Evaluation of the APOH Gene as a Positional Candidatefor prcd in Dogs

WeiKuan Gu,1 Kunal Ray,1'2 Sue Pearce-Kelling,1 Victoria J. Baldwin,1

Amelia A. Langston,5 Jharna Ray,1 Elaine A. Ostrander,5 Gregory M. Acland,1 andGustavo D. Aguirre1

PURPOSE. Progressive rod-cone degeneration (prcd) is an autosomal recessive retinal degenerationof dogs characterized by abnormalities in lipid metabolism. It has recently been mapped to thecentromeric region of canine chromosome 9, homologous to human 17q, which contains theapolipoprotein H (apoH, protein; APOH, gene) gene involved in lipid metabolism and regulation oftriglycerides. The present study was undertaken to evaluate APOH as a positional candidate forprcd.

METHODS. Expression of APOH in the retina was examined by reverse transcription-polymerasechain reaction (RT-PCR) and by immunocytochemistry in normal and/>rcd-affected dogs. The levelof apoH in the plasma was determined by western blot analysis. Intragenic polymorphic markerswere identified and typed in the prcd pedigree. Canine-rodent hybrid cell lines were analyzed todetect canine APOH.

RESULTS. ApoH has been localized to the photoreceptor outer segment layer by immunocytochem-istry. Its expression in the retina of normal and prcd-affected dogs was confirmed by RT-PCR. Thelevels of antihuman apoH cross-reacting material in plasma were similar in all dogs, regardless ofdisease status. Finally, linkage analysis of the APOH gene with the disease locus in the prcd pedigreedetected 3 recombinants among 70 informative offsprings (lod score 15.09 at 0 = 4.3 centimorgan[cM]).

CONCLUSIONS. APOH is expressed in the retina and tightly linked to the prcd locus. However, despiteits potential role in phenotypes of abnormal lipid metabolism associated with prcd, the gene hasbeen excluded as a primary candidate for prcd by linkage analysis. (Invest Ophthalmol Vis Set.1999;40:1229-1237)

Progressive rod-cone degeneration (prcd) is the mostprevalent retinal disorder of dogs among several distinctinherited diseases recognized clinically and collectively

as progressive retinal atrophy. In homozygous affected dogs,the retina develops and functions normally. Subsequently,however, rods and cones undergo a topographically defined

From the 'James A. Baker Institute, College of Veterinary Medi-cine, Cornell University, Ithaca, New York; and the 3Fred HutchinsonCancer Research Center, Seattle, Washington.

Supported in part by The Morris Animal Foundation and TheSeeing Eye, Inc., Morristown, NJ; Grant EYO6855 from the National EyeInstitute, Bethesda, Maryland; The Foundation Fighting Blindness,Hunt Valley, Maryland; the Cornell Center for Advanced Technologyand The Baker Institute PRA Research Fund, Ithaca, New York; aresearch grant from the Canine Health Foundation of the AmericanKennel Club (EAO), Aurora, Ohio; Grant VM-165 from the AmericanCancer Society, Atlanta, Georgia (EAO). EAO is the recipient of anAmerican Cancer Society Junior Faculty Award; AAL is the recipient ofPhysician Scientist Award Kl 1 HD00936 from the National Institutes ofHealth, Bethesda, Maryland.

Submitted for publication June 23, 1998; revised January 6, 1999;accepted Febaiary 17, 1999-

Proprietary interest category: N.2Present address: Department of Human Genetics, Indian Institute

of Chemical Biology, Calcutta, India.Reprint requests: Gustavo D. Aguirre, James A. Baker Institute for

Animal Health, College of Veterinary Medicine, Cornell University,Ithaca, NY 14853-6401.

sequence of disease and degeneration that eventually leads toretinal atrophy and blindness. The standard variant of the prcdphenotype was first characterized in miniature poodle dogs.1

Since then, phenotypic variations of the disorder have beenobserved to occur in several other breeds, and these includevariation in the age of onset, rate of progression, and otherclinical and morphologic characteristics.2'3

Recent development of a low-resolution framework mapof the canine genome,4 and physical and linkage mapping ofhuman chromosome 17q loci to canine chromosome 9 (CFA9)by Werner et al.5 has permitted us to establish a linkage map ofthe prcd locus.6'7 Identification of a linkage group flankingprcd ([TK1, GALK1, prcd] — [MYL4, C09.173, C09.2263] —[RAKA, RAPD]—C09.250—C09.474—NF1) localizes the dis-ease locus close to the centromeric end of CFA9.

Identification of synteny between CFA9 and HSA17q hasbeen helpful to refine further the linkage map of prcd andidentify potential positional candidates. One such candidate isapolipoprotein H (apoH, protein; APOH, gene),8'9 also calledj3-2 glycoprotein I. The APOH gene codes for a glycoprotein ofapproximately 50 kDa consisting of 326 amino acids and richin cysteine and proline residues. A canine cDNA sequencefrom the liver has been reported (GenBank accession no.X72933), and its derived amino acid sequence showed highhomology to the human homologue.

