Identification and characterization of Pkhd1, the mouse orthologue of the human ARPKD gene

Identification and Characterization of Pkhd1, the MouseOrthologue of the Human ARPKD Gene

YASUYUKI NAGASAWA,* SONJA MATTHIESEN,† LUIZ F. ONUCHIC,‡

XIAOYING HOU,§ CARSTEN BERGMANN,¶ ERNIE ESQUIVEL,� JAN SENDEREK,¶

ZHIYONG REN,§ RAOUL ZELTNER,� LASZLO FURU,� ELLIS AVNER,#

MARKUS MOSER,† & STEFAN SOMLO,� LISA GUAY-WOODFORD,§

REINHARD BUTTNER,† KLAUS ZERRES,¶ and GREGORY G. GERMINO**Department of Medicine and Genetics, Johns Hopkins University, Baltimore, Maryland; †Institut ofPathology, University of Bonn, Bonn, Germany; ‡Department of Medicine, University of Sao Paulo, SaoPaulo, Brazil; §Department of Medicine and Pediatrics, University of Alabama at Birmingham, Birmingham,Alabama; ¶Institute of Human Genetics, Technical University of Aachen, Aachen, Germany; �Department ofMedicine and Genetics, Yale University, New Haven, Connecticut; #Department of Pediatrics, Rainbow BabiesChildren Hospital, Cleveland, Ohio; &Max-Planck-Institute for Biochemistry, Martinsried, Germany.

Abstract. PKHD1, the gene mutated in human autosomal re-cessive polycystic kidney disease has recently been identified.Its translation products are predicted to belong to a superfamilyof proteins involved in the regulation of cellular adhesion andrepulsion. One notable aspect of the gene is its unusuallycomplex pattern of splicing. This study shows that mousePkhd1 and its translation products have very similar propertiesto its human orthologue. Mouse Pkhd1 extends over approxi-mately 500 kb of genomic DNA, includes a minimum of 68nonoverlapping exons, and exhibits a complex pattern of splic-ing. The longest ORF encodes a protein of 4059aa predicted tohave an N-terminal signal peptide, multiple IPTs and PbH1repeats, a single transmembrane span (TM), and a short cyto-plasmic C-terminus. Although the protein sequence is gener-

ally well conserved (approximately 73% average identity), theC-termini share only 55% identity. The pattern of Pkhd1 ex-pression by in situ hybridization was also examined in devel-oping and adult mouse tissues over a range of ages (E12.5 to 3mo postnatal). High levels of expression were present in renaland biliary tubular structures at all time points examined.Prominent Pkhd1 signals were also found in a number of otherorgans and tissues. Tissue-specific differences in transcriptexpression were revealed through the use of single exonprobes. These data show that key features of human PKHD1are highly conserved in the mouse and suggest that the com-plicated pattern of splicing is likely to be functionallyimportant.

Autosomal recessive polycystic kidney disease (ARPKD)(MIM 263200) is an inherited disorder of the kidney and liverwith an estimated incidence of 1 in 20,000 live births (1).Affected individuals often present in utero with enlarged andechogenic kidneys as well as oligohydramnios secondary to apoor urine output (2). Up to 50% of the affected neonates dieshortly after birth as a result of severe pulmonary hypoplasiaand secondary respiratory failure. Those who survive the peri-natal period express widely variable disease phenotypes, with

severe hypertension, renal insufficiency, and portal hyperten-sion due to portal tract fibrosis as the most common clinicalfeatures (3,4).

Mutations at a single locus, PKHD1 (polycystic kidney andhepatic disease 1), are responsible for all typical forms ofhuman ARPKD. Using a conventional positional cloning ap-proach, we recently identified human PKHD1 and determinedthat it is a novel gene with a minimum of 86 exons that areassembled in a complex pattern of alternative splice variants(5). The longest open reading frame is encoded by a 67-exontranscript and predicted to yield a protein of 4074 amino acidswith multiple IPT domains, PbH1 repeats, and a single TMnear its carboxyl terminus. Multiple other polypeptides, includ-ing a set of secreted products, might also be encoded by thislocus if any of its alternative transcripts are translated. Thepredicted translation products are novel proteins that sharedomain architecture with a superfamily of proteins involved inthe regulation of cell proliferation and cellular adhesion andrepulsion. We named the gene products polyductin.

Contemporaneous with our efforts, another group used adifferent strategy to identify PKHD1 (6). These investigators

Yasuyuki Nagasawa and Sonja Matthiesen contributed equally to this work.Received June 4, 2002. Accepted July 9, 2002.Correspondence to Dr. Klaus Zerres, Institute of Human Genetics, TechnicalUniversity of Aachen, Pauwelsstr. 30, D-52074 Aachen, Germany. Phone:0241-8080-178; Fax: 0241-8082-580; E-mail: [email protected]; or Dr.Gregory G. Germino, Johns Hopkins University School of Medicine, Divisionof Nephrology, Ross 958, 720 Rutland Avenue, Baltimore, MD 21205. Phone:410-614-0089; Fax: 410-614-5129; E-mail: [email protected]

1046-6673/1309-2246Journal of the American Society of NephrologyCopyright © 2002 by the American Society of Nephrology

DOI: 10.1097/01.ASN.0000030392.19694.9D

J Am Soc Nephrol 13: 2246–2258, 2002

discovered that the locus responsible for ARPKD in the pck ratmapped to a region syntenic to that of human ARPKD (7,8).Taking advantage of the directed genetic mapping capability ofthe rat, they refined the position of PKHD1 and identified boththe rat and human genes. They named the predicted translationproduct fibrocystin.

