Proc. Natl. Acad. Sci. USA Vol. 87, pp. 1606-1610, February 1990 Biochemistry Identification of a distinct type IV a chain with restricted kidney distribution and assignment of its gene to the locus of X chromosome-linked (basement membranes/glomerulus/hereditary nephritis/) SIRKKA LIISA HOSTIKKA*, ROGER L. EDDYt, MARY G. BYERSt, MATTI HOYHTYA*, THOMAS B. SHOWSt, AND KARL TRYGGVASON* *Biocenter and Department of Biochemistry, University of Oulu, 90570 Oulu, Finland; and tDepartment of Human Genetics, Roswell Memorial Park Institute, New York State Department of Health, Buffalo, NY 14263 Communicated by Marilyn Gist Farquhar, November 6, 1989

ABSTRACT We have identified and extensively charac- mosome-linked disease primarily characterized by hematuria terized a type IV collagen a chain, referred to as a5(IV). Four and patchy splitting of the glomerular overlapping cDNA clones isolated contain an open reading (GBM) (10-12). There is evidence that, at least in certain frame for 543 amino acid residues ofthe carboxyl-terminal end kindreds, the pathological mechanism is an alteration or even of a collagenous domain, a 229-residue carboxyl-terminal the complete absence of a type IV collagen-like a chain (13, noncollagenous domain, and 1201 base pairs coding for a 3' 14). However, the defect in the Alport syndrome cannot be untranslated region. The collagenous Gly-Xaa-Yaa repeat se- in the al(IV) or a2(IV) genes since they have been mapped quence has five imperfections that coincide with those in the to 13qter (15, 16). corresponding region of the al(IV) chain. The noncollagenous domain has 12 conserved cysteine residues and 83% and 63% sequence identity with the noncollagenous domains of the MATERIALS AND METHODS al(IV) and a2(IV) chains, respectively. The a5(IV) chain has Isolation and Characterization of cDNA Clones. A human less sequence identity with the putative bovine a3(IV) and placenta cDNA library in Agtll (Clontech) was screened (3) a4(IV) chains. Antiserum against an a5(IV) synthetic peptide using an oligonucleotide mixture TG(TC)CA(AG)GT(ACG- stained a polypeptide chain of about 185 kDa by immunoblot T)TG(TC)ATG (32 permutations) that codes for a type IV analysis and immunolocalization of the chain in human kidney collagen NC domain consensus sequence Cys-Gln-Val- was almost completely restricted to the glomerulus. The gene Cys-Met. The filters were rehybridized with cDNA clones was assigned to the Xq22 locus by somatic cell hybrids and in HT-21 (17) and HD-4 (18), which code for the human al(IV) situ hybridization. This may be identical or close to the locus of and a2(IV) chains, respectively. Clones that were positive for the X chromosome-linked Alport syndrome that is believed to the oligonucleotide mixture but not for the al(IV) and a2(IV) be a type IV . cDNAs were isolated, subcloned into M13 (19) vector, and sequenced (20) for initial characterization. For Nothern blot Type IV collagen, the major component of basement mem- analysis (21), total RNA was isolated from frozen tissues branes, consists primarily of al(IV) and a2(IV) chains (1). using an acid guanidinium thiocyanate/phenol/chloroform The complete primary structure of the al(IV) and a2(IV) extraction procedure (22). chains from man (2, 3) and mouse (4, 5) and an a(IV) chain Synthetic Peptides and Preparation of Antibodies to an of Drosophila (6) have been reported. All these chains a5(IV)-Derived Peptide. An a5(IV) peptide Ser-Asp-Met- contain a highly conserved carboxyl-terminal noncollage- Phe-Ser-Lys-Pro-Gln-Ser-Glu and an al(IV) peptide Ser- nous domain (NC domain) of 227-231 residues and a collag- were made. For the enous domain of 1398-1456 residues. Unlike fibrillar colla- Glu-Met-Phe-Lys-Lys-Pro-Thr-Pro-Ser numerous inter- preparation of antiserum, the a5(IV) peptide was coupled to gens, the type IV collagen a chains contain ovalbumin (23) and a rabbit was immunized. ruptions in the otherwise continuous Gly-Xaa-Yaa repeat skin sequence. The existence of two distinct type IV collagen Immunological