Abstract Gaucher disease, an inherited glycolipid storagedisorder, is caused by a deficiency of the catabolic enzymeglucocerebrosidase (EC 3.2.1.45). The gene for humanglucocerebrosidase is located on chromosome 1q21 andhas a highly homologous pseudogene situated 16 kbdownstream. We report two novel polymorphic sequencesin the glucocerebrosidase gene region: the first consists ofa variable number of dinucleotide (CT) repeats located 3.2kb upstream from the glucocerebrosidase gene, and thesecond is a tetranucleotide (AAAT) repeat found betweenthe glucocerebrosidase gene and its pseudogene, 9.8 kbdownstream from the functional gene. These polymorphicsequences, along with a previously reported PvuII poly-morphism in intron 6 of the glucocerebrosidase gene, wereanalyzed in patients with Gaucher disease (n=106) and intwo normal control populations, one of Ashkenazi Jewishancestry (n=72) and the second comprising non-Jewish in-dividuals (n=46). In these samples, strong linkage disequi-librium was found between mutations N370S,c.84–85insG, and R463C and specific haplotypes; no sig-nificant linkage disequilibrium was found when examininghaplotypes of patients with the L444P mutation. Studies ofthese polymorphic sites in several instances also led to therecognition of genotyping errors and the identification ofunusual recombinant alleles. These new polymorphic sitesprovide additional tools for mutational screening andfounder effect studies of Gaucher disease.

E. K. Lau · N. Tayebi · S. L. Winfield · V. Koprivica · D. L. Stone E. I. Ginns · E. Sidransky (✉)Clinical Neuroscience Branch, NIMH, Bldg. 49, Room B1EE16, 49 Convent Drive, MSC4405, Bethesda, MD 20892-4405, USAe-mail: sidranse@irp.nimh.nih.gov, Tel.: +1-301-496-0373, Fax: +1-301-402-6438

L. J. IngrahamCenter for Professional Psychology, George Washington University,Washington, DC, USA

A. ZimranGaucher Clinic, Department of Medicine, Shaare Zedek MedicalCenter, Jerusalem, Israel

Introduction

Gaucher disease results from an inherited deficiency of thelysosomal enzyme glucocerebrosidase (EC 3.2.1.45). Al-though the gene for glucocerebrosidase was mapped tochromosome 1q21 and cloned over a decade ago (Sorge etal. 1985b; Tsuji et al. 1986; Horowitz et al. 1989), our un-derstanding of the genetic component leading to the phe-notypic variation encountered in Gaucher disease is stillincomplete. The existence of a highly homologouspseudogene located only 16 kb downstream from the hu-man glucocerebrosidase gene complicates genotypic anal-yses (Beutler and Grabowski 1995; Tayebi et al. 1996).

Hum Genet (1999) 104:293–300 © Springer-Verlag 1999

ORIGINAL INVESTIGATION

Elaine K. Lau · Nahid Tayebi · Loring J. Ingraham Suzanne L. Winfield · Vuk Koprivica · Deborah L. StoneAri Zimran · Edward I. Ginns · Ellen Sidransky

Two novel polymorphic sequences in the glucocerebrosidase gene region enhance mutational screening and founder effect studiesof patients with Gaucher disease

Received: 5 December 1998 / Accepted: 14 January 1999

While more than 100 different point mutations have beenidentified in patients with Gaucher disease (Beutler andGrabowski 1995; Grabowski and Horowitz 1997; Beutlerand Gelbart 1998), three point mutations, N370S,c.84–85insG (also known as 84GG), and L444P are thosemost commonly observed. The frequency of specificGaucher mutations varies between patients of differentethnic groups. In the Ashkenazi Jewish population, inwhich the incidence of Gaucher disease is greatly in-creased, approximately 90% of the mutations are eitherN370S or c.84–85insG, whereas in the non-Jewish popula-tion L444P is commonly observed (Beutler and Grabowski1995). Mutant and fusion alleles arising from recombina-tion events between the gene and the pseudogene have alsobeen reported in both populations (Beutler and Grabowski1995; Zimran et al. 1990b; Tayebi et al. 1998). Further-more, the region surrounding the humanglucocerebrosidase gene appears to be very “gene rich.” Anovel, convergently transcribed gene, metaxin, is locatedimmediately downstream of the glucocerebrosidasepseudogene, and its pseudogene lies within the intergenicregion (Long et al. 1996). Recently, three additional genes,clk2, propin1, and cote1, have been identified within the32-kb region upstream of the glucocerebrosidase gene(Winfield et al. 1997). The possible contribution of thesecontiguous genes to the phenotypes encountered in pa-tients with Gaucher disease is currently under investiga-

tion. Because of both the large number of possible muta-tional events and the current limitations in predicting phe-notype from genotype, genetic counseling for the disordercan be quite challenging.

