Chimpanzee C4 and CYP21 genes

Eur. J. Immunol. 1990. 20: 739-745

739

Hiroshi Kawaguchiov, Mladen Golubic., Felipe Figueroa. and Jan Kleinoo

Organization of the chimpanzee C4-CYP21 region: implications for the evolution of human genes*

Max-Planck-Institutfur Biologie., Abteilung Immungenetik and Department of Microbiology and Immunologyo, University of Miami School of Medicine, Miami

We prepared a cosmid library from chimpanzee DNA and screened it with a mouse probe specific for the complement component 4 (C4)-encoding gene. We isolated 29 clones and constructed restriction maps for 20 of these. The clones could be arranged into two overlapping clusters covering the entire C4 region of both chromosomes in this particular heterozygous chimpanzee. The region is about 100 kilobases (kb) long and contains two C4 and two CYP21 genes, the latter coding for the enzyme 21-hydroxylase. Using oligonucleotide probes we identified the genes as corresponding to human C4A, C4B, CYP21 and CYP21P genes.The last gene apparently contains an 8-base pair (bp) deletion (as does the corresponding human gene), which renders it a pseudogene. The genes are arranged in the order C4A . . . CYP21P. . . C 4 B . . . CYP21. Each of the two C4 genes is 16 kb long and thus corresponds to the short version of the human C4 genes.We suggest that the duplication of the basic C4-CYP21 unit that generated the standard arrangement of the human C 4 - U P 2 1 region occurred before the separation of the evolutionary lineages leading to humans and chimpanzees (i.e., more than five million years ago). We suggest further that the original form of the C4 gene was of the long variety and was generated by the insertion of a 6.8-kb element into one of the C4 introns.The element was subsequently excised in the ancestors of the chimpanzees and in at least one lineage of the human C4B gene. We speculate that the presence of the 6.8-kb insert in the human C4A and some C4B genes might largely be responsible for the great instability of this chromosomal region which leads to frequent duplications and deletions, some of which cause 21-hydroxylase deficiency.

1 Introduction In mammals, the clusters of class I and class I1 loci in the MHC (reviewed in [l]) are separated from each other by a chromosomal segment containing a number of unrelated loci [2-51. Prominent among these loci are C4 and CYP21, coding for the fourth complement component and the cytochrome P450, respectively. The C4 component is a serum glycoprotein composed ofthree chains, a,p and y, held together by disulfide bonds and generated by the post-translational cleavage of a single polypeptide, the pro-C4 [6]. The a chain contains, in a region referred to as C4d, a thiolester group which, when exposed by the removal of the chain’s N-terminal fragment (C4a), readily forms bonds with other substances. The bond formation anchors the remainder of the C4 molecule (the C4b fragment) to a surface on which the rest of the complement cascade can then take place. Cytochrome P450 is an enzyme (21-hydroxylase, E.C. 1.14.99.10) which catalyzes the conversion of progesterone t o 1l-deoxycorticosterone and of 17-a-hydroxy-progesteroneto 1l-deoxycortisol in

[I 80331

*

This work was supported in part by grant No. 1 RO1 A123667 from the National Institutes of Health, Bethesda, Maryland, and by a grant from the Fonds der Chemischen Industrie, Frankfurt, FRG . Permanent address: Department of Dermatology, Yokohama City University School of Medicine, Yokohama, Japan.

Correspondence: Jan Klein, Max-Planck Institut fiir Biologie, Abteilung Immungenetik, Corrensstr. 42, D 7400 Tubingen, FRG 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990

the biosynthetic pathway of adrenal corticosteroids. A 21-hydroxylase deficiency is one of the most common causes of congenital adrenal hyperplasia (for reviews see [7,8]. Several autoimmune disorders, such as SLE and Grave’s disease, have been associated with certain C4 alleles. In the human MHC, the H L A complex, the C4 and CYP21 genes occupy adjacent loci [9-131 and the CYP21 gene is the closest of all the known unrelated loci to the most proximal class I1 gene (HLA-DRA).The distance between CYP2I and DRA is approximately 300 kb.The number of copies of the human CYP21 and C4 genes varies among individuals, but most individuals have two copies of C4 and two copies of CYP21 [ l l , 12,141.The two C4genes are designated C4A and C4B, whereas the two cytochrome P450 genes are referred to as GYP21 and CYP21PThe loci are arranged in the order HLA-DRA . . . CYP21 . . . C 4 B . . . CYP21P.. . C4A, followed by unrelated loci G l l , RD, BF, C2 and others, and then by class I loci, in the direction from the centromere to the telomere. The arrangement probably arose by the duplication of a chromosomal segment that was about 30 kb long and contained a single pair of CYP21 and C4 genes [9].The CYP22P is a pseudogene as a result of three changes: an 8-bp deletion in the third exon, l-bp insertion in the seventh exon and a point mutation in the eighth exon. The first two changes cause a shift in the reading frame and the third change results in premature chain termination [15-17].The two C4 genes are both active but seem to have differentiated in their function [18, 191. The product of the C4A gene binds more efficiently to protein antigens (the thiolester group reacts more readily with amino acids forming amide bonds), whereas the C4B molecule attaches preferentially t o carbohydrates (the thiolester group reacts, in this case, more readily with 0014-2980/90/0404-0739$02.50/0

