A. INTRODUCTION

Lists of hemoglobin (Hb) variants have appeared in many hematology and related text books (1,2), and since 1978 have been published once each year in the journal HEMOGLOBIN. A syllabus of Hb variants that lists the structural variation for each abnormality, its many (abnormal) properties, methodology used for detection and characterization, its incidence and occurrence, has been published twice. The first was by Lehmann and Kynoch (3), published in 1976, which listed some 200 variants, and the second by R.N. Wrightstone (4) in 1984. The present syllabus finds its basis in Wrightstone's work; however, the description of each variant is greatly modified and often shortened, which was necessary because of the need to accommodate almost 700 variants over several different categories (Table 1). These include single base mutations in the alpha2- or alpha1-globin genes, and in the beta-, G-gamma-, A-gamma-, and delta-globin genes (variants of the zeta or epsilon chain have not been reported); two-point mutations observed in alpha- and beta-globin genes; Hb variants resulting from the fusion of parts of delta- and beta-, or beta- and delta-, or A-gamma- and beta-globin genes; variants with elongated alpha or beta chains, at both the C-terminal and at the N-terminal ends, and Hb variants with alpha or beta chains with a deletion and/or insertion. Not included are deletions and mutations that result in an alpha-, or beta-, or gamma-, or delta-beta-, or gamma-delta-beta-, or epsilon-gamma-delta-beta-thalassemia (thal), even if such a change (point mutation or frameshift) occurs in one of the coding regions of a gene. Information regarding these abnormalities can be found in the repositories that are regularly published mainly in the journal HEMOGLOBIN [the latest beta- and delta-thal repository was published in 1995, issues 3&4, pages 213-236; the deletional hereditary persistence of fetal Hb (HPFH) and delta-beta-thal repository appeared in 1993, issue 6, pages 569-573; information about the alpha-thalassemias can be found in Ref. 5].

B. METHODOLOGY FOR DETECTION AND CHARACTERIZATION

I. Screening. Most Hb variants have been detected by electrophoretic methods (descriptions in Refs. 6 and 7). Hb S was observed by moving boundary electrophoresis in the late 1940s (8). Mass screening became possible with paper, cellulose acetate, and starch gel electrophoresis. In most institutions, these procedures have been replaced by isoelectrofocusing (IEF) (9). A commercial form of IEF that uses agar plates has been developed by Isolab, Inc. (Akron, OH, USA). This method has been used by many laboratories for several years with great success; it is presently the screening method of choice for both the Hbs in cord blood as well as adult blood samples.

Chromatographic procedures never played a significant role for screening until micro procedures were developed (reviewed in Ref. 10). These, however, were rapidly replaced by high performance liquid chromatography (HPLC) methodology because chromatograms developed on such columns could be completed in a short period of time, and the method could readily be automated. Mostly cation exchange material is used (for details, see Refs. 7 and 11).

Clinically significant Hb variants, however, are usually first observed by routine hematological procedures. A low Hb level, microcytosis, hypochromia, reticulocytosis, bilirubinemia, ahaploglobinemia may suggest the possible presence of an alpha- or beta-thal, or a (severely) unstable Hb type. Furthermore, a high Hb level (erythrocytosis) together with appropriate clinical observations may suggest a Hb variant with an amino acid replacement that lead to a (greatly) increased oxygen affinity. Indeed, most unstable Hb variants and high (or low) oxygen affinity variants were initially discovered this way. Their presence is then confirmed by one of the several stability tests and by the measurement of the oxygen affinity of whole blood or red cell lysate (6).

With the advance of DNA methodology it has become common practice in some laboratories to sequence the three coding regions of the alpha- and beta-globin genes. Most of the so-called silent mutations have been detected by this approach.

II. Characterization. Identification of an amino acid replacement in a Hb variant requires the isolation of the abnormal Hb and of the abnormal chain (alpha or beta or gamma or delta), digestion with a proteolytic enzyme (usually trypsin), separation of the smaller fragments, determination of the composition of each peptide with an amino acid analyzer, and sequencing of the peptide with the amino acid replacement. Isolation of the Hb is usually done by chromatography (10), occasionally by cation exchange HPLC, preparative starch-block electrophoresis or preparative IEF (9,11). The chromatographic procedure on CM-cellulose with phosphate-8 M urea developers is most popular for the separation of the two types of chains, although more recently (semi-) preparative reversed phase HPLC methodology is preferred by some laboratories, particularly when micro methodology is available for the entire structural analysis. Rarely, it is not possible to separate the abnormal Hb from its normal counterpart or the abnormal chain from the normal chain; in such cases, the study is continued with a mixture of the two (alpha^A+alpha^X or beta^A+beta^X).

