Single Base Mutation in &5(W) Collagen Chain Gene Converting a Conserved Cysteine to Serine in Alport Syndrome

JING ZHOU,* DAVID F. BAfwm,t SIRKKA LIISA HOSTIKKA,* MARTIN C. GREGORY,* CURTIS L. ATKIN,*~ AND KARL TRYGGVASON*

*Biocenter and Department of Biochemistry, University of O&I, SF-90570 Oulu, Finland; and Departments of tMedica/ lnformatics, §t?iochemistry, and $Med/one, University of Utah School of Medicine,

50 North Medical Drive, Salt Lake CYty, Utah 84132

Received July 13, 1990, revlsed October 4, 1990

We have identified a point mutation in the type IV colla- gen o5 chain gene (COL4A5) in Alport syndrome. Variant PstI (Barker et al., 1990, Science 248, 1224-1227). and BgZII restriction sites with complete linkage with the Al- port phenotype have been found in the 3’ end of the COL4A5 gene in the large Utah Kindred P. The approxi- mate location of the variant sites was determined by restric- tion enzyme mapping, after which this region of the gene (1028 bp) was amplified with the polymerase chain reac- tion (PCR) from DNA of normal and affected individuals for sequencing analysis. The PCR products showed the ab- sence or presence of the variant PstI and Bgl II sites in DNA from normal and affected individuals, respectively. DNA sequencing revealed a single base change in exon 3 (from the 3’ end) in DNA from affected individuals, chang- ing the TGT codon of cysteine to the TCT codon for serine. This single base mutation also generated new restriction sites for PstI and Bgl II. The mutation involves a cysteine residue that has remained conserved in the carboxyl-end noncollagenous domain (NC domain) of all known type IV collagen (Y chains from Drosophila to man. It is presumably crucial for maintaining the right conformation of the NC domain, which is important for both triple-helix formation and the formation of intermolecular cross-links of type IV collagen molecules. I<: 1991 Aeademir Press, Inc.

INTRODUCTION

X-chromosome-linked Alport syndrome is a domi- nantly inherited progressive hematuric kidney dis- ease often accompanied by hearing loss and eye le- sions (Alport, 1927; Atkin et al., 1988a). The disease has been classified into six different types (Atkin et al., 1988a): Type I is a dominant juvenile form with deafness, type II is an X-linked dominant juvenile disease with deafness, type III is an X-linked domi- nant adult form with deafness, type IV is an X-linked dominant adult type without deafness, type V is an

Copyright ft. 1991 by Academic Press, Inc. All rights of reproduction in any form reserved

autosomal dominant type nephritis with deafness and thrombocytopathia, and type VI is an autosomal dom- inant juvenile disease with deafness. Using restriction fragment length polymorphism (RFLP) markers, the defective gene has been mapped to the Xq22-24 re- gion in several studies (Atkin et al., 198813; Brunner et al., 1988; Flinter et al., 1989; Kashtan et al., 1989). Ultrastructurally, the disease is characterized by ab- normalities in the glomerular basement membrane (GBM) such as patchy thickening and thinning with splitting of the lamina densa. These findings led to the suggestion of a defect in the major structural com- ponent, type IV collagen (Spear, 1973), and immuno- staining with antibodies against GBM-derived type IV collagen indicated alterations or absence of a type IV collagen N chain (Kashtan et al., 1986, 1989; Sav- age et al., 1989; Kleppel et al., 1989).

Type IV collagen is a complex protein composed of three TV chains that can be of types (ul(IV), cr2(IV), (u3(IV), a4(IV), or &i(IV) (Timpl and Dziadek, 1986; Butkowski et al., 1987; Saus et al., 1988; Gunwar et al., 1990; Hostikka et al., 1990) and that are assembled into a triple-helical molecule. Each N chain contains a carboxyl-terminal noncollagenous domain (NC do- main) and an amino terminal collagenous domain containing a frequently interrupted Gly-X-Y repeat sequence. The NC domain, which has a highly con- served primary structure, is probably crucial for the assembly of three CY chains and folding into the triple- helix conformation. It also participates in head-to- head cross-links between individual collagen mole- cules (Siebold et al., 1988). The major molecular for- mula of type IV collagen is probably [nl(IV)],a2(IV) but, at present, little is known about combinations involving the other three N chains. The trimeric mole- cule has a large carboxyl-terminal globular region formed by the NC domain of the three cy chains and an about 400.rim-long kinky rod-like collagenous do- main. In contrast to fibrillar collagens, type IV colla-

TYPE IV COLLAGEN DEFECT IN ALPORT SYNDROME 11

gen forms a network-like structure into which other components such as laminin, entactin, and proteogly- cans are bound in a still unknown fashion (see Timpl and Dziadek, 1986; Yurchenco, 1990).

