PEDS Advance Access originally published online on February 20, 2004
Protein Engineering Design and Selection 2004 17(2):175-182; doi:10.1093/protein/gzh024
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2004 Oxford University Press
Distribution of proline-rich (PxxP) motifs in distinct proteomes: functional and therapeutic implications for malaria and tuberculosis
Malaria Group, International Centre for Genetic Engineering and Biotechnology, Aruna Asaf Ali Marg, New Delhi-110067, India
1 To whom correspondence should be addressed. e-mail: asharma{at}icgeb.res.in
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
We have conducted a survey of proline-rich (PxxP) motifs in the proteomes of human, mouse, yeast, Mycobacterium tuberculosis and Plasmodium falciparum. Our analyses reveal a strikingly high occurrence of these motifs in each organism, suggesting a wide dependence on protein–protein interaction networks in cellular systems. All proteomes considered have an abundance of PxxP motifs which can potentially participate in binding to SH3 domain-containing proteins. A large fraction of these motifs can be assigned to structurally conserved types of class I and class II sequences. We propose that while maintaining the primary biochemical function, many proteins are likely to participate in additional interactions involving molecular cross-talk with other proteins using proline-rich and other motifs. We have also identified PxxP-containing motifs that are unique to P.falciparum and M.tuberculosis. These sequences may serve as leads for the development of peptidomimics that specifically target these organisms. We propose a novel drug target selection strategy where shared PxxP-containing motifs can be used to direct the development of inhibitors that focus on multiple targets in the cell. Screening for such unique PxxP-containing motifs in the P.falciparum proteome yielded highly conserved sequences in the variant surface antigen family that can be used to initiate design of peptidomimics that may potentially abrogate parasite cytoadherence during malaria infections.
Keywords: peptidomimics/protein–protein interactions/PxxP motifs/SH3 modules/signal transduction
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The assignment of biological function to new proteins remains a formidable problem. Integration of proteins into biochemical and signaling pathways is an essential requisite for understanding each protein. The physiological relevance of each protein can be highlighted in view of its interactions with numerous other proteins that co-exist in the cellular milieu. Although we now have genomic information on a variety of organisms, our understanding of the biological function(s) of most new proteins remains limited. This point is best illustrated by analysis of the Plasmodium falciparum proteome where
65% of the proteins are of unknown function. Rapid protein sequence divergence often makes sequence comparisons difficult, leading to a failure in detecting evolutionary conservation of function (Florens et al., 2002
). Further, genomic and structural information may be insufficient in assigning all the biochemical attributes to a protein, given that multiple cellular functions may be associated with the same protein fold.
Protein–protein interactions underpin most molecular cross-talk in cells and these interactions can occur over short regions, often <10 amino acids in length (Cohen et al., 1995; Kay et al., 2000; Mayer, 2001). Aside from a primary biochemical role, many proteins have additional functional attributes that are utilized in networking with cellular proteins. To define the numerous protein–protein interactions within a cell, researchers still rely on laborious laboratory techniques such as yeast 2-hybrid systems, c-DNA expression library screening and co-immunoprecipitation (Cohen et al., 1995; Sudol, 1998; Kay et al., 2000; Mayer, 2001). Interactions mediated by modules such as the Src homology (SH) 2 and 3 domains, WW domains, post-synaptic density/disc-large/ ZOI (PDZ) domains and Eps 15 homology domains (EH domains) display a typical mode of interaction by recognizing a linear region of 3–9 amino acids (Sudol, 1998; Kay et al., 2000; Mayer, 2001). The amino acid proline attains paramount importance in many protein–protein interactions (Sudol, 1998; Kay et al., 2000; Mayer, 2001). SH3 modules along with several other domains prefer ligand sequences that are proline-rich. SH3 domains are 60 residue modules which often occur in signaling and cytoskeletal proteins (Dalgarno et al., 1997; Sudol, 1998; Kay et al., 2000). These structurally conserved domains are ubiquitous in biological systems, including in a bacterium such as the Mycobacterium (Ponting et al., 1999; Feese et al., 2001). Binding of SH3 domains to simple peptides folded in to proline-rich II (PPII) helix is likely to govern the formation of a large number of protein complexes (Tong et al., 2002). Most proteins known to interact with SH3 domains contain at least one copy of the motif PxxP. SH3 domains bind to sequences that adopt a left-handed PPII helical structure in which the two invariant proline residues are found on the same face of the peptide and participate in hydrophobic interactions (Feng et al., 1994; Lim et al., 1994). Recent data suggest that SH3 domains can also interact with non-PXXP motifs (Kang et al., 2000; Kami et al., 2002), although the binding sites for non-PxxP-containing peptides and for the classical PxxP motifs do not overlap (Kami et al., 2002).