Methods. For blotting, normal human a3(IV) and a4(IV) chains has also been reported (7, 8), which fibroblasts were extracted with SDS buffer and electropho- indicates that this must have several molecular com- resed under reducing conditions (24). The separated positions. Immunological studies indicate that the major form were transferred to a nitrocellulose filter for immunostaining. of collagen IV containing the al(IV) and a2(IV) chains is a Immunohistological staining was carried out on 4-,um-thick ubiquitous basement membrane component (see ref. 1). methanol-fixed cryosections from an adult human kidney However, this may not always be the case, as the presence obtained at autopsy. Fluorescein isothiocyanate- (Janssen of other type IV collagen a chains in these antigens has not Biochimica) or peroxidase- (Zymed Laboratories) conju- been ruled out. Evidence for region specificity, at least in the gated anti-rabbit IgG was used for immunostaining. Peroxi- kidney, has been obtained with antibodies against the puta- dase activity was visualized using 4-chloro-1-naphthol (25) tive a3(IV) chain (9). and H202 as substrates. Diseases, such as the Goodpasture syndrome and the Chromosomal Assignment. Chromosomal assignment of Alport syndrome, that seem to be associated with type IV the human a5(IV) gene was carried out using a panel con- collagen defects provide models for studies of the minor sisting of 36 human-mouse hybrids derived from 14 unrelated chains. The Goodpasture antigen has been proposed to be the human cell lines and 4 mouse cell lines (26, 27). Southern blot a3(IV) NC domain (9). The Alport syndrome is an X chro- analysis was performed with 10 ,g of HindIII-digested DNA from each cell line by using 32P-labeled MD-6 cDNA as probe The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" Abbreviations: NC, noncollagenous; GBM, glomerular basement in accordance with 18 U.S.C. §1734 solely to indicate this fact. membrane.

1606 Downloaded by guest on September 26, 2021 Biochemistry: Hostikka et al. Proc. Natl. Acad. Sci. USA 87 (1990) 1607

(28). In situ chromosome hybridization was performed as COLLAGENOUS DOMAIN NC-DOMAIN 3' UNTRANSLATED described (29, 30) using 3H-labeled MD-6 cDNA as probe. 5, PL-31 B C M 6 k TAA PC-4 K K T+A AAAAAAA RESULTS k PL-35 K T*A AAAAAATTAAA Isolation of cDNA Clones Coding for a5(IV), Type IV K H Collagen a Chain. The strategy for the isolation of cDNA 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 kb clones coding for potential type IV collagen a chains was to make use of the fact that there are short identical amino acid FIG. 1. Partial restriction map of four overlapping cDNA clones sequences in the carboxyl-terminal NC domain of the human coding for the human a5(IV) chain. The translation stop codon (TAA), two potential polyadenylylation signals (AATATA and AAT- and mouse al(IV) and a2(IV) chains (17, 18, 31, 32) as well TAAA), and a poly(A) tail are indicated. The regions of the mRNA as in the evolutionarily distant Drosophila a(IV) chain (6). It encoded by the clones are indicated at the top. Restriction sites for was concluded that these sequences are essential for the BamHI (B), HinclI (Hc), HindIlI (H), and Kpn I (K) are indicated. structure or function of the protein, and they could, there- fore, be expected to be present in all other potential a chains Comparison of the Human a5(IV) Chain with the al(IV), of type IV collagen. One such sequence is Cys-Gln- a2(IV), a3(IV), and a4(IV) Chains. The 772-residue car- Val-Cys-Met. A 15-mer synthetic oligonucleotide coding for boxyl-terminal end sequence of a5(IV) was compared with this sequence was used as probe for screening of the cDNA the corresponding sequence ofthe human al(IV) and a2(IV) library. chains (Fig. 2). All the chains are remarkably similar but the Primary screening of 360,000 plaque-forming units yielded a5(IV) chain is more closely related to the a 1(IV) chain than 720 duplicate positive signals. Since it could be anticipated to the a2(IV) chain. In the NC domain the