Re: What happened in human history to make O blood dominant in our society?

Hi Jon,

You are asking two related questions. To paraphrase, the first question is, "what happened to make the ABO O blood type (phenotype) so common?" and the second question is, "why do dominant alleles (like the ABO A and B alleles) not result in phenotypes that predominate in a population?” Your first question will take a long time to answer thoroughly, so I will answer the second question first.

The terms dominant and recessive are terms of Mendelian genetics that are primarily useful in explaining the genotypes and phenotypes of individual organisms. When you are considering multiple organisms (i.e., a population or a species), you need to think in terms of population genetics, which is more concerned with allele frequency than phenotype. I think that in your question, you are confusing the genetic term dominant with the word predominant. These words do not mean the same thing, and they cannot be used interchangeably; the reason why is explained in the answer you mentioned in your question.The only time you should be using the Mendelian term dominant is when you are trying to determine the phenotype for a given genotype. There is no reason why a dominant allele should result in a predominant phenotype.

The first part of your question, what happened to make the ABO O phenotype so common? assumes that the O phenotype has not always been prevalent in our species. This may or may not be a valid assumption, but for now, lets assume that it is. At this point, I should also make it clear that, while we can make a lot of inferences about what is going on with these alleles, we don’t know for certain what the exact events were that gave rise to the situation we see today. We probably can't ever know, but we can put together a good theory as to what is likely.

So, before getting to your question, I'm going to quickly recap the way the ABO blood system works. It is easy to oversimplify this system, but to thoroughly address your question I'll have to be very specific. I hope you can bear with me (I warned you that this answer would be long) if you already know what follows.

An ABO blood type is determined by the presence or absence of A or B antigens. When only an A or B antigen is present, the blood type (phenotype) is either A or B, and when both are present, the phenotype is AB. However, an O blood type does not automatically mean that there are no antigens present. The A and B antigens are oligosaccharides that are synthesized by terminal glycosyltransferase enzymes (from here on out, I’ll just call these the A enzyme and the B enzyme). The A and B enzymes synthesize the A and B antigen from a common oligosaccharide precursor, known as the H antigen. An O blood type usually means that the precursor H antigen is present instead of the A or B antigen. Because of this, some people refer to the ABO antigens as the ABH antigens.

The ABH antigens are just 3 of almost 200 oligosaccharide antigens that are found on our cell surfaces, and in addition to the A and B enzymes, there is a host of other enzymes that produce these antigens. In fact, some people lack the enzyme that synthesizes the H antigen in saliva and on some skin cells. These people (called non-secretors) appear to have the O phenotype if you test their saliva instead of their blood, and comprise about 20% of the population of Europe (as well as people of European ancestry). However, for the purpose of your answer, I am only going to focus on the alleles of the ABO locus; just remember that some people who seem to have the O phenotype are actually non-secretors who would otherwise (if they were secretors) have A, B, or AB phenotypes.

The A and B enzymes are encoded by different alleles at the ABO locus. If you have a functional copy of the A or B allele, you will make A or B antigen from H antigen (resulting in A, B, or AB phenotypes). In order to have an O phenotype, you must have two non- functioning alleles (O alleles) at the ABO locus (with the exception of the caveats above). This means that the O phenotype is a recessive Mendelian trait and, your question becomes, why has the frequency of the non-functional O allele risen over the course of the history of the human species?

With your question phrased this way, we can see that it is a question of evolution, because evolution is defined as a change in allele frequency over time. There are three principal reasons why the frequency of an allele changes. In no particular order, these reasons are mutation, natural selection, and genetic drift. There are many good discussions of natural selection and genetic drift already available in the MadSci Archives and elsewhere, and I won’t go into great detail about them here (because I just cited two of my own answers that give detailed examples).

Lets take a look at each of these phenomena.

Mutation results when there is a change in the DNA sequence of particular gene. The most common mutations are single nucleotide changes, and these are thought to occur in a given gene with a regular frequency. A new mutation that gets passed on to the next generation automatically results in a change in allele frequency, simply because the new allele did not exist in the previous generation.