Twelve polymorphic sites have been previously report-ed in the glucocerebrosidase gene region (Beutler et al.1992). However, only two major haplotypes, termedPv1.1+ and Pv1.1– based upon the presence or absence ofa PvuII restriction site in intron 6 of theglucocerebrosidase gene (Sorge et al. 1985a), are de-scribed in Caucasians. Linkage disequilibrium has beenshown between the common Ashkenazi Jewish mutationN370S and the Pv1.1–haplotype (Zimran et al. 1990a;Lacerda et al. 1994; Amaral et al. 1997; Cormand et al.1998; Demina et al. 1998; Rockah et al. 1998) while muta-tion c.84–85insG is in disequilibrium with the Pv1.1+ hap-lotype (Beutler et al. 1992; Rockah et al. 1998). However,both of these Pv1.1 polymorphisms are common in thenormal Caucasian population (Beutler et al. 1992; Glennet al. 1994), limiting their usefulness in determining an an-cestral chromosome. A G→A transition at nucleotide 5470in intron 7 (g.5470G→A), resulting in the disruption of aBsu36I restriction site, was also found to be polymorphic(Amaral et al. 1997), but this polymorphism has only beendescribed in Gaucher patients from Portugal.

Studies of a highly polymorphic trinucleotide repeat inthe gene for pyruvate kinase (PKLR), which is closelylinked to the glucocerebrosidase gene (Glenn et al. 1994),reveal that 96% of alleles carrying N370S also carry oneallelic form of the PKLR polymorphism, A1, which ispresent in only 6.7% of the control population (Rockah etal. 1998). In another report a PKLR polymorphism at nu-cleotide 1705 was evaluated in 195 Gaucher patients withgenotype N370S/N370S, and all were found to share ho-mozygosity for the –/– PKLR haplotype (Demina et al.1998). Several microsatellite markers located within a 3.2-cM region surrounding the human glucocerebrosidasegene have also recently been analyzed in Spanish patientswith Gaucher disease in an attempt to establish a putativehaplotype associated with the N370S mutation (Cormandet al. 1998).

In this study, we have identified two novel polymorphicsequences in the glucocerebrosidase gene region (Fig. 1).The first polymorphic sequence, 5GC3.2, is a dinucleotiderepeat (CT) located 3.2 kb upstream from theglucocerebrosidase gene. The second polymorphic site,ITG6.2, is a tetranucleotide repeat (AAAT) located be-tween the glucocerebrosidase gene and its pseudogene, 9.8kb downstream of the functional gene. The novel polymor-phic sites described in the report, together with the previ-ously reported PvuII polymorphism, are representative of

a very defined region immediately surrounding the humanglucocerebrosidase gene on 1q21, and are valuable for theidentification of founder chromosomes that are linked tospecific mutated alleles in affected patients. Strong linkagedisequilibrium between certain mutations and a particularfounder chromosome in a defined population can facilitatethe molecular diagnosis of Gaucher disease, especially incases where common mutations are absent.

Subjects and methods

Subjects

Samples from patients with Gaucher disease were collected with in-formed consent. Genomic DNA was isolated from blood, fibroblastcell lines, and/or Epstein-Barr virus-transformed lymphoblasts aspreviously reported (Sambrook et al. 1989). The patients included 75subjects with type 1 Gaucher disease, 10 with type 2 Gaucher dis-ease, and 21 with type 3 Gaucher disease. Of the type 1 patients, 64were of Ashkenazi Jewish ancestry while all of the type 2 and type 3patients were non-Ashkenazi. All 29 of the patients with genotypeN370S/N370S and all 22 patients with genotypeN370S/c.84–85insG were of Ashkenazi Jewish ancestry.

The DNA from the Ashkenazi control group was extracted fromcord blood samples collected in hospitals in Israel as previously de-scribed (Zimran et al. 1991). The DNA was screened for two com-mon mutations, N370S and c.84–85insG, and all samples positivefor either mutation were excluded from this study. The DNA fromthe non-Jewish control group was obtained with informed consentfrom individuals without Gaucher disease who were participating inother studies and from Bios Laboratories (New Haven, Conn.).