The two studies reported similar findings for many keyfeatures of PKHD1 and its predicted translation product. Bothgroups determined that PKHD1 is a very large gene extendingover �450 kb of genomic DNA. The two reports identifiedessentially the same 4074–amino acid protein as the translationproduct of the longest ORF of PKHD1. Finally, both groupsshowed a diffuse signal strongest in adult and fetal kidney butalso present in liver and pancreas on Northern using PKHD1cDNA fragments as probes. There were some important dif-ferences between the two groups findings, however. WhereasWard et al. (6) described the occasional amplification of cDNAproducts with alternative exon usage, we found that the ma-jority of PKHD1 cDNA fragments generated from adult humankidney were comprised of a variable set of exons. We attrib-uted the diffuse Northern blot signal for PKHD1 to the gene’scomplicated pattern of gene splicing. Given the extraordinaryabundance of alternatively spliced products, we reasoned thatthis property was likely functionally important.

In this report, we describe the complete sequence of themouse homologue of PKHD1 (Pkhd1) and show that it under-goes a similarly complex pattern of splicing. We also present acomprehensive survey of its expression pattern in mouse tis-sues by in situ hybridization and show differential expressionpattern for a subset of exons.

Material and MethodsIsolation of Pkhd1 cDNA

We initially designed oligonucleotide primers on the basis of thesequence of human PKHD1 (5) and mouse ESTs with homology tohuman PKHD1 and used them to amplify cDNA fragments frommouse adult kidney double-stranded cDNA purchased from Clontech(Marathon Ready cDNA; Clontech, Palo Alto, CA). Mouse genomicsequence was used as it became available to design additional primers.Some primer pairs were designed on the basis of the sequence ofPkhd1 as it was identified. The 5' RACE and 3' RACE experimentswere performed according to the manufacturer’s instructions (Clon-tech). All primer sequences used to amplify the set of cDNA productsare shown in Supplementary Information Table S1.

The genomic sequence of Pkhd1 was obtained from both public(Genbank) and proprietary (Celera, Rockville, MD) databases.

Sequence Analysis and Protein ModelingThe genomic structure and orientation of Pkhd1 was established by

aligning the confirmed expressed sequences with the interval genomicsequence using BLAST2. BLAST2 was also used to identify regionsof homology between the mouse and human gene sequences. Allcoding exons, untranslated regions (UTRs), introns, alternatively usedexons, and Celera transcripts hCT1642763 and hCT1646988 weresubject to such analysis. SMART (Simple Modular Architecture Re-search Tool) (9,10) and PROSITE were used to identify domainarchitecture and protein motifs. All analyses were performed usingdefault parameters.

Northern AnalysisProbes were amplified using cloned gene fragments as template,

gel purified, 32P-labeled using the multi-prime method, and hybrid-ized to both commercial (Clontech and RNWAY [Seegene, Seoul,South Korea] mouse adult MTN blots) and self-made Northern blots.For the latter, 10 to 20 �g of total RNA that had been extracted fromfreshly harvested adult tissues using Trizol was loaded in each lane.Some of the blots were prepared using 1 to 2 �g of oligo d(T)-enriched RNA/lane (poly-A kit; Qiagen, Chatsworth, CA). Hybrid-izations were performed at 68°C using ExpressHyb (Clontech) for16 h and washed in 2�SSC/0.05% sodium dodecyl sulfate (SDS) atroom temperature for 40 min and 0.1�SSC/0.1% SDS at 50°C for 40min. Images were obtained using a PhosphorImager (Molecular Dy-namics, Sunnyvale, CA).

In Situ Hybridization of Mouse Pkhd1In situ hybridization to paraffin-embedded tissue sections was

performed as has been previously described (11). Sense and antisense33P-UTP labeled cRNA probes spanning exons 5 and 41 were used asprobes.

ResultsThe Genomic Organization of Pkhd1 and Identificationof its Transcripts

A search of public databases identified multiple mouse ESTswith very high homology to human PKHD1 (Figure 1C) (5).We used the mouse EST and human PKHD1 sequences todesign a series of primers that could be used in conjunctionwith 5' and 3' RACE reactions, to amplify a series of cDNAfragments spanning approximately 13 kb composite lengthfrom adult mouse kidney (Supplementary Information, TableS1). The longest continuous open reading frame (ORF) wasdetermined on the basis of the sequence of a set of sevenoverlapping cDNA fragments (Figure 1). The sequences of allexons were determined by double-strand sequencing of PCRproducts and by comparison with the sequence of mouse ESTsand with the mouse genomic sequence of the interval. Anydiscrepancies were resolved by re-sequencing additionalcDNA clones.

The results are presented in Figure1 and SupplementaryInformation Figure S1. Like its human orthologue (5,6), thetranscript with the longest ORF of mouse Pkhd1 has 67 exonsand is predicted to encode a translation product of 4059 aminoacids. The human and mouse cDNA sequences share an overallidentity of 80% by BLAST2 analysis but have segments withsignificantly higher (i.e., 96%, exon 35) and lower homology(5' and 3' UTR) (Table 1). As was the case for human PKHD1,we identified multiple transcripts using distinct 3' UTRs, eachof which ended with a polyadenylation signal and a polyade-nylation sequence. The longest 3'UTR was approximately 2.5kb in length, approximately 1.3 kb shorter than that reportedfor human PKHD1 by Ward et al. (6). The predicted sizes ofmRNAs that encode the longest ORF of Pkhd1 range betweenapproximately 13 and approximately 15 kb.

