Expression, Purification, and Physicochemical Characterization of a ...

. 267, No. 33, Issue of November 25, PP. 23759-23766,1992

THEJOURNAL OF BIOLOGICALCHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc

Printed in U.S.A.

Expression, Purification,and Physicochemical Characterizationof a Recombinant Yersinia Protein Tyrosine Phosphatase* (Received for publication, July 6, 1992)

Zhong-Yin ZhangS,James C. ClemensS, Heidi L. SchubertS, Jeanne A. Stuckeys, Mark W. F. Fischers, Daniel M. Humej, Mark A. SaperSj, andJack E. DixonSfl From the SDepartment of Biological Chemistry and the §Biophysics Research Division, Universityof Michigan, Ann Arbor, Michigan 48109

The Yersinia protein tyrosinephosphatase (PTPase) Yop51, a C235R point mutation (Yop51*), and a protein lacking the first 162 amino acids at the NH2 terminus (Yop51*A162) have been overexpressedin Escherichia coli and purified tohomogeneity through the use of CM Sephadex C25 cation exchange chromatography followed by Sephadex G-100 gel filtration. Greater than 50 mg of homogeneous Yop51*and Yop51*A162 can be obtained from a single liter of bacterial culture, whereas the same procedure yields only 5 mg of pure Yop51. Large, diffraction-quality crystals havebeen obtained forYop51*A162. Size exclusion chromatography,sedimentationequilibrium, and enzyme concentration dependence experiments have established that the Yersinia PTPases exist and function as monomers in solution, Yop5l and Yop51* display identical UV, CD, and fluorescence spectra and have identical kinetic and structural stability properties. These full-length Yersinia PTPases have 31%ahelix, an emission maximum of 342 nm, a turn-over number of 1200 s” at pH 5.0, 30 “C, and anunfolding AG value of 6 kcal/mol at 25“C. Yop51*A162 has very similar kinetic and fluorescence characteristics to the full-length molecules, whereas its CD and UV spectra show noticeable differences due to the elimination of 162 NHz-terminal residues. The Yersinia PTPases are by far the most active PTPases known, and their kinetic parameters are extremely sensitive to the ionic strength of reaction medium.