Analysis of phage display libraries have demonstrated that individual SH3 domains have distinct specificities for potential ligands and can discern subtle differences in the ligand primary structure (Sparks et al., 1996). Much of this specificity comes from amino acids flanking the core PxxP motif (Sparks et al., 1996). The role of PxxP-mediated protein–protein interactions in a multitude of cellular processes is therefore central. We reasoned that dissecting the proteomes for the presence of PxxP sequences would open new vistas in understanding sets of protein–protein interactions and signaling pathways. We therefore performed extensive in silico sieving of the complete proteomes of two pathogenic organisms, P.falciparum and Mycobacterium tuberculosis, along with the proteomes of Schizosaccharomyces pombe, Mus musculus and Homo sapiens for the prevalence of PxxP motifs. We defined a set of PxxP motifs that have been experimentally verified to bind to SH3 domains (Cesareni et al., 2002; Tong et al., 2002) and screened these five proteomes. We used two strategies to identify putative SH3 ligands in these proteomes: (a) screening with a consensus motif and (b) screening with ‘specific’ motifs. Our results indicate an abundance of class I and class II PxxP motifs in the proteomes of two major pathogenic organisms and three higher eukaryotic organisms. Our analyses have also identified shared motifs in the proteomes of M.tuberculosis and P.falciparum which can potentially be used for the development of peptidomimics against both these organisms. We propose a novel strategy of drug target selection where multiple proteins in the cell can be simultaneously inhibited using peptidomimics that target common PxxP motifs. In line with this strategy, we have identified several motifs that are highly conserved in the cytoplasmic (the erythrocyte cytoplasm) domains of P.falciparum variant surface antigens (also called P.falciparum erythrocyte membrane proteins or PfEMPs). These PxxP motifs are absent from the human proteome. The PfEMP1 family proteins undergo antigenic variation in P.falciparum and are crucially responsible for the cytoadherence associated mortality in P.falciparum malaria (Smith et al., 1995; Su et al., 1995). The development of peptidomimics that may disrupt the normal functioning of PfEMPs can provide a new focus for the development of anti-malarials that target cytoadherence of P.falciparum.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The proteomes should harbor a specific number of PxxP motifs depending on the average size of the protein and the size of the proteome. To address this issue, we used the simple probability concept. We calculated the probability of the occurrence of a particular motif in any random sequence by the multiplication rule of probability. For example, the probability of occurrence of the motif PxxPxR is P = (1/Pro)x1x1x(1/Pro)x(1/Arg) taking into account the expected frequency for each amino acid (represented by Pro, Arg in this case). This calculation is based on the ATGC content of each genome studied and on codon degeneracy. Further, the number of occurrences of PxxPxR in any protein of average size n (amino acids) is A = nP. To find the expected number of these motifs in the entire proteome, the above value (A) is multiplied by the total number of proteins in the proteome. Occurrences of each motif in the five proteomes were determined and expected distributions were computed. This allowed us to calculate the difference between observed and expected values. We also calculated chi-square values for the occurrence of each motif and observed significant deviations between observed and expected frequencies. The observed and expected frequencies along logarithmic values of respective motif probabilities are represented graphically in Figure 1. These analyses resulted in PxxP motifs that were either under- or over-represented, suggesting a non-random distribution. Statistical under- or over-representation indicates that the motif is likely to be of functional significance. Since the probability values are very low and therefore difficult to represent graphically, their logarithmic values are presented (logarithm to base 10). Chi-square values were calculated for goodness of fit between observed and expected numbers using the formula (O – E)2/E, where O is observed number and E is expected number. The chi-square values were calculated with 99% confidence interval.
