THE JOURNAL 0 1991 by The American Society for Biochemistry and Molecular Biology,

OF BIOLOGICAL CHEMISTRY Inc.

VOl. 266, No. 7, Issue of March 5, pp. 4315-4321,1991 Printed in U. S A .

The Product of the rap2 Gene, Member of the ras Superfamily BIOCHEMICAL CHARACTERIZATION AND SITE-DIRECTED MUTAGENESIS*

(Received for publication, June 25, 1990)

Isabelle LeroseyS, Pierre Chardin$, Jean de Gunzburg, and Armand Tavitian From the Znstitut National de la Santi et de la Recherche Medicale U-248, Faculte de Medecine Lariboisiire-Saint Louis, 10 avenue de Verdun, 75010 Paris. France

The human rap2 gene encodes a 183 amino acid protein that shares 46% identity with the K-ras p21. Its cDNA was engineered and inserted into the bacte- rial expression vector ptac; this allowed the production of high levels of soluble recombinant protein in Esch- erichia coli that was purified to near homogeneity. The rap2 protein binds GTP and exhibits a low intrinsic GTPase activity (rate constant of 0.5 % lo-’ min”). It exchanges its bound GDP with a half-life of 18 min at 37 “C in the presence of 10 mM Mg2’. Under the same conditions, the dissociation of bound GTP was at least 25-fold slower showing that the rap2 protein has a much higher affinity for GTP than GDP. The contri- bution of individual domains of the protein to its bio- chemical activities was investigated by site-directed mutagenesis. Substitution of Val for Gly at position 12 results in a 2-fold decrease in the GDP dissociation rate constant and GTPase activity. Replacement of the Ser at position 17 by Asn severely impairs the GTP binding ability of the protein and points to an impor- tant role of this residue in the coordination of Mg2+. Mutation of Thr-35 to Ala results in a decreased affin- ity for GTP and a reduction (3-fold) of the GTPase activity. Finally, substitution of Thr-145 by Ile leads to an imperfect binding of guanyl nucleotides as ex- emplified by an increase in their dissociation rate con- stants and reduction of the GTPase activity of the protein. These properties of the normal and mutant rap2 proteins are compared with those of ras p21 carrying similar substitutions and are discussed in re- lation to the structural models proposed for ras p21.

Ras oncogenes and ras related genes constitute a large family which, based on sequence comparisons of the encoded proteins, can be divided into three main branches in mammals (1): the first one contains the three ras proto-oncogenes (H- ras, K-ras, and N-ras; for a general review, see Ref. 2) as well as the closely related R-ras (3), rul (4), and rap (5-7) genes. Rho (8, 9) and rub (10-13) genes, respectively, constitute the second and third branches. Members of this family have not only been found in higher eucaryotes, but also in simpler

* This work was supported by grants from Institut National de la Sant i e t de la Recherche Medicale and the Association pour la Recherche sur le Cancer. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Recipient of a fellowship from the Ecole Normale Superieure de

09-36-35; Fax: 33-1-42-05-96-90. Lyon. To whom correspondence should be addressed. Tel.: 33-1-42-

§ Present address: Inst. de Pharmacologie Moleculaire et Cellu- laire, Centre de Valbonne, UPR 0411, 06560 Valbonne, France.

organisms such as insects, mollusks, and even in lower eu- caryotes like slime molds and yeasts (2).

Ras genes encode 21-kDa proteins (p21)’ that bind GDP and GTP, exhibit a weak intrinsic GTPase activity, and reside on the inner face of the plasma membrane (2). Analysis of the three-dimensional structure of the H-ras protein as well as mutational analysis have shown that at least six regions of p21 bear essential functions. Regions 10-17, 57-63, 110-120, and 143-147 constitute the nucleotide binding site; the two first ones interact with the ,8 and y phosphates of GTP, whereas the two latter bind the guanine base (14, 15). The structure of this guanine nucleotide binding site is well con- served in all proteins of the ras superfamily. Region 32-42 is thought to constitute the region of interaction of ras p21s with their putative effector (16, 17). Finally, the C-terminal region, constituted by a C-A-A-X motif (where C is a cysteine, A an aliphatic residue, and X any amino acid), plays an important role in the membrane localization of ras proteins which is necessary for their biological activity (18, 19). It has recently been shown that this C-terminal region undergoes a complex series of postranslational modifications consisting in the acylation of the cysteine residue of the C A M motif by an isoprenoid farnesyl group (20-22), followed by cleavage of the three last residues and carboxymethylation of the now C- terminal cysteine (23). In the case of the H-ras and N-ras proteins that contain additional cysteine residues proximal to the C terminus, those amino acids are acylated by a palmitate moiety (23). These modifications result in tight binding of the proteins to the plasma membrane.

Similarly to other guanine nucleotide binding proteins such as elongation factors and signal-transducing G proteins, ras proteins should exert their actions by cycling between an inactive GDP-bound and an active GTP-bound form (24). By analogy with yeast where the product of the CDC25 gene controls the activity of ras proteins (25, 26), an exchange factor, such as that recently described from bovine and rat brain (27, 28), would induce the dissociation of GDP and subsequent binding of GTP resulting in a conformational change of the ras protein (15). This active form of the protein would be able to interact with its yet unindentified target(s) in order to promote its biological effects. Hydrolysis of GTP to GDP by the GTPase activity of ras p21s, normally stimu- lated by the interaction with a GTPase-activating protein or GAP, returns the protein to its inactive state (29). In fact, oncogenic ras proteins are found in vivo to be complexed with GTP, whereas their normal counterparts contain GDP. The transforming mutations (at positions 12, 59, and 61) abolish stimulation of ras GTPase activity by GAP, thereby leaving

The abbreviations used are: p21, 21-kDa protein; IPTG, isopro- pyl-P-D-thiogalactoside; BSA, bovine serum albumin; DTT, dithio- threitol; GAP, GTPase-activating protein; bp, base pair(s); SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis.

4315

by guest on April 8, 2018

http://ww

w.jbc.org/

Dow

nloaded from

the protein in a permanently activated GTP bound state (30, 31). GAP may be considered as a negative regulator of ras activity, since it promotes its return to the inactive GDP- bound form; but since it has been shown to interact with the effector region of ras p21 (30, 32, 33), GAP may be part of the effector complex responsible for the biological effects of these proteins (34).