Investigative Ophthalmology & Visual Science, May 1999, Vol. 40, No. 6Copyright © Association for Research in Vision and Ophthalmology 1229

1230 Gu et al. IOVS, May 1999, Vol. 40, No. 6

TABLE 1. Primers Used for PCR

Primer Sequence (5' to 3') Direction

ForwardReverseForwardReverseReverseForwardReverseReverseReverseForwardForwardForward

Location*

Intron 3Intron 3Exon 3/276-302Exon 5/476-450Exon 8/1043-1020Exon/407-432Exon 4/402-382Intron 3Intron 3Intron 3Intron 3Intron 3

APOH-1APOH-2APOH-4APOH-5APOH-6APOH-7APOH-8APOH-9APOH-10APOH-11APOH-12APOH-13

TCTGGGACAAAGCTTGGCATGAGAAGCGTCTGTCTCTTCTTCAGGAATCTTAGAAAATGGAGCTGTACGCTCACTAAGTGTTGCAAACTTAGGTACGGGCTTTACATCTGATGCATCGGTTTGTTGACCTTCCTGTCTGTACTCGTGCATTTTCCTTCCTCGGTGCATGCCTTTGCTCATAAGGCACATAAGCGGTGCATTTAGCAGAGCTACTTCCCCTGCTGATTGGAAAGAGGGATTTGGCTTTGCTCCACATAAGAAGCTGTGCATTACAAAATAGGGGGAAAGGC

* Location of the primers, designed from coding sequence, is represented based on canine APOH cDNA sequence (GenBank accession no.X72933). Exon and intron numbers are assigned arbitrarily based on the human APOH gene (GenBank accession no. Y11493-Y11498).

The protein (apoH) has been implicated in a variety ofphysiologic pathways including phospholipid metabolism,but its involvement in lipid metabolism, and its physicallocation in HAS17q23-qter directs our attention to considerAPOH to be a positional candidate for prcd. The protein issynthesized primarily in liver, although recent studies havefound APOH mRNA expression in intestinal epithelial celllines and tissues.l0 APOH is involved in lipid metabolism andregulation of triglycerides, and these functions are deter-mined by genetic structural polymorphisms represented byalleles AP0H1-4 in man ." 1 2 1 3 Regarding lipoprotein me-tabolism, 6% of the protein is in chylomicrons and the VLDLfraction.14 Based on studies using isoelectric focusing, struc-tural alleles have been identified in patients that are associ-ated with specific patterns of cholesterol and triglycerides inthe plasma.1 l l 2

We have previously reported abnormalities in lipid me-tabolism in prcd-affected dogs.1516 The plasma levels ofdocosahexaenoic acid (DHA; 22:6n-3) and cholesterol inaffected dogs were significantly lower than those of normalanimals sharing the same environment and fed the samediet. No other differences in plasma fatty acids, lipids, tri-glycerides, or levels of the fat-soluble vitamins A and E werefound between normal and affected dogs.15 More recently,we have found that short- and long-term administration ofDHA-enriched supplements to prcddffected dogs resulted ina sustained three- to fourfold elevation of DHA in plasma andliver but that the rod outer segment (ROS) DHA levelsremained unchanged.16 We speculated that these could bethe result of a primary reduction in the synthesis of DHA-containing phospholipid in the ̂ red-affected retinas or, al-ternatively, secondary to a reduction in DHA uptake, trans-port, or storage within the retinal pigment epithelium-photoreceptor complex.

To our knowledge, no studies have been reportedon the expression of apoH in the retina or its plausible rolein DHA-phospholipid metabolism, particularly in the ROS.In this study, we analyzed the association between theAPOH gene and prcd disease using linkage analysis andexamined the expression and the localization of apoH in theretina.

MATERIALS AND METHODS

Animals

The dog colony on which this study is based is part of anNational Eye Institute-National Institutes of Health-sponsoredproject (Grant EY06855, Models of Hereditary Retinal Degen-eration) at the Retinal Disease Studies Facility (Kennett Square,PA). The dogs used encompass nine related multigenerationalpedigrees with 70 informative progenies in the third genera-tion. All dogs derive from the research colony of purebredminiature poodles in which the phenotype and inheritance ofprcd was originally characterized.' '2' '7 To generate informativepedigrees, prcd-affected dogs were bred to homozygous nor-mal unrelated miniature poodles, beagles, and beagle-cross-bred dogs, and their heterozygous offspring were then back-crossed to prcd-affected dogs to yield litters segregating for the/>ra/-phenotype. There are only two genotypes, homozygousaffected and heterozygous nonaffected dogs, in the third gen-eration of this colony. The phenotype of the dogs was deter-mined by a combination of ophthalmoscopic, electrophysio-logical, and retinal morphologic evaluations. The diagnosticcriteria for the disease has been previously described.1'2 Allprocedures involving animals were performed in adherence tothe ARVO Resolution for the Use of Animals in Ophthalmic andVision Research.

Polymerase Chain Reaction

DNA was isolated, using standard methods18 from blood sam-ples collected in citrate anticoagulant tubes or splenic samplesfrom deceased clogs, maintained at — 70°C and stored in TE (10mM Tris-HCl [pH 8.0] and 0.1 mM EDTA) buffer. Initially,primers were designed (Table 1) from canine APOH cDNAsequence (GenBank accession no. X72933). Additional primerswere designed for characterization of the gene by sequencingand identification of intragenic polymorphisms.

All polymerase chain reactions (PCRs) were performed for30 to 35 cycles in a volume of 25 /xl to 50 /utl containing 50 mMKC1; 10 mM Tris-HCl (pH 8.3); 1.5 mM to 2 mM MgCl2; 0.2 mMeach deoxyadenosine triphosphate, deoxycytidine triphos-phate, deoxyguanosine triphosphate, and deoxythymidinetriphosphate, (dATP, dCTP, dGTP, and dTTP); 0.2 p,M oligo-nucleotide primer; 200 ng template DNA; and 0.5 U Taq DNA