We have identified and characterized the &(IV) chain from cDNA clones and localized its gene (COL4A5) to the region of Alport syndrome on Xq22 (Hostikka et al., 1990). Furthermore, immunohisto- logical studies showed this chain to be located in the kidney almost solely in the GBM (Hostikka et al., 1990; Sariola et al., manuscript in preparation). More importantly, we have recently demonstrated a large deletion in the COL4A5 gene in Utah Kindred EP with Alport syndrome (Barker et al., 1990). This dele- tion involves the deletion of about 15 kb from the 3’ half of the gene, resulting in the loss of six exons (5-10 as counted from the 3’ end). Analysis of the gene structure (Zhou et al., 1991) has revealed that this mutation would generate a short probably nonfunc- tional a5(IV) chain whose presence could interfere with the normal triple-helix formation and supramo- lecular assembly of type IV collagen molecules. The deletion is thus likely to be sufficient for causing the structural defects of the GBM in the patients in- volved. We have also identified a mutation generating a new PstI restriction site in the COL4A5 gene (Barker et al., 1990) in the large Utah Kindred P with type III Alport syndrome (Atkin et al., 1988a). This mutation showed complete linkage with the defective Alport gene.

In the present study we have shown that the muta- tion that generates the new PstI restriction site in Utah Kindred P also generates a new BglII site and is the result of a single base change in the third last exon of the COL4A5 gene. The mutation converts a con- served cysteine residue in the NC domain to serine. The implications of this mutation on the protein cuFi(IV) chain are discussed.

MATERIALS AND METHODS

Pedigree Material

Subjects included 122 members of Utah Kindred P, which carries Type III Alport syndrome. They were chosen to yield meioses potentially informative for genetic linkage of the Alport gene with RFLP markers (Hasstedt and Atkin, 1983). Type III Alport is distinguished by well-established X-linked inheri- tance, the presence of progressive sensorineural hear- ing loss, and the supervention of renal failure in mid- dle-aged males (Atkin et al., 1988a). Kindred P is by far the largest kindred studied: X-linkage has been proven by likelihood analysis (Hasstedt and Atkin, 1983) and linkage studies (Barker et al., 1989), hear- ing loss has been characterized by standard and ultra-

high-frequency audiometry (Wester, 1990), and the intrakindred mean age of end-stage renal disease is 33 f 7 years (n = 20) (Atkin et al., 1988a).

Southern Blot Analysis

DNA blot analyses of 122 members of Alport syn- drome Kindred P were carried out as previously de- scribed (Barker et al., 1990). The 2.2-kb genomic P.stI fragment, which corresponds to the fragment that is found to be altered in affected members of Kindred P (Barker et al., 1990), was isolated from a PstI digest of the genomic clone MG-2 (Zhou et al., 1991) (see Fig. 1). The 2.2-kb PstI fragment was identified on hybrid- ization with the MD-6 cDNA clone (Hostikka et al., 1990) (Fig. 1) and subcloned into the PstI site of pUC8.

Polymeruse Chain Reaction

For detecting and analyzing the mutations in indi- vidual patients, the polymerase chain reaction was applied to the amplification of specific segments of genomic DNA (Scharf et al., 1986; Erlich et al., 1988). The synthetic oligonucleotide primers chosen were (i) P, (5’-GACTCTAGAAAGGCCATTGCACTGGTT- 3’), containing a sequence upstream of the normally existing PstI site near exon 3; and (ii) P, (5’-AGC- GAATTCCTGACCTGAGTCATGTAT-3’), with a sequence downstream of exon 2. Each of the primers contained modified 5’ ends to produce XbuI and EcoRI restriction sites, respectively, for subsequent subcloning into the sequencing vector (Messing, 1983; Williams, 1989). The DNA segment of interest was amplified by mixing 1 pg of genomic DNA as template and 25 pmol of each primer in the polymerase reac- tion kit buffer (Perkin Elmer Cetus). Samples were initially melted at 94°C for 10 min and chilled on ice for 3 min, after which 0.5 ~1 of TuqI polymerase (Per- kin-Elmer/Cetus) was added to start the reaction. This was followed by 35 reaction cycles, each of which included 1.5 min of denaturation at 94”C, 2 min of annealing at 65”C, and 2 min of extension at 72°C. The amplified products were extracted with phenol/ chloroform, concentrated by ethanol precipitation, digested with XbaI and EcoRI, and electrophoresed in a 1% Nusieve GTG (FMC BioProducts) low-melting- point agarose gel. The expected fragments of about 1 kb were excised and purified, and portions were li- gated with XbaI/EcoRI-cut Ml3mp19 and M13mp18 vectors in the presence of 1 mM hexamine cobalt chlo- ride (Murray, 1986). Insert-containing clones were isolated and sequenced.

