Simplifying genetic locus assignment of HLA-DRB genes

viewpoint

Simplifying genetic locus assignment of HLA-DRB genes G6ran Andersson, Leif Andersson, Dan Larhammar, Lars Rask and Sunna Sigurdard6ttir The DR haph}td'pes {}f the human major histocompatibility complex have been arranged in fi:'e haph}typic groups based on genomic cloning and sequence ,malvses. To date, the expressed DRB sequences have been assig~:ed to four different loci: DRBI, 3, 4 and 5. DRBI alleles are present in all hapl,}td,pes, whereas DRBd, 4 and 5 are present only m some l..at,h~t)'pes. Here, G6r,m Ande, s~,on and coUeagues suggest that DRBd, 4 ,rod S sequences may be treated as a single allelic series. They argu, that such a model is appropriate, since DRBd, 4 ,rod 5 sequences are inherited m an .dh,lic ?~lshion, have simiDr genomic localization;, exhibit similar h'vcls of gene expression and :;re, with a feu, rare exceptions, not present in the same haplotype. Human HLA-I)RB genes of the major histocompatibility complex (MHC) exhibit both allelic and haplotypic po!ynmrphism. The former is characterized bv multiple amino acid replacements among alleles, and by the presence of many alleles at the same locus. Moreover, no allele dominates in the population. The haplotype polymorphism is mamfested by the existence of variable numbers of expressed as well as nonexplcssed DRB genes. Since DRB gene expression differs quantitatively, Lb.is feature may qualitatively affect the ,~utcome of a normal immune response. An under~:tanding of the complexity within the DRB gene system is therefore of central immunological importance. Moreover, the ass{}ciation of certain autoimmune diseases with specific DR types is likely to depend on the expression of different DRB genes. Accordin!, :,~ the curren: nomenclature, HI A I)RB gent,,; have been assigned to nine different loci c{mlprising four expressed loci IDRBI, 3, 4 and 51 and fi-'c pseudogcnes {I)P,B2 and 6-9i {Rc;. 1). Thi, nomenclature appears ~ore complicated than available data requires. Theref-re, in this paper, a simplification ir proposed, whereb~ genes are assigned to only t~o different expressed DRB loci. This simplified nomenclature should be more informative and make the important literature on DRB diversit,~ and function more co}reprehensible to non-HLA specialists.

expresses two DRB genes and the products of each can heterodimerize with the DRc~ chain. The two expressed DRB genes present on one chromosome are transcribed at different levels {see below). The genomic organization of DRB genes from three different haplotypes has been characterized by molecular cloning 4 s. From these results, as well as eDNAand polymerase chai,i reacnon (PCR)-derived nucleotlde sequence information, the DRB genes have been classified into nine different loci (see above and Ref. 1). Five main groups of DR haplotypes have been identified on the basis of the the polymorphic serological specificity expressed by the DRBI locus (see Ref. 1 and Fig. 1 ). The five major DR haplotypic groups are: I1} the I)R1 group, expressing the DR1, 10 and 103 serological specificities; {2} the DR51 group, expressing the DRI5 and 16 serological specificities; (3) the DR52 group expressing the DR3, 1i, 12, 13, 14, 14{)3 and 1404 serological specificities; (4} th,~ DR8 group, expressing the DR8 specificity; and (5) the DR53 group, expressing the DR4, 7 and 9 serological specificities ~. Despite the varying number of DRB genes within each haplotypic group, some similarities are evident in their organization The DRBI locus, which encompasses the most highly expressed and polymorphic DRB genes~ is located at the centromeric end of the Genomic organization of DRB genes DR region, with the single DRA gene at its telomeric DR genes are enc,}ded ,}n the :~.h,.)rtarm of chromq~- end. The DRB9 IcJcus, which seems to be present in all some 6, in the HLA class II region of the human MHC haplotypes and which only constitutes a free DRB {Ref. 21. The number ot DRB genes varies from two to exon 2, is located centromeric to DRA. The second five between haplotypes ~. The DR region also contains and more weakly expressed DRB locus in the DR52 a single monomorphic DRA gene coding for the DRa and DR53 haplotypes {DRB3 and DRB4, respectively) chai:l. Translated DRA and DRB mRNA generate c~ is located centromeric to DRB9. The DRB5 gene in and j3 chains that form glycosylated a/J~ cell-surface DRSI haplotypes is also located centromeric to DRB9 heterodimers. In most individuals, each chromosome {for expression of this gene see discussion below). In