sequence identity that a large portion of the signals were derived from al(IV) between the a5(IV) chain and a 1(IV) is 83% whereas it is 63% and a2(IV) cDNA clones, the clones HT-21 (17) and HD-4 with the a2(IV) chain. The NC domain of the a5(IV) and (18) coding for these chains, respectively, were used to al(IV) chains has 229 residues whereas it is 227 residues in rescreen the same filters under stringent hybridization and the a2(IV) chain. The sequence available from the collage- washing conditions (21). This resulted in eight negative nous domain of the a5(IV) chain is 58% identical with the clones that were subcloned into M13 vectors and sequenced corresponding region of the al(IV) chain but only 46% from both ends for initial characterization. Six ofthose coded identical with that of the a2(IV) chain. All the interruptions for a unique type IV collagen-like a chain. The longest clone, in the collagenous domain match in their location between the MD-6 [1.4 kilobases (kb)], encoded a short part of a 3' al(IV) chain and the a5(IV) chain (Fig. 2). untranslated region, a complete NC domain, and a part ofthe Short sequences around the presumable junction between collagenous domain (Fig. 1). the collagenous domain and the NC domain of the putative The MD-6 clone was then used to rescreen the same cDNA bovine a3(IV) and a4(IV) chains have been reported (7, 8). library and this yielded, among others, one 2.0-kb clone Comparison of those with the corresponding region of the (PL-31) that reached 955 base pairs (bp) further upstream and a5(IV) chain shows sequence identity with the a3(IV) ofonly two clones of 2.4 kb (PC-4) and 1.8 kb (PL-35) that coded for 46%, when the figure for the putative a4(IV) chain is only the 3' untranslated region, the NC domain, and a part of the 24%. This extensive difference cannot be explained by in- collagenous domain (Fig. 1). terspecies differences because, for example, the human and Nucleotide and Amino Acid Sequence of the Human a5(IV) murine al(IV) chains have complete sequence homology in Chain. The complete nucleotide sequencer ofthe overlapping this region (4, 17). clones for the type IV a chain we have termed a5(IV) We, therefore, conclude that the cDNA clones described revealed a 2316-bp open reading frame coding for 772 amino here code for a5(IV), a distinct chain of the type IV collagen acid residues and a 1201-bp 3' untranslated region (Fig. 2). a chain family. The PC-4 cDNA clone has 1167 bp of a 3' untranslated region Immunoblot Analysis and Indirect Immunofluorescence sequence and a 22-nucleotide poly(A) tail that is preceded by with Antisera Against an a5(IV) Chain-Derived Peptide. Total two potential polyadenylylation signals, AATTAAA or protein from cultured human fibroblasts was used for immu- AATATA (33), instead of the typical AATAAA consensus noblot analysis. Filter strips containing the blotted protein sequence (Fig. 2). The PL-35 cDNA clone that has no poly(A) were incubated with the antibodies made against the a5(IV)- tail reaches 34 nucleotides further downstream of PC-4, chain-derived peptide. The antiserum stained a band of the indicating that at least two sizes of mature mRNAs exist. same size as the a 1(IV) chain (Fig. 4, lane B). This stain was The cDNA-deduced amino acid sequence contains 543 blocked by preincubation of the antiserum with the antigen residues of a collagenous Gly-Xaa-Yaa repeat sequence that (Fig. 4, lane C). The stain was not blocked by preincubation has five imperfections of 2-5 residues and a 229-residue of the antiserum with a corresponding al(IV) peptide con- carboxyl-terminal NC domain. The NC domain contains 12 jugate (Fig. 4, lane D). Preimmune serum was completely cysteine residues. The conserved Cys-Gln-Val-Cys-Met se- negative (Fig. 4, lane A). Nonspecific stain of a band about quence was located at the same site as in the other known the size of the a2(IV) chain and of lower molecular weight type IV collagen a chains (Fig. 2). Northern blot analysis with could be seen in all samples (lanes A-D). In