DNA Sequencing

Amplified DNA containing Ml3 clones were se- quenced with the Sanger dideoxynucleotide chain ter-

12 ZHOll El‘ Al,

A MD-6 cDNA

COL4A5

MG-2

B H

1.9 kb y Variant (Kindred P)

w 2.2 kb

y Normal

3 2 1

P P’ P

* *

P3 P4

FIG. 1. Schematic diagram ofthe 3’end of the human C’OL-IAS gene and normal and variant I’st I restriction sites. Exons (numbered from the 3’end) are indicated by shaded boxes and introns and 3’ flanking regions by solid interconnecting lines. (A) Scheme showing relationship of the MD-6 cDNA and MG-2 genomic h clones with the 3’ end of the gene. Locations of EcoRI restriction sites are indicated. (B) Pstl sites (P) in the 3’end of the gene. The variant PstI site found in Utah Kindred P is marked P*. The locations of the normal 2.2.kb and the variant 1.9.kb PstI fragments are shown above. The locations ofthe two polymerase chain reaction primers (I’, and P,) used for amplification ofthe variant region are indicated below

mination method (Sanger et al., 1977) using “Sequen- ase” (U.S. Biochemicals) with the Ml3 “universal primer” or specific sequencing primers designed on the basis of the known gene sequence and made in an Applied Biosystems, Inc., DNA synthesizer.

RESULTS

Novel PstI and B&II Sites Associated with the COMA5 Mutation in Kindred P

Southern hybridization of genomic DNA was carried out using a clone of the 2.2 PstI genomic frag- ment which spans exons 2 and 3 and the 5’ end of exon 1 (Zhou et al., 1991; Barker et al., 1990). This probe revealed only the 2.2- or 1.9-kb COL4A5-specific PstI fragments in the Kindred P individuals (data not shown), whereas the 0.3-kb fragment could not be de-

tected with Southern hybridization. As previously de- scribed (Barker et al., 1990), the 1.9-kb fragment is completely associated with the disease allele in the 122 tested members of Kindred P and no variant frag- ment of this length was observed on 102 other normal chromosomes examined. In particular, all 23 gene- carrier mothers appear to have one copy each of the 1.9- and 2.2-kb fragments, which segregated to a total of 45 sons in complete linkage with the presence or absence, respectively, of the Alport syndrome pheno- type.

When DNA blots of BglII digests from the same 122 individual members of Kindred P were probed with the 2.2-kb PstI fragment clone, a BglII site vari- ant that also showed complete concordance with the presence of the disease allele in this kindred was de- tected (data not shown). The normal BglII pattern is a single 13.2-kb fragment. The variant BgZII pattern

TYPE IV COLLAGEN DEFECT IN ALPORT SYNDROME 13

A B

1 2 3 4 5 123456789

I I I

PS, I SCJI II

FIG. 2. Agarose gel electrophoresis of PCR-amplified DNA from normal and Kindred P individuals. (A) Without digestion. 1Bl After digestion with PstI or BglII restriction enzymes. Lane 1, $X174 DNA digested with HueIII; lanes 2 and 6, DNA amplified from genomic clone MC-2 (39); lanes 3 and 7, from normal geno- mic DNA; lanes 4 and 8, DNA of affected male from Kindred P; lanes 5 and 9, DNA amplified from a gene-carrier mother.

consists of fragments of 10.4 and 2.8 kb (data not shown), indicating that a new BglII site occurs within the limits of the 2.2-kb PstI segment, either in close association with or at the site of the disease-causing mutation in this kindred. The variant BglII frag- ments were not observed on 102 normal chromosomes that have been studied.