hnmun,dogy "l'fMay

58

VoL I T N,}. 2 1994

viewpoint DRB1

DR1 haplotype DR1, DR10, DR103, DR15*

DRS6

// DRB1

DR51 haplotype DR15, DR16, DR1*

DRB6

/

DRB5

DRB9

DRA

DRB9

DRA

DRB9

DRA

DRB9

DRA

DRB9

DRA

DR51

DRB1

DR52 haplotype DR3, DR11, DR12, DR13, DR14, DR1403, DR1404

DRB2

DRB3 DR52

DRB1

DR8 haplotype DR8

DRB1

DR53 hapiotype DR4, DR7, DR9

/

//

DRB7

DRB8

// DRB4 r)R53

Fig. 1. Schematic representation of the five main DR haplotypic groups. Organization of the DR i and I)t,~8 haplutypes are in/erred/rom ~he known genomic organization of the DRS1, DR52 and DR53 hapto~pes 4.~'. The haplotypic desig,:;.:ons are indicated to the left. DRA and DRB genes are shown as boxes. The expressed serological specificities are indicated be!ou, the haph)typie gnmp. Distances hrzween genes are ,,~ drawn to scale. Presur,~ed allelic genes are shown in identical shades. Brackets in the lines are introduced m ~,rder to verticallx align putatwe allelic genes. Nomenclature is according to Bodmer eta].. ~ denotes DRB pseudoge::es. * denotes rar,,Iv ohsen,ed ~emlogtcal spect/tcates.

ail haplotypic groups, except DR8, 9ne or two additional pseudogenes are located telomeric to DRBI. The DRB2 pseudogene lacks exon 2, the DRB6 pseudogene lacks exon 1, and the DRB8 pseudogene

(a)

apparently lacks both exon 1 and exon 2 (Refs 4, 9-14). Tl~e DRB7 pseudogene has no deletion, but does have several inactivating point mutatio~s ~~*. These highly divergent pseudogene sequences have not

o~u,o,o,

(b)

[ DRB4*0101 - - - ~ ' ~ DRB4+

DRB? [

DP,BIR}IO1 D~RBI~IOO1

DRB2

~IB7

DRBI*1001

[ DRBI"I~I

ORBI"I~I

[

DRBlq);q~ DRBI"0~O'I

~

D~B~0101

,4- ~ ' q ~

----t ~ D~I"ISI~ u DRBI"1801 DRBl"11~ DRBI'~I~I01 DRBI'~01 DRBI"0~

DR~ Bovi;'~

-_ DRBI~O702 ~ - DRBI"0101 J ~_ ' DRBPOIQ1 i -- DRB1"O401

DR~ 1

~ D ~

DRBP01Ol

~

- -

DRBI"0702

oR..olol

II

~

I

~

DRBI"1001

DRBI"0301

~'L--[---~

rlRB1"11~2

DRBI"DS~ I

J

l

l

~

I

d

~

~

~

I 0.1 ~ i t

Fig. 2. Phylogenetic t~ees of human DRB ~equences constructed using the neigbhour-jomi,lg metbod I~'. Fro" DRB mtclet~tide sequence references see B.dmer et al.~. Comparisons were made lbr: ~a) exam 2: (b) exon 3: and (c) exon 2 after tbe c,dons for polymorphic amino acids in the antigen-wc~gniti.,, site (AR.g) u,ere excluded. These include ARS codrms 9, 11, 13, 28, 30, 37, 38, 57, 67. 70, 71, 74, 78 and 86. + denotes sequences nut present in the recent nomenclature J. Scale bars indicate genetic distal,lee as defined by the freque;zcy of nucleotide substitutions per site where each vertical bar represents 0.01 arbitrary uni~= P:mel (a) uses the same scale bar as shown below panel (c).

D R ~

.~

~

RBlt'lff01 DRBl*18~ 1"1601

DRB~

~ ~

~

0..,

~.._ DRB3~201 [-|

r

[ DRB4*0101

__4

(c)

DRWP'11(I~

~'- DRB1"1301 - - DRBI*O~]2 r~ DR B3"0~21)1

n

l

l

:

DRB3 Bovlr~

l

l

O

.

l

~

Immunology Today

59

Vol. 15 No. 2 1994

viewpoint @

(a)

I

~8,

g "5_

G)

~ ~ ee 0 ~1~.0~ 0~"

~o

14

o

• 0 O• e,• °o o o ~, '~o~ ~' . % 0,,i~ o4o O.$ o •

~2

0 O 0 ee @~° 0

8

@0 0

,,)

8 o

o

©

PhyHgenetic analyses Currendy, more than 60 alle!es have been assigned to the DRB1 locusL The three weakly expressed DRB loci show less aiielic polymorphism, with only four DRB3, two DRB4 and four DRB5 alleles identified. It may be quemoned whether there are stronger arguments for assigning the most highly expressed DRB gene of each haplotype into a single locus, DRB1, than to organize the more weakly expressed DRB3, DRB4 and DRB5 genes into another single alle!ic series. To investigate this, phylogenetic analyses of a~ailablc exon sequences h~we been carried out using parsimony I~' and neighbo,w-joining programs (see

o

o

it

o

10

i5

25

2G

% Silent substitutions

g

8

o

•°

.