740

H. Kawaguchi, M.Golubic, E Figueroa and J. Klein

carbohydrate hydroxyl groups forming ester bonds). This difference in biological activity is determined by the amino acid sequence in the C4d region. The human C4A gene is about 22 kb long; the C4B gene occurs in two versions, long (22 kb) and short (16 kb; [ l l , 12]).The difference between the long and the short gene is attributed to the presence of a 6.8-kb fragment approximately 3 kb downstream from the 5' end (probably in the intron) of the long C4B gene [ l l , 12, 201. The short and long C4B genes exist as polymorphisms in human populations. The variation in the number of C4 and CYP21 genes probably results from rare instances of unequal crossing-over involving this chromosomal region. Some forms of congenital adrenal hyperplasia are caused by the deletion of the active CYP21 gene as a result of unequal crossing-over [21-241. The variation in gene copy number, the insertion polymorphism at the C4B loci, the slight functional difference between the C4A and C4B genes and the presence of the CYP21P pseudogene all offer an excellent opportunity to study the evolution of a short but highly active (in terms of structural rearrangements) chromosomal region. By determining approximately the time at which some of these changes occurred, one may obtain information about the relationship between structural changes and the origin of species. Recent studies [25-281 support the hypothesis that gene polymorphisms are sometimes passed from one species to another [29, 301 and the possibility exists that similar trans-species evolution may also apply to short chromosomalregions.The sequence similarity between the C4A and C4B, as well as that between the CYP21 and CYP21P loci suggests that the presumed duplications of these loci occurred relatively recently. It can therefore be expected that the study of close relatives of Homo sapiens will shed light on the origin and evolution of the C4-CYP21 region. Because of these considerations,we have initiated a systematic investigation of this region in the great apes and other higher primates. In this communication, we report the results of our studies on the chimpanzee C 4 - U P 2 1 region.

2 Materials and methods 2.1 Cell lines The human EBV-transformed B lymphoblastoid line (Lie) was provided by Dr. Marion Schneider, Universitat Diisseldorf, Institut fiir Blutgerinnungswesen und Transfusionsmedizin, Diisseldorf, FRG. The EBV-transformed B lymphoblastoid lines derived from chimpanzees Hugo, Jakob, Jolanda and Yvonne were provided by Drs. Ronald E. Bontrop and Margreet Jonker, TNO Primate Center, Rijswijk, The Netherlands.

2.2 Probes

Eur. J. Immunol. 1990.20: 739-745 fragment specificfor the human CYP2I gene was purchased from Appligene (Illkirch, France). Oligonucleotide probes were synthesized for us by Dr. J. Pohlner, Max-PlanckInstitut f i r Biologie,Tiibingen.The5' G A G C A G A G A C C A A C G A C A G 3' oligonucleotide was complementary to the sequence flanking the 8-bp deletion in the human CYP21P gene [16].The 5' T G T C T A A C A C T G G A C A G G G 3' oligonucleotide was specific for the nucleotide sequence 3398-3416 of the C4d region in the human C4A cDNA clone [191.The 5' T A T G T A T C A C T G G A G A G A G 3' oligonucleotide was specific for the corresponding sequence in the human C4B cDNA clone. Oligonucleotides were labeled using Y~~P-ATP and T4 polynucleotide kinase [32]. All other probes were oligolabeled using a32P-CTF' (Pharmacia oligolabelingkit, Uppsala, Sweden).

2.3 Construction and screening of cosmid library Genomic DNA was isolated from the Hugo cell line according to Davis and colleagues [33] with modifications. A cosmid library was constructed and screened as described by Steinmetz and co-workers [34]. It contained 1.5 x 106 independent clones. After hybridization, filters were washed with 0.1 x SSC, 0.1% SDS at 60 "C.

