gene in three siblings with Bernard-Soulier syndrome ...

From bloodjournal.hematologylibrary.org by guest on July 14, 2011. For personal use only.

1993 81: 2339-2347

Double heterozygosity for mutations in the platelet glycoprotein IX gene in three siblings with Bernard-Soulier syndrome SD Wright, K Michaelides, DJ Johnson, NC West and EG Tuddenham

Information about reproducing this article in parts or in its entirety may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requests Information about ordering reprints may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprints Information about subscriptions and ASH membership may be found online at: http://bloodjournal.hematologylibrary.org/site/subscriptions/index.xhtml

Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036. Copyright 2011 by The American Society of Hematology; all rights reserved.

From bloodjournal.hematologylibrary.org by guest on July 14, 2011. For personal use only.

Double Heterozygosity for Mutations in the Platelet Glycoprotein IX Gene in Three Siblings With Bernard-Soulier Syndrome By Stephen D. Wright, Katerina Michaelides, Daniel J.D. Johnson, Nicholas C. West, and Edward G.D. Tuddenham Bernard-Soulier syndrome (BSS) giant platelets have defective and/or deficient glycoprotein (GP) Ib/lX complexes, causing absent ristocetin-induced aggregation, defective interaction with von Willebrand factor, morphologic abnormality, and a clinical bleedingtendency. Recently several mutations have been described in the platelet GPlba gene in individuals exhibiting the BSS phenotype. We have studied a family with classical BSS, and have excluded lesions at the GPlba locus by restriction fragment length polymorphism linkage analysis. Analysis of the genes for two other components of the platelet GP1b:IX complex, namely GPlbfi and GPIX, showed t w o different missense mutations in the coding region of the GPlX gene: an A- G transition in codon 21 results in conversion of an aspartic acid to glycine and an A G change in codon 45 converts an asparagine res-

idue to serine. Three affected individuals are doubly heterozygous for these mutations, which alter conserved residues in or flanking the GPlX leucine-rich glycoprotein motif. Both mutations create new recognition sites for the enzyme Fnu 4H1; therefore, this enzyme was used to screen 60 normal subjects (120 alleles). Neither mutation was detected in any subject other than direct relatives of the affected individuals. Although low levels of GPlb were demonstrable by both flow cytometry and immunoblot analysis in an affected individual’s platelets, there was no evidence of GPlX immunoreactivity. We propose that expression of abnormal GPlX prevents stable assembly of the GPlb/lX complex, causing BSS in the doubly heterozygous individuals in this family. 0 1993 by The American Society of Hematology.

B

not demonstrable, platelets from patients with the BSS phenotype may sometimes have a normal complement of GPIb: IX complex components.20 Several mutations in the GPIba gene have recently been identified in patients with BSS: a nonsense mutation, producing a truncated protein lacking a transmembrane domain”; substitution of Ala1 56 by Valine in BSS type Bolzano impairs vWF binding2‘; a mutation affecting a conserved residue of one ofthe LRG repeats (Leu57 to Phe) associated with an autosomal dominant form of BSS.” There have been no reports to date of mutations affecting the other components of the GPIbIX complex, although current evidence implicates at least one other ~ o c u s . In ~,~~ this study, we report two missense mutations in the GPIX gene and present evidence supporting the premise that these mutations impede the expression of GP1b:IX complexes in the platelet membrane and are, therefore, responsible for the clinical bleeding tendency in a BSS kindred.

-

ERNARD-SOULIER syndrome (BSS) is a rare congenital bleeding disorder that is usually inherited in an autosomal recessive manner.’ It is characterized by morphologically abnormal giant platelets and reduced or absent ristocetin-induced platelet aggregation (RIPA), not reversed by the addition of normal plasma. The platelet membrane glycoprotein (GP) IbIX complex has consistently been shown to be deficient or defective in platelets from patients with BSS. This is thought to account for the defective, von Willebrand factor (vWF)-mediated platelet adhesion to exposed subendothelium manifested by BSS platelets. At least four separate gene products may fail to be expressed normally in the BSS platelet membrane, ie, GPIba chain, GPIbP chain, GPIX, and GPV.2-6 The a chain of GPIb has a molecular weight of approximately 140 Kd. Its extracellular domain has been described as a flexible rod that projects from the platelet membrane and also contains binding sites for both thrombin and v W F . ~ , ~ The GPIbP chain is smaller (24 Kd) and disulphide-linked to GPIba.’ GPIX (22 Kd) is noncovalently bound to GPIb and recent evidence suggests that GPV (82 Kd) is also associated with the complex in a noncovalent manner.’O All four glycoproteins are members of the leucine-rich glycoprotein (LRG) family characterized by the leucine-rich repeat motifs of which they contain various numbers (LRG repeats). Consensus sequences have also been identified in the regions flanking the LRG repeats.“ The primary structure of the gene for GPIba has been determined12.13and cDNAs for GPIbP and GPIX have been cloned and s e q ~ e n c e d ? , ~The ~ ~ ”striking feature of the GPIba gene is its intron-depleted structure, with one small intron immediately upstream of the initiation codon and the coding sequence contained in its entirety within exon 2. Although partial amino-acid sequence has been determined for GPV,’6317 the gene has yet to be cloned. However, there is evidence suggesting that GPV, unlike GPIba, GPIbP, and GPIX, may not be essential for membrane expression of the GP1b:IX complex.” In BSS, there are often demonstrable levels of residual subunit proteins, but the relative quantities of each can vary from pedigree to ~ e d i g r e e . ~ -Although ~-” generally each protein is significantly reduced in quantity or Blood, Vol81, No 9 (May 1). 1993: pp 2339-2347

