- Email: [email protected]
Heterologous expression and characterization of recombinant purple acid phosphatase from red kidney bean
ABB Archives of Biochemistry and Biophysics 401 (2002) 164–172 www.academicpress.com
Heterologous expression and characterization of recombinant purple acid phosphatase from red kidney beanq Andreas Vogel,a,b Torsten B€ orchers,c Katrin Marcus,d Helmut E. Meyer,d Bernt Krebs,b and Friedrich Spenera,* b
a Department of Biochemistry, University of M€unster, Wilhelm-Klemm-Strasse 2, 48149 M€unster, Germany Department of Inorganic Chemistry, Wilhelm-Klemm-Strasse 8, University of M€unster, 48149 M€unster, Germany c Institute for Chemical and Biochemical Sensor Research, Mendelstrasse 7, 48149 M€unster, Germany d Institute of Physiological Chemistry, University of Bochum, 44780 Bochum, Germany
Received 7 August 2001, and in revised form 5 March 2002
Abstract Purple acid phosphatases (PAPs) are dinuclear metallohydrolases of widespread occurrence. In a first step to understand structure–function relationship of PAP from red kidney bean (kbPAP), we cloned its cDNA and functionally expressed the enzyme in insect cells. kbPAP cDNA encodes a protein of 459 amino acids with 99% identity to the published primary structure (T. Klabunde et al., Eur. J. Biochem. 226 (1994) 369–375). N-terminally the cDNA encodes 27 amino acids with characteristics for a signal directing the nascent protein to the endoplasmic reticulum. A baculovirus vector was constructed containing cDNAs of kbPAP and green fluorescent protein, the latter to serve as transfection and infection marker. Heterologous expression in High Five insect cells afforded a dimeric, disulfide-linked phosphatase of 110 kDa, identical to the mass of native kbPAP. Purification in three steps yielded 1.5 mg recombinant protein per liter of culture medium with a specific activity of 266 units/mg, slightly exceeding that of native kbPAP. The recombinant protein was functionally indistinguishable from native kbPAP, despite differences in glycosylation and sensitivity to redox reagents. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Purple acid phosphatase; Plant; Tartrate-resistant acid phosphatase; Metalloenzyme; Heterologous expression; Baculovirus
Purple acid phosphatases (PAPs)1 are dinuclear metalloenzymes which have been isolated from mammalian, plant, and fungal sources. Their characteristic purple color is due to a charge transfer transition caused by a tyrosine residue coordinating a ferric ion (for re-
q The complete cDNA sequence of kbPAP has been deposited in the EMBL/GenBank/DDBJ nucleotide sequence database under the accession number AJ001270. Supplementary data for this article are available on IDEAL (www.idealibrary.com). * Corresponding author. Fax: +49-251-833-2132. E-mail address: [email protected] (F. Spener). 1 Abbreviations used: PAP, purple acid phosphatase; kbPAP, PAP from red kidney bean; spPAP, sweet potato PAP; TRAP, tartrateresistant acid phosphatase; RACE, random amplification of cDNA ends; GFP, green fluorescent protein; pNPP, para-nitrophenylphosphate; pNP, para-nitrophenolate; NTA, nitrilotriacetic acid; ECL, enhanced chemiluminescence; endo H, endoglycosidase H; PNGase F, peptide:N-glycosidase F; MALDI, matrix-assisted laser desorption/ ionization; PSD, postsource decay.
view, see [1,2]). Next to this chromophoric iron, the monomeric 36-kDa mammalian enzymes always contain a second iron ion, which can, in contrast to the former, easily change its redox state [3,4]. The di-iron PAPs exhibit phosphatase activity only in the mixed-valent Fe3þ –Fe2þ state, whereas they are inactive when oxidized to the Fe3þ –Fe3þ state [5,6]. In contrast to this, such a redox susceptibility is not found for the dimeric 110-kDa plant PAPs, due to a redox-stable Zn2þ ion next to the chromophoric iron, found in the enzymes isolated from red kidney bean (kbPAP), soybean and sweet potato (spPAP) [7–9]. Another spPAP isoenzyme was shown to contain an iron-manganese core, but its redox chemistry has not been investigated yet [8]. Interestingly, the zinc ion in kbPAP can be exchanged by iron to yield a redox-active enzyme with spectroscopic characteristics similar to those of the mammalian enzymes [10,11]. By the same token, redox-stable iron–zinc mammalian PAPs can be generated [4,12].
0003-9861/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 0 0 0 3 - 9 8 6 1 ( 0 2 ) 0 0 0 4 6 - 2
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
The first X-ray structure solved for this class of enzymes was that for kbPAP [13,14]; it revealed an oxobridged dinuclear core with the metals coordinated by a N/O donor set. After the first publications of crystal structures of mammalian PAPs [15–17] it became clear that in plant and mammalian PAPs the respective catalytic domains are formed by two sandwiched b-sheets surrounded by a-helices consisting of identical amino acid residues for metal coordination [2]. The striking difference between plant and mammalian PAPs is the additional N-terminal domain of the former, which has no contact with the active site and features a protein fold resembling that of a fibronectin type III domain [18]. PAPs display a broad substrate specificity, preferring activated phosphoric acid monoesters and anhydrides, like para-nitrophenylphosphate, a-naphthylphosphate, polyphosphates, ATP, ADP, and phosphorylated peptides/proteins [7,19]. In contrast to other acid phosphatases they are not inhibited by L (+)-tartrate; hence in the animal kingdom PAPs are also known as tartrate-resistant acid phosphatases (TRAP) [20]. PAPs hydrolyze phosphoesters by a direct transfer of a hydroxide to the phospho group by an in line SN 2-type displacement step [21]. The nature of the hydroxide is still under debate [17,22] and crystallographic evidence clarifying this item is still missing. It was proposed by Klabunde et al. [13] that, in addition to the metal ions, three histidine residues which do not coordinate the metals, play a role in phosphoesterhydrolysis for kbPAP. Only two of the histidines have their structural counterparts in mammalian PAPs [15]. With the ultimate aim to elucidate the reaction mechanism of plant PAPs by refined crystallographic and enzymatic studies with wild-type and mutant enzymes, we set out to clone the cDNA of kbPAP and to develop a heterologous expression system. To date, heterologous expression systems are available for mammalian PAPs only [23–25]. Here we report our studies to establish such a system for the more complex kbPAP and can confirm that the recombinant enzyme is functionally equivalent to the native enzyme.
Materials and methods Materials. Except otherwise stated, all proteins and fine chemicals were purchased from Sigma (Deisenhofen, Germany). Restriction enzymes were from New England Biolabs (Frankfurt, Germany), Roche (Basel, Switzerland), and Genecraft (M€ unster, Germany). Materials for chromatography were from AmershamPharmacia Biotech (Uppsala, Sweden). Oligonucleotides were synthesized by Roth (Karlsruhe, Germany) and Genset (Paris, France). DNA sequencing was performed by TOPLAB (Martinsried, Germany) and Seqlab (G€ ottingen, Germany).
