Protein Analysis

In 1956, by fractionating the peptide products of digestion of human Hb with the proteolytic enzyme, trypsin, Ingram found 30 discrete peptide fragments. Trypsin cuts polypeptide chains at the amino acids arginine and lysine. Analysis of the 580 amino acids of human Hb had previously shown there to be a total of 60 arginine and lysine residues, suggesting that Hb was made up of two identical peptide chains with 30 arginine and lysine residues on each chain.

At about the same time, a family was reported in which two hemoglobin variants, Hb S and Hb Hopkins II, were both present in some family members. Several members of the family who possessed both variants had children with normal Hb, offspring who were heterozygous for only one Hb variant, as well as offspring who, like their parents, were doubly heterozygous for the two Hb variants. These observations provided further support for the suggestion that at least two different genes were involved in the production of human Hb.

Shortly thereafter, the amino-terminal amino acid sequence of human Hb was determined and showed valine–leucine and valine–histidine sequences in equimolar proportions, with two moles of each of these sequences per mole of Hb. This was consistent with human Hb being made up of a tetramer consisting of two pairs of different polypeptides referred to as the α- and β-globin chains.

Analysis of the iron content of human Hb revealed that iron constituted 0.35% of its weight, from which it was calculated that human Hb should have a minimum molecular weight of 16,000 Da. In contrast, determination of the molecular weight of human Hb by physical methods gave values of the order of 64,000 Da, consistent with the suggested tetrameric structure, α₂β₂, with each of the globin chains having its own iron-containing group—heme (Figure 10.1).

FIGURE 10.1 Diagrammatic representation of one of the globin chains and associated porphyrin molecule of human hemoglobin.

Subsequent investigators demonstrated that Hb from normal adults also contained a minor fraction, constituting 2% to 3% of the total Hb, with an electrophoretic mobility different from the majority of human Hb. The main component was called Hb A, whereas the minority component was called Hb A₂. Subsequent studies revealed Hb A₂ to be a tetramer of two normal α chains and two other polypeptide chains whose amino-acid sequence resembled most closely the β chain and was designated delta (δ).

Protein Studies

Amino acid sequencing of the various globin polypeptides in the 1960s showed that the α chain was 141 amino acids long compared with the β chain’s 146 amino acids. Their amino acid sequences were similar but by no means identical. The sequence of the δ chain differs from the β chain by 10 amino acids, and analysis of the γ chain showed that it also most closely resembles the β chain, differing by 39 amino acids. In addition, it was found that there were two types of HbF, in which the γ chain contain either the amino acid glycine or alanine at position 136, designated (G)γ and (A)γ, respectively. Partial sequence analyses of the ζ and ε chains of embryonic Hb suggest that ζ is similar in amino acid sequence to the α chain, whereas ε resembles the β chain.

Thus, there are two groups of globin chains, the α-like and β-like, possibly derived from an ancestral Hb gene that has undergone a number of gene duplications and diverged over evolutionary time.

Globin Gene Mapping

The first evidence for the arrangement of the various globin structural genes on the human chromosomes was provided by analysis of the Hb electrophoretic variant, Hb Lepore. Comparison of trypsin digests of Hb Lepore with Hb from normal people revealed that the α chains were normal, whereas the non-α chains appeared to consist of an amino-terminal δ-like sequence and a carboxy-terminal β-like sequence. It was therefore proposed that Hb Lepore could represent a ‘fusion’ globin chain that had arisen as a result of a crossover coincidental with mispairing of the δ- and β-globin genes during meiosis as a result of the sequence similarity of the two genes and the close proximity of the δ- and β-globin genes on the same chromosome (Figure 10.3). If this hypothesis was correct, it was argued that there should also be an ‘anti-Lepore’ Hb (i.e., a β–δ-globin fusion product in which the non-α-globin chains contained β-chain residues at the amino-terminal end and δ-chain residues at the carboxy-terminal end). In the late 1960s, a new Hb electrophoretic variant, Hb Miyada, was identified in Japan, whereby analysis of trypsin digests showed it to contain β-globin sequence at the amino-terminal end and δ-globin sequence at the carboxy-terminal end, as predicted.

FIGURE 10.3 Mechanism of unequal crossing over which generates Hb Lepore and anti-Lepore.

(Adapted from Weatherall DJ, Clegg JB 1981 The thalassaemia syndromes. Blackwell, Oxford.)

Further evidence at the protein level for the physical mapping of the human globin genes was provided by the report of another Hb electrophoretic variant, Hb Kenya. Amino acid sequence analysis suggested it was a γ–β fusion product with a crossover having occurred somewhere between amino acids 81 and 86 in the two globin chains. For this fusion polypeptide to have occurred, it was argued that the γ-globin structural gene must also be in close physical proximity to the β-globin gene.

Little evidence was forthcoming from protein studies about the mapping of the α-globin genes. The presence of normal Hb A in individuals who, from family studies, should have been homozygous for a particular α chain variant, or obligate compound (double) heterozygotes (p. 114), suggested there could be more than one α-globin gene. In addition, the proportion of the total Hb made up by the α chain variant, in those heterozygous for those variants, was consistently lower (less than 20%) than that seen with the β chain variants (usually more than 30%), suggesting there could be more than one α-globin structural gene.

Globin Gene Structure

The detailed structure of globin genes has been made possible by DNA analysis. Immature red blood cells, reticulocytes, provide a rich source of globin messenger RNA (mRNA) for the synthesis of complementary DNA (cDNA)—reticulocytes synthesize little else! Use of β-globin cDNA for restriction mapping studies of DNA from normal individuals revealed that the non-α, or β-like, globin genes are located in a 50-kilobase (kb) stretch on the short arm of chromosome 11 (Figure 10.4). The entire sequence of this 50-kb stretch containing the various globin structural genes is known. Of interest are non-functional regions with sequences similar to those of the globin structural genes—i.e., they produce no identifiable message or protein product and are pseudogenes.

FIGURE 10.4 The α- and β-globin regions on chromosomes 16 and 11 showing the structural genes and pseudogenes (ψ) and the various hemoglobins produced.

(Adapted from Carrell RW, Lehman H 1985 The haemoglobinopathies. In: Dawson AM, Besser G, Compston N eds. Recent advances in medicine 19, pp. 223–225. Churchill Livingstone, Edinburgh, UK.)

Studies of the α-globin structural genes have shown that there are two α-globin structural genes—α₁ and α₂—located on chromosome 16p (see Figure 10.4). DNA sequencing has revealed nucleotide differences between these two structural genes even though the transcribed α-globin chains have an identical amino acid sequence—evidence for ‘degeneracy’ of the genetic code. In addition, there are pseudo-α, pseudo-ζ, and ζ genes to the 5′ side of the α-globin genes, as well as an additional theta (θ)-globin gene to the 3′ side of the α₁-globin gene. The θ-globin gene, whose function is unknown, is interesting because, unlike the globin pseudogenes, which are not expressed, its structure is compatible with expression. It has been suggested that it could be expressed in very early erythroid tissue such as the fetal liver and yolk sac.

Structural Variants/Disorders

In 1975, Ingram demonstrated that the difference between Hb A and Hb S lay in the substitution of valine for glutamic acid in the β chain. Since then, more than 300 Hb electrophoretic variants have been described due to a variety of types of mutation (Table 10.2). Some 200 of these electrophoretic variants are single amino acid substitutions resulting from a point mutation. The majority are rare and not associated with clinical disease. A few are associated with disease and relatively prevalent in certain populations.

Table 10.2 Structural Variants of Hemoglobin