BSCI 1511L Statistics Manual: The molecular basis of Mendelian alleles

In the BSCI 111a course we examined the structure of the ABO glycosyltransferase gene and the nature of the mutations that caused several allelic variants at the molecular level.  In particular we focused on the single nucleotide polymorphism (SNP) identified as OMIM "Blood group O" locus 110300.0001.  This SNP involves the deletion of a guanine at position 258 in the protein coding region of the gene.  When the gene contains the allele without the deletion, the protein product is a functional enzyme.  However, when the gene contains the allele with the deletion, a frameshift mutation occurs which results in a premature stop codon.  If this gene is translated, the protein product is greatly truncated in comparison to the protein that results if the deletion is not present.  The truncated protein is NOT a functional enzyme. 

Fig.7.  Action of variant glycosyltransferase enzymes in attaching carbohydrates to red blood cell (RBC) transmembrane glycoproteins.

Next week we will begin our investigation of Mendelian genetics.  In classic Mendelian genetics, an allele is a sort of "black box" that is identified by a letter such as "T" or "t".  Combinations of alleles produce particular visible phenotypes, but the mechanism by which this occurs is not specified.  Many students have previously studied the ABO gene as a complex three locus Mendelian system that results in the blood types A, B, O, and AB.  This week provides an opportunity to examine how the molecular details that we studied in the fall semester result in a Mendelian system such as those we will utilize extensively in the coming weeks. 

Fig. 7 shows how the functional glycosyltransferase enzymes attach monosaccharide epitopes (N-acetylgalactosamine or galactose) to the polysaccharides extending from the surface of the red blood cell membranes.  The protein formed from translation of the gene having the position 258 deletion is not capable of attaching any monosaccharide and the polysaccharide remains unmodified. 

Whether a particular glycosyltransferase enzyme transfers N-acetylgalactosamine (creating the "A" antigen) or galactose (creating the "B" antigen) depends on the state of seven SNPs found within the gene and identified collectively as the OMIM "Blood group A/B polymorphism" locus 110300.0002 .  When one set of basepairs is present at the SNPs, the enzyme coded for is alpha 1-3-N-acetylgalactosaminyltransferase (a.k.a. transferase A).  When a different set of basepairs is substituted at the seven SNPs, the enzyme coded for is alpha 1-3-galactosyltransferase (a.k.a. transferase B).  However, the set of basepairs present at the seven SNPs becomes irrelevant if the 258 guanine deletion is present because the portion of the gene containing the seven SNPs is downstream of the premature stop codon formed by the frameshift and never gets transcribed.  This information is summarized in Table 1.

Table 1.  Relationship between SNPs on the molecular level and Mendelian ABO alleles

At the molecular level, there are actually eight SNP loci involved: the seven basepair substitutions involved in the Blood group A/B polymorphism and the single G deletion involved in the Blood group O polymorphism.  But because the seven substitution SNPs are always linked and can only have an effect when there is no deletion, they have a single effect (to produce either transferase A or transferase B) and hence in Mendelian "black box" terms they are identified as a single allele (A or B).  Since the G deletion always has the same effect (production of a non-functioning enzyme) regardless of the state of the seven substitution SNPs, it is considered a third allele (called "O").  So in Mendelian terms, the eight molecular loci boil down to a single locus ("the ABO locus") having three alleles (Table 1). 

The connection between these three Mendelian alleles and the condition of RBC surface antigens (Fig. 5) would be relatively straightforward, except for the complicating fact that humans are diploid.  The presence of two homologous chromosomes 9 means that the "ABO locus" need not have the same Mendelian allele for both copies of the ABO gene.  If one chromosome has the A allele and the other has the B allele, then both transferase A and transferase B enzymes would result from translation of the two versions (Fig.8, bottom), resulting in type AB surface antigens (bottom of Fig.5).   If both homologous chromosomes 9 have the O allele, then all of the resulting proteins would be non-functional (Fig. 8, top) and none of the polysaccharides would have added glycosyl groups (Fig. 5, top).  If the A allele is present on one chromosome, it doesn't matter whether the other chromosome also contains the A allele or if it contains an O allele.  Since the DEL 258G protein coded for by the O allele is non-functional, processing of the polysaccharide is done exclusively by the functional transferase A enzyme, so the effect is the same as if both chromosomes contained A alleles: all polysaccharides will be capped with N-acetylgalactosamine to form the A antigen (Fig. 5, type A).  The situation is similar with the B allele. 

Fig 8. Types of protein products formed by possible combinations of Mendelian genotypes

In classical Mendelian genetics, the mechanism by which an allele exerts its influence on the phenotype of the organism is unknown.  It is a given that a diploid organism has two alleles at a particular locus.  If both of those alleles are the same (i.e. the individual is a homozygote), then the effect of the alleles is clear: the organism will have the phenotype associated with that allele.  For example, in Mendel's classic pea example, if a pea plant has two R (round) alleles (genotype: RR) it will have round seeds and if a pea plant has two r (wrinkled) alleles (genotype: rr) it will have wrinkled seeds.  Mendel did not know what the R and r allele were, nor did he know how they caused seeds to be round or wrinkled.  The issue of dominance depends on what happens if an organism has one allele of each kind (i.e. is a heterozygote with genotype: Rr).  Clearly, a pea cannot be both round and wrinkled at the same time.  It can be determined empirically that a heterozygote has round seeds, so we say that the R allele is dominant since it's phenotypic effect "wins" over the phenotypic effect of the r allele when an organism has one of each. 