|
2 values are as follows: (a) M.tuberculosis, 1K, 2K and 1@: 565.5, 556 and 54.96; (b) P.falciparum, 1K, 2K and 1@: 867.7, 552.8 and 470.8; (c) S.pombe, 1K, 2K and 1@: 233.3, 268.3 and 124.6; (d) M.musculus, 1K, 2K and 1@: 2802.4, 2709.7 and 64.9; and (e) H.sapiens, 1K, 2K and 1@: 353.6, 476.32 and 556.6.Simple Unix and PERL scripts were used to automate the motif searches using data from public access genomic and proteomic databases. The extent of solvent exposure of PxxP motifs in three-dimensional protein structures was assessed by using both PROCHECK and MODELLER programs (Sali et al., 1995
; Laskowski et al., 1996). |
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
An abundance of PxxP motifs in various proteomes
We identified all proteins with at least one occurrence of the PxxP motif in the proteomes of human, mouse, S.pombe, M.tuberculosis and P.falciparum. The resulting motifs were delineated into classes and subclasses keeping in view published nomenclatures (Rickles et al., 1995; Sudol, 1998; Cesareni et al., 2002; Tong et al., 2002). A large fraction of the diversity observed in SH3I domains is represented in the SH3 repertoire of Saccharomyces cerevisiae (Cesareni et al., 2002; Tong et al., 2002). Therefore, we selected the SH3 domain containing proteins of S.cerevisiae and their PxxP motif recognition specificities as a guide to analyze the five proteomes considered. We followed the PxxP motif nomenclature published recently (Cesareni et al., 2002) and classified the PxxP sequences into three major classes: 1K, 2K and 1@. All other atypical PxxP motifs are grouped under Class X. The various subclasses have been defined based on highly probable consensus motifs which have also been experimentally verified for binding to different SH3 domains (Kay et al., 2000; Mayer, 2001; Cesareni et al., 2002; Tong et al., 2002). We used the consensus sequences for various classes and subclasses as queries and filtered out the hits for each proteome. Using this motif-based proteome scanning approach, we identified all PxxP motifs in the annotated and predicted proteins (Figure 2a). The number of PXXP motifs in each proteome was plotted as a percentage of total number of proteins (Figure 2a). The P.falciparum proteome revealed the lowest fraction of PxxP motifs amongst the proteomes (25.9%). The M.tuberculosis proteome revealed a relatively high fraction of proteins contained the PxxP motif (58.4%). Surprisingly, the S.pombe proteome contained the maximum fraction of proteins with PxxP motifs (70.2%). Amongst the mammalian species, both human and mouse proteomes had a comparable fraction of PxxP motifs (65.2 and 61.8%, respectively). These data reveal the abundance of PXXP motifs in proteins from evolutionary distant species.
|
|
The AT-rich genome of P.falciparum is constrained in harboring prolines as only two of the 12 nucleotides that encode proline are either A or T (the rest are G or C). Based on the P.falciparum genome ATGC content, we expect 1% occurrence of prolines (observed occurrence 1.9%), whereas in the GC-rich genome of M.tuberculosis the expected occurrence of prolines is 12%.
Omnipresence of PxxP-containing proteins
We next classified the PxxP-containing proteins from each proteome into categories based on the likely cellular localization of each protein (Table II) and by gene ontology (Table III). The data indicate that PxxP motifs are prevalent in cytoplasmic, nuclear and surface proteins. We found an abundance of PxxP motifs in various gene ontology groups of proteins including enzymes, cytoskeletal proteins, nucleic acid-binding proteins, transport proteins, splicing factors, metal-binding proteins and ribosomal proteins (Table III). This wide occurrence across diverse functional classes in the five proteomes indicates evolutionary conservation of protein–protein networks centered on PxxP sequences. The prevalence of PxxP motifs in biological systems is consistent with the ubiquitous nature of SH3 domains and together the SH3-PxxP interactions are central in a diverse array of cellular communication processes (Sudol, 1998; Kay et al., 2000; Mayer, 2001).
|
|
Functional significance of PxxP classes and subclasses
The number of major classes of PxxP motifs in each organism was plotted as a percentage of total number of PxxP motifs in that respective organism (Figure 2b and Table IV). Analysis of these data reveals that the majority of the motifs can be classified either as a Class I (1K) or a Class II (2K) motif (Mayer, 2001; Cesareni et al., 2002). The Class I PxxP motifs are defined as sequences that fit the amino acid stretch +xxPxxP (where + is a basic residue and x is any amino acid). Class II PxxP motifs are sequences which match PxxPx+ and 1@ represents sequences that fit the motif Px@xxPxxP (where @ is an aromatic).