We also determined the genomic organization of Pkhd1using both public (Genbank) and private (Celera) sources. Weidentified all splice-site junctions (Supplementary Information,Figure S1) and determined the approximate length of the

J Am Soc Nephrol 13: 2246–2258, 2002 Pkhd1, the Mouse ARPKD Gene 2247

Figure 1. Structure of full length Pkhd1 and its splicing variants. (A) Genomic organization of mouse Pkhd1. Exons identified by a wholenumber are part of the transcript that encodes the longest open reading frame (ORF), and those marked by a number and letter (i.e., 50a) areoverlapping but utilize alternative splice donor or acceptor sites. Exon 55.1 is the only exon that is completely unique and not part of thetranscript encoding the longest ORF. * identifies exons not currently present in mouse genomic databases. The physical scale is indicated abovethe genomic segment representation. Many of the intron sizes are estimated on the basis of the available sequence. (B) Pkhd1 exons identifiedin this study. The set of 68 nonoverlapping exons that spans the full-length of PKHD1 is shown in the top row. Seven additional overlappingexons (gray boxes) that use different splice sites are presented below. The single exon that neither overlaps any of the others nor contributesto the longest ORF is unfilled. (C) Primer sets and Pkhd1 cDNAs. The approximate location of each primer set used to amplify various cDNAsis shown, and a representative set of amplified products is indicated below each schema. White boxes indicate noncoding exons in thecorresponding transcripts, and gray boxes identify exons with alternative boundaries. SC identifies the approximate location of stop codons,and ORF indicates that an open reading frame extends throughout the length of the fragment. The last line identifies the alternative transcriptencoded by IMAGE clones 1481021, 1481154, and 1432609, the sequences of which are nearly identical. The longest potential ORF wasidentified on the basis of the amplified products 1.1, 2.1, 4.1, 5.2, 7.1, 8.2, 9.1, 10.2, the public EST sequence information, and the structureof the longest ORF of human PKHD1.

2248 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 2246–2258, 2002

Table 1. Mouse/human exon comparisons

Exon Number(Mouse)

Length (bp)(Mouse)

Length (bp)(Human) Identity (%)a Exon Number

(Human)

Exon1 139 No similarityExon2 137 No similarityExon3 78 78 91 human Exon3Exon4 142 151 91 human Exon4Exon5 109 109 85 human Exon5Exon6 58 58 79 human Exon6Exon7 79 79 64 human Exon7Exon8 75 75 82 human Exon8Exon9 65 65 81 human Exon9Exon10 40 40 82 human Exon10Exon11 71 71 83 human Exon11Exon12 102 102 85 human Exon12Exon13 96 96 84 human Exon13Exon14 142 142 87 human Exon14Exon15 115 115 89 human Exon15Exon16 279 279 86 human Exon16Exon17 90 90 86 human Exon17Exon18 91 91 76 human Exon18Exon19 143 143 73 human Exon19Exon20 128 128 80 human Exon20Exon21 176 176 78 human Exon21Exon22 139 139 83 human Exon22Exon23 128 128 73 human Exon23Exon24 185 185 82 human Exon24Exon25 123 123 87 human Exon25Exon26 106 106 80 human Exon26Exon27 276 276 78 human Exon27Exon28 131 131 81 human Exon28Exon29 136 136 86 human Exon29Exon30 196 196 71 human Exon30Exon31 68 68 85 human Exon31Exon32 1602 1608 76 human Exon32Exon33 144 144 90 human Exon33Exon34 220 220 78 human Exon34Exon35 151 151 96 human Exon35Exon36 157 157 82 human Exon36Exon37 213 213 86 human Exon37Exon38 208 211 71 human Exon39Exon39 158 158 79 human Exon40Exon40 192 192 86 human Exon41Exon41 126 126 86 human Exon42Exon42 57 57 59 human Exon43Exon43 131 131 81 human Exon44Exon44 113 113 80 human Exon45Exon45 106 106 87 human Exon46Exon46 135 135 88 human Exon47Exon47 133 136 85 human Exon48Exon48 247 247 82 human Exon49Exon49 178 178 82 human Exon50Exon50 196 196 80 human Exon51Exon51 66 66 75 human Exon52Exon52 129 129 84 human Exon53Exon53 138 138 85 human Exon54Exon54 114 114 83 human Exon55Exon55 88 88 86 human Exon56Exon56 155 155 85 human Exon57Exon57 153 153 85 human Exon58Exon58 876 879 79 human Exon59Exon59 169 169 85 human Exon60Exon59a 153 153 85 human Exon60aExon60 158 158 84 human Exon61Exon61 1018 1018 76 human Exon65Exon62 136 136 81 human Exon66Exon63 88 88 79 human Exon67Exon64 108 108 81 human Exon68Exon65 159 159 81 human Exon69Exon66 123 120 83 (1–37bp)–(38–123bp) human Exon70Exon67 931 614 85 (1–87bp)–(88–142bp) human Exon71

83 (143–284bp)–(285–931bp)

a Percent identity as determined by BLAST2 under default conditions. Numbers in bold identify exons for which sizes differ betweenthe two species.


genomic interval. Analysis of the available mouse genomicsequence suggests that mouse Pkhd1 is similar to its humanorthologue in that it extends over approximately 500 kb andhas the same general organization (Figure 1A). However, thereare presently numerous gaps in the available sequence, thusprecluding a more accurate estimation of its size. A compari-son of mouse and human exons revealed that most are ofidentical length, though there are a few exceptions (Table 1).

In the course of assembling the complete cDNA, we iden-tified a large number of distinct transcripts that had uniquecombinations of Pkhd1 exons. Representative examples arepresented in Figure 1C. In most cases, the alternatively splicedproducts lacked a variable number of exons and were thuslikely derived from shorter mRNAs. A subset of clones wasfound to include exons whose splice-site boundaries differedfrom those used to create the cDNA that encoded the longestORF. In one clone, we found a unique exon not present in the67-exon ORF transcript. It should be emphasized that in allcircumstances conventional splice site donor and acceptor se-quences were used. Interestingly, several of the most dramat-ically spliced products were identified as ESTs derived frommouse kidney IMAGE cDNA clones (AI16884 from clone1481021; AI118496 and AI119022 from 1481154; andAI043040 and AI049325 from clone 1432609). These productsare particularly notable because they link the most 5' and 3'exons of Pkhd1 in a single transcript of only 3.1 kb. Given thatthese clones were isolated from a cDNA library generated inthe conventional manner, it suggests that the unusual pattern ofalternative splicing that we have observed is a general propertyof Pkhd1 and not simply an artifact resulting from use ofPCR-based methods.