Amplification and Sequencing of the Region Containing the Novel PstI and BglII Restriction Sites

The appearance of novel PstI and BglII restriction sites could be the result of deletion, insertion, inver- sion, or point mutation. A large deletion or insertion could be ruled out, however, since the restriction frag- ment patterns observed with EcoRI, PstI, and TaqI digestion (Barker et al., 1990), in addition to the BglII pattern described here, showed no changes in lengths of fragments from the same part of the gene. There- fore, a point mutation or a change involving a few nucleotides was considered most likely. As we have reported previously (Barker et al., 1990), the new PstI restriction site must reside within or in the close prox- imity to the 5’ end of exon 3 (see Fig. 1). To analyze this, we sequenced the region from 236 bp upstream of exon 3 to 162 bp downstream of exon 2 in genomic DNA from normal individuals, one affected Alport male, and one carrier mother. For this purpose, we first determined the sequence (data not shown) in the normal gene from a subclone of the previously iso- lated genomic clone MG-2 (Zhou et al., 1991). This yielded a nucleotide sequence spanning a 1038-bp re- gion, 241 and 682 bp, respectively, upstream and downstream of exon 3 (not shown). This region also

contains a normal PstI site located 215 bp upstream of exon 3 (see Fig. 1). Based on the sequence obtained, two oligonucleotide primers were made to amplify this region of the gene in the individual DNAs using the polymerase chain reaction. According to the se- quence in the normal gene, the amplified fragment should have the size of 1046 bp. Furthermore, all am- plified DNA from the affected male and half of that from the mother should contain the variant PstI site and possibly also the BglII site. In contrast, the am- plified fragment from control DNA and half of that from the mother should only have the normal PstI site, resulting in 28-bp (invisible) and 1013-bp frag- ments after digestion with PstI. Also, they should not have a BglII site.

The results of the PCR amplification are shown in Fig. 2A. The apparent size of the amplified products was the same, or about 1046 bp, in two control DNAs (lanes 2 and 3), the affected male (lane 4), and the carrier mother (lane 5). This demonstrated that the mutation(s) probably did not involve a deletion or in- sertion but rather a small inversion, point mutation, or changes of only a few nucleotides. The results of the digestion of the PCR fragments with PstI and BglII are shown in Fig. 2B. The DNA from controls had no visible band change (lanes 2 and 3). In con- trast, the digestion of DNA from the affected male (lanes 4 and 8) patient gave two visible fragments of about 240 and 790 bp both with PstI and BglII. This suggested that the mutation is a point mutation giv- ing rise to a new restriction site for both enzymes. Digestion of the amplified fragment from the carrier mother resulted in the digestion of about one-half of the DNA to fragments the same size as those from the male patients, whereas half of the sample remained undigested. These results completely agree with the results obtained with genomic DNA (Barker et al., 1990).

Sequencing of the PstI and BglII Polymorphic Sites

To determine the exact mutation(s), PCR-ampli- fied DNA fragments were subcloned into the Ml3 vectors and sequenced (Fig. 3). The results demon- strated a single base mutation in the fragment ampli- fied from the affected male, changing a G to a C. This mutation changed the TGT codon of cysteine to the TCT codon for serine (Figs. 3 and 4). The mutation was verified from 18 clones of two different PCR prod- ucts that were all sequenced. In the case of the carrier mother, 10 of the 16 amplified fragments did not con- tain the mutation, whereas the other 6 did.

Examination of the sequence showed that the point mutation was present at the 5’ end of exon 3 (Fig. 4) so that both a PstI and a BglII restriction site were gen- erated. This mutation causes a codon change from

14 ZHOU ET AL.

G

Genomic clone

Control DNA

Carrier Attected mother

male (10116)

GATC GATC GATC GATC

Carrier mother (6116)

m

GATC

T

r G A C G T

I

c = sar T A

g a

1

c c

FIG. 3. Partial nucleotide sequence of PCR-amplified DNA from normal and affected alleles. The mutation in the affected male of Kindred P was shown to be a single base change from G to C converting a codon for cysteine to the codon for serine. Sequencing of PCR-amplified DNA from the carrier mother (heterozygote) that was subcloned into the Ml3 vector (16 clones sequenced) gave the sequence from both the normal allele (10/16) and the variant allele (6/16).

cysteine to serine. The cysteine residue involved is the eighth from the carboxyl terminus of the NC do- main which contains 12 cysteine residues (Hostikka et al., 1990).

DISCUSSION

The present study has demonstrated that the vari- ant PstI and BglII restriction sites in the COL4A5 gene associated with the Alport phenotype in Utah Kindred P are not neutral variants, but in fact the result of a single base mutation in the coding sequence that causes an amino acid substitution. This is the second demonstration of a mutation in the COUA5 gene in Alport syndrome leading to changes in the

primary structure of the highly GBM-specific cu5(IV) collagen chain. The first mutation identified, in Utah Kindred EP, involved the deletion of exons 5-10, leading to shortening of the (u5(IV) chain by 240 amino acid residues, 47 in the NC domain, and 193 in the collagenous domain (Barker et al., 1990). The point mutation analyzed in this study involves an evolutionarily conserved cysteine residue in the NC domain and it can have severe effects, both on the triple-helix formation and on intermolecular assem- bly of type IV collagen molecules containing the (u5(IV) chain.

