Molecular basis for a severe case of LAD

Eur. J. Irnrnunol. 1992. 22: 1877-1881

1877

Molecular basis for a severe case of leukocyte adhesion deficiency*

Angel L. Corbi., Alicia Vara., Angeles Ursa., Mari Cruz Garcia Rodriguez+, Gumersiudo Fontan+ and Francisco Sanchez-Madrid. Unidad de Biologia Molecular. and Seccion de Inmunologia., Hospital de La Princesa, and Unidad de Inmunologia+, Hospital La Paz, Madrid

The leukocyte integrins LFA-1, Mac-1 and p150,95 (CDlla/CD18, CDllb/CD18, CDlldCD18) mediate crucial leukocyte adhesive functions in immune and inflammatory reactions. Leukocyte adhesion deficiency (LAD) disease is caused by the defective expression of these adhesion molecules on leukocytes, and is characterized by recurrent infections and impaired pus formation due to the blockade of leukocyte migration into inflamed tissues. LAD is originated by heterogeneous mutations affecting the CDlX gene and, based on the severity of the deficiency, two phenotypes (severe and moderate) have been defined. Biochemical and genetic studies have allowed the classification of five different types of LAD. We have identified a type V LAD patient (severe phenotype, and normal size and levels of both CD18 precursor and CD18 mRNA), and determined its molecular basis. Reverse transcription-polymerase chain reaction and cloning and sequencing of CD18 cDNA derived from this patient revealed three silent mutations and a missense mutation that leads to the substitution of glycine at position 169 for an arginine. Analysis of patient-derived cDNA clones revealed the concomitant presence of aberrant splicingwithin the 5’ region of the CD18 gene. The description of an identical mutation at residue 169 in an unrelated severe LAD patient raises the possibility that severe LAD typeV is caused by a unique genetic defect.

1 Introduction

mediated functions, such as neutrophil and monocyte mobilization into inflammatory sites [2-51. The level of Cell adhesion processes are absolutely required for the leukocyte integrin expression is heterogeneous among induction and regulation of immunological and inflamma- LAD patients and correlates with the severity of the clinical tory responses [l]. Most leukocyte adhesive functions are symptoms and the loss of leukocyte adhesive functions. For dependent on three leukocyte-specific cell surface hetero- this reason, two LAD phenotypes have been defined, dimeric glycoproteins, LFA-1, Mac-1, and p150,95 (or severe and moderate, depending on whether their leukoCDllaKD18, CD11WCD18 and CDllcKD18) [l]. Struc- cyte integrin expression is undetectable or ranges between turally, LFA-1, Mac-1 and p150,95 belong to the integrin 3 % -30 % of normal levels [6]. Biosynthetic studies [7-101 family of adhesion proteins, and are composed of a and reconstitution experiments using human common (3 subunit (CD18 or integrin (32) non-covalently LAD X mouse lymphocyte hybrids [ l l ] have shown that associated to homologous a subunits, LFA-la (CDlla), the cell surface expression of the three leukocyte integrins Mac-la (CDllb), and p150,95a (CDllc) [1].The relevance requires the association of their corresponding a subunit of the functions mediated by the leukocyte integrins is precursors to the precursor of the common (3 chain. The emphasized by the identification of a rare autosomal availability of specific probes for CD18 [12, 131 has led to recessive disease termed leukocyte adhesion deficiency the demonstration that LAD is caused by heterogeneous mutations which alter the expression and/or structure of the (LAD) ~ 231. , common f~subunit [S-10, 14-16]. In fact, five distinct types LAD is characterized by a defective cell surface expression of LAD have been defined based on: (a) the size and level of the 02 leukocyte integrins and is normally diagnosed by of expression of the CD18 mRNA and CD18 precursor the occurrence of recurrent bacterial infections of soft protein, and (b) the final phenotype, severe or moderate, of tissues, delayed umbilical cord separation, impaired pus each patient [lo]. LAD moderate phenotype has been formation and granulocytosis [2, 31. All these clinical shown to be caused by point mutations [15-181 or exon symptoms are secondary to defective leukocyte integrin- deletion by aberrant splicing [14]. In addition, an in-frame insertion, which has been detected in a moderate LAD patient, appears to lead to a complete absence of CD18 membrane expression [17]. For severe LAD, a point [I 104641 mutation [16] and a single nucleotide deletion [18] have :k This work was supported by Programa Nacional de Salud (grant been demonstrated as the molecular basis for the two cases SAL 89-0883) and Universidad Aut6noma de Madrid (to analyzed to date. ~

~~

~~~~~~

A.L.C.) and a Fundacion Ramon Arecesgrant and FISS 9170259 (to F.S.M.).