2.4 Isolation and analysis of cosmid clones DNA was isolated from bacterial colonies according to Maniatis and colleagues [35] and digested with restriction enzymes. Fragments were separated on agarose gel and transferred to Hybond-N membranes (Amersham, Braunschweig, FRG). Some of the fragments were subcloned into pUC8 or pUC19 plasmid vectors according to Davis and colleagues [37].

2.5 DNA blot analysis Hybridizations with cloned cosmid DNA were carried out at 42°C in 50% deionized formamide, 5 x SSC, 1 x Denhardt's solution, 0.2% SDS, 0.02 M phosphate buffer (pH 6.8) and 100 pg/ml of denatured salmon sperm DNA. After 16 h, filters were washed with 0.1 x SSC, 0.1% SDS at 60 "C. Genomic DNA was hybridized as above, but the filters were washed with 2 x SSC, 0.1% SDS at 50 "C.When oligonucleotides were used as probes, agarose gels were dried and directly hybridized at 45 "C. Gels were washed with 6 x SSC at room temperature. Filters and gels were exposed to XAR5 films (Kodak, Stuttgart, FRG) with intensifying screens for 1 to 3 days at - 70 "C.

3 Results 3.1 Restriction map of the C4-CYP21region

The mouse C4 probe was a 1.9-kb Bam HI fragment derived from the mouse cDNA clone pMC4/7 [31], which was given to us by Dr. Tommaso Me0 (Institut Pasteur, Paris, France). The human 5' and 3' end C4 probes were 500-bp and 2.2-kb Bam HI - Kpn I fragments, respectively [22], provided by Dr. Peter Schneider, Institut fiir Rechtsmedizin, Universitat Mainz, FRG. The 2.0-kb Bam HI

A cosmid library was prepared from DNA of a chimpanzee named Hugo, and was screened with a mouse C4 and human CYP21 probes. The former probe, designated pMC4/7 [3l],was derived from a cDNA clone that encoded the carboxy-terminalhalf of the C4a chain and almost the entire length of the y chain. (The precursor, pro-C4, is

Eur. J. Immunol. 1990. 20: 739-745

Chimpanzee C4 and CYP21 genes

synthesized as a single polypeptide chain comprising the fl, a and y chains, in that order.) The screening yielded 29 independent cosmid clones which, when isolated and digested with Taq I, produced 4.3- and 0.8-kb, 3.5- and 0.8-kb, 4.3-, 3.5 and 0.8-kb, or 0.8-kb pMC417-hybridizing fragments. Some of the clones also yielded a 2.5-kb pMC4/7-weakly hybridizing Taq I fragment. Twenty clones C4A

CYP2lP

H

741

were then selected for further characterization. They were digested with restriction endonucleases Cla I, Sal I, Bgl 11, Hind 111,Eco RI,Kpn I and Taq I, singly or in combinations of two enzymes, and restriction maps of the individual clones were constructed. The clones could be arranged in two clusters, A and B, containing 15 and 5 overlapping clones, respectively (Fig. 1A). In addition, the digests were

C4B

cYP21

H

ClOI

SalI

BglII HindIII EcoRI Kpn I

ToqI

C62 I

I

c11

cob

1

I

I

I

C25

I

I

C1B P c11 P C51 I CI1 1 I C16 I C52 I c22 I I C26 I C31 I I C32 I c12 I I

C4A

CYP21P

H

C4B

I I I I

I

CYP2l

H

ClaI

Sir1

BglII HindIII EcoRI KpnI Toq I

CLl I

I

c93 I

I

ca3 I C65

I I

1

C61

kb

(A)

0 I

10.

I

1

20

30

LO

50

60

70

80

90

100

I

I

I

I

I

I

I

I

I

CIA

I

I

EcoRI I

1.L

KpnI

I

Barn H I ToqI I (TaqI I

I

I 1.L I 2.2

I I

'

6.4 A

I I

9.5

I

6.6

11.0 1.61

3.3

6.L A

I

1.51 1.0

2.0

2.6

I

2.6

10.81

5.8

I L.3*

10.81

2.5

11.01

3.5+4+

10.81

2.5

11.01

)

U

C4B

I Eco R I

KpnI Barn H I TaqI

I I

I I

2.2 3.3

5.6 A

1.6 I

I 11.0 1.51

I

2.6

2.0

1.51 1.0

(B)

kb

I

5.8

I

10.81

3 . 5 W b 10.81 U

U

I

9.5

I

6.6 .