Among the ras superfamily, the rap proteins (raplA, raplB, and rap2) whose cDNAs were cloned in our laboratory by homology to the Drosophila Dras 3 gene (35), exhibit striking similarities with ras p21s (5-7). In particular, the "effector" domain is conserved between these two proteins families, which had prompted speculations that rap proteins could antagonize the effects of ras proteins, possibly by competing for a common effector. In agreement, a cDNA was identified by M. Noda's group (36) on the basis that its expression could revert the phenotype of K-ras transformed murine fibroblasts; this gene called Kreu-1, exhibiting such a ras-transformation suppressor activity, encodes the same protein as raplA. The closely related rap2 gene encodes a protein of 183 amino acids, which is 60% identical to the rap 1A protein and 50% identical to the product of the H-ras gene (5). The six principal func- tional domains of ras proteins are well conserved in the rap2 protein, suggesting that they should possess similar biochem- ical properties (see Fig. 1). With a view of better understand- ing the role of the rap2 protein, we have expressed it in E. coli, purified the recombinant product, and investigated its biochemical properties. We report that similarly to ras p21, the rap2 protein binds GDP, GTP, and exhibits an intrinsic GTPase activity. In addition, we have introduced point mu- tations at crucial positions in different domains involved in guanine nucleotide binding or interaction with a putative effector in order to understand their respective functional contributions. The results are compared with the properties of ras proteins and displayed in the context of the ras p21 structure.

MATERIALS AND METHODS

In Vitro Mutagenesis-Site-directed mutagenesis was performed on single-stranded M13 templates with oligonucleotides synthesized on a Biosearch Cyclone DNA synthesizer. Universal M13 sequencing primer (New England Biolabs) and oligonucleotides for mutagenesis (see below) were phosphorylated by the T4 polynucleotide kinase. 10

1

* *** ** I* * * * * * * I * * * * * * * * . * * X *** f * A-ras MTEYKLWV GGGVGK~ALTIQLIQNHFVDEYDPTIEDSYRXQWIDGETCLLD Rap2 MREYlCVJVL k G G V G K S ALTVQFVTGTFIEKYDPTIEDEYRXEIEVDSSPSVLE

10" 20 30 40 50

4 7 5 1 N,, A35

60 70 80 90 100 H-ras I L FTATQq EYSAMRDQYMRTGEGFLCVFAINNTKSFEDIHQYREQIKRVKDSDIJ Rap2 ILIDTAGTEIQFASMRDLYIKNGQGFILVYSLVNQQSFQDIKPMRDQIIRVKRYEK

**,**** * , "tt * * ** * " ** ** * ** * * *

130 H-ras V PMVLVGNKCDL AA-RTVESRQAQDLARSYGIPY IETSA KTRQGVEDAFYTL

150 *mo * * * ** *:40"J* " . Rap2 v PVILVGNKVDL ESEREVSSSEGWEWGCPF METSA KSKTMVDELFAEI

1145

160 170 180 H-ras VREIRQHKLRKLNPPDESGPGCMSCKCVLS

Rap2 YRQWAAQPDKDDPCCSA-------CNIQ ** * * * *

FIG. 1. Sequence comparison of the human H-ras and rap2 proteins. *, identical residues. Boxed sequences represent the nu- cleotide binding domain; stippled and full lines show the regions interacting with the p- and -,-phosphates (10-17, 57-63) and the guanine base (110-120, 143-147), respectively. The effector domain is underlined. Mutations performed on the rap2 p21 in this study are shown.

4316 Biochemical Characterization of the rap2 Protein

pmol of oligonucleotide and 5 pmol of sequencing primer were hy- bridized with the appropriate M13 vector (500 ng) a t T, = 2 AT + 3 CG. The second strand was synthesized with the Klenow fragment of DNA polymerase I in the presence of 0.2 mM dNTP, 0.01 mM ATP, and T4 DNA ligase to bridge the different fragments, in a final volume of 10 pl. 0.5 r l was used to transform J M 105 Escherichia coli; M13 plaques were transferred to nitrocellulose filters and hybridized to the "'P-labeled oligonucleotide. Hybridization was carried out a t T = 2 AT + 3 CG in 5 X Denhardt's, 5 X SSC, 0.1% SDS for 4 h. Filters were washed in 2 X SSC, 0.1% SDS for 15 min at room temperature and for 15 min a t I", + 2 "C approximately. Single stranded DNA was prepared from positive plaques and sequenced to confirm the presence of the mutation.

Construction of theptac-rap2 Bacterial Expression Vector-Normal and mutant rap2 proteins were produced in E. Coli using the ptac 32 vector as described for production of the u-H-ras protein (37). Plasmid ptac32-u-H-ras, a generous gift from Dr. A. Wittinghofer (Max- Planck Institute, Heidelberg, Germany), was digested with restriction endonucleases EcoRI and HindIII in order to remove u-H-ras coding sequences and the 4.4-kilobase pair fragment, corresponding to the vector carrying ampicillin resistance, was electroeluted. Site-directed

ATG initiator codon of the rap2 cDNA; indeed, the 5' GAA TTC T mutagenesis was used to introduce an EcoRI site just upstream the

ATG 3' sequence, where ATG is the initiator codon, had been shown to yield high amounts of ras p21 from a ptac32-ras vector in E. coli (37). A 980-bp EcoRI fragment from the human rap2 cDNA, including the complete coding sequence of 549 bp, 33 bp of 5' noncoding region and 398 bp of 3' noncoding region, was subcloned in M13 mpll and served for site-directed mutagenesis. First, an NdeI site was intro- duced at the ATG initiator codon with oligonucleotide 5' C GGA GGG CAT ATG CGC GAG TAC 3'. The corresponding double- stranded form of the M13 vector was prepared, cut with NdeI, and the 5' protruding ends were filled by the Klenow fragment of DNA polymerase I. The rap2 cDNA was released by digestion with HindIII and cloned in a 5' blunt end/HindIII M13 m p l l vector, thereby eliminating the 5' noncoding region. Single-stranded DNA from this clone was then mutated with a second oligonucleotide, 5' C TGT GAA TTC T ATG CGC GAG 3', in order to introduce an EcoRI site just upstream of the ATG initiator codon necessary for cloning into the ptac32 expression vector. We also used oligonucleotide-driven site-directed mutagenesis on this M13 single-stranded template to introduce amino acid substitutions into the rap2 protein; glycine 12 was changed to valine, serine 17 to asparagine, threonine 35 to alanine, and threonine 145 to isoleucine with oligonucleotides 5' C G T G G G C T C G G T C G G G G T A G G C A 3 ' , 5 ' C G T A G G C A A A AAC GCC CTG ACC 3', 5' TAC GAC CCC GCC ATC GAG GAC 3', and 5' C TTT ATG GAA ATT TCC GCT AAG 3', respectively.

