ABO BLOOD GROUP SYSTEM LECTURE SLIDE 001

What you had better know about the ABO blood groups

ABO BLOOD GROUP SYSTEM LECTURE SLIDE 002

(From history to molecular genetics) This slide presentation was given at the IMPPC on October 5, 2009.

ABO BLOOD GROUP SYSTEM LECTURE SLIDE 003

ABO phenotype is a hereditary trait.

ABO BLOOD GROUP SYSTEM LECTURE SLIDE 004

Bernstein proposed the one gene locus-three alleles model in 1924.

A and B alleles are co-dominant against recessive O alleles.

Yamamoto family

The ABO blood types of Yamamoto family members

Potential ABO genotypes deduced from the ABO phenotypes

Mother's genotype cannot be A/A because her son has the O phenotype.

Therefore, the mother's genotype should be A/O.

Daughter's genotype cannot be A/A because her father has the O phenotype.

Therefore, the daughter's genotype should be A/O. 

In this family all the ABO genotypes were determined based on the phenotypes and the inheritance.

Here is another example of the ABO blood groups of another family.

Potential ABO genotypes deduced from the ABO blood groups are shown.

Mother's genotype cannot be B/B because her son has the A type.

Therefore, the mother's genotype should be B/O.

Son's genotype cannot be A/A because his mother has the B blood type.

Therefore, the son's genotype should be A/O. Father's genotype is either A/A or A/O. 

However, only with the inheritance information, we cannot determine the genotype.

Another example of ABO inheritance is shown. Father and mother exhibit weak A and B subgroup phenotypes, A2 and B3, respectively. 

Whereas their son has regular O phenotype, their daughter exhibits the rare A2B3 phenotype.

Potential ABO genotypes are deduced from the ABO blood types.

Father and mother cannot be homozygous (A2/A2, B3/B3) because their son has the O phenotype.

Therefore, the father and mother should have the ABO genotypes of A2/O and B3/O, respectively. 

In this family also, the ABO genotypes of the family members are determined only with the ABO blood groups and the inheritance.

A cis-AB inheritance case (right)

Father and his daughter have the A2B3 phenotype whereas the mother and son have the O blood group.

If we assume that A2B3 phenotype is caused by the presence of 2 alleles of A2 and B3, ....

If ...., there is a contradiction in the inheritance of the alleles. A father of A2/B3 cannot have a son with the O blood group. 

A mother of O/O cannot have a daughter with the A2B3 blood group, either.

In order to explain the inheritance, it is necessary to assume that the A2B3 phenotype is specified by a single gene, A2B3.

Because both A2 and B3 phenotypes are inherited as the single unit in the cis manner, it was named as cis-AB allele.

Contrastingly, the regular AB type is specified by 2 alleles (A and B) on 2 different homologous chromosomes. 

Therefore, it may be called as trans-AB although this naming is not used often.

ABO phenotypes were initially recognized as ABO blood groups defined by the agglutination pattern of human red blood cells (RBCs).

Later A and B antigens were also identified on a variety of cells and tissues, including cells in the gastrointestinal digestive tract.

A and B antigens are also expressed on the endothelial cells forming blood vessels.

Therefore, the ABO matching is also important in transplantation, in addition to blood transfusion.

Crime scene evidences such as blood, sweat, seminal fluid, and hairs may exhibit A and B antigens depending on the ABO genotype.

The ABO phenotypes of the suspects may be determined.

The ABO phenotypes of the evidences may also be determined and used to exclude certain suspects.

The potential suspects will be further narrowed down if......

The ..... if the ABO genotypes are also determined of the suspects and evidences.

In this example, the ABO blood groups and genotypes of the evidences do not match with those of the subjects 1, 3, 4, and 5.

Therefore, these subjects are excluded.

ABO blood groups were originally discovered with human RBCs.

Later ABO blood groups were also found in such primates as chimpanzees and gorillas.

A and B antigens were also found in such mammals as mice, rats,

, dogs,

, elephants, pigs, and others.

Humans have four major groups of A, B, AB, and O. 

However, depending on the species, the kinds of the ABO phenotype are different. 

Also many species do not express A and B antigens on RBCs.

The ABO blood groups were discovered by Landsteiner in 1900.

He separated cellular component and liquid component of blood from himself and his colleagues and mixed in various combinations.

When the cellular component and liquid component from the same individuals were mixed, nothing happened.

However, hemagglutination (agglutination of RBCs) was observed after mixing the cellular component and liquid component in some combinations.

Landsteiner found out that the individuals could be separated into groups depending on the hemagglutination pattern. 

