First | Previous  <   Home | Targets | Evaluation | Contact   >   Next | Last

T0467

© 2008 Grishin lab/HHMI/UTSW

updated on 2008-12-09

T0467

hypothetical membrane-associated protein BC4932 from Bacillus cereus

Target sequence:

>T0467 BcR97A, Bacillus cereus, 97 residues
MDLNRMGKDEYYVQITVDGKEVHSKADNGQKYKDYEYKLTGFDKDGKEKELEFTAQKNLRKEAFLRVYHSDKKGVSAWEEVKKDELPAKVKEKLGVK

Structure:

Method: NMR

Determined by: NESG

PDB ID: 2k5q

Domains:  PyMOL of domains

Single domain protein with poorly structured termini and loops. Structurally variable regions were removed from NMR models for evaluation of predictions.

Structure classification:

OB-fold. 5-stranded β-barrel, partly open between the 3rd and the 5th β-strands. The five strands are shown in the ribbon diagram below as dark blue, cyan-green, lime-green, yellow and orange. In addition to the 5-stranded OB-core, this structure is elaborated with the long insertion between β-strands 1 and 2 housing an α-helix, and a C-terminal α-helix.

CASP category:

Fold recognition. Fold similarity not detectable statistically by sequence methods known to us.

Closest templates:

The first COMPASS hit to OB-fold proteins is to the Nucleic acid - binding domain superfamily:

Subject = d1quqb_ b.40.4.3 Replication protein A 14 KDa (RPA14) subun
length=114  filtered_length=112  Neff=10.250  Smith-Waterman score = 38  E-value = 2.43e+01
QUERY         38     KLTGFDKDGKEKELEFTAQKNLRKEAFLRV==YHSDKKGVSAWEEVK  82
CONSENSUS_1   37     KLKGYDKDGKEKEVKFTADKKLKKGAYLKL==YVNKGKGVKSWEEVK  81
                     +L + + DG++ +V+++ + +L +G+Y+++  +VN++ +++  + ++
CONSENSUS_2   38     TLTLSSGDGGNVTVELNPPLDLEIGGYVEVIGKVNGDLTIRALSYID  84
d1quqb_       38     MFILSDGEGKNGTIELMEPLDEEISGIVEVVGRVTAKATILCTSYVQ  84

However, this is the hit #24 in the total list of hits and its E-value is around 25 (which indicates BTW how good COMPASS E-value estimate is), so it can hardy be called statistically significant. Moreover, SH3 domain proteins were found by COMPASS prior to this hit. Nevertheless, 9 consensus match residues in the above alignment shown as "dots" on the two OB-fold structures map to structurally equivalent positions validating the COMPASS alignment:

According to structure, proteins from 'Nucleic acid - binding' superfamily are also most similar to T0467, and those include phage ssDNA - binding proteins with an open barrel.

Target sequence - PDB file inconsistencies:

NMR models contain the C-terminal His-tag residues 98-LEHHHHHH-105, not present in the target sequence. This regions was removed from NMR models due to disorder anyway.

T0467    1 MDLNRMGKDEYYVQITVDGKEVHSKADNGQKYKDYEYKLTGFDKDGKEKELEFTAQKNLRKEAFLRVYHSDKKGVSAWEEVKKDELPAKVKEKLGVK-------- 97
           *******|||||||||||||||||||***|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||***~~~~~~~~
2k5qA    1 MDLNRMGKDEYYVQITVDGKEVHSKADNGQKYKDYEYKLTGFDKDGKEKELEFTAQKNLRKEAFLRVYHSDKKGVSAWEEVKKDELPAKVKEKLGVKLEHHHHHH 105

Sequence classification:

DUF1093 in Pfam, which includes just a few very close homologs. Below is PROMALS3D alignment of several most distant homologs detectable by PSI-BLAST. This alignment was not edited manually and may have errors. Biological sequences contain signal peptide (predicted as a helix at the N-terminus) not included in the target sequence.