4 Q @

rr

0

0 0 0

2

O'

0

• ~*

•

I~

0

Q.

e@

0 0

@

0

10

o

2

4

6

• @

12

14

% Silent substitutions Fig. 3. Representatam of the relatwe t~equencws of silent versus replacement substitutio~ in DRB coding sequences. The relative frequencies are shown ilt Fainetse compansmts ot DRBI (OL DRB3 (÷1, DRB4 cx). attd DRB5 (~) sequences; the two DRB5 sequences were included as mma]lelic comp.;is,,ns. The frequencies u'ere estmlated according to the method o( Nei and (;IVoDorfl" applying ]ukes ,rod (?dnlor'.~2~' ~,,rre~ti,m fiw muhiph' bits. The sequences used are those given 'n Fig. 2 excluding t.t.e [)RRL, DRB7 and DRB~; geues. ~a! exlm 2; (b! exlm L r;.,:s figu;e is .::i,,[ :ed iron, Ret. 14.

been exposed to any ,election pressure and, consequently, have a~umuiated ;uhstantial sequence heterogeneity aP~cr :heir in:,_,:tivation. Therefore, they will not be grouped genetically for the purpose of this discussion. [hc DR8 haplot.vpe is unique since it contains a single DRB gene. This gene was probably generated by a gene-~ontractMn event that occurred in a primordial DR52-1ike haplotype, since the DR8 DRB gene is similar to DRB1 at the o'-end and similar to a DRB3 gene at the 3"-end ]:-r. The presence of such an ancestral t,,:T~!o-!-pc in the gorilla MHC supports this h t;,,ti-,cslsL~. Unfortunately, no genomic sequence is available for the DRB gene in the DR8 haplotype. Until then, for simplicity, this gene will be grouped within the DRB1 locus.

lmmunoh,gy Today

60

Figure 2 shows dcndrograms constructed using tb~ neighbour-jommg method .8. Comparisons were made separately for exons 2 and 3 (Fig. 2a,b). For exon 2, comparisons were also made after the codons for polymorphic amino acids in the antigen-recognition site tARS) had been omitted, to avoid the possibility that the selection acting on these codons may interfere with the phylogenetic analyses (Fig. 2c). However, the phylogenetic trees for exon 2 sequences, with or without ARS-codons, were very similar (Fig. 2a,c). As expected, DRB1 sequences from closely related DR haplotypes branch close to each other in ~hese analyses. Furthermore, DRB3, DRB4 and DRB5 sequences are on separate branches. The most striking finding is that presumed all°lie DRBI sequences from different haplotypic grou W do not branch together in any comparison, rather, they mingle with DRB3, 4 and 5 sequences. For instance, allelic DRB3 sequences can be considered as closely related alleles within a more comtqex locus that also includes DRB4 and DRB5, and represent sequences from closely related haplotypes :1, the same way that DRB1 sequences do (Fig. 11. Fhese analyses show that the generic differences among the DRB3, 4 and 5 sequences are not higher than those found within the DRB1 atlelic series. A similar conclusion was reached from calculations of the relative frequencies of silent and replacement substitutions in pairwise combinations of available expressed DRB sequences. This analysis was performed separately on exon 2 (encoding the polymorphic first domain) and exon 3 (encoding the second domain) (Fig. 3a,b). A striking observation derived from this analysis is that data from comparisons between presumed DRB1 all°lie seouences greatly overlap with data from comparisons of presumed nonaiieiic sequences. This holds true for both exons analysed (Fig. 3a,b). These data clearly show that expressed DRB genes do not exhibit distinct locus characteristics i:l their coding sequences. The genet;c analyses of DRB coding sequences reported here are in agreement with previous restrictirol fragment length polymorphism (RFLP) analyses23 indicating that DRB genes show a high degree of similarity w;thin haplotypic groups. These phylogenetic analyses are consistent with the view that different DR haplotypic groups are evolutionarily ancient ~3. Thus, DRB alleles at each expressed locus (DRB1, 3, 4 and 5) are very similar within each haplotypic group.

Vol. I5 No. 2 1994

viewpoint Expression of DRB genes An additional line of data supports the notion that DRB3, DRB4 and DRB5 may indeed be allelic. First, the expression of DRB3 and DRB4 gene products is significandy lower than that of DRBI allelic products "t'-'-'. Th~s is substantiated by the frequency of isolated DRB cDNA clones from different DRB genes ";. In addition, more recent studies have shown that DRB1 mRNA is more anund'_ant than DRB3 ov DRB4 mRNAs, and that DRBI genes are differentially regulated at the transcriptional level as compared with DRB3 or 4 genes"