Those 3 groups became A, B, and O. The fourth group (AB) was found in the next year by his colleagues.

The other and more important finding is that blood transfusion should be avoided in the combination that will result in the hemagglutination. 

He explained the hemagglutination pattern, by postulating the presence of antigens, A and B, on RBCs.

He also postulated antibodies against these antigens in the sera of the individuals that do not express the antigens.

Landsteiner assumed that the A individuals express A antigen on RBCs and contain anti-B antibody in their sera.

The B individuals express B antigens on RBCs and contain anti-A antibody in sera.

The AB individuals express both A and B antigens on RBCs, but contain neither of anti-A or anti-B antibodies.

The O individuals do not express A or B antigens on RBCs, but contain both anti-A and anti-B antibodies in sera.

The A and B antigens were found to be oligosaccharide antigens. They are similar, but also different.

The A antigen has a GalNAc residue whereas the B antigen has a galactose. 

These two sugars are different by the side group at the C2 position (-NHCOCH3 in GalNAc and -OH in galactose).

The H antigen, which lacks the GalNAc or galactose, was found present in the O individuals.

Careful examination of the structures of A, B, and H antigens led to ....

.... the hypothesis of the biosynthetic pathways of A and B antigens from the common precursor of H antigen.

We cloned cDNAs encoding the presumptive human A transferase.

We took a unique approach to utilize degenerate oligo primers to amplify cDNA fragment encoding the protein.

Rather than oligonucleotide probes, we used the 98 bp DNA fragment obtained by PCR as a probe to screen the cDNA library.

cDNA library was constructed, using RNA from the MKN45 cells that express A antigens and exhibit strong A transferase activity.

The cDNA library was screened, and several clones were identified.

The Northern hybridization was performed, using RNAs from cells that exhibited different ABO phenotypes.

Strong signal was observed with RNA from MKN45 cells as anticipated.

Signal was also detected with COLO 205 cells that exhibited the O phenotype.

Signal was also detected with cells that exhibited the AB phenotype.

Signal was also detected with cells that exhibited B phenotype.

The nucleotide sequence of one (59-5) of the cDNA clones from the MKN45 library was chosen as the standard and is schematically shown.

This clone was temporarily assumed to represent the A allele.

Two additional cDNA libraries were constructed and screened.

The cDNA clones from SW48 library were separated into 2 groups depending on the differences in the nucleotide sequence.

The cDNA clones, which were more homologous to 59-5, were assumed to represent the A alleles. 

The cDNA clones from the other group were assumed to represent the B alleles.

The cDNA clones from SW948 library showed the same nucleotide sequence. 

Compared with the 59-5 clone, these clones contained a single nucleotide deletion, which may abolish the enzyme activity by frameshifting.

Therefore, those cDNA clones from SW948 were assumed to represent the O alleles, which fits well with the non-functionality hypothesis of O alleles.

Two additional cDNA libraries were constructed and screened.

The cDNA clones from COLO 205 all contained the single nucleotide deletion. Additionally, they contained additional nucleotide substitutions.

Because of the single nucleotide deletion, the cDNA clones from COLO 205 were assumed to represent the O alleles.

The cDNA clones from the SW1417 library were divided into 2 groups.

The cDNA clones from one group showed the same sequence as the cDNA clones that were assumed to represent the O allele. 

The cDNA clones of the other group were identical to the SW48 clones that were assumed to represent the B allele.

Based on the results, these cDNA clones were shown to represent the ABO alleles. 

Additionally, the genotype of the individual from whom the SW1417 was derived was demonstrated to be BO.

The deduced amino acid sequences are compared among representative A, B, and O alleles.

Due to the single nucleotide deletion, the O allele encodes a truncated protein.

Four amino acid substitutions are observed between the proteins encoded by A and B alleles.

The presence and absence of the O allele-specific single nucleotide deletion result in the cleavage sites of the restriction enzymes, KpnI and BstEII, respectively. 

Therefore, we performed RFLP and Southern hybridization of genomic DNAs from the cells that were used for the cDNA library constructions. 

The BstEII digested 2 alleles of MKN45 cells and SW48 cells and 1 allele of SW1417 cells.

On the other hand, both alleles of the SW948 and COLO 205 cells and 1 allele of the SW1417 cells were cleaved with KpnI. 

Based on those results, it was shown that MKN45 cells and SW48 cells do not possess the single nucleotide deletion. 

The SW1417 is heterozygous, and SW948 and COLO 205 cells are homozygous of the single nucleotide deletion.

Three of the four locations of amino acid substitutions between A and B alleles also result in RFLP. 

For example, BssHII and NarI cut A/O alleles and B allele, respectively. 