tyrosine phosphorylation (Hunter, 1989; Tonks andCharbonneau, 1989). Collectively, the PTPases appear to fall into at least two “subfamilies.” One subfamily has variable extracellular domains, a transmembrane-spanningregion, and usually two cytoplasmic PTPase domains. Members of this subfamily includes CD45, LAR, and PTPl8 (Guan and Dixon, 1990). This subfamily of receptor-like PTPases is expanding rapidly, and there are more than 20 currently known (Fischer et al., 1991). HPTP-P (Krueger et al., 1990) and DPTlOD (Yang et al., 1991; Tian et al., 1991) are members of this subfamily but have only a single PTPase domain. The second subfamily of PTPases consist of enzymes whichhave a single PTPase domain and have an intracellular location. Members of this subfamily include PTP1,PTPlB,as well astheT-cell PTPases (Tonks et al., 1988; Guan et al., 1990; Cool et al., 1990). Allof the PTPases have a conserved domain of approximately 300 residues, which is responsible for catalysis. Sequences located outside of the catalytic domain of the PTPases appear tobe involved in localization and/or regulation (Woodford-Thomas et al., 1992; Frangioni at al., 1992). A Cys residue in PTP1, Cys”‘, has been shown to be essential for catalysis (Guan and Dixon, 1991). Site-directed mutagenesis was used to alter this residue to either Ala or Ser, and both changes resulted in atotal loss of enzyme activity. Utilizing conserved amino acids within the active site of the PTPase, Guan and Dixon searched for proteins which would have sequence identity to the mammalian PTPases. One of the proteins identified was Yop51. Yop51shares a significant degree of identity to the mammalian PTPases (Guan and Dixon, 1990),even though it is found in a pathogenic bacteria Protein tyrosine phosphorylation plays a crucial role in the (Yersinia). Yersinia is the bacterium responsible for plague, or Black regulation of cell proliferation and differentiation (Hunter Death. Guan and Dixon (1990) demonstrated that Yop51 is a and Cooper, 1987; Yarden and Ullrich, 1988). The levels of PTPase which selectively hydrolyzes only phosphotyrosinetyrosine phosphorylation appear to be modulated within the cell by both tyrosine kinases and protein tyrosine phospha- containing proteins and which contains a Cys residue essentases (PTPases).’ The PTPases constitute a family of en- tial for catalysis. These features arealso shared by the mamzymes that function in regulating the extent and duration of malian PTPases.In addition, Bliska et al. (1991) demonstrated that it is in fact the phosphatase activity of Yop51 * This work was supported by a grant from the National Institutes which is necessary for bacterial pathogenesis. of Health, National Institute of Diabetes and Digestive and Kidney There are alimited number of papers addressing structural Diseases Grant 18849,and the Walther Cancer Institute. The crys- properties and mechanisms of the PTPases. This is in part tallographicstudies were supportedinpart by theUniversity of due to thelack of sufficient amounts of pure enzyme. Isolation Michigan Program in Protein Structure and Design (grants to J. A. S. and M. A. S.) and the University of Michigan Cancer Center (grant of receptor like PTPases requires detergent solubilization, t o M. A. S.).The costs of publication of this article were defrayed in and P T P l B is rapidly altered by proteolysis (Fischer et al., part by the payment of page charges. This article must therefore he 1991). Recently, recombinant techniques have been used to hereby marked “advertisement” in accordance with18 U.S.C. Section produce recombinant catalytic domains of receptor-like 1734 solely to indicate this fact. PTPases whose mechanism have been examined (Cho et al., 7 To whom correspondence should be addressed. 1992). In order to facilitate the structural and mechanistic The abbreviations used are: PTPase, protein tyrosine phosphastudies of PTPases, we have developed a rapid and simple tase; MES, 2-(N-morpholino)ethanesulfonicacid; PCR, polymerase procedure to obtain large quantities of homogeneous recomchain reaction; PEG, polyethylene glycol; p-NPP,p-nitrophenyl phosphate; PAGE, polyacrylamide gel electrophoresis. binant Yersinia PTPase. Thecoding sequence of the PTPase