|
All five organisms displayed essentially equivalent distributions of Class I and II motifs (Figure 2b), which occur in a statistically relevant and non-random fashion (Figure 1a–e). The mouse and human proteomes are more abundant in PxxP motifs than the proteomes of the prokaryote Mycobacterium or of the simpler eukaryotes P.falciparum and S.pombe. The relative abundance of 1@ motifs was comparatively less across all five proteomes. Further, a substantial fraction of PxxP motifs in each proteome varied from the classical definition of Class I and II motifs. These sequences may represent motifs that have evolved to suit particular biological requirements by providing additional specificity in their interactions with SH3 domains.
Subtle differences in binding affinity and specificity in PxxP–SH3 interactions have been attributed to residues outside of the PxxP motifs (Cesareni et al., 2002). Several studies have used phage display libraries to identify the specific ligand of different SH3 domains (Rickles et al., 1994, 1995; Sparks et al., 1996). The optimal ligand preference for each SH3 domain varies around the PxxP core (Table IV). We reasoned that scanning the five proteomes with greater stringency would yield information on the appropriate SH3 receptors. The subclass delineation was based on definition of PxxP-containing sequences (Cesareni et al., 2002). Analysis of the data (Table IV) clearly indicates a preference for a subclass of motifs within each major PxxP class, suggesting reliance on protein–protein interactions based around some SH3 domain- containing proteins. We also plotted the number of different 1K ligands in the five proteomes considered as a percentage of the total number of proteins containing PxxP motifs (Figure 3a). Amongst the Class I motifs (Table IV), the PxxP ligands with preference for SH3 domain-containing proteins Rvs167, Pex13, Hck and SHO1 are highly represented. This distribution is likely to be of biological significance. The yeast Rvs167 protein plays critical roles in actin cytoskeleton organization and endocytosis (Breton et al., 2001). It interacts with myosin and participates in vesicle traffic and maintenance of cell integrity (Breton et al., 2001). The yeast Pex13 encodes for an SH3 domain-containing peroxisomal membrane protein required for the import of proteins into peroxisomes (Winkler et al., 2002). Hck is a prototypical member of the tyrosine kinase family and the yeast SHO1 protein is a membrane-bound osmosensor that is involved in stress activation (Winkler et al., 2002). Together, these data suggest that a substantial number of proteins in each of the five proteomes may use SH3–PxxP interactions to drive essential cellular processes of signal transduction, cytoskeleton organization, protein transport and osmoregulation.
|
The distribution of Class II (2K) motifs and its subclasses are shown in Figure 3B and Table IV. The occurrences of PxxP motif specific for the SH3 domain-containing protein Ysc84 is very high compared with other motifs. Ysc84 is a yeast protein that is expressed abundantly during sporulation of the S.cereviseae, but its biological significance has not yet been elucidated (Rocco et al., 1993
). Within Class II motifs, the subclass motif for interaction with the C-terminal domain of Grb2 has the highest occurrence (Figure 3c). The frequencies for these motifs have been plotted separately. Grb2 is an adapter protein that interacts with multiple proteins and is an essential constituent of numerous signaling pathways (Tari and Lopez-Berestein, 2001). Grb2 interacts with RAS in the signaling pathway leading to DNA synthesis. It is also involved in translocating guanine nucleotide exchange factors (Tari and Lopez-Berestein, 2001). Therefore, interaction with Grb2 via the PxxP motifs is likely to be central in numerous signaling cascades.
The distribution of 1@ subclasses is shown in Figure 3d. The highest occurrences are observed for spectrin-binding motifs followed by those for myosin and the tyrosine kinase protein Abl. The tyrosine kinase Abl participates in several processes including in the apoptotic response of cells to DNA damage (Yoshida et al., 2002) and is specific for atypical PxxP motifs. The SH3 domain of spectrin is implicated in a variety of myosin- and spectrin-based cytoskeleton organization (Ziemnicka-Kotula et al., 1998; Geli et al., 2000).