Mutational analysis

DNA samples from the patients with Gaucher disease were screenedfor Gaucher mutations including N370S, c.84–85insG, L444P,R463C, a 55-bp deletion (g.5296–5350 del), IVS2+1G>A, andD409H, as previously described (Sidransky et al. 1994, 1996; Tayebiet al. 1996, 1998). DNA from patients identified as having an allelewith the mutation L444P was subsequently sequenced to determinewhether the mutation was part of a complex recombinant allele. Mu-tations V394L, D399N, P415R, R163X, R170C, R120W, S196P,R131L, R257Q, and H255Q were detected by polymerase chain re-action (PCR) amplification of specific exons, followed by sequenc-ing, as previously described (Tayebi et al. 1998).

Genotyping at specific polymorphic sites

The two new polymorphic sites, 5GC3.2 and ITG6.2, were initiallyidentified by sequence analysis of a 75-kb region surrounding thehuman glucocerebrosidase gene (Winfield et al. 1997). Fluorescentlylabeled PCR primers were used to amplify the regions containing therepeats. Genotyping was performed using an ABI 373 sequenc-er/genotyper, using 373A Data Collection Software (PE-AppliedBiosystems). Samples were heat denatured and electrophoresed at 30W for 12 h on a 4.75% acrylamide gel using filter set B and a

294

Fig. 1 The locations of threepolymorphic sites, 5GC3.2,PvuII, and ITG6.2, on chromo-some 1q21. GC gluco-cerebrosidase gene, ψGCglucocerebrosidase pseudogene.The subscript number denotesthe range of repeat copiesencountered

TAMRA-labeled Genescan-2500 size standard. The data were ana-lyzed with the Genescan Analyser and the Genotyper 1.1.1 software.The region containing the 5GC3.2 polymorphic site was amplifiedwith the following primers: forward, 5′-TTCAATCGCCCCCATCC-ACC-3′; and reverse, HEX-labeled 5′-TCAGAGCCCTTCCTCAAG-3′. The PCR reactions, in 10-µl volumes, contained 200–500 ngDNA, 0.4 µM of each primer, 1 µl 10×PCR buffer (100 mM TRIS-HCl, 15 mM MgCl2, 500 mM KCl, pH 8.3, 2 µM of each dNTP, and0.25 U Taq polymerase (Boehringer-Mannheim). The PCR was per-formed with a 4-min denaturation at 94°C, followed by 30 cycles of94°C for 30 s, 62°C for 45 s, 72°C for 1 min, with a final extensionat 72°C for 10 min.

Screening for the ITG6.2 polymorphism was performed by PCRamplification using the following primers: forward 5′-CACATGAG-GTCAGGTGTTTG-3′ and reverse TET-labeled 5′-GCAAAGG-AGTGGTGAACTTC-3′. PCR reactions were prepared in 10-µl vol-umes, containing 200–500 ng DNA, 0.4 µM of each primer, 1 µl 10×standard PCR buffer (Boehringer-Mannheim), and 2 µM of eachdNTP. The Taq polymerase (0.25 U) was pretreated with TaqStartantibody (ClonTech) for 5 min before amplification. The PCR ampli-fication conditions were as follows: 4 min denaturation at 94°C, 30cycles of 94°C for 30 s, 64.5°C for 45 s, 72°C for 1 min, with a finalextension at 72°C for 10 min.

The PvuII polymorphic site was evaluated by PCR amplificationusing a forward primer located in exon 6 (5′-TCAAGACCAATGGA-GCGGTG-3′) and a reverse primer from intron 6 (5′-GAAAGG-TCATGAATGATCCG-3′). Each PCR reaction was prepared in a totalvolume of 25 µl containing 500 ng DNA, 0.5 µM of each primer, 2.5µl 10× standard PCR buffer (Boehringer-Mannheim), 2 µM eachdNTP, and 2.0 U Taq polymerase. The PCR was performed with a 4-min denaturation at 94°C, followed by 30 cycles of 94°C for 30 s,60°C for 45 s, 72°C for 1 min, with a final extension at 72°C for 10min. The PCR product (200 ng) was digested with 10 U of the restric-tion enzyme PvuII (Gibco BRL) for 3 h, electrophoresed on a 2% aga-rose gel, and visualized by ethidium bromide staining.

Screening for the presence of the restriction site polymorphismg.5470G>A was performed by a Bsu36I (Gibco BRL) restriction di-gestion of PCR-amplified DNA from the region (Amaral et al.1997), and confirmed by direct sequencing as previously described(Tayebi et al. 1998).

Sequencing

The amplified PCR product was gel electrophoresed, the band puri-fied using Gene Clean (BIO101), and 50 ng of this PCR product wascycle sequenced using the Prism dye-terminator kit with each primerat a final concentration of 0.5 µM. The samples were column cleanedusing a Sephadex G50 spin column, heat denatured, and then elec-trophoresed on a 4.75% acrylamide gel at 30 W for 14 h using anABI 373A Automated Sequencer and Data Collection Software (PE-Applied Biosystems).

