Article Amino-acid type identification in 15 N-HSQC spectra by combinatorial selective 15 N-labelling Peter S. C. Wu, Kiyoshi Ozawa, Slobodan Jergic, Xun-Cheng Su, Nicholas E. Dixon & Gottfried Otting* Research School of Chemistry, Australian National University, Canberra, ACT, 0200, Australia Received 9 September 2005; Accepted 2 November 2005 Key words: 15 N-HSQC, cell-free protein synthesis, combinatorial 15 N-labelling, DNA polymerase III, resonance assignment, subunit s Abstract The efficiency of cell-free protein synthesis combined with combinatorial selective 15 N-labelling provides a method for the rapid assignment of 15 N-HSQC cross-peaks to the 19 different non-proline amino-acid types from five 15 N-HSQC spectra. This strategy was explored with two different constructs of the C-terminal domain V of the s subunit of the Escherichia coli DNA polymerase III holoenzyme, s C 16 and s C 14. Since each of the five 15 N-HSQC spectra contained only about one third of the cross-peaks present in uniformly labelled samples, spectral overlap was much reduced. All 15 N-HSQC cross-peaks of the backbone amides could be assigned to the correct amino-acid type. Availability of the residue-type information greatly assisted the evaluation of the changes in chemical shifts observed for corresponding residues in s C 16 vs. those in s C 14, and the analysis of the structure and mobility of the C-terminal residues present in s C 16 but not in s C 14. Abbreviations: s C 14 – residues 499–625 of the s subunit of E. coli DNA polymerase III with an additional N-terminal methionine; s C 16 – same as s C 14, but including the C-terminal 18 residues 626–643 of s. Introduction 15 N-HSQC spectra provide well-resolved finger- print information and are the cornerstone of backbone resonance assignments of 15 N-labelled proteins. With the advent of high-yield cell-free protein synthesis systems, the preparation and NMR spectroscopic analysis of selectively 15 N-la- belled proteins has become both fast and inexpen- sive (Kigawa et al., 1995; Ozawa et al., 2005a). While maximal information could be obtained from 19 different samples, where each of the 19 non-proline residues is selectively 15 N-labelled (Yamazaki et al., 1991; Ozawa et al., 2004), the same information can be retrieved with much less effort by the use of combinatorial 15 N-labelling, where several amino acids are simultaneously 15 N- labelled in a limited number of samples (Shortle 1994; Parker et al., 2004). Combinatorial isotope labelling results in ambiguities only if chemical shift degeneracies lead to perfect superposition of the cross-peaks of two or more amino-acid residues. Here we present a different combinatorial 15 N-labelling scheme which delivers complete residue-type identifications for the 15 N-HSQC cross-peaks from all 19 non-proline residues and where each 15 N-HSQC spectrum produces only a third of the peaks generated by uniformly 15 N-labelled protein. It was used for the structural *To whom correpondence should be addressed. E-mail: gottfried.otting@anu.edu.au Journal of Biomolecular NMR (2006) 34: 13–21 Ó Springer 2006 DOI 10.1007/s10858-005-5021-9