10VS, May 1999, Vol. 40, No. 6

A 1 2 3 4 5 6 M

B

48KDa

FIGURE 1. Expression of APO11 in normal and prcd-affected dogs. (A)Detection of APOH mRNA in normal and />rctf-affected retina by RT-PCR. Total RNA from normal Claries 1 and 2) and /jra/-affected dogretinas (Janes 3 and 4) was used for analysis. Total RNA from liver ofnormal clogs (lanes 5 and 6) was used as control. The PCR productswere electrophoresed in a 2% agarose gel. Lane M represents DNAmarkers (100-bp ladder). (B) Detection of plasma apoH in normal and/wctf-affected dogs. Serum proteins (280 /xg) were electrophoresed ina denaturing sodium dodecyl sulfate-polyacrylamide gel (12.5%) fromhomozygous normal (+ /+ ; lane /), prcd-dffectcd (prcdlprcd; lanes 3and 4), and/>rctf-carrier (+/prcd; lanes 5 and 6) dogs. Commerciallyavailable purified human apoH (1 fig) was used as control (lane 2).Separated proteins were transferred to nylon membrane and cross-reacted with rabbit anti-human antibody as described in the Materialsand Methods section.

polymerase (Life Technologies, Gaithersburg, MD). Amplifica-tion parameters were 94°C for 1 minute; annealing for 1 to 2minutes at a suitable temperature, depending on the meltingtemperature (Tm) of the primer pairs; and elongation at 72°Cfor the appropriate time, depending on the size of the ampli-fication product, which was usually 1 minute per kilobase. PCRproducts were separated in agarose or acrylamide gels depend-ing on the size of the amplified DNA fragments.

Reverse Transcription and PCRTotal RNA from retinal tissues of normal and affected dogs andfrom livers of normal dogs was isolated as described.19 Reversetranscription (RT) was conducted using an RNA PCR kit (Per-

APOH and prcd 1231

kin Elmer, Foster City, CA), as recommended by the manufac-turer. Five microliters of the RT-product was used in the PCRto amplify a portion of the canine APOH cDNA using primerpairs APOH-4 and APOH-6 (Table 1). Reaction was obtained ina total volume of 50 JU.1 at a 2-mM MgCl2 concentration (35cycles of 94°C for 1 minute, 55°C for 1 minute, and 72°C for 2minutes). PCR products were analyzed by 2% agarose gelelectrophoresis.

Identification of Polymorphic Markers

To amplify intronic regions from the APOH gene to identifypolymorphisms, different primers were designed according tothe canine cDNA sequence. Combinations of these primerpairs were used in the PCR amplifications with genomic DNA(data not shown). Many of them amplified small single ormultiple bands, but a unique 4-kb fragment, which containedapproximately 3 8 kb of intronic sequence, was amplified witha primer pair APOH-4 and APOH-5 (Table 1). This fragment,therefore, was further analyzed for polymorphisms. Amplifica-tion was performed for 35 cycles of 94°C for 30 seconds, 58°Cfor 1 minute, and 72°C for 4 minutes under the conditionsdescribed earlier, using 1.5 mM MgCl2. The identity of the DNAfragment as the APOH gene was confirmed by nested PCR andby partial sequencing through the coding region of the gene.Digestion of the 4-kb fragment with 22 restriction enzymesrevealed no restriction fragment length polymorphism. Next,the same region of the APOH gene was amplified as twofragments, 1.6 kb and 2.4 kb, using two pairs of primers,APOH-7/APOH-5 and APOH-4/APOH-8, respectively (Table 1).These two DNA fragments were digested with restriction en-zymes, and heteroduplex analysis was performed by themethod of Gu et al.20

The DNA fragments suspected to contain a heteroduplexwere purified from the gel (Qiagen, Chatsworth, CA), andsequenced directly by Taq cycle sequencing using DyeDeoxyterminators in an automated DNA sequencer (ABI 377; AppliedBiosystems, Foster City, CA) at the core sequencing facility ofCornell University (Ithaca, NY). Whenever the sequenced re-gion revealed the presence of a nucleotide change, it wasexamined for alteration of any restriction site to create a re-striction fragment length polymorphism. Manipulation andcomparison of the DNA sequences was performed using com-puter programs (GCG; Genetics Computer Group, Madison,WI; and DNA STAR; DNASTAR, Madison, Wl). Primers weredesigned flanking the polymorphic site and used to amplify thetarget DNA fragments from dogs of different genotypes in theprcd pedigree.

Linkage Analysis

Once a polymorphism in the APOH gene was identified, par-ents of informative prcd litters and progeny of known prcdgenotype from doubly informative parents were tested fortheir genotypes. All the respective grandparents from whichDNA was available were also tested and scored and haplotypesassigned. Linkage analyses was performed using the LINKAGEpackage21 as described by Terwilleger and Ott.22

Physical Mapping of the APOH Gene on Canine-Rodent Hybrid Cell Lines

Determination of the physical location of the APOH gene incanine-rodent hybrid cell lines was conducted by PCR ampli-

1232 Gu et al. IOVS, May 1999, Vol. 40, No. 6

FIGURE 2. lmmunocytochemistry of l-p,m tliick diethylene glycol disteanite (DGD)-embedded retinafrom a normal adult dog reacted with polyclonal antibody against the apoH protein. (A) Low powershowing immunoreactivity in the outer segment layer. Open arrow, immunoreactivity within an innerretinal vessel; *aitifactual space secondary to poor DGD penetration. (B) Higher power shows labeling inthe outer segment layer and entrapped within strands {thin arrows) of the insoluble IPM. Arrowhead,external limiting membrane. ONL, outer nuclear layer; INL, inner nuclear layer. Magnification, (A) X370;(B) X880.

fixation as described previously by Langston et al.23 A pair ofprimers, APOH-11 and APOH-8 (Table 1), was used to developa canine-specific sequence-tagged site (STS) for the APOHgene. Canine-specific PCR product of expected size (315 bp)was obtained by amplification at 30 cycles of 94°C for 30seconds, 55°C for 1 minute, and 72°C for 1 minute, using 1.5mM MgCI2.