165
RNA isolation and cDNA synthesis. RNA of freshly harvested wet red kidney bean seeds were isolated by a modification of the method used by Hall et al. [26]. Seeds (4 g) were ground in liquid N2 and after thawing mixed with 10 ml of boiling borate buffer (0.2 M borate, pH 9.0, 30 mM EGTA, 1% (w/v) SDS, 0.5% (v/v) bmercaptoethanol). After homogenization with an Ultraturrax 50 ll proteinase K solution (20 mg/ml) was added and incubated at 37 °C for 1 h. The suspension was centrifuged at 10,000g for 30 min and the supernatant was stored at )70 °C. An aliquot of this crude extract was further purified using columns supplied with the RNeasy RNA isolation kit (Qiagen, Hilden, Germany) according to the clean-up recommendations of the manufacturer. Poly(A) RNA was isolated with the oligotex mRNA kit (Qiagen). cDNA pools, primed with anchorðdTÞ18 (50 -AACTGGAAGAATTCGCGGCCGCAGGAAT18 30 ) or gene-specific primer kbPAP-400 (50 -CCTATGAG ACCAAATGTA-30 ), respectively, were generated from poly(A) RNA using superscript II reverse transcriptase (Life Technologies, Rockville, MD). Cloning of complete kbPAP cDNA. The anchorðdTÞ18 primed cDNA pool was used for amplification of a kbPAP cDNA fragment with degenerate primers kbPAP+580 (50 -TAYCARCCNTGGATHTGGAC-30 ) and kbPAP-920 (50 -ACRAACCANGCYTCRAAYTT30 ) which were derived from kbPAP amino acid sequence [27]. The PCR product was cloned into the pCR II vector (Invitrogen, Leek, The Netherlands) via T/A cloning and sequenced. Based on this sequence a gene-specific primer was designed (3RACEkbPAP, 50 -CACTTCATGGAAG GGGAA-30 ) and used for 30 -RACE-PCR [28], in conjunction with the anchor primer (50 -AACTGGAAGAAT TCGCGG-30 ). The coding region of the mature kbPAP cDNA was amplified with primers derived from the known N-terminal amino acid sequence of kbPAP (NTSma, 50 -AACCCGGGTTYGTGCGGAAAACAA AYAAG-30 ) and a primer containing the stop codon (CTBam 50 -TGGGATCCTTATGTGGAATCATCAA CTGG-30 ) introducing restriction sites (underlined) for subcloning. The PCR product was cloned into the pGEM-T easy vector (Promega, Heidelberg, Germany) resulting in pGEM-kbPAPF1 and sequenced. The 50 RACE was performed essentially as described by Frohman [28] with the following modification according to the manual of the 50 =30 -RACE kit by Roche. After removing excess primer of the kbPAP-400-primed cDNA pool by duplicate passage through a 2-ml Centricon-100 (Amicon, Witten, Germany), the single-stranded cDNA was tailed with terminal deoxynucleotidyl transferase (Roche) in the presence of dATP. This cDNA served as template for the first 50 -RACE-PCR with primer pair anchorðdTÞ18 and kbPAP-280 (50 -CTGATGGTGGTGTGATGAAT30 ). The PCR product obtained was diluted 200-fold and subjected to a second 50 -RACE-PCR using primer pair
166
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
anchor(dT18 ) and kbPAP-160 (50 -CCGGTTCATCCAT AGTCA-30 ). The PCR product was cloned by T/A cloning into the pGEM-T easy vector, resulting in pGEM5RACE, and sequenced. Generation of recombinant baculovirus. Recombinant baculoviruses were produced using the Bac-to-Bac expression system (Life Technologies), which is based on site-directed transposition in Escherichia coli, as described by Luckow et al. [29]. The complete coding region of kbPAP including native signal sequence was cloned into pFastBac Dual vector under the control of the polyhedrin promoter. In brief, PCR was carried out with primer pair EcokbPAPsig (50 -GTGAATTCAAAA TGGGTGTTGT-30 ) and CTBam using Pfu DNA polymerase (Stratagene) and a mixture of linearized pGEM-5RACE and pGEM-kbPAPF1 as template. The PCR product was cleaved with EcoRI and BamHI and ligated into pPICZ A-vector (Invitrogen), restricted with EcoRI and BsmBI, resulting in pPIC-kbPAP. From this vector the insert was released by restriction with EcoRI and NotI and ligated into pFastBac Dual, resulting in vector pFB-phkbPAP. Vector pEGFP-N1 (Clontech, Heidelberg, Germany) was used as a gene source for green fluorescent protein (GFP). The insert was liberated by digestion with BamHI and XbaI and ligated into pFB-phkbPAP digested with BbsI and NheI, resulting in pFB-p10GFP-phkbPAP. Recombinant bacmids were produced by site-specific transposition in E. coli DH10Bac according to the recommendations given in the Bac-to-Bac manual. Spodoptera frugiperda Sf9 cells (Invitrogen) were transfected with the purified bacmids using the transfection reagent Cellfectin (Life Technologies). The resulting recombinant baculoviruses were harvested from the culture medium 3 days after transfection. Cell culture and estimation of virus titer. Sf9 and High Five insect cells (Invitrogen) were cultured at 27 °C in Sf900 II medium (Life Technologies) containing 10 lg=ml gentamycin according to standard procedures [30]. Suspension cultures were maintained in Spinner flasks (Wheaton, Millville, NJ). The virulence of a given virus solution containing GFP cDNA was assessed by the content of green insect cells two days after infection. Different amounts of the virus were added to insect cells in a 24-well plate and the ratio of infected cells was estimated by visual inspection with a fluorescence microscope. An equivalent of the virus amount that yielded more than 70% infection in a 24-well plate was used for protein production in a Spinner flask. Protein determination and enzyme assay. Protein concentrations during purification were determined according to Bradford [31] using ovalbumin as standard. To quantify purified kbPAP an absorbance coefficient of 1% E280 nm ¼ 21:8 was used [11]. Activity of kbPAP was measured in 100 mM Mes, pH 6.0, 200 mM KCl, and with 50 mM para-nitrophenyl-
phosphate (pNPP) as substrate. For measurements of PAP activity in insect cell extracts 10 mM tartrate was added (TRAP activity). When recombinant enzyme was assayed, 10 mM ascorbate and 0.2 mM Fe(II) were added. The discontinuous assay was carried out in 96-well plates by mixing 5–20 ll of sample with 200 ll of substrate solution and stopping the reaction after 1–15 min with 50 ll of 0.5 M NaOH. Subsequently para-nitrophenolate (pNP) formed was measured at 405 nm using epNP ðpH > 10Þ ¼ 16; 600 M1 cm1 . One unit of PAP activity corresponds to 1 lmol of pNP liberated per minute at room temperature. The continuous assay was carried out in a spectrophotometer at 37 °C by measuring the absorbance at 405 nm over a period of 2 min after starting the reaction; the reaction rate was determined from initial slopes using epNP ðpH 6Þ ¼ 1954 M1 cm1 . One unit of PAP activity corresponds to 1 lmol of pNP liberated per min at 37 °C. kbPAP expression and purification. For heterologous protein production High Five cells in Spinner flasks were infected at a density of 1:6 106 cells/ml. Five days after infection, cells were removed by centrifugation at 2000g for 10 min and the supernatant was treated with bulk S-Sepharose material (20 ml/500 ml culture medium) overnight at 4 °C. After washing the cation exchanger with 20 mM Hepes, pH 7.5, 0.15 M NaCl, kbPAP was eluted by increasing the NaCl concentration up to 2 M in the same buffer. The buffer of the eluate was exchanged to 20 mM Mes, pH 6.0, 0.5 M NaCl, and loaded onto a Blue-Sepharose column (XK 16/20 filled with 20 ml chromatography material) equilibrated with the same buffer. A linear gradient from 0.5 to 1.25 M NaCl was applied to elute recombinant kbPAP. The enzymatically active fraction was then loaded onto a Mono Q HR5/5 column using a linear gradient from 0.2 to 0.6 M NaCl in 20 mM Bis–Tris, pH 6.5. Active fractions eluting from the column were immediately mixed with equal volumes of 20 mM Bis–Tris, pH 6.5, 1 M NaCl to maintain NaCl concentration above 0.5 M, which prevents denaturation of kbPAP. Generation of antibodies. Antibodies were raised in New Zealand rabbits by injection of purified native kbPAP using Freund’s adjuvant. Serum was collected and kbPAP-specific antibodies were purified following the method of Gu et al. [32]. In brief, the catalytic domain of kbPAP (corresponding to amino acid residues 123–423) was expressed as His-tag fusion protein using pET28-vector (Novagen) in E. coli Bl21(DE3). Inclusion bodies were purified over a Ni–NTA column. Recombinant protein, attached to the column, served as ligand for antibodies specific for nonglycosylated kbPAP. SDS–PAGE and Western blotting. SDS–PAGE was performed with 10 and 12% gels according to Laemmli [33]. Proteins were transferred onto nitrocellulose membrane (Serva, Heidelberg, Germany) by electroblotting in 10 mM Caps, pH 11.0, 20% methanol. The