2.5

11.0

I

Figure 1. (A) Restriction maps of the C4-CYP21 regions in the two chromosomes of the chimpanzee named Hugo. The 4.3kb and 3.5-kb Taq I fragments detected by the mouse C4 probe are marked b y * and **, respectively; the 0 and 0 symbols indicate 3.2-kb and 3.7-kb Taq I fragments of the CYP22 genes, respectively. (B) Expanded restriction map of the two C4 genes. The A and A symbols indicate the 6.4-kb (C4A) and 5.4-kb (C4B) Taq I fragments, respectively. The brackets show the region detected b y the C4d-specific oligonucleotide. The Taq I restriction sites in parentheses are those of the allelic C4A gene. The asterisks indicate the same as in (A). Numbers between vertical bars indicate the length of fragments produced by the action of the restriction enzyme on the left-hand side.The scale in kb is given at the bottom.

742

Eur. J. Immunol. 1990.20: 739-745

H. Kawaguchi, M. Golubic, F. Figueroa and J. Klein

also hybridized with a probe specific for the CYP21 gene. The probe detected almost the entire coding sequence of the U P 2 1 gene and in the Taq I digests of the individual clones hybridized with3.2-kb, 3.7- plus 1.4-kb, or 3.7-, 3.2and 1.4-kb fragments. The position of the C4 and CYP21 genes in the clones was determined by hybridization of the individual restriction enzyme digests with the pMC4/7 and CYP21 probes. Clones C11 and C18 proved to be particularly informative in regard to the localization of the C4 and CYP21 genes. Upon digestion with Taq I, the C11 clone yielded two pMC4/7hybridizing fragments, 4.3 kb and 3.5 kb, mapping to different positions and thus presumably identifying two physically linked C4 genes. This presumption was further supported by the hybridization pattern of the C18 clone with the CYP21 probe. The clone produced two Taq I CYP21-hybridizing fragments, 3.2 kb and 3.7 kb. On the restriction map, these two fragments flanked the 4.3-kb and 3.5-kb, C4-hybridizing fragments, respectively. We interpret these data as indicating the presence of two C4 and two CYP21 genes on the same chromosome, arranged in the order C 4 . . . CYP21 . . . C4 . . . CYP21 (Fig. 1A). Analysis of other clones belonging to cluster A supports this interpretation. The restriction maps of the five clones belonging to cluster B (clones C41, C61, C65, C83 and C93) were similar to those of clones C62, C11, C04, C25 and C18 in cluster A. The one conspicuous difference was that one of the Taq I , C4-hybridizing fragments in cluster B clones was 3.5 kb long rather than 4.3 kb long, as was the case in the cluster A clones. It appears likely, therefore, that cluster B clones cover part of the second C4-CYP21 chromosome in the heterozygous chimpanzee from whose DNA the cosmid library was constructed. If so, the 3.5-kbTaq I fragment is derived from a C4 gene that is allelic to the C4 gene characterized by the 4.3-kb Taq I fragment in the first chromosome. Since the restriction maps of the cluster B clones are different from those of cluster A C51, C81, C16 C52, C22, C26, C31, C32 and C12 clones, the C4 gene borne by cluster B clones cannot be an allele of the C4 gene characterized by the 3.5-kb Taq I fragment on the first chromosome. No allelic variant of this second C4 gene was found among the isolated clones, indicating that either the second chromosome carries only one C4 (and one CYP21) gene or the second C4 (and CYP21) gene on the second chromosome is indistinguishable, by restriction fragment analysis, from the second gene of the first chromosome.We consider the latter possibility more likely than the former because had there been only one C4-CYP21 unit on one of the two chromosomes and two units on the other chromosome, clones C65 and C61 could be expected to show a different distribution of restriction sites compared to their counterparts in cluster A.The fact that the restriction maps in this region were identical between clusters A and B suggests that some of the clones in cluster A are in fact derived from the first and others from the second chromosome of the heterozygous chimpanzee.

C4-bearing fragments of the cosmid clones were subcloned in pUC8 or pUC19 plasmid vectors, digested with Eco RI, Kpn I, Bam HI and Taq I restriction endonucleases, and relatively precise restriction maps of the genes were constructed (Fig. lB).To locate the ends of the genes, the digests were hybridized with the 5'- and 3'-specific probes. In all three genes, the 5' end was defined by the 3.3-kb Bam HI fragment and the 3' end by the 5.8-kb Kpn I fragment in Fig. 1B. From these results, we estimate that all three C4 genes are of the same length, namely 16 kb.