Western Blot Analysis

Serum samples from normal, carrier, and affected dogs ofsimilar age were prepared as previously described.24 Purifiedhuman apoH sample (AKZO Nobel; Perlmmune, Rockville,MD) was used as a control. Protein concentration was deter-mined using a protein assay kit (Bio-Rad, Hercules, CA). Thesamples were electrophoresed in a 12.5% sodium dodecylsulfate-polyacrylamide gel under reducing condition,25 andthen electrotransferred to a nitrocellulose membrane (Schlei-cher & Schuell, Keene, NH). Because the APOH coding se-quences of humans and dogs are the same size and the de-duced amino acid sequences shares 89% homology (GenBankP337O3 and P02749), western analysis was performed using acommercially available polyclonal anti-human apoH immuno-globulin (AKZO Nobel, Perlmmune) and alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin as a secondaryantibody. The apoH immune complex was detected by color-imetric method using nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Promega, Madison, WI).

ImmunocytochemlstryImmunocytochemical localization of apoH was achieved usingnormal and/?ro/-affected retinas obtained by enucleation fromlight-adapted dogs under intravenous anesthesia with sodiumpentobarbital. The retinas were fixed in 4% paraformaldehydein 0.1 M phosphate-buffered saline (PBS) and embedded indiethylene glycol distearate (Polysciences, Warrington, PA) aspreviously described.26 One- to 2-ju.m sections were cut on amicrotome (model 2065; Reichert, Leica, Inc., Deerfielcl, IL)microtome using glass knives and mounted on silane-coatedslides. Retinal tissue sections were dewaxed with toluene andrehydrated, and nonspecific staining was blocked by a 15-

minute incubation in 0.01 M PBS containing 2.5% bovine se-rum albumin and 0.25% Triton X-100. The same rabbit anti-human polyclonal antibody was used and incubated overnightat room temperature in a humidified light-tight chamber at a1:25 dilution in 0.01 M PBS. Biotinylated goat anti-mouse IgGwas used as a secondary antibody (Zymed, San Francisco, CA)to visualize the immunoreactions. All sections were examinedand photographed using a Zeiss Axioplan microscope (CarlZeiss, Oberkochen, Germany) and Kodak Technical Pan film(Eastman Kodak, Rochester, NY).

RESULTS

Expression of APOH Gene by RT-PCR, WesternBlot Analysis, and ImmunocytochemistryAn 800-bp APOH cDNA fragment was amplified from retinalmRNA by RT-PCR from homozygous normal ( + / + ) and prcd-affected (prcd/prcd) samples (Fig. 1A). Sequencing of the PCRproduct indicated that the amplimer represented authenticAPOH sequence. In a western blot analysis, human anti-apoHpolyclonal antibody cross-reacted with a protein (48 kDa) sim-ilar in size to human apoH (Fig. IB). In addition to the majorband, some weak bands were detected, possibly because of theprotein polymorphisms. This finding was confirmed on re-peated experiments, and no differences were observed be-tween normal and /jratf-affected dogs.

In the 1- to 2-/nm thick retinal sections used for immuno-cytochemistry, immunoreactivity was present in the lumen ofthe choroidal and retinal vessels and in the photoreceptorouter segment layer (Fig. 2A, left). The remainder of the pho-toreceptor cell body and the perinuclear and synaptic regionsshowed only weak background staining that was no differentfrom that found in the remaining retinal layers. Control sec-tions, without the primary antibody, showed no immunoreac-tivity. In sections in which the RPE and retina had just slightlyseparated, moderate immunoreactivity was observed in thestrands associated with the insoluble interphotoreceptor ma-trix (IPM; Fig. 2B, right). Because the 1PM collapses onto thephotoreceptors, particularly the outer segments, with this

IOVS, Msiy 1999, Vol. 40, No. 6

A

APOH and prcd 1233

1 2 3 M

bl/b2-

1 2 3 M

200

3 AF0H-4-->1 GAATCTTAGA AAATGGAGCT GTACGCTATA CAACTTTTGA ATATCCCAAC51 ACGATCAGTT TTGCTTgtaa cactgggtaa ggacgtctag ccaacagata101 canaacattc tgtgccttta agctaaacat caanacatat tttattttta151 gccttcaggc tggacttgct gaatacaaat caattatagt cctgttttta201 atctatacca cagtggtatt ggggttaggg aaaccataac aactcgtgaa251 atgtgaagtt gctactgcat taaaactttt ttctcccaat taattactta301 tctgaggagg tacactggaa tgattaaaac tggataataa ttcctaatca

APOH-1-->351 tctctqggac aaagcttgqc atqaaggtta anatacttta ggaattggac401 tgggctttag ttatgaacct aatttgattg actgacctgg gagtctccag451 aaggagtttt gcttttgttt ttttttaatt tattcaagac acagaaagag501 agagagagag agaggcagag acacaggcag agggagaagc aggctccctg551 caggaagcct gacacaggac tccatcctgg atctccagga tcacaacctg601 gactgaagac ggtactaaag cactgagcca cctgggcttc ccctagaagg651 gagttctttg gtgaaggttg gcgtgtggtg gtataaaaca tacttctgtt7 01 aatcctcaca caatgcagga cagcattccc agcaagaaag catgtacagc