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
membrane was incubated with affinity-purified rabbit anti-kbPAP antibodies (see below, 2 lg=ml) for 30 min. Staining was performed using the ECL system (Amersham-Pharmacia, Braunschweig, Germany) after incubation with protein A-horseradish peroxidase conjugate (Bio-Rad, Munich, Germany) diluted 1/2000. Endoglycosidase digestions. Protein (5 lg) was subjected to digestion with endoglycosidase H (endo H) and peptide:N-glycosidase F (PNGase F) (New England Biolabs), respectively, under denaturing conditions for 16 h overnight. Mass spectrometric analysis. Protein samples were digested with trypsin (Roche) in digestion buffer 1 (10 mM NH4 HCO3 , pH 7.8, 1 U trypsin/1 lg protein) and AspN (Roche) in digestion buffer 2 (50 mM sodium phosphate, pH 8.0, 1/20 enzyme/protein), respectively. After addition of respective protease, digestion was maintained for 10–12 h at 37 °C. Peptides were concentrated with Poros 10 R1 beads (Perseptive Biosystems, Cambridge, England). After evaporation of the solvent, peptide-containing Poros beads samples were suspended in 1 ll matrix solution (a-cyano-4-hydroxycinnamic acid (Sigma) in 0.1% trifluoroacetic acid/acetonitrile) and transferred onto the matrix-assisted laser desorption/ ionization (MALDI) target. The resulting peptides were analyzed using a Reflex III mass spectrometer (Bruker–Daltonik, Bremen, Germany) equipped with a Scout 384 ion source. For the analysis in the reflector mode the acceleration voltage was set to 20 kV, the reflection voltage to 21.6 kV, ion source acceleration to 13.5 kV, and the reflector–detector voltage to 1.6 kV. MALDI experiments in the linear mode were performed with an acceleration voltage of 20 kV, an ion source acceleration of 18.3 kV, and a linear detector voltage of 1.45 kV. Postsource decay (PSD) spectra were obtained with a parent ion selection of 40 Da. The reflection voltage was reduced in 14 steps from 21.6 to 0.65 kV. Data acquisition was performed on a SUN Ultra using the XACQ software, Version 4.0.2. Postanalysis was done using the XMASS software. The statistical error in the trypsin digest was 50 ppm using internal calibration and 100 ppm for the AspN digest with external calibration.
Results and discussion Cloning of complete kbPAP cDNA Guided by the amino acid sequence of kbPAP [27] a PCR product of the expected size could be obtained with a pair of degenerate PCR primers. The amino acid sequence inferred from this PCR product showed all but two identical amino acids to the published amino acid sequence. From this cDNA sequence, a gene-specific
167
primer was designed for 30 -RACE-PCR and the deduced amino acid sequence revealed one further difference. In order to get the complete cDNA-encoding mature kbPAP, a 50 -primer was designed on the basis of the Nterminal peptide of the enzyme. Two subsequent rounds of 50 -RACE-PCR with gene-specific primers kbPAP-280 and kbPAP-160 (Fig. 1) yielded the missing 50 -region including the start codon. Taken together, four independent PCRs from the kidney bean seed cDNA pool led to clones of which at least two in each case were sequenced in both directions; they covered the whole kbPAP cDNA sequence of 1624 bp length (Fig. 1). The putative start codon is the first ATG in the kbPAP reading frame at position 125. The stop codon TAA at position 1504 corresponds exactly to the C-terminus determined by Edman degradation of the protein [27] and MALDI-MS fingerprint analysis (see below). A polyadenylation signal AATAAA is found 46 bp downstream of the stop codon in the 123-bp-long 30 untranslated region. In addition to the coding region for the mature protein, a reading frame for a 27-amino-acid N-terminal signal peptide was found which has all characteristics of an endoplasmic reticulum leader peptide. The cleavage site predicted by von Heijne rules [34] most likely is at position Gly5 and thus does not correspond to the N-terminus found by amino acid sequencing of peptides obtained after bromocyano cleavage (Pheþ1 in Fig. 1). The three differences to the published amino acid sequence alluded to above are residues His253 ; Ile254 , and Asp342 for which codons for Tyr253 ; Ser254 , and Asn342 were found in the cDNA. A reexamination of the original amino acid sequence data and the electron density map of kbPAP indicates agreement with the present data that are based on cDNA cloning (T. Klabunde, personal communication). Expression of kbPAP in E. coli Several vector constructs were generated to express the complete coding region as well as the catalytic domain only of kbPAP in E. coli. All attempts led to inclusion bodies only and refolding failed (data not shown).2 Nevertheless, the inclusion bodies obtained were successfully used as standard for nonglycosylated kbPAP in electrophoretic separations and for antibody purification, which allowed for the first time the generation of monospecific antibodies and their purification from rabbit immunoserum. Heterologous expression of kbPAP in the yeast Pichia pastoris was also tested, but without success (data not shown)2.
2 Experimental conditions applied are available from the authors on request.
168
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
Fig. 1. Nucleotide and deduced amino acid sequence of full-length cDNA-encoding kbPAP. Potential signal sequence cleavage site predicted by von Heijne rules (.) or peptide analysis (O) are indicated. The putative start codon is marked in bold. The stop codon is indicated by an asterisk and polyadenylation signal is double underlined. Amino acids differing from sequence obtained by peptide analysis are underlined. Position of primers used for cloning of complete kbPAP cDNA are boxed and the orientation is indicated by an arrow. Positive amino acid numbering starts with the Nterminal residue (Phe1 ) found by peptide analysis.
Expression of kbPAP in High Five cells Infection of High Five cells with the baculovirus containing kbPAP cDNA resulted in TRAP activity in the culture medium, which reached maximum levels at day 5 postinfection. Coexpression of GFP as a proof for successful baculoviral transfection and infection process did not affect TRAP activity (data not shown). Western blot analysis revealed kbPAP protein in the medium, but not in the lysate of the insect cells (data not shown).
Under reducing conditions of SDS–PAGE analysis the protein migrated to about 55 kDa, under nonreducing conditions, to 110 kDa representative for disulfide bond formation between two identical monomeric units as observed for native kbPAP. All this indicated functionality of the native signal sequence for the secretion of the protein into the medium. Moreover, signal peptide cleavage and glycosylation (see below) are also in support of a secretory pathway of the recombinant protein as opposed to protein liberation by cell lysis.