MATERIALS AND METHODS

Patients. The propositus (11-5) is a 59-year-old female with a lifelong bleeding disorder who presents frequently with episodes of mu-

cosal bleeding, particularly melena. Two of her siblings are similarly affected (11-3 and 11-6, Fig 1). RIPA was consistently absent in all three affected individuals and thrombocytopenia with morphologically abnormal giant platelets was also noted. Immunochemical staining of platelet membrane antigens demonstrated a near absence of GPIb,

From the Department of Haematology, Royal PostgraduateMedical School, Hammersmith Hospital, London; the Haemostasis Research Group, Clinical Research Centre, Harrow, Middlesex; and West Cumberland Hospital, Whitehaven, Cumbria, UK. Submitted June 9, 1992; accepted December 15, 1992. Supported by the Medical Research Council. Address reprint requesis to Edward G.D. Tuddenham. MD, Haemostasis Research Group, Clinical Research Centre, Walford Road, Harrow, Middlesex. HA1 3UJ United Kingdom. The publication costs ofthis article were dejayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. section 1734 solely to indicate this fact. 0 1993 by The American Society of Hematology. 0006-4971/93/8109-0006$3.00/0 2339

From bloodjournal.hematologylibrary.org by guest on July 14, 2011. For personal use only. 2340

WRIGHT ET AL West

West

Table 1. Primers Used to Amplify Segments of GPlba, B, and IX Genes Oligonucleotide

Sequence 5 ' t o 3

Coordinates'

GPlbal GPI ba2 GP1ba3 GP1ba4 GP9a GP9b/9bbio GPBl GPB2/B2bio GP9c GP9d GP9e GP9f

TGCAGTGTGACAATTCAGAC GAGACCCAAGACATAGAAGC CACATGCATCTATGCACAGA AGGCTGGAGTTCAGTGGTAC CGGTGTCCAGAGACAGTTAG CAGTTTATTCTGGGCTTCTG CTGAGCTTACTGCTCCTGCT GGGTTTATTCAGCACCAGAG CAGCAGAGGCCACCAAG GCACACTCCACCTTCC GCAGGCAGGGCCGTG GCTGCAGGTCCGCTGTG

3853-3872 4604-4585 5238-5257 5669-5650 109-128 8 6 1 -842 47-66 939-920 260-276 136-151 365-351 546-562

I

I

I

7

8

9

10

11

12

13

Fig 1. Pedigree of the family studied. The propositus is subject 11-5. Solid symbols indicate individuals with BSS. Half-solid symbols indicate obligatory heterozygotes (assuming recessive inheritance). A diagonal stroke on male (square) or female (round) symbols indicates that the subject is deceased.

* Numbering according t o EMBL data base entry for gene

consistent with a diagnosis of BSS. Blood samples were taken from the affected patients, one unaffected sister (11-I), and the youngest daughter of the propositus (111-6). Informed consent was obtained from each participant. DNA anal.vsis. Genomic DNA was extracted from whole blood collected into EDTA as described elsewhere.23Oligonucleotides for polymerase chain reaction (PCR) and for sequencing were designed according to the published sequences of GPlbu, GPlbB, and GPIX,9.'2,'5and synthesised on an ABI-PCR mate instrument (Applied Biosystems Inc, Warrington, UK). Primer pairs and their coordinates as entered on the European Molecular Biology Laboratory (EMBL)(Heidelberg,Germany) database are shown in Table 1. GP9b and GPB2 were also synthesized with a biotinylated 5' end to yield GP9bbio and GPB2bio, respectively. PCR reactions were performed on a DNA thermal cycler (Perkin Elmer Cetus, Norwalk, CT). For each complementary pair of oligonucleotides the conditions were optimized to ensure maximum yield and specificity. One hundred-microliter volumes were used, containing 200 to 600 ng of DNA, 2 U of Taq polymerase (Promega, Madison, WI), and 300 ng of each primer. When a biotinylated primer was used, only 100 ng of each primer was added. Control tubes to which no DNA was added were included with each batch of reactions. PCR buffer was supplied by Promega and was made up with each dNTP to a concentration of 200 pmol/L. Cycling conditions for the primer pairs were 5 minutes at 94°C. then 32 cycles with 35 seconds at 94"C, 1 minute at the annealing temperature (see below), and I minute at 72°C (plus 3 additional seconds per cycle). All reactions were completed with I O minutes at 72°C. Annealing temperatures were asfollows: GPlbal/2 and GP9a/b, 53°C; GPIba3/4,56"C; and GPBI/2, 48°C. Consistent amplification with GPBlj2 required dimethyl sulfoxide (DMSO) in the reaction mix at a concentration of 10%.Reaction products were checked by agarose gel electrophoresis and subsequent restriction digests were performed in 20 pL reactions for 1 hour in the appropriate buffers and at the appropriate temperatures. Singlestranded conjbrmation polymorphism (SSCP) analysis. We have developed a modification of previously described metho d ~that ~ we ~ have , ~ termed ~ consecutive SSCP. The principle is illustrated in Fig 2. PCR reactions were performed as above, except that '*P-labeled dATP or dCTP was added to the reaction mixes (approximately 2 pCi in each reaction tube). For consecutive SSCP, a biotinylated antisense primer was used and the product was concentrated by binding to streptavidin-coatedmagnetic beads (seebelow; Dynabeads; Dynal AS, Oslo, Norway). This allowed the residual reaction products including unincorporated radiolabeled nucleotides to be removed using a magnetic separator. The beads were then

