Transcript Unknown function, JCSG

Comparative Analysis of Novel Proteins from the CATH Family of Zinc Peptidases
Debanu Das1,2, Abhinav Kumar1,2, Lukasz Jaroszewski1,3 and Ashley Deacon1,2
1Joint
Center for Structural Genomics, 2Stanford Synchrotron Radiation Laboratory, Menlo Park, CA 94025,
3Burnham Institute, La Jolla, CA, 92037
I. Introduction
In March 2007, the JCSG initiated a study of the CATH homologous superfamily of zinc peptidases (3.40.630.10), which
belong to the phosphorylase/hydrolase-like fold in SCOP. These proteins show significant sequence divergence and have
a broad phylogenetic distribution across both prokaryotes and eukaryotes. At the time, despite 40 unique experimentally
determined structures in the PDB, only half of the family members had reliable homology models.
JCSG is improving the structural coverage by determining novel structures which share <30% sequence identity with
those in the PDB. Hidden Markov Models from the CATH database were used to identify sequences in the JCSG
genome pool. PSI-Blast seeded with these sequences was used to find additional proteins. These two sets yielded 226
unique targets. After removing targets with more than 30% sequence identity to any PDB structure or to any crystallized
target from a structural genomics center, 161 targets remained. Further clustering at 90% (in order to avoid nearly
identical sequences), produced a set of 137 targets. To date, JCSG has contributed 6 new structures to the family and 7
other targets have been crystallized.
We present our progress towards complete structural coverage of this family, highlighting common and variant structural
features that support different molecular and cellular roles, focusing on active site residues, ligand binding, protein size
and oligomerization state. This analysis may provide insights into structural themes that dictate protein function and also
allows modeling of protein structures related by sequence. Our structures serve as a nucleation point for the design of
further structure-based experiments to probe the biochemical and biomedical roles of these proteins.
III. General structure and biochemistry
II. Background and Significance
CATH 3.40.630.10 proteins are distributed across 8 PFAM families, which form the large
peptidase_MH clan (CL0035). Also known in the MEROPS peptidase database as the clan
MH/MC/MF of metallopeptidases:
PF04952
Succinylglutamate desuccinylase / Aspartoacylase family
(AstE-AspA )
458 proteins
PF02127
Aminopeptidase I Zinc metalloprotease M18
227
1 JCSG structures, 5 all other SG
4 all other SG
PF01546
Peptidase family M20/M25/M40
3779
3 JCSG structures, 7 all other SG
6 non-SG
PF00246
Zinc carboxypeptidase M14
1013
2 JCSG structures
10 non-SG
PF04389
Peptidase family M28
812
PF00883
Cytosol aminopeptidase family, catalytic domain
827
1 all other SG
1 non-SG
PF05343
M42 Glutamyl aminopeptidase
427
1 JCSG structures, 1 all other SG
1 non-SG
PF05450
Nicastrin (eukaryotic, not known to be peptidase, part of γsecretase complex, no structures)
48
None
None
5 non-SG
These proteins are involved in a variety of proteolytic activities, have a range of substrate
specificities and are present in numerous microbial organisms, many of which are important
human pathogens like S. aureus, S. typhimurium, T. vaginalis, M. tuberculosis, N.
gonorrhea, N. meningitidis, C. trachomatis, G. intestinalis, and E. coli. Several of these
proteins have been investigated for their therapeutic potential and diseases roles (Canavan’s
disease, cancer therapy and prohormone/propeptide processing).
IV. Progress of structure determination
These metallopeptidases show a high degree of structural conservation in the CATH
domain which has a α/β/α sandwich architecture. The
active site usually comprises of histidines and
carboxylates interacting with two zinc ions.
Despite the variety of molecular functions
and substrate specificities of these proteins, the
catalysis most likely involves a hydroxyl ion
ligand involved in a nucleophilic attack. The full
proteins often oligomerize and display some
differences in their oligomerization state, however,
the exact role of the oligomer in the molecular function
is still unclear. In some cases, dimer formation results in
assembly of a productive catalytic site. Dimerization is
usually mediated by a dimerization domain. Higher
oligomeric forms such as tetramers or octamers are also
observed for some proteins.
Distribution of selected targets
across Pfam families
Current status of 137 targets
140
120
Targets assigned
in PfamA
Targets unassigned
in PfamA *
100
PF04952
32
3
80
60
PF02127
0
0
40
PF01546
56
1
All targets selected in March 2007
20
0
PF00246
9
8
PF04389
10
7**
PF00883
2
0
PF05343
5
1
PF05450
0
7**
Selected
Cloned
Expressed
Purified
Crystallized Diffracted
Solved
PDB
Deposit
* PFAM assigned based on sequence homology detected with FFAS http://ffas.ljcrf.edu/ffas-cgi/cgi/ffas.pl
There are 3 targets not assigned by PfamA or FFAS.