3.3 Identification of the C4 genes The C4d region of the human C4 genes, specifically the part coding for amino acids 1100-1107, contains sequences differentiating the C4A and C4B genes [lo]. On the assumption that this region might be conserved between humans and chimpanzees, we prepared oligonucleotide probes A and B specific for the human C4A and C4B genes, respectively, and hybridized them with restriction fragments derived from selected C4 clones isolated from the chimpanzee library. Oligonucleotide A hybridized with 6.4-kb Kpn I fragments of chimpanzee clones C11, C41, C62 and C65 (Fig. 2A). Oligonucleotide B hybridized with 6.4-kb Kpn I fragments of clones C11, C18, and C51 (Fig. 2B).This result reveals that clone C11 bears both the C4A and C4B genes; clones C41, C62 and C65 bear the C4A gene; and clones C18 and C51 bear the C4B gene (see Fig. 1A; the C18 clone apparently bears the 3' end of the C4A gene but not the part corresponding to the C4d region, and therefore it does not hybridize with the C4A-specific oligonucleotide). The conclusion that the C11 clone contains both the C4A and C4B genes is further supported by the observation that in a Taq I digest of this clone, a 4.3-kb fragment hybridized with the C4A-specific oligonucleotide, and a 3.5-kb fragment hybridized with the C4B-specific oligonucleotide (data not shown). The data also support our interpretation that the gene borne by cluster B clones is allelic to one of the genes borne by cluster A clones: both are C4A genes.

6 . 4 kb

6 . 4 kb

-

(W

3.2 Characterization of the C4 genes

To determine the lengths of the three C4 genes, probes specific for the 5' and 3' ends of the gene were used. The

Figure 2. Identification of C4A and C4B genes on selected chimpanzee cosmid clones through hybridization with oligonucleotide probes. (A) Southern blot of Kpn I-digested clones hybridized with C4A-specific oligonucleotide probe. (€3) Southern blot of Kpn I-digested clones hybridized with C4B-specific probe.

Eur. J. Immunol. 1990.20: 739-745

Chimpanzee C4 and CYP21 genes

743

3.4 Identification of the CYP21P gene

3.5 The C4-CYP21 region in other chimpanzees

As mentioned earlier, an 8-bp deletion renders one of the two CYP2Z genes a pseudogene [16].To determine whether a similar deletion might be present also in one of the chimpanzee CYP21 genes, we prepared another oligonucleotide which spans the deleted region and encompasses 19 bp of flanking sequence. When tested on selected chimpanzee cosmid clones, the probe hybridized with 3.2-kb Taq I fragments obtained from clones C11, C18, C62, and C65; it did not hybridizewith a similar fragment of clone C32 (Fig. 3A).When theTaq I digests of these clones were tested by the CYP2Z-specific cDNA probe, the 3.2-kb fragment was demonstrated in the C11, C62 and C65 clones; a 3.7-kb fragment was demonstrated in the C32 clone; and both fragments were demonstrated in the C18 clone (Fig. 3B).We interpret this result as an indication that clone C32 bears the CYP2Z gene, clones C11, C62 and C65 bear the CYP2ZP gene and clone C18 bears both genes (Fig. 1A). This interpretation is again consistent with the arrangement of C4 and CYP2Z genes as depicted in Fig. 1A.

To determine whether variation in the organization of the C4-CYP2Z occurs among individual chimpanzees as it does among humans, we tested Taq I digests of genomic DNA isolated from three other chimpanzees (Jolanda, Jakob and Yvonne) and compared the patterns obtained with those of Hugo (Jakob is Hugo's father) and those of human (Lie cell line) DNA. Hybridization with the 5' end C4 probe did not reveal any differences among the four chimpanzees (Fig. 4A): all four seem to carry the short (16 kb) version of both the C4A and C4B genes.The human Lie DNA, on the other hand, appears to contain a long version of the C4A gene (producing the 7.0-kb fragment) and both long and short versions of the C4B gene (producing the 6.0-kb and 5.4-kb fragments, respectively). Hybridization with the mouse C4 cDNA probe revealed a difference between Jolanda and the other three chimpanzees: Jolanda's hybridization pattern was the same as that found with the human DNA (Fig. 4B). One possible interpretation of the patterns is that Jolanda is homozygous for the 3.5-kbTaq I fragment of the C4 gene, Hugo is a 3.514.3 heterozygote, and Jakob as well as Yvonne are either 4.3/4.3 homozygotes or, like Hugo, 3.514.3 heterozygotes. No difference in the large Taq I fragments was detected among the chimpanzees when their DNA were hybridized with the CYP2Z probe (Fig. 4C). Differences between these and the human DNA in the smaller fragments are difficult to interpret for technical reasons. On the whole, however, no major variations could be detected in the organization of the C4-CYP2Z region among the four chimpanzees.

