Methods to Identify the NMR Resonances of the ¹³C-Dimethyl N-terminal Amine on Reductively Methylated Proteins

Nuclear magnetic resonance (NMR) spectroscopy is a valuable structure elucidation tool for proteins¹. NMR spectroscopy can be used to determine the solution structure of a protein in its native state. To overcome the low natural abundance of stable magnetic isotopes, it is necessary to incorporate ¹³C and ¹⁵N into the protein of interest. The most common method employed is recombinant expression in a bacterial host^2-3. However, two disadvantages of bacterial host over-expression are it cannot produce post-translational modifications and does not work for all proteins^3-4. When bacterial expression is not a viable route for protein production, over-expression in nonbacterial hosts can be used, but isotopic labeling is difficult and expensive⁵. Alternative expression methods for incorporating ¹³C and ¹⁵N isotopes into proteins for NMR analysis include sparse labeling techniques using metabolic precursors for methyl labeling⁶ and single ¹³C, ¹⁵N amino acids^7-9. A chemical approach to sparse labeling used herein is the well-established reductive ¹³C-methylation reaction (Figure 1), where the primary amino groups on a protein – the N-terminal α-amine and the lysine, side chain ε-amines – are methylated. Once monomethylamines are formed, the amine readily undergoes methylation again, due to the higher pKa value, to form dimethylamines.

Reductive methylation was first introduced as a method to chemically modify proteins by Means and Feeney¹⁰. The advantages of this reaction are its broad applicability and mild reactions conditions at buffered, physiological pH and low temperatures^11,12. In the presence of formaldehyde and a reducing agent, such as dimethylamine borane complex (DMAB), the lysine ε-amino groups and the N-terminal α-amino group are selectively methylated to produce dimethylated amines. Although formaldehyde is known to cross-link proteins through the formation of methylene bridges, this process is blocked by the reducing agent^13,14.

Reductive methylation has been successfully used to study proteins with both NMR and x-ray crystallography. Reductive methylation is used to facilitate the crystallization of otherwise intractable proteins¹⁵. Hen egg white lysozyme was the first protein crystallized in its dimethylated form. The root-mean square difference between the heavy atoms in the methylated and unmethylated lysozyme structures is 0.40 Ź⁶. This comparison demonstrates that the protein structure can be maintained after reductive methylation, making the reaction a viable, labeling tool for structure elucidation.

By using ¹³C-labeled formaldehyde in the reductive methylation reaction, ¹³C-dimethylated amines are produced. The ¹³C-dimethylamines are NMR-detectable probes that have been widely used to study protein dynamics, structure, and function. NMR and reductive ¹³C-methylation have been used to study protein-ligand and protein-protein interactions for the β₂ adrenergic receptor¹⁷, ribonuclease A¹⁸, lysozyme^19, fd gene 5 protein²⁰, and cytochrome c²¹. Similarly, structural and functional properties have been studied of reductively ¹³C-methylated ribonuclease A²², lysozyme²³, fd gene 5 protein²⁴, Clostridium pasteurianum ferredoxin²⁵, Fc fragment of IgG²⁶, apolipoprotein A-I²⁷, and MIP-1α²⁸ The dynamics of the ¹³C-dimethylamino groups have been studied on concanavalin A^29-30 and calmodulin³¹.

Even though reductive ¹³C-methylation has been used widely to study proteins with NMR, the labeling method has always been limited by the difficulties of assigning the NMR resonances²⁶. Most assignment strategies for reductively ¹³C-methylated proteins have relied on small numbers of sites,^20,28 known structural properties^{11,19,23,26,31,32}. or extensive genetic modifications³³. None of these studies successfully assigned all the ¹³C-dimethylamine peaks except for the calmodulin study, where the peaks were assigned by site-directed mutagenesis of each lysine³³. In the study of MIP-1α dimer formation, the use of mass spectrometry (MS) to aid in the NMR assignment of reductively ¹³C-methylated amines was first reported²⁸. Matrix assisted laser desorption ionization time-of-flight (MALDI-TOF) MS was used to identify the lysine at the interface of the dimer. The partially methylated lysines of human MIP-1α tryptic peptides were identified with MS and correlated with the appearance of mono- and dimethylamine signals of the intact protein observed in 2D ¹H-¹³C HSQC NMR spectra²⁸. Our group expanded on the use of MS and presented an assignment method that requires no prior knowledge of the protein's structure or properties other than the amino acid sequence³⁴. This method is applicable to most proteins. One exception is when a protein has an N-terminal lysine because the MS isotopic profile of the N-terminal lysine α- and ε-¹³C-dimethylamines cannot be independently measured.

Here we present two methods, one chemical and one enzymatic, to identify the N-terminal α-dimethylamine and the side chain ε-dimethylamine sites of an N-terminal lysine residue. The first method was inspired by Córdova et al. who used pH to control the selectivity of protein acetylation³⁵. The reaction favors the N-terminal α-amine at low pH and the side chain ε-amines at high pH, allowing the protein α- and ε-amino groups to be distinguished, in their studies, with capillary electrophoresis³⁵. We demonstrate how high and low pH is used to alter the labeling of protein amino groups using the reductive ¹³C-methylation reaction and to allow application of the MS-assisted assignment strategy. The second method to assign the N-terminal α- and ε-¹³C-dimethylamines takes advantage of the selective removal of the N-terminal residue(s) using recombinant Aeromonas proteolytica aminopeptidase.