The insert was excised from double-stranded M13 with EcoRI and PstI and subcloned into pUC 8 in order to eliminate the 3' EcoRI site. The EcoRI/HindIII fragment containing the complete rap2 coding sequence including the new EcoRI site upstream of the ATG initiator codon was released by digestion with EcoRI and HindIII and introduced into the ptac 32 vector prepared as described above.

Expression of rap2 Proteins in E. coli-The ptac-rap2 vectors were used to transform E. coli strain HB 101 containing plasmid pDMI that overexpresses the lac repressor and provides kanamycin resist- ance (38). Colonies were grown overnight in LB medium containing 80 mg/liter ampicillin and 25 mg/liter kanamycin. Bacteria were then diluted 1/20, incubated for 1 h a t 37 "C, and induced overnight with 200 PM IPTG. After centrifugation, bacteria were resuspended in SDS sample buffer and lysed by boiling for 5 min. Extracts from induced and noninduced cultures were analyzed by electrophoresis in polyacrylamide gels in the presence of SDS (39). Clones expressing rap2 proteins were identified by Coomassie Blue staining as well as western blotting using monoclonal antibody M90 directed against the GTP binding site of ras proteins (40).

Purification of rap2 Proteins-10 ml of an overnight culture of the appropriate recombinant bacterial clone were diluted in 200 ml of LB medium containing 80 mg/liter ampicillin and 25 mg/liter kanamy- cine and grown at 37 "C to an ODeoo., of 0.8. A 4-liter fermenter was inoculated with these 200 ml, incubated for 3 h, and production of the recombinant protein was induced with 200 p~ IPTG. Bacteria were harvested 16 h later by centrifugation and washed once with phosphate-buffered saline (10 mM sodium phosphate buffer, pH 7.5, 150 mM NaCl), yielding approximately 100 g of cell paste which was frozen in liquid nitrogen and stored a t -70 "C. Bacterial lysis and purification of recombinant rap2 proteins were performed as de- scribed (37) for ras p21, by ion exchange chromatography on QAE-

by guest on April 8, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Biochemical Characterization of the rap2 Prote in

Sepharose fast-flow (Pharmacia LKB Biotechnology Inc.) and gel filtration on Ultrogel AcA54 (IBF Industries). In the case of the Thr + Ala 35 mutant, remaining contaminants were eliminated by addi- tional purification on a Mono Q column (Pharmacia LKB Biotech- nology Inc.). The final purity of the proteins was greater than 90% as judged by SDS-PAGE. Protein concentration was measured ac- cording to Bradford using BSA as a standard (41). As determined by [8-:'H]GDP binding assays, our purified preparations contained 50- 70% active protein.

Kinetics of CDP Dissociation-1 p~ rap2 p21 was incubated for 15 min a t 37 "C with 5 p~ [8-'H]GDP (11.6 Ci/mmol, Amersham Corp.), 2.5 mM EDTA, 50 mM Tris buffer, pH 7.5, and 1 mM DTT, in a total volume of 200 pl. When indicated, M $ + concentration was brought to 10 mM with MgC1,. A 100-fold excess of unlabeled GDP was added and 10-111 portions were removed at the indicated times. They were immediately diluted in 2 ml of washing buffer at 4 "C (50 mM Tris, pH 7.5,l mM DTT, 1 mM NaNa, 5 mM MgCI,, 100 mM NaCI), filtered through nitrocellulose filters (Schleicher and Schuell BA 85, pore size 0.45 pm), and washed twice with 2 ml of the same buffer. Radioactivity on the filter corresponding to ['HIGDP-bound protein was deter- mined by liquid scintillation counting. All measures were performed in duplicate. The resulting graph is an exponential where k- , = In 2/ h, , k- , being the GDP dissociation rate constant and tr the time after which 50% of labeled GDP is dissociated.

GTP Binding and Dissociation-Kinetics of GTP dissociation were determined as for GDP dissociation except that [y-:"P]GTP (2000 cpm/pmol, Amersham Corp.) was used instead of [&'H]GDP. In the presence of EDTA, the diminution of radioactivity on the filter is only due to the exchange of [y-"'PIGTP with unlabeled GTP. How- ever in the presence of M e , the measured decrease in radioactivity is due to the exchange reaction as well as to the hydrolysis of the [y- "'Plphosphate of GTP coming from the intrinsic GTPase activity of the rap2 protein. The GTP dissociation rate constants measured in the presence of M e must therefore be corrected accordingly. The observed rate constant (kJ is related to the GTP dissociation rate constant ( k - J and the rate of GTPase activity (kcTpnsp) by the equation: k, = k-, + k ~ ~ p ~ ~ (for determination of k ~ . ~ l , ~ ~ ~ , see below).

GTP binding in the presence of Mg" was also assessed by a blot overlay technique, using the following procedure. 500 ng of each purified rap2 protein was submitted to SDS-PAGE and transferred to nitrocellulose. The filter was successively incubated for 1 h in 10 mM Tris buffer, pH 7.5, containing 0.15 M NaCI, 5% BSA, and 0.05% Tween-20 in order to allow renaturation of the proteins, then 15 min in 20 mM Tris buffer, pH 7.5, containing 10 mM MgCl,, 2 mM DTT, 0.3% BSA, and 0.1% Nonidet P-40 and 30 min in the same buffer with 30 pCi of [m-:"P]GTP (3000 Ci/mmol, Amersham Corp.). The filter was washed three times for 10 min in 10 mM Tris buffer, pH 7.5, containing 0.15 M NaC1,0.5% BSA, 0.5% Triton X-100, and 0.2% SDS, dried, and autoradiographed. All incubations and washes were performed a t room temperature.