Conservation:    9   7                                               9      5                 7    6      5      5               555                 9     6     5                                              

T0467         1  M---------------------------------------DLNR--MGKDEYYVQITVDGKEVHSKADNGQKYKDYEYKLTGFDKD---GKEKELEFTAQ-----KNLRKEAFLRVYHSDK---------KGVSAWEEVKKDELPAKVKEKLGVK------------------------------------   97

T0468         1  M--ERA----------------------------------SLNR--IGKDVYYMQIKGEGTIEKVDG-----RNLRNYTLPAYDED---GVKKQITFRSTKKENDHKLNKYAFLRLYVDQDDNSKNEISSIEVKSYEEIQKADLPEKVKDKFTIK------------------------------------  109

gi_163762713  1  M--KKT--------IIILISTLVSLAVIAVVII---ATI-DFNR--VGKDHAYTLITENGDTEETVTGSGEVFTRYHYSQSAFTDE---GEEITVSFSSH-----KNLRHGAYLRLYMNH----------TDVTSYDEVQTEDVPAQALETLMQE------------------------------------  121

gi_116334877  1  ----KG-----------------------------IGTIDRFDPF-LSYKVGYTKLPSEKEISPNNK------ATNGYWTAWFTDNTMGNEWTKLTLTTGL----NPGVKNPYVMAYHKG----------SYVKVARIVPRTYINQTYVSNLAALDALLNHKGNTATNNHDLKKYKTQIHISDTVQQLVTE  137

gi_49398117   1  M--KRT--------LIIV--TIVLLIIILSLLLVRNEHLDRFNPL-LKEKVSYAKVPKDT--------------QDYRNITVYSKN---GEKKSYKLDFLG-----YDPTKEYVKVNHKG----------KYVRSIEYIT-KIFLAF--------------------------------------------  101

gi_116492027  1  M--RKG--------MIIS--GVVVVILLITF----GSAYFWYNSN-YGTQDYYTQIVDKGTKFVEKDDSNRSYVNYAYHQPAYNQA---SQKITVKFNGNL---ERPLKLHAYLKIGYNAKR--------NQVISWSKVDKKDVPKAALKQINP-------------------------------------  123

gi_69249191   1  M--KK---------LIGV--LVVL---GIFL----GGSFAAYHYF-YGGEAYYTKITTNGEMSATQADNGEKFTTYTYKQAAYNKD---GEKKQVTLREER---EHPLKMNAYLKLKVNPR---------KGVLSWEEIKASEVPKNAAEKIDN-------------------------------------  118

gi_116873589  1  MITKKR-------VLISL-----ATVLLAVCA-----IWEYAMPTDAAVKSYYLKVAEAGKPVKKN-----QFKGYEYTSKVYDDK---GKQKEIKFYSEK-----ELTKNEQFKVMIGEN---------KMVVNYKKIK--NVPDKIKYLAEK-------------------------------------  113

gi_65320434   1  M--NKRRNDYEDFVLCIL--GIGLPIYLLPFIL--RGDFDRFNPI-AEEKNVYAIAKGYGVPDYH------HKGRAMYSLKSFDES---GSRNEYTVGTST---PNDFIRKTYLRIHVKG----------KYVYSYEAISEKDIPEKIRKQLEV------EIK----------------------------  128

Consensus_aa:    M..c+.................................h..hp...h..cs.Yhbl...sp...pp......b....Yp..s@scp...tpp.phph........p....p.@l+l.hp...........p.V.shcblppp.ls..h.pph.......................................

Consensus_ss:                     h  hh hhhhhhhh                  eeeeeee     ee          eeeeeeeeee        eeeeeee              eeeeeee            eeeehhh  hhh  hhhhhhh                                       

Comments:

This protein is a close homolog of T0468:

PSIPRED  ------------EEEEEEE---EEEE-------EEEEEEEEEEEE-----EEEEEEE--------------EEEEEE--------------EEEEEEE------HHHHHHH---

T0467    MD---LNRMGKDEYYVQITVDGKEVHSKADNGQKYKDYEYKLTGFDKDGKEKELEFTAQK-----NLRKEAFLRVYHSDK---------KGVSAWEEVKKDELPAKVKEKLGVK

T0468    MERASLNRIGKDVYYMQIKGEGTIEKV-----DGRNLRNYTLPAYDEDGVKKQITFRSTKKENDHKLNKYAFLRLYVDQDDNSKNEISSIEVKSYEEIQKADLPEKVKDKFTIK

