, are 
METHOD discussed in a manuscript submitted to BBRC before 
METHOD the deadline for T0063. -> Gerloff, Joachmiak, Cohen, 
METHOD Cannarozzi, Chamberlin and Benner. Structure prediction 
METHOD in a post-genomic environment. A secondary and tertiary 
METHOD structural model for the initiation factor 5A (IF-5A) 
METHOD family. Abstract. Two predictions have been prepared for 
METHOD the conformation of initiation factor 5A (IF-5A) starting 
METHOD from a set of homologous sequences. The first secondary 
METHOD structural model was predicted in 1994, when only eleven 
METHOD homologs (and no eubacterial homologs) had been sequenced, 
METHOD the second recently, after genome projects had generated 
METHOD a total of 33 sequences for the protein family, from species 
METHOD of all three kingdoms of life. With the second set of se- 
METHOD quences, but not the first, it was possible to predict that 
METHOD the N-terminal domain of the protein may fold in a, possibly 
METHOD open, beta-barrel/sandwich core structure, with a short helix  
METHOD capping one side of the barrel. [...]. 
METHOD 
METHOD We derived a set of expected characteristics for the folded 
METHOD structure of T0063 (#1-87 or #1-98) based on (a), a strongly  
METHOD predicted sequence of core secondary structure elements,  
METHOD S1a-S1b-S2-S3-[H]-S4-S5-(S6?), (S=strand, H=helix), (b) on 
METHOD an analysis of highly conserved alignment positions in the  
METHOD different kingdoms, and (c) on comparison with our expectations 
METHOD regarding possible folding topologies with the standard out- 
METHOD puts generated by the publicly accessible prediction tools 
METHOD at the UCLA-DOE Laboratory (D.Fischer, D. Rice, D. Eisenberg) 
METHOD and through the Procyon (H. Floeckner, M. Sippl) PRoFIT pro- 
METHOD gram.  
METHOD Altogether, these observation indicate that the folded struc- 
METHOD ture of the first domain is composed of a highly twisted, an- 
METHOD tiparallel beta-sheet and a single helix (which may be reduced 
METHOD to one or two turns in the archae structures), and that the  
METHOD residues involved in functional interactions may be located 
METHOD at both "ends" of the resulting barrel/sandwich structure. 
METHOD With respect to the strand order in the sheet, we have a  
METHOD weak preference for a "Greek Key" topology, or similar, with 
METHOD the predicted helix in the long connection. For 5 core seg- 
METHOD ments and a closed barrel structure, this topology is exempli- 
METHOD fied in the "OB-fold", described by Murzin (Murzin (1993). 
METHOD EMBO J. 12:861-867). However, our model would suggest some 
METHOD noticeable deviations from the typically conserved features 
METHOD of this ancient fold, dominated by the unusual location of 
METHOD the highly conserved stretch in functional residues in the 
METHOD eukaryotic subfamily, in the loop connecting core strands 
METHOD 2 and 3 (according to Murzin's proposed numbering of OB- 
METHOD fold core segments). While this is unusual, we have found a 
METHOD precedent for such an insertion in 1rip (ribosomal protein 
METHOD S17), which is the target structure for our first model 
METHOD submitted in threading alignment format. The very open barrel 
METHOD (rather a strongly twisted sheet) is also compatible with 
METHOD our observation that it would be difficult to avoid placing 
METHOD aromatic side-chains into the interior of the core, which 
METHOD would be more difficult in the small, tightly wound structures 
METHOD making the "classic" OB-folds, like 1csp (Murzin, EMBO J.). 
METHOD Finally, a deletion mutant has identified a fragment of human 
METHOD eIF-5A as a minimum substrate for hypusinylation that would 
METHOD correspond to positions 22-72 in the target sequence. This 
METHOD could be too short to include the entire open barrel domain 
METHOD suggested through our analysis. Reasons for this inconsistency 
METHOD could be a misassignment or misalignment of the core secondary 
METHOD structures, supersecondary structure formation in the fragment, 
METHOD or misprediction of the tertiary structure altogether. 