We used PCR and RFLP to examine the genomic DNAs from those cells that were used to construct the cDNA libraries.

Another location that discriminates A/O alleles and B allele can also be examined by PCR and RFLP using HpaII and AluI.

In addition to the cells that were used for the cDNA library constructions, it was important to examine whether the same differences existed in blood specimens or not. 

We performed the RFLP and deduced the ABO genotypes of the 14 blood specimens.

The presence or absence of the O allele-specific single nucleotide deletion was determined by RFLP using KpnI and BstEII and Southern hybridization.

The discrimination between A/O alleles and B allele was done by PCR and RFLP using the combination of BssHII and NarI, and separately the combination of HpaII and AluI.

Possible ABO genotypes were deduced based on the RFLP results.

We compared the blood groups of those specimens with the deduced ABO genotypes, and found no contradictions.

This suggested that the allele-specific differences are commonly present.

The Central dogma of biology is that the genetic information flows from DNA to RNA, and then to protein.

In case of the ABO, the final products are not proteins but the oligosaccharide antigens. 

Therefore, the Central dogma may be modified to "from DNA to RNA, protein, and to oligosaccharide antigen".

By the cDNA cloning and characterization, we were able to correlate the nucleotide sequences of the A, B, and O alleles with the expression of the A, B, and H antigens.

In other words, we successfully demonstrated the Central Dogma of ABO.

By 1993 we determined the differences in the nucleotide sequences of several ABO alleles.

We found 4 amino acid substitutions that discriminate A and B transferases.

We found the single nucleotide deletion in a majority of O alleles. We also found the O alleles that lack the single nucleotide deletion but contain amino acid substitutions.

We also identified the A2 allele-specific single nucleotide deletion. We also found missense mutations in the A3, Ax, and B3 alleles.

We also identified amino acid substitutions that specified phenomena named cis-AB and B(A).

The number of ABO alleles characterized has increased drastically during the past decade.

This table prepared by Prof. Blumenfeld shows more than a hundred of alleles.

Only the differences in the nucleotide sequences from the standard A101 allele are shown. 

The orientation of the gene is from left to right, and the coding sequences are highlighted in gray.

The coding region of the ABO gene is carried on 7 exons.

The A alleles are indicated.

There are several cis-AB and B(A) alleles.

The B alleles are indicated.

The O alleles are indicated.

The four arrows indicate the locations of the four amino acid substitutions that discriminate A and B transferases.

The arrow indicates the location of the single nucleotide deletion present in a majority of O alleles.

In order to show the huge size of the table, the area surrounded by the rectangle ....

 ... was enlarged, and is shown below.

Using the Zoo blot of genomic DNA and hybridization with the human A transferase cDNA probe, we found the presence of homologous sequence in mouse.

And probably in chicken DNA, too.

We also examined genomic DNA from mammalian species.

The results showed the presence of homologous sequences in genomic DNA of all the species examined.

Dog, cat, and rabbit exhibited surprisingly strong signals of hybridization.

We determined the partial nucleotide sequences of the ABO genes from primates.

All the A antigen-expressing animals possessed the conserved leucine and glycine residues at the last 2 of the 4 amino acid substitutions that discriminate human A and B transferases.

On the contrary, the corresponding amino acid residues from the B antigen-expressing animals were methionine and alanine as was the human B transferase.

Based on the nucleotide sequences, a phylogenetic tree of the ABO gene was constructed.

B alleles are shown red in the figure.

Apparently, A to B transition seems to have occurred at three different occasions in this lineage.

We also determined the ABO gene of mice. Mice express A antigens in the gastrointestinal tract in vivo. 

We found that mice produce an enzyme which can transfer both GalNAc and galactose to synthesize A and B antigens in the in vitro experiments.

We elucidated the molecular genetic basis of the pig AO system. The pig O allele seems to lack most of the coding sequence.

Rats have both A and B genes, but they are not alleles.

Genome DNA sequences have been determined of a variety of species. Many species shown in the figure were demonstrated to possess the ABO gene equivalents.

The phylogenetic tree of the ABO gene was constructed, using the sequence information that was available from the public databases.

The deduced amino acid sequences of the ABO genes surrounding the codons 266 and 268 of the human A and B transferases and corresponding sequences from the other species are shown.

The positions of the codons 266 and 268 of the human A and B transferases are marked with asterisks.

In order to assess the effects of mutations identified in the ABO alleles, we employed the DNA transfection assays.

Firstly, A and B transferase cDNA expression constructs were prepared.

The constructs were transfected to the human cancer cells of uterus, HeLa cells.

The HeLa cells express H antigens on cell surface.