4 Discussion

Figure 3. (A) Southern blot analysis of ?liq I-digested chimpanzee cosmid clones with oligonucleotide probe specific for the 8-bp deletion in the human CYP21 P gene. (€3) Control blot hybridized with a cDNA CYP21 probe.

We have cloned about 100 kb of the C4-CYP2Z region of the chimpanzee Hugo. The region contains two C4 genes corresponding to the human C4A and C4B genes, and two CYP22 genes, one of which is a pseudogene. The genes are arranged in the order C4A . . . CYP2IP. . . C 4 B . . . CYP21, whichis the standard arrangement found in many human c4 haplotYPes. The identification of the two c 4 genes as C4A and C4B is in agreement with the data of Granados and his colleagues [36].These investigators studied genetic

Figure 4. Southern blot hybridization of genomic chimpanzee and human (Lie) Taq I-digested DNA with human 5' end C4 probe (A), mouse C4 cDNA probe (B) and human cDNA CYP21 probe (C).

I

744

H. Kawaguchi, M. Golubic, E Figueroa and J. Klein

polymorphism in the chimpanzee by agarose gel electrophoresis of desialated plasma and immunofixation with antiserum to human C4. They found multiple partially overlapping bands of which only the most cathodal ones had strong hemolytic activity. In analogy to human C4, they designated the latter C4B and those with little hemolytic activity C4A. Our data obtained by hybridization of C4Aand C4B-specific oligonucleotide probes establish the existence of these two paralogous genes in the chimpanzee and indicate a conservation of sequence between humans and chimpanzees in the section of the C4d region responsible for isotypic differences. The existence in the chimpanzee of genes orthologous to the human C4A and C4B genes suggests that the presumed duplication of the ancestral C4-CYP21 unit occurred before the spearation of the evolutionary lineages leading to humans and chimpanzees, an event estimated to have occurred between 5 and 7 million years ago [37-391. It is therefore probably not correct to assume, as some investigators have done, that the rare human haplotypes with a single C4-CYP21 unit are ancestral to those with duplicated C4-CYP21 units. If the duplication occurred before human speciation, the single-unit haplotypes must be derived secondarily from the duplicated ones by the subsequent deletion of one 30-kb unit. (An alternative possibility is, however, that both the single- and the double-unit chromosomes existed as polymorphism in the common ancestor of humans and chimpanzees and that this polymorphism was handed on to humans but not to chimpanzees.) We have prepared rather precise restriction maps of the C4 genes carried by Hugo and have demonstrated that the C4A and C4B genes of this chimpanzee are each 16-kblong, which is the length of the short version of the human C4B gene. This observation has an important implication for the origin of the human long C4 gene. As mentioned earlier, all human C4A genes are of the long variety (22 kb), whereas C4B genes can be both long and short (16 kb) [ l l , 12, 201. The insertion of 6.8 kb of sequence into an intron in the 5' region of the C4 gene (an insertion thought to be responsible for the generation of the long version of the C4 gene) must have occurred before the duplication that generated C4A and C4B genes. Otherwise, one would have to postulate that the same insertion occurred twice, once into the C4A and a second time into the C4B gene, both presumably at the same position. Such an event is highly improbable. If the insertion occurred only once, it must have done so before the C4A and C4B genes separated. Since our data suggest that the separation predated human speciation, we must argue that the insertion predated this event as well.We suggest, therefore, that a long version of a single ancestral C4 gene existed in the primate lineage at least 5-7 million years ago.The long gene then duplicated and through diversification produced the C4A and C4B ancestral genes, both of the long variety. In the lineage leading to the present-day chimpanzees, the 6.8-kb element was subsequently excised in both C4 genes, converting them into their short versions. In the human lineage, the 6.8-kb element has thus far been excised in only some C4B genes; all C4A genes have retained their long form. If this hypothesis is correct and the insertion occurred long before the separation of the human and chimpanzee lineages, it might still be possible to find the long version of the C4 gene in other apes. In fact, preliminary results