GTPase Activity-Rap2 proteins (10 p ~ ) were loaded with 50 p~ [m-:"P]GTP (4000 cpm/pmol) in 50 mM Tris, pH 7.5, containing 1 mM EDTA and 1 mM DTT for 15 min a t 37 "C in a total volume of 10 pl. The GTPase reaction was initiated by adding MgCI, to a final concentration of 10 mM. At the indicated times, 2-pl portions were removed and mixed with 2 111 of a solution containing 0.2% SDS, 5 mM EDTA, 50 mM GDP, and 50 mM GTP a t 4 "C. Samples were incubated a t 70 "C for 2 min to dissociate protein-bound nucleotides, and 1-pl aliquots were spotted onto polyethyleneimine-cellulose-cov- ered thin layer chromatography plates (29). They were developed in 0.6 M sodium phosphate buffer, pH 3.4, for 30 min, dried, and autoradiographed. The GTP and GDP spots were excised and radio- activity was measured by liquid scintillation counting.

RESULTS

Expression of the rap2 Protein in E. coli-In order to produce the human rap2 protein in E. coli, neither as a truncated nor as a fusion product, we have engineered its cDNA to position it under the control of the IPTG-inducible tac promoter, such as constructed by Tucker et al. (37) to produce ras p21 in a soluble form in E. coli. When introduced into bacteria overexpressing the lac repressor and upon in- duction with IPTG, the rap2 protein accumulates to a level representing approximately 3-8% of total E. coli proteins (Fig. 2, lanes 2 and 3 ) . The identity of this protein was confirmed

1 2 3 69 )I)*-

21.5)

4317

4 5 6 7 8

""

14.3,

FIG. 2. Expression and purification of rap2 proteins. Total extracts of noninduced and induced bacteria carrying the ptac-rap2 vector as well as 5 pg of each purified proteins were treated with SDS sample buffer and separated on SDS-PAGE. The gel was stained with Coomassie Brilliant Blue. Lane 1, molecular mass markers in kilo- daltons as indicated; lane 2, noninduced bacteria; lane 3, induced bacteria; lane 4, wild-type rap2; lane 5, Val-12 rap% lane 6, Asn-17 rap2; lane 7, Ala-35 rap2; lane 8, Ile-145 rap2. Arrows indicate bands corresponding to the rap2 proteins.

by using an antiserum raised against a peptide specific to the rap2 protein? The protein was also recognized by monoclonal antibody M90 directed against the GTP binding site of ras p21 (40), very well conserved in the rap2 protein (data not shown). Mutant rap2 proteins were expressed in E. coli to a similar extent as the normal protein. After lysis and high speed centrifugation of induced bacteria carrying the appro- priate vector, about 80% of the rap2 proteins (normal and mutant) was present in the supernatant which indicated that they were produced in a soluble form (data not shown). This enabled us to purify them, without prior denaturation, by chromatography on QAE-Sepharose and gel filtration on U1- trogel AcA54 columns. As shown in Fig. 2 (lanes 4-8), the rap:! proteins (normal and mutant) were purified to near homogeneity as judged by SDS-PAGE. The purified rap2 proteins (normal as well as mutant) bound ['HIGDP with a stoichiometry of 0.5-0.7 mol/mol of protein showing that our preparations contained 50-70% active molecules.

Kinetics of GDP Dissociation from Normal and Mutant rap2 Proteins-The dissociation of ['HIGDP from the normal rap2 protein was followed a t 37 "C in the absence (Fig. 3A) or presence (Fig. 3B) of M$+ as described under "Materials and Methods.'' Under those conditions, the dissociation curve followed a single exponential, as expected for a single class of GDP binding sites. As shown in Table I, the dissociation rate constant was calculated to be 3.8 X lo-' min" in the presence of M e , a value &fold higher than that reported for H-ras p21 under similar conditions (0.79 x lo-' min") (42). That the rap2 protein releases GDP 5-fold faster than ras p21 may not be so surprising, since one of the important features of the guanyl nucleotide binding domain of rap proteins is the presence of a threonine instead of the glutamine found in ras proteins at the highly sensitive position 61 (5-7). In the absence of M e , the dissociation of GDP from the rap2 protein was increased 18-fold. Indeed, it had already been shown that the magnesium ion was involved in binding the p- and y-phosphates of GTP to ras proteins and that the nucleotide exchange rate was greatly increased in the absence of M e (14, 15, 42, 43). The similar observation that we report here with the rap2 protein suggests that, as in ras p21, the M e ion is exposed to the solvent in the GTP binding site of the rap2 protein.

In order to investigate the respective contributions of var- ious domains of the rap2 protein to its biochemical activities, we have examined the effects of four amino acid substitutions in the rap2 protein, localized a t crucial positions of different

I. Lerosey and J. de Gunzburg, unpublished results.

by guest on April 8, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4318 Biochemical Characterization of the rap2 Protein A

TIME (mln)

FIG. 3. Kinetics of GDP dissociation from normal and mu- tant rap2 proteins in the absence or presence of 10 mM MgC12. A, in the absence of Mg2‘. 1 p M rap2 protein and 5 p M [8-’H]GDP were incubated with 2.5 mM EDTA for 15 min at 37 “C. A 100-fold excess of GDP was added and the exchange of labeled GDP was measured as described under “Materials and Methods.” B, in the presence of 10 mM Me. The experiment was performed as described above except that the concentration of Mg2‘ was adjusted to 10 mM just before a 100-fold excess of GDP was added.

TABLE I GDP dissociation rate constants for normal and mutant rap2 proteins

The dissociation rate constants were determined as described in the legend to Fig. 1 and “Materials and Methods.” They represent average of three independent experiments; the standard deviation was less than 10%.

In the presence of In the absence 10 mM Mg2+ of M$+

t% k-, x 10’ t* k-, x loz Protein

min min” min min” Wild-type 18 3.8 1 69 Val-12 42 1.6 1.5 46 Asn-17 0.75 92 1 69 Ala-35 18 3.8 0.9 76 Ile-145 5 13.8 0.25 276

regions of the protein (see Fig. 1). The substitution of Gly-12 by Val decreases the GDP

dissociation rate roughly 2-fold in the presence of 10 mM MgC12 as shown in Table I. The effect is very similar to that described for ras p21 where such a substitution leads to an approximately 3-fold decrease of the GDP dissociation rate constant (42). The Val-12 rap2 protein, like the wild-type protein, exchanges its GDP more rapidly in the absence than in the presence of M e .