The globular NC domain at the carboxyl-terminal end of all type IV collagen (Y chains is thought to serve the same two major functions: First, it is believed to

CTGCAG Pst I

aqATCT

Bgl II

FIG. 4. Location of the point mutation in the COIAA5 gene in Kindred P. The single base change is located at the 5’ end of exon 3 (counted from the 3’end of the gene). The mutation changes a codon of cysteine to the codon of serine and generates both new PstI and BglII restriction sites. Capital letters, exon sequences; lower case letters, intron sequences.

TYPE IV COLLAGEN DEFECT IN ALPORT SYNDROME 15

H2N+------- Collegsnous

domain

NC-domaln

1

Tentative rearrangement of dlsulflde bonds and head-to-head dimerlzatlon of NC-domalns of two separate molecules

FIG. 5. Schematic illustration of normal intrachain disulfide bonds in the NC domain of a single type IV collagen cul chain and potential rearrangement of inter-chain disulfide bonds after dimerization of chains from two separate triple-helical molecules via their NC domains. For simplification the bonds are shown only for single chains of the triple-helical molecule. The NC domain is indicated by a solid black line and the collagenous domain (carboxyl-terminal end) by a dotted line. The positions of cysteine indicate the number of residues from the amino-terminal end of the NC domain. The exact organization of the disulfide bond between pairs 20/53 and 108/111 and pairs 130/164 and 222/225 has not been elucidated. Modified from (31).

be crucial for the correct alignment of three LY chains prior to their folding into the triple-helical conforma- tion in a manner similar to that in type I collagen (Doege and Fessler, 1986). Second, it is essential for the head-to-head cross-linking between the carboxyl terminals of two triple-helical molecules during for- mation of the type IV collagen network (Weber et al., 1984; Siebold et al., 1988). These functions require the right conformation of the NC domain that is main- tained by disulfide bonds. The primary structure of the NC domain of type IV collagen CY chains has been remarkably well conserved during evolution, as shown for the chains from distant species such as Dro- sophila, Caenorhabditis elegans, and mammals (Blumberg et al., 1987; Cecchini et al., 1987; Hostikka

and Tryggvason, 1988; Guo and Kramer, 1989). This is not surprising since the NC domain of the type IV collagen LY chains has similar biological functions re- gardless of species or kinds of chain. The length of the NC domain varies from 227 to 231 in the mammalian (uB(IV1 and Drosophila w(IV) chains, respectively. The domain has two homologous subdomains each containing six cysteine residues in identical positions (see Fig. 5). The cysteine residues participate in in- trachain disulfide bonds, but they have not been shown to form interchain bonds between the three chains of the triple-helical molecule (Fessler and Fessler, 1982; Weber et al., 1984). After the head-to- head aggregation of two molecules, a rearrangement of the disulfide bonds occurs and intermolecular

16 ZHOU ET AL.

HzN4------- Collagenous

domaln

A NC-domain

H,N.------ Collagenous

domaln

NC-domain

FIG. 6. Hypothetical consequences of the COL4A5 gene mutation in Kindred P. The mutation changes the cysteine at position 168 to serine, abolishing a disultide bond connecting cysteine 108 to either cysteine 53 (A) or cysteine 20 (B) (31). This, in turn, causes a change in the conformation of the NC domain with possible rearrangement of intrachain bonds and subsequent interference of helix formation. The mutation may also seriously influence the rearrangement of the disulfide bonds and the formation of interchain bonds, with the simplest change reducing the number of cross-links from 8 to 6. However, the mutation might interfere more seriously with dimerization by even completely inhibiting the format.ion of interchain cross-links between NC domains.

cross-links are formed (Siebold et al., 1988). The ar- rangement of intrachain disulfide bonds has been elu- cidated by Siebold et al. (1988). Each of the two homol- ogous subdomains within a single (Y chain is stabilized by an identical set of disulfide bonds as illustrated in Fig. 5. Accordingly, cysteines at positions 20 and 53 (as counted from the amino-terminal end of the NC domain) are connected with cysteines at positions 108 and 111, but the exact organization of this pair of bonds is not known. The third bond is formed be- tween cysteines 65 and 71. A similar arrangement is present in the other subdomain with cysteines 130 and 164 connected with cysteines 222 and 225, and the cysteine residues 176 and 182 are bound to each other. The rearrangement that occurs after the head- to-head aggregation involves the cysteine residues at positions 20,53, 108, 111, 130,164,222, and 225, each of which forms a new disulfide bond with cysteine residues in the NC domain of another triple-helical molecule as illustrated in Fig. 5 (Siebold et al., 1988).