We had observed the same phenomenon in our studies ofhuman PKHD1 (5). In the case of the human gene, however,we had identified both a larger number of unique exons (71versus 68) and exons with alternative boundaries (15 versus 6).Although this may reflect true species differences, it is morelikely the result of a less exhaustive analysis of every primercombination using mouse cDNA.

The observation that the human and mouse genes sharedcomplex patterns of splicing prompted us to examine whetherspecific features were conserved. We compared the specificpatterns of splicing, the relative position and sequence of exonsthat occasionally used alternative splice sites, and the relativeposition and sequence of unique exons that were not part of the67 exon longest ORF transcript. We used both cDNA andgenomic sequences for this analysis to detect segments ofhomology not yet represented in our clone set. We found thatneither the patterns of exon assembly nor the DNA sequencesof most of the alternative exons were similar between the twospecies. Moreover, no significant homology could be foundbetween the two species even when the translated sequences ofalternative exons were compared. Likewise, most of the exonsof Celera human transcripts hCT1642763 and hCT1646988that were not present in any of the published PKHD1 tran-scripts also lacked homologues in the mouse sequence (5,12).

There were several exceptions, however. Mouse exon 59,like its human counterpart exon 60, is subject to the same

alternative splice-site usage. In addition, human PKHD1 exon38, which is not part of the longest ORF, and hCT1642763exon B (5) have very high sequence identity to the correspond-ing segments of the mouse genomic sequence. Significant butlower homology was also detected for human exon 64, anotherexon that is not part of the longest ORF.

Finally we searched for regions of homology between thegenomic sequence of human and mouse Pkhd1, looking forpotential exons that have not yet been isolated as part of eithercDNA set. We identified 48 segments that had a minimumlength of �200 bp and a minimum of 80% sequence identity.Reanalyzing the genomic segments with exon prediction pro-grams, we determined that a few members of the set wereidentified as possible exons. Attempts at linking these seg-ments with adjacent exons by PCR of kidney cDNA have beenunsuccessful, however.

Expression Features of Pkhd1One of the more striking aspects of the studies by Ward et al.

(6) and Onuchic et al. (5) was the hybridization pattern ofPKHD1 on Northern blots of human tissue. We thereforeexamined the pattern of Pkhd1 expression in mouse tissuesusing two different cDNA probes derived from exons 5 and 41of mouse Pkhd1, respectively. Representative results are pre-sented in Figure 2. Both probes detected a predominant band ofapproximately 13 kb strongly expressed in kidney and weaklyin liver, heart, stomach, intestine, muscle, uterus, and placenta(Figure 2, A and B). Interestingly, the exon 41 probe alsorecognized a small, approximately 1-kb Pkhd1 transcript intestis that was not detected with the exon 5 probe. Hybridiza-tion of a blot of kidney tissues with probes for mouse Pkd1 andPkhd1 confirmed that the predominant message of the latterwas smaller than 14 kb, as previously reported by Ward et al.(6) (Figure 2, A through C).

To analyze Pkhd1 expression in more detail, we examinedthe pattern of its expression by in situ hybridization in devel-oping mouse tissues over a range of ages (E12.5 to 3 mopostnatal; Figures 3 to 6). High levels of expression werepresent in renal tubular structures at all time points beginningin the mesonephrios (E12.5) and extending to several monthspostbirth (Figure 3). During later stages of kidney develop-ment, strong signals were detected in the branching uretericbud but absent in comma- and s-shaped bodies as well as in themetanephric mesenchyme. In postnatal kidney sections, thestrongest Pkhd1 expression levels were observed in the col-lecting ducts, with somewhat lower levels detected in theproximal and distal tubular epithelia. Expression was notablyabsent from glomeruli at all time points (Figure 3K). Takentogether, Pkhd1 expression in renal collecting ducts and tubulioverlaps precisely with the anatomic structures that undergopathologic changes during cystic transformation in ARPKDpatients.

Another hallmark of ARPKD patients is fibrocystic liverdisease, characterized by malformation of intrahepatic bileducts associated with portal fibrosis. In situ hybridization toliver sections of E15 mouse embryos revealed strong signalsspecifically confined to the developing bilary ducts (Figure 4).


The surrounding hepatic cells and the embryonic blood cellsdid not express Pkhd1. This expression pattern persistedthroughout later stages of hepatic development and after birth.Liver tissue sections from 3-mo-old mice confirmed that Pkhd1expression is restricted to the intra- and extrahepatic bile ductsin the mature organ.

Strong Pkhd1 signals were also found in a number of organsand tissues other than kidney and liver (Figure 5). High levelsof expression were detected in the muscular wall of largevessels, including the thoracic and abdominal aorta, the pri-mordial testis, and dorsal root ganglia. We also detected weaksignals in the embryonic lung mesenchyme in close proximityto the respiratory epithelium, in pancreatic ducts, in the devel-oping trachea and in skeletal muscle (data not shown).

Given the results of our Northern blot analyses and cDNAstudies, we reasoned that specific cell types or entire organsmight express transcripts with unique patterns of exon assem-bly. We confirmed this hypothesis by comparing the in situhybridization results obtained using the exon 41 probe with onethat detects only exon 5 (Figure 6). These studies demonstratethat transcripts expressed in the wall of large vessels includeexon 41 but lack exon 5. Exon 5 was also not detectable inPkhd1 transcripts in the developing lung or trachea, though itwas abundantly present in kidney and liver.