The amino acid at position 17 in ras p21 has been shown to be essential for proper binding of GTP; indeed, the substi- tution of Asn for Ser in p21 ras dramatically decreases the affinity for GTP but has less pronounced effects on the affinity for GDP (44). Moreover, the three-dimensional struc-

ture of the ras p21-GTP complex shows that this amino acid participates in the coordination of the magnesium bound to GTP (14,15). Consequently, we were interested in investigat- ing the properties of a rap2 protein carrying the same Asn-17 mutation. Our results show that the GDP dissociation rate of this mutant protein was unaffected by the presence or absence of M e and was similar to that obtained with the normal rap2 protein in the absence of M C (Fig. 3 and Table I). This suggests that similarly to ras p21, the serine 17 residue is involved in the coordination of Mg2+ and that the substitution of Ser by Asn greatly disturbs this relation.

Position 35 is part of the 32-42 region which constitutes the putative site of interaction of ras p21 with its effector and is very well conserved in the rap2 protein (See Fig. 1) (5). The crystal structure of ras p21 shows that threonine 35 is also involved in M$+ coordination and proper binding of GTP (14, 15). Our experiments demonstrate that the GDP ex- change rate constant of the rap2 protein is not altered by the Thr to Ala-35 substitution, neither in the presence nor in the absence of M F (Table I). In contrast with our observation with the substitution a t position 17, these results indicate that the Thr to Ala substitution a t position 35 seems to have no effect on the coordination of M C by the rap2 protein in the GDP bound form.

A Thr to Ile mutation a t position 144 has been described in ras p21 as a substitution resulting in a decreased affinity for GTP, attributable to an increased rate of dissociation of the nucleotide (45). We were interested to examine the effect of such a mutation in the rap2 protein and, therefore, created an Ile-145 mutant which corresponds to the Ile-144 mutant of ras p21 (see Fig. 1). Fig. 3, A and B, and Table I show that the Ile-145 rap2 mutant is characterized by increased GDP exchange rate constants, both in the absence and in the presence of Mg2+ (four times faster than wild-type protein in the presence of M F ) . The observed effect of this mutation is of much lower magnitude than for ras p21.

GTP Binding and Dissociation-The replacement of GDP by GTP constitutes an essential step in the activity of ras proteins, as the GDP-bound form represents the inactive state and the GTP-bound form the active state. Consequently, we investigated the GTP binding ability of normal and mutant rap2 proteins.

Purified rap2 proteins were loaded with [y3’P]GTP in solution in the absence of M F as described under “Materials and Methods.” The wild-type, Val-12, and Ile-145 rap2 pro- teins bound 0.5-0.7 mol of GTP/mol of protein, consistent with what had been established with [3H]GDP. By contrast, the Ala-35 protein bound only 10% of the expected GTP, whereas the Asn-17 had no measurable GTP-binding activity. GTP binding was also assayed in the presence of M F by a blot overlay technique. Autoradiography of the filter (Fig. 4) shows that except for the Asn-17 mutant, all of the rap2 proteins bound GTP, but to different extents. In particular,

1 2 3 4 5

c

14.3,

FIG. 4. GTP binding activity of normal and mutant rap2 proteins. 500 ng of each purified rap2 proteins were boiled 3 min in SDS sample buffer, separated on SDS-PAGE, and electrotransferred to nitrocellulose. The filter was probed with [a-”PIGTP and washed as described under “Materials and Methods.” Lane I, normal rap$ lane 2, Val-12 rap$ lane 3, Asn-17 rap$ lane 4, Ala-35 rap$ lane 5, Ile-145 rap2.

by guest on April 8, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Biochemical Characterization of the rap2 Protein 4319

the Ala-35 mutant displayed a decreased and the Ile-145 mutant an increased labeling compared with the normal rap2 protein. These first results indicate that, irrespective of the presence of M$+, the Asn-17 and Ala-35 substitutions affect the GTP binding ability of the rap2 protein.

We measured the GTP dissociation rate constants of wild- type, Val-12, Ala-35 and Ile-145 rap2 proteins, first in the absence of M P (constants are listed in Table 11). The rate constants for the normal and Val-12 rap2 proteins were similar. The Ile-145 mutation led to a 4-fold increase of the GTP dissociation rate, and the Ala-35 mutation had an even more pronounced effect. In the case of the Asn-17 mutant, our failure to measure GTP binding activity is probably the consequence of a very high dissociation rate (see below).

We also wished to investigate the dissociation of GTP from the rap2 protein in the presence of Mg+, since GTP is normally complexed with M e in the cell. As described under “Material and Methods,” the observed dissociation rate con- stants had to be corrected for the loss of [ -“P]phosphate from the protein due to its intrinsic GTPase activity. The latter was either measured as described under “Materials and Meth- ods” or by filtration under the same conditions used to assay for GTP dissociation (except that no excess of unlabeled nucleotide was added) with similar results. The resulting GTP dissociation rate constants are listed in Table 11. For normal and Val-12 rap2 proteins, the GTP dissociation rate constants (k -J could not be determined precisely because of a very slow nucleotide exchange. Indeed, after correction for GTPase activity, it appeared that only about 15% of bound GTP had dissociated after 2 h. GTP exchange from the normal rap2 protein is therefore much slower than GDP exchange. In the case of the Asn-17 mutant, no GTP binding activity could be measured in solution even in the presence of 10 mM Me. Very poor binding of GTP was measured for the Ala-35 mutant irrespective of the presence of M e ions; only 10% of the sites could be occupied and GTP dissociation from this mutant was a t least 25-fold faster than from the wild-type rap2 protein. These results confirm that the Asn-17 and Ala- 35 substitutions greatly disturb the GTP binding ability of the rap2 proteins. Finally, the Ile-145 mutant displayed an approximately 4-fold increased GTP dissociation rate con- stant when compared with the normal rap2 protein. However, as for GDP, the effect is of lower magnitude than that de- scribed for the Ile-144 ras p21.