Eur. J. Immunol. 19%. 20: 739-745 indicate that the long version of the C4 gene is present in the orang-utang (H. Kawaguchi and J. Klein, unpublished data). The C4-CYP21 segment is also duplicated in the house mouse, but there is evidence that in this species, the duplication occurred independently from that in humans and in the chimpanzee [31]. On the other hand, other species, including the horse [@I, guinea pig [41] and cat [42], may not have duplicated their C4 loci. Hugo appears to be heterozygous at the C4A and homozygous, within the limits of our analysis, at the C4B locus. The limited study of other chimpanzees suggests that some of their genes, too, might be polymorphic but that they all carry two C4 and two CYP21 genes. Granados and his colleagues [32] found three C4A and two C4B variants in their study of a larger panel of chimpanzees. Although they reported the identification of half-null haplotypes with C4B*Q0, they provided no evidence that the apparent absence of the C4B protein was due to a deletion of the C4B gene. The overall impression, and it is certainly no more than that at this stage, is that the chimpanzees are more homogeneous in terms of C4 gene copy number than humans are. If this impression could be confirmed by further studies, one could then speculate that the relatively high instability of the human C4-CYP22 haplotypes, reflected in frequent occurrences of duplications and deletions, is largely due not so much to sequence homology between the duplicated genes themselves (as has been postulated, see for example [9]), as to the presence of the 6.8-kb insert. It is possible that the inserted sequence promotes recombination when it is present in both the C4A and C4B genes. Pairing between the 6.8-kb inserts in the two genes might lead t o frequent cases of unequal crossingover and consequently to duplications and deletions [43]. Such extreme instability of a chromosomal region might be selectively disadvantageous, particularly if it encompasses vital genes such as CYP22, and the evolutionary trend might therefore be toward loss of the insert. Perhaps the chimpanzees have already managed this and humans are only in the early phase of the process. This possibility, of course, brings little comfort to the many patients suffering from 21-hydroxylase deficiency. The presence of the 8-bp deletion in one of the chimpanzee GYP21 genes, as implied by the oligonucleotide hybridization data, suggests that this gene is functionally inactive, just as it is in humans. The CYP21P pseudogene must therefore have been silenced at least 5 to 7 million years ago. Attempts to use oligonucleotide probes to identify two other defects in the chimpanzee which are known to exist in the human CYP22P gene (a frame-shift and chain termination mutations) failed for technical reasons. Apparently, the chimpanzee and human sequences diverged in these regions to such a degree that species differences mask the two mutations when oligonucleotides are used as probes. Sequencing will have to be carried out to determine the point at which the other two mutations occurred in the evolution. We thank all investigators mentioned in Section 2 for providing us with probes or ceN lines. We thank Ms. Lynne Yakes for editorial assistance.

Received October 19, 1989; in revised form December 4, 1989.

Eur. J. Immunol. 1990. 20: 739-745

5 References 1 Klein, J., Natural History of the Major Histocompatibility Complex, John Wiley & Sons, New York 1986. 2 Dunham, I., Sargent, C. A.,Trowsdale, J. andcampbell, R. D., Proc. Natl. Acad. Sci. USA 1987. 84: 7237, 3 Carroll, M. C., Katzman, p., Alicot, E. M., Koller, B. H., Geraghty, D. E., Orr, H. T., Strominger, J. L. and Spies, T., Proc. Natl. Acad. Sci. USA 1987. 84: 8535. 4 Spies,T., Blanck, G., Bresnahan, M., Sands, J. and Strominger, J. L., Science 1989. 243: 214. 5 Sargent, C. A., Dunham, I. and Campbell, R. Duncan, EMBO J. 1989. 8: 2305. 6 Reid, K. B. M. and Porter, R. R., Annu. Rev. Biochem. 1981. 50: 433. 7 White, P. C., New, M. I. andDupont, B., N . Engl. J. Med. 1987. 316: 1580. 8 Miller, M. J., A m. J. Human Genet. 1988. 42: 4. 9 White, P. C., Grossberger, D., Onufer, B. J., Chapfin, D. D., New, M. I., Dupont, B. and Strominger,J. L., Proc. Natl. Acad. Sci. USA 1985. 82: 1089. 10 Carroll, M. C., Campbell, R. Duncan, Bentley, D. R. and Porter, R. R., Nature 1984. 307: 237. 11 Carroll, M. C., Campbell, R. Duncan and Porter, R. R., pro,-. Natl. Acad. Sci. USA 1985. 82: 521. 12 Carroll, M. C., Palsdotti, A., Belt, K. T. and Porter, R. R., EMBO J. 1985. 4: 2547, 13 Uring-Lambed, B., Goetz, J., Tondo, M. M.,Mayer, s. and Hauptmann, G., Tissue Antigens 1984. 24: 70. 14 McLean, R. H., Donohoue, P. A., Jospe, N., Bias, W. B., Van Dop, C. and Migeon, C. J., Genomics 1988. 2: 76. 15 Whilte, P. C., New, M. I. andDupont, B., Proc. Natl. Acad. Sci. USA 1986. 83: 5111. 16 Higashi, Y , Yoshioka, H., Yanae, M., Gotoh, 0.and FujiiKuriyama,Y., Proc. Natl. Acad. Sci. USA 1986. 83: 2841. 17 Rodrigues, N. R., Dunham, l.,yung,yu, c., carroll, M: c., Porter, R. R. and Campbell, R. D., EMBO J. 1987. 6: 1653. 18 Law, S. K. A., Dodds, A.W. and Porter, R. R., EMBO J. 1984. 3: 1819. 19 Belt, K. T., Carroll, M. C. and Porter, R. R., Cell 1984. 36: 907. 20 Prentice, H. L., Schneider, l? M. and Strominger, J. L., Imrnunogenetics 1986. 23: 274. 21 Boehm, B. O., Rosak, C., Boehm,T. L. J., Kuehnl, €!,White, P. C. and Schoffling, K., Mol. Biol. Med. 1986.3: 4137. 22 Schneider, P. M., Carroll, M. S., Alper, C. A., Rittner, C., Whitehead, A. S. ,Yunis, E. J. and Colten, H. R., J. Clin. Inv. 1986. 78: 650.