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
169
Table 1 Purification of recombinant kbPAP from High Five insect cellsa Purification step
Total protein (mg)
Total activityb (units)
Sp. act. (units/mg)
Purification factor
Yield (%)
Medium (1 liter) S-Sepharose Blue-Sepharose Mono Q
660.0 44.7 5.0 1.5
955c 638 538 398
1.5 14.3 108.6 265.8
1 10 75 184
100 67 56 42
a
Representative example; final yield and specific activity were reproducible. Continuous assay with addition of tartrate, Fe2þ , and ascorbate as described under Materials and Methods. c Background activity of medium from wild-type baculovirus-infected cells was subtracted. b
Purification and characterization of recombinant kbPAP Purification. After SDS–PAGE analysis of protein from culture medium of kbPAP-transfected insect cells a Coomassie-stained band corresponding to kbPAP was not found. Recombinant kbPAP became visible, however, after the first purification step (S-Sepharose) and the specific activity increased about 10-fold (Table 1). Affinity chromatography on Blue-Sepharose, the main step in purification of native kbPAP [27], led to a fraction with the recombinant protein being the main band. The final anion-exchange chromatography step (Mono Q) yielded pure recombinant kbPAP by the criterion of SDS–PAGE. Although a strict comparison of the final purity of the enzyme preparations could not be made, it is worth mentioning that the specific activity of 266 units/mg of the recombinant enzyme exceeded that of native kbPAP (230 units/mg). Taken together, from 1 liter of culture medium 1.5 mg of pure recombinant kbPAP was obtained with an overall yield of 42% (Table 1). Primary structure analysis. After proteolytic cleavage of recombinant and native kbPAP with trypsin and AspN, respectively, several mass peaks could be assigned by MALDI-MS fingerprint analysis to the peptide backbone for both enzymes covering about 70% of the sequence, this confirmed the identity of the heterologously expressed protein (supplementary Tables 1 and 2 and supplementary Fig. 1). The identity of recombinant kbPAP was further supported by PSD analysis of selected peptides from the tryptic digest. Residual regions not unambiguously identified corresponded to glycopeptides or to high-mass fragments. Mass peaks corresponding to the predicted C-terminal peptide were detected, indicating no proteolytic processing at this region. In both samples peaks were not obtained that could be attributed to a peptide starting with Phe1 as N-terminal amino acid, but mass peaks corresponding to peptides Ser3 –Arg3 (observed mass 709.3 Da/calculated mass 709.4 Da) were detected after the tryptic digest. Moreover, after the tryptic digestion of recombinant kbPAP a mass peak could be assigned to peptide Gly5 –Lys7 (1365.5/1365.8) which would correspond to the cleavage site predicted by von Heijne rules. The detection of Pheþ1 as N-terminal amino acid
in the plant enzyme [27], obtained by sequencing peptides after BrCN cleavage, most likely resulted from unspecific proteolysis during enzyme preparation or BrCN cleavage. Interestingly, N-terminal sequencing of recombinant as well as of native protein by Edman degradation was not possible; blocking by acetylation or myristoylation could be an explanation, but corresponding mass peaks were not found by MALDI-MS. In conclusion, the MALDI-MS analysis does not support the N-terminal amino acid reported previously [27], rather indicates a signal peptide cleavage following von Heijne rules. Moreover, an N-terminal modification, yet to be defined, is very likely. Analysis of sugar composition. It was shown that native kbPAP contains five N-linked glycosylation sites at Asn 81, 109, 143, 211, and 396. The glycans have a complex-type xylose-containing structure [35]. The glyco content was reported to be 10% of the protein mass which was confirmed here by MALDI-MS after reduction of the protein samples with b-mercaptoethanol. The Table 2 Assignment of mass peaks obtained by MALDI-MS fingerprint analysis of glycopeptides derived from recombinant and native kbPAP by trypsin and AspN digestion Peptide
MHþ observed (Da)
MHþ calculated (Da)
Carbohydrate compositiona; b
Recombinant 72–94 79–93 79–93 79–93 79–93 109–112 141–149
kbPAP 3670.6 2703.3 2849.4 2865.4 3011.5 1853.9 2383.5
3670.7 2703.2 2849.3 2865.3 3011.3 1853.7 2383.9
GlcNAc2 Man3 GlcNAc2 Man2 Fuc GlcNAc2 Man2 Fuc2 GlcNAc2 Man3 Fuc GlcNAc2 Man3 Fuc2 GlcNAc2 Man5 Fuc GlcNAc2 Man4 Fuc2
Native kbPAP 79–93 2997.3 79–93 3200.3 109–112 1718.7 109–113 2386.1 393–408 3194.0
2997.3 3200.4 1718.7 2386.0 3194.4
GlcNAc2 Man3 FucXyl GlcNAc3 Man3 FucXyl GlcNAc3 Man3 Xyl GlcNAc4 Man3 FucXylGal GlcNAc2 Man3 FucXyl
a Derived from molecular mass with possible carbohydrate compositions GlcNAc2 Man2–9 Fuc0–2 and those detected by Stahl et al. [35] for native kbPAP. b GlcNAc, N-acetylglucosamine; Man, mannose; Fuc, fucose; Xyl, xylose; Gal, galactose.
170
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
Fig. 2. Digestion of kbPAP by endoglycosidases. Coomassie-stained SDS–PAGE (10%) under reducing conditions of kbPAP samples (5 lg). Recombinant and native kbPAP were analyzed after incubation with endoglycosidases under denaturing conditions. Recombinant kbPAP expressed as His-tag fusion protein and purified from E. coli inclusion bodies with a molecular mass of 52 kDa was applied for comparison.
latter analysis revealed glycosylation of the recombinant protein to a similar extent. Whereas the recombinant enzyme showed one peak with a molecular mass of 54,838 Da, the spectrum of native kbPAP showed two peaks with molecular masses of 54,442 and 55,847 Da (data not shown). The heterogeneous composition of the native enzyme became also apparent by the double band seen after SDS–PAGE (Fig. 2), and can be attributed to residue Asn81 , which is only partially glycosylated [35]. In order to compare the carbohydrate content of recombinant and native kbPAP, samples were analyzed by SDS–PAGE following digestion with endoglycosidases. A shift in molecular mass for recombinant kbPAP was observed after PNGase F digestion, which confirms glycosylation of recombinant kbPAP (Fig. 2). The band of the resulting protein migrated almost identical as nonglycosylated kbPAP, which had been heterologously expressed in E. coli as His-tag fusion protein. Digestion of recombinant kbPAP with endo H, which is specific for high-mannose glycans with more than six mannose residues, resulted in a small shift only, whereas digestion of native kbPAP had no effect on protein migration. However, PNGase F digestion of native kbPAP resulted in a slight shift in molecular mass. This is in agreement with the glycan structure analysis by Stahl et al. [35], which identified four of the five sugar side chains containing an a1-3 fucose attached to the asparagine-linked N-acetylglucosamine, preventing removal of these glycans by PNGase F. For three of the five potential glycosylation sites in recombinant kbPAP MALDI-MS fingerprints revealed glycan modification (Table 2). In agreement with the PNGase F experiment, the carbohydrates that could be assigned for the recombinant enzyme were all of the high mannose type with up to two additional fucose residues. Five different sugar compositions were derived for position Asn81 , indicating heterogeneous glycosylation at this site. Mass peaks for glycans linked to the native protein were not found and vice versa. As a proof for the quality of the glycan analysis carried out here, we could reproduce six glycan structures for the native protein which have been described earlier by Stahl et al. [35].