23759

23760

Expression Characterization and

of Yersinia PTPase

NdeI and EcoRI restriction endonuclease sites. A NdeI restriction site was designed to include the ATG initiator codon of the PTPase. An EcoRI site was positioned downstream of the Yop51 termination codon. The PCR primers used to generate this DNA fragment were 5'ATTAAGGAGGGACATATGAACTTATCATT and 5'GTCGGATCCTGAATTCGAATAAATATTTACATTAGC for the 5' and 3' primer, respectively. PCR using these primers and pYop51/GEX-KG as a template produced a 1.4-kilobase pair DNA fragment (Guan and Dixon, 1990). This fragment was subcloned into the NdeI and EcoRI sites of pT7-7 (Tabor and Richardson, 1985) to generate pYop51*/ pT7. This construct was verified by restriction digestion and the DNA sequence determined. Here we noted that a single nucleotide change results in a C235R mutation. The mutation arose most probably via PCR. The protein encoded by the C235R mutation has been called Yop51*. To correct the substitution at position 235 a NarIKpnI fragment from the "wild type" sequence was substituted for the region containing the Cys/Arg mutation (referred to asYop51). A similar procedure was followed to generate a clone where DNA sequences encoding 162 residues at theamino terminus of the Yop51* were deleted, thus generating Yop51*A162. The same 3' primer as described above was used, but a different 5' primer was designed 5'CCGTTGCATATGCGTGAACGACCACACACT. This primer anneals 435 bases downstream from the start of translation, but upstream of the sequence coding for the conserved phosphatase domain. The primer contains sequences designed to create a new, in-frame EXPERIMENTALPROCEDURES start codon within an NdeI restriction site that is in frame with the Materials-p-Nitrophenyl phosphate (p-NPP), bovine serum al- coding sequence. PCR using these primers and pYop51*/pT7 as a bumin, Folin reagent, Sephadex G-100, CM-Sephadex C-25, gel fil- template produced a DNA fragment that was 960base pairs inlength. tration molecular weight marker (kit MW-GF-70) and Fluorinert FC- This fragment was subcloned into the NdeI and EcoRI sites of pT740 were from Sigma. Materials for SDS-polyacrylamide gel electro- 7. This construct was verified by restriction mapping and sequencing. phoresis were purchased from Bio-Rad. Polyethylene glycol 1500 was Expression and Purification of Yersinia PTPme-To purify the from Fluka. Buffers were prepared using deionized and distilled water. Yersinia PTPase, an overnight culture (10 ml) wasgrownfrom a All other chemicals were of the highest purity and were used without single colony and was used to inoculate 1 liter of 2 X YT (containing further purification. 100 pg/ml ampicillin). This culture was grown at 37 "C to an optical Construction of the Expression Plasmids-PCR was used to gen- density between 0.6 and 0.9 at 660 nm, induced with isopropyl p-Derate a DNA fragment containing the Yop5l sequence flanked by thiogalactoside to a final concentration of 0.4 mM and grown for an additional 3-4 h. The cells were harvested by centrifugation at 4800 X g for 10 min. The resulting pellets were resuspended in 40 mlof A PTPWUAIN 100 mM acetate, 100 mM NaCl, 1 mM EDTA, pH 5.7 (Buffer A), and lysed by two passages through a French press, maintaining pressures 1 468 above 1200p.s.i. at 4 "C. The soluble fraction was isolated by centriYop51* fuging the lysates at 27,000 X g for 20 min at 4 "C. The resultant clear supernatant was added to 15 ml of CM Sephadex C-25 equili163 468 brated in Buffer A. Binding of the Yersinia PTPase to the cation Yop51*A162 exchange resin was accomplished by gentle shaking at 4 "C for 1 h. 335 1 The Sephadex resins were then pelleted using a clinical centrifuge, YPTPI and the supernatantwas removed. After washing the gel three times 1 432 with 40 ml of Buffer A, the gel waspacked into acolumn. The column r n 1 was continually washed with Buffer A until the eluent had an optical density of 0 at 280 nm. The enzyme was eluted with a 150-ml linear 1 926 salt gradient from 100 to 500 mM NaCl in 100 mM acetate, 1 mM I V-I m M E G EDTA, pH 5.7. The enzyme was purified further by passing the B pooled and concentrated fractions over a Sephadex G-100 column 1 LINLSLSDLBR RYSRLYOPES GmcTogLRQN YAANXFRFQ QLTIASAXE (2.5~108cm) equilibrated in Buffer A. The protein concentration of the fractions was monitored by A280,and the enzyme activity was 51 - A m SBVANIYLTP EDTAKLUST-Y RSYQtUNSll followed using p-NPP asa substrate. Enzyme Assay-The Yersinia PTPase activity was assayed at 30 "C 1 0 1 YSWlSIXlnTL ( D I Wl(IEs3AXm YSsBSHsYLa mvxm in a reaction mixture (0.2 ml) containing 10 mM p-NPP assubstrate and 100 mM acetate, 1 mM EDTA, pH 5.5 buffer; the ionic strength 151 RSHLDPRTPP W P W H T S Q U E Q A W TWSTYSPPQ P-SSR of the buffer was adjusted using NaCl to I = 0.15 M. The reaction * 201 LTTLRmAP ATNDEWLQA C W E K L A W R DIOCCRGTAY XADLNANYIQ was initiated by addition of enzyme and quenched after 2-3 min by addition of 1 ml of 1 N NaOH. The nonenzymatic hydrolysis of the 251 3 substrate was corrected by measuring the increase in optical density without the addition of enzyme. The amount of product (p-nitro301 YFR(1soTpQS ITVEs1[IITQoYBLrmaIMAD IRTLTIREAQ QKTISYPYYE phenol) was determined from the absorbance at 405 nmusing a molar extinction coefficient of 18,000 M" cm" (Zhang and Van Etten, 351 YQNWPDCiTAV S S E Y T U U S LYDRT-R NUYESKQSSA YADDSKLRF'Y 1991). One unit of activity is defined as the amount of enzyme that is needed to hydrolyze 1 pmol of p-nitrophenyl phosphate/min at 4 401 IBCP.A(NQRT AQLIQAllCIIN DSXNSQLSVE DHYS(1MIYQR WIHTRKDEQ 30 "C. Specific activity is defined as thenumber of enzyme units/mg of protein. 451 LDYLIMRPLWIS The effect of enzyme concentration on the hydrolysis of p-NPP was investigated at pH 7.0, 25 "C, in 50 mM 3,3-dimethylglutarate, 1 FIG. 1. PTPase alignment. A, schematic alignment of Yop51*, Yop51*A162, yeast PTPase 1 (YPTPI),rat brain PTPase-1 mM EDTA, I = 0.15 M, with [SI, 50 mg of homogeneous protein/liter of bacteria.This amount of pure protein makesit possible to explore detailed studiesof PTPases. This paper describes the overexpression, purification, detailed physicochemical characterization, and crystallization of the Yersinia PTPase.