Statistical analysis

Haplotype frequencies were estimated using the program EM-HAPFRE (Excoffier and Slatkin 1995). This method uses the expec-tation-maximization (EM) algorithm to determine the gametic phaseof haplotypes in diploid individuals who are heterozygous at morethan one locus (Excoffier and Slatkin 1995). The EMHAPFRE pro-gram was also used to calculate the frequencies of selectedhaplotypes associated with specific mutations. The linkage disequi-librium was calculated using the equation δ=(FD–FN)/(1–FN), whereFN is the frequency of the haplotype of a given allele in the controlpopulation and FD the frequency of the haplotype of an allele associ-ated with a specific mutation in the patient population (Bengtssonand Thomson 1981). The significance of differences in haplotypefrequencies between patient and control groups was analyzed using aone-sided Z-test (Sokal and Rohif 1969).

Results

Allelic frequencies in the control populations

The allelic frequencies of the 5GC3.2, PvuII, and ITG6.2polymorphisms were determined in the Ashkenazi (144 al-leles) and in the non-Jewish (92 alleles) control samples.Four different allelic forms (218, 220, 222, 224 bp) wereobserved for the 5GC3.2 polymorphism in both controlpopulations while an additional allele (226 bp) was en-countered in one Ashkenazi control individual (Fig. 2A).In both control groups there were four allelic forms of theintergenic ITG6.2 polymorphism (314, 318, 322, 326 bp),while an additional form (330 bp) was observed in threenon-Jewish controls (Fig. 2B). Comparison of the allelicdistribution of each polymorphism revealed that the 318-bp form of the ITG6.2 allele was more prevalent in theAshkenazi group (frequencies 0.36 vs 0.25), and the 326-bp form of the ITG6.2 allele was less prevalent in theAshkenazi group (frequencies 0.11 vs 0.24). The frequen-cies of the alleles of the PvuII polymorphism in both thenon-Jewish control group (0.43 for the Pvu1.1+ allele;0.57 for the Pvu1.1– allele) and the Ashkenazi controlpopulation (0.39 for the Pvu1.1+ allele; 0.61 for thePvu1.1– allele) were similar to those previously describedfor a general population control group (0.35 for thePvu1.1+ allele; 0.65 for the Pvu1.1– allele) (Sorge et al.1985a).

Allelic frequencies in patients with Gaucher disease

The allelic frequencies of polymorphisms of 5GC3.2, PvuIIand ITG6.2 were determined in 106 affected individuals, in-cluding 29 patients with mutations N370S/N370S, 22 pa-tients with N370S/c.84–85insG, 10 patients withL444P/L444P, and 6 patients with N370S/L444P (Fig. 2A,B). Neither the 5GC3.2/226-bp allele nor the ITG6.2/314-bp or 330-bp alleles were observed in any of the patients.The ITG6.2/318-bp allele was seen more frequently in thepatient group (0.73 vs 0.25 and 0.36 in the non-Jewish andAshkenazi control groups, respectively). In patients withN370S/N370S, the allelic frequency of 5GC3.2/222 bp,Pvu1.1– and ITG6.2/318 bp differed markedly from con-trols being 1.00 vs 0.60 and 0.71 for 5GC3.2/222, 1.00 vs0.57 and 0.61 for Pvu1.1– and 0.98 vs 0.25 and 0.36 forITG6.2/318, for patients, and non-Jewish and Ashkenazicontrols, respectively.

Table 1 shows the 5GC3.2, PvuII, and ITG6.2 allelesfound in 67 patients with the more common mutations.The frequency of alleles 222/222, –/–, and 318/318 dif-fered dramatically among the Ashkenazi patients with mu-tations N370S/N370S as compared with the Ashkenazicontrol population, at 0.97 vs 0.11, respectively. Also, thefrequency of alleles 220/222, –/+, and 318/318 were mark-edly different among the Ashkenazi patients with muta-

295

tions N370S/c.84–85insG as compared with the Ashkenazicontrol group, being 1.0 vs 0, respectively.

There were seven combinations of 5GC3.2/PvuII/ITG6.2 alleles represented among ten L444P/L444P pa-tients (Table 1). Similar to the non-Jewish control sample,none of these were predominant in the L444P/L444P pa-tient group.