The endoglycosidase digest together with the fingerprint analysis indicated that the glycan structures of both enzymes are different and that the glycan structure of the recombinant enzyme is in agreement with that found for glycoproteins that have been heterologously expressed in High Five cells [36]. Enzymatic properties. PAPs with an iron–zinc center show no influence on mild reductants and oxidants, in contrast to di-iron PAPs [3,11]. The latter are activated by a mixture of Fe2þ and ascorbate and inactivated by H2 O2 . In agreement with that, native kbPAP with an iron–zinc center was not activated upon reduction of the enzyme with ferrous ions and ascorbate, but the activity of recombinant kbPAP was increased about fourfold (Table 3). Addition of H2 O2 decreased activity of either enzyme. This could indicate a partial replacement of the iron–zinc active site in the recombinant enzyme by a diiron core. The pH dependence of phosphatase activity was investigated for both enzymes, yielding an almost identical curve under nonreducing conditions with a pH optimum at pH 6.1. The pH optimum for recombinant kbPAP under reducing conditions was a rather sharp maximum at pH 6.2 (data not shown). It is known that tetraoxoanions inhibit PAPs because they mimic the substrate [13], but PAPs are insensitive to inhibition by tartrate, in contrast to other acid phosphatases [20]. EDTA is able to extract zinc from kbPAP [11], although an activation effect at low concentration
Table 3 Effects of redox reagents on phosphatase activity of kbPAPa Additive
Recombinant kbPAP (%)
Native kbPAP (%)
None Fe2þ =ascorbate (0.2/10 mM) H2 O2 (1 mM) H2 O2 (10 mM)
28 4 100 19 22 1 20 4
100 4 64 9 93 10 84 7
a
Discontinuous assay in 0.1 M Mes, pH 6.0, 0.2 M KCl, 50 mM pNPP, and redox reagents in concentrations as indicated. The highest activity obtained was set to 100% (2.3 for recombinant and 3.5 units/ml for native kbPAP). Mean values SD (N ¼ 3).
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
171
Fig. 3. Recombinant and native kbPAP show the same effects on inhibitors. Discontinuous assay for acid phosphatase activity. The reaction buffer was 0.1 M Mes, pH 6.0, 0.2 M KCl, the substrate 50 mM pNPP. In the case of recombinant enzyme Fe2þ (0.2 mM) and ascorbate (10 mM) were added. Inhibitors were added 10 min before starting the assay with substrate addition. 100% activity corresponds to 2.3 and 3.5 units/ml for recombinant and native enzyme, respectively. Values are means SD (N ¼ 3).
was reported [37]. The influence of these inhibitors on activity of recombinant and native kbPAP was investigated and shown to be similar for both enzymes (Fig. 3). In conclusion, expression of kbPAP in baculovirusinfected insect cells has been shown to be the system of choice, as expression in either E. coli or P. pastoris did not result in functional enzyme. The enzyme characteristics of the recombinant protein were comparable to those of the native enzyme, despite differences in the glycan structure and reactivity toward redox agents. Future studies aimed at the elucidation of mechanism and structure–function relationship can now be based on mutant enzymes as well. Moreover, this first description of functional heterologous expression of a plant PAP offers opportunities for other enzymes of this family.
Acknowledgments This work was financially supported by the EC biotechnology program and the Deutsche Forschungsgemeinschaft (SFB 424). This work is part of the Ph.D. thesis of A.V. at the University of M€ unster.
References [1] T. Klabunde, B. Krebs, Struct. Bond. 89 (1997) 177–198. [2] A. Vogel, F. Spener, B. Krebs, in: A. Messerschmidt, R. Huber, T. Poulos, K. Wieghardt (Eds.), Handbook of Metalloproteins, vol. 2, Wiley, Chichester, 2001, pp. 752–767.
[3] P.G. Debrunner, M.P. Hendrich, J. DeJersey, D.T. Keough, J.T. Sage, B. Zerner, Biochim. Biophys. Acta 745 (1983) 103–106. [4] J.C. Davis, B.A. Averill, Proc. Natl. Acad. Sci. USA 79 (1982) 4623–4627. [5] M. Merkx, B.A. Averill, Biochemistry 37 (1998) 11223–11231. [6] M. Dietrich, D. M€ unstermann, H. Suerbaum, H. Witzel, Eur. J. Biochem. 199 (1991) 105–113. [7] J.L. Beck, L.A. McConachie, A.C. Summors, W.N. Arnold, J. de Jersey, B. Zerner, Biochim. Biophys. Acta 869 (1986) 61–68. [8] G. Schenk, Y. Ge, L.E. Carrington, C.J. Wynne, I.R. Searle, B.J. Carroll, S. Hamilton, J. de Jersey, Arch. Biochem. Biophys. 370 (1999) 183–189. [9] A. Durmus, C. Eicken, B.H. Sift, A. Kratel, R. Kappl, J. H€ uttermann, B. Krebs, Eur. J. Biochem. 260 (1999) 709–716. [10] J.L. Beck, M.J. McArthur, J. de Jersey, B. Zerner, Inorg. Chim. Acta 153 (1988) 39–44. [11] H. Suerbaum, M. K€ orner, H. Witzel, E. Althaus, B.D. Mosel, W. M€ uller-Warmuth, Eur. J. Biochem. 214 (1993) 313–321. [12] D.T. Keough, D.A. Dionysius, J. de Jersey, B. Zerner, Biochem. Biophys. Res. Commun. 94 (1980) 600–605. [13] T. Klabunde, N. Str€ater, R. Fr€ ohlich, H. Witzel, B. Krebs, J. Mol. Biol. 259 (1996) 737–748. [14] N. Str€ater, T. Klabunde, P. Tucker, H. Witzel, B. Krebs, Science 268 (1995) 1489–1492. [15] J. Uppenberg, F. Lindqvist, C. Svensson, B. Ek-Rylander, G. Andersson, J. Mol. Biol. 290 (1999) 201–211. [16] Y. Lindqvist, E. Johansson, H. Kaija, P. Vihko, G. Schneider, J. Mol. Biol. 291 (1999) 135–147. [17] L.W. Guddat, A.S. McAlpine, D. Hume, S. Hamilton, J. de Jersey, J.L. Martin, Struct. Fold. Des. 7 (1999) 757–767. [18] I. Tsyguelnaia, R.F. Doolittle, J. Mol. Evol. 46 (1998) 612–614. [19] J. Ljusberg, B. Ek-Rylander, G. Andersson, Biochem. J. 343 (1999) 63–69. [20] J.B. Vincent, B.A. Averill, FASEB J. 4 (1990) 3009–3014. [21] E.G. Mueller, M.W. Crowder, B.A. Averill, J.R. Knowles, J. Am. Chem. Soc. 115 (1993) 2974–2975. [22] M. Merkx, B.A. Averill, J. Am. Chem. Soc. 121 (1999) 6683– 6689.
172
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
[23] A.R. Hayman, T.M. Cox, J. Biol. Chem. 269 (1994) 1294– 1300. [24] K. Marshall, K. Nash, G. Haussman, I. Cassady, D. Hume, J. de Jersey, S. Hamilton, Arch. Biochem. Biophys. 345 (1997) 230– 236. [25] B. Ek-Rylander, T. Barkhem, J. Ljusberg, L. Ohman, K.K. Andersson, G. Andersson, Biochem. J. 321 (1997) 305–311. [26] T.C. Hall, Y. Ma, B.U. Buchbinder, J.W. Pyne, S.M. Sun, F.A. Bliss, Proc. Natl. Acad. Sci. USA 75 (1978) 3196–3200. [27] T. Klabunde, B. Stahl, H. Suerbaum, S. Hahner, M. Karas, F. Hillenkamp, B. Krebs, H. Witzel, Eur. J. Biochem. 226 (1994) 369–375. [28] M.A. Frohman, Methods Enzymol. 218 (1993) 340–356. [29] V.A. Luckow, S.C. Lee, G.F. Barry, P.O. Olins, J. Virol. 67 (1993) 4566–4579.
