 
PFRMAT TS 
TARGET T0052 
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, M.Cueto, F.E.Cohen & S.A.Benner. 
METHOD Important comment: We submit T0052 as a "modeling 
METHOD exercise" as the lack of homologous sequences in the 
METHOD databases precluded the use of the systematic methods 
METHOD developed by members of our team. In particular, the 
METHOD prediction is NOT based on the SAINT secondary structure 
METHOD prediction system (Benner et al., U.of Florida). 
METHOD To illustrate this fact, we submit coordinates only. 
METHOD 
METHOD Methodology used in this prediction for T0052 (CN-V): 
METHOD The prediction process was mostly manual and based on 
METHOD different, sometimes contradicting, clues found through  
METHOD sequence analysis and in the literature, as described 
METHOD below.  
METHOD Firstly, we inspected the automated secondary structure 
METHOD predictions available to us and potentially useful even 
METHOD if used in the absence of homologous sequence information 
METHOD (PHD and J.M.Chandonia, personal communication). The in- 
METHOD consistencies between the outputs by different methods 
METHOD prompted us to disregard the predicted secondary structure 
METHOD in this case! Instead, we attempted to conclude from the 
METHOD position of the DISULFIDE BRIDGES and the indications in  
METHOD the sequence on the number and approximate position of  
METHOD SEGMENTS OF POSSIBLE SECONDARY STRUCTURE. On these grounds,  
METHOD we favor a model with a total of up to 7 short segments, or half- 
METHOD segments, probably separated only by a bend in the protein chain  
METHOD (in contrast to a sharp turn). Thus the 7 short segments are  
METHOD likely to be found in approximately five stretches of secondary 
METHOD structure. Further, like others before us, we found a putative  
METHOD INTERNAL REPEAT in the target sequence (#1-50, 51-101).  
METHOD Interestingly, this piece of information was IN CONFLICT 
METHOD WITH A POSSIBLE, DISTANT HOMOLOGY TO 1PMD that was found through 
METHOD "second frame" homology searches. 
METHOD With this information, wea ttempted to perform an EXHAUSTIVE 
METHOD COMBINATORIAL ANALYSIS of the possible folding topology of T0052, 
METHOD assuming that the repeated domains are superimposable and contain 
METHOD predominately (exclusively) beta-sheet structure. 
METHOD However, the analysis using a similar scheme as we applied 
METHOD quite successfully in CASP1 on synaptotagmin (Proteins 
METHOD 1995, 22:299-310), FAILED TO REDUCE THE NUMBER OF POSSIBLE 
METHOD FOLDING TOPOLOGIES to the desired level for submission at 
METHOD at CASP3. This is because we could not find sufficiently 
METHOD strong and reliable "filters", for example a prediction 
METHOD of tight turns vs. wide turns/connections as is often 
METHOD possible with the help of multiple sequence information. 
METHOD 
METHOD Therefore, and due to time pressure, we submit our model 
METHOD as an "educated guess", illustrating that we decided to 
METHOD rely on the postulated internal repeat rather than on the 
METHOD the conflicting secondary structures. Further, we disre- 
METHOD garded the output of the UCLA-DOE fold recognition server 
METHOD programs as no significant matches were returned.  
METHOD Instead, the packing constraints for the putative 50 res. 
METHOD domains seem to FAVOR AN ALL-BETA FOLD PREDICTION. 
METHOD The possible homology to 1pmd (163-263) was, again,  
METHOD conflicting with an internally repeated structure made of 
METHOD two lobes. Intriguingly, however, we are able to propose 
METHOD a model predicting partial structural similarity of T0052 
METHOD with a small domain in 1prc (photosythetic react. center) 
METHOD which, if duplicated as putatively in T0052, has some 
METHOD resemblance in shape with the domain of interest in 1pmd. 