Our analysis of four patients heterozygous for theIVS2+1G>A mutation revealed that they did not share thesame haplotype at the three polymorphic sites. In contrastto previous studies indicating that mutation IVS2+1G→Ais always linked to a Pvu1.1– allele (Glenn et al. 1994;Rockah et al. 1998), we found one individual with geno-type IVS2+1G>A/R463C who was homozygous for thePv1.1+ allele.

296

Fig. 2A, B Frequency of allelesin controls and groups of pa-tients with Gaucher disease: A5GC3.2 B ITG6.2

B

A

Patients with other Gaucher genotypes were also in-cluded in this study. There was at least one patient witheach of the following genotypes: N370S/R463C,N370S/g.5296–5350del, N370S/complex, R120W/S196P,R131L/R131L, g.5296–5350del/R257Q, D399N/R463C,L444P/R170C, L444P/R163X, L444P/R463C, L444P/x,R463C/c.84–85insG, R463C/IVS2+1G>A, N370S/IVS2+1G>A, D399N/R463C, D409H/D409H, D409H/com-plex, L444P/V394L, L444P/P415R, L444P/complex, com-plex/complex. Direct sequencing revealed that these in-cluded several different complex alleles with the sites ofrecombination occurring in intron 8, exon 9, intron 9, orexon 10 (Tayebi et al. 1998).

Among three patients carrying the 55-bp deletion,g.5296–5350del, there were at least two haplotypes at5GC3.2, PvuII and ITG6.2, including 220/+/322 and222/–/318. Examination of one patient who was homozy-gous for a complex allele (Rec NciI) that encompassedmutations L444P, A456P, and V460V and four others het-erozygous for a complex allele indicated that two or morehaplotypes (222/–/318, and 220/+/322) were represented.Among the ten patients carrying the mutation R463C, asingle haplotype (220/+/322) could be deduced by exam-ining the alleles present in N370S/R463C, L444P/R463C,and R463C/c.84–85insG patients. Since there were noR463C/R463C homozygotes, these conclusions are basedupon the assumption that mutations N370S andc.84–85insG are associated with the specific haplotypesdiscussed above.

Of great interest is the finding that two patients actuallyhad more than two ITG6.2 alleles. The 318-bp, 322-bp,and 326-bp alleles were detected in each of these individu-als. Southern blot analyses using the restriction enzymeSspI have demonstrated that these patients may have a du-plication in the glucocerebrosidase gene region (Sidranskyet al. 1998).

The 5470G>A polymorphism

Only 1 of the 29 patients with genotype N370S/N370Sscreened for the g.5470 G>A polymorphism in intron 7was positive, and only on one allele. Interestingly, this pa-tient was the single individual with N370S/N370S whowas not homozygous for the 222/–/318 haplotype.

Haplotype frequencies

The haplotype frequencies at the three polymorphic siteswere calculated using the EMHAPFRE program and com-pared in the Ashkenazi controls, Ashkenazi patients, non-Jewish controls and non-Jewish patients (Table 2). AmongAshkenazi individuals, the frequencies of two haplotypes,222/–/318 (P<0.001) and 220/+/318 (P<0.001), were sig-nificantly higher in patients than in the corresponding con-trols. Among the non-Jewish patients, haplotype222/–/318 also occured more frequently than in the corre-sponding control group (P<0.001). However, the

297

Tab

le 1

The

fre

quen

cy o

f 5G

C3.

2, P

vuII

and

IT

G6.

2 al

lele

s ob

serv

ed in

67

Gau

cher

pat

ient

s w

ith

com

mon

gen

otyp

es a

nd in

con

trol

s

218/

222,

–/–

220/

220,

+/+

22

0/22

2,–/

+

220/

222,

–/+

22

2/22

2,–/

– 22

2/22

2,–/

+

222/

222,

–/+

22

2/22

2,–/

– 22

2/22

2,+

/–

Oth

er31

8/32

232

2/32

231

8/31

831

8/32

231

8/31

831

8/32

232

2/32

232

2/32

232

2/32

2

N37

0S/N

370S

n=29

00

00

0.97

0.03

00

00

N37

0S/c

.84–

85in

sGn=

220

01

00

00

00

0L

444P

/L44

4Pn=

100.

10.

10

0.1

0.3

0.1

00.

10.

20

N37

0S/L

444P

n=6

00

00.

170.

50.

170.

170

00

Ash

kena

zi c

ontr

ols

n=72

0.03

00

0.11

0.11

0.13

0.04

0.04

0.03

0.50

Non

-Ash

kena

zi c

ontr

ols

n=46

00.

020.

040.

090.

040.

110

0.02

0.02

0.65

222/–/318 haplotype segregated more often with theAshkenazi controls than in the non-Jewish control popula-tion (P<0.001).