Correspondence: Francisco Sinchez-Madrid, Secci6n de Inmunologia, Hospital de la Princesa, Diego de Le6n 62, E-28006 Madrid, Spain Abbreviations: LAD: Lcukocyte adhesion deficiency PCR: Polymerase chain reaction RT-PCR: Reverse transcription-polyrnerase chain reaction

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1992

In the present report, we describe the determination of the CDl8 mRNA sequence in a severe-type V LAD patient (severe phenotype and normal size and levels of CD18 precursor and CD18 mRNA). Analysis of reverse transcription-polymerase chain reaction (RT-PCR) products and patient-derived CD18 cDNA clones revealed the simultaneous occurrence of a point mutation affecting residue 169

1878

Eur. J. Immunol. 1992. 22: 1877-1881

A. L. Corbi, A.Vara, A. Ursa et al.

and aberrant splicing events within the CD18 gene. The previous identification of an identical mutation in a unrelated severe LAD patient suggests that typeV LAD might be caused by a single molecular defect.

2 Materials and methods 2.1 Patient The patient was a 2-year-old Caucasian female born from second-degree related healthy parents. She had a delayed umbilical cord separation, and a history of recurrent infections including omphalitis, gingivitis, bilateral otitis and aseptic meningitis. Her blood leukocyte counts were extremely high (up to 60 OOO/pl), with percentages of neutrophils ranging between 75 YO and 85 YO. 2.2 Cell culture and flow cytometry analysis

Cultured cell lines were grown in RPMI medium (Flow Laboratories, Irvine, Scotland) supplemented with 5 % FCS (Flow Laboratories), 2 mML-glutamine, and 50 pg/ml gentamicine. Peripheral blood lymphocytes (PBL) were obtained from heparinized venous blood by Ficoll-Hypaque centrifugation, and patient PBL were used to establish an EBV-transformed cell line. For PHA blasts, PBL were cultured in the presence of 1YO PHA for 3 days. For neutrophils, removal of erythrocytes was done by sedimentation in 1.3 % (w/w) dextran at room temperature followed by hypotonic lysis. Expression of cell surface antigens was measured on an FACScan cytofluorometer (Becton Dickinson, Mountain View, CA). Briefly, control and patient neutrophils, PBL and patient-derived EBV-transformed cells were incubated with 100 yl of the following hybridoma culture supernatants: TS1/11 and TP1/40 (anti-CDlla), Bear-1 (anti-CDllb), HC1/1 (anti-CDllc), TS1118 (antiCDlS), TP1/36 (anti-CD43), and D3/9 (anti-CD45) (19-21). After washing, bound antibodies were detected using fluorescein-conjugated F(ab')2 goat anti-mouse Ig. The supernatant from the myeloma P3X63 was included as negative control.

2.4 Northern blot Total RNA from PHA blasts and cultured cell lines was isolated using guanidinium isothiocyanate as denaturant and a cesium chloride ultracentrifugation step [22]. RNA (20 yglsampfe) was separated on formaldehyde-containing agarose gels. After ethdium bromide staining, RNA was transferred onto nitrocellulose and the membrane was prehybridized and hybridized with the 1-kb Eco RI fragment of CD18 as probe [12, 131. After washing, the nitrocellulose filter was exposed to film overnight at - 10 "C with intensifying screens.

2.5 Reverse transcription-Polymerasechain reaction Poly A + RNA from the EBV-transformed LAD cell line was purified by affinity chromatography on oligodT-cellulose [22]. cDNA first-strand synthesis was performed on 1 yg of poly A+ RNA, using random hexamers as primer, in a totalvolume of 40 pl, as described [23]. Amplification was carried out by 35 cycles of denaturation (94 "C, 1 min) and annealing/extension (72 "C, 3 min) plus a final 10-min extension step at 65"C, using 10 ~1 of the cDNA reaction mixture and two primer pairs derived from the published CD18 cDNA sequence [12]: CD18#1 (51-74)/CD18#2 (1286-1262) and CD18#3 (1204-1228)/CD18#4 (24182394). PCR products were blunted with the Klenow fragment of E. coli DNA polymerase I and ligated into the Sma 1 site of pUCBM21. Sequencing was done by the Sanger method [24], using CD18- and plasmid-specific primers.