[30] D.R. O’Reilly, L.K. Miller, V.A. Luckow, Baculovirus Expression Vectors. A Laboratory Manual, Oxford University Press, New York/Oxford, 1994. [31] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254. [32] J. Gu, C.G. Stephenson, M.J. Iadarola, BioTechniques 17 (1994) 257–262. [33] U.K. Laemmli, Nature 227 (1970) 680–685. [34] G. von Heijne, J. Membr. Biol. 115 (1990) 195–201. [35] B. Stahl, T. Klabunde, H. Witzel, B. Krebs, M. Steup, M. Karas, F. Hillenkamp, Eur. J. Biochem. 220 (1994) 321–330. [36] P.M. Rudd, A.K. Downing, M. Cadene, D.J. Harvey, M.R. Wormald, I. Weir, R.A. Dwek, D.B. Rifkin, P.E. Gleizes, Biochemistry 39 (2000) 1596–1603. [37] A.G. Cashikar, R. Kumaresan, N.M. Rao, Plant Physiol. 114 (1997) 907–915.
Heterologous expression and characterization of recombinant purple acid phosphatase from red kidney beanq Andreas Vogel,a,b Torsten B€ orchers,c Katrin Marcus,d Helmut E. Meyer,d Bernt Krebs,b and Friedrich Spenera,* b
a Department of Biochemistry, University of M€unster, Wilhelm-Klemm-Strasse 2, 48149 M€unster, Germany Department of Inorganic Chemistry, Wilhelm-Klemm-Strasse 8, University of M€unster, 48149 M€unster, Germany c Institute for Chemical and Biochemical Sensor Research, Mendelstrasse 7, 48149 M€unster, Germany d Institute of Physiological Chemistry, University of Bochum, 44780 Bochum, Germany
Received 7 August 2001, and in revised form 5 March 2002
Abstract Purple acid phosphatases (PAPs) are dinuclear metallohydrolases of widespread occurrence. In a first step to understand structure–function relationship of PAP from red kidney bean (kbPAP), we cloned its cDNA and functionally expressed the enzyme in insect cells. kbPAP cDNA encodes a protein of 459 amino acids with 99% identity to the published primary structure (T. Klabunde et al., Eur. J. Biochem. 226 (1994) 369–375). N-terminally the cDNA encodes 27 amino acids with characteristics for a signal directing the nascent protein to the endoplasmic reticulum. A baculovirus vector was constructed containing cDNAs of kbPAP and green fluorescent protein, the latter to serve as transfection and infection marker. Heterologous expression in High Five insect cells afforded a dimeric, disulfide-linked phosphatase of 110 kDa, identical to the mass of native kbPAP. Purification in three steps yielded 1.5 mg recombinant protein per liter of culture medium with a specific activity of 266 units/mg, slightly exceeding that of native kbPAP. The recombinant protein was functionally indistinguishable from native kbPAP, despite differences in glycosylation and sensitivity to redox reagents. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Purple acid phosphatase; Plant; Tartrate-resistant acid phosphatase; Metalloenzyme; Heterologous expression; Baculovirus
Purple acid phosphatases (PAPs)1 are dinuclear metalloenzymes which have been isolated from mammalian, plant, and fungal sources. Their characteristic purple color is due to a charge transfer transition caused by a tyrosine residue coordinating a ferric ion (for re-
q The complete cDNA sequence of kbPAP has been deposited in the EMBL/GenBank/DDBJ nucleotide sequence database under the accession number AJ001270. Supplementary data for this article are available on IDEAL (www.idealibrary.com). * Corresponding author. Fax: +49-251-833-2132. E-mail address: [email protected] (F. Spener). 1 Abbreviations used: PAP, purple acid phosphatase; kbPAP, PAP from red kidney bean; spPAP, sweet potato PAP; TRAP, tartrateresistant acid phosphatase; RACE, random amplification of cDNA ends; GFP, green fluorescent protein; pNPP, para-nitrophenylphosphate; pNP, para-nitrophenolate; NTA, nitrilotriacetic acid; ECL, enhanced chemiluminescence; endo H, endoglycosidase H; PNGase F, peptide:N-glycosidase F; MALDI, matrix-assisted laser desorption/ ionization; PSD, postsource decay.
view, see [1,2]). Next to this chromophoric iron, the monomeric 36-kDa mammalian enzymes always contain a second iron ion, which can, in contrast to the former, easily change its redox state [3,4]. The di-iron PAPs exhibit phosphatase activity only in the mixed-valent Fe3þ –Fe2þ state, whereas they are inactive when oxidized to the Fe3þ –Fe3þ state [5,6]. In contrast to this, such a redox susceptibility is not found for the dimeric 110-kDa plant PAPs, due to a redox-stable Zn2þ ion next to the chromophoric iron, found in the enzymes isolated from red kidney bean (kbPAP), soybean and sweet potato (spPAP) [7–9]. Another spPAP isoenzyme was shown to contain an iron-manganese core, but its redox chemistry has not been investigated yet [8]. Interestingly, the zinc ion in kbPAP can be exchanged by iron to yield a redox-active enzyme with spectroscopic characteristics similar to those of the mammalian enzymes [10,11]. By the same token, redox-stable iron–zinc mammalian PAPs can be generated [4,12].
0003-9861/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 0 0 0 3 - 9 8 6 1 ( 0 2 ) 0 0 0 4 6 - 2
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
The first X-ray structure solved for this class of enzymes was that for kbPAP [13,14]; it revealed an oxobridged dinuclear core with the metals coordinated by a N/O donor set. After the first publications of crystal structures of mammalian PAPs [15–17] it became clear that in plant and mammalian PAPs the respective catalytic domains are formed by two sandwiched b-sheets surrounded by a-helices consisting of identical amino acid residues for metal coordination [2]. The striking difference between plant and mammalian PAPs is the additional N-terminal domain of the former, which has no contact with the active site and features a protein fold resembling that of a fibronectin type III domain [18]. PAPs display a broad substrate specificity, preferring activated phosphoric acid monoesters and anhydrides, like para-nitrophenylphosphate, a-naphthylphosphate, polyphosphates, ATP, ADP, and phosphorylated peptides/proteins [7,19]. In contrast to other acid phosphatases they are not inhibited by L (+)-tartrate; hence in the animal kingdom PAPs are also known as tartrate-resistant acid phosphatases (TRAP) [20]. PAPs hydrolyze phosphoesters by a direct transfer of a hydroxide to the phospho group by an in line SN 2-type displacement step [21]. The nature of the hydroxide is still under debate [17,22] and crystallographic evidence clarifying this item is still missing. It was proposed by Klabunde et al. [13] that, in addition to the metal ions, three histidine residues which do not coordinate the metals, play a role in phosphoesterhydrolysis for kbPAP. Only two of the histidines have their structural counterparts in mammalian PAPs [15]. With the ultimate aim to elucidate the reaction mechanism of plant PAPs by refined crystallographic and enzymatic studies with wild-type and mutant enzymes, we set out to clone the cDNA of kbPAP and to develop a heterologous expression system. To date, heterologous expression systems are available for mammalian PAPs only [23–25]. Here we report our studies to establish such a system for the more complex kbPAP and can confirm that the recombinant enzyme is functionally equivalent to the native enzyme.