METHOD 
METHOD Further, in our comparisons with the standard fold recognition 
METHOD servers, we found that 1rip had been ranked 11 with a subsignificant 
METHOD z-score by the UCLA-gonnet+predss method (Fischer & Eisenberg). 
METHOD One of the folds that was found recurrently and highly-ranked on 
METHOD the UCLA-lists was 2rsp_A, rous sarcoma viral protease. We submit 
METHOD model two as a threading submission to this parent structure. In 
METHOD our favorite sequence-structure alignment, however, the differences 
METHOD between the two resulting models should not be drastic (both Greek- 
METHOD Key type topologies, modifications would concern predicted core  
METHOD strand 3 mostly). Finally a third target we considered for sub- 
METHOD mission in the CASP experiment is the first domain of the RepA 
METHOD protein (1jmc_A, res.183-298), because of its "jaw-like" binding 
METHOD site which could provide one explanation for our finding of a  
METHOD split, or bipartite functional site in the eukaryotic eIF-5A 
METHOD proteins, and because of the space neighbouring the strand2-3 
METHOD connection (OB-fold nomenclature) which is held in 1jmc by an 
METHOD insertion later in the fold, which does not seem to be present 
METHOD in 1jmc. 
METHOD ---------------------------------------------------------------- 
METHOD IMPORTANT COMMENT REGARDING MODEL 1 (BASED ON 1rip): 
METHOD Model 1 of the 9 NMR structure models combined in the PDB-entry 
METHOD has an obvious defect in its structure (a KNOT!!). Models 2 ff. 
METHOD are also not convincing, structurally, but not knotted, at least! 
METHOD ---------------------------------------------------------------- 
METHOD Due to time constraints, we may however not be able to submit 
METHOD an alignment for all three predictions. We submit model 1 as 
METHOD a recognized parent structure only, for structural alignment 
METHOD of our secondary structure prediction in general, please refer 
METHOD to the submitted manuscript. (OB-domain core predicted to include 
METHOD target residues 10-88 in our preferred model. 
MODEL  2 
PARENT 2rsp_A 
S   21   L   1 
Y   22   A   2 
V   23   M   3 
V   24   T   4 
I   25   M   5 
D   26   E   6 
G   27   H   7 
E   28   K   8 
P   29   D   9 
C   30   R   10 
R   31   P   11 
V   32   L   12 
V   33   V   13 
E   34   R   14 
I   35   V   15 
E   36   I   16 
K   37   L   17 
S   38   T   18 
K   39   N   19 
T   40   T   20 
G   41   G   21 
K   42   S   22 
H   43   H   23 
S   45   P   24 
A   46   V   25 
K   47   K   26 
A   48   Q   27 
R   49   R   28 
I   50   S   29 
V   51   V   30 
A   52   Y   31 
V   53   I   32 
G   54   A   34 
V   55   L   35 
F   56   L   36 
D   57   D   37 
G   58   S   38 
G   59   G   39 
K   60   A   40 
R   61   D   41 
T   62   T   43 
L   63   I   44 
S   64   I   45  
L   65   S   46 
P   66   E   47 
V   67   E   48 
D   68   D   49 
A   69   W   50 
Q   70   P   51 
V   71   T   52 
E   72   D   53 
V   73   W   54 
P   74   P   55 
I   75   V   56 
I   76   M   57 
E   77   E   58 
K   78   A   59 
F   79   A   60 
D   89   G   70 
I   91   I   71 
Q   92   P   72 
L   93   M   73 
M   94   R   74 
D   95   K   75 
M   96   S   76 
R   97   R   77 
D   98   D   78 
Y   99   M   79 
K   100  I   80 
T   101  E   81  
TER 
END 