washed once with sterile distilled water and resuspended in 17 pL of the same. To this was added 1 p L of restriction enzyme and 2 pL of the appropriate IOX buffer. After 1 hour of incubation, the cleaved fragments were separated (again using the magnetic separator) from the remaining fragments that were still attached to the beads by their biotinylated end (the antisense primer). The beads were again washed and resuspended in sterile distilled water in preparation for the next restriction digest. In this way it was possible to run fragments of optimal length for SSCP analysis in separate lanes of a polyacrylamide gel, simplifying interpretation and allowing identifiable bandshifts to be directly ascribed to the correct fragment. The final fragment can be analyzed either by cleaving the PCR product close to the biotinylated end or alternatively by disrupting the biotin-streptavidin bond directly, which requires heating to 95°C or more for I O minutes. Each cleavage product was analyzed separately as follows. One to three microliters ofthe product was made up to 10 pL with SSCP buffer (80%formamide, 0.1% Bromphenol Blue, 0.1% XyleneCyanol, I mmol/L EDTA, 10 mmol/L NaOH); samples were denatured at 90°C for 3 minutes, transferred to ice, and then 4 pL was loaded onto a 4.5% polyacrylamide gel (39: 1 acrylamide to bisacrylamide) that was run at 4°C and at 40 W constant power; gels were then dried under vacuum at 80°C and exposed to X-ray film at room temperature ( I to 3 days). Seqtimnng of amplifrcd DNA Amplified DNA was sequenced as described.26Briefly, the PCR product was bound to magnetic beads

Enzvme 1

2

3

W

Biotinylated 5' end of antisense strand Fig 2. The principle of consecutive SSCP. PCR for SSCP is performed using a biotinylated primer. Product is bound to magnetic beads, allowing it to be washed and purified between successive restriction enzyme digests. Through judicious choice of restriction enzymes, used in a consecutive fashion (enzymes 1 , 2 , and 3).the original PCR product can be subdivided into fragments (1, 2, and 3) suitable for SSCP and run separately to facilitate interpretation (see text).


GLYCOPROTEIN IX AND BERNARD-SOULIER SYNDROME

MW

N1

N2

II-3

II-5

I1 -I+

I-

bp 432

-252 - 180 I

1 1

iI Fig 3. Two percent agarose gel electrophoresis of Taq I-digested, PCR product of 452 bp including the Tsq Ipolymorphism 3 to the GPlba coding sequence. When the enzyme recognition sequence is absent ( - / - ), only a single 432-bp band is seen, whereas subjects who are homozygous for the restriction site ( / ) show two bands of 252 bp and 180 bp. Heterozygotes ( / - ) show all 3 bands, as illustrated by N1. Two affected siblings (11-3 and 11-5)are homozygous for different alleles at this locus.

-1-

+I+

+/+-

b

I11

+I+

Fig 4. Allele distribution for the Taq Ipolymorphism 3 to GPlba gene in selected members of the family under investigation. The presence ( ) or absence ( - ) of Tsq I restriction site at nucleotide 5491 of GPlba is indicated. This distributionexcludes an abnormality in the GPlba gene as the cause of abnormal platelet function in this family.

+

blood cells. Each sample was centrifuged at 1.600 rpm for 5 minutes and the supernatant was removed. Pellets were resuspended in icecold PGB (PRScontaining20 mmol/L glucose and 0.5% bovine serum albumin) and the wash was repeated. Second antibodies were added at this stage and the process was the same as with the first antibody. except that no further erythrolyse was added. For FMC-25. a rabbit antimouse IgG FlTC conjugate was added as the second antibody and. before the addition of further antibodies. mouse serum was added

++ +

according to the manufacturer’s instructions (Dynabeads). The strands were separated by adding 0. I mol/L NaOH and removing the supernatant while using the magnetic separator. The beads. to which were attached the biotinylated strands. were then washed and resuspended in IO pL sterile distilled water. Sequencing reactions were performed using T7 DNA polymerase according to the protocol sup plied by the manufacturers (Pharmacia. Milton Keynes. UK). Either the original primers used for amplification or specific internal sequencing oligonucleotides were annealed to the template DNA (Table I ) and the reactions were mixed by gentle pipette action only. The reaction products were heated to 72°C for 3 minutes before loading on a denaturing polyacrylamide gel. The supernatant was separated from the beads by using the magnetic separator once more immediately before loading the samples. Fhw. cj*/onw/r,r. Venous blood was taken into heparin and left mixing gently at room temperature. Aliquots of whole blood were diluted with phosphate-buffered saline (PRS).pH 7.2, according to the platelet count such that each aliquot of 100 pL contained IO6 platelets. The following antibodies were used to demonstrate platelet membrane GP expression: anti-GPllb/llla fluorescein isothiocyanate (F1TC)conjugate. clone no. M 148 (Serotec. Oxford. UK): anti-GPlb R. Phycoerythrin conjugate. clone no. AK2 (Serotec): anti-GPIX. clone no. FMC-25 (Serotec). Samples were incubated with saturating concentrations of antibody for I hour at room temperature. Two milliliters of erythrolyse (Serotec) was then added to lyse the red