The first method available for the separation of the proteolytic fragments was the fingerprinting technique, i.e. a combination of an electrophoretic and chromatographic separation. This technique was slowly replaced by macrochromatographic methods with cation exchangers (6), and more recently, reversed phase HPLC methodology has become popular. The latter method is fast, accurate, and most reproducible, and provides enough material of each peptide for further analysis (7). There are numerous suitable HPLC columns on the market that are excellent for this particular purpose. Amino acid analyses are made with automated amino acid analyzers; some 40 years ago one such analysis took two weeks and can at present be accomplished in 30-60 minutes. Instrumentation for automatic sequencing of peptides and proteins are available, and a simple manual procedure (12) will also provide the required data.

The development of polymerase chain reaction (PCR) methodology and sequencing techniques based on the procedure developed by Sanger et al (13) have revolutionized the characterization of Hb variants. Presently, it is common practice in quite a few laboratories to isolate DNA from 5-10 ml freshly collected blood, identify the chain (alpha, beta, gamma, or delta) where an amino acid replacement can be expected through protein analysis, amplify the mutated gene with an appropriate set of primers, and sequence the exons 1, 2 and 3 with the Sanger methodology. This approach has led to the discovery of some unexpected amino acid replacements that are often silent (for instance: Ala->Gly; Leu->Met, etc.) but greatly affect physicochemical or functional properties. The resulting Hb types can either not be separated from their normal counterparts or are so severely unstable that they are not detectable at the protein level. Furthermore, several neutral but "innocent" amino acid replacements have been detected with this approach and an increase in these cases is anticipated.

REFERENCES

C. THE ALPHA AND BETA CHAIN VARIANTS WITH SINGLE AMINO ACID SUBSTITUTIONS

I. Number and Types of Mutations. This group of variants comprises more than 75% of all abnormal Hbs that have been detected. Since the sequences of the two alpha-globin genes (alpha2- and alpha1-) and the beta-globin gene have been known for quite some time, it is possible to list the different mutations at the appropriate locations. Tables 2 and 3 provide this information for the alpha2- and alpha1-globin genes (there are no sequence differences between the coding regions of the alpha2 and alpha1 genes) and the beta-globin gene. Many of the mutations are deduced from the known sequences of the normal genes and the protein analytical data; slightly more than 10% of the mutations have been determined through DNA sequencing. Occasionally a discrepancy is observed, such as the Ala->Asp replacement at position 19 of the alpha chain (a GCC->GAC mutation has to be assumed rather than a GCG->GAC change), and at positions 50 and 67 of the beta chain (Thr->Lys at beta50 and Val->Asp at beta67). In reviewing these data the reader should use the information provided in Table 4 because the different amino acid replacements could not be added to Tables 2 and 3. Of the 141 codons of the alpha genes, as many as 99 have been found to be mutated; for several, three or four mutations have been discovered, while five mutations are known for codons 23, 75, and 94, and even six for codon 141. The mutations appear to occur at random; thus, either one of the three bases are replaced in the 199 known alpha-globin gene mutants. As many as 138 of the 146 codons of the beta-globin gene have been mutated and 335 different base changes have been recorded; five mutations are known for six different codons (22, 67, 97, 121, 143, and 146), six mutations for codon 92, and seven mutations for codon 99. It can be calculated that nearly 2,600 single base changes can occur in the alpha-globin gene (counting the alpha2 and alpha1 gene as one because of their identical sequences) with 141 codons, and in the -globin gene with 146 codons. Of these, 534 or slightly more than 20% have been discovered.