Mouse PolyductinThe full-length translation product of Pkhd1 is predicted to

be 4059 aa in length (Supplementary Information, Figure S2).We selected the first in-frame methionine as the likely initia-tion codon, though an adjacent one could serve in the samecapacity. Neither codon has flanking sequences that exactlymatch the Kozak consensus. Like its human orthologue, mousepolyductin is predicted by SMART to be an integral membraneprotein with a 3851 amino acid extracellular amino terminus, asingle transmembrane domain, and a very short carboxyl ter-minus (Figure 7). It has the same general domain structure ashuman polyductin, with multiple IPT domains (Ig-like, plexin,transcription factor) and PbH1 repeats (parallel beta-helix) inits extracellular segment. Human and mouse polyductin share73% identity over their complete length, though there aresegments with considerably higher (87%) and lower values(40%).

Interestingly, two of the segments with the highest similarity(excluding the TM) are sites where the SMART predictionswere discordant for the two species (Figure 7). For example,SMART did not identify a PbH1 repeat at the position ofmouse polyductin that corresponds to the fourth PbH1 repeat ofthe human protein, yet the sequences are over 84% identical.Likewise, over 85% identity was observed between the two

Figure 2. Northern blot analyses of Pkhd1 in mouse tissues. (A and B) Mouse multiple tissue Northern blot (Seegene, RNWAY) probed withPkhd1 exon 5 (A) and exon 41 (B). Br, brain; ht, heart; lu, lung; li, liver; kd, kidney; st, stomach; sin, small intestine; skm, skeletal muscle;sk, skin; ty, thymus; te, testis; ut, uterus; pl, placenta. (C) Northern blot of mouse adult kidney tissue probed with Pkhd1 and Pkd1. The sameblot was used for both hybridizations.


Figure 3. Developmental expression profile of Pkhd1 in mouse kidney. A, C, E, G, I, and K show dark field photographs in parallel to the brightfield views of hematoxylin-stained sections in B, D, F, H, J, and L. Strong expression is present in the mesonephros (A) and overlaps withdeveloping tubular structures in the embryonic kidney (C, E, G). Persistent expression is observed in the kidney postnatally and overlapsprecisely with tubular structures and collecting ducts (I, K). No expression is observed in the glomeruli (K). A and B, E12.5; C and D, E14.5;E through H, E16.5; I through L, day 7 postnatal. Exon 41 was used as probe. Magnifications: �100 for A through F, I, and J; �200 for Gand H; �250 for K and L.


species for the sequence that corresponds to the eighth PbH1repeat of mouse polyductin. Although it is theoretically possi-ble that the discordant SMART predictions are biologicallyrelevant, it is far more likely that the results are due to limi-tations of the software rather than to true functional differ-ences. In support of this conclusion, SMART failed to identifya predicted signal peptide at the extreme amino terminus ofhuman polyductin, previously reported by Ward et al. (6),though it did so for the mouse orthologue.

Mouse and human polyductin sequence also differ consid-erably in several segments. For example, despite having anidentical number of IPT domains, the homology between thetwo species for the segment of polyductin that contains thecluster of five domains is relatively low at approximately 66%.Likewise, a short segment between PbH1 repeats #2 and #3 ofhuman and the corresponding portion of mouse (i.e., betweenrepeat #1 and #2) is only 42% identical. The most strikingfinding, however, is that the sequence of the cytoplasmictermini of the two proteins is only 55% identical.

We also sought to determine whether any of the motifsidentified by PROSITE were conserved between the species. Infact, several potentially interesting motifs were found common

to human and mouse polyductin. For example, two of the threeputative cAMP/cGMP phosphorylation sites (PS0004) identi-fied by PROSITE within the carboxyl terminus of humanpolyductin are conserved in the mouse orthologue (5). PROS-ITE also identified a putative EF-hand signature sequence(PS00018) in mouse polyductin (820 to 832 aa) that wasinitially missed in the human orthologue but can be found ifone slightly changes the search parameters (822 to 834 aa; 85%similarity to consensus sequence). Using the modified searchconditions, a second EF-hand also is predicted by PROSITE inthe human protein (3266 to 3278 aa; 91% similarity). Thesignificance of the findings in either species is unclear becausenone of the motifs are flanked on both sides by alpha-helicaldomains and all are positioned in extracellular domains. An-other finding of uncertain significance is the putative RGDmotif (PS00016) of human polyductin, because the mouseorthologue lacks the same.

We previously predicted that the translation products ofhuman polyductin would fall into two classes (5). Members ofthe first group, which includes the longest continuous ORF,have a single TM domain and are thus likely to be associatedwith the plasma membrane (polyductin-M), whereas members

Figure 4. Expression of the Pkhd1 transcript during liver development and in mature tissue. A, C, E, G, I, K, M, and O show dark fieldphotographs in parallel to the bright field views of hematoxylin-stained sections in B, D, F, H, J, L, N, and P. Signals overlap with the ductalplate (A and C), small and large intrahepatic bile ducts (E, G, I, and K), and also with extrahepatic bile ducts (O). A through D, E15.5; E throughH, day 1 postnatal; I through P, 3 mo postnatal. Exon 41 was used as probe. Magnifications: �100 for A through J, M, and N; �200 for K,L, O, and P.