Haplotype frequencies of the three polymorphic siteswere also considered by examining the Gaucher mutation(Fig. 3A, B). The distribution varied widely between theAshkenazi and non-Jewish patient groups. Two muta-tion/haplotype combinations, N370S with 222/–/318 andc.84–85insG with 220/+/318, account for 88% of patientalleles in the Ashkenazi patients, but only 11.4% of the al-leles in non-Jewish patients. Among the non-Jewish pa-tients there was no significant mutation/haplotype associa-tion (P<0.001).

Linkage disequilibrium

The degree of linkage disequilibrium between patientswith mutation N370S and the haplotype 222/–/318 was0.972 (FD=0.97, FN=0.28) while the value calculated formutation c.84–85insG and haplotype 220/+/318 was 1.00(FD=1, and FN=0.05). The degree of linkage between mu-tation R463C and 220/+/322 in our samples is 0.872(FD=0.89 and FN=0.14). Linkage disequilibrium calcula-tions were not performed for the other mutations becauseof limitations in sample size.

Discussion

In this study two novel highly polymorphic sites, 5GC3.2and ITG6.2, were identified in the glucocerebrosidasegene region. By studying these two polymorphic sites andthe previously described PvuII polymorphism, a back-ground haplotype could be better established. These analy-ses enable the more consistent identification of ancestralchromosomes, facilitate the tracking of mutations in afamily or a specific population and enhance genotype/phe-notype analyses helpful for genetic counseling. Addition-ally, since the 5GC3.2 and ITG6.2 polymorphic sites arelocated 3′ and 5′ to the glucocerebrosidase gene, respec-tively, we found that in several instances they led to the

recognition of genotyping errors and to the identificationof unusual recombinant, duplication, and/or deletion alle-les.

In our studies of two control populations and 106 pa-tients with Gaucher disease, linkage disequilibrium wasdemonstrated between certain Gaucher mutations and spe-cific haplotypes. Of 29 N370S/N370S patients, 28 werehomozygous for the 222/–/318 haplotype, confirming pre-vious studies indicating that the N370S mutation has aconserved haplotype on 1q21 in the Ashkenazi population(Beutler et al. 1992; Demina et al. 1998; Rockah et al.1998). These findings support the hypothesis that mutationN370S in the Ashkenazi population may have originatedfrom a single founder. However, the genotype of oneAshkenazi patient studied included both the 222/–/318 andthe 222/–/322 haplotypes. Upon further investigation, wefound that this heterozygous patient also carried the un-common g.5470 G>A polymorphism in intron 7 (Amaralet al. 1997). This is the first report of a patient with the

298

Table 2 Frequencies of specific haplotypes in patients and controls

Haplotype Ashkenazi Non-Jewish

Controls Patients Controls Patientsn=72 n=64 n=46 n=42

222/+/322 0.133 0 0.046 0.026222/–/322 0.189 0.010 0.207 0.132222/–/318 0.327 0.783 0.219 0.553222/–/236 0.031 0 0.098 0.052220/+/322 0.155 0.021 0.160 0.160220/+/318 0.021 0.185 0.031 0220/+/326 0.052 0 0.110 0Other 0.092 0 0.067 0.077

Fig. 3A, B Haplotype frequency in patients with Gaucher diseaseassociated with specific Gaucher mutations: A Ashkenazi patients, Bnon-Jewish patients

A

B

g.5470 G>A polymorphism identified outside of the Portu-guese population. This N370S/N370S patient, heterozy-gous at two polymorphic sites, could support the hypothe-sis that a separate, possibly later, event occurred that gaverise to the second N370S allele. Another possible explana-tion is that the patient inherited this allele from a non-Ashkenazi ancestor.

Our study of these polymorphic sites uncovered severalgenotyping errors, demonstrating that these polymorphicsites may be useful in verifying Gaucher mutations. Onepatient, diagnosed by another laboratory as homozygousfor the N370S mutation, had alleles 220/222, –/+, and318/322. Upon direct sequencing of exons 9–11 of theglucocerebrosidase gene, we discovered that this patientwas actually heterozygous for N370S and g.5296–5350del(Beutler et al. 1993; Tayebi et al. 1996). This erroneousmolecular diagnosis occurred because the original PCRscreening, looking only for the most common Gauchermutations, was performed using primers that were comple-mentary to the deleted sequence.