2.6 Construction and screening of LAD patient-derived cDNA library

Poly A + RNA ( 5 pg) was reverse transcribed with AMV reverse transcriptase using oligo(dT)17 as primer, and double-stranded cDNA synthesis completed as described [23]. After blunting, methylation, linker ligation and Eco RI digestion, dscDNA was size-selected, separated from digested linkers by gel filtration on Sepharose CL-4B, and cloned into the Eco RI site of h gt1O. Packaging of the ligation reaction yielded 1.5 x lo6 primary recombinants. For screening of the library, 5 x 10s primary recombinants were plated at a density of 5 x lo4 phage/plate. Replica 2.3 Cell surface and biosynthetic radiolabeling filters were hybridized with the more 5' Eco RI fragment For cell surface labeling, patient and control neutrophils from the CD18 cDNA, as previously described [23]. After (15 x 106cells) were washed twice in PBS and added into two rounds of subcloning, positive clones were grown and Iodo-gen-coated glass vials containing 1 mCi of Na12sI. their cDNA inserts ligated into the Eco RI site of pUC18. After extensive washing in PBS, neutrophils were lysed in a Sequencing was done with plasmid- and CD18-specific PBS solution containing 1YOTritonX-100,l YOhemoglobin oligonucleotides. and 1mM PMSF. For biosynthetic labeling, patient and control PHA blasts (1 X 107-3 x 10' at 5 x 106celldm1 in RPMI 1640, 2 r n L-glutamine, ~ 5 YO dialyzed FCS) were 3 Results and discussion pulse-labeled for 30 min with 100 pCi/ml ["sSS]methionine and ["Slcysteine. Chase was done for 30 min, 4 h and 24 h, Flow cytometry analysis of immunodeficient patient blood by the addition of unlabeled methionine, and the cells were cells allowed the identification of a 2-year-old girl whose lysed as for cell surface labeling. For immunoprecipitation, leukocytes showed a complete absence of expression of the cell lysates were mixed with 100 pl hybridoma culture (32 integrins. The level of expression of LFA-1, Mac-1 and supernatant or a 1:100 dilution of polyclonal antisera. p150,95 on the patient neutrophils was undetectable, Immune complexes were isolated with protein A-Sepha- whereas other leukocyte differentiation markers showed a rose (Pharmacia, Uppsala, Sweden) and analyzed by normal profile (Fig. 1A, and not shown). Flow cytometry SDS-PAGE and autoradiography. analysis of resting and PHA-activated patient PBL also

Eur. J. Immunol. 1992. 22: 1877-1881

Molecular basis for a severe case of LAD

demonstrated the defective expression of the leukocyte integrins (data not shown). Immunoprecipitation after cell surface 1251-labelingconfirmed the absence of leukocyte integrin polypeptides on the cell surface of the patient neutrophils (Fig. lB), suggesting a defect in the expression of the common (32 subunit. The lack of PMA- and NKL16induced leukocyte aggregation (data not shown), which are LFA-1 mediated, further indicated that the patient was suffering from severe LAD.

1879

Imrnunoprecipitations after pulse-chase metabolic labeling on patient-derived PHA blasts revealed the presence of normal-size precursors for both C D l l a and CDl8 (Fig. 2A and B). However, both precursors were not processed to mature polypeptides and were absent at longer chase times (Fig. 2A and B). The presence of a normal-abundance, normal-size CD18 mRNA in the patient’s leukocytes was shown by Northern blot analysis on PHA blasts and patient-derived B lymphoblastoid cell line (Fig. 2C, and

B

A CONTROL

PATIENT

PATIENT

CELL NUMBER

200 -

15095 -

I

I

1 2 3 4 5 6 7 8

CDlle

1 2 3 4 6 1 7 8

CDllc

FLUORESCENCE INTENSITY

Figure I . Cell surface expression of LFA-1, Mac-1 and p1S0,95. A . Immunofluorescence flow cytometry was performed on control and patient’s granulocytes. Cells were labeled with anti-CD45, -CDlla, -CDllb, -CDllc and -CD18 monoclonal antibodies (solid lines), or with the ncgative control P3X63 (dotted line), and stained with an FITC-goat anti-mouse IgG. The number of cells is plotted against fluorescence intensity (log scale). B. Immunoprecipitation of cell surface antigens from control and patient granulocytes. Lysates were precipitated with control P3X63 (lane 1): anti-CD45 (lane 2), antLCD43 (lane 3), anti-CDlla (lanes 4 3 , anti-CDllb (lane 6), anti-CDllc (lane 7) and anti-CD18 (lane 8) monoclonal antibodies. Reduced samples were subjected to a 7 % SDS-PAGE.