<--APOH-9751 gggtcatata tagcttatqt gccttatgag caaaqqcact gccagtgaca801 ttgtcaatag cacgtcgtct gtatcctagg gcaaaatttc actgagactt851 cttgctcgga attggaaatt aagaaatact atgtgttaaa gagaattgga901 ctcctttcat ctgaaacatc aattaatttg tgggtctgat gacaaagtac951 cttgaaacaa ttcgcttcga ggagacagtg ttttcagtgg aaatatactg1001 cattgttttt acccatttat gtcagtagac tgaatcgttt cagaaaatgg1051 atagcagcca tgactcattt gattcacttt tgtgatccct gaagaagaga

APOH-12-->1101 cagacgcttc tttqctccac ataaqaaqct qttaatttqq cctgctgatt1151 ggaaaagagg gatttggggt tttggcttct atgntcatag ctatggaaac12 01 ttctcattnt tctgtgagac ctcngcagcc ctgggttggg aaatttgatg

APOH-13 — >1251 agcctnaaat gagtgctgag caactcagga aaacagcatt acaaaataqq

(g)1301 qqqaaaqtc(g/a) gtagcattaa tctgaagcaa tgctcagtgg agaagacatt1351 tgaaangcca tcaatggatt cttttactgc ttgatggttc cattttccct1401 tttcagGTTT TATCTAAATO QAAGTAGCTC TQCTAAATCC ACCGAGGAAG

<—(APOH-10) <--APOH-81451 GAAAATG

FIGURE 3. Identification of polymorphisms in the APOH gene. (A) The first polymorphism (left) wascaused by insertion of a tetranucleotide (tgac). ifcm/ll-digested DNA from samples homozygoiis for alleleal (.lane 1) containing the insert and for allele a2 (lane 2) without the insert and heterozygous for al anda2 (lane 3) were separated using a 6% polyacrylamide gel. The extra two bands in lane 3 were caused byheteroduplex formation between alleles al and a2. The second polymorphism (middle) was caused byinsertion of a single nucleotide a. Samples homozygous for allele bl (+a; lane 1) and allele b2 (—a; lane2) and heterozygous for bl and b2 (lane 3) were separated using a 6% polyacrylamide gel. Note thatsamples homozygous for bl and b2 (lanes 1, 2, respectively) have similar mobility and that the heterozy-gous sample (lane 3) is distinguished by the presence of the slower migrating heteroduplex. The thirdpolymorphism (right) was caused by a single nucleotide polymorphism (g/a). /4c*I-digested DNA fromsamples homozygous for allele cl (lane 1) and allele c2 (lane 2) and heterozygous for cl and c2 (lane J)were separated using a 6% polyacrylamide gel. Only allele cl was digested by creation of a restriction siteby the mismatch primer (APOH-13). (B) The sequence of the region of the APOH gene containing thepolymorphisms. The region of the gene representing intron 3 is shown in lowercase and the gt agat the splice junctions are indicated by bold letters. Three polymorphic sites are shown in bold letters. Thefirst polymorphism (insertion/deletion; tgac) is located from nucleotide 429 to 432, the second polymor-phism (insertion/deletion; a) is located at the nucleotide 1156, and the third polymorphism (SNP; eitherg or a) is located at the nucleotide 1310. Sequence of all the primers used in the PCR are identified. Thenucleotide g in parentheses above t (nucleotide 1308) represents the sequence of the mismatch primerAPOH-13 at that location to create a restriction site. The sequence has been submitted to GenBank(accession no. AFO69O52).

1234 Gu et al.

A P775(1.3) (2,5)

f SI(2,3)1 (2,4)

(2,3) (2,4) (2,4) (2,3)

IOVS, May 1999, Vol. 40, No. 6

E301 P492(4,4) (2,4) (2,4)

9

(2,4)

P898(2,4)

(2,2)

AW AM AM(2,4) (2,2) (2,4) (2,2) (2,4) (2,2) (2,4) (2,4) (2,2)

• • • • • •>• • .<• • •(3.4) (2,3) (3,3) (2,3) (3,3) (2,3) (3.3) (2,4) (3.3) (2,3) (2,4) (3,3) (3,3)(2,3)

BP775(2,5)

P780E30I (3,4)(4,4)

P7O3(2,3) (2,3)

P(2,4) (3,4)

(2,3) (3,4) (3,4) (3,4) (3,4) (3,4)*

AW AW 4W A & A(3,3) (2,3) (3,3) (3,3) (2,3) (3,3) (2,3) (3,3) (3,3) (3,3) (2,3) (3,3) (2,3) (2,3)

P492

<2.3)(£ (3,4)(^ [^(2,3) A (2,4)

(3,3) (3,3)

AJ44*(2,4) (2,4) (2,2) (2,2) (3,4)

(3,3)

mrr(3,3) (2,3) (2,3) (2,3) (3,3) (2,3)

AAMAMAJJ9(3,4) (3,4) (3,3) (3,4) (3,3) (3,4) (3,4) (3,4) (3,3)

FIGURE 4. Pedigree informative for prcd showing genotypes for the disease and APOH gene. Circles represent females, squares represent males,open symbols indicate homozygous normal at prcd locus, solid symbols indicate prcd-affected, half-filled symbols indicate prcd heterozygous. Thegenotype for APOH alleles for each dog is shown in parentheses. Six different APOH alleles were detected depending on the three polymorphicmarkers (as shown in Fig. 3) identified in the gene (allele 1: al bl cl; allele 2: al bl c2; allele 3: al b2 cl; allele 4: a2 bl c2; allele 5: a2 b2 c2;allele 6: al b2 c2). The three clogs identified as recombinants between the prcd locus and the APOH marker are marked by asterisks.

method of fixation and embedding,27 it was not possible todetermine whether apoH also is a constituent of the IPM or isrestricted to the outer segments. Identical results were ob-tained with the ̂ red-affected retinas (data not shown).