MW N1 N2 N3 II-3 II-5 II-6 MW

bP 460 202 90

Fig 5. Agarose gel electrophoresisofAvs Il-digestedPCR product obtained by amplication using the primer pair GP1bal and GP1ba2. The region amplified corresponds to the macroglycopeptidedomain of GPlba and predicts a 752-bp product (460 202 90 bp after Ava II digestion). Fragment length polymorphism is seen in the larger fragment for N1 and N3 (controls), but not for the BSS-affected patients (11-3.11-5. and 11-6).The polymorphic natureofthe repetitive motif demonstrated here was first reported by Ware et al.”

+

+

From bloodjournal.hematologylibrary.org by guest on July 14, 2011. For personal use only. WRIGHT ET AL

2342

Table 2. lntmic Sequence Identified 5’ to the GPlX Coding Region GCTGAGACCCGAGAAGGGAGgtgagtgcaccccgtcccatgtcaggctccgcta-

catccccagtgcttgccgtccctgaggatcggtccaggctgccaggccctccctcacagcccctctctctgcagCCAGCCTGTCCCEG

lntronic sequence is shown in lower case letters. The ATG initiation codon is underscored.

to block any excess antibody. Control samples that were incubated with secondary antibody only were included along with normal control samples. Cell suspensions in a final volume of 250 pL were analyzed in a FACStar flow cytometer ( W o n Dickinson, Mountain View, CA). Levels ofexpression of the target proteins were determined by measuring red (GPlb) and green (GPIX and GPllb/llla) emission. Immitnoblotring. Platelet-rich plasma was prepared by allowing blood taken into anticoagulant to settle at room temperature and pipetting off the plasma fraction. After two washes with PBS, the platelets were pelleted and resuspended in 50 mmol/L Tris buffer containing 2% sodium dodecyl sulfate (SDS). Ten micrograms of protein from each platelet extract was run in nonreduced form on a polyacrylamide gel. Gels were electrophoretically transferred to nitrocellulose and left to soak for 30 minutes in buffer containing 50 mmol/L Tris, I50 mmol/L NaCI, 5 mmol/L EDTA, 0.05% Triton X-100, and 0.25% gelatin. The nitrocellulose membranes were incubated with anti-GPlb or anti-GPIX antibodies for I to 2 hours

A

N1

N2

N3

11-3

II-5

B

II-6

and then washed three times ( I O minutes each) in fresh buffer. A goat antirabbit IgG antibody. conjugated with horseradish peroxidase (BioRad, Hemel Hempstead. UK) was added as the secondary antibody. After incubating for I further hour. the three washes were repeated. The nitrocellulose membranes were coated with Amenham ECL detection reagents (Amenham. Amenham. UK) and exposed to x-ray film for I to 30 minutes. RESULTS

Gflba. Two regions of the GPlba gene, both of which have been shown to be polymorphic, were selected for amplification: firstly, the 3’ end of the gene spanning the Tuq I restriction site polymorphism reported by Finch et a1.22( I ) A 432-bp fragment was amplified using GPI ba3 and GPI ba4, incorporating the Tuq I polymorphism at nucleotide 5491 (EMBLdatabase). The presence or absence of the Tuq I restriction site was determined by agarose gel electrophoresis ofthe PCR products after digestion with Tuq I, with cleavage at this site resulting in fragments of 252 bp and I80 bp. Within the family studied. this polymorphism was clearly not linked to the BSS phenotype because 11-3 and 11-5 were shown to be homozygous for different alleles (Fig 3). Results obtained from these and other family members are summarized in Fig 4. Secondly. we looked at the region coding for the heavily glycosylated portion of the extracellular domain known as the macroglycopeptide domain. This region has been shown

II-5

N1

Fig 6. (A) SSCP analysis of DNA fragments containing the coding region for the GPlX gene amplified from genomic DNA as described. The resulting products (861 bp) were digested with Taq I to give fragments of 461 bp and 400 bp, a total of four single strands. A solitary bandshift was observed for t h e two BSS samples (11-3 and 11-5) when compared with three normals (N1 through 3). The experiment was repeated using t h e consecutive SSCP method outlined above (E) to identify with certainty the fragment responsible for the bandshift. (6) The 5 fragment was run separately, demonstrating that this fragment was responsible for the bandshift observed in (A). The two conformations seen, probably representing different strands of 461 bp (sense) and 463 bp (antisense). corresponded to the upper two (slower migrating) bands of (A). Two separate bandshifts were discernable here (seetext).