-

Expression and Characterization of Yersinia PTPase sorption a t 420 nm using a Perkin-Elmer X6 spectrophotometer, and the first order rate constantwas calculated by analysis of the experfit algorithm (Yamaoka imental data through a nonlinear least square e t al., 1981). In all cases, BSA was included in the assay buffer a t 0.5 mg/ml to minimize irreversible enzyme absorption to the quartzcell walls. Protein concentrations of the homogeneous enzyme preparations were determined by amino acid analysis using a Applied Biosystem model 420H amino acid analyzer. Total protein estimates during the purification was determined by the Lowry method (Lowry et al., 1951). The NH2-terminal sequences were determined using aApplied Biosystem model 470A sequenator. Circular Dichroism and Fluorescence Spectroscopy-Circular dichroism spectra were recorded with a Jasco 710 spectropolarimeter calibrated with ammonium d-camphor-10-sulfonate. Each spectrum was obtained as the average of three scans to optimize the signal to noise ratio and was corrected for background using the buffer solution over the range of260-180 nm using a 0.5-mm path-length cell a t 25 "C. Proteinconcentrations were 0.438 mg/ml for Yop5l and Yop51* and 0.288 mg/ml for Yop51*A162. The buffer used were 10 mM Tris-acetate, pH 6.0. The data were reported in terms of mean residue ellipticity. Fluorescence experiments were performed on a Perkin Elmer LS50 fluorimeter. Spectral measurements were made with either 280 or 295 nm excitation (slit width 3.5 nm) and emission spectra were recorded either from 290 to 400 or from 300 to 400 nm (slit width 5 nm). Emission spectra were corrected for solvent background and Raman scattering. UV spectra were recorded using a Perkin Elmer X6 spectrophotometer. Sedimentation Equilibrium-Sedimentation equilibrium experiments were performed a t 20 "C in a Beckman TL-100 tabletopultracentrifuge equipped with a TL-55 swinging bucket rotor. Thick-wall polycarbonate tubes containing 350 pl of 6.89 p M PTPase in 50 mM 3,3-dimethylglutarate, 1 mM EDTA, I = 0.15 M, pH 7.0, buffer and 200 pl of Fluorinert FC-40, were first centrifuged for 4 h a t 20,000 rpm and thencentrifuged a t 15,000 rpm for 44 h. After the sedimentation, 50-pl aliquots were collected from the meniscus, and the amount of the PTPasewas determined by the measurement of enzyme activity and by absorbance a t 280 nm. The molecular weight value was calculated from In c uersus ? plot (Chervenka, 1973). Gel Filtration-Molecular size was determined by gel filtration on a Sephadex G-100 column (2.5 X 108 cm) equilibrated with 100 mM acetate, 100 mM NaCI, 1 mM EDTA, pH 5.7, a t 4 "C. The column was standardized using gel filtration molecular mass standards: bovine serum albumin (66 kDa), carbonic anhydrase (29 kDa), cytochromec (12.4 kDa), and aprotinin (6.5 kDa)(MW-GF-70 from Sigma). The flow rate was 8 ml/h, and theelution was monitored by absorbance a t 280 nm. Urea Denaturation-Urea denaturation was studied by monitoring the change in fluorescence a t 340 nm (slit width 5 nm) with an excitation wavelength a t 295 nm (slit width 3.5 nm). Typically, ) added to appropriate urea concentration preenzyme (3.47 p ~ was of 50 mM, pared in pH 6.0 buffer with a final succinate concentration 1mM EDTA, and I = 0.15 M. After incubation a t 25 "C for a t least 6 h, the fluorescence of the samples was measured. After the fluorescence measurements, enzyme activity of all the samples were determined a t their corresponding urea concentration using p-NPP as a substrate.