When we look at the DNA sequences of the ABO alleles of many different people we see something very interesting. First of all, there are several different alleles that code for functional A or B enzyme, and a great number of alleles that code for non-functional ABO enzyme (O alleles). In addition, some of these A and B alleles seem to code for proteins with greatly reduced enzyme activity, such that homozygotes for these alleles appear to have an O phenotype, even though they produce small amounts of A and B antigens. So, right off, the answer to your question is going to be pretty complex, because there is not simply one O allele. The DNA sequences of A alleles are all very similar to one another. The DNA sequences of B alleles are all very similar to one another. However, there seem to be at least three distinct "families" of O alleles, and two of these appear to be derived from (once functional) A alleles.

These O alleles have been generated by mutations in the A (and B) alleles. Since single nucleotide mutations occur with a regular frequency, we would expect O alleles to be generated from A (and B) alleles at a very slow, random rate. If ABO allele frequencies were otherwise unaffected by natural selection and genetic drift (discussions to follow), mutation alone could result in a gradual rise in the frequency of O alleles (and therefore the O phenotype) over time.

In large populations, genetic drift usually acts to counter this process of accumulation, because high frequency alleles have a better chance of being contributed to the next generation (simply because there are more of them), while low frequency alleles (such as new O alleles) have a better chance of being lost.

However, when dramatic examples of genetic drift occur (as in the case of a founder effect or population bottleneck), alleles that were at low frequencies in the parent population can be raised to high frequencies in the colony. In the extreme case of a population bottleneck, where a population is reduced to a few individuals, the alleles of the survivors have a good chance of rising to high frequencies in the populations of their descendants. This is thought to be the case for Native Americans, who seem to be descended from a small founding population. The frequencies of O alleles are extremely high in Native American groups, and this high frequency can be explained if one or a few of the founders who colonized the Americas had O alleles. This is likely the case for high frequencies of O alleles in other regions of the world that are thought to have been colonized by small groups of people.

To understand how natural selection affects ABO allele frequencies, we have to take a look at the biological function of the ABH (and other oligosaccharide) antigens. It is still not completely clear why we go to all the trouble to synthesize these oligosaccharide antigens, but it looks like they are involved in interaction between cells (cell adhesion and cell motility), and some of them seem to confer a resistance to programmed cell death (apoptosis). The A and B antigens are also present on the von Willebrandt blood clotting factor (this factor is larger than a platelet), and seem to increase its lifespan and blood clotting ability.

Right here, there seems to be some potential for the influence of natural selection on ABO allele frequencies. For example, lower levels of von Willebrandt factor are observed in O phenotype individuals, but these individuals also have a lower risk of coronary heart diseases (possibly because of less effective von Willebrandt factor). Perhaps because of the increased resistance to apoptosis connected to the presence of A and B antigens, blood type A, B and AB individuals have higher incidences of gastro-intestinal cancers than blood type O individuals. However, these conditions (coronary heart disease and gastro-intestinal cancers) tend to become problematic after people have reproduced, so they might not have that big an effect on the frequency of the O phenotype.

In addition to these native functions, some pathogens make use of these same antigens to infect our cells, and some people think that terminal glycosyl groups like the A and B antigens evolved to mask the H antigen from pathogens that recognized it. For example, blood type O people have an increased risk of developing gastrointestinal ulcers, which result from an infection of Helicobacter pylori. Norwalk viruses (which cause acute non-bacterial gastroenteritis) bind to H antigens, but do not bind well to B antigens. Non- secretors (who do not produce H antigen) are immune to infection by Norwalk viruses, while O phenotype individuals have increased risk of infection by Norwalk viruses (relative to A, B and AB phenotype individuals).

It seems likely that some pathogens eventually evolved an ability to use the A and B antigens to infect cells. For example, Staphylococcus saprophyticus preferentially causes urinary tract infections in blood type A and AB women. In addition, some pathogens present these same antigens on their own cell surfaces. For example, Escherichia coli strain 086 presents the B antigen, and this antigen is readily recognized by anti-B antibodies in blood type A and O people, protecting them from infection.

These pathogen-related observations have given rise to a theory that explains the current diversity of ABO alleles. Because there are so many different pathogens that can make use of the ABH antigens to infect cells, natural selection favors populations comprised of individuals with many different ABH phenotypes. If all pathogens used only the H antigen to infect cells, then there would be a selective pressure for everyone in the population to have only A or B antigens, but since there are pathogens that use H, A or B for infection, the population can survive by having some individuals with no H, some with no A and some with no B antigens. Similarly, individuals with no A antigen can make antibodies against pathogens with A antigen, etc. In this way, even if some individuals are infected and die, the population survives. This form of selection (called balancing selection) results in relatively high allele frequencies for multiple alleles (as we see at the ABO locus), and can work to counter the action of genetic drift in eliminating low frequency alleles.

