PhosSNP for systematic analysis of genetic polymorphisms that influence protein phosphorylation

Abstract

We are entering the era of personalized genomics as breakthroughs in sequencing technology have made it possible to sequence or genotype an individual person in an efficient and accurate manner. Preliminary results from HapMap and other similar projects have revealed the existence of tremendous genetic variations among world populations and among individuals. It is important to delineate the functional implication of such variations, i.e. whether they affect the stability and biochemical properties of proteins. It is also generally believed that the genetic variation is the main cause for different susceptibility to certain diseases or different response to therapeutic treatments. Understanding genetic variation in the context of human diseases thus holds the promise for "personalized medicine." In this work, we carried out a genome-wide analysis of single nucleotide polymorphisms (SNPs) that could potentially influence protein phosphorylation characteristics in human. Here, we defined a phosphorylation-related SNP (phosSNP) as a non-synonymous SNP (nsSNP) that affects the protein phosphorylation status. Using an in-house developed kinase-specific phosphorylation site predictor (GPS 2.0), we computationally detected that approximately 70% of the reported nsSNPs are potential phosSNPs. More interestingly, approximately 74.6% of these potential phosSNPs might also induce changes in protein kinase types in adjacent phosphorylation sites rather than creating or removing phosphorylation sites directly. Taken together, we proposed that a large proportion of the nsSNPs might affect protein phosphorylation characteristics and play important roles in rewiring biological pathways. Finally, all phosSNPs were integrated into the PhosSNP 1.0 database, which was implemented in JAVA 1.5 (J2SE 5.0). The PhosSNP 1.0 database is freely available for academic researchers.

Fig. 1.

Computational procedure of phosSNPs detection. In addition to ab initio prediction of kinase-specific phosphorylation sites using GPS 2.0 (22), we also detected potential phosSNPs by exact string matching with 23,978 experimentally identified human phosphorylation sites (Exp. Data) from a recent analysis (26). In total, there were 64,035 potential phosSNPs identified in 17,614 sequences.

Fig. 2.

Five types of phosSNPs with typical examples. A, Type I PhosSNP. The K897T nsSNP of KCNH2/ERG1 creates a new AKT-specific phosphorylation site (Type I (+)), whereas the S421F nsSNP of HTR2A removes the phosphorylation site at Ser-421 (Type I (−)). B, Type II PhosSNP. The G4561D nsSNP of AHNAK might render its nearby Thr-4564 residue as a potential phosphorylation site (Type II (+)), whereas the P830L nsSNP of BRCA1 might prohibit Ser-832 phosphorylation (Type II (−)). C, Type III PhosSNP. The P47S nsSNP of p53 might induce changes of PK types for multiple adjacent phosphorylation sites. D, Type IV PhosSNP. The S412Y nsSNP of F2R might induce a change of its upstream serine/threonine PKs into tyrosine PKs. E, Type V PhosSNP. The E600Stop nonsense nsSNP might remove its following phosphorylation sites. TM, transmembrane.

Fig. 3.

Different PKs exhibit different substrate preferences on Ser or Thr residue. From the Phospho.ELM 8.2 database (19), we collected 377, 347, 36, and 38 experimentally verified phosphorylation sites for four well studied PK groups, including AGC/PKA, CMGC/CDK, AGC/PDK1, and AGC/DMPK/ROCK, respectively. For AGC/PKA (A) and CMGC/CDK (B), the Ser is the preferred residue, whereas the Thr residue is more preferred for AGC/PDK1 (C) and AGC/DMPK/ROCK (D).

Fig. 4.

One nsSNP could be mapped on different genes. A, the human CFH and CFHR1 were located on Chromosome (Chr.) 1. B, by sequence comparison, two proteins share 68% identities in their C terminus. C, one nsSNP (rs425757) in complement factor H and complement factor H-related 1 is Y1058H and H157Y, respectively.