The different mutations are scattered over the alpha and beta genes without certain segments being more highly "mutatable" than others. Figs. 1 and 2 illustrate the distribution of the amino acid substitutions over the alpha chain and beta chain, respectively; no specific segment of these protein molecules has a significantly higher number of amino acid replacements. How have the many Hb variants been detected? Changes in functional or physicochemical properties leading to sickling, increased affinity for oxygen and erythrocytosis, and instability have played a most significant role, and indeed, numerous variants with these abnormal properties have been detected (some are listed in Tables 6, 7, 8, and 9). Furthermore, the use of electrophoretic methods for mass screening has resulted in the detection of numerous Hb variants with pK values different from that of Hb A; most of these variants have normal functional and/or physicochemical properties. Finally, the availability of methodology to rapidly screen the exons of the alpha and beta genes for mutations has resulted in the discovery of silent mutations not leading to clinical abnormalities. This is illustrated by the data presented in Table 5 which lists the 32 mutations involving codons AAA or AAG for lysine, and the 26 mutations involving codons GAA and GAG for glutamic acid in the beta-globin gene only. In both instances, numerous replacements have been found primarily because of charge differences and (to a lesser extent) of changes in affinity for molecular oxygen. Lys->Arg and Glu->Asp replacements are rarely observed. The situation is quite different for the neutral amino acids leucine (32 mutations) and glycine (28 mutations). Numerous silent or neutral replacements are possible, but these are not observed simply because DNA sequence methodology needs to be used. The majority of the amino acid substitutions involve Leu->Pro, Leu->Arg, Gly->Arg, and Gly->Asp replacements, leading to the synthesis of an unstable protein and/or one with a difference in charge allowing detection with an electrophoretic method.

II. Alpha And Beta Chain Variants With Altered Functional Properties. A review of the descriptions of the individual Hb variants will show that many do exhibit some changes in oxygen binding properties or stability. Often, the difference is minor and negligible for some Hb variants; occasionally the reported changes might be due to the methodology that was used for the isolation of the variant. Therefore, only the variants that show a distinct difference in the listed property will be listed in the following tables and figures.

The most striking and unique Hb variant is of course Hb S, or alpha₂beta₂6(A3)Glu->Val, discovered in 1949. Concentrated Hb S, when deoxygenated, forms birefringent gels made up of several strands of Hb S molecules. X-ray diffraction methods have been used to study crystals of deoxyHb S; the data showed, among others, that the hydrophobic side chain (valine at position 6 of the beta-S chain) fits into a pocket formed mainly by leucine at beta88 and phenylalanine at beta85 of an adjacent Hb S tetramer. Most likely, the resulting polymers are part of the polymer fibers seen in the sickled cell upon deoxygenation.

Table 6 lists all alpha and beta chain variants that are reported to have a greatly increased affinity for molecular oxygen (so-called high oxygen affinity variants) and cause a significant erythrocytosis in its carrier. Nearly all subjects with such an abnormality are heterozygotes; homozygotes for Hb Tarrant [alpha126(H5)Asp->Asn ] have been discovered. The fact that the majority of the carriers of high oxygen affinity variants likely have four alpha-globin genes makes it understandable that the presence of such an abnormality causes a less clinically significant condition than that of a high oxygen affinity beta chain variant. Figs. 3 and 4 identify the locations of the variants. Most of the high oxygen affinity alpha chain variants have mutations at the C-terminus or at positions occupied by amino acids essential for the alpha1 beta1 and alpha1 beta2 contacts. The numerous high oxygen affinity chain variants have mutations mainly at the C-terminus (residues 139 through 146), at the alpha1 beta2 and heme contacts mainly between the F and G helices (residues 94 through 103), at the alpha1 beta1 and alpha1 beta2 contacts of the B and C helices, while the replacement of hydrophilic residues in the heme pocket by hydrophobic residues may also cause a change in functional properties.

Table 7 lists the three alpha chain variants and the 23 beta chain variants that are reported to have a decrease in their affinity for oxygen. Quite a few variants are associated with a cyanosis in its carrier (mainly the M Hbs; Hb Denver; Hb Saint Mande) or with a mild-to-severe hemolytic anemia. Many amino acid replacements concern residues participating in the contact with heme, and four replacements in the beta chain involve the introduction of a proline residue in helix B (B13), C (C4), E (E20), and G (G19). Most of the variant Hbs are distinctly unstable and determination of its oxygen carrying properties is not always accurate.

III. Unstable Hbs. The term instability is difficult to define as several methods are available for its determination, such as Heinz body formation in red cells, heat instability in red cell lysates, instability of the protein in an environment of water and an organic compound (isopropanol, n-butanol) etc. Sometimes, an instability can be detected with one method but not with another. It is, therefore, not surprising to see that many Hb variants are listed in the literature as being mildly, moderately, or severely unstable. Table 8 lists the alpha and beta chain variants that are considered severely unstable; their presence is associated with a severe hemolytic anemia in the heterozygote. Quite a few severely unstable variants (mainly alpha chain variants) are not included in this list because their presence is associated with a thalassemic condition rather than a severe hemolytic anemia. These Hb variants (examples: Hb Agrinio, alpha29 Leu->Pro; Hb Quon Sze, alpha125 Leu->Pro; Hb Tunis-Bizerte, alpha129 Leu->Pro) are so unstable that their presence is most difficult to detect and DNA sequence analyses are required for their identification.