23761

TABLE I Purification of the Yersinia PTPases ~

~~

PTPase Step Protein

Activity

w

units

E!:;:i

Yield Purification

units/

%

w

-fold

Yop51 Lysate 455 CM 14.6 5.81 GI00

6462 4810 4642

14.2 100 32.9 74.4 799 71.8

1 23 56

Yop51* Lysate 392 CM 52. GlOO 44.8

51382 37348 36125

Lysate 385 65.7 CM GI00 61.2

89470 232 78240 1192 75460 1233

131 718 806

100 72.7 70.3

1 5.5 6.2

100

1 5.1 5.3

Yop51*A162

A

B

87.4 84.3

C

1 2 3 4 ,

FIG.2. SDS-PAGE analysis of Yersiniu PTPases. A , Yop51*; B, Yop51; C, Yop51*Al62. Lane 1, Bio-Rad SDS-PAGE molecular size standards, from top to bottom: rabbit muscle phosphorylase b, 97.4 kDa; bovine serum albumin, 66.2 kDa; hen egg white ovalumin, 45 kDa; bovine carbonic anhydrase, 31 kDa, soybean trypsin inhibitor, 21.5 kDa; hen egg white lysozyme, 14.4kDa. Lane 2, cell lysate. Lane 3, post-CM Sephadex C25. Lane 4, post-G100. Protein samples were fractionated on 12.5% acrylamide gels and stained with Coomassie Blue.

Yop51*Al62. The CM Sephadex C25 step is particularly efficient since the PTPasebinds specifically to thenegatively charged resins while most bacterial proteins are not retained on the resin under these conditions. This single step usually results in enzyme which is 95% pure. The second step of the purification resulted in an overall purification of 56-fold for Yop51,6.2-foldforYop51*, and 5.3-foldforYop51*A162. Each purified recombinant protein appeared as a single Coomassie-Blue stained band on a SDS-PAGE gel with the predicted mass of 51 kDa for Yop51 and Yop51* and 33 kDa for Yop51*A162 (Fig. 2). The homogeneous Yop51 and Yop51* have a specific activity of 800 units/mg, and the purified RESULTS Yop51*A162 has a specific activity of 1200 units/mg at pH 5.5, I = 0.15 M, 30 "C whenp-NPP is used as a substrate. The Overexpression and Purification of the Yersiniu PTPaseThe coding sequences ofYop51,Yop51*, and Yop51*A162 higher specific activity of Yop51*A162 relative to Yop51 and were cloned behind the bacteriophage T7 promoter in the Yop51* can be accounted for by the lower molecular weight plasmid pT7-7. Induction of these constructs by addition of of Yop51*A162 (33.5 kDa). In fact, Yop51*Al62 has thesame isopropyl P-D-thiogalactoside to the exponentially growing $.,values as those of Yop51 and Yop51* under identical cultures Escherichia coli (DE3) (Studier and Moffatt, 1986) conditions (Table 11). The purified enzyme could bestored at gavehighlevel of expression ofYop51* and Yop51*Al62, a protein concentration greater than 0.1 mg/ml in 100 mM while under the same conditions, Yop51 was expressed to a acetate, pH5.7,l mM EDTA, I = 0.15 M buffer at 4 "C without much lesser degree (Table I). Yop51* and Yop51*A162 were loss of activity over a period of several months. expressed at 10-fold greater concentration than Yop51 (Fig. In order to validate the enzyme preparation and toevaluate 2, lane 2). Almost all of the phosphatase activity resided in purity, Yop51,Yop51*, and Yop51*A162were subjected to the clear supernant of the bacterial cell lysate and homoge- amino acid sequencing. NH2-terminal sequencing of 19 resineous enzyme is obtained followingbatch ion exchange chro- dues of Yop51 and Yop51* and 13 residues of Yop51*Al62 matography and gel filtration (Fig. 2). Table I summarizes gave the anticipated sequence MNLSLSDLHRQVSRLVQQE the purification procedure for Yop51, Yop51* and for Yop51 and Yop51*, and MRERPHTSGHHGA for