** 7 targets indicated show significant FFAS match to both PF04389 and PF05450, and could possibly be distant
bacterial homologs to the exclusively eukaryotic nicastrin family (PF05450).
Figure of the representative CATH structure fro
http://cathwww.biochem.ucl.ac.uk/cgi-bin/cath/GotoCath.pl?cath=3.40.630.10
V. Structures solved by JCSG
2QJ8.pdb (HP10622H), 2.0Å, R/Rf= 20.7/25.4%,
Unknown function, PF04952
Homolog involved in Canavan’s disease
2QVP.pdb (HP10645A), 2.0Å, R/Rf= 16.1/21.3%
Unknown function, PF04952
Structure suggests target may be closer in homology
To PF00246 family
VI. Phylogenetic tree and structure tree
3B2Y.pdb (HP10645E), 1.74Å, R/Rfr=17.45/21.51%
Unknown function, PF04952, Ni+2 bound
Structure suggests target may be closer in homology
To PF00246 family
2RB7.pdb (HP1666A), 1.6Å, R/Rfr=15.4/18.0%
Unknown function, PF01546
48 close homologs from important human pathogens
Potential in cancer therapy
VIII. Comparison of two proteins with >30% sequence
identity within the same Pfam
PF01546: 1CG2, 2RB7
Sequence with
>30% identity
within
a particular
Pfam also
cluster
together in
structure space
Target was selected based on 30% sequence id over full length
protein and despite similairty in CATH domain, more diversity
is observed in the dimerization domain.
2QYV.pdb (HP9625C), 2.11Å, R/Rf= 22.0, 24.4%
Putative Xaa-His dipeptidase, PF01546, Zn+2 bound
7 close homologs from important human pathogens
XI. Elucidation of a unique oligomeric form
2RB7 (cyan) and 1CG2, PF01546. Proteins in this Pfam with solved structures and >30% seq id with one another have
functions which include succinyl-diaminopimelate desuccinylase activity; Carboxypeptidase G2 which cleaves C-terminal
glutamate moiety from folic acid and its analogues, such as methotrexate; N-acetyl-L-citrulline deacetylase and Peptidase T
tripeptidase.
The 2QYV (PepD, MEROPS M20.007, clan MH, subfamily C) monomer is
very similar in structure to the 1LFW monomer (PepV, MEROPS M20.004,
subfamily A). Both are dipeptidases belonging to PF01546. However, 1LFW is
known to function as a monomer in which the molecular structure mimics that
of a dimer seen in most other proteins in this Pfam. PepD in E. coli and
PrevotellaAcknowledgements
albensis are seen to function as dimers. 2QYV represents the first
crystal structure of a PepD, revealing it to be dimeric in the crystal structure
(monomers in magenta and gold) as well as by size exclusion chromatography
and shows the structural nature of the dimer. This novel structure serves as a
starting point for further experiments to probe the effect of this unique dimer
formation on protein
function.
Based on this information, it would now be
possible to perform targeted biochemical
assays to determine substrate for 2RB7, to
try to understand the structural basis for
substrate selection and specificity and to
exploit this information for its therapeutic
potential. For example, can 2RB7
hydrolyse methotrexate? Can it do so more
efficiently? Can active site engineering
based on structural information produce a
more potent enzyme?
Active site in 2RB7
1CG2:C-terminal glutamate moiety
from folic acid and its analogues,
such as methotrexate
Common core ~290 aa, RMSD ~3.0 Å
http://fatcat.burnham.org/POSA
For structures that cluster together at 30% level, structural conservation in the common core is the highest,
Generally only slight rearrangement of secondary structural elements is observed (within the domain).
VII. Suggestion of PfamA assignment based on structure
IX. Proteins with <30% sequence id. within the same Pfam
HP10645A (2QVP) and HP10645E (3B2Y) sequences are assigned to PF04952 in PfamB. However,
structural comparisons of the CATH domain show a stronger similarity to a member of PF00246 (1QMU,
left) than to a member of PF04952 (2QJ8, center). This is also supported by structure & phylogenetic trees
and FFAS. Also, like 1QMU, HP10645A/E lacks an ~70 amino acid insertion that forms a “C-terminal
domain” (right, black circle), which is present in all PF04952 proteins and is important for biochemical
function. These two pieces of evidence suggest and support the assignment of HP10645A/E in PF00246
in PfamA.
PF01546: 2RB7, 2QYV (green)
Common core ~250 aa, RMSD ~3.0 Å
HP10625B, 2.3Å, work in progress
PF01546
50 close homologs from important human pathogens
Potential in cancer therapy
X. Active site study may lead to structural basis of substrate specificity
2RB7: Unknown function, JCSG
http://www.phlogeny.fr
2FVG.pdb (TM1049), 2.01Å, R/Rf= 20.3/24.4%
Endoglucanase, PF05343
27 close homologs from important human pathogens
Active site is 1CG2 is
H112, D141, E200, E176, H385
Based on this, putative active
site in 2RB7 is H72, D99, D100,
E138, E139, D162
Hydrolysis of methotrexate by 1CG2
XII. Inferences and further work
•In the quest for increasing structural coverage across protein families, it is expected that proteins similar in sequence within a protein family will be similar in structure. Increasing structural coverage
provides better templates for modeling other proteins. The comparative structural analysis presented here provides experimental verification of the validity of this approach.