If the constructs encode functional proteins with A and/or B transferase activity, A and/or B antigens will be synthesized.

Those newly synthesized A and/or B antigens will be immunologically detected using the antibodies against A and/or B antigens and the plant lectins that are specific to the terminal sugars.

The results of DNA transfection experiments are shown.

The transfection of the A and B transferase constructs resulted in the appearance of cells expressing A and B antigens, respectively.

When A2 allele or A3 allele-specific mutations were introduced into the original A transferase expression construct, the A antigen expression decreased.

When the O allele-specific single nucleotide deletion or the O allele-specific amino acid substitutions were introduced into the original A transferase expression construct, the A transferase activity was lost.

We constructed 14 A-B transferase chimeras that possess A allele or B allele-specific amino acid residues at the locations of the four amino acid substitutions that discriminate human A and B transferases.

The constructs containing leucine and glycine at the 3rd and 4th positions allowed the expression of only A antigens.

When the constructs contained methionine and alanine at the 3rd and 4th locations allowed the expression of only B antigens.

When the constructs contain leucine and alanine at the 3rd and 4th positions, the specificity depended on the second position. 

When it is glycine, the constructs showed only A transferase activity.

On the other hand, when the 2nd position is serine, in addition to A transferase activity, the constructs expressed weak B transferase activity also.

When the constructs contain methionine and glycine at the 3rd and 4th positions, they directed the expression of both A and B antigens.

We also constructed the in vitro mutagenesis constructs that contained any one of 20 amino acid residues at codon 268 of the A transferase. 

The activity and specificity of the constructs were determined by DNA transfection.

The original A transferase possesses glycine residue at codon 268. Only A transferase activity was detected with the construct.

When the glycine was replaced by alanine, weak B transferase appeared, in addition to the strong A transferase activity.

When the glycine was replaced by cysteine or serine, which possess relatively small side groups, the constructs exhibited B transferase activity although A activity diminished.

When the glycine was replaced by asparagine or threonine, the A activity was lost while weak B activity was still present.

The differential activity and specificity by the amino acid substitutions may be explained by assuming the cavity around the codon 268. 

When codon 268 is glycine, GalNAc fits well to the cavity.  Galactose may not be held tight because the cavity is too big to fit in.

The construct with alanine at codon 268 may allow galactose, in addition to GalNAc, to fit in the cavity because alanine is a little bigger than glycine.

The cysteine and serine constructs may also allow the use of both galNAc and galactose although the efficiency drops with GalNAc.

The asparagine and threonine constructs may still allow the use of galactose although it is very inefficient. 

Those constructs may not transfer GalNAc because the cavity became too small.

Similar experiments were performed using the in vitro mutagenized B transferase constructs at codon 268.

The original B transferase has alanine at codon 268.

When the alanine was replaced by glycine, the protein exhibited not only B transferase activity, but also A transferase activity.

The cysteine and serine substitution constructs exhibited strong B transferase activity.

The B transferase activity was a bit decreased with the leucine, threonine, and valine substitution constructs.

The comparison between the asparagine and glutamine constructs (and the aspartic acid and glutamic acid constructs) suggests the importance of the size of the side group of codon 268. 

The comparison between the asparagine and aspartic acid constructs (and glutamine and glutamic acid constructs) suggests the importance of the charge of the side group.

The differences in the specificity and activity of the codon 268 amino acid substitution constructs of the human B transferase can also be explained by assuming a cavity around codon 268. 

With the original alanine, galactose fits well, but GalNAc does not.

The glycine construct has wider space that may accommodate either GalNAc or galactose.

The cysteine and serine constructs still possess strong B transferase activity.

The side groups of asparagine, asparatic acid, glutamine, and glutamic acid are compared.

The three dimensional structures of A and B transferases were determined in 2002.

The amino acid residue at codon 176, the first of the four amino acid substitutions between A and B transferases, is found far from the catalytic domain.

The amino acid residue at codon 235, the second of the four amino acid substitutions between A and B transferases, is found close to the catalytic domain.

And the amino acid residue at codon 266 is located at the center of the catalytic domain.

The amino acid residue at codon 268 is also situated in the center of the catalytic domain.

There is a CpG island around the initiation site(s) of transcription of the ABO gene. The methylation status differs among the cell lines.

There is a reverse correlation between the DNA methylation of the CpG island and the ABO gene expression. 

When the promoter region is methylated, the gene transcription is suppressed.

There are several different levels of control for the A/B antigen expression.

The transcription of the ABO gene is regulated by cis and trans elements.

There are more than a hundred ABO alleles that have been documented to date.