In addition, we found the alleles 220/222, –/–, and318/318 present in three patients reported by another labo-ratory screening for common Gaucher mutations to havethe Gaucher genotype N370S/c.84–85insG. Because of thediscrepancy, we performed direct sequencing, and demon-strated that none of the three patients actually had thec.84–85insG mutation. Thus our study confirms that thec.84–85insG mutation is linked to a single haplotype aspreviously reported (Glenn et al. 1994). Mutationc.84–85insG may also have originated from a singlefounder in the Ashkenazi population. The initial errors ingenotyping underscore the fallibility of the commonlyused screening methods for Gaucher disease mutations.Although direct sequencing of the glucocerebrosidasegene is a more accurate method for detecting and verifyingmutations, it is also more time consuming, expensive, andrather impractical for large-scale screening. Alternatively,the characterization of these three polymorphic sites maybe a more practical means for validating mutational analy-ses by looking for haplotypes that are inconsistent with thepurported mutations.

Haplotype analyses of DNA from parents of patients,when available, may be particularly valuable. This is espe-cially true when homozygosity for a particular mutation isidentified and it is essential to establish that neither chro-mosome has a total or partial gene deletion.

There are several other inconsistencies brought to lightby our current results. The existence of several differenthaplotypes in patients with L444P/L444P supports the hy-pothesis that the L444P mutation arose in several ancestralchromosomes and that the site or region surroundingL444P may be a hotspot for mutation. However, our find-ings are in contrast to a previous report concluding thatmutation L444P is linked to the Pv1.1– genotype in thenon-Jewish population (Tuteja et al. 1993). Our data alsoprovide an exception to previous reports that the mutationIVS2+1G>A is linked to the Pv1.1– allele (Glenn et al.1994; Rockah et al. 1998). Our results may differ fromthose published because we studied a very ethnically di-

verse patient sample. In fact, unlike the patients describedin the previous reports, the individual with IVS2+1G>Awho was homozygous for the Pv1.1+ allele was actuallynon-Jewish.

It appears that certain other Gaucher mutations, includ-ing g.5296–5350del and complex alleles with mutationsL444P, A456P, and V460V, may have derived from sepa-rate ancestral alleles or mutational events. These findingswould support our previous studies indicating that com-plex alleles arise by different mechanisms and sites of re-combination (Tayebi et al. 1998). Despite difficulties indetermining haplotype phase, our findings suggest thatmutation R463C is another mutation associated with aconserved haplotype. Because of the limited sample size,it remains to be seen whether there is linkage disequilibri-um between these and other Gaucher mutations and cer-tain haplotypes, and whether there is a strong correlationwith phenotype. However, because of the efficiency andinformativeness of genotyping the polymorphic sites de-scribed in the report, we suggest that these haplotypescould be exploited to provide a more powerful mutationaldetection strategy.

References

Amaral O, Marcao A, Pinto E, Zimran A, Sa Mirinda MC (1997)Distinct haplotype in non-Ashkenazi Gaucher patients withN370S mutation. Blood Cells Mol Dis 23:415–416

Bengtsson BO, Thomson G (1981) Measuring the strength of associ-ations between HLA antigens and diseases. Tissue Antigens18:356–363

Beutler E, Gelbart T (1998) Hematologically important mutations:Gaucher disease. Blood Cells Mol Dis 24:2–8

Beutler E, Gelbart T, West C (1993) Identification of six newGaucher disease mutations. Genomics 15:203–205

Beutler E, Grabowski GA (1995) Gaucher disease. In: Scriver CR,Beaudet A, Sly W, Valle D (eds) The metabolic and molecularbasis of inherited disease, 7th edn. McGraw-Hill PublishingCompany, New York, pp 2641–2670

Beutler E, West C, Gelbart T (1992) Polymorphisms in the humanglucocerebrosidase gene. Genomics 12:795–800

Cormand B, Grinberg D, Gort L, Chabas A, Vilageliu L (1998) Mo-lecular analysis and clinical findings in the Spanish Gaucher dis-ease population: putative haplotype of the N370S ancestral chro-mosome. Hum Mutat 11:295–305

Demina A, Boas E, Beutler E (1998) Structure and linkage relation-ships of the region containing the human L-type pyruvate kinase(PKLR) and glucocerebrosidase (GBA) genes. HematopatholMol Hematol 11:63–71

Excoffier L, Slatkin M (1995) Maximum-likelihood estimation ofmolecular haplotype frequencies in a diploid population. MolBiol Evol 12:921–927

Glenn D, Gelbart T, Beutler E (1994) Tight linkage of pyruvate ki-nase (PKLR) and glucocerebrosidase (GBA) genes. Hum Genet93:635–638

Grabowski GA, Horowitz M (1997) Gaucher’s disease: molecular,genetic and enzymological aspects. Baillieres Clin Haematol10:635–656