Materials and methods Materials. Except otherwise stated, all proteins and fine chemicals were purchased from Sigma (Deisenhofen, Germany). Restriction enzymes were from New England Biolabs (Frankfurt, Germany), Roche (Basel, Switzerland), and Genecraft (M€ unster, Germany). Materials for chromatography were from AmershamPharmacia Biotech (Uppsala, Sweden). Oligonucleotides were synthesized by Roth (Karlsruhe, Germany) and Genset (Paris, France). DNA sequencing was performed by TOPLAB (Martinsried, Germany) and Seqlab (G€ ottingen, Germany).
165
RNA isolation and cDNA synthesis. RNA of freshly harvested wet red kidney bean seeds were isolated by a modification of the method used by Hall et al. [26]. Seeds (4 g) were ground in liquid N2 and after thawing mixed with 10 ml of boiling borate buffer (0.2 M borate, pH 9.0, 30 mM EGTA, 1% (w/v) SDS, 0.5% (v/v) bmercaptoethanol). After homogenization with an Ultraturrax 50 ll proteinase K solution (20 mg/ml) was added and incubated at 37 °C for 1 h. The suspension was centrifuged at 10,000g for 30 min and the supernatant was stored at )70 °C. An aliquot of this crude extract was further purified using columns supplied with the RNeasy RNA isolation kit (Qiagen, Hilden, Germany) according to the clean-up recommendations of the manufacturer. Poly(A) RNA was isolated with the oligotex mRNA kit (Qiagen). cDNA pools, primed with anchorðdTÞ18 (50 -AACTGGAAGAATTCGCGGCCGCAGGAAT18 30 ) or gene-specific primer kbPAP-400 (50 -CCTATGAG ACCAAATGTA-30 ), respectively, were generated from poly(A) RNA using superscript II reverse transcriptase (Life Technologies, Rockville, MD). Cloning of complete kbPAP cDNA. The anchorðdTÞ18 primed cDNA pool was used for amplification of a kbPAP cDNA fragment with degenerate primers kbPAP+580 (50 -TAYCARCCNTGGATHTGGAC-30 ) and kbPAP-920 (50 -ACRAACCANGCYTCRAAYTT30 ) which were derived from kbPAP amino acid sequence [27]. The PCR product was cloned into the pCR II vector (Invitrogen, Leek, The Netherlands) via T/A cloning and sequenced. Based on this sequence a gene-specific primer was designed (3RACEkbPAP, 50 -CACTTCATGGAAG GGGAA-30 ) and used for 30 -RACE-PCR [28], in conjunction with the anchor primer (50 -AACTGGAAGAAT TCGCGG-30 ). The coding region of the mature kbPAP cDNA was amplified with primers derived from the known N-terminal amino acid sequence of kbPAP (NTSma, 50 -AACCCGGGTTYGTGCGGAAAACAA AYAAG-30 ) and a primer containing the stop codon (CTBam 50 -TGGGATCCTTATGTGGAATCATCAA CTGG-30 ) introducing restriction sites (underlined) for subcloning. The PCR product was cloned into the pGEM-T easy vector (Promega, Heidelberg, Germany) resulting in pGEM-kbPAPF1 and sequenced. The 50 RACE was performed essentially as described by Frohman [28] with the following modification according to the manual of the 50 =30 -RACE kit by Roche. After removing excess primer of the kbPAP-400-primed cDNA pool by duplicate passage through a 2-ml Centricon-100 (Amicon, Witten, Germany), the single-stranded cDNA was tailed with terminal deoxynucleotidyl transferase (Roche) in the presence of dATP. This cDNA served as template for the first 50 -RACE-PCR with primer pair anchorðdTÞ18 and kbPAP-280 (50 -CTGATGGTGGTGTGATGAAT30 ). The PCR product obtained was diluted 200-fold and subjected to a second 50 -RACE-PCR using primer pair
166
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
anchor(dT18 ) and kbPAP-160 (50 -CCGGTTCATCCAT AGTCA-30 ). The PCR product was cloned by T/A cloning into the pGEM-T easy vector, resulting in pGEM5RACE, and sequenced. Generation of recombinant baculovirus. Recombinant baculoviruses were produced using the Bac-to-Bac expression system (Life Technologies), which is based on site-directed transposition in Escherichia coli, as described by Luckow et al. [29]. The complete coding region of kbPAP including native signal sequence was cloned into pFastBac Dual vector under the control of the polyhedrin promoter. In brief, PCR was carried out with primer pair EcokbPAPsig (50 -GTGAATTCAAAA TGGGTGTTGT-30 ) and CTBam using Pfu DNA polymerase (Stratagene) and a mixture of linearized pGEM-5RACE and pGEM-kbPAPF1 as template. The PCR product was cleaved with EcoRI and BamHI and ligated into pPICZ A-vector (Invitrogen), restricted with EcoRI and BsmBI, resulting in pPIC-kbPAP. From this vector the insert was released by restriction with EcoRI and NotI and ligated into pFastBac Dual, resulting in vector pFB-phkbPAP. Vector pEGFP-N1 (Clontech, Heidelberg, Germany) was used as a gene source for green fluorescent protein (GFP). The insert was liberated by digestion with BamHI and XbaI and ligated into pFB-phkbPAP digested with BbsI and NheI, resulting in pFB-p10GFP-phkbPAP. Recombinant bacmids were produced by site-specific transposition in E. coli DH10Bac according to the recommendations given in the Bac-to-Bac manual. Spodoptera frugiperda Sf9 cells (Invitrogen) were transfected with the purified bacmids using the transfection reagent Cellfectin (Life Technologies). The resulting recombinant baculoviruses were harvested from the culture medium 3 days after transfection. Cell culture and estimation of virus titer. Sf9 and High Five insect cells (Invitrogen) were cultured at 27 °C in Sf900 II medium (Life Technologies) containing 10 lg=ml gentamycin according to standard procedures [30]. Suspension cultures were maintained in Spinner flasks (Wheaton, Millville, NJ). The virulence of a given virus solution containing GFP cDNA was assessed by the content of green insect cells two days after infection. Different amounts of the virus were added to insect cells in a 24-well plate and the ratio of infected cells was estimated by visual inspection with a fluorescence microscope. An equivalent of the virus amount that yielded more than 70% infection in a 24-well plate was used for protein production in a Spinner flask. Protein determination and enzyme assay. Protein concentrations during purification were determined according to Bradford [31] using ovalbumin as standard. To quantify purified kbPAP an absorbance coefficient of 1% E280 nm ¼ 21:8 was used [11]. Activity of kbPAP was measured in 100 mM Mes, pH 6.0, 200 mM KCl, and with 50 mM para-nitrophenyl-
phosphate (pNPP) as substrate. For measurements of PAP activity in insect cell extracts 10 mM tartrate was added (TRAP activity). When recombinant enzyme was assayed, 10 mM ascorbate and 0.2 mM Fe(II) were added. The discontinuous assay was carried out in 96-well plates by mixing 5–20 ll of sample with 200 ll of substrate solution and stopping the reaction after 1–15 min with 50 ll of 0.5 M NaOH. Subsequently para-nitrophenolate (pNP) formed was measured at 405 nm using epNP ðpH > 10Þ ¼ 16; 600 M1 cm1 . One unit of PAP activity corresponds to 1 lmol of pNP liberated per minute at room temperature. The continuous assay was carried out in a spectrophotometer at 37 °C by measuring the absorbance at 405 nm over a period of 2 min after starting the reaction; the reaction rate was determined from initial slopes using epNP ðpH 6Þ ¼ 1954 M1 cm1 . One unit of PAP activity corresponds to 1 lmol of pNP liberated per min at 37 °C. kbPAP expression and purification. For heterologous protein production High Five cells in Spinner flasks were infected at a density of 1:6 106 cells/ml. Five days after infection, cells were removed by centrifugation at 2000g for 10 min and the supernatant was treated with bulk S-Sepharose material (20 ml/500 ml culture medium) overnight at 4 °C. After washing the cation exchanger