Philip Dei
Thu, Jul-18-02, 01:17
What is HLA stands for Human Leucocyte Antigen, it is
the human equivilant of the Mouse Major
Histocompatibility Complex.
HLA is composed of Major and minor genes
The major genes Fall into Class I -A -B -C -e -f -g
Class II (D) -DR -DQ -DP -DM
Class III (compliment proteins and factors) -C4 -C2 -Bf -TNF
(tumor necrosis factor)
Highly Oversimplified Biology, class I generally presents
infected cells proteins to cell for the purpose of both immune
activation but primarily as a suicide bomb, the trigger is the
HLA, but the explosive is the C4, C2 and Bf.
Class II are on surface of immune cells (so called
'professional' APC) which present nasty antigens to immune
system for the purpose of eliciting or suppressing a
T-cell response, which in turn expands B-cells which
recognize an antigen in a certain context, and cause
antibodies to be produced.
By now each of you should have a copy of Parham and Ohta, 1996
Science Vol 272, 5 April. Page 67 shows map.
I am going to reverse the order. The region is about 4 million
base pairs in length, a few centimorgans to the true
geneticist out there, The intergenic space is large.
I am reversing the order and approximating position, I will
start with A = 0. Position of genes relative to A
A - 0 C - 1.2 mbp B - 1.35 TNF- 1.6
C4/2,Bf- 2.0 DR - 2.4 DQ - 2.7 TAP/LMP- 2.9 DM - 3.0 DP - 3.2
We are not going to look at some of these so I will shorten it
A - 0 C - 1.2 mbp B - 1.35
C5/2,Bf- 2.0 DR - 2.4 DQ - 2.7 DP - 3.2
DR is composed of an alpha and beta. -alpha is functionally
invariable, despite a probable heterozygous selection
coefficient, it was fixed in the last constriction, no
variation remained, and only a single mutation has appeared
-Beta is encoded primarily by the Beta1 gene (DRB1) See ayala,
science december 1995 'myth about eve'. There are now over 200
alleles for this loci, it has been described as the fastest
evolving gene in humans. For good reason, with the DQ and DP
variation is composed from alpha and beta; with beta generally
predominating the evolution, we tend to see combinations of
100s for alpha/beta, since alpha was fixed in DR this isnt
possible for DR1 so the beta chains are playing catchup. -Beta
chains are also encoded by DRB3, B4 (psuedogene), B5 and a
growin list of psuedogenes.
Therefore you can have a protein DRBA0101-DRB10405 or in
the same human DRBA0101-DRB30101 and if there are 2
sister chromosomes two more for each allele. Not all
humans have DRB3.
DQ is composed of -A variable alpha chain -A variable beta
chain Q alpha is closely linked to Q beta and recombination of
alpha and beta chains is rare.
DP is composes of -A variable alpha chain -A variable beta
chain P alpha is closely linked to P beta and recombination of
alpha and beta chains is rare.
P alpha was almost fixed in the last constriction, variation
is slowly rebounding.
How do HLA genes evolve. HLA class II genes can recombine
alpha and beta, but since they are closely linked this takes
forever. In addition HLA A, B and C are composed of single
chains and an invariable B2microglobulin.
In order to increase the rate of evolution nature has cleverly
used 'gene conversion' to accelerate this process. HLA genes
are the epitome of gene conversion, intragenic recombination
are rare, but gene conversion is common, In unpublished
analysis of HLA DRB1 in 1996 I determined that all of the 160
or so alleles can be explained via recombination of 14
starting alleles in the human population, except a single
allele which had an obvious point mutation. Not that point
mutations in these alleles have not occurred, but gene
conversion is so pervasive it is difficult to distinquish
which alleles are convertants and which are the result of
novel point mutations (sound familiar). Another interesting
character of HLA D is that new allelic variants are typically
mutations of hypervariable segment, but even more important,
variation of the critical contact residues in the primary
structure of the protein. This differs from HLA class I, but
the function of variation in class I has not been critically
determined. For more information of gene/function relationship
of class II I can give references privately.
Serotyping. A long time before PCR was even a concept people
where doing their dog best to cypher these alleles within the
HLA, they used antibodies and they still do.
When one uses antibodies one get an typing of (serotyping)
A1 or B8 or Cw7 (eliminate the w in your mind). Which is OK
for the class I but if we take a serotyped individual who is
C6/BBL and sequence them we get a genotype.
C6/A1A (afor null) Some of the antiboides are pretty good
and can distinguish a 0802 from a 0801 in the
right hands.