lane B some total RNA from human embryonic kidney and lung demon- possible degradation products are also visible. strated that the gene codes for a transcript(s) of about 6.5 kb Immunofluorescence studies on cryosections from a hu- and the signal is about three times more intense with RNA man adult kidney with the a5(IV) peptide antiserum gave a from kidney than from lung (Fig. 3). Using kidney total RNA, reaction that was highly restricted to the GBM whereas the the ratio of the bands was about 40:20:1, respectively, with basement membrane of the Bowman's capsule was com- al(IV), a2(IV), and a5(IV) cDNA probes of similar size and pletely negative (Fig. SA). Furthermore, little if any reaction specific activity (data not shown). Therefore, the a5(IV) was seen in the tubular or vascular basement membranes. No chain is a minor basement membrane component in whole reaction was obtained with the preimmune serum (data not kidney. shown). In contrast, antiserum prepared against the NC domain of type IV collagen isolated from human placenta MThe sequence reported in this paper has been deposited in the showed a strong stain of the basement membranes of Bow- GenBank data base (accession no. M31115). man's capsule and tubuli as well as the GBM (Fig. SB). Downloaded by guest on September 26, 2021 1608 Biochemistry: Hostikka et al. Proc. Natl. Acad. Sci. USA 87 (1990)

G R S G V P G L KGD D G L Q G Q P G L P G P T G EK G S K G E P G L P G PP G P MID P N L LIG S 897 -A. P -L --E - - - H -F P -S S -P R -D P -L - - D --D V - -- - K - - S -I- K V D M - - M 938 --P -F --S --E A -F F -I ---K -L A --P - F --S R - D ------P IF - V I - P G M- 148 ~ A AAGGT LI 50 K G E KG E P G L P G I P G V S G P KG Y Q GL P GD P G Q P G L S G Q P G L P G PP G P K GN P 947 I-Q - - D Q EK Q I -P I -E - - S R -D --T --V - - K D - - A -Q --Q D - 987 - D I -- - K -D E- P M -L K -Y L - A K -I Q-M -- I - - --- I - - R --H I[- V K 295 99 G L P G Q P G L I G PP G LK G T I GD M G F P G P Q G V E G P P G P S G V P G Q P G S P G L P G Q 996 - I S- T --A P -L --P -- S V -G - - L - - T P -E K- V -- I P- P Q -S - - L -- D K -A 1037 -D I -V --I P -L --F P - V A -P P - I T - F P -F I -S R - D K -A --R A - L Y - E I -A 445 149 K GD K G D P G I S GL P GL P G P K G E P G L P G Y P G N P G I K G S V G D P G L P G L P 1046 - - E - - Q A -P P G -I - --R -E - - D Q -I A -F - - S - - E --E K -S I -I --M - 1087 T --F --I -D T I lLP-R - - - K -E R - T T -I --L K - FF - E --T E --I -F --I T 589 197 G T P G A K G Q P G L P G F P G T P G PP G P K G I S G P P G N P G L P G E P G P V GG GG H P G Q 1094 -S - - L - - S - - S V -Y - - S --L - - E - - D K -L - - L D - I - - V K -E A -L P -T --P 1137 -V T - V Q - P - - - K -Q T - F --L T --P - S Q -E L - R I - - - - G K -D D -W P -A --L 739 247 P G P P G E KG K P G Q DG I P G P A Q K G E P G Q P N P G P P G L P GG L S G Q KG D G G 1144 T --A -Q - - E -- S - - -- - S - - E - - -- - L - Rl- F --F --A K -D K -S - - E V - 1187 --F --L R - I R - L H - L --T K - F P - S -- S D -D --F --P- -E R- D P - E A 886 296 L P G I P GN P G L P G P K G E P G F H G F P G V Q G P P G P P G S P G P A G P K G N P G P Q 1193 F --L A -S --I - - S - - - Q -- M -P --P - - Q - - L --- - - H - ----D R - - - 1237 - P V -V -- Q K - D Q - A - -E R -P --S P - L Q - F -- I TIP--S N I S - A P - D K - A P 1033 345 G P P G R P G L P G P E G PP G L P G N G G I KG E K G N P G Q P G L P G L P G L K G D Q G P P G L 1242 -Q - - L ------M ------I D - V - - D - - -- - W - - A - - V - - P - - - P - F Q - M 1287 -I F - L K - Y R -- P - - - - S AIALP S - - D T - -- - A - - T - - T K - W A - - S - - Q-R 1183 395 Q GN P G R P G L NG M K G D P G L P G V P G F P G M K G P S G V P G S A G P E G E P G L I G P P G 1292 P - I G - S - - I T - S - -- M - P ------Q - P - - L P - L Q - I K - D Q - D Q - V P - A K - 1337 P - V F - L - - E K - P R - E Q - F M - N T - P T - AV - D R - P K - P K - D P - F -- A P - T V - 1333 445 PP G L P G P S G Q S G D A G P P G I P G Q P G L KGL P G P Q G P Q G L P G P T G P P G D 1342 L - - P - P - P YID? - -I - E P - L - - P E - P ------Q - L P - - K - Q Q - V - - L V - I 1387 A -- I A - I PIOQ K I A - Q P - T V - - Q - R R - P - - A P - E I - - - - - P - E - - F R - A -- K 1483 495 P G R N G L P G F D G A GG R K G D P G L P G Q P G T R G L D G P P G P D G L Q G P P G P P G T S S 1392 - - P P - I ------P - Q - - E M - P A - P T - P - - F P ------P - S M - - - - - P - 1437 A - P Q-R G - V S - F R - - E - P I - H Q - P I - Q E - A -- R P - S P - L - - M - - R f 1633 545 V A H G F L I T R H S Q TTD A P QIE1P Q G T L Q V Y E G F S L L Y V Q G N K R A H G Q D L G T A G 1442 -D --- - V ------I - D - -_-S - - K I L - H - Y ------E ------1486 -S I - Y - L V K - - -- D Q E - M - V - M N K L W S - Y - - -- F E - Q E K -- N - - - - L - - 1783 595 S L R R F S T M P F M F N I N N V N F A S R N D Y S Y W L S T P E P M P M S M Q P L K G Q S I 1492 - - K ------L ------A - I T - E N - -_ - -- -U_ U 1536 A ------L Y P G D - YY - - - - - K ------T A - L-- - - V A E D E - 1933 645 Q P F S R A V E A P AV V I A V HS Q T I Q I P H P Q G W D S L W I G Y S F M M H T S A G A 1542 R - - ______----M - M ------P _-S - - S ------V ------1584 K - y ____S - _- - - - I A ------D V S - -- _-A - - R ------L - -- A - - D 2083 695 E G S G Q A L A SP G S L E E F R S A P F I E H G R G T N Y Y A N S Y S F W L A T V D V S D 1592 _ - - _ _ - -- - - A ------I E R - E 1634 - - G - - S - V - - - - _- - D - - A T - - - - _N - G - -- _H - - - - K - - - - - T - I P E Q S 2230 744 M F S K P Q S E T L K A G D LR T R I S R Q V Q M K R T * 772 a5(IV) 1641 - - K - - T P S .- - - E --- H V - - _ _- - R - - * 1669 al(IV) 1684 F Q G S - S A D - - - - - L I - H - -- -- N L * 1712 u2 (IV) 2380 2530 2680 2830 2980 __AA__h'ILAVALTnr_TCP- 3130 3280 'im~ ~ ~ ~ ~ p 3430

FIG. 2. Nucleotide sequence and the derived amino acid sequence of a5(IV) chain coding cDNA clones and comparison with the amino acid sequence of the human a1(IV) and a2(IV) chains. Rows: 1, nucleotide sequence numbered beginning from the 5' end; 2, deduced amino acid sequence of the a5(IV) chain; 3, corresponding sequence of the al(IV) chain where it differs from the a5(IV) chain; 4, sequence of the a2(IV) chain where it differs from the a5(IV) chain. The cysteine residues are circled and the interruptions in the collagenous Gly-Xaa-Yaa repeat sequence are boxed. The conserved sequence Cys-Gln-Val-Cys-Met (CQVCM) used for design of screening primers as well as the two sequences used to make synthetic peptides specific for the a5(IV) and al(IV) chains are underlined. The translation stop codon (TAA) is indicated by an asterisk and the potential polyadenylylation signals are indicated by double underlines. The 3' end of the PC-4 cDNA with a poly(A) tail is shown below the 3' end of the entire nucleotide sequence. The single-letter amino acid code is used. Chromosomal Assignment. Chromosomal location of the gation of human COLAS in the cell hybrid panel cocorre- human a5(IV) gene (COL4AS) was determined by hybrid- lated with the distribution of human chromosome X. All other ization of the 1400-bp MD-6 cDNA clone to DNA from a chromosomes segregated discordantly with COMAS. An panel of 36 human-mouse somatic cell hybrids. The segre- examination of cell hybrids retaining different human X- Downloaded by guest on September 26, 2021 Biochemistry: Hostikka et al. Proc. Natl. Acad. Sci. USA 87 (1990) 1609

kb 23.1- 9I4- 6.6- 4.3 -

2.3 - 2.0 -

FIG. 3. Northern blot analysis oftotal RNA from human embryonic kidney and lung tissues. RNA (10,ug) was separated on a 1.0%o agarose gel, blotted to nitro- cellulose, and hybridized with a 1.2-kb 32P-labeled fragment from the 3' untrans- lated region of the PL-35 cDNA clone. Lanes: 1, lung; 2, kidney. Migration of 1 2 size markers is indicated. autosome translocations with various lengths of the X chro- mosome further localized COLAAS to the Xq22-26 region (cell hybrid data not presented, but are available from au- thor). In situ hybridization confirmed the localization of COLA5 to chromosome Xq22 (Fig. 6). No other chromo- some had counts above background.