From bloodjournal.hematologylibrary.org by guest on July 14, 2011. For personal use only. GLYCOPROTEIN IX AND BERNARD-SOULIER SYNDROME

2343

Homology between NH, GPlba GPlb$ GPK LRG

M

- termlnrl flrnklng sequences

V N C D K R~ N LT [ : T ~ J P D V D C G R RG L TWA S L P T A F P V D T T E V D C R G H G L T - A L P - A L P A R T R H A E T P G Y L P A D T V H

B

Fig 7. (A and B) Different regions of the same sequencing gal showing GPIX coding sequence for the propositus (11-5) and a noma1 subject (Nl). (A] G transition was identified in A heterozygous A codon 21 from the propositus. This predicts alteration of a conserved residue in the amino-terminal flanking sequence ofthe single LRG domain of GPIX, shown below aligned with homologous sequences of LRG NH,-flanking sequences (after Roth”). Mutated residue is shown in bold type. (B) A second heterozygoussingle base-pair substitution, also an A + G transition, was identified in codon 45. predicting the alteration of a conserved residue of the core LRG motif, shown below aligned with homologous sequences (after Roth”). Mutated residue is shown in bold type.

-

-.

C C G A C G/A A C A A C C G

Homology between LRG sequences

to contain a repeat motif corresponding to 13 amino-acid residues (codons 369-38 I of published sequence), present in variable number ( I to 3).” (2) A 752-bp fragment was amplified using GPlbal and GPIba2. It was possible to demonstrate fragment length polymorphism by agarose gel electrophoresis of the amplified products. Digestion of the PCR products with Avu 11 delineated the alleles more clearly, although the polymorphism was uninformative with respect to the family under investigation (Fig 5). GPIX. It was considered a likely possibility that the genes for GPIX and GPIbfi were structurally related to the GPIba gene. The GPIX gene is known to be small and located on chromosome 3.15 Primers were selected (GWa and GP9b) that spanned the coding region in its entirety. The region immediately upstream of the initiation codon was also included as this region is known to contain the sole intron in the GPlba gene. A fragment was amplified that was approximately 110 bp longer than the corresponding cDNA se-

45 a m + ser

quence (753 bp). To confirm the origin of the fragment, the 5’ end was sequenced and a small intron of 108 bp was identified 12 bp upstream of the start codon. This sequence is presented in Table 2. SSCP analysis (Fig 6) after a Tuq I digest of the GPIX gene identified a single bandshift, which appeared to be linked to the BSS phenotype. When the two fragments generated by Tuq I digestion were separated by consecutive SSCP analysis, an additional (apparently heterozygous) bandshift was observed that was also observed for the daughter of the propositus. The entire gene was subsequently sequenced for each family member who had provided a sample and for four normal subjects. Two separate single base-pair substitutions were identified in all three siblings with BSS such that they were heterozygous for both mutations (Fig 7A and B). These substitutions predict the following alterations at the amino-acid level: codon 2 l Asp + Gly (GACGGC) and codon 45 Asn + Ser (AAC-AGC). The two unaffected family members who were tested (11-1 and 111-6)

~ ~


MW

WRIGHT ET AL

N1

N2 11-3 II-5 II-6

m-6

or absence of an intron upstream of the initiation codon. Consistent amplification required the presence of 10% DMSO in the reaction mix, possibly a consequence of the high G C content of the gene (75% of the first 900 nucleotides). The DMSO was subsequently found to interfere with SSCP analysis and therefore consecutive SSCP analysis was performed on biotinylated product (see above). In this way it was possible to remove the DMSO and concentrate the radiolabeled product. Although polymorphisms were identified within the GPlbS coding region. none showed linkage to the BSS phenotype (Fig 9). Flow c.vIomefty. Platelets from the propositus were analyzed for membrane expression ofGPlb, GPIX, and GPllb/

+

bP 11751169

- 151 -(118)

N2 11-3 N3 N1 11-6 II-5 N4 A Fig 8. Fnu4HI digest of the amplified GPIX gene in normals (N1 and N2). affected siblings (11-3, 11-5, and 11-6). and a heterozygote carrying the codon 21 mutation only (111-6) analyzed in agarose gel electrophoresis stained with ethidium bromide. New bands are seen where the A G transitions create additional recognition sites for Fnu4HI: a 118-bpfragment corresponding to the codon 21 mutation and a 105-bp fragment corresponding to the codon 45 mutation. Sixty normal subjects were analyzed in this way, and neither substitution was detected outside the immediate family.

-.

were found to be heterozygous only for the codon 21 substitution. Despite full sequence analysis. it is not clear why the most slowly migrating band on SSCP analysis resolves in the patients as a single-shifted band as might be expected for a homozygous sequence irregularity. Neither substitution was found in the normal subjects, but a heterozygous neutral substitution was identified in one subject at codon 28 (ACG-ACA. Thr). This latter substitution was presumed to have given rise to a shift observed on SSCP analysis of the same individual (not shown). Both of the mutations found in the family with BSS were predicted to create restriction sites for the enzyme Fnu4H I. New bands could be identified on agarose gel electrophoresis of Fnu4H I-digested PCR product corresponding to each of the substitutions (Fig 8). Therefore, it was possible to rapidly screen 60 subjects with normal platelet function for both the codon 21 and the codon 45 substitutions simultaneously. Neither substitution was identified in any individual unrelated to the family under investigation. GPlhS. The GPlbB gene was also amenable to amplification from genomic DNA. suggesting that it had an introndepleted primary structure similar to that of the gene for GPlba. Using primers GPBl and GPBZ. it was possible to amplify a fragment that at 893 bp was the size predicted from the published cDNA sequence.’ Because only limited sequence was available at the 5’ end of the cDNA, the sense primer (GPBI ) corresponded to sequence coding for the signal peptide and thus it was not possible to establish the presence