Collectively, these data suggest but do not confirm thatmutations in APOH are not the major underlying genetic defectresulting in prcd. The RT-PCR experiment only indicates ex-pression of the 800-bp fragment in normal and affected retinaand would not detect any subtle but consistent difference inthe level of expression. Additionally, the western blot analysisas performed on plasma proteins may not reveal a mutantprotein of normal size. For that reason, we attempted to eval-uate unequivocally the causal association of the APOH genewith prcd by linkage analysis.

Identification of Polymorphisms in the APOHGene in the prcd Pedigree

A 4-kb fragment of the APOH gene, spanning exon 3 to exon 5,was amplified in two segments, a 1.6-kb DNA fragment con-

taining intron 3 and a 2.4-kb DNA fragment containing intron4, and examined for any nucleotide change. Three polymor-phisms were identified in the 1.6-kb DNA fragment and thealleles characterized in all dogs (Fig. 3).

The first polymorphic site, insertion of a tetranucleotide(TGAC), was screened by amplification of a 430/434-bp frag-ment using APOH-1 and APOH-9 primers (Table 1) and thendigested with BsmAl (Fig. 3A, left). The digested DNA con-tained a polymorphism, either 100 bp (allele al) or 96 bp(allele al), and several other nonpolymorphic DNA fragments.

The second polymorphic site (±A) was detected in a347-bp fragment amplified using primers APOH-12 and APOH-8(Table 1). The two alleles, with and without the polymorphicA, were identified as bl and b2, respectively (Fig. 3A, middle).For typing the dogs in the prcd pedigree, PCR was performedon all samples to identify dogs heterozygous for this polymor-phism by detecting the heteroduplex. PCR product from thosedogs producing a single band was mixed with the PCR productfrom a known control dog homozygous for the bl allele and

IOVS, May 1999, Vol. 40, No. 6

similarly separated on a polyacrylamide gel. This resulted inidentification of the alleles in the dogs homozygous for eitherthe b J or the b2 allele. Those that contained only the b2 alleleresulted in heteroduplex by mixing with the PCR productsfrom the bl control sample, whereas those that contained thebl allele did not result in a heteroduplex when mixed with thebl control sample.

The third polymorphic site represented a single nucleo-tide polymorphism (G/A) and did not alter any restriction site.A restriction site was therefore created by using a mismatchedprimer (APOH-13, ATTACAAAATAGGGGGAAAGGC) in whicha base (G) was changed from T in the original sequence. Thiscreated an Act1 site (5'-GGC'G-3') in the PCR product from theallele (ci) containing a G. Amplification of the other allele (c2)containing an A in the polymorphic site (5'-GGCA-3') wouldnot create the restriction site. For screening, a 158-bp fragmentwas amplified from the genomic DNA by PCR using primersAPOH-10 and APOH-13 (Table 1) and digested with Acil, whichproduced 133-bp and 25-bp fragments when the polymorphicsite contained a G (Fig. 3A, right).

Linkage of APOH Gene to the prcd Locus andIts Localization Using Canine-Rodent HybridCell Lines

Different subsets of the colony were informative for at leastone polymorphism, which resulted in six haplotypes (Fig. 4).These haplotypes then were used to analyze the linkage be-tween the APOH gene and the prcd disease locus. The resultsrevealed 3 recombinants among the 70 informative dogs (lodscore = 15.09; 8 = 0.043). The location of the APOH gene onCFA9 in relation to other markers and genes previously de-scribed in the prcd interval6 was determined ([TK1, GALK1,prcd) — [MYL4, APOH, C09.173, C092263] — [RARA,RAPD] —C09.250—C09.474—NF1) and is illustrated in Figure 5.

The microsatellite markers (C09.173, C09-250, andC09-474) of this linkage group, and canine-specific STS for twogenes (breast cancer 1 susceptibility gene, BRCA1; and cyclicguanosine monophosphate phosphodiesterase y-subunit gene,PDE6G) have been typed in canine-rodent hybrid cell lines.23

The location of BRCA1 in CFA9 has been confirmed,5 and therelative location of PDE6G in the linkage group will be deter-mined once markers associated with the gene are detected.Identification of a canine-specific STS for the canine APOHgene in the same cell lines (Table 2) provided physical evi-dence of location of the gene in the same canine chromosomecontaining the prcd locus and is consistent with the observedlinkage of the gene with the disease locus.

APOH and prcd 1235

CFA9 - prcd interval

Distance fromprcd (cM)

o

0.0

3.3

ti5.1

9.1

11.5

15.6

24.4

32.0

>TK1GALKI, prcd

/MYL4APOH

RARA

RAPD

250

474

NF1

FIGURE 5- Linkage map of theprcrf-interval on canine chromosome 9.The location of APOH is shown relative to that of prcd and other locicorresponding to gene-based and anonymous markers. Loci listed inbrackets indicate no recombinations observed between these markersin the present study. Order and distances shown are based on two-point linkage analyses of the data in this study. Figure modified fromAcland GM, Ray K, Mellersh CS, et al. Linkage analysis and comparativemapping of canine progressive rod- cone degeneration (prcd) estab-lishes potential locus homology with retinitis pigmentosa (RP17) inhumans. Proc Natl Acad Sci USA. 1998;95:3048-3053.