The nucleotide sequence information of those ABO alleles are deposited in the ABO system portion of the Blood Group Antigen Gene Mutation Database.

The A alleles include A1, A2, A3, Ael, Aint, Am, Aw, and Ax alleles.

There are several cis-AB and B(A) alleles deposited in the database.

The B alleles include B (B1), B3, Bel, Bw, and Bx alleles.

There are many O alleles.

The differences in the sequence may be divided into several different categories. These include missense mutations.

Nonsense mutation due to nucleotide substitution has not been identified.

However, frameshift mutations due to the insertion or deletion have been identified.

The mutations that affect splicing have also been identified.

One mutation was identified to change the initiation codon.

Mutations in the promoter region have been identified. Whether the mutations affect the transcription or not has yet to be determined for most of them.

Combinations of mutations were found in several alleles, suggesting the recombination in the past.

A mutation resulting in the mislocalization (not to the Golgi apparatus but to the cytoplasm) of the protein has been identified.

In addition to A and B transferases, there are additional enzymes with similar specificities.

One of them is the enzyme to synthesis the alpha 1-3 Gal epitope.

This enzyme, alpha 1,3-galactosyltransferase, is encoded by the GGTA1 gene. The gene is nonfunctional pseudogene in Homo sapiens. 

Humans cannot produce the alpha 1-3 Gal epitope, but contains the antibody against the epitope in the serum.

Another enzyme catalyzes the synthesis of iGb3 Cer.

This enzyme, isogloboside 3 synthase, is encoded by A3GALT2 gene.

Another example is the enzyme to synthesize the Forssman antigen.

The Forssman synthetase encoded by GBGT1 gene is responsible for the synthesis.

The phylogenetic tree of the ABO and other alpha 1,3 galactose (GalNAc) transferase genes was constructed, using the sequence information from the databases. 

It was demonstrated that the genes encoding these enzymes are evolutionarily related.

In addition to the ABO, GGTA1, A3GALT2, GBGT1 families of genes, another group of genes were identified to possess sequence homology, and they were named as GLT6D1 genes.

The amino acid sequence surrounding codons 266 and 268 of the human ABO gene and corresponding sequences from the related genes from a variety of species were aligned and are shown in the figure.

The ABO gene portion was enlarged, and is shown with the asterisks at codon s 266 and 268 of the human ABO gene.

The GLT6D1 gene portion was enlarged. No glycosyltransferase activity has been confirmed of this family of genes.

The histidine and alanine residues are well conserved in the GGTA1 genes among different species.

The histidine and alanine residues are also conserved in the A3GALT2 genes.

The GBGT1 genes encode the Forssman synthetases to transfer GalNAc. 

The corresponding amino acid residues are glycine and alanine, rather than histidine and alanine in the GGTA1 and A3GALT2 genes that encode enzymes with galactose specificity.

In addition to the alpha 1, 3 GalNAc (galactose) transferases such as A and B transferases, there are more than a hundred genes encoding glycosyltransferases with different specificities.

Glycosyltransferases are categorized, depending on the sugar the enzymes transfer.

In this table, B transferase is indicated by an arrow.

We developed the Systematic Multiplex RT-PCR (SM RT-PCR) method to analyze the expression of multiple genes.

Using the method, we examined the gene expression of dozens of glycosyltransferases.

These included sialyltransferases, fucosyltransferases, galactosyltransferases, GalNAc transferases, glucosyltransferases, and GlcNAc transferases.

We examined the gene expression in a variety of human tissues.

The tissue expression of the A/B transferases is indicated by an arrow.

Based on the SM RT-PCR results, the tissues were sorted by the glycosyltransferase gene expression, using the hierarchical clustering method.

The A/B transferase genes are indicated by an asterisk.

Anatomically related tissues of colon and rectum exhibited similar expression patterns and were clustered.

Penis and vulva were also clustered because of similar glycosyltransferase gene expression.

Brain cerebrum and cerebellum also showed similarity in glycosyltransferase expression.

In summary, we cloned the human ABO genes and elucidated the molecular genetic basis of the blood group ABO system. 

We also demonstrated the Central Dogma of ABO.

We also characterized mutations in the human ABO alleles.

We showed that the ABO genes are present in a variety of organisms, and studied the evolution of the ABO genes.

We studied the structural basis of the human A and B transferases and characterized mutations that affect the activity and specificity of the enzymes.

We cloned the genomic DNA of the human ABO gene and defined the promoter region.

We determined the expression of several dozens of glycosyltransferase genes in various human tissues.

Functionality of the ABO and related genes remains to be elucidated.

This lists the collaborators of my research on ABO.