An NMR approach to structural proteomics

Abstract

The influx of genomic sequence information has led to the concept of structural proteomics, the determination of protein structures on a genome-wide scale. Here we describe an approach to structural proteomics of small proteins using NMR spectroscopy. Over 500 small proteins from several organisms were cloned, expressed, purified, and evaluated by NMR. Although there was variability among proteomes, overall 20% of these proteins were found to be readily amenable to NMR structure determination. NMR sample preparation was centralized in one facility, and a distributive approach was used for NMR data collection and analysis. Twelve structures are reported here as part of this approach, which allowed us to infer putative functions for several conserved hypothetical proteins.

Figure 1

Histogram of the number of proteins cloned (blue), expressed (red), and soluble (gray) from each organism.

Figure 2

Distribution chart of the HSQC classifications for soluble, purifiable proteins. (a) M. thermoautotrophicum (M. Th.); (b) E. coli; (c) S. cerevisiae (S. ce.); (d) T. maritima (T. ma.); (e) Myx. Spectra are classified as good (blue), promising (red), mostly unfolded (gray), poor (purple), and those for which no HSQC could be obtained because of a loss of protein during the concentration procedure (green).

Figure 3

¹⁵N HSQC spectra and the backbone ribbon representations of the 12 structures presented in this paper. All HSQCs are plotted from 6.0–10.5 ppm in the ¹H dimension (x axis) and from 107 to 133 ppm in the ¹⁵N dimension (y axis). The number of residues for each protein is indicated on the HSQC spectrum. β-sheets are shown in cyan, and α-helices are shown in red. N-terminal residues 1–20 of yedF_ecoli and Myxv156r are unstructured and not shown. C-terminal residues 198–208 of Mth1692 are unstructured and not shown. All structure diagrams were created by using the MOLAUTO program within MOLSCRIPT (17).