Multiple Alleles and Blood Groups

An allelic series (or "multiple alleles") describes a single gene locus that exists in three or more alternative DNA sequence forms circulating in a population. This is not a violation of Mendelian rules — it is Mendelian genetics applied to a locus with a richer mutational history. Every alteration to a gene's sequence — a nucleotide substitution, a small deletion, a frameshift — can create a new allele. Over evolutionary time, mutation, recombination, and selection accumulate many such variants at a locus. Crucially, an individual diploid organism still inherits only two alleles at that locus (one from each parent), so ordinary Mendelian segregation and dominance rules still apply to any single cross; it is only across the whole population that "multiple" alleles are visible. The ABO locus in humans exemplifies this perfectly, and it layers on two additional Mendelian concepts already established for two-allele systems. The alleles I^A and I^B each encode a functional but differently specific enzyme, and both are expressed simultaneously in a heterozygote (I^A I^B) — this is codominance, because both antigens appear on the cell surface, visibly and separably, rather than blending into an intermediate phenotype. The i allele, by contrast, is a loss-of-function (null) allele: it fails to make a functional enzyme. Both I^A and I^B are dominant over i, so I^A i individuals are phenotypically type A and I^B i individuals are phenotypically type B — indistinguishable at the phenotype level from I^A I^A and I^B I^B homozygotes, respectively, though distinguishable by genotyping or by family pedigree analysis. This gives four phenotypes (A, B, AB, O) generated from six genotypes, all governed by three alleles at one locus:

The gene and its product. The ABO locus sits on the long arm of chromosome 9 (9q34.2) and encodes a glycosyltransferase — an enzyme that transfers a sugar molecule onto an acceptor substrate. All red blood cells (and many other tissues) display a common precursor structure on their surface called the H antigen, itself built by a separate enzyme (encoded by the FUT1 gene, at a different locus) that adds fucose to a glycolipid/glycoprotein backbone. The ABO enzyme acts downstream of this H-antigen precursor.

• The I^A allele encodes an N-acetylgalactosaminyltransferase, which adds N-acetylgalactosamine to the H antigen, creating the A antigen.
• The I^B allele encodes a D-galactosyltransferase, differing from the A enzyme by only four amino acid residues, which instead adds galactose to the H antigen, creating the B antigen.
• The i allele carries a single-nucleotide deletion near the start of the coding sequence, causing a frameshift that truncates the protein into a catalytically dead product. No sugar is added, so the H antigen remains unmodified — this is blood type O.

In an I^A I^B heterozygote, both alleles produce their respective functional enzymes independently, in the same cell, and each modifies a separate pool of H-antigen substrate molecules on the same red blood cell surface. The result is a cell surface displaying both A and B antigens side by side — the molecular definition of codominance: two gene products, each fully and separately detectable.

Why this matters immunologically. The immune system distinguishes "self" from "non-self" surface antigens. A person's plasma spontaneously contains antibodies against whichever ABO antigen(s) they themselves lack (these are called naturally occurring antibodies, arising from early-life exposure to structurally similar antigens on gut bacteria). This is the basis of the classic transfusion table:

Type O is the "universal donor" for red cells because its cells carry no A or B antigen for a recipient's antibodies to attack; type AB is the "universal recipient" because its plasma carries no anti-A or anti-B antibodies to attack incoming donor cells.

Segregation across a cross. Because ABO is a single autosomal locus, ordinary Mendelian segregation applies: each parent passes one of their two alleles to each gamete, and Punnett square logic predicts offspring genotype ratios exactly as with any two-allele gene — the only twist is tracking three possible alleles and translating genotypes to phenotypes through the dominance hierarchy (I^A = I^B > i).

Karl Landsteiner's serum-mixing experiments (1900–1901). Before Landsteiner, transfusions were essentially a gamble: some patients tolerated donor blood, others suffered violent, sometimes fatal reactions involving fever, kidney failure, and clumping (agglutination) of red blood cells. Physicians had no systematic way to predict which donor–recipient pairs were safe. Landsteiner, working at the Pathological-Anatomical Institute in Vienna, took blood samples from himself and five colleagues. He separated each sample into serum (the antibody-containing liquid fraction) and red blood cells, then systematically mixed every person's serum with every person's red cells in all combinations and observed under the microscope whether the red cells clumped together (agglutinated) or stayed dispersed. The results were strikingly consistent and non-random: agglutination occurred in some serum–cell combinations but never in others, and the pattern sorted every individual into one of what Landsteiner initially called groups A, B, and C (C was later renamed O, from the German "ohne," meaning "without," since these cells carried neither A nor B antigen). Two of Landsteiner's students, Alfred von Decastello and Adriano Sturli, subsequently identified a fourth, rarer group — AB — in 1902, whose cells agglutinated with neither anti-A nor anti-B serum from group A or B individuals, but whose own serum agglutinated everyone else's cells.

What this proved. The reciprocal, all-or-none pattern of agglutination showed that red blood cells carry discrete, heritable surface markers (antigens) and that plasma carries corresponding antibodies against whichever markers a person's own cells lack. This was a functional assay for a genetic trait decades before DNA sequencing existed. Crucially, when the ABO blood types were later tracked through families (by Felix Bernstein in 1924, using population and pedigree data), the phenotype ratios did not fit a simple two-independent-gene model (which some earlier researchers, including von Dungern and Hirszfeld, had proposed, treating A and B as separate loci). Bernstein's statistical analysis of large population samples showed the data fit a single locus with three alleles far better, correctly predicting that an AB parent could never have an O child, and that two A-type parents could sometimes have an O child (if both were I^A i heterozygotes) — outcomes the two-locus model could not cleanly explain. This resolved a genuine scientific controversy and cemented the one-locus, three-allele model that stands today. The clinical payoff was immediate and enormous: cross-matching blood by ABO type became standard practice, transfusion mortality plummeted, and blood banking became feasible for the first time — work for which Landsteiner received the 1930 Nobel Prize in Physiology or Medicine.