with 20 mM Hepes, pH 7.5, 0.15 M NaCl, kbPAP was eluted by increasing the NaCl concentration up to 2 M in the same buffer. The buffer of the eluate was exchanged to 20 mM Mes, pH 6.0, 0.5 M NaCl, and loaded onto a Blue-Sepharose column (XK 16/20 filled with 20 ml chromatography material) equilibrated with the same buffer. A linear gradient from 0.5 to 1.25 M NaCl was applied to elute recombinant kbPAP. The enzymatically active fraction was then loaded onto a Mono Q HR5/5 column using a linear gradient from 0.2 to 0.6 M NaCl in 20 mM Bis–Tris, pH 6.5. Active fractions eluting from the column were immediately mixed with equal volumes of 20 mM Bis–Tris, pH 6.5, 1 M NaCl to maintain NaCl concentration above 0.5 M, which prevents denaturation of kbPAP. Generation of antibodies. Antibodies were raised in New Zealand rabbits by injection of purified native kbPAP using Freund’s adjuvant. Serum was collected and kbPAP-specific antibodies were purified following the method of Gu et al. [32]. In brief, the catalytic domain of kbPAP (corresponding to amino acid residues 123–423) was expressed as His-tag fusion protein using pET28-vector (Novagen) in E. coli Bl21(DE3). Inclusion bodies were purified over a Ni–NTA column. Recombinant protein, attached to the column, served as ligand for antibodies specific for nonglycosylated kbPAP. SDS–PAGE and Western blotting. SDS–PAGE was performed with 10 and 12% gels according to Laemmli [33]. Proteins were transferred onto nitrocellulose membrane (Serva, Heidelberg, Germany) by electroblotting in 10 mM Caps, pH 11.0, 20% methanol. The
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
membrane was incubated with affinity-purified rabbit anti-kbPAP antibodies (see below, 2 lg=ml) for 30 min. Staining was performed using the ECL system (Amersham-Pharmacia, Braunschweig, Germany) after incubation with protein A-horseradish peroxidase conjugate (Bio-Rad, Munich, Germany) diluted 1/2000. Endoglycosidase digestions. Protein (5 lg) was subjected to digestion with endoglycosidase H (endo H) and peptide:N-glycosidase F (PNGase F) (New England Biolabs), respectively, under denaturing conditions for 16 h overnight. Mass spectrometric analysis. Protein samples were digested with trypsin (Roche) in digestion buffer 1 (10 mM NH4 HCO3 , pH 7.8, 1 U trypsin/1 lg protein) and AspN (Roche) in digestion buffer 2 (50 mM sodium phosphate, pH 8.0, 1/20 enzyme/protein), respectively. After addition of respective protease, digestion was maintained for 10–12 h at 37 °C. Peptides were concentrated with Poros 10 R1 beads (Perseptive Biosystems, Cambridge, England). After evaporation of the solvent, peptide-containing Poros beads samples were suspended in 1 ll matrix solution (a-cyano-4-hydroxycinnamic acid (Sigma) in 0.1% trifluoroacetic acid/acetonitrile) and transferred onto the matrix-assisted laser desorption/ ionization (MALDI) target. The resulting peptides were analyzed using a Reflex III mass spectrometer (Bruker–Daltonik, Bremen, Germany) equipped with a Scout 384 ion source. For the analysis in the reflector mode the acceleration voltage was set to 20 kV, the reflection voltage to 21.6 kV, ion source acceleration to 13.5 kV, and the reflector–detector voltage to 1.6 kV. MALDI experiments in the linear mode were performed with an acceleration voltage of 20 kV, an ion source acceleration of 18.3 kV, and a linear detector voltage of 1.45 kV. Postsource decay (PSD) spectra were obtained with a parent ion selection of 40 Da. The reflection voltage was reduced in 14 steps from 21.6 to 0.65 kV. Data acquisition was performed on a SUN Ultra using the XACQ software, Version 4.0.2. Postanalysis was done using the XMASS software. The statistical error in the trypsin digest was 50 ppm using internal calibration and 100 ppm for the AspN digest with external calibration.
Results and discussion Cloning of complete kbPAP cDNA Guided by the amino acid sequence of kbPAP [27] a PCR product of the expected size could be obtained with a pair of degenerate PCR primers. The amino acid sequence inferred from this PCR product showed all but two identical amino acids to the published amino acid sequence. From this cDNA sequence, a gene-specific
167
primer was designed for 30 -RACE-PCR and the deduced amino acid sequence revealed one further difference. In order to get the complete cDNA-encoding mature kbPAP, a 50 -primer was designed on the basis of the Nterminal peptide of the enzyme. Two subsequent rounds of 50 -RACE-PCR with gene-specific primers kbPAP-280 and kbPAP-160 (Fig. 1) yielded the missing 50 -region including the start codon. Taken together, four independent PCRs from the kidney bean seed cDNA pool led to clones of which at least two in each case were sequenced in both directions; they covered the whole kbPAP cDNA sequence of 1624 bp length (Fig. 1). The putative start codon is the first ATG in the kbPAP reading frame at position 125. The stop codon TAA at position 1504 corresponds exactly to the C-terminus determined by Edman degradation of the protein [27] and MALDI-MS fingerprint analysis (see below). A polyadenylation signal AATAAA is found 46 bp downstream of the stop codon in the 123-bp-long 30 untranslated region. In addition to the coding region for the mature protein, a reading frame for a 27-amino-acid N-terminal signal peptide was found which has all characteristics of an endoplasmic reticulum leader peptide. The cleavage site predicted by von Heijne rules [34] most likely is at position Gly5 and thus does not correspond to the N-terminus found by amino acid sequencing of peptides obtained after bromocyano cleavage (Pheþ1 in Fig. 1). The three differences to the published amino acid sequence alluded to above are residues His253 ; Ile254 , and Asp342 for which codons for Tyr253 ; Ser254 , and Asn342 were found in the cDNA. A reexamination of the original amino acid sequence data and the electron density map of kbPAP indicates agreement with the present data that are based on cDNA cloning (T. Klabunde, personal communication). Expression of kbPAP in E. coli Several vector constructs were generated to express the complete coding region as well as the catalytic domain only of kbPAP in E. coli. All attempts led to inclusion bodies only and refolding failed (data not shown).2 Nevertheless, the inclusion bodies obtained were successfully used as standard for nonglycosylated kbPAP in electrophoretic separations and for antibody purification, which allowed for the first time the generation of monospecific antibodies and their purification from rabbit immunoserum. Heterologous expression of kbPAP in the yeast Pichia pastoris was also tested, but without success (data not shown)2.
2 Experimental conditions applied are available from the authors on request.
168
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
Fig. 1. Nucleotide and deduced amino acid sequence of full-length cDNA-encoding kbPAP. Potential signal sequence cleavage site predicted by von Heijne rules (.) or peptide analysis (O) are indicated. The putative start codon is marked in bold. The stop codon is indicated by an asterisk and polyadenylation signal is double underlined. Amino acids differing from sequence obtained by peptide analysis are underlined. Position of primers used for cloning of complete kbPAP cDNA are boxed and the orientation is indicated by an arrow. Positive amino acid numbering starts with the Nterminal residue (Phe1 ) found by peptide analysis.
Expression of kbPAP in High Five cells Infection of High