Class II is a different problem. Antibodies may detect an
alpha/beta combination which is more or less senstive to beta
chain and oblivious to alpha, or it may detect both. Since
alpha and beta are closely linked alpha is seldomly displaced
by a different alpha. Now the human population is what, 6
billion and 10,000 generations out of the last constriction,
let me guess, how seldom does seldom have to be not to create
every imaginable combination of alpha beta.
If you see DR1 it means DRB1-01xx DRA1-0101(or 02) you don't
care about alpha.
If you see DQ1 it means DQB1-0601 DRA1-0101 and don't
ask me why.
So you will noticed in the dataset the HLA A B C are serotyped
but as we migrate in the DR, DQ, DP they are genotyped,
because frankly serotyping of Q and P aint worth c-r-a-p. In
the study of autoimmune diseases as single amino acid
difference that an antibody is oblivious to, can mean the
difference between normal and disease.
So you have to use your imagination a bit.
So let us now go through the haplotypes (a collection of
specific alleles passed as a group)
A24 Cw7 B52 DR15 DRB3BL DQ2 This is a serotype DR1502
DRB3(absent) DQA0101 DQ0201 DPA0101 B0401
A-24 B-52 C-07 DRA-0101 DRB1-1502 DRB3-Absent DRB5-Absent
DQA-0101 DQB-0201 DPA-0101 DPB-0401
This represents an haplotype of the chromosome of a certain
percentage of individuals. This is alot of work for one
person. Really. Because they also have another chromosome.
Imagine serotyping a person who has the genetic haplotype
QA0201/QB0601 on one chromosome and QA0102/QB0502 That
individual can produce
QA0201:QB0601, QA0201:QB0502, QA0102:QB0601, QA0102:QB0502 Two
genes, four proteins, you think even a good serologist will
correctly identify the components. Thats why genotyping is a
must for HLA class II. Previously we could not find linkages
to genetic diseases even though the models suggested they were
their. The state of the haplotyping today can allow the
specific association of a DXalpha, with a DXbeta without any
guess work.
Each allele cost about 300$ to type a single individual is
about $2000 imagine how much effort was involved in typing
100s of individuals in each of 20 to 30 populations.
Post per request, hope this clears up some of the confusion.
Parham and Ohta is a must. Read it and light bulbs should
start going crazy. Philip <pdeitik at bcm.tmc.edu
the human equivilant of the Mouse Major
Histocompatibility Complex.
HLA is composed of Major and minor genes
The major genes Fall into Class I -A -B -C -e -f -g
Class II (D) -DR -DQ -DP -DM
Class III (compliment proteins and factors) -C4 -C2 -Bf -TNF
(tumor necrosis factor)
Highly Oversimplified Biology, class I generally presents
infected cells proteins to cell for the purpose of both immune
activation but primarily as a suicide bomb, the trigger is the
HLA, but the explosive is the C4, C2 and Bf.
Class II are on surface of immune cells (so called
'professional' APC) which present nasty antigens to immune
system for the purpose of eliciting or suppressing a
T-cell response, which in turn expands B-cells which
recognize an antigen in a certain context, and cause
antibodies to be produced.
By now each of you should have a copy of Parham and Ohta, 1996
Science Vol 272, 5 April. Page 67 shows map.
I am going to reverse the order. The region is about 4 million
base pairs in length, a few centimorgans to the true
geneticist out there, The intergenic space is large.
I am reversing the order and approximating position, I will
start with A = 0. Position of genes relative to A
A - 0 C - 1.2 mbp B - 1.35 TNF- 1.6
C4/2,Bf- 2.0 DR - 2.4 DQ - 2.7 TAP/LMP- 2.9 DM - 3.0 DP - 3.2
We are not going to look at some of these so I will shorten it
A - 0 C - 1.2 mbp B - 1.35
C5/2,Bf- 2.0 DR - 2.4 DQ - 2.7 DP - 3.2
DR is composed of an alpha and beta. -alpha is functionally
invariable, despite a probable heterozygous selection
coefficient, it was fixed in the last constriction, no
variation remained, and only a single mutation has appeared
-Beta is encoded primarily by the Beta1 gene (DRB1) See ayala,
science december 1995 'myth about eve'. There are now over 200
alleles for this loci, it has been described as the fastest
evolving gene in humans. For good reason, with the DQ and DP
variation is composed from alpha and beta; with beta generally
predominating the evolution, we tend to see combinations of
100s for alpha/beta, since alpha was fixed in DR this isnt
possible for DR1 so the beta chains are playing catchup. -Beta
chains are also encoded by DRB3, B4 (psuedogene), B5 and a
growin list of psuedogenes.