Example 1 — a three-generation ABO pedigree with an unexpected genotype revealed. A woman of blood type A and a man of blood type B have four children with blood types A, B, AB, and O. What are the parents' genotypes, and what is the probability that a fifth child would be type O? Reasoning: the appearance of an O child (ii) proves that both parents must carry a hidden i allele, since ii requires one i from each parent. So the mother, despite being phenotype A, must be I^A i (not I^A I^A). The father, phenotype B, must be I^B i (not I^B I^B). This single observation — the O grandchild — unambiguously resolves both parental genotypes that phenotype alone could not distinguish. Cross: I^A i × I^B i Gametes from mother: I^A, i (each ½) Gametes from father: I^B, i (each ½)

Offspring ratio: 1 AB : 1 A : 1 B : 1 O, i.e., each phenotype at probability ¼. This matches the four types actually observed among the four children, and it tells us that any fifth child has a ¼ (25%) probability of being type O — the same probability the first four children faced independently each time, since meiosis has no memory of prior outcomes.

Example 2 — a disputed-paternity style reasoning problem (and its real limits). A child is blood type O. The mother is blood type A. Two men are considered as possible fathers: Man 1 is type AB; Man 2 is type B. Which man, if either, can be excluded as the biological father on genetic grounds alone? Reasoning: an O child has genotype ii, meaning the child inherited an i allele from each biological parent. The mother, type A, could be I^A I^A or I^A i — but since her child is ii, she must be I^A i (she must carry a hidden i to pass one on). Now examine the candidate fathers:

• Man 1 (AB) has genotype I^A I^B. He has no i allele at all to contribute. He cannot be the father — he is excluded with certainty.
• Man 2 (B) could be I^B I^B or I^B i. Only the I^B i genotype could contribute an i allele to produce an O child. So Man 2 is not excluded, but exclusion-by-ABO-typing can never prove paternity — it can only rule candidates out. Man 2 remains merely consistent with being the father; establishing him as the actual father requires much more discriminating evidence (today, STR/DNA profiling).

This example captures both the power and the sharp limit of classical blood typing in forensics: it is an excellent exclusion tool (Man 1 is genetically impossible) but a poor inclusion tool (Man 2 is only "not ruled out," shared by a large fraction of the population with compatible types). This is precisely why blood typing was historically used in paternity disputes only to exonerate, never to convict or confirm — a distinction real courts have had to grapple with.

Karl Landsteiner (1868–1943) trained as a physician in Vienna and spent much of his early career in pathology and serology, a discipline then focused on immune reactions and the chemistry of blood and antibodies. He was meticulous and famously restless, moving between institutions (Vienna, then the Netherlands, then the Rockefeller Institute in New York) as political and economic upheaval in Europe disrupted his research environment through two world wars. His defining insight was methodological as much as biological: rather than assuming all human blood was immunologically identical (as many contemporaries did, attributing transfusion reactions to disease or technical error), he suspected that individual variation itself might be the cause, and he tested this directly and systematically by cross-mixing sera and cells from healthy colleagues — a simple, elegant design that any well-equipped lab could have run, but that required the conceptual leap of expecting to find heritable human diversity in something as seemingly uniform as blood. Landsteiner went on to co-discover the Rh blood group system in 1937–1940 (with Alexander Wiener), extending the same logic of antigen–antibody typing to a second, independently inherited red cell marker, and made foundational contributions to virology (including work related to poliovirus) and to the general theory of antigens and haptens.

Transfusion medicine remains the most direct application: every unit of donor blood is ABO- and Rh-typed and cross-matched before transfusion, and ABO incompatibility is still a leading preventable cause of acute hemolytic transfusion reactions when protocols fail. Organ and tissue transplantation likewise requires ABO compatibility between donor and recipient, since mismatched ABO antigens on donor vascular endothelium trigger rapid antibody-mediated rejection; transplant matching programs explicitly track ABO type alongside HLA typing.

Hemolytic disease of the newborn can involve ABO incompatibility between mother and fetus (though it is typically milder than Rh-based disease), motivating routine prenatal blood typing. Population and evolutionary genetics use ABO allele frequency variation across human populations as one of many classic markers for studying migration history, though modern ancestry work has moved to genome-wide SNP data. Glycobiology and cancer research study altered glycosyltransferase expression (including ABO-pathway enzymes) as a marker of tumor cell surface changes. Forensic and paternity contexts still teach ABO exclusion logic as the conceptual ancestor of modern DNA profiling, even though STR and SNP-based DNA testing has almost entirely replaced blood typing for actual casework. Finally, the ABO locus is a favorite teaching bridge toward more advanced Mendelian topics: extending "one locus, several alleles" thinking prepares students for Lethal Alleles (where certain allele combinations are inviable rather than merely differently expressed) and for Pleiotropy (where a single gene, like the ABO locus itself, has effects reaching beyond red cell typing into subtle associations with clotting factors and disease susceptibility).

Ask the tutor: How would the expected offspring ratios change if I were also tracking the Rh blood group locus alongside ABO in the same cross?