SH2/PTB–ErbB interactions

Click below to download a Microsoft Excel file that provides equilibrium dissociation constants (Kd's) for the interaction of human SH2 and PTB domains with peptides representing experimentally-verified sites of tyrosine phosphorylation on the four human ErbB receptors (EGFR, ErbB2, ErbB3, ErbB4).

The data provided in version 01 are reported in:
Jones RB, Gordus A, Krall JA, MacBeath G (2005). A quantitative network for the ErbB receptors using protein microarrays. Nature 439: 168-174; advance online publication, 6 November 2005 (doi:10.1038/nature04177).

Sample microtiter plate (jpeg identifying peptides and concentrations)
Rhodamine image (149 Mb, 16-bit tiff file)
Cy5 image (149 Mb, 16-bit tiff file)

Click below to download a Microsoft Excel file that provides Kd's for the interaction of human SH2 and PTB domains with phosphopeptides representing every intracellular tyrosine residue on EGFR, FGFR1, and IGF1R.

These data are reported in:
Kaushansky A, Gordus A, Chang B, Rush J, & MacBeath G (2008). A quantitative study of the recruitment potential of all intracellular tyrosine residues on EGFR, FGFR1 and IGF1R. Mol. BioSyst. 4: 643-653.

PDZ–peptide interactions

Click below to download a Microsoft Excel file that provides equilibrium dissociation constants (Kd's) for the interaction of mouse PDZ domains with peptides representing the C-terminal tails of selected mouse proteins.

The data provided in version 01 are reported in:

Stiffler MA, Chen JR, Grantcharova VP, Lei Y, Fuchs D, Allen JE, Zaslavskaia LA, MacBeath G (2007). PDZ domain binding selectivity is optimized across the mouse proteome. Science 317: 364-369.

The full text of this paper can be downloaded from the publications section of this web site.

This paper also reports a multidomain selectivity model (MDSM) that predicts to which PDZ domains a peptide will bind, given its sequence. The model includes 74 mouse PDZ domains. In the model, a peptide is predicted to bind to PDZ domain i if:

where is a binding score, A is an indicator of peptide sequence ( if the amino acid at position p of the peptide is q and  otherwise), and  is a scoring threshold, specific to each domain. The model takes into account the last five residues of the PDZ ligand (p can have the values -4, -3, -2, -1, and 0). To determine if your peptide sequence is likely to bind to a given PDZ domain, calculate for your peptide sequence and that domain by adding up the five relevant values of (see Microsoft Excel file below). If , the peptide is predicted to bind to that domain. The Excel file provides three different values of for each domain (m=5, 10, 20). If m=5, the sensitivity (true positive rate) of the model is 48% and the specificity (1 - false positive rate) is 88%. If m=20, the sensitivity is 35% and the specificity is 93%. For further details about the model and the meaning of m, please see the paper.