B

Fig 9. Consecutive SSCP analysis of the coding region of the GPlW gene, amplified using primers GPBl and GPBPbio. Reactions were performed in the presence of 3aP-labeled dCTP and 10% DMSO. The products were subsequentlyconcentrated and purified by binding with streptavidin-coatedmagnetic beads and a magnetic separator (supplied by Dynal AS). Consecutive cleavages were made from the immobilized products using Not Iand Taq 1, removing the cleaved fragments on each occasion before subjectingthem to SSCP analysis. (A) The Not I-cleaved fragments representing strands of 355 and 351 bases in length. (B) The Taq I-cleavedfragments representing strands of 306 and 304 bases in length. Bandshifts can be seen between samples in both (A) and (8).but there is no linkage to the BSS phenotype, and these shifts are therefore due to polymorphism unrelated to disease. No shifts at all were observed in the remaining fragment, representing the 3 noncoding region (not shown).


GLYCOPROTEIN IX AND BERNARD-SOULIER SYNDROME

Fig 10. FACS analysisof normal and BSS platelets labeled with fluorescent antibody conjugates. (A) and (B) show fluorescence intensity of cells incubated with anti-GPlb-phycoerythrin (FU) and anti-GPllb/llla-FITC (FL1). (A) Normal control platelets (SW): a uniform populationof cells positive for both spectra of emission. (B) BSS platelets (subject 115, EB): the population is less homogeneous, and largely positive for only the FlTC spectrum representing anti-GPllb/llla. (C) and (D) show merged population distribution curves for fluorescence intensities (x axis) after incubation with anti-GPlb or anti-GPIX. The curves relating to the BSS platelets are labeled EB and those for the normal control are labeled SW. (C) Anti-GPlb: while clearly reduced in the propositus (EB), a secondary peak was observed representing approximately 25% of cells in which there was a low level of expression. (D) Anti-GPIX: uniformly reduced emission is seen for the BSS platelets when compared with m a l platelets. The BSS platelet fluorescence distribution was identical to that observed for platelets not exposed to the anti-GPIX primary antibody (data not shown).

1 10’ FLl +

1

205kDa

+

116kDa

+

2

lo3

lo4

1 10‘ FLl +

lo2

lo3

lo4

D

C

I 101 102 103 Fluorescence intensity +

io0

llla (Fig IO). There was no significant difference in levels of expression ofGPllb/llla between the propositus and normal controls. GPlX expression was undetectable in the platelets of the propositus, with fluorescence intensity being the same as that for controls lacking the primary antibody (FMC-25). However. there was evidence of residual GPlb expression on the patient’s platelets. the results for which were clearly different from those for the control sample and also from the pattern obtained from another patient with BSS whose platelets were analyzed at the same time (data not shown).

A

lo2

100

10’

lo2

lo3

‘1 lo4

ftntniinoblo~titi~.In concurrence with the flow cytometric data. there was evidence of residual GPlb in the platelet extracts from the propositus (Fig I I). A band was seen that corresponded to a molecular weight of 160 Kd. This band was observed for platelet extracts of both normal and BSS platelets. although the signal was clearly less intense for the BSS platelet extract. When antibody directed against GPlX was used, a band was observed in the normal control sample corresponding to a molecular weight ofapproximately 32 Kd. This is somewhat

B

32kDa

lo4

1

-

2

Fig 11. Westem blots of platelet protein extracts from the propositus and a normal control run in nonreducing conditions. (A) Extractsprobed with monoclonal antibody AK2 that recognizes an epitope on the heterodimeric GPlb molecule. (B) Extracts probed with antiGPlX antibody FMC-25. Track 1, normal platelet extract; track 2, BSS platelet extract. GPlb was detected in both normal and BSS platelet extracts, although there was significantly less in the latter sample. No signal was obtained for GPlX in the BSS platelet extracts. The band obtained for the normal sample is at a position consistent with a molecular weight of approximately 32 Kd (see text).


WRIGHT ET AL

higher than previously reported and presumably indicates anomalous migration in the polyacrylamide gel system used. There was no evidence of any band when a BSS platelet extract was analyzed in identical conditions. DISCUSSION