•The structures for the proteins HP10645A and HP10645E suggest that they should be assigned to PF00246 in PfamA instead of the current suggestion of belonging to PF04952 by PfamB.
• The 7 structures presented here provide a basis for enhancing the modeling of 2177 out of 7591 proteins (~29%) belonging to this Pfam clan. Furthermore, 3 of these JCSG structures provide the first
examples of structures for proteins within a particular sequence cluster (2QYV, 2QJ8 and 3B2Y) and thus provide the basis for modeling 384 unique proteins (10 from organisms listed as top human
pathogens) belonging to these 3 clusters from 2 different Pfams (PF01546 and PF04952).
Common core of 226 aa, RMSD 2.45 Å
Common core of 191 aa, RMSD 2.49 Å
GNF & TSRI
(Crystallomics Core)
UCSD & Burnham
(Bioinformatics Core)
John Wooley
Lukasz Jaroszewski
Lian Duan
Natasha Sefcovic
Andrew Morse
Tamara Astakhova
Cindy Cook
Adam Godzik
Slawomir Grzechnik
Sri Krishna Subramanian
Piotr Kozbial
Prasad Burra
Josie Alaoen
Dana Weekes
Superimposition of all 6
structures in PF04952:
1YW4, 1YW6, 2BCO,
2G9D, 2GU2 and 2QJ8
Scott Lesley
Dennis Carlton
Marc Deller
Polat Abdubek
Julie Feuerhelm
Hope Johnson
Sebastian Sudek
Glen Spraggon
Charlene Cho
Jessica Canseco
Mark Knuth
Heath Klock
Thomas Clayton
Kevin D. Murphy
Daniel McMullan
Christina Trout
Claire Acosta
Linda M. Columbus
Joanna C. Hale
Thamara Janaratne
Linda Okach
Edward Nigoghossian
Aprilfawn White
Bernhard Geierstanger
Ylva Elias
Sanjay Agarwalla
Bi-Ying YehAnna Grzechnik
Mimmi Brown
Larger rearrangements and extensions of secondary
structural elements. Inserts and novel features more common.
Common core ~190 aa, RMSD ~3.0 Å
PF04952: 2QJ8, 3B2Y (cyan)
• 2QYV/HP9625C represents the first crystal structure of a dipeptidase PepD showing a dimer.
• Further analysis will be performed to try to understand evolutionary relationships between these proteins based on sequence-based phylogenetic trees and structure-based trees.
• Attempts will be made to investigate use of these structures and their comparative analyses in understanding structural basis for enzyme function and substrate specificities by analysis of active site amino
acids, and to attempt to exploit information for therapeutic purposes.
Stanford /SSRL
(Structure Determination Core)
Keith Hodgson
Mitchell Miller
Hsiu-Ju (Jessica) Chiu
Christopher Rife
Silvya Oommachen
Henry van den Bedem
Christine Trame
Ashley Deacon
Debanu Das
Kevin Jin
Qingping Xu
Scott Talafuse
Ronald Reyes
TSRI
(NMR Core)
Kurt Wüthrich
Reto Horst
Maggie Johnson
Amaranth Chatterjee
Michael Geralt
Wojtek Augustyniak
Pedro Serrano
Bill Pedrini
William Placzek
TSRI
(Admin Core)
Ian Wilson
Marc Elsliger
Gye Won Han
David Marciano
Henry Tien
Xiaoping Dai
Lisa van Veen
Scientific Advisory Board
Sir Tom Blundell
Univ. Cambridge
Homme Hellinga
Duke University Medical Center
James Naismith
The Scottish Structural Proteomics facility
Univ. St. Andrews
Soichi Wakatsuki
Photon Factory, KEK, Japan
James Wells
UC San Francisco
Robert Stroud
Center for Structure of Membrane Proteins
Membrane Protein Expression Center
UC San Francisco
James Paulson
Consortium for Functional Glycomics
The Scripps Research Institute
Todd Yeates
UCLA-DOE Inst. for Genomics and Proteomics
The JCSG is supported by the NIH Protein Structure
Initiative (PSI) Grant U54 GM074898 from NIGMS
(www.nigms.nih.gov). Portions of this research were
carried out at the Stanford Synchrotron Radiation
Laboratory (SSRL). The SSRL is a national user facility
operated by Stanford University on behalf of the U.S.
Department of Energy, Office of Basic Energy Sciences.
The SSRL Structural Molecular Biology Program is
supported by the Department of Energy, Office of
Biological and Environmental Research, and by the NIH.
Annual meeting with SAB 2007