GTPase Actiuity-Since GTP hydrolysis plays an important

TABLE I1 GTP dissociation rate constants for normal and mutant rap2 proteins

The rate constants were determined as described under “Material and Methods;” in the presence of M e the observed rate constants were corrected for the loss of the [y-’”Plphosphate due to the intrinsic GTPase activity of the rap2 protein.

In the presence of In the absence 10 mM Mg2‘ of M$+

t“. k-, x 10’ tnk k-, x 10’ Protein

rnin min” min min” Wild-type >500” C0.14” 2.5 27 Val-12 >500” <0.14a 2 34 Asn-17 NDb NDb NDb ND‘ Ala-35 20 3.3 0.25 270 Ile-145 120 0.36 0.8 86

a Normal and Val-12 rap2 proteins exchange GTP very slowly. The GTP dissociation rate constants for these proteins were calculated by extrapolation of the kinetics determined during 3 hours.

* ND, not determined. For the Asn-17 mutant, irrespective of the presence or absence of M e , no GTP binding activity could be measured. This probably reflects a very high dissociation rate (see “Discussion”).

role in the regulation of GTP-binding proteins, we have investigated the intrinsic GTPase activity of the normal and mutated rap2 proteins. Equal amounts of active proteins were incubated with [a-:”P]GTP and GTP hydrolysis was followed by the generation of GDP after adjusting the MgC12 concen- tration to 10 mM. As shown in Fig. 5, the normal and mutant rap2 proteins were all capable of hydrolyzing GTP, but with different velocities. The GTPase rate constant of the normal rap2 protein was measured to be 0.5 X 10” min” which represents a 2-fold diminution compared with ras proteins. The Val-12, Ala-35, and Ile-145 mutations led to a decrease in the GTP hydrolysis rate of 2-, 3-, and 3.5-fold, respectively, compared with the normal rap2 protein. The GTPase rate constants ( k C . ~ p ~ . ~ ) for these mutants are 0.25 X 10“ rnin”, 0.16 X lo-* rnin”, and 0.14 x lo-* rnin”. Although we were unable to detect GTP binding to the Asn-17 rap2 protein, it hydrolyzed GTP to a rate comparable to that of the wild-type protein. This mutant is therefore able to bind GTP, but with a very high dissociation rate.

DISCUSSION

During the last 6 years, an increasing number of ras-related genes encoding small molecular weight GTP-binding proteins have been identified. Among these, the products of the rap genes are of great interest as they share striking similarities with ras p21s such as the region of interaction with their putative effector and the C-terminal domain responsible for postranslational processing and membrane binding (5-7). We describe here the biochemical characterization of the rap2 protein as well as the effects of four different mutations, localized in distinct domains of the protein. The human rap2 cDNA was cloned into the ptac bacterial expression vector, allowing the production in E. coli of high levels of soluble recombinant protein which was purified without the use of denaturing agents. In the same way, we have produced and purified mutated rap2 proteins, containing the following sub- stitutions: Gly-12 to Val, Ser-17 to Asn, Thr-35 to Ala, and Thr-145 to Ile (see Fig. 1). The availability of high amounts of active purified proteins allowed us to perform a detailed biochemical characterization of the product of the rap2 gene.

We show there that the rap2 protein actually binds GDP and GTP and also exhibits a low intrinsic GTPase activity. The GDP dissociation rate constant of the rap2 protein in the presence of 10 mM M e is approximately five times faster than that of ras p21 (42). This difference could be due to the presence of a threonine a t position 61 in the rap2 protein, whereas ras p21s possess a glutamine. Indeed, the amino acid a t position 61 is part of the nucleotide binding site, and it has

A

# 4 . # . # o t u ’ a b @ * . e4.W WM”

W ~ / ( w w

B C

D 0 5 10 1530 60 0 5 1015 3060 0 5 10 15 30 60

E

0 5 1 0 1 5 3 0 6 0 0 5 I 0 1 5 3 0 6 0

FIG. 5. GTPase activity of rap2 proteins. rap2 protein (10 pM) was preincubated with [CY-”~P]GTI-‘ (50 p ~ ; 4000 cpm/pmol) in the presence of 1 mM EDTA for 15 min at 37 “C. The GTPase activity was initiated by adding MgC12 to a final concentration of 10 mM. Aliquots were removed from the mixture at indicated times (minutes) and analyzed by thin layer chromatography as described under “Ma- terials and Methods.” A, wild-type rap2; B, Val-12 rap2; c, Asn-17 rap2; D, Ala-35 rap2; E, Ile-145 rap2.

by guest on April 8, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4320 Biochemical Characterization of the rap2 Protein

TABLE I11 Comparative effects of the mutations on the GTPIGDP binding

wooerties of the rau2 and ras oroteim Protein

Mutation rap2 p21" H-ras p21

Gly + Val-12 GDP exchange rate 2-fold reduced

GTP exchange similar to the wild-type protein

Ser 4 Asn-17 GDP exchange unaffected by the presence of Mg2+

No measurable GTP bind- ing (very high dissocia- tion rate)

Thr Ala-35 GDP exchange similar to the wild-type protein

Poor GTP binding (in- creased dissociation rate >25-fold)

Thr 4 Ile-145 Increased exchange rate for GDP and GTP (3-4-fold)

3-fold decrease of the GDP dissociation

&fold decrease of the rate

GTP dissociation rate'

No effect on GDP binding

Decreased affinity for GTP due to an in- creased dissocia- tion rate'

Not determined

Decreased affinity for GTP due to an in- creased rate of dis- sociation (25-f0ld)~

Results for the rap2 protein come from this study. Data from Ref. 42.

e Data from Ref. 44. Data from Ref. 45.

been shown for ras p21 that a Leu 61 mutation increases the GDP dissociation rate (14); our results are therefore consist- ent with the data obtained for ras proteins. Moreover, the release of GDP from the rap2 protein is drastically increased when Mg2+ ions are removed this reveals that the M$+ ion is probably coordinated in the same way than in ras p21 and is also exposed to the solvent. An additional interesting fea- ture of the rap2 protein is that the release of GTP is much slower (at least 25-fold) compared with that of GDP. This differs widely from the case of H-ras p21 that had been shown to exchange GTP %fold faster than GDP (42) and suggests that the rap2 protein exhibits a higher affinity for GTP than GDP. The GTPase activity of the rap2 protein has been calculated to be 2-fold reduced compared with ras p21. This is probably due to the presence of a Thr at position 61, since it has been proposed that the Gln-61 of ras p21 plays an essential role in the hydrolysis of GTP, by activating a water molecule for the nucleophilic attack of the y-phosphate (14, 46, 47).