Recent studies have characterized mutations affecting GPIba in patients with BSS.'9-2'Mutations that modify vWF binding may result in an alternative phenotype, ie, pseudoor platelet-type von Willebrand's d i ~ e a s e . ~ ~In' ~the ' latter cases, the mutations are dominant acting and situated within the vWF binding region of the molecule. In BSS, however, the mutations described are more heterogenous. The missense mutations that were identified in the current investigation indicate that lesions within the GPIX gene can also give rise to the BSS phenotype. Furthermore, in a Chinese hamster ovary (CHO) cell expression system, it has been shown that GPIX is a necessary requirement for membrane expression of the GP1b:IX complex, and the same seems to be true for both GPIba and GPIbP, but not apparently for GPV.18 In the family reported here, involvement of the GPIba gene was excluded by the (fortunate) distribution of alleles at this locus. The intron-depleted structure of the GPIbP gene enabled us to screen for mutations using SSCP. Only polymorphic band shifts were found that helped to exclude this locus. Analysis of the GPIX gene, similar to that of the GPIbP gene, was facilitated by its intron-depleted structure and we were able to identify a single small intron of 108 bp. This intron was situated in a region showing close homology to the corresponding region in the GPIba gene, immediately upstream of the initiation codon.I5 The two single base-pair substitutions identified within the GPIX coding region both predict amino-acid substitutions affecting conserved residues within or flanking the single LRG repeat. The codon 45 asparagine to serine mutation changes a conserved residue within the core LRG motif, whereas the codon 21 aspartic acid to glycine mutation changes a conserved aspartic acid residue within the amino-terminal flanking sequence. Screening for these changes at DNA level failed to identify either mutation in 60 normals (120 alleles). It is likely that these missense mutations cause severe disruption of GPIX structure and it supports evidence that the GPIX molecule, although small, is of critical importance in the formation ofthe intact, functional complex. Mutant GPIX may be either rapidly degraded or else be present in a nonfunctional form, not recognized by the monoclonal antibody FMC-25. The bulk of the GPIX molecule lies outside the platelet membrane. There is a very small intracellular domain (6 residues) and a transmembrane domain that can be acylated with palmitic acid through a thioester linkage.3' The extracellular domain is largely composed of the single LRG motif and its flanking sequences. The current work serves to highlight the potential importance of this region, although it is not yet clear whether the LRG domain contributes directly to the structural integrity of the complex, perhaps by interacting with the single LRG domain ofGPIbP. The only other mutation affecting a residue from an LRG domain (to date) was identified in GPIbcu." This mutation was predicted to alter a conserved leucine residue and, interestingly, appears

to cause a significant bleeding tendency and morphologically abnormal platelets in a dominant manner. The GP1b:IX complex seems also to have an important role in the structural development of the platelet, possibly through its provision of an effective attachment between the membrane and, through actin-binding protein, the membrane skeleton. The platelets from the patients presented here have the characteristic morphologic abnormalities associated with BSS. The results of the flow cytometry suggest that the residual GPIb detected was present in the membrane of a distinct subpopulation of cells, (approximately 25% of cells counted) which can be observed as a secondary peak in the merged graph in Fig 10. This subpopulation may represent the younger platelets that then lose remaining GPIb during their circulatory life span, which is generally reduced in BSS. Further studies are required to test this hypothesis. It is probable that the BSS phenotype represents the end result of diverse mutations affecting at least two loci involved in the assembly of the GPIb/IX complex on the platelet surface. Further studies on BSS kindred are likely to uncover more heterogeneity and potentially will increase our understanding of the structural requirements for proper assembly and function of the GPIb/IX receptor. ACKNOWLEDGMENT

Arthur Stackpoole assisted with the FACS analysis. REFERENCES

1. Berndt MC, Fournier DJ, Castaldi PA: Bernard-Soulier syndrome. Baillieres Clin Haematol 2585, 1989 2. Clemetson KJ, McGregor JL, James E, Deachavanne M, Luscher E F Characterizationof the platelet membrane glycoprotein abnormalities in Bernard-Soulier syndrome and comparison with normal by surface-labelling techniques and high resolution two-dimensional electrophoresis. J Clin Invest 70:304, I982 3. Berndt MC, Gregory C, Chong BH, Zola H, Castaldi PA: Additional glycoprotein defects in Bernard-Soulier's syndrome: Confirmation of genetic basis by parental analysis. Blood 62:800, 1983 4. Nurden AT, Didry D, Rosa J-P Molecular defects of platelets in the Bernard-Soulier syndrome. Blood Cells 9:333, 1983 5. Drouin J, McGregor JL, Parmentier S, Izaguirre CA, Clemetson KJ: Residual amounts of glycoprotein Ib concomitant with nearabsence of glycoprotein IX in platelets of Bernard-Soulier patients. Blood 72:1086, 1988 6. De Marco L, Mazzucato M, Fabris F, De Roia D, Coser P, Girolami A, Vincente V, Ruggeri ZM: Variant Bernard-Soulier syndrome type Bolzano: A congenital bleeding disorder due to a structural and functional abnormality of the platelet glycoprotein Ib-IX complex. J Clin Invest 86:25, 1990 7. Fox JEB, Aggerbeck LP, Berndt MC: Structure of the glycoprotein 1b:IX complex from platelet membranes. J Biol Chem 263: 482, 1988 8. Lopez JA, Chung DW, Fujikawa K, Hagen FS, Papayannopoulou T, Roth GJ: Cloning of the a chain of human platelet gly-

coprotein Ib: A transmembrane protein with homology to leucinerich a-2-glycoprotein.Proc Natl Acad Sci USA 845615, 1987 9. Lopez JA, Chung DW, Fujikawa K, Hagen FS, Davie EW, Roth GJ: The a and p chains of human platelet glycoprotein Ib are both transmembrane proteins containing a leucine-rich amino acid sequence. Proc Natl Acad Sci USA 85:2135, 1988 10. Moddermann PT, Admiraal LG, von dem Borne AEGKr: GlycoproteinsGPIb/IX and GPV form a non-covalent complex in