Expression Characterization and

23762

of Yersinia PTPase

TABLE I1 Comparison of kinetic parameters of Yop51, Yop51*, and Yop51*A162 All measurements were made at 30 "C, using p-NPP as a substrate. Buffers used were as follows: pH 5.0 and 5.5, 100 mM acetate; pH 6.0, 50 mM succinate; pH 7.0, 50 mM 3,3-dimethylglutarate; pH 8.0, 100 mM glycinamide. In all the buffer systems, 1 mM EDTA was included and theionic strength was kept at 0.15 M, adjustedby NaCl. YOp61

pH

kcat S"

5.0 5.5 6.0 7.0 6.22 8.0

1235 f 36 683 k 20 281 f 7.1 42.4 k 1.2 k4.15 0.18

Yop51*

Km mM 2.90 f 0.27 2.15 f 0.17 2.45 f 0.19 2.31 f 0.08 f 0.14

K, mM 2.70 f 0.23 f 0.15

kcat

S-1

1230 f 33 7332.04 f 20 295 f 2.32 6.2 46.3 f 1.5 6.40 f 0.20

Yop51*A162. No detectable impurities were observed (data not shown). Monomeric Nature of the Yersinia PTPase-Size estimates of Yop51, Yop51*, and Yop51*A162 were performed by gel filtration on a Sephadex G-100 (2.5 X 108 cm) column. The column was calibrated with molecular mass standardsas outlined under "Experimental Procedures." Both Yop51 and Yop51* eluted just prior to the elution position of BSA corresponding to anapparent molecular mass of 74 kDa, whereas Yop51*A162 eluted as a 35-kDa protein (data not shown). Since Yop51 and Yop51* display identical chromatographic behavior both in CM Sephadex C25 and Sephadex G-100, and since both Yop51 and Yop51* show a higher molecular mass than the expected based upon the mass calculated from its primary structure, we further explored the oligomeric structure of Yop 51* using sedimentation equilibrium. The advantage of this method is that the system is being studied at equilibrium and thus there is no dependence on the shape of the molecule or the viscosity of the media. From measurements of the distribution of concentration, c, of the macromolecule asa function of distance, r, along the cell (i.e. distance from the axis of rotation), the M, can be calculated provided that we know u, the partial specific volume of the solute,and p , the density of the solution, M, = (2RT/(1 -up)w2) dlncld?, where R is the gas constant, T is the temperature, and w is the angular velocity. The value of (dm/ d?) is obtained from the slope of lnc against ?. The partial specific volume of the Yop51* (0.726cm3/g) is calculated from its amino acid composition (Cohn and Edsall, 1943) by using the known values of the molar volumes of each increment of structure of the various amino acids. At pH 7.0, 50 mM 3,3dimethylglutarate, 1 mM EDTA, I = 0.15 M, 20 "C, Yop51* hasa M, of 49,800 f 1700 in good agreement with the predicted monomer Mr. Our results demonstrate that Yop51* exists as aactive monomer in solution. Monomer and dimer equilibrium can also be probed through enzyme concentration dependence of enzyme activity (Shen et al., 1975; Zhang et al., 1991). Enzyme concentration dependence is best studied under conditions where the substrate, which serves as aprobe for the active enzyme, does not perturb any potential equilibria between different enzyme species. pNPP is ideally suited for these conditions because of its relatively high value of K, and large absorption change upon hydrolysis. The time course of the enzymatic reaction with [SI,