The mutation in Kindred P that changes cysteine residue 108 to serine would abolish one of the intra- chain disulfide bonds. The severity of the mutation may depend on whether cysteine 108 is bound to cys- teine 20 or 53, which has not been clarified (see Fig. 6). In any case, the mutation would undoubtedly alter the conformation of the NC domain. To date, the a5( IV) chain itself or molecules containing it have not been isolated and it is not known whether the chain is present in homo- or heterotrimers. However, the con- sequences of the mutation can be surmised on the basis of the knowledge available on other type IV col- lagen N chains. First, the conformational changes caused by the loss of one disulfide bond could influ- ence the formation of the triple helix, whether the molecule is a homotrimer or heterotrimer. The rate of helix formation could be slowed down or abolished completely, or the alignment of Gly-X-Y repeats might not be in the right register. Second, even though a triple-helical molecule might be formed and

TYPE IV COLLAGEN DEFECT IN

secreted from the cell, the loss of a cysteine could dra- matically affect the formation of cross-links between two molecules. It is possible that some cross-links can be formed but they are fewer and, therefore, probably weaker than normal. The cross-links may also be formed between wrong cysteine residues or they may not occur at all. In any case, the consequences of the mutation can almost certainly explain the generation of the disrupted glomerular basement membrane found in affected males of Kindred P. Further under- standing of the basement membrane defects that oc- cur in Alport syndrome will require the characteriza- tion of additional mutations.

From a clinical point of view, the major benefit of the present work is that we now have a specific diag- nostic DNA probe for this particular mutation. There- fore, it can be applied directly for pre- or postnatal diagnosis in the large kindred known to have this mu- tation. It is notable that all of the three mutations identified to date have been found only once in the set of 20 independent families that we have studied thus far. Although we cannot at present rule out the exis- tence of a “major mutation” accounting for a signifi- cant fraction of the disease, it appears likely that Al- port syndrome is highly heterogeneous, involving mulple different mutations in the COL4A5 gene in the same way that osteogenesis imperfecta is character- ized by many different mutations in the type I colla- gen genes (Prockop and Kivirikko, 1984; Byers, 1989). In addition, the possibility exists that a second type IV collagen gene is located head-to-head on the COLIAS locus, in an arrangement similar to that of the COL-IAI and COL4A2 genes on chromosome 13 (Piischl et al., 1988; Soininen et al., 1988). Mutations in such a gene might cause some instances of Alport as well as other related genetic diseases.

ACKNOWLEDGMENTS

This work was supported by grants from the Academy of Fin- land (K.T.1 and the Sigrid Juselius Foundation (K.T.), by NIH Grant DK 39497 (C.L.A.1, by DOE Grant DE-FG02-88ER 60689 (D.F.B.), and by grants from the Hereditary Nephritis Foundation (C.L.A.1. C.L.A. acknowledges further support from PHS Grant MO1 RR 00064 from the Division of Research Resources.

REFERENCES

1. ALPORT, A. C. (1927). Hereditary familial congenital haem- orrhagic nephritis. Brit. Med. J. (Clin. Kes.1 1: 5044506.

2. ATKIN, C. L., GREGORY, M. C., AND BORDER, W. A. (1988a). Alport syndrome. In “Diseases of the Kidney” (R. W. Schrier and C. W. Gottschalk, Eds.), 4th ed., Chap. 19, pp. 617-641, Little, Brown, Boston.

3. ATKIN, C. L., HASSTEDT, S. J., MENLOVE, L., CANNON, L., KIRSCHNER, N., SCHWARTZ, C.. NGUYEN, K., AND SKOLNICK,

4.

5.

6.

7.

a.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

ALPORT SYNDROME 17

M. (1988b). Mapping of Alport syndrome to the long arm of the X chromosome. Amer. J. Hum. Genet. 42: 249-255.

BARKER, D., HOSTIKKA, S. L., ZHOU, d., CHOW, L. T.. OLI- PHANT, A. R., GERKEN, S. C., GREGORY, M. C., SKOLNICK, M. H., ATKIN, C. L., AND TRYGGVASON, K. (19901. Identifica- tion of mutations in the COL4A5 collagen gene in Alport syn- drome. Science 248: 1224-1227.