DISCUSSION The present work demonstrates the existence ofa previously unknown polypeptide chain ofbasement membrane (type IV) collagen. The amino acid sequence derived from the cDNA clones isolated in this study provided about 50% of the entire amino acid sequence from the carboxyl-terminal end of this FIG. 5. Immunolocalization of the human a5(IV) chain in human chain, which we refer to as a5(IV). The a5(IV) chain is more kidney. (A) The a5(IV) antiserum shows stain highly restricted to the closely related to the al(IV) chain than to the a2(IV) chain. GBM. No stain is seen in the Bowman's capsule basement membrane This is evident from the exactly same size and high sequence and little if any stain can be seen in the tubular basement membrane. identity of the NC domains and also because the locations of No stain was obtained with the preimmune serum (data not shown). all imperfections in the helical domain are conserved. To (B) For comparison, antiserum against the NC domain of human placental type IV collagen shows intense stain in the GBM as well as date, there are only minor protein and sequence data avail- the basement membranes of the Bowman's capsule and tubuli. able from the potential a3(IV) and a4(IV) chains (7, 8). It is (Nikon Optiphot microscope; x 14.) evident from the present study that the sequence of the a5(IV) chain is distinct from the sequences of those chains. Antiserum to a 10-residue synthetic peptide with a se- quence from the NC domain was made. The peptide antigen had a 50% sequence identity with the sequence from the _- - a1 (IV) al(IV) chain, which was the lowest possible sequence iden- - (t2(IV) tity for a 10-peptide sequence. Immunoblot analysis against total protein from human fibroblasts showed that the a5(IV) antiserum stained a band of the same size as the a 1(IV) chain gm (185 kDa). Since the stain could be blocked by preincubation A B _ of the antiserum with the a5(IV)-peptide-ovalbumin antigen but not by preincubation with the a l(IV)-peptide-ovalbumin

., ' zW conjugate we conclude that the antiserum does not detect a 1(IV). A B C D E Interestingly, immunostaining of kidney was restricted to FIG. 4. Immunoblot analysis with antiserum against an a5(IV) the GBM, indicating a highly tissue-specific distribution of chain synthetic peptide. Total cellular protein from cultured human the a5(IV) chain. There was no apparent stain in the tubular skin fibroblasts was size fractionated by SDS/PAGE, transferred to basement membranes or in the Bowman's capsule. Antibod- nitrocellulose, and stained with antiserum. Lanes: A, preimmune ies against the putative NC domain of the a3(IV) and a4(IV) serum; B, a5(IV) chain antiserum detects a 185-kDa band; C, chains have shown strong reaction in the GBM and in the preincubation of the a5(IV) peptide antiserum with the antigen (100 Bowman's capsule (9). Reactivity of our a5(IV) chain anti- ,ug/2 ml of 1:100 diluted antiserum) completely blocks the staining bodies with these chains cannot be ruled out. The antiserum reaction; D, preincubation of the a5(IV) peptide antiserum with an human equal amount of a 1(IV) peptide conjugate does not block the staining against placenta type IV collagen stained the GBM, reaction; E, two bands of the al(IV) and a2(IV) chains of 185 kDa Bowman's capsule and tubular basement membranes. The and 170 kDa, respectively, are seen in the [3H]proline-labeled type placental antigen consists presumably mainly of a 1(IV) and IV collagen from Engelbreth-Holm-Swarm (EHS) tumor electro- a2(IV) chains but the presence of other type IV collagen phoresed on the same gel. chains cannot be excluded. Accordingly, the al(IV) and Downloaded by guest on September 26, 2021 1610 Biochemistry: Hostikka et al. Proc. Natl. Acad. Sci. USA 87 (1990)