Horowitz M, Wilder S, Horowitz Z, Reiner O, Gelbart T, Beutler E(1989) The human glucocerebrosidase gene and pseudogene:structure and evolution. Genomics 4:87–96

Lacerda L, Amaral O, Pinto R, Aerts J, Sa Mirinda MC (1994) TheN370S mutation in the glucocerebrosidase gene of Portuguesetype 1 Gaucher patients: linkage to the PvuII polymorphism. JInherit Metab Dis 17:85–88

299

Long GC, Winfield S, Adolph KW, Ginns EI, Bornstein P (1996)Structure and organization of the human metaxin gene (MTX)and pseudogene. Genomics 33:177–184

Rockah R, Narinsky R, Frydman M, Cohen IJ, Zaizov R, WeizmanA, Frisch A (1998) Linkage disequilibrium of common Gaucherdisease mutations with a polymorphic site in the Pyruvate Ki-nase (PKLR) gene. Am J Med Genet 78:233–236

Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a lab-oratory manual, 2nd edn. Cold Spring Harbor Laboratory Press,Cold Spring Harbor, New York

Sidransky E, Bottler A, Stubblefield B, Ginns EI (1994) DNA muta-tion analysis of type 1 and type 3 Gaucher patients: How well domutations predict phenotype? Hum Mutat 357:407–410

Sidransky E, Tayebi T, Stubblefield BK, Eliason W, Klineburgess A,Pizzolato P, Cox JN, Porta J, Bottani A, DeLozier-Blanchet CD(1996) The clinical, molecular, and pathological characterizationof a family with two cases of lethal perinatal type 2 Gaucher dis-ease. J Med Genet 33:132–136

Sidransky E, Tayebi N, Stone D, Koprivica V, Krasnewich D, FillanoJJ, Ginns EI (1998) Phenotypic and genotypic characterizationof a patient with Gaucher disease and parkinsonian symptoms.Am J Hum Genet 63:A274

Sokal RR, Rohif FJ (eds) (1969) Introduction to biostatistics. WHFreeman and Company, NY

Sorge J, Gelbart T, West C, Westwood B, Beutler E (1985a) Hetero-geneity in type I Gaucher disease demonstrated by restrictionmapping of the gene. Proc Natl Acad Sci U S A 82:5442–5445

Sorge J, West C, Westwood B, Beutler E (1985b) Molecular cloningand nucleotide sequence of the human glucocerebrosidase gene.Proc Natl Acad Sci U S A 82:7289–7293

Tayebi N, Stern H, Dymarskaia I, Herman J, Sidransky E (1996) 55-base pair deletion in certain patients with Gaucher disease com-

plicates screening for common Gaucher alleles. Am J Med Gen-et 66:316–319

Tayebi N, Cushner S, Kleijer W, Lau E, Damschroder-Williams PJ,Stubblefield B, Den Hollander J, Sidransky E (1997) Prenatal le-thality of a homozygous null mutation in the humanglucocerebrosidase gene. Am J Med Genet 73:41–47

Tayebi N, Reissner KJ, Lau EK, Stubblefield BK, KlineBurgess AC,Martin BM, Sidransky E (1998) Genotypic heterogeneity andphenotypic variation among patients with type 2 Gaucher dis-ease. Pediatr Res 43:571–578

Tsuji S, Choudary PV, Martin BM, Winfield S, Barranger JA, GinnsEI (1986) Nucleotide sequence of cDNA containing the com-plete coding sequence for human lysosomal glucocerebrosidase.J Biol Chem 261:50–53

Tuteja R, Bembi B, Agosti E, Baralle FE (1993) 1448C mutationlinked to the Pv1.1– genotype in Italian patients with Gaucherdisease. Hum Mol Genet 2:781–784

Winfield SL, Tayebi N, Martin BM, Ginns EI, Sidransky E (1997)Identification of three additional genes contiguous to theglucocerebrosidase locus on chromosome 1q21: implications forGaucher disease. Genome Res 7:1020–1026

Zimran A, Gelbart T, Beutler E (1990a) A Linkage of the PvuIIpolymorphism with the common Jewish mutation for Gaucherdisease. Am J Hum Genet 46:902–905

Zimran A, Sorge J, Gross E, Kubitz M, West C, Beutler E (1990b) Aglucocerebrosidase fusion gene in Gaucher disease: implicationsfor molecular, anatomy, pathogenesis, and diagnosis of this dis-order. J Clin Invest 815:219–222

Zimran A, Gelbart T, Westhaver B, Grabowski GA, Beutler E (1991)High frequency of the Gaucher disease mutation at nucleotide1226 among Ashkenazi Jews. Hum Genet 49:855–859

300