Conformations of both proline (backbone angle 
–65°) and the flanking residues are limited because of the bulk of the N-substituent in proline (MacArthur and Thornton, 1991). Therefore, polyproline sequences tend to adopt the PPII helix conformation which is generally solvent exposed and amphipathic in globular proteins (Stapley and Creamer, 1999). We evaluated 4531 entries (which contained PxxP motifs) from the protein structure database (PDB) and found that 1115 entries have PxxP motifs which are 50% solvent exposed, 370 entries have PxxP motifs which are 70% solvent exposed and 36 entries have PxxP motifs which are 90% solvent exposed (the programs MODELLER and PROCHECK were used to detect residue accessibility of PxxP motifs). These solvent-exposed PxxP motifs have the potential to interact with SH3 modules (some examples of accessible Class I and II motifs are shown in Figure 4). Indeed, our recent crystal structure analysis of Pfg27 from P.falciparum revealed a striking distribution of exposed PxxP motifs in the protein, which readily interact with SH3 modules (Sharma et al., 2003).
|
Drug targets and development of peptidomimics
The development of peptidomimics that specifically inhibit interaction of PxxP motifs with their cognate receptors is feasible given the unique structure of the proline residue (Nguyen et al., 1998; Oneyama et al., 2002, 2003). Therefore, we surveyed the total number of PxxP motifs that are unique to either P.falciparum or M.tuberculosis (Table V) with the aim of identifying sequences that could serve as leads for the development of peptide-based inhibitors. Indeed, there are a large number of unique PxxP motifs in both of these organisms. With the aim of identifying lead sequences for the design of drugs against malaria and tuberculosis, we found PxxP motifs that are common to both P.falciparum and M.tuberculosis but are absent from human proteins. The following sequences are not present in humans but are common to M.tuberculosis and P.falciparum: PAPAAPSS (found in PPE family protein and PfEMP), IGPNCPGI (found in succinyl-CoA synthetases) and SSPNTPGL (found in dihydroorotate dehydrogenases). Experimental verification of the biological necessity of these motifs may pave the way for their inclusion in drug development efforts against malaria and tuberculosis.
|
We also catalogued the unique PxxP motifs found in the proteomes of P.falciparum and M.tuberculosis. Intriguingly, several motifs in both organisms occur in multiple proteins. This offers an opportunity to design peptidomimics that may target either families of proteins or different proteins simultaneously. Inhibition of the potential interaction of these motifs with cognate receptors may disrupt multiple protein–protein interactions and therefore subvert a wide array of signaling networks. The peptides LYIPIYPYMH, KKKPKSPVDL, KYGPKAPTSW, IYAPRAPKYK and IYVPGSPKYK are unique to the P.falciparum proteome. These sequences are present in the PfEMP family of antigens (Table VI). PfEMPs are crucially responsible for the cytoadherence of P.falciparum and the resultant mortality in malaria (Smith et al., 1995
; Su et al., 1995). These unique PxxP sequences are present in the highly conserved cytoplasmic domains of PfEMPs. The latter interact with the host erythrocyte cytoskeletal proteins spectrin (which contains the SH3 module) and actin, thereby assisting anchorage of PfEMP in the erythrocyte skeleton (Oh et al., 2000). It remains to be verified experimentally whether these interactions are mediated by the conserved PxxP motifs that we have identified. Disruption of PfEMP interactions with proteins of the host cytoskeleton in infected erythrocytes may offer a unique strategy to abrogate the cytoadherence associated with malignant malaria.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The organization of each cell depends on complex networks that involve multitudes of protein–protein interactions. Signaling between proteins underpins essential cellular processes including metabolism, cell movement, differentiation, development, cytoskeletal arrangement and apoptosis. In many cases, this molecular cross-talk is organized around PxxP motifs and protein modules such as SH3 domains, which recognize them. Typically, SH3–PxxP complexes are formed as low-affinity but high-specificity interactions (Mayer, 2001
). Such weak interactions are ideally suited for transient signaling between proteins in the cellular milieu. Most proteins have additional ancillary functions aside from their primary biochemical activity. The relevance of newly discovered proteins in understanding human diseases such as malaria and tuberculosis will be clear once the biological attributes of most proteins have been highlighted.
Close to 65% of the P.falciparum proteome is unannotated and >3000 proteins remain hypothetical (Florens et al., 2002). Here, we have used a simple PxxP motif search to predict putative protein–protein interactions in P.falciparum and other organisms. We had previously highlighted the interaction of P.falciparum gametocyte protein Pfg27 with different SH3 domains and discovered a striking structural display of PxxP motifs in the Pfg27 underbody (Sharma et al., 2003). Our analyses suggest that the distribution of PxxP motifs in P.falciparum proteins is therefore likely to be of significance in understanding parasite biology.