22.3 so We thank Juha and Leila Risteli for generously providing antibod- 22.2 ies against human type IV collagen and Helena Autio-Harmainen for 22.1 providing the human kidney tissues. This work was supported in part 21.3 by grants from the National Research Council for Natural Sciences, the National p 1 Academy of Finland, The Juselius Foundation, and 22111.4 Institutes of Health and 11.3 0 (GM20454 HD05196). 11.23 0 1. Timpl, R. & Dziadek, M. (1986) Int. Rev. Exp. Pathol. 29, 1-112. 0Sm 2. Soininen, R., Haka-Risku, T., Prockop, D. J. & Tryggvason, K. 12 (1987) FEBS Lett. 225, 188-194. 13 FIG. 6. Idiogram of chromo- 3. Hostikka, S. L. & Tryggvason, K. (1988) J. Biol. Chem. 263, 21.1 s some X showing silver grain dis- 19488-19493. 21.2 s tribution using the COL4A5 4. Muthukumaran, G., Blumberg, B. & Kurkinen, M. (1989) J. Biol. 21.3 probe. One hundred metaphases Chem. 264, 6310-6317. 5. Saus, J., Quinones, S., MacKrell, A., Blumberg, B., Muthukuma- q _ggg q 23 domly examined. The distribution ran, G., Pihlajaniemi, T. & Kurkinen, M. (1989) J. Biol. Chem. 264, 24 of chromosomal silver grains iden- 6318-6324. J. 25 tified a region on the long arm of 6. Blumberg, B., MacKrell, A. J. & Fessler, J. H. (1988) Biol. chromosome X that encodes Chem. 263, 18328-18337. 26 COL4AS. The average number of 7. Butkowski, R. J., Langeveld, J. P. M., Wieslander, J., Hamilton, J. 27 grains was 1.7 per metaphase and & Hudson, B. G. (1987) J. Biol. Chem. 262, 7874-7877. 28 8.9% of the total grains were lo- 8. Saus, J., Wieslander, J., Langeveld, J. P. M., Quinones, S. & cated at Xq22 and 37.5% of the Hudson, B. G. (1988) J. Biol. Chem. 263, 13374-13380. grains on chromosome X were lo- 9. Butkowski, R. J., Wieslander, J., Kleppel, M., Michael, A. F. & x cated at Xq22. Fish, A. J. (1989) Kidney Int. 35, 1195-1202. 10. Atkin, C. L., Hasstedt, S. J., Menlove, L., Cannon, L., Kirschner, a2(IV) chains might also have various distributions in dif- N., Schwartz, C., Nguyen, K. & Skolnick, M. (1988) Am. J. Hum. Genet. 42, 249-255. To date there are no ferent basement membranes of the body. 11. Flinter, F. A., Cameron, J. S., Chantler, C., Houston, I. & Bobrow, immunohistological data on the tissue specificity of a1(IV) M. (1988) Lancet, 1005-1007. and a2(IV). Thus type IV collagen appears to be a consid- 12. Brunner, H., Schroder, C., van Bennekom, C., Lambermon, E., erably more complex protein than previously thought with Tuerlings, J., Menzel, D., Olbing, H., Monnens, L., Wieringa, B. & tissue and functions. Ropers, H.-H. (1988) Kidney Int. 34, 507-510. various chain compositions, specificity, 13. Kleppel, M. M., Kashtan, C. E., Butkowski, R. J., Fish, A. J. & The results of this study establish that the human a5(IV) Michael, A. F. (1987) J. Clin. Invest. 80, 263-266. chain gene (COL4AS) is on chromosome X at band q22, 14. Kashtan, C., Fish, A. J., Kleppel, M., Yoshioka, K. & Michael, which is particularly intriguing with respect to the Alport A. F. (1986) J. Clin. Invest. 78, 1035-1044. This disease has four X chromosome-linked 15. Boyd, C. D., Weliky, K., Toth-Fejel, S. E., Deak, S. B., Chris- syndrome. phe- tiano, A. M., Mackenzie, J. W., Sandell, L. J., Tryggvason, K. & notypes of which at least two have been mapped to Xq22-24 Magenis, E. (1986) Hum. Genet. 74, 121-125. in three studies (10, 12). As described above, the GBM is 16. Griffin, C. A., Emanuel, B. S., Hansen, J. R., Canevee, W. K. & structurally defective and some evidence suggests that there Myers, J. C. (1987) Proc. Natl. Acad. Sci. USA 84, 512-516. 17. Pihlajaniemi, T., Tryggvason, K., Myers, J. C., Kurkinen, M., is an alteration or even complete absence of a type IV Lebo, R., Cheung, M.