BARKER, D. F., DIEZ-BAND, J. N., GOLDGAR, D. E., NGUYEN, K. N., TURCO, A. E., DONALDSON, C. W., WILLARD, H. F., SKOLNICK, M. H., AND ATKIN, C. L. (1989). Mapping of the gene for Alport syndrome (ASLN) within a physically and genetically mapped cluster of X RFLP markers. Human gene mapping 10. Cytogenet. Cell Genet. 51(1-41: 957.

BLUMBERG, B., MACKRELL, A. ,J., OLSON, P. F., KURKINEN, M., MONSON, J. M., NATZLE, J. E., AND FESSLER, J. H. (1987). Basement membrane procollagen IV and its specialized car- boxy1 domain are conserved in Drosophila, mouse and human. J. Riol. Chem. 262: 5947-5950.

BRUNNER, H., SCHR~DER, C., VAN BENNEKOM, C., LAMBER- MON, E., TUERLINGS, J., MENZEL, D., OLBING, H., MONNENS, L., WIERINGA, B., AND ROPERS, H.-H. (1988). Localization of the gene for X-linked Alport’s syndrome. Kidney Int. 34: 507-510.

BUTKOWSKI, R. J., LANGEVELD, J. P. M., WIESLANDER, J., HAMILTON, .J., AND HUDSON, B. G. (1987). Localization of the Goodpasture epitope to a novel chain of basement membrane collagen. J. Riol. Chem. 262: 7874-7877.

BYERS, P. H. (1989). Inherited disorders of collagen gene structure and expression. Amer. J. Med. Genet. 34: 72-80.

CECCHINI, J.-B., KNIBIEHLER, B., MIRRE, C., AND LE PARCO, Y. (1987). Evidence for a type-IV-related collagen in Drosoph- ila melanogastert Evolutionary constancy of the carboxyl-ter- minal noncollagenous domain. Eur. J. Biochem. 165: 587- 593.

DOEGE, K. J., AND FESSLER, J. H. (1986). Folding of carboxyl domain and assembly of procollagen I. J. Biol. Chem. 261: 8924-8935.

ERLICH, H. A., GELFAND, D. H., AND SAIKI, R. K. t 1988). Spe- cific DNA amplification. Nature (London) 314: 461-462.

FESSLER, L. I., AND FESSLER, J. H. (1982). Identification of the carboxyl peptides of mouse procollagen IV and its impli- cations for the assembly and structure of basement mem- brane procollagen. J. Riol. Chem. 257: 9804-9810.

FLINTER, F. A., ABBS, S., AND BOBROW, M. (1989). Localiza- tion of the gene for classic Alport syndrome. Genomics 4: 335- 338.

GUNWAR, S., SAUS, J., NOELKEN, M. E., AND HUDSON, B. G. t 1990). Glomerular basement membrane: Identification of a fourth chain, tu4, of type IV collagen. J. &of. Chem. 265: 5466-5469.

GUO, X., AND KRAMER, J. M. (1989). The two C’aenorhabditis elegans basement membrane (type IV) collagen genes are lo- cated on separate chromosomes. J. Biol. Chem. 264: 17,574- 17,582.

HASSTEDT, S. J., AND ATKIN, C. L. (1983). X-linked inheri- tance of Alport syndrome: Family P revisited. Amer. J. Hum. &net. 35: 1241-1251.

HOSTIKKA, S. L., AND TRYGGVASON, K. (1988). The complete primary structure of the cu2 chain of human type IV collagen and comparison with the al(IV1 chain. J. Hiol. (‘hem. 263: 19,488-79,493.

HOSTIKKA, S. L., EDDY, R. L., BYERS, M. G., H~YHTY;~, M., SHOWS, T. B., AND TRYGGVASON, K. (1990). Identification of

ZHOU ET AL 18

20.

21.

22.

23.

24.

25.

26.

27.

28.

a distinct type IV collagen (Y chain with restricted kidney dis- tribution and assignment of the gene to the locus of X chro- mosome-linked Alport syndrome. Proc. N&l. Acad. Sci. USA 87: 1606-1610.

KASHTAN, C. E., ATKIN, C. I,., GREGORY, M. C., AND MI- CHAEL, A. F. (1989). Identification of variant Alport pheno- types using an Alport-specific antibody probe. Kidney Int. 36: 669-674.