ABO phenotype may also play an important role in fertility, in a phenomenon known as intrauterine selection. The idea here is that the mother's immune system will produce anti- bodies to foreign blood groups, which could attack a fetus with a different ABO phenotype. Quite a few studies have concluded that the blood types of the mother and the fetus can affect the chance of a spontaneous miscarriage, several have inferred selection against the A allele, and at least one study suggested that natural selection would favor a male population with high frequencies of the O phenotype (because the fetus would have an ABO phenotype similar to the mother's).

So, natural selection could operate to increase the frequency of O alleles in a number of ways; the O phenotype may be advantageous in some cases, balancing selection could operate to maintain elevated levels of O alleles, and intrauterine selection could favor O phenotype men.

So, as I promised, that was a long answer. To recap, O alleles are non-functional versions of A and B alleles, and are generated from A and B alleles via mutation. These O alleles accumulate in populations over time, rising in frequency. Genetic drift counters this gradual rise in frequency, except in dramatic cases such as founder effects and bottlenecks, where O alleles can rise to extremely high frequencies. Finally, there are several ways that natural selection could affect O allele frequency, with pathogen-directed balancing selection and intrauterine selection maintaining levels O allele that are elevated relative to what would be expected under mutation and genetic drift alone.

Finally, I want to suggest that the current situation (with a relatively high frequency of O alleles) may not be that different from the way things have been for millions of years. Looking at other primate species (apes, old world monkeys, and new world monkeys), we can see appreciable frequencies of A, B and O alleles as well (although there are some primate species that have only A or B phenotypes). With the exception of B alleles in humans and gorillas, the ABO alleles in other primate species differ greatly from human ABO alleles. It seems likely that (because they can be generated by mutation) O alleles have always been present in primate populations. The ABO alleles in our species appear to be millions of years old. So, the alternative answer to your question might be that the frequency of O alleles has always been sort of high, and that it has not necessarily risen over the course of human history.

Here are some references that go into even greater detail about this stuff.

ABH/oligosaccharide antigen function and disease association

Marionneau S, Cailleau-Thomas A, Rocher J, Le Moullac-Vaidye B, Ruvoen N, Clement M, Le Pendu J. (2001) ABH and Lewis histo-blood group antigens, a model for the meaning of oligosaccharide diversity in the face of a changing world. Biochimie. Jul;83(7):565-73. This is an excellent review that covers the function and biochemistry of oligosaccharide antigens.

Lindesmith L, Moe C, Marionneau S, Ruvoen N, Jiang X, Lindblad L, Stewart P, LePendu J, Baric R. (2003) Human susceptibility and resistance to Norwalk virus infection. Nature Medicine. May;9(5):548-53.

ABO sequence diversity and evolution

Seltsam A, Hallensleben M, Kollmann A, Blasczyk R. (2003) The nature of diversity and diversification at the ABO locus. Blood. Oct 15;102(8):3035-42.

Saitou N, Yamamoto F. (1997) Evolution of primate ABO blood group genes and their homologous genes. Molecular Biology and Evolution. Apr;14(4):399-411.

Sumiyama K, Kitano T, Noda R, Ferrell RE, Saitou N. (2000) Gene diversity of chimpanzee ABO blood group genes elucidated from exon 7 sequences. Gene . Dec 23;259(1-2):75-9.

Corvelo TC, Schneider H, Harada ML. (2002) ABO blood groups in the primate species of Cebidae from the Amazon region. Journal of Medical Primatology. Jun;31(3):136-41.

Intrauterine selection

Schaap T, Shemer R, Palti Z, Sharon R. (1984) ABO incompatibility and reproductive failure. I. Prenatal selection. American Journal of Human Genetics. Jan;36(1):143-51.

Millard AV, Berlin EA. (1983) Sex ratio and natural selection at the human ABO locus. Human Heredity. 1983;33(2):130-6.

Satyanarayana M, Vijayalakshmi M, Rao CS, Mathew S. (1978) ABO blood groups and fertility -- with special reference to intrauterine selection due to materno-fetal incompatibility. American Journal of Physical Anthropology. Nov;49(4):489- 96.