Four single amino-acid changes were introduced in regions of the rap2 protein important for guanine nucleotide binding and hydrolysis; a comparison with the effects of similar mu- tations in ras p21 is summarized in Table 111. The Gly-12 to Val substitution in the rap2 protein leads to a 2-fold decrease of the GDP dissociation rate constant and of the GTPase activity. The affinity for GTP is not modified and GTP dissociates from the protein very slowly, as observed for the normal rap2 protein. These properties, except for the slow GTP dissociation, are comparable with those of a Val-12 ras p21 (see Table 111). Nevertheless, it is of interest to note that this substitution does not confer oncogenicity to the rap2 p r ~ t e i n . ~

Substitution of the Ser at position 17 for Asn, in ras p21, decreases the affinity for GTP without modifying the affinity for GDP (44). The rap2 protein, as well as all the small GTP

I. Lerosey, unpublished observation.

binding proteins, translation elongation factors and a sub- units of heterotrimeric G proteins also possess a Ser at the position homologous to 17 in ras p21 (48). We show here that the mutation of this residue to Asn in the rap2 protein has profound effects on the GDP dissociation rate in the presence of M$+ as well as on GTP binding. Indeed, the Asn-17 rap2 protein, irrespective of the presence or absence of M$+, releases its GDP as fast as the wild-type rap2 protein in the absence of M P . Moreover, we failed to demonstrate GTP binding for this mutant, even in the presence of M$+. Never- theless, the Asn-17 rap2 protein exhibits a GTPase activity slightly lower than that of the normal protein which leads us to conclude that the protein is able to bind GTP, but with a very high dissociation rate. All these results show that the residue at position 17 in the rap2 protein is involved in the coordination of the M$+ ion and proper binding of GTP. This is consistent with the structure of ras p21, where Ser-17 was identified as one of the ligands of the M F ion (14, 15). One would then expect that the rap2 protein, as ras p21, binds M$+ as a monodentate @-phosphate complex in the GDP bound form, and as a bidendate @-y-phosphate complex, in the GTP-bound form, Ser-17 being essential for a correct coordination of the ion.

We have studied the effect of the Thr to Ala substitution at position 35, in a region which is thought to constitute the effector domain of proteins from the ras family. Determina- tion of the three-dimensional structure of ras p21 pointed out that residue 35 is also involved in the coordination of Mg2+ and the amide nitrogen of Thr-35 forms an interaction with an oxygene of the y-phosphate group (14, 15). Our results demonstrate that the GDP exchange from this Ala-35 rap2 mutant is not modified compared to the wild-type, neither in the presence nor in the absence of M P . However, the Ala- 35 mutant binds GTP poorly independently of the presence of M F and has a reduced GTPase activity. These results suggest that, as in ras p21, the residue at position 35 in the rap2 protein is important for GTP binding and interacts with the y-phosphate; this could explain the decrease of the GTPase rate constant of the Ala-35 mutant. However, it is likely that threonine 35 plays a lesser role in the coordination of the Mg2+ ion than serine 17. As this mutation at position 35 only affects GTP binding without modifying GDP ex- change, we can hypothesize that region 32-42 should have a different conformation whether the protein is in the active GTP bound form or in the inactive GDP bound form. This shift, which has been demonstrated in the case of ras p21 (15, 47), is probably important for the interaction of the rap2 protein with its putative effector(s).

Finally, substitution of Thr-145 by Ile in the rap2 protein increases the nucleotide dissociation rates, both for GDP and GTP. This effect is much less pronounced than that reported for ras p21 (45). On the other hand, the GTPase activity of this mutant is reduced. These data show that the residue at position 145 in the rap2 protein is probably part of or close to the guanine nucleotide binding site. In the case of ras p21, residue 144 is not directly implicated in the binding of the guanine base but its substitution probably affects the local conformation of the 143-147 region (14). The GTPase activity of an Ile-144 ras p21 has, to our knowledge, never been investigated. We can suppose that the Ile-145 substitution in the rap2 protein leads to a conformational change that shifts the GTP out of its optimal position at the catalytic site for GTPase activity.

Taken together, these results suggest that the overall ar- chitecture of the rap2 protein resembles that of ras p21, with minor differences, and that similar structural elements are

by guest on April 8, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Biochemical Characterization of the rap2 Protein 4321

implicated in the conformational change between GDP- and 21. Schafer, W. R., Kim, R., S t e m , R., Thorner, J., Kim, S.-H., and GTP-bound forms. Characterization of the biochemical prop- erties of the rap:! protein is essential to understand its phys- 22. Casey, P. J., Solski, P. A., Der, C. J., and Buss, J. E. (1989) Proc.

rently being used to search for potential effectors. (1989) EMBO J. 8,1093-1098

Rine, J. (1989) Science 245, 379-385

Natl. Acad. Sci. U. S. A. 86,8323-8327 iological function(s) and availability of pure protein is cur- 23. Gutierrez, L., Magee, A. I., Marshall, C. J., and Hancock, J. F.

24. Gilman, A. G. (1987) Annu. Reu. Biochem. 56,615-649 Acknowledgments-We thank Dr. A. Wittinghofer for the generous 25. Broek, D., Toda, T., Michaeli, T., Levin, L., Birchmeier, C.,

gift of the ptac-u-H-ras vector and Dr. J. C. Lacal for monoclonal Zoller, M., Powers, S., and Wigler, M. (1987) Cell 48, 789-799 antibody M90. We are indebted to Vkronique Pizon, Nicolas Touchot, 26. Robinson, L. C., Gibbs, J. B., Marshall, M. S., Sigal, I. S., and and Ahmed Zahraoui for helpful discussions, as well as Catherine Tatchell, K. (1987) Science 235, 1218-1221 Rehoux for secretarial assistance. 27. West, M., Kung, H. F., and Kamata, T. (1990) FEBS Lett. 259,

245-248

1. 2. 3.

4. 5.

6.

7.

