The Role of Separations Methodology

Beside its modest size, insulin was also available to the Sanger group in relatively large amounts in homogenous form, a prerequisite for chemical sequencing studies. It is therefore not surprising that most of the studies that followed Sanger's pi­oneering effort were with proteins that were also relatively easy to purify and obtain in substantial amounts (see Schroeder, 1968, for a description of a representative listing of these early efforts). These tended to be extracellular moieties that were found in fluids or other conveniently obtained rich sources.

Nonetheless, the substantial improvements in separations technology that followed the Sanger work played a very sig­nificant role in making available more and more proteins for sequence analysis. Two of the most important of these were ion-exchange resins suitable for protein separations, such as substituted celluloses and gel filtration media.

These latter resins, consisting of cross-linked dextrans (Sephadex) or polymerized and cross-linked acrylamide (Bio-gel), basically fractionated protein mixtures on the basis of size, and gave a new dimension to purification schemes that was largely in­dependent of separations based on charge. Performed in tan­dem they were very effective. Of course there were many additional methods that were modified or introduced, such as gel electrophoresis, isoelectric focusing, reverse phase high­performance liquid chromatography (HPLC), and affinity chromatography among others, that substantially increased the arsenal of separations procedures.

When any or all of these were used in conjunction with the older bulk methods, such as alcohol or ammonium sulfate precipitation, sophisticated multistep protocols to purify even relatively rare proteins began to be fairly commonplace by the mid-1960s. It was not till recombinant technology became available that these often laborious protocols became largely passe and the need to se­quence proteins by chemical means (other than to obtained partial sequence data to construct oligonucleotide probes for cloning experiments), for the most part, disappeared.

The development of separations technology also played a very significant role in preparing peptides and fragments for sequencing. In early studies the fragments generated by en­zymatic cleavage were generally small and soluble (although there was invariably an insoluble core that often contained longer fragments, who's further fractionation always presented significant challenges).

In fact, until automated sequencing became routinely available, peptides much longer than 10 residues would require some sort of further reduction in size since manual Edman degradation rarely went much further than that. Thus it was common to use several types of en­zymes, with differing specificities, as a means to generate enough soluble peptides to get complete coverage of the se­quence being determined.

Since, after trypsin, the specificity of the proteases available was less precise and, therefore, less reliable this often proved to be a problem.

For the most part, mixtures of soluble peptides from enzymatic digests were amenable to fractionation by the same resins used for separ­ating amino acids (Dowex 50, its basic counterpart, Dowex 1, and Amberlite IRC 50, which was a carboxylate resin and hence a weak anionic exchanger). Gel filtration resins, with low-molecular weight cutoffs, and in later years, HPLC, were also very useful.

The wide-scale introduction of the automated protein sequencer in the early 1970s had a material impact on strategy and hence the technology employed. At first, it was most valuable in determining the N-terminal sequence of whole proteins but that soon lead to the realization that by creating bigger fragments these could be analyzed as if they were proteins in their own right (and unlike the whole protein, which was often blocked by an acetyl group if it was a cytoplasmic protein from a eukaryotic cell, they had an ac­cessible N-terminus).

Employing cleavage strategies that produced fewer cuts and thus larger pieces had already ap­peared in more ambitious studies of significantly larger pro­teins. Many of the more successful methods for creating bigger fragments depended on chemical reagents (as opposed to enzymes).

The most important and widely used of these was cyanogen bromide (CNBr) (Gross and Witkop, 1961), a highly reactive and toxic substance that in strongly acidic so­lutions, for example 70% formic acid, reacts only with the thioether of methionine side chains to produce a sulfonium salt that then results in cyclization leading to cleavage of the adjacent peptide bond (and leaves homoserine lactone in the place of the methionine in the C-terminal position of each fragment generated, except the one derived from the C-terminus of the protein).

Generally the large fragments produced were due to the relatively rare occurrence of methionine. Various other methods, often depending on the introduction of one or more modifications first, were also developed (Blackburn, 1970).

The creation of big fragments that were easy to analyze by automated sequencing and were effective in subdividing large sequences into more manageable sized pieces also produced their own challenges.

Being derived from denatured proteins and therefore essentially without structure themselves, they were often difficult to solubilize and purification schemes for them usually required more extreme solvent conditions than the fractionation of the smaller soluble mixtures. Gel filtration was particularly useful in these applications because it was better suited to tolerate such solvents (including detergents).

It is noteworthy that the spinning cup technology of Begg and Edman (1967) became much more amenable to the an­alysis of smaller peptides with the discovery that polybrene added to the reaction mixture materially reduced losses of sample, thus greatly enhancing its value in large-scale sequence projects.