Chimpanzee C4 and CYP21 genes

745

23 Werkmeister, J.W., New, M. I., Dupont, B. and White, P. C., A m. J. Hum. Genet. 1986. 39: 461. 24 Collier, s.,Sinnott, P. J., Dyer, l? A., Price, D. A., Harris, R. and Strachan, T., EMBO J. 1989. 8: 1393. 25 Figueroa? F., GUnther, E. and Klein, J., Nature 1988. 335: 265. 26 MaYer,W. E.,Jonker, M., Klein, D., Ivanyi, P.,Van Seventer, G. and Klein, J., EMBO J. 1988. 7: 2765. 27 Lawlor, D. A., Ward, F. E., Ennis, P. D., Jackson, A. P. and Parham, P., Nature 1988. 335: 268. 28 McConnel1,T. J. ,Talbot,W. S., McIndoe, R. A. and Wakeland, E. K., Nature 1988. 332: 651. 29 Klein, J., in Fougereau, M. and Dausset, J. (Eds.), Immunology 80: Progress in Immunology Iv Academic Press, London 1980. 30 K1ein, J.y Hum, zmmunol. 1987- 19: l55. 31 Tosi, M., LCvi-Straws, M., Duponchel, C. and Meo, T., Phil. Trans. R. SOC. Lond. 1984. 306: 389. 32 Conner,B.J., Reyes, A. A., Morin, C., Itakura, K.,Teplitz, R. L- and 19g3- 80: R. B., proc. Natl, Acad. sci. 272. 33 Davis, L. G., Dibner, M. D. and Battey, J. F., Basic Methods in Molecular Biology, Elsevier, New York 1986. 34 Stehmetz, M., Stephan, D., Dastronikoo, G. R., Gibb, E. and Romaniuk, R., in Lefkovits, I. and Pernis, B. (Eds.), Immunological Methods, vol. 3 Academic Press, New York 1985, p.2. 35 ManiatiS, T., Fritsch, E. F. and Sambrook, J., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York 1982. 36 Granados, J., Awdeh, Z. L., Chen, J.-H., Giles, C. M., Balner, H.zYunis, E- J. and Alper, c. A., Immunogenetics 1987. 26: 344. 37 Sarich,V. M. and Wilson, A. c . , Science 1967. 158: 1200. 38 Sibley, C. G. and Ahlquist, J. E., J. Mol. Biol. 1984. 20: 2. 39 Nei, M-, in Ohta,T. and Aoki, K. (Eds.1, Population Genetics and Molecular Evolution, Jap. Sci. SOC.Press,TokyolSpringerVerlag, Berlin 1985, p. 41. 40 Kay, P. H., Dawkins, R. L., Bowling, A. T. and Bernoco, D., J. Immunogenet. 1987. 14: 247. 41 Tosi, M., LCvi-Straws, M., Georgatson, E., Amor, M. and Meo, T., Imrnunol. Rev. 1985. 87: 151. 42 DeKroon, A. I. P. M., Doxiadis, L. I. and Hensen, E. J., zmmunogenetics 1986. 24: 20243 Carrol, M. C. and Alper, C. A., Brit. Med. Bull. 1987. 43: 50.