A CONTROL Chase

-

10.5

I

4

PATIENT

1

24

6 7 8

1 2 3 4 5

Chase

-

I

0.5

1 2 3 4

I

I

4

5

6

24

7 8

B CONTROL Chase 0 . 5

I

PATIENT

4

Chase

0.5

1

4‘

- 28s

4

b

- 18s 1 2 3 4

1 2 3 4

Figure 2. Biosynthesis of LFA-I a and fi subunits. A. Immunoprecipitation of the LFA-la precursor from control and patient’s PHA blasts. Cells were pulse-labeled and chased at 0 , 0 . 5 , 4 or 24 h, lysed and immunoprecipitated with an anti-LFA-la monoclonal antibody (evcn lanes) or with the X63 negative control (odd lanes). Arrows indicate the position of the precursor and mature forms of LFA-la. B. Immunoprccipitation of the (32 subunit precursor. Labeled cells were chased for 0.5 or 4 h, lysed and immunoprecipitated with an anti-(32 rabbit antiserum (odd lanes) or with normal rabbit serum (even lanes). Arrows indicate the position of the precursor and mature forms of the LFA-1 13 subunit. C. Northern blot analysis of fi2 subunit mRNA on 20 pg of total RNA from PMA-treated HL60 cells (lane l ) , uninduced HL60 (lane 2), patient’s PHA blasts (lane 3) and control PHA blasts (lane 4).

Eur. J. Immunol. 1992. 22: 1877-1881

A . L. Corbi, A.Vara, A . Ursa et al.

1880

not shown).These results demonstrated that the CD18gene was actively transcribed and translated in the patient's leukocytes. Taken together, all the above data indicate that this patient fits the characteristics of typeV LAD patients, according to the classification of Kishimoto et al. [lo]: severe phenotype, and normal size and levels of both CD18 precursor and CD18 mRNA. To locate the genetic defect within the CD18 mRNA, we performed RT-PCR using two primer pairs which together covered the whole coding region of the CD18 mRNA. Analysis of the PCR products revealed differences between the sequence of the patient CD18 mRNA and the previously reported CD18 cDNA [12, 131. Three silent mutations were detected on residues Gly273(GGG for GGA),VaP7 (GTA for GTC), and (GTC for GTT). The fourth mutation affected residue Gly169,where the first nucleotide of the codon GGG was replaced for an A , thus leading to a change of Glycine for Arginine. Therefore, the sequence of the PCR products suggested that the Gly/Arg change at residue 169 might be the molecular basis for this case of severe-type V LAD. The three detected silent mutations probably represent polymorphisms, since two of them ( G ~ Y "and ~ Val441)have been also detected in the CD18 gene from an unrelated individual [25]. To rule out the possibility that the detected mutations were due to the high mutation rate of Taq DNA polymerase (5 x lop6 errors/nucleotide incorporatedkycle, [26]),we screened for CD18 cDNA clones from a cDNA library constructed from the patient-derived B lymphoblastoid cell line. Restriction mapping of the isolated clones was used to select eight primary recombinants whose inserts were approximately 2.5 kb and, therefore, should encode residue 169. Complete sequencing of three cDNA clones (1.4.1.1, 1.2.1.8, and 2.2.1.3) showed that all of them encoded an Arginine at residue 169, and exhibited the

185 158 N E I T E S O R I O I G S I V D K T V L P F V N T E P D R1 17L. ~ ~ T ~ = D ~ P I C ~ C S ~ V E K T V Y P Y l 101S T _. . 01

B

B

I

CD18

I TM

HIGHLY CONSERVED REGION

SP

polymorphisms identified by RT-PCR Val367and Val441, Fig. 3). Moreover, the comparison of the 5' sequences of the three analyzed cDNA clones revealed differences among them. Apart from the three silent mutations and the Gly/Arg change, the 2.2.1.3 cDNA clone was identical to the wild-type CD18 and extended from nucleotide 20 [12] to the poly-A tail (Fig. 3B). In contrast, cDNAclones 1.4.1.1 and 1.2.1.8 covered the whole coding region starting at nucleotides 131 and 220, respectively, but their 5' sequences differed from the wild-type CD18 [12] (Fig. 3B).The CD18 gene exons 3 and 4 start at nucleotides 131 and 220 [25]and, therefore, the sequences from clones 1.4.1.1 and 1.2.1.8 diverge from the wild-type CD18 sequence precisely at the junctions between intron 2/exon 3 and intron 3/exon 4 [25], respectively. Based on the CD18 gene structure [25], clone 1.4.1.1 includes exons 3-16 