-C., Prockop, D. J. & Boyd, C. D. (1985) J. collagen-like a chain (13). Kashtan et al. (14) have reported Biol. Chem. 260, 7681-7687. that an Alport (X chromosome-linked) patient who obtained 18. Hostikka, S. L., Kurkinen, M. & Tryggvason, K. (1987) FEBS Lett. a renal allograft developed antibodies against a 26-kDa NC 216, 281-286. domain-like component. This is the same size as that of the 19. Messing, J. (1983) Methods Enzymol. 101, 20-78. 20. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. a that the NC domain of the l(IV) chain (14). It is obvious Sci. USA 74, 5463-5467. chain possibly missing in Alport patients is not al(IV) since 21. Maniatis, T., Fritsch, E. & Sambrook, J. (1982) Molecular Cloning: its gene has been localized to 13q. Consequently, the 26-kDa A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring component detected by the Alport antiserum could be the NC Harbor, NY). 22. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162, 156-159. as it has almost the same size as domain of a5(IV), exactly 23. Kagen, A. & Glick, M. (1979) in Methods of Hormone Radioim- that of the al(IV) chain. munoassay, eds. Jaffe, B. B. & Behrman, H. R. (Academic, New Current data strongly suggest that one or even more type York), pp. 328-329. 24. Laemmli, U. K. (1970) Nature (London) 227, 680-685. IV collagen a chains are defective in the GBM of certain patients with X chromosome-linked Alport syndrome. This 25. Moeremans, M., Daneels, G., Van Dijck, A., Langanger, G. & De J. (1984) J. Immunol. Methods 74, 353-360. for Mey, could be due to a mutation involving gene(s) coding type 26. Shows, T., Eddy, R., Haley, L., Byers, M., Henry, M., Fujita, T., IV collagen a chain(s), but no type IV collagen gene has Matsui, H. & Taniguchi, T. (1984) Somat. Cell. Mol. Genet. 10, previously been assigned to this chromosome. Since 315-318. COL4AS is located in the locus of the Alport syndrome, it is 27. Shows, T. B. (1983) in Isoenzymes: Current Topics in Biological and quite possible that it is involved in the generation of one type Medical Research, eds. Rattazzi, M. C., Scandalios, J. G. & Whitt, G. S. (Liss, New York), Vol. 10, pp. 323-339. disease. Southern of the X chromosome-linked However, 28. Naylor, S. L., Sakaguchi, A. Y., Shows, T. B., Law, M. L., Hoe- blot analysis on DNA from one male patient with X chro- ddel, D. V. & Gray, P. W. (1983) J. Exp. Med. 57, 1020-1027. mosome-linked Alport syndrome did not reveal an abnormal 29. Nakai, H., Byers, M. G., Shows, T. B. & Taggart, R. T. (1986) pattern with several restriction enzymes (data not shown), Cytogenet. Cell Genet. 43, 215-217. 30. D. G. & H. P., eds. (1985) ISCN 1985: An no a deletion in the 3' Harnden, Klinger, indicating major rearrangement-e.g., International for Human Cytogenetic Nomenelature another IV System end of the gene. There might also be type collagen (Karger, Basel), pp. 1-117. a chain on this locus in analogy with the genes for the human 31. Oberbaumer, I., Laurent, M., Schwarz, U., Sakurai, Y., Yamada, al(IV) and a2(IV) chains on 13q (34). The definite answer Y., Vogeli, G., Voss, T., Siebold, B., Glanville, R. W. & Kuhn, K. (1985) Eur. J. Biochem. 147, 217-224. about the involvement of the a 5(IV) locus or another possibly 32. U., Oberbaumer, I. & Kuhn, K. (1986) FEBS linked locus for a IV a chain in Alport Schwarz-Magdolen, closely type collagen Lett. 208, 203-207. studies on the syndrome, therefore, requires further detailed 33. Birnstiel, M. L., Busslinger, M. & Strub, K. (1985) Cell 41, 349-359. entire a5(IV) gene locus in both normal individuals and 34. Soininen, R., Huotari, M., Hostikka, S. L., Prockop, D. J. & Alport patients. Tryggvason, K. (1988) J. Biol. Chem. 263, 17217-17220. Downloaded by guest on September 26, 2021