From bloodjournal.hematologylibrary.org by guest on July 14, 2011. For personal use only. GLYCOPROTEIN IX AND BERNARD-SOULIER SYNDROME

the platelet membrane. Thromb Haemost 1991:1071, 1991 (abstr 1352) I 1. Roth GJ: Developing relationships: Arterial platelet adhesion, glycoprotein Ib, and leucine-rich glycoproteins. Blood 775, 1991 12. Wenger RH, Kieffer N, Wicki AN, Clemetson KJ: Structure of the human blood platelet membrane glycoprotein Ib a gene. Biochem Biophys Res Commun 156:389, 1988 13. Wenger RH, Wicki AN, Kieffer N, Adolph S, Hameister H, Clemetson KJ: The 5’ flanking region and chromosomal localization of the gene encoding human platelet membrane glycoprotein Ib a. Gene 85517, 1989 14. Hickey MJ, Williams SA, Roth GJ: Human platelet glycoprotein IX: An adhesive prototype of leucine-rich glycoproteins with flank-center-flank structures. R o c Natl Acad Sci USA 6:6773, 1989 15. Hickey MJ, Deaven LL, Roth GJ: Human platelet glycoprotein IX. Characterization of cDNA and localization of the gene to chromosome 3. Fed Eur Biochem Sci 274:189, 1990 16. Shimomura T, Fujimara K, Maehama S, Takemoto M, Oda K, Fujimoto T, Oyama R, Suzuki M, Ichihara-Tanaka K, Titani K, Kuramoto A: Rapid purification and characterization of human platelet glycoprotein V: The amino acid sequence contains leucinerich repetitive modules as in glycoprotein Ib. Blood 75:2349, 1990 17. Roth GJ, Church TA, McCullen BA, Williams SA: Human platelet glycoprotein V: A surface leucine-rich glycoprotein related to adhesion. Biochem Biophys Res Commun 170:153, 1990 18. Lopez JA, Leung B, Fox JEB Expression of functional human glycoprotein IbIX complex in hamster cells: Subunit requirements for membrane expression. Circulation 82597, 1990 (abstr) 19. Ware J, Russell SR, Vincente V, Scharf RE, Tomer A, McMillan R, Ruggeri ZM: Nonsense mutation in the glycoprotein Iba coding sequence associated with Bernard-Soulier syndrome. Proc Natl Acad Sci USA 87:2026, 1990 20. Miller JL, Lyle VA, Cunningham D: Mutation of leucine-57 to phenylalanine in a platelet glycoprotein Iba leucine tandem repeat occumng in patients with an autosomal dominant variant of BernardSoulier disease. Blood 79:439, 1992

2347

21. Ware J, Russell S, Murata M, Mazzucato M, DeMarco L, Ruggeri ZM: Ala1 56-Val substitution in platelet glycoprotein Iba impairs von Willebrand factor binding and is the molecular basis of Bernard-Soulier syndrome type Bolzano. Blood 79:278a, 1992 (abstr) 22. Finch CN, Miller JL, Lyle VA, Handin RI: Evidence that an abnormality in the glycoprotein Iba gene is not the cause of abnormal platelet function in a family with classic Bernard-Soulier disease. Blood 75:2357, 1990 23. Blin N, Stafford DW: Isolation of high molecular weight DNA. Nucleic Acids Res 3:2303, 1976 24. Orita M, Iwahana H, Kanazawa H, Hayashi K, Sekiya T: Detection of polymorphisms of human DNA by gel electrophoresis as single-strand conformation polymorphisms. Proc Natl Acad Sci USA 86:2766, 1989 25. Orita M, Suzuki Y, Sekiya T, Hayashi K: Rapid and sensitive detection of point mutations and DNA polymorphisms using the polymerase chain reaction. Genomics 5974, 1989 26. Thein SL, Hinton J: A simple and rapid method of direct sequencing using Dynabeads. Br J Haematol 79: 1 13, 199 I 27. Ware J, Russell S, Ruggeri ZM: Genetic bases for the molecular polymorphisms of platelet glycoprotein Iba. Thromb Haemost 65: 770a, 1991 (abstr) 28. Miller JL, Cunningham D, Lyle VA, Finch CN: Mutation in the gene encoding the a chain of platelet glycoprotein Ib in platelettype von Willebrand disease. Proc Natl Acad Sci USA 88:476 I , 1991 29. Pincus MR, Dykes DC,Carty RP, Miller JL: Substitution of valine for glycine at residue 233 ofplatelet glycoprotein Iba in platelettype von Willebrand disease causes a discrete conformational change. Thromb Haemost 65:770a, 1991 (abstr) 30. Russell SD, Roth GJ: A mutation in the platelet glycoprotein Iba gene associated with pseudo-von Willebrand disease. Blood 78: 281a, 1991 (abstr, suppl I ) 3 I. Muszbek L, Laposata M: Glycoprotein Ib and glycoprotein IX in human platelets are acylated with palmitic acid through thioester linkages. J Biol Chem 264:9716, 1989