DISCUSSION

We examined the APOH gene as a positional candidate forprcd. This was prompted by our recent finding that places prcdin a region homologous to the RP17 locus, in the MYL4 andTK1 interval,6 and by the comparative mapping studies thatshow conservation of gene order between dog chromosome 9and human chromosome 17q.5 By two-point analysis, we haveplaced the APOH gene on the prcd linkage map. Using canine-rodent somatic cell hybrids, we have shown that the gene isphysically located in the same chromosome that has the. prcdlocus. Although the results were not conclusive, we failed todetect a difference in the expression of APOH between ho-

mozygous normal (+/+) and j5ra/-affected (prcdlprcd) dogsby RT-PCR, western blot, and immunocytochemistry. It is pos-sible, however, that a missense mutation in the gene, causallyassociated with the disease, would not be detected by thesestrategies. To establish an association between the APOH geneand the disease locus, we identified polymorphic markers inAPOH and by linkage analysis excluded the gene, because wehad identified 3 recombinants out of 70 in the study pedigree.Although there was a strong linkage between the gene and thedisease locus (lod score 15.09), the linkage distance betweenthe two was approximately 4.3 cM.

Our previous work in the prcd model examined the role ofplasma lipid abnormalities in the disease process.1516'28 These

1236 Gu et al. IOVS, May 1999, Vol. 40, No. 6

TABLE 2. Physical Localization of APOH Gene in theCanine-Rodent Hybrid Cell Lines Containing OtherMarkers Linked to prccl

Cell Line*

Marker/Genef mDE-6 mDE-15

RAPDAPOHBRCA1PDE6GC09.173CO9.25OC09.474

• Cell lines identified according to Langston et al.23

f Marker: RAPD, C0SU73, C09.250, and C09474; Gene: BRCA1,and PDE6G.

studies focused on DHA, although other lipid abnormalitiesalso are present in prccl—for example, a 20% to 30% decreasein serum cholesterol, and reduced secretion by the liver ofvery-low-density-lipoprotein- containing DHA.2930 We knowthat lipid abnormalities (low plasma cholesterol and DHA, lowROS DHA levels) are inherited and associated with prcd. How-ever, there are animals with prcd and normal plasma DHAlevels without supplementation, and these have a slower dis-ease phenotype (Aguirre and Acland, unpublished data, 1998).

In recent experimental matings (affected, low DHA Xnormal high DHA), we have found that the low-DHA pheno-type is not expressed in Fl obligate heterozygous dogs, andresegregates in the F2 generation with /?ra/-affected dogs hav-ing a lower DHA (P < 0.05) level. This indicates that thelow-DHA phenotype is clearly inherited as a recessive traitassociated with the prcd locus.3' However, recombinationsbetween the prcd and low DHA phenotypes occur in the F2generation, and the number of recombinants (1/17) is similarto that found between APOH and prcd. In our laboratory, weare examining those genes that can regulate DHA levels inplasma and the ROS; because of its role in phospholipid me-tabolism, l l l 2 " ! its retinal expression, and its close linkage tothe prcd locus, the APOH gene merits further scrutiny in thesestudies.

Our current studies have found that the APOH transcriptis expressed in retina and that the protein is localized to thephotoreceptor outer segments and possibly the IPM. Littleinformation is available, however, on the localization of theprotein in the IPM and the expression and regulation of APOHin the RPE and retina or whether it plays a role in phospholipidmetabolism in the retina, particularly in the ROS. More re-cently, we have found that APOH is expressed in the RPE byRT-PCR (data not shown). An affirmative answer would indi-cate that, at least in prcd and the associated low DHA pheno-type, APOH may play a role as a modifying gene locus, assuggested for apolipoprotein E alleles in human patients withAlzheimer disease.32 Concerning the retina, results of studiesin the late 1980s suggested disease-specific differences in se-rum lipoproteins in retinitis pigmentosa,33'34 and, more re-cently, it has been proposed that the e4 allele of apoE producesa retinitis punctata albescens phenotype rather than an retinitispigmentosa phenotype in patients with an Argl35Trp muta-

tion in the rhodopsin gene.35 Studies on APOH and retinal DHAmetabolism are currently in progress.

Acknowledgments

The authors thank Ilyas Kamboh for helpful discussions and commentsand Jill Czarnecki for help with the illustrations.

References1. Aguirre G, Alligood J, O'Brien P, et al. Pathogenesis of progressive

rod- cone degeneration in miniature poodles. Invest OphthalmolVisSci. 1982;23:6l0-630.

2. Aguirre GD, Acland GM. Variation in retinal degeneration pheno-type inherited at the prcd locus. Exp Eye Res. 1988;46:663-687.

3. Aguirre GD, Acland GM. Inherited retinal degeneration in theLabrador retriever dog: a new animal model of RP? Invest Oph-thalmol Vis Sci. 1991;32:889.

4. Mellersh CS, Langston AA, Acland GM, et al. A linkage map of thecanine genome. Genomics. 1997;46:326-336.

5. Werner P, Raducha MG, Prociuk U, et al. Physical and linkagemapping of human chromosome 17 loci to dog chromosomes 9and 5. Genomics. 1997;4274-4282.

6. Acland GM, Ray K, Mellersh CS, et al. Linkage analysis and com-parative mapping of canine progressive rod-cone degeneration(prcd) establishes potential locus homology with retinitis pigmen-tosa (RP17) in humans. Proc Natl Acad Sci USA. 1998;95:3048-3053.

7. Gu WK, Acland GM, Langston AA, et al. Identification of a RAPDmarker linked to progressive rod-cone degeneration in dogs.Mamm Genome. 1998;9:740-744.

8. Haagerup A, Kristensen T, Kruse TA. Polymorphism and geneticmapping of the gene encoding human beta2-glycoprotein I tochromosome 17. Cytogenet Cell Genet. 1991;58:2005.

9. Steinkasserer A, Cockburn DJ, Black DM, et al. Assignment ofapolipoprotein H (APOH: beta-2 glycoprotein I) to human chro-mosome 17q23-qter determination of the major expression site.Cytogenet Cell Genet. 1992;60:31-33.