PSIPRED  ------------EEEEEE----EEE-----------EEEEEEEEE-----EEEEEEE--------------EEEEEEE------------EEEEHHH--HHH--HHHHHHH---

This OB-fold was predicted de-novo by ROSETTA, which correctly indicated an open barrel. This prediction was very suggestive of the correct structure, because ROSETTA is biased towards local β-strand pairing, and OB-fold has a crossing loop to form H-bonds between β-strands 1 and 4, not to mention barrel closure present in most OB-folds with H-bonds between β-strands 3 and 5. Metaserver bioinfo.pl fails to find similarity to other OB-fold proteins, and incorrectly provides SH3-like predictions. Although some servers used OB-folds as templates, it is unclear how significant those predictions are, as the templates were closed barrels and many severs used SH3-fold templates as well. The bias towards SH3 fold is likely caused by the C-terminal region, which shows strong local conformational similarity to Sso7d Chromo-domain DNA-binding proteins.

This similarity covers about 30 residues (half of Sso7d SH3-fold) and spans through a β-hairpin and two α-helices. The first helix is a single-turn helix characteristic of SH3 fold. Such a helix frequently structured as a 310-helix is present at this spatial location in the majority of SH3 fold proteins. Mutual orientations of the four secondary structural elements between the two fragments (from OB and SH3) is very similar, and sequence similarity reflects this.

2k5q   64 FLRVYHSDKKGVSAWEEVKKDELPAKVKEKLGV 96

2bf4_A 30 ISFTYDEGG-GKTGRGAVSEKDAPKELLQMLEK 61

          EEEEEEEE  EEEEEEEEHHHH HHHHHHHHHH   

Therefore it is not surprising that servers find this similarity. However, upon closer inspection there are several positions with charged residues (highlighted red in the above alignment) aligned to hydrophobs (yellow). These positions have different exposure to solvent in the two structures and hint that the local similarity may not translate to global fold similarity. Indeed, the two C-terminal helices are essential core elements is SH3 structure of Sso7d, but are peripheral surface helices not contributing to the core in OB-fold of T0467. In addition, the surface of the hairpin buried in SH3 fold is exposed in T0467. These inconsistencies in hydrophobic patterns are very suggestive of global structural differences:

DALI-LITE does not see this local similarity, but aligns the central 3-stranded meander β-sheet in two proteins (shown in the back on the images above: blue-cyan-green in T0467, and green-yellow in Sso7d), albeit with a very low Z-score of 0.3 (33 residues, RMSD 2Å). We agree that this is a globally meaningful alignment (only residues in capital letters are aligned), as it superimposes hydrophobic cores of both proteins.

DSSP   ....lllllllleeeeeELLLllLEEHhhhllllllLEEEEEEEEEllllLEEEEEEEELL..LLLLlleeeeeelllllllleeeelllllllhhhhhllllllllllll

2k5q   ....mdlnrmgkdeyyvQITVdgKEVHskadngqkyKDYEYKLTGFdkdgKEKELEFTAQK..NLRKeaflrvyhsdkkgvsaweevkkdelpakvkeklgvklehhhhhh

2bf4_A atvkfkykgeekevdisKIKK.vWRVG.........KMISFTYDEG....GGKTGRGAVSEkdAPKE..................................llqmlekqkk

DSSP   leeeeeelleeeeeehhHEEE.eEEEL.........LEEEEEEELH....HHLEEEEEEELllLLHH..................................hhhhhhhlll

As a summary, superposition of locally similar fragments does not result in a global superpositions of structural cores of the two folds, and does not result in a reasonable fold prediction. Vice versa, global superposition of the cores leaves the two locally similar fragments as non-equivalent parts of the two proteins, as they occupy very different spatial locations and carry out different structural roles.

Global similarity between OB and SH3 folds has been noticed previously1), and explained in terms of very distant evolutionary relationship (homology). It is possible that the two folds shared a common ancestor and are indeed homologous over the 3-stranded curved meander sheet, although definitive evidence for this is lacking. The meaning of the local fragment similarity between T0467 and Sso7d-like chromo-domain OB-fold proteins is unclear. It is conceivable that these similar fragments originated independently and conformational resemblance between them is due to chance.

Server predictions:

T0467:pdb 8-94:seq 8-94:FR;   alignment

click on a score in the table below to display the model in PyMOL