Therefore you can have a protein DRBA0101-DRB10405 or in
the same human DRBA0101-DRB30101 and if there are 2
sister chromosomes two more for each allele. Not all
humans have DRB3.
DQ is composed of -A variable alpha chain -A variable beta
chain Q alpha is closely linked to Q beta and recombination of
alpha and beta chains is rare.
DP is composes of -A variable alpha chain -A variable beta
chain P alpha is closely linked to P beta and recombination of
alpha and beta chains is rare.
P alpha was almost fixed in the last constriction, variation
is slowly rebounding.
How do HLA genes evolve. HLA class II genes can recombine
alpha and beta, but since they are closely linked this takes
forever. In addition HLA A, B and C are composed of single
chains and an invariable B2microglobulin.
In order to increase the rate of evolution nature has cleverly
used 'gene conversion' to accelerate this process. HLA genes
are the epitome of gene conversion, intragenic recombination
are rare, but gene conversion is common, In unpublished
analysis of HLA DRB1 in 1996 I determined that all of the 160
or so alleles can be explained via recombination of 14
starting alleles in the human population, except a single
allele which had an obvious point mutation. Not that point
mutations in these alleles have not occurred, but gene
conversion is so pervasive it is difficult to distinquish
which alleles are convertants and which are the result of
novel point mutations (sound familiar). Another interesting
character of HLA D is that new allelic variants are typically
mutations of hypervariable segment, but even more important,
variation of the critical contact residues in the primary
structure of the protein. This differs from HLA class I, but
the function of variation in class I has not been critically
determined. For more information of gene/function relationship
of class II I can give references privately.
Serotyping. A long time before PCR was even a concept people
where doing their dog best to cypher these alleles within the
HLA, they used antibodies and they still do.
When one uses antibodies one get an typing of (serotyping)
A1 or B8 or Cw7 (eliminate the w in your mind). Which is OK
for the class I but if we take a serotyped individual who is
C6/BBL and sequence them we get a genotype.
C6/A1A (afor null) Some of the antiboides are pretty good
and can distinguish a 0802 from a 0801 in the
right hands.
Class II is a different problem. Antibodies may detect an
alpha/beta combination which is more or less senstive to beta
chain and oblivious to alpha, or it may detect both. Since
alpha and beta are closely linked alpha is seldomly displaced
by a different alpha. Now the human population is what, 6
billion and 10,000 generations out of the last constriction,
let me guess, how seldom does seldom have to be not to create
every imaginable combination of alpha beta.
If you see DR1 it means DRB1-01xx DRA1-0101(or 02) you don't
care about alpha.
If you see DQ1 it means DQB1-0601 DRA1-0101 and don't
ask me why.
So you will noticed in the dataset the HLA A B C are serotyped
but as we migrate in the DR, DQ, DP they are genotyped,
because frankly serotyping of Q and P aint worth c-r-a-p. In
the study of autoimmune diseases as single amino acid
difference that an antibody is oblivious to, can mean the
difference between normal and disease.
So you have to use your imagination a bit.
So let us now go through the haplotypes (a collection of
specific alleles passed as a group)
A24 Cw7 B52 DR15 DRB3BL DQ2 This is a serotype DR1502
DRB3(absent) DQA0101 DQ0201 DPA0101 B0401
A-24 B-52 C-07 DRA-0101 DRB1-1502 DRB3-Absent DRB5-Absent
DQA-0101 DQB-0201 DPA-0101 DPB-0401
This represents an haplotype of the chromosome of a certain
percentage of individuals. This is alot of work for one
person. Really. Because they also have another chromosome.
Imagine serotyping a person who has the genetic haplotype
QA0201/QB0601 on one chromosome and QA0102/QB0502 That
individual can produce
QA0201:QB0601, QA0201:QB0502, QA0102:QB0601, QA0102:QB0502 Two
genes, four proteins, you think even a good serologist will
correctly identify the components. Thats why genotyping is a
must for HLA class II. Previously we could not find linkages
to genetic diseases even though the models suggested they were
their. The state of the haplotyping today can allow the
specific association of a DXalpha, with a DXbeta without any
guess work.
Each allele cost about 300$ to type a single individual is
about $2000 imagine how much effort was involved in typing
100s of individuals in each of 20 to 30 populations.
Post per request, hope this clears up some of the confusion.
Parham and Ohta is a must. Read it and light bulbs should
start going crazy. Philip <pdeitik at bcm.tmc.edu