However, we are in the 21^st century, not the 19^th century, and we are not stuck with the R and r alleles being "black boxes" having an unknown function.  We have the capability of determining on a molecular level how the alleles cause the seeds to be round or wrinkled.  In round peas, it was observed that starch was present in larger amounts and the starch grains were large and simple.  In wrinkled peas, less starch was present and the grains were small and deeply fissured.  The wrinkled peas had higher levels of free sucrose which caused them to have higher osmotic pressure when fresh and to subsequently lose more water when dry (and therefore become wrinkled).  Thus it became clear that the underlying cause of the wrinkles was related to starch metabolism.  Subsequent research showed that loss of activity of starch-branching enzyme prevented amylase from being converted to amylopectin resulting in a reduction in starch synthesis and the observed differences of wrinkled peas described above.  In the wrinkled allele, the gene coding for the starch-branching enzyme has an 800 bp insertion which causes the loss of the last 61 amino acids in the protein, resulting in reduced activity of the enzyme (Bhattacharyya et al. 1990). 

So at the level of Mendelian genetics, based on the visible phenotype the locus is called the "wrinkled seed" locus and the two alleles are called R (round) and r (wrinkled).  However, at the molecular level the actual locus for the trait is the SBEI (starch branching enzyme) gene with the two alleles being the non-insertion (normal) coding sequence and the 800 bp insertion (non-functional) sequence.  The connection between the molecular alleles and the Mendelian alleles (non-insertion → R, and 800 bp insertion → r) is indirect but is typical for dominant/recessive traits.  In many such traits, a mutation which causes loss of function of an enzyme which controls some aspect of a metabolic or developmental pathway causes a cascade of molecular events which ultimately results in one or more visible phenotypic effects.  It is actually a bit of an oversimplification to just say that the r allele (i.e. the 800 bp insertion) causes wrinkled seeds because the r allele also causes higher sucrose levels, reduced starch grain size, and any of the other measurable effects of decreased SBEI activity. 

Returning to the question of dominance, it should now be clear why the R allele is dominant over the r allele.  The normal (unmutated) allele of the SBEI gene (i.e. the R allele) codes for an enzyme that is functional.  As long as either homologous chromosome contains that allele, some functional enzyme will be present in the peas.  It doesn't really matter if all of the enzymes are functional (RR genotype) or of half of them don't work well (Rr genotype).  As long as there are some that work, that is enough for the pea to be round.  However, if both homologous chromosomes contain alleles of the SBEI gene having the 800 bp insertion (i.e. the r allele), then only the non-functional enzyme will be present in the peas, the peas will have messed up starch metabolism, and be wrinkled.

So a mutation that causes an enzyme to be non-functional or to have reduced functionality is a common underlying cause for recessive alleles.  If we return to the case of the ABO blood groups, we see that is the case with the Blood group O locus.  The Mendelian "Type O blood" allele has as its underlying cause the deletion, frameshift mutation, and premature stop codon that causes the glycosyltransferase enzyme to be non-functional.  We call the "Type O blood" allele recessive because as long as it is not the only kind of allele present on the homologous chromsosomes 9, the glycosyltransferase enzyme coded for on the other chrosome can get the job of modifying the RBC surface antigens done. 

In order for complete dominance to occur, having half the number of functional enzymes (in the case of a heterozygote) must be just as good as having all functional enzymes (i.e. homozygous dominant).  Often this is the case, particularly if the enzyme having the mutation performs a controlling function in some early part of a pathway.  Just turning on the pathway is enough.  However, if the enzyme itself must be present in sufficient quantities to produce the visible phenotype, then reduced amounts of enzyme may result in a heterozygote phenotype that is intermediate between the phenotypes of the two homozygotes.  This is illustrated by case of the mutant a^f allele of the CHS-D gene in Ipomoea purpurea (common morning glory) flowers.  The mutant allele fails to be transcribed with the result that heterozygotes have a reduced level of CHS-D mRNA and CHS-D enzyme.  The reduced level of enzyme limits the rate of anthocyanin (red) pigment biosynthesis and causes the flowers to be pink rather than red (Johzuka-Hisatomi et al. 2011). 

In some cases, the gross phenotype of the heterozygote appears the same as the homozygous dominant but differences in the level of enzymes or their products can be detected chemically. Such differences may allow for tests to detect heterozygous carriers of recessive genetic diseases. 

In the ABO gene example, when the SNP 258 deletion does not occur, the glycosyltransferase enzyme will be functional and can act on either N-acetylgalactosamine or galactose depending on the state of the seven basepair substitutions that form the "Blood group A/B polymorphism" locus 110300.0002 .  Neither form of the glycosyltransferase enzyme has an effect on the other, and an individual's red blood cell surface antigens will be modified by all of the forms that are present in the organism.  An AA homozygote will use only N-acetylgalactosamine to produce A antigens, a BB homozygote will use only galactose to produce B antigens, and an AB heterozygote will use both substrates to make some of each of the two antigens.  Thus we say that the Mendelian A and B alleles are codominant, since their phenotype (the presence of A antigens and the presence of B antigens respectively) will be exhibited any time that the allele is present. 

In the simplified Mendelian view of the ABO blood group system, there is a single locus with three alleles, with O recessive to A, O recessive to B, and A codominant with B.  The preceding section explains the molecular basis for why the O allele is recessive and why the A and B alleles are codominant.  Although this explanation is complex, it makes it clear that all of the Mendelian genetics that we will study for the rest of the semester is underlain by the molecular genetics that we studied the first semester.  Despite the high degree of simplification present in Mendelian genetics, it is remarkable that it provided a powerful predictive tool for over a hundred years before the underlying mechanism was fully understood.  It remains a useful means to describe complex genetic phenomena at a high level.