Figure 5. Expression of Pkhd1 transcript in organs other than kidney and liver. A, C, E, G, I, and K show dark field photographs in parallel to thebright field views of hematoxylin-stained sections in B, D, F, H, J, and L. Strong signals overlap with the muscular wall of large vessels includingthe thoracic and abdominal aorta (A and C). Particularly prominent expression was observed in the aortic outflow tract. Other sites of expressioninclude the primordial testes (E), lateral sympathetic ganglia (G), the pancreas (I), and the trachea (K). A and B, E14.5; C and D, day 7 postnatal; Eand F, E12.5; G and H: day 7 postnatal; I and J, E14.5; K and L, E16.5. Exon 41 was used as probe. Magnification, �100 for all images.


of the other group, which lacked the TM domain, are likely tobe secreted (polyductin-S). Given that mouse Pkhd1 undergoesa complicated pattern of splicing similar to that of humans, wepredict its translation products will include both membrane-associated and secreted proteins.

DiscussionIn this report we describe Pkhd1, the mouse orthologue of

the gene mutated in human ARPKD. We have found that mostof the features previously reported for the human gene areconserved in the mouse. Both genes are very large, extendingover hundreds of kilobases of genomic DNA. The longest ORFexpressed by each is encoded in a 67-exon transcript. Each ofthe genes is subject to complicated patterns of exon assemblyand alternative exon usage. The pattern of hybridization ofPKHD1 probes on Northern blot is also similar for the twospecies. Lastly, the translation products of each gene are pre-dicted to be very similar and to include polypeptides that areboth membrane-associated and secreted.

A number of small differences also were noted. For exam-ple, several of the domains and motifs identified by SMARTand PROSITE were found to be present in one species but notin the other. Although the absence of these potentially func-tional elements may in fact reflect true biologic differencesbetween the species, we think the discordant results morelikely reflect the limitations of the analytical methods that wereused. The consensus sequence patterns used to identify some ofthe motifs (such as the cell-attachment domain, RGD, or thecAMP/cGMP phosphorylation sites) are very short and thusprone to having a high false positive rate. In the case of theputative EF-hand, we could identify the structure in an identi-cal position in both species only after we lowered the strin-gency of the search conditions. It is clear that detailed func-tional analyses will be required to establish which if any ofthese predictions are valid.

One somewhat surprising finding was the relatively lowsequence homology that was observed between the cytoplas-mic carboxyl terminus of polyductin of the two species. If

Figure 6. Tissue-specific differences in expression of Pkhd1 transcripts. Dark field (A and B) and bright field (C and D) images of serialsections of E15.5 embryos hybridized to exon 5 (A and C) and exon 41 (B and D) probes. Both probes produce strong signals in the renal tubuliand collecting ducts, but only the exon 41 probe detects strong signals in the aortic wall and the muscular wall of renal lobular arteries.


polyductin functions as a receptor, one might have expectedmore conserved regions where the protein interacts with com-ponents of intracellular signaling pathways. The observationthat the cGMP/cAMP phosphorylation sites are conserved de-spite the relatively low homology of the C-termini may indi-cate that they are in fact functionally important.

One of the most striking features of human PKHD1 is itsunusual pattern of splicing. More than twenty different types oftranscripts were identified in our initial phase of cDNA isola-tion, and this is likely a gross underestimate (Onuchic L, et al.,2002; unpublished observations). The discovery that mousePkhd1 has the same properties and undergoes cell-type–spe-cific alternative splicing supports our hypothesis that the com-plicated splicing pattern is likely biologically important. Sim-ilar findings and conclusions have been reported for theneurexins (13–17). This family of genes encodes cell-surfaceproteins that are thought to be important for neuronal cell-cellrecognition. Just three genes are thought to give rise, throughalternative splicing, to thousands of isoforms. Interestingly, thepattern of alternative exon and splice-site usage is very similarto what we have observed for human and mouse Pkhd1. In fact,both Pkhd1 and the neurexins are thought to encode twoclasses of proteins: a set with a single TM and a second inwhich members are likely secreted. Cell-type–specific expres-sion of various isoforms is another feature shared by the twogroups of genes. In the case of neurexins, alternative splicinghas been shown to result in products with different ligandbinding properties (13,14). It will be interesting to see if Pkhd1translation products have similar properties.

The in situ expression data suggest that Pkhd1 may also beimportant in other cell types and organ systems. As expected,prominent expression of Pkhd1 was detected in even the ear-liest stage kidney (mesonephros) and was persistent through

postnatal life. Likewise, Pkhd1 was expressed in the develop-ing and adult biliary tract, as would be predicted on the basisof the clinical phenotype of individuals with ARPKD. How-ever, moderately high levels of expression were also detectedin a number of organs not generally thought to be dysfunctionalin the disease. One implication of these findings is that onemust reconsider whether there might be other covert or sub-clinical disease manifestations in ARPKD patients. It is per-haps interesting to note at least two reports of an association ofaneurysms and ARPKD (18,19). It also should be noted thatour expression data were generated using only two probes.Given the high degree of alternative exon usage in this gene,one might predict that additional probes that detect other exonsmight yield different results. A comprehensive survey may berequired to understand fully the functions of the variousisoforms.

Finally, we have described the genomic organization ofmouse Pkhd1. These data will prove invaluable for the identi-fication of important regulatory elements. Some of the non-coding conserved elements may fall into this category. Thegenomic structure will also be essential for the creation ofgenetically faithful models of human ARPKD. This is neces-sary because none of the currently available mouse models ofARPKD map to the region of mouse chromosome 1 thatcontains Pkhd1 (20). We can also use this information todetermine whether the complete null phenotype differs appre-ciably from that observed in humans or rats. Given the com-plex pattern of splicing of human and mouse Pkhd1, it ispossible that many of the mutant alleles may yield partiallyfunctional products. If this is the case, the phenotype of a truenull may be more severe and have additional phenotypic man-ifestations. It also will be interesting to cross Pkhd1 mutant

Figure 7. Comparative analysis of mouse and human polyductin. Schematic representations of the various domains and motifs of human andmouse polyductin are presented on top. Individual domains of the proteins from the two species were compared using the SIM Alignment Toolfor Proteins, and the output is graphically represented at the bottom. “Unpredicted PbH1 domain” identifies a region of polyductin with highsequence identity between the two species that was identified as a likely PbH1 domain by SMART in one species and not in the other.


mice with mice with ARPKD due to mutations of other genesto determine whether any are members of common pathways.