Protein–protein interaction networks have been studied in S.cerevisiae with respect to SH3 modules. Tong et al. showed that using a set of 18 SH3 modules in 2-hybrid analysis there are 233 interactions involving 145 PxxP-containing proteins (Tong et al., 2002). Using the technique of phage display, they observed 394 SH3–PxxP interactions from 206 proteins. These data imply that a vast number of protein–protein interactions are mediated using the SH3–PxxP cross-talk.
Our comparative distribution analyses of PxxP motifs across various proteomes, which include the proteomes of infectious organisms such as M.tuberculosis and P.falciparum, can underpin ongoing efforts to dissect and delineate the plethora of signaling pathways in these organisms. It also provides a basis for further biological characterization of PxxP motifs that can be utilized for the design of peptidomimics. Specific sequence motifs offer a platform for the design of peptidomimics that specifically target ‘multiple’ pathways in cells. This novel strategy may provide a new focus for the development of anti-tuberculosis and anti-malarial agents.
|
Acknowledgements |
|---|
We thank Chetan Chitnis for discussions and suggestions. A.S. is funded by a Wellcome Trust Senior Reseach Fellowship.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Breton,A.M., Schaeffer,J. and Aigle,M. (2001) Yeast, 18, 1053–1068.[CrossRef][Web of Science][Medline]
Cesareni,G., Panni,S., Nardelli,G. and Castagnoli,L. (2002) FEBS Lett., 513, 38–44.[CrossRef][Web of Science][Medline]
Cohen,G.B., Ren,R. and Baltimore,D. (1995) Cell, 80, 237–248.[CrossRef][Web of Science][Medline]
Dalgarno,D.C., Botfield,M.C. and Rickles,R.J. (1997) Biopolymers, 43, 383–400.[CrossRef][Web of Science][Medline]
Feese,M.D., Ingason,B.P., Goranson-Siekierke,J., Holmes,R.K. and Hol,W.G. (2001) J. Biol. Chem., 276, 5959–5966.
Feng,S., Chen,J.K., Yu,H., Simon,J.A. and Schreiber,S.L. (1994) Science, 266, 1241–1247.
Florens,L. et al. (2002) Nature, 419, 520–526.[CrossRef][Medline]
Geli,M.I., Lombardi,R., Schmelzl,B. and Riezman,H. (2000) EMBO J., 19, 4281–4291.[CrossRef][Web of Science][Medline]
Kami,K., Takeya,R., Sumimoto,H. and Kohda,D. (2002) EMBO J., 21, 4268–4276.[CrossRef][Web of Science][Medline]
Kang,H., Freund,C., Duke-Cohan,J.S., Musacchio,A., Wagner, G. and Rudd,C.E. (2000) EMBO J., 19, 2889–2899.[CrossRef][Web of Science][Medline]
Kay,B.K., Williamson,M.P. and Sudol,M. (2000) FASEB J., 14, 231–241.
Laskowski,R.A., Rullmannn,J.A., MacArthur,M.W., Kaptein,R. and Thornton,J.M. (1996) J. Biomol. NMR, 8, 477–486.[Web of Science][Medline]
Lim,W.A., Richards,F.M. and Fox,R.O. (1994) Nature, 372, 375–379.[CrossRef][Medline]
MacArthur,M.W. and Thornton,J.M. (1991) J. Mol. Biol., 218, 397–412.[CrossRef][Web of Science][Medline]
Mayer,B.J. (2001) J. Cell Sci., 114, 1253–1263.[Abstract]
Nguyen,J.T., Turck,C.W., Cohen,F.E., Zuckermann,R.N. and Lim,W.A. (1998) Science, 282, 2088–2092.
