Fifty-eight years of hemoglobin analysis.

Hemoglobinopathies are the archetypal molecular disease, and from 1949, when Linus Pauling et al. first demonstrated that hemoglobin (Hb) S could be separated from Hb A by paper electrophoresis (1 ), the detection and identification of Hb variants has provided a testing ground for new methods of protein analysis. The actual molecular basis of sickle cell anemia was not finally elucidated until 1959, when Vernon Ingram perfected the techniques of tryptic peptide fingerprinting (2 ). These modest beginnings led to the evolution of more sensitive analytical techniques, and the number of structurally characterized Hbs has now expanded to more than 1000 known variants. During the 1960s and 1970s, 2-dimensional peptide mapping on paper and ion exchange mapping on Dowex resins led to the identification of many new variants until these techniques were eventually superseded by reversed-phase HPLC in the 1980s. Along with the introduction of reversed-phase chromatography, the need for more readily automated procedures for the quantification of HbA2 and HbF for thalassemia assessment and HbA1c in relation to diabetic control led to the introduction of cation exchange HPLC systems as a means of separating Hb species. Because thalassemia is the most common genetic disease world wide, and because of the prevalence of diabetes in western cultures, many ion exchange separations are performed annually, leading to the incidental detection of large numbers of “interfering” Hb variants. In the 1990s, refinements of soft-ionization techniques, such as fast atom bombardment, electrospray, and matrixassisted laser desorption, heralded the introduction of protein mass spectrometry, and those working on Hb analysis were among the first to exploit the power of this new technology in a diagnostic setting. In this issue of Clinical Chemistry, Kleinret and colleagues (3 ) provide an interesting account of their experience in the analysis of Hb samples from 2105 individuals referred to their laboratory over a 5-year period. Each sample was analyzed by cation exchange HPLC, reversed-phase HPLC, and mass spectrometry to detect variants with altered charge, hydrophobicity, or mass, respectively. The application of cation exchange chromatography showed quantitative variations in HbA2 and HbF, indicating some 204 cases of thalassemia, and the sequential use of ion exchange chromatography, reversed-phase chromatography, and mass spectrometry led to the detection of some 226 genetic variants of Hb. Of these, 4 contained substitutions (Asn3Ser, Thr3Ser, or Gly3Ser) that lacked charge or hydrophobicity changes and could be detected only by mass spectrometry, and a further 4 contained variants detectable only by their hydrophobicity and mass differences. Although these statistics indicate that ion exchange alone would detect 218 of the 226 consecutive variants, it should be emphasized that the variants were not unique. An alternative interpretation of the data gives a better appraisal of the value of the combined approach. Of the 18 unique variants that were identified, the use of ion exchange alone detected 12, ion exchange plus reversed-phase detected 15, and the combined use of all 3 methods led to the detection of all 18 variants. The traditional ion exchange and alkaline cellulose acetate electrophoresis methods are readily capable of detecting the majority of common variants, such as HbC, HbD, or HbE. However, the substitutions involved in these, and similar charge variants (Glu7Lys, Glu7Gln, Asp7Asn, and Lys7Gln) involve mass changes of 1 Da or less, and are not detectable by mass spectrometry. So the combined use of mass spectrometry with electrophoresis, or ion exchange, becomes very powerful because the procedures detect entirely different subsets of mutations. All the mutations reported by Kleinret et al. could have been detected by these 2 methods alone, so why include reversed-phase HPLC in the protocol? The practical answer to this question is revealed by Kleinret et al.’s comparative study of different types of mass spectrometer. Both electrospray ionization/quadrupole and MALDI-TOF instruments can resolve 50/50 mixtures of globin species if their mass difference is 6 Da. Therefore the allowable point substitution of Asn3 Ile (–1 Da) would go undetected if only charge and mass were considered. Indeed, other mutations above the 6 Da threshold may also be missed if reversed-phase HPLC is not included in the analysis; those expressed in lower amounts, such as -chain variants, and some unstable, or thalassemic, -chain variants. Another interesting observation was that Fourier transform ion cyclotron resonance mass spectrometry offered no practical advantage over the more modestly priced alternatives in relation to variant detection. As well as providing insights from their practical experience of Hb analysis, Kleinret and colleagues asked the question, how many different variants are out there, and how many of them have been found? When they considered all possible point substitutions of both the and -globin genes, they found there were 1695 possible variants of adult Hb; 733 of which have been Clinical Chemistry 54:1 8 –10 (2008) Editorial