preceded by 300 bp that do not correspond to the 3' sequence of the intron 2. Conversely, clone 1.2.1.8 contains CD18 exons 4-14, with intron 3 immediately preceding exon 4. Although we cannot rule out the possibility that these cDNA clones represent incompletely spliced RNA, the absence of the 3' sequences from intron2 in clone 1.4.1.1. suggests that it is derived from previously, but aberrantly, spliced RNA. All the above data demonstrate a Gly/Arg substitution at residue 169 in the patient CD18 gene and suggest the existence of aberrant splicing within its 5' region. The parents consanguinity and the detection of identical mutations on all the analyzed cDNA clones and RT-PCR products strongly suggests the homozygosity of the patient for the altered CD18 gene. So far, only one severe LAD type V patient has been genetically characterized [ 161. Strikingly, the molecular basis for the deficiency was shown to be a Gly/Arg change at residue 169 by functional analysis of the transfected mutant CD18 cDNA [16]. These data further demonstrate that the point mutation at residue 169 is the molecular basis for the type V LAD patient here reported.The familiar relationship between both patients is highly unlikely since (a) they have different geographical origins, and (b) RNAse protection assays on the former [16]did not reveal any of the polymorphisms that we Tpatient P I have detected (Fig. 3B). The fact that two independent severe-type V LAD patients exhibit a common point mutation raises the possibility that some specific types of LAD may be due to a single genetic defect. Therefore, this mutation has occurred more than once in two different ethnic groups, a phenomenon already described for sickle cell anemia (27).

e" :A 72213

~~

I

0

0

0

0

0

0

1-

1 I

0

I

I

lntrmic sequdnces

0

Gly-Arg mutanon

1411 1218

0 Silent mutations

FLgure 3 Identification and location of mutations in the patient's CD18 mRNA. A. Sequence of the seven identified integrin p subunits and comparison with the mutant CD18 sequence. A highly conserved sequence is boxed, and the mutated residue is indicated by an arrow. B. Schematic representation of CD 18 and location of the mutations detected in the patient-derived CD18 cDNA clone.The location of the Gly/Arg substitution (filled circle) and silent mutations (empty circles) is indicated. Aberrantly spliced cDNA clones are shown and the intronic sequences rcpresentcd by black boxes. TM = Transmembrane region; SP = Signal peptide.

The residue Gly169is invariant in all the sequenced integrin b subunits [28] (Fig. 3A) and is located within a 250-amino acid region which exhibits the highest level of identity among the different integrin fisubunits [28] and contains the ligand-binding site in p3 [2Y]. The finding that a mutation at Gly169is the basis for a severe LAD points out an additional role for the highly conserved region either in the correct folding of the CD18 polypeptide or in its appropriate association with the a subunits. Although exon deletion by aberrant splicing has also been shown to cause LAD of moderate phenotype [14], the contribution of the aberrant splicing event to the final outcome of the deficiency here characterized is difficult to evaluate due to the concomitant presence of the mutated residue 169. Interestingly, the concomitant presence of aberrant splicing and

Eur. J. Immunol. 1992. 22: 1877-1883

missense mutation has been recently described in a moderate LAD patient [17]. In summary, detailed analysis of the CD18 mRNA from a severe LAD patient has revealed a point mutation affecting a residue within the highly conserved region of integrin p subunits, as well as the possible existence of a defective splicing in the 5’ region of the gene. The detection of a common point mutation in this and an additional unrelated type V severe LAD patient raises the possibility that this particular type of LAD might be caused by a single molecular defect. The authors gratefully acknowledge Drs. Miguel A . Vega, Teresa Belldn and Joaquin Eixido for critical reading of the manuscript and helpful suggestions. Received March 13, 1992; in revised form April 6, 1992.