Oh,S.S., Voigt,S., Fisher,D., Yi,S.J., LeRoy,P.J., Derick, L.H., Liu,S. and Chishti,A.H. (2000) Mol. Biochem. Parasitol., 108, 237–247.[CrossRef][Web of Science][Medline]
Oneyama,C., Nakano,H. and Sharma,S.V. (2002) Oncogene, 21, 2037–2050.[CrossRef][Web of Science][Medline]
Oneyama,C., Agatsuma,T., Kanda,Y., Nakano,H., Sharma,S.V., Nakano,S., Narazaki,F. and Tatsuta,K. (2003) Chem. Biol., 10, 443–451.[CrossRef][Web of Science][Medline]
Ponting,C.P., Aravind,L., Schultz,J., Bork,P. and Koonin, E.V. (1999) J. Mol. Biol., 289, 729–745.[CrossRef][Web of Science][Medline]
Rickles,R.J., Botfield,M.C., Weng,Z., Taylor,J.A., Green, O.M., Brugge,J.S. and Zoller,M.J. (1994) EMBO J., 13, 5598–5604.[Web of Science][Medline]
Rickles,R.J., Botfield,M.C., Zhou,X.M., Henry,P.A., Brugge, J.S. and Zoller,M.J. (1995) Proc. Natl Acad. Sci. USA, 92, 10909–10913.
Rocco,V., Daly,M.J., Matre,V., Lichten,M. and Nicolas,A. (1993) Yeast, 9, 1111–1120.[CrossRef][Web of Science][Medline]
Sali,A., Potterton,L., Yuan,F., van Vlijmen,H. and Karplus, M. (1995) Proteins, 23, 318–326.[CrossRef][Web of Science][Medline]
Sharma,A., Sharma,I., Kogkasuriyachai,D. and Kumar,N. (2003) Nat. Struct. Biol., 10, 197–203.[CrossRef][Web of Science][Medline]
Smith,J.D., Chitnis,C.E., Craig,A.G., Roberts,D.J., Hudson-Taylor,D.E., Peterson,D.S., Pinches,R., Newbold,C.I. and Miller,L.H. (1995) Cell, 82, 101–110.[CrossRef][Web of Science][Medline]
Sparks,A.B., Rider,J.E., Hoffman,N.G., Fowlkes,D.M., Quillam,L.A. and Kay,B.K. (1996) Proc. Natl Acad. Sci. USA, 93, 1540–1544.
Stapley,B.J. and Creamer,T.P. (1999) Protein Sci., 8, 587–595.[Web of Science][Medline]
Su,X.Z., Heatwole,V.M., Wertheimer,S.P., Guinet,F., Herrfeldt,J.A., Peterson,D.S., Ravetch,J.A. and Wellems,T.E. (1995) Cell, 82, 89–100.[CrossRef][Web of Science][Medline]
Sudol,M. (1998) Oncogene, 17, 1469–1474.[CrossRef][Web of Science][Medline]
Tari,A.M. and Lopez-Berestein,G. (2001) Semin. Oncol., 28, 142–147.[CrossRef][Web of Science][Medline]
Tong,A.H. et al. (2002) Science, 295, 321–324.
Winkler,A., Arkind,C., Mattison,C.P., Burkholder,A., Knoche, K. and Ota,I. (2002) Eukaryot. Cell, 1, 163–173.
Yoshida,K., Komatsu,K., Wang,H.G. and Kufe,D. (2002) Mol. Cell. Biol., 22, 3292–3300.
Ziemnicka-Kotula,D., Xu,J., Gu,H., Potempska,A., Kim,K.S., Jenkins,E.C., Trenkner,E. and Kotula,L. (1998) J. Biol. Chem., 273, 13681–13692.
Received October 22, 2003; revised January 6, 2004; accepted February 2, 2004 Edited by Alan Fersht
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  V. M. Navadgi, B. R. Chandra, P. C. Mishra, and A. Sharma The Two Plasmodium falciparum Nucleosome Assembly Proteins Play Distinct Roles in Histone Transport and Chromatin Assembly J. Biol. Chem., June 23, 2006; 281(25): 16978 - 16984. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Jenwitheesuk and R. Samudrala Identification of Potential Multitarget Antimalarial Drugs JAMA, September 28, 2005; 294(12): 1490 - 1491. [Full Text] [PDF]
|
|
|
|
|
|
A. Rao, S. Chopra, G. Ram, A. Gupta, and A. Ranganathan Application of the "Codon-shuffling" Method: SYNTHESIS AND SELECTION OF DE NOVO PROTEINS AS ANTIBACTERIALS J. Biol. Chem., June 24, 2005; 280(25): 23605 - 23614. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||