PDZ domain binding selectivity is optimized across the mouse proteome

Science

Volumn 317, Issue 5836, 2007, Pages 365-369

  Stiffler, Michael A ^a   Chen, Jiunn R ^b   Grantcharova, Viara P ^a,c   Lei, Ying ^a   Fuchs, Daniel ^a   Allen, John E ^a   Zaslavskaia, Lioudmila A ^a,d   MacBeath, Gavin ^a  

^a HARVARD UNIVERSITY   (United States)

^b HARVARD UNIVERSITY   (United States)

^c MERRIMACK PHARMACEUTICALS   (United States)

^d Lifecodes Corporation   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

PDZ PROTEIN; PROTEOME; PEPTIDE;

CHEMICAL BINDING; GENOME; MODELING; PEPTIDE; PROTEIN; PROTEOMICS; RODENT;

ACCURACY; ARTICLE; BINDING AFFINITY; FLUORESCENCE POLARIZATION; MOUSE; NONHUMAN; PRIORITY JOURNAL; PROTEIN ANALYSIS; PROTEIN BINDING; PROTEIN DOMAIN; PROTEIN FUNCTION; PROTEIN MICROARRAY; PROTEIN PROTEIN INTERACTION; ALGORITHM; AMINO ACID SEQUENCE; ANIMAL; BIOLOGY; CHEMISTRY; COMPUTER SIMULATION; METABOLISM; PROTEIN TERTIARY STRUCTURE;

ALGORITHMS; AMINO ACID SEQUENCE; ANIMALS; COMPUTATIONAL BIOLOGY; COMPUTER SIMULATION; FLUORESCENCE POLARIZATION; MICE; PEPTIDES; PROTEIN ARRAY ANALYSIS; PROTEIN BINDING; PROTEIN STRUCTURE, TERTIARY; PROTEOME;

EID: 34547127612     PISSN: 00368075     EISSN: 10959203     Source Type: Journal    
DOI: 10.1126/science.1144592     Document Type: Article

References (33)

1. T. Pawson, Nature 373, 573 (1995).
2. T. Pawson, P. Nash, Science 300, 445 (2003).
3. E. Kim, M. Niethammer, A. Rothschild, Y. N. Jan, M. Sheng, Nature 378, 85 (1995).
4. H. C. Kornau, L. T. Schenker, M. B. Kennedy, P. H. Seeburg, Science 269, 1737 (1995).
5. C. Nourry, S. G. N. Grant, J.-P. Borg, Sci. STKE 2003, re7 (2003).
6. B. Z. Harris, F. W. Lau, N. Fujii, R. K. Guy, W. A. Lim, Biochemistry 42, 2797 (2003).
7. Z. Songyang et al., Science 275, 73 (1997).
8. I. Bezprozvanny, A. Maximov, FEBS Lett. 509, 457 (2001).
9. E. Song et al., Mot. Cell. Proteomics 5, 1368 (2006).
10. I. Letunic et al., Nucleic Acids Res. 34, D257 (2006).
11. J. Schultz, F. Milpetz, P. Bork, C. P. Ponting, Proc. Natl. Acad. Sci. U.S.A. 95, 5857 (1998).
12. The "genomic" mode of the Simple Modular Architecture Research Tool (SMART) database (10, 11) currently lists 240 PDZ domains identified from the mouse genome sequence. We obtained sequence-verified clones for 203 of them. In addition, we cloned 18 PDZ domains that are listed only in the "normal" mode of the SMART database. Each domain was purified in a single step from a large-scale bacterial culture, its purity was assessed by SDS-polyacrylamide gel electrophoresis (fig. S1), and its aggregation state was assessed by analytical gel filtration (table S1). Soluble protein of the correct molecular weight was obtained for 157 PDZ domains. Of these, 151 were monomeric, 4 were dimeric, and 2 were a mixture of monomers and dimers (table S1).
13. G. Fuh et al., J. Biol. Chem. 275, 21486 (2000).
14. Y. Zhang et al., J. Biol. Chem. 281, 22299 (2006).
15. We derived 57 of the training-set peptides from proteins that had previously been shown to interact with PDZ domains. To allow for the possibility of discovering sequences that fall outside the established view of peptide-binding selectivity, we derived the other peptides from different members of 13 families of membrane proteins, regardless of whether their C termini feature canonical PDZ domain binding motifs (table S2).
16. R. B. Jones, A. Gordus, J. A. Krall, G. MacBeath, Nature 439, 168 (2006).
17. M. A. Stiffler, V. P. Grantcharova, M. Sevecka, G. MacBeath, J. Am. Chem. Soc. 128, 5913 (2006).
18. d was <100 μM.
19. d for all positive interactions. A comparison of the resulting interaction matrix with the microarray data showed a false-negative rate of 6.6%.
20. J. C. Obenauer, L. C. Cantley, M. B. Yaffe, Nucleic Acids Res. 31, 3635 (2003).
21. M. Sandberg, L. Eriksson, J. Jonsson, M. Sjostrom, S. Wold, J. Med. Chem. 41, 2481 (1998).
22. Materials and methods are available as supporting material on Science Online.
23. S. van Dongen, thesis, University of Utrecht, Netherlands (2000).
24. Full-length sequences of 31,302 unique mouse proteins (including splicing variants) were downloaded with BioMart from data set NCBIM36 (Mus musculus genes) of Ensembl 44. The C-terminal sequence of each entry was extracted using a Python script.
25. Interactions in table S5 are model predictions with m set to 20. On the basis of the results of our model validation efforts, we estimate these predictions to have a true-positive rate of 35%, a false-positive rate of 7%, and a TP/FP ratio of 0.83.
26. This single-point assay measures the difference between FP at 20 nM peptide, 20 μM PDZ domain and FP at 20 nM peptide, 0 μM PDZ domain. An analysis of 1710 FP titration curves shows that applying a threshold of 40 mP units to this single-point assay correctly identifies 91% of the positives and 96% of the negatives (fig. S3). Thus, instead of performing an additional 1170 titration curves, we used this single-point assay with a threshold of 40 mP units to evaluate interactions between the 48 test peptides and the 74 PDZ domains in the MDSM.
27. R. Jansen et al., Science 302, 449 (2003).
28. A. Tropsha, Annu. Rep. Comput. Chem. 2, 113 (2006).
29. G. Birrane, J. Chung, J. A. Ladias, J. Biol. Chem. 278, 1399 (2003).
30. F. Jaulin-Bastard et al., J. Biol. Chem. 277, 2869 (2002).
31. F. Jaulin-Bastard et al., J. Biol. Chem. 276, 15256 (2001).
32. A. Zarrinpar, S. H. Park, W. A. Lim, Nature 426, 676 (2003).
33. We thank A. Tropsha for valuable suggestions and the Faculty of Arts and Sciences Center for Systems Biology for support with instrumentation and automation. This work was supported by awards from the Smith Family Foundation, the Arnold and Mabel Beckman Foundation, and the W. M. Keck Foundation and by a grant from the NIH (1 RO1 GM072872-01). M.A.S. was supported in part by the NIH Molecular, Cellular, and Chemical Biology Training Grant (5 T32 GM07598-25), and J.R.C. was the recipient of a Corning CoStar fellowship.