KASHTAN, C., FISH, A. J., KLEPPEL, M., YOSHIOKA, Ii.. AND MICHAEL, A. F. (1986). Nephritogenic antigen determinants in epidermal and renal basement membranes of kindreds with Alport-type familial nephritis. J. Clin. Inuest. 78: 10X- 1044.

KLEPPEL, M. M.. KASHTAN, C., SANTI, P. A., WIESLANDER, J., AND MICHAEL, A. F. (1989). Distribution of familial nephritis antigen in normal tissue and renal basement mem- branes of patients with homozygous and heterozygous Alport familial nephritis. Relationship of familial nephritis and Goodpasture antigens to novel collagen chains and type IV collagen. Lab. Znuest. 61: 278-289.

MESSING, ,J. (1983). New Ml3 vectors for cloning. In “Meth- ods in Enzymology” (R. Wu, L. Grossman, and K. Moldave, Eds.), Vol. 101, pp. 20-78, Academic Press, San Diego.

MURRAY, J. A. H. (1986). HCC ligation: Rapid and specific DNA construction with blunt ended DNA fragments. Nucleic Acids. Res. 14: 10.118.

PROCKOP, D. .J., AND KIVIRIKKO, K. 1. (1984). Heritable dis- eases of collagen. N. En,&. J. M&. 311: 376-386.

PBSCHL, E., POLLNER, R., AND KUHN, K. (1988). The genes for the al(W) and c~2(1V) chains of human basement mem- brane collagen type IV are arranged head-to-head and sepa- rated by a bidirectional promoter of unique structure. EMBO J. 7: 2687-2695.

SANGER, F., NICKLEN, S.. AND COULSON, A. R. (1977). DNA sequencing with chain terminating inhibitors. Proc. Nat/. Acad. Sci. 1 ISA 74: 5463-5467.

SAUS, d., WIESLANDER, d., LANGEVELD, J. P. M., QUINONES, S., AND HUDSON, B. G. (1988). Identification of the Goodpas-

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

ture antigen as the tu3(IV) chain of collagen IV. J. Biol. (‘h(,m. 263: 13,374-13,380.

SAVAGE. C. 0. S., NOEL, L.-H., CRUTCHER, E., PRICE, S. R. G., GRUNFELD, d. P., AND LOCKWOOD, C. M. (1989). Hereditary nephritis: Immunoblotting st,udies of the glomer- ular basement membrane. Lab. Inuest. 60: 613-618.

SCHARF, S. J., HORN, G. T., AND ERLICH, H. A. (1986). Direct cloning and sequence analysis of enzymatically amplified ge- nomic sequences. Science 233: 7076&1078.

SIEBOLD, B.. DEUTZMANN, R., AND KUHN, K. (1988). The ar- rangement of intra- and intermolecular disulfide bonds in the carhoxyterminal, non-collagenous aggregation and cross- linking domain of basement-membrane t,ype IV collagen. Eur. J. Riochem. 176: 617-624.

SOININEN, R., HUOTARI, M., HOSTIKKA, S. L.. PROCKOP, D. .J.. AND TRYGGVASON, K. ( 1988). The structural genes for (~1 and tu2 chains of human t.ype IV collagen are divergently encoded on opposite DNA strands and have an overlapping promoter region. J. Rio/. C’hem. 263: 17.217-17,220.

SPEAR, G. S. (1973). Alport’s syndrome: A consideration of pathogenesis. C/in. Ne&rol. 1: 336-337.

TIMPL, R., AND DZIADEK, M. (1986). Structure, development and molecular pathology of basement membranes. Znt. Ret). Exp. Potho/. 29: 304~306.

WEBER, S.. ENGEL. J., WIEDEMANN, H.. GLANVILLE, R. W., AND TIMPI,, R. (1984). Subunit structure and assembly of the globular domain of basement-membrane collagen type IV. Eur. J. Biochem. 139: 401-410.

WESTER, D. C. (1990). “Sensorineural Hearing Loss in Al- port Syndrome.” Doctoral thesis, Department of Communica- tion Disorders, University of Utah, Salt Lake City, 1990.

WILLIAMS, J. F. (1989). Optimization strategies for the poly- merase chain reaction. Rio7’echniques 7: 762-767.

YURCHENCO, P. D. (1990). Assembly of basement mem- branes. Ann. N. 1.. Acad. Sci. 580: 19%21:1.

THOU, ,J., HOSTIKKA, S. L., CHOW, L. T., AND TRYGGVASON, K. (1991). Characterization of the 3’ half of the human type IV collagen tu5 gene that is affected in Alport syndrome. &no- micx 9:1-g.