METHOD As the crystal structure 1pmd is at low resolution and 
METHOD undoubtedly faulty, this could be an indication for how 
METHOD to correct 1pmd, if correct. However, according to our 
METHOD combinatorial analysis (see above), there are several 
METHOD other topologies that would meet all the "filtering" 
METHOD constraints we could bring forward. For example, a  
METHOD model made of two symmetry-equivalent SH3-like domains 
METHOD would also be conceivable. 
METHOD 
METHOD Both of the lobes of our model are constructed partly 
METHOD onto the topology of 1prc (res. 155h-196h). The remaining 
METHOD parts of the model, and the possible ROTATIONAL SYMMETRY 
METHOD between the two lobes was modeled manually, ab initio.  
METHOD (coordinates generated using SYBYL, Tripos). The ab initio 
METHOD fragments are mostly to accommodate the necessary changes 
METHOD in the structure for adopting a two-lobe structure with 
METHOD a shared part of the core. As such, the overall folding 
METHOD arrangement (topology) is more important in our prediction  
METHOD than the detail conformation of the polypeptide  
METHOD chain of the shortfragments, in the case of T0052! 
METHOD Backbone and C-beta coordinate are submitted, but the 
METHOD model is to be considered at the resolution equivalent 
METHOD to not more than a wire model... 
MODEL 1 
REMARK     T0052 Cyanovirin-N, Nostoc ellipsosporum 
PARENT N/A               
ATOM    713  N   ASP    95      -1.998  -2.076   7.482  1.00 10.00               
ATOM    714  CA  ASP    95      -1.346  -1.921   8.781  1.00 10.00               
ATOM    715  C   ASP    95       0.010  -2.590   8.934  1.00 10.00               
ATOM    716  O   ASP    95       0.009  -3.746   9.308  1.00 10.00               
ATOM    717  CB  ASP    95      -2.259  -2.673   9.790  1.00 10.00               
ATOM    721  N   GLY    96       1.178  -1.940   8.710  1.00 10.00               
ATOM    722  CA  GLY    96       2.484  -2.573   8.944  1.00 10.00               
ATOM    723  C   GLY    96       2.723  -4.043   8.642  1.00 10.00               
ATOM    724  O   GLY    96       3.404  -4.341   7.673  1.00 10.00               
ATOM    725  N   THR    97       2.189  -4.958   9.487  1.00 10.00               
ATOM    726  CA  THR    97       2.217  -6.404   9.211  1.00 10.00               
ATOM    727  C   THR    97       1.036  -6.847   8.348  1.00 10.00               
ATOM    728  O   THR    97       1.191  -7.768   7.562  1.00 10.00               
ATOM    729  CB  THR    97       2.069  -7.132  10.564  1.00 10.00               
ATOM    732  N   LEU    98      -0.127  -6.170   8.513  1.00 10.00               
ATOM    733  CA  LEU    98      -1.318  -6.286   7.689  1.00 10.00               
ATOM    734  C   LEU    98      -1.837  -7.569   7.108  1.00 10.00               
ATOM    735  O   LEU    98      -1.148  -8.555   6.903  1.00 10.00               
ATOM    736  CB  LEU    98      -1.079  -5.356   6.486  1.00 10.00               
ATOM    740  N   LYS    99      -3.127  -7.437   6.717  1.00 10.00               
ATOM    741  CA  LYS    99      -3.608  -8.346   5.707  1.00 10.00               
ATOM    742  C   LYS    99      -4.347  -7.681   4.570  1.00 10.00               
ATOM    743  O   LYS    99      -4.752  -6.529   4.645  1.00 10.00               
ATOM    744  CB  LYS    99      -4.528  -9.377   6.364  1.00 10.00               
ATOM    749  N   TYR   100      -4.537  -8.488   3.502  1.00 10.00               
ATOM    750  CA  TYR   100      -5.312  -8.036   2.360  1.00 10.00               
ATOM    751  C   TYR   100      -6.660  -8.707   2.367  1.00 10.00               
ATOM    752  O   TYR   100      -6.883  -9.656   3.105  1.00 10.00               
ATOM    753  CB  TYR   100      -4.542  -8.433   1.082  1.00 10.00               
TER 
END 