The mechanisms responsible for the decreased stability of Hb variants are many, such as replacement of a hydrophobic residue in the interior of the molecule by a hydrophilic residue; replacement of the essential distal and proximal histidines; etc. Of considerable interest is the effect of the introduction of a proline residue as well as that of a proline replacement (Table 9). The introduction of a proline residue often disrupts secondary structures because this amino acid cannot participate in an alpha helix except in the first three positions. As many as 39 variants have been reported that are characterized by the substitution of an Ala (or Leu, or Ser, or Arg, or Thr, or His) to Pro (11 in the alpha chain; 28 in the beta chain). Indeed, most of these variants are (severely) unstable. In contrast, a Pro->Arg (or Leu, or Ser, or Ala, or Thr) replacement does not often result in a decreased stability but more frequently in changes in functional properties, i.e. an increase in their affinity for oxygen.

D. THE GAMMA AND DELTA CHAIN VARIANTS WITH SINGLE AMINO ACID REPLACEMENTS

The number of gamma chain variants has increased to 68. As many as 38 are the result of mutations of the G-gamma gene and 20 that of the A-gamma gene. Interesting are the A-gamma-T variants with an additional mutation in the A-gamma gene; their detection is entirely based on DNA sequencing data. None of the gamma variants have abnormal properties that seriously affect the health of the carrier. Only a few have changes in functional properties leading to a (relatively mild) cyanosis in the baby. All variants occur at an (extremely) low frequency except Hb F-Sardinia with the A-gamma-T chain (A-gamma75 Ile->Thr) which is observed in nearly all populations, and often at a frequency of 0.15 and higher.

As of January 1996, 28 delta chain variants have been reported. Several are so-called silent variants, i.e. abnormalities not easily detectable by electrophoresis but detected by either a reversed phase HPLC experiment or by sequencing of amplified DNA. The latter procedure plays an increasingly important role in the identification of the delta chain abnormalities; more than one-third of the known variants have been identified with this method.

E. HBS WITH HYBRID CHAINS

The most common type is Hb Lepore-Boston-Washington which is found in Italians. Hb Lepore-Baltimore is found mainly in some Spanish families, and Hb Lepore-Hollandia in Papua, New Guinea. All three are associated with a beta-thal phenotype. Hbs Miyada, P-Congo, P-Nilotic, Lincoln Park, and P-India are so-called anti-Lepore Hb types. The abnormal beta-delta hybrid gene is located between a normal delta- and a normal beta-globin gene. Hb P-Nilotic has the highest frequency of these variants. The non-alpha chain of Hb Kenya is the product of an A-gamma-beta hybrid gene with a crossover between residues 81 and 86. A Hb Kenya heterozygote has an HPFH phenotype (G-gamma-HPFH) with an increased Hb F level and a normal Hb A2 level.

F. HB VARIANTS WITH TWO AMINO ACID SUBSTITUTIONS

G. HB VARIANTS WITH ELONGATED ALPHA OR BETA CHAINS

This is a group of interesting variants (seven alpha and six variants) with elongations at the C-terminus or at the N-terminus. Most important are the five alpha variants with mutations in the terminating codon TAA because all five are associated with an alpha-thalassemic phenotype. Common is Hb Constant Spring, found mainly in Oriental populations. The extended alpha chain (+31 residues) is rather unstable and occurs at low quantities in both heterozygotes and homozygotes. Hb Koya Dora occurs in certain Indian populations, while the other three are rare occurrences, found in only a few families. The other variants with an extension at the C-terminus are also rare; all four are the result of different frameshifts and the extension to each protein is therefore different.

Certain mutations in the first or second codons of the alpha- or beta-globin genes affect the initiation codon ATG (= methionine). This methionine is not released and, thus, the polypeptide chain is extended with one amino acid residue.

H. HBS WITH DELETIONS AND/OR EXTENSIONS IN THE ALPHA OR THE BETA CHAINS

This large group of 22 rare variants have highly characteristic changes in their alpha or chains, most frequently caused by crossovers between misaligned chromosomes resulting in loss of some codons or the introduction of some additional nucleotides (nts). In one instance, Hb Natal with the deletion of two residues from the C-terminus of the alpha chain, the change is due to a mutation of a TAC codon to a terminating codon TAA (codon 140). Some additional variants belonging to these categories have been reported but are not listed here because the proteins have never been isolated and analyzed.