10. Averna M, Paravizzini G, Marino G, et al. Liver is not the unique siteof synthesis of beta-2-glycoprotein I (apolipoprotein H): Evidencefor an intestinal localization. IntJ Clin Lab Res. 1997;27:207-212.

11. Kamboh MI, Ferrell RE. Apolipoprotein H polymorphism and itsrole in lipid metabolism. Adv Lipid Res. 1991;1:9-18.

12. Cassader M, Ruiu G, Gambino R, et al. Influence of apolipoproteinH polymorphism on levels of triglycerides. Atherosclerosis. 1994;11:45-51.

13. Sanghera DK, Wagenknecht DR, Mclntyre JA, et al. Identificationof structural mutations in the fifth domain of apolipoprotein H(beta2-glycoprotein I) which affect phospholipid binding. HumMol Genet. 1997;6:3H-3l6.

14. Polz E, Kostner GM. The binding of /32-glycoprotein-I to humanserum lipoproteins: distribution among density fractions. FEESLett. 1979;90:1305-1312.

15. Anderson RE, Maude MB, Alvarez RA, et al. Plasma lipid abnormal-ities in the miniature poodle with progressive rod- cone degener-ation. Exp Eye Res. 1991 ;52:349-356.

16. Aguirre GD, Acland GM, Maude MB, et al. Diets enriched indocosahexaenoic acid fail to correct progressive rod-cone degen-eration (prcd) phenotype. Invest Ophthalmol Vis Sci. 1997;38:2387-2407.

17. Aguirre GD, O'Brien P. Morphological and biochemical studies ofcanine progressive rod-cone degeneration: 3H-fucose autoradiog-raphy. Invest Ophthalmol Vis Sci. 1986;27:635-655.

18. Sambrook J, Fritsch EF, Maniatis T. Molecular Clotting: A Labora-tory Manual. Cold Spring Harbor, NY: Cold Spring HarborLaboratory; 1989;280-281.

19. Chomczynski P, Sacchi N. Single-step method of RNA isolation byacid guanidinium thiocyanate-phenol-chloroform extraction. AnalBiochem. 1987;l62:156-159.

20. Gu WK, Aguirre GD, Ray K. Detection of single nucleotide poly-morphisms. BioTechniques. 1998;24:836-837.

IOVS, May 1999, Vol. 40, No. 6 APOH andprcd 1237

21. Lathrop GM, Lalouel JM. Easy calculations of LOD scores andgenetic risks on small computers. Am]Hum Genet. 1984;36:460-465.

22. Terwilliger JD, Ott J. Handbook of Human Genetic Linkage.Baltimore: Johns Hopkins University Press; 1994.

23. Langston AA, Mellersh CS, Neal CL, et al. Construction of a panelof canine-rodent hybrid cell lines for use in partitioning of thecanine genome. Genomics. 1997;46:317-325.

24. Verdugo ME, Ray J. Age-related increase in activity of specificlysosomal enzymes in the human retinal pigment epithelium. ExpEye Res. 1997;65:231-240.

25. Laemmli UK. Cleavage of structural proteins during the assemblyof the head of the bacteriophage T4. Nature. 1970;22:680-685.

26. Huang JC, Chesselet MF, Aguirre GD. Decreased opsin mRNA andimmunoreactivity in progressive rod-cone degeneration (prcd):cytochemical studies of early disease and degeneration. Exp EyeRes. 1994;58:17-30.

27. Gropp KE, Huang JC, Aguirre GD. Differential expression of pho-toreceptor-specific proteins during disease and degeneration inthe progressive rod- cone degeneration (prcd) retina. Exp Eye Res.1997;64:875-886.

28. Anderson RE, Maude MB, Alvarez R, et al. Plasma lipid abnormal-ities in />rcd-affected miniature poodles and Abyssinian cat. In:Degenerative Retinal Disorders: Clinical and Laboratory Inves-tigations. Boca Raton, FL: CRC Press; 1991;131-l42.

29. Bazan NG, Rodriguez-DeTurco EB. Pharmacological manipulationof docosahexaenoic-phospholipid biosynthesis in photoreceptorcells: implications in retinal degeneration. / OCM/ Pharmacol.1994;10:591-604.

30. Willmott NJ, Hussain AA. Non-retinal abnormalities associated withprogressive retinal atrophy (PRA) in the miniature poodle. Exp EyeRes. 1996;63:527-533.

31. Aguirre GD, Anderson RE, Maude M, et al. Heritability studies oflow plasma docosahexaenoic acid in the. prcd-dog [ARVO Ab-stract]. Invest Ophthalmol Vis Sci. 1996;37(3):S1046. Abstract nr4800.

32. Corder EH, Saunders AM, Risch NJ, et al. Protective effect ofapolipoprotein E type 2 allele for late onset Alzheimer disease. NatGenet. 1994;7:180-184.

33- Jahn CE, Oette K, Esser A, et al. Increased prevalence of apoli-poprotein e-2 in patients with retinitis pigmentosa. OphthalmicRes. 1987; 19:285-288.

34. Converse CA, Keegan WA, Huq L, et al. Further epidemiologicalstudies on lipid metabolism in retinitis pigmentosa. In: LaVail MN,Anderson RE, Hollyfield JG, eds. Progress in Clinical and Biolog-ical Research. New York: Alan R. Liss; 1989;39-48.

35. Souied E, Soubrane G, Benlian P, et al. Retinitis punctata albescensassociated with the Argl35Trp mutation in the rhodopsin gene.Am J Ophthalmol. 1996;121:19-25.