Electronic Database InformationAccession numbers and URLs for data in this article are as

follows:OMIM: http://www3.ncbi.nlm.nih.gov/Omim/. Genbank: http://

www.ncbi.nlm.nih.gov/Genbank/ for the sequences of all 68exons and the composite cDNA with the longest ORF (acces-sion number AY130764). Unigene: http://www.ncbi.nlm.nih.gov/UniGene/. BLASTP/N/X/2: http://www.ncbi.nlm.nih.gov/blast/. SMART: http://smart.embl-heidelberg.de/. PROSITE:http://ca.expasy.org/. FGenesh: http://genomic.sanger.ac.uk/gf/gf.shtml. GenScan: http://bioweb.pasteur.fr/seqanal/interfaces/genscan.html.

Accession Numbers: PbH1 repeat, SMART #SM0710;Sema domain, SMART #SM0630; PSI domain, SMART#SM0423; IPT domain, SMART #SM0429.

SIM-Alignment tool for proteins: http://us.expasy.org/tools/sim-prot.html.

AcknowledgmentsThis work was supported by NIH R01 DK51259, FAPESP 2000/

00280–3, and the Deutsche Forschungsgemeinschaft. GGG is theIrving Blum Scholar of the Johns Hopkins University School ofMedicine, Baltimore, Maryland.

References1. Zerres K, Mucher G, Becker J, Steinkamm C, Rudnik-

Schoneborn S, Heikkila P, Rapola J, Salonen R, Germino GG,Onuchic L, Somlo S, Avner ED, Harman LA, Stockwin JM,Guay-Woodford LM: Prenatal diagnosis of autosomal recessivepolycystic kidney disease (ARPKD): Molecular genetics, clinicalexperience, and fetal morphology. Am J Med Genet 76: 137–144,1998

2. Reuss A, Wladimiroff JW, Stewart PA, Niermeijer MF: Prenataldiagnosis by ultrasound in pregnancies at risk for autosomalrecessive polycystic kidney disease. Ultrasound Med Biol 16:355–359, 1990

3. Zerres K, Rudnik-Schoneborn S, Deget F, Holtkamp U, BrodehlJ, Geisert J, Scharer K: Autosomal recessive polycystic kidneydisease in 115 children: Clinical presentation, course and influ-ence of gender. Arbeitsgemeinschaft fur Padiatrische. Nephrolo-gie Acta Paediatr 85: 437–445, 1996

4. Zerres K, Rudnik-Schoneborn S, Mucher G: Autosomal reces-sive polycystic kidney disease: Clinical features and genetics.Adv Nephrol Necker Hosp 25: 147–157, 1996

5. Onuchic LF, Furu L, Nagasawa Y, Hou X, Eggermann T, Ren Z,Bergmann C, Senderek J, Esquivel E, Zeltner R, Rudnik-Schoneborn S, Mrug M, Sweeney W, Avner ED, Zerres K,Guay-Woodford LM, Somlo S, Germino GG: PKHD1, the poly-cystic kidney and hepatic disease 1 gene, encodes a novel largeprotein containing multiple IPT domains and PbH1 repeats. J AmHum Genetic 70: 1305–1317, 2002

6. Ward CJ, Hogan MC, Rossetti S, Walker D, Sneddon T, WangX, Kubly V, Cunningham JM, Bacallao R, Ishibashi M, MillinerD, Torres VE, Harris PC: The gene mutated in autosomal reces-

sive polycystic kidney disease encodes a large, receptor-likeprotein. Nat Genet 30: 1–11, 2002

7. Lager DJ, Qian Q, Bengal RJ, Ishibashi M, Torres VE: The pck rat:A new model that resembles human autosomal dominant polycystickidney and liver disease. Kidney Int 59: 126–136, 2001

8. Sanzen T, Harada K, Yasoshima M, Kawamura Y, Ishibashi M,Nakanuma Y: Polycystic kidney rat is a novel animal model ofCaroli’s disease associated with congenital hepatic fibrosis. Am JPathol 158: 1605–1612, 2001

9. Schultz J, Milpetz F, Bork P, Ponting CP: SMART, a simplemodular architecture research tool: Identification of signalingdomains: Proc Natl Acad Sci USA 95: 5857–5864, 1998

10. Letunic I, Goodstadt L, Dickens NJ, Doerks T, Schultz J, Mott R,Ciccarelli F, Copley RR, Ponting CP, Bork P: Recent improve-ments to the SMART domain-based sequence annotation re-source. Nucleic Acids Res 30: 242–244, 2002

11. Moser M, Imhof A, Pscherer A, Bauer R, Amselgruber W,Sinowatz F, Hofstadter F, Schule R, Buttner R: Cloning andcharacterization of a second AP-2 transcription factor: AP- 2beta. Development 121: 2779–2788, 1995

12. Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, SuttonGG, Smith HO, Yandell M, Evans CA, Holt RA, Gocayne JD,Amanatides P, Ballew RM, Huson DH, Wortman JR, Zhang Q,Kodira CD, Zheng XH, Chen L, Skupski M, Subramanian G,Thomas PD, Zhang J, Gabor Miklos GL, Nelson C, Broder S,Clark AG, Nadeau J, McKusick VA, Zinder N, Levine AJ,Roberts RJ, Simon M, Slayman C, Hunkapiller M, Bolanos R,Delcher A, Dew I, Fasulo D, Flanigan M, Florea L, Halpern A,Hannenhalli S, Kravitz S, Levy S, Mobarry C, Reinert K, Rem-ington K, Abu-Threideh J, Beasley E, Biddick K, Bonazzi V,Brandon R, Cargill M, Chandramouliswaran I, Charlab R,Chaturvedi K, Deng Z, Di Francesco V, Dunn P, Eilbeck K,Evangelista C, Gabrielian AE, Gan W, Ge W, Gong F, Gu Z,Guan P, Heiman TJ, Higgins ME, Ji RR, Ke Z, Ketchum KA, LaiZ, Lei Y, Li Z, Li J, Liang Y, Lin X, Lu F, Merkulov GV,Milshina N, Moore HM, Naik AK, Narayan VA, Neelam B,Nusskern D, Rusch DB, Salzberg S, Shao W, Shue B, Sun J,Wang Z, Wang A, Wang X, Wang J, Wei M, Wides R, Xiao C,Yan C, Yao A, Ye J, Zhan M, Zhang W, Zhang H, Zhao Q,Zheng L, Zhong F, Zhong W, Zhu S, Zhao S, Gilbert D, Baum-hueter S, Spier G, Carter C, Cravchik A, Woodage T, Ali F, AnH, Awe A, Baldwin D, Baden H, Barnstead M, Barrow I, BeesonK, Busam D, Carver A, Center A, Cheng ML, Curry L, DanaherS, Davenport L, Desilets R, Dietz S, Dodson K, Doup L, FerrieraS, Garg N, Gluecksmann A, Hart B, Haynes J, Haynes C, HeinerC, Hladun S, Hostin D, Houck J, Howland T, Ibegwam C,Johnson J, Kalush F, Kline L, Koduru S, Love A, Mann F, MayD, McCawley S, McIntosh T, McMullen I, Moy M, Moy L,Murphy B, Nelson K, Pfannkoch C, Pratts E, Puri V, Qureshi H,Reardon M, Rodriguez R, Rogers YH, Romblad D, Ruhfel B,Scott R, Sitter C, Smallwood M, Stewart E, Strong R, Suh E,Thomas R, Tint NN, Tse S, Vech C, Wang G, Wetter J, WilliamsS, Williams M, Windsor S, Winn-Deen E, Wolfe K, Zaveri J,Zaveri K, Abril JF, Guigo R, Campbell MJ, Sjolander KV,Karlak B, Kejariwal A, Mi H, Lazareva B, Hatton T, NarechaniaA, Diemer K, Muruganujan A, Guo N, Sato S, Bafna V, Istrail S,Lippert R, Schwartz R, Walenz B, Yooseph S, Allen D, Basu A,Baxendale J, Blick L, Caminha M, Carnes-Stine J, Caulk P,Chiang YH, Coyne M, Dahlke C, Mays A, Dombroski M,Donnelly M, Ely D, Esparham S, Fosler C, Gire H, Glanowski S,Glasser K, Glodek A, Gorokhov M, Graham K, Gropman B,


Harris M, Heil J, Henderson S, Hoover J, Jennings D, Jordan C,Jordan J, Kasha J, Kagan L, Kraft C, Levitsky A, Lewis M, LiuX, Lopez J, Ma D, Majoros W, McDaniel J, Murphy S, NewmanM, Nguyen T, Nguyen N, Nodell M, Pan S, Peck J, Peterson M,Rowe W, Sanders R, Scott J, Simpson M, Smith T, Sprague A,Stockwell T, Turner R, Venter E, Wang M, Wen M, Wu D, WuM, Xia A, Zandieh A, Zhu X: The sequence of the humangenome. Science 291: 1304–1351, 2001

13. Ichtchenko K, Hata Y, Nguyen T, Ullrich B, Missler M,Moomaw C, Sudhof TC: Neuroligin 1: A splice site-specificligand for beta-neurexins. Cell 81: 435–443, 1995

14. Ichtchenko K, Nguyen T, Sudhof TC: Structures, alternativesplicing, and neurexin binding of multiple neuroligins. J BiolChem 271: 2676–2682, 1996

15. Missler M, Sudhof TC: Neurexins: Three genes and 1001 prod-ucts. Trends Genet 14: 20–26, 1998

16. Ushkaryov YA, Hata Y, Ichtchenko K, Moomaw C, Afendis S,Slaughter CA, Sudhof TC: Conserved domain structure of beta-

neurexins. Unusual cleaved signal sequences in receptor-like neu-ronal cell-surface proteins. J Biol Chem 269: 11987–11992, 1994

17. Ullrich B, Ushkaryov YA, Sudhof TC: Cartography of neurex-ins: More than 1000 isoforms generated by alternative splicingand expressed in distinct subsets of neurons. Neuron 14: 497–507, 1995

18. Neumann HP, Krumme B, van Velthoven V, Orszagh M, ZerresK: Multiple intracranial aneurysms in a patient with autosomalrecessive polycystic kidney disease. Nephrol Dial Transplant 14:936–939, 1999

19. Lilova MI, Petkov DL: Intracranial aneurysms in a child withautosomal recessive polycystic kidney disease. Pediatr Nephrol16: 1030–1032, 2001

20. Xiong H, Chen Y, Yi Y, Tsuchiya K, Moeckel G, Cheung J,Liang D, Tham K, Xu X, Chen XZ, Pei Y, Zhao ZJ, Wu G: Anovel gene encoding a TIG multiple domain protein is a posi-tional candidate for autosomal recessive polycystic kidney dis-ease. Genomics 80: 96–104, 2002


Identification and characterization of Pkhd1, the mouse orthologue of the human ARPKD gene

Documents