4 References 1 Springer, T. A., Nature 1990. 346: 425. 2 Anderson, D. C. and Springer, T. A., Annu. Rev. Med. 1987. 38: 175. 3 Dana. N. and Arnaout, M. A., Bailliere’s Clin. Immunol. Al1er.w 1988. 2: 453. 4 Arnaout, M. A., Lanier, L. L. and Faller, D. F., Cell. Physiol. 1988. 137: 305. 5 Anderson, D. C., Miller, L. J., Schmalstieg, F. C., Rothlein, R. and Springer, T. A., J. Immunol. 1986. 137: 15. 6 Anderson, D. C., Schmalstieg, F. C., Finegold, M. J., Hughes, B. J.. Rothlcin, R., Miller, L. J., Kohl, S.,Tosi, M. F., Jacobs, R. L.. Waldrop, T. C., Goldman, A. S., Shearer, W. T. and Springer, T. A., J. Infect. Dis. 1985. 152: 668. 7 Springer, T. A., Thompson, W. S., Miller, L. J., Schmalstieg, F. C. and Anderson, D. C., J. Exp. Med. 1984. 60: 1901. 8 Dimanche, M. T., LeDeist, F., Fisher, A., Arnaout, M. A., Griscelli, C. and Lisowska-Grospierre, B., Eur. J. Immunol. 1987. 17: 417.

Molecular basis for a severe case of LAD

1881

9 Dana, N., Clayton, L. K., Tenen, D. G., Pierce, M. W., Lachmann, P. J., Law, S. K. A. and Arnaout, M. A., J. Clin. Invest. 1987. 79: 1010. 10 Kishimoto, T. K., Hollander, N., Roberts, T. M., Anderson, D. C. and Springer, T. A , , Cell 1987. 50: 193. 11 Marlin, S. D., Morton, C. C., Anderson, D. C. and Springer, T. A.,.I. Exp. Med. 1986. 164: 855. 12 Kishimoto,T. K., OConnor, K., Lee, A., Roberts,T. M. and Springer,T. A., Cell 1987. 48: 681. 13 Law, S. K. A., Gagnon, J., Hildreth, J. E. K., Wells, C. E., Willis, A. C. and Wong, A. J., EMBO J. 1987. 6: 91.5. 14 Kishimoto, T. K., O’Connor, K. and Springer, T. A., J. Biol. Chem. 1989.264: 3588. 15 Arnaout, M. A., Dana, N., Gupta, S. K., Tenen, D. G. and Fathallah, D. M., J. Clin. Invest. 1990. 85: 977. 16 Wardlaw, A. J., Hibbs, M. L., Stacker, S . A. and Springer, T. A., J. Exp. Med. 1990. 172: 335. 17 Nelson, C., Rabb, H. and Arnaout, M. A., J. Biol. Chem. 1992. 267: 3351. 18 Sligh, J. E., Hurwitz, M. Y., Zhu, C., Anderson, D. C. and Beaudet, A. L., J. Biol. Chem. 1992. 267: 714. 19 Sanchez-Madrid, F., Nagy, J. A., Robbins, E., Simon, P. and Springer, T. A., J. Exp. Med. 1983. 158: 1785. 20 Lacal, P., Rulido, R., Sanchez-Madrid, F., Cabaiias, C. and Mollinedo, F., Biochem. Biophys. Res. Commun. 1988. 154: 641. 21 Campanero, M. R., Pulido, R., Alonso, J. L., Pivel, J. P., Pimentel-Muiiios, F. X., Fresno, M. and Sanchez-Madrid, F., Eur. J. lmmunol. 1991. 21: 3045. 22 Maniatis, T., Fritsch, E. F. and Sambrook, J., Molecular cloning: A laboratory manual. Cold Spring Harbor Laboratory Laboratory Press, Cold Spring Harbor 1982. 23 Corbi, A. L., Miller, L. J., O’Connor, K., Larson, R. S. and Springer, T. A., EMBO J. 1987. 6: 4023. 24 Sanger, F., Nicklen, S. and Coulson, A. R., Proc. Natl. Acad. Sci. USA 1977. 74: 5463. 25 Weitzman, J. B.,Wells, C. E.,Wrigth, A. H., Clark, P. A , , Law, S. K. A., FEBS Lett. 1991. 294: 97. 26 Fucharoen, S., Fucharoen, G., Fucharoen, P. and Fikumaki,Y., J. Biol. Chem. 1989. 264: 7780. 27 Nagel, R. L. and Ranney, H. M., Semin. Hematol. 1990. 27: 342. 28 Erle, D. J., Riiegg, C., Sheppard, D. and Pytela, R., J. Biol. Chem. 1991.266: 11009. 29 D’Souza, S. E., Ginsberg, M. H., Burke, T. A., Lam, S. C-T. and PEow, E. F., Science 1988. 242: 91.