8. 9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

-~~ -~~

REFERENCES 28. Wolfman, A., and Macara, I. G. (1990) Science 248,67-69

Chardin, P. (1988) Biochimie 70,865-868 Barbacid, M. (1987) Annu. Reu. Biochem. 56, 779-827

S. L., and Goeddel, D. V. (1987) Cell 48, 137-146

29. Trahey, M., and McCormick, F. (1987) Science 238, 542-545 30. Adari, H., Lowy, D. R., Willumsen, B. M., Der, C. J., and

Lowe, D. G.9 Capon, D, J.7 Delwart, E., SakaWchi, A. y.7 Naylor, 31. Gibbs, J. B., Schaber, M. D., Allard, W. J., Sigal, 1. S., and McCormick, F. (1988) Science 240,518-521

Chardin, P., and Tavitian, A. (1986) EMBO J. 5, 2203-2208 Scolnick, E. M. (1988) Proc. Natl. Acad. Sci. U. S. A. 85,5026-

Pizon, V., Chardin, P., Lerosey, I., Olofsson, B., and Tavitian, A. 32. CalBs, c., Hancock, J. F., Marshall, c. J., and ~ ~ 1 1 , A. (1988) 5030

Nature 332,548-551

M. S., Scolnick, E. M., Sigal, I. S., and Gibbs, J. B. (1988) Nature 335, 90-93

(1988) Oncogene 3,201-204

Acids Res 16,7719

and Takai, Y. (1988) J. Biol. Chem. 263,18965-18971

Pizon, v., LeroseY, 1.9 Chardin, p.7 and Tavitian, A. (1988) Nucleic 33. Vogel, U. S., Dixon, R., Schaber, M. D., Diehl, R. E., Marshall,

Kawata, M., Matsui, Y., Kondo, J., Hishida, T., Teranishi, Y.,

Madaule, P., and Axel, R. (1985) Cell 41,31-40 Yeramian, p., Chardin, p., Madauk p., and Tavitian, A. (1987) 36. Kitayama, H., Sugimoto, y., Matsuzaki, T., Ikawa, Y., and No&,

Nucleic Acids Res 15,1869 Touchoty N . ~ p., and Tavitian, A. Proc. NatL. 37. Tucker, J., Sczakiel, G., Feuerstein, J., John, J., Goody, R. S.,

M. (1989) Cell 56, 77-84

Acad. Sci. U. S. A. 84,8210-8214 Zahraoui, A., Touchot, N., Chardin, p., and Tavitian, A. (1988) 38. Certa, u., Bannwarth, w. , Stuber, D., Gentz, R., Lanzer, M., Le

and Wittinghofer, A. (1986) EMBO J. 5, 1351-1358

Nucleic Acids Res. 16,1204 Matsui, Y., Kikuchi, A., Kondo, J., Hishida, T., Teranishi, Y.,

Grice, S., Guillot, F., Wendler, I., Hunsmann, G., Bujard, H.,

and Takai, Y. (1988) J. Biol. Chem. 263, 11071-11074 and Mous, J. (1986) EMBO J. 5,3051-3056

39. Laemmli, U. K. (1970) Nature 227, 680-685 Zahraoui, A., Touchot, N., Chardin, p., and Tavitian, A. (1989) 40. Lacal, J. C., and Aaronson, S. A. (1986) Proc. Natl. Acad. S C ~ . U.

J. Biol. Chem. 264,12394-12401 Psi, E. F., Kabsch, w., Krengel, u., Holmes, K. c., John, J., and 41. Bradford, M. M. (1976) A d . Biochem. 72, 248-254

Wittinghofer, A. (1989) Nature 341, 209-214 42. John, J., Frech, M., and Wittinghofer, A. (1988) J. Biol. Chem. Milburn, M. V., Tong, L., De Vos, A. M., Brunger, A., Yamaizumi,

z., Nishimura, s., and Kim, S.-H. (1990) Science 247,939-945 43. Hall, A., and Self, A. J. (1986) J. BWl. Chem. 261,10963-10965 Sigal, 1. s., Gibbs, J. B., D'Alonzo, s. J., and Scolnick, E. M. 44. Feig, L. A., and Cooper, G. M. (1988) Mol. Cell. Biol. 8, 3235-

Willurnsen, €3. M., Papageorge, A. G., Kung, H. F., Bekesi, E., 45. Feig, L. A., Pan, B.-T., Roberts, T. M., and Cooper, G. M. (1986)

46. Feuerstein, J., Goody, R. S., and Webb, M. R. (1989) J. Biol.

34. McCormick, F. (1989) Cell 56, 5-8 35. Schejter, E. D., and Shilo, B.-Z. (1985) EMBO J. 4, 407-412

S. A. 83,5400-5404

263,11792-11799

(1986) Proc. Natl. Acad. Sci. U. S. A. 83,4725-4729 3243

Robins, T., Johnsen, M., Vass, W. C., and Lowy, D. R. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,4607-4611 Mol. Cell. Biol. 6, 2646-2654

Willumsen, B. M., Christensen, A., Huhbert. N. L., PaDaaeorae, chem. 264. mm-619n A. G., and Lowy, D. R. (1984) Nature 310,583-586

and Lowy, D. R. (1984) EMBO J. 3,2581-2585 G. A., and Goody, R. S. (1990) Nature 345, 309-315

(1989) Cell 57, 1167-1177 390-395

- - - . ~ ~ ~~ -, ~ ~ - - ~ - - -

19. Willumsen, B. M., Norris, K., Papageorge, A,, Hubbert, N. L., 47. Schlichting, I., Almo, S. C., Rapp, G., Wilson, K., Petratos, K.,

Lentfer, A., Wittinghofer, A., Kabsch, W., Pai, E. F., Petsko,

20. Hancock, J. F., Magee, A. I., Childs, J. E., and Marshall, C. J. 48. Powers, S., ONeill, K., and Wigler, M. (1989) Mol. Cell. Biol. 9,

by guest on April 8, 2018

http://ww

w.jbc.org/

Dow

nloaded from

I Lerosey, P Chardin, J de Gunzburg and A Tavitiancharacterization and site-directed mutagenesis.

The product of the rap2 gene, member of the ras superfamily. Biochemical

1991, 266:4315-4321.J. Biol. Chem. 

  http://www.jbc.org/content/266/7/4315Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/266/7/4315.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on April 8, 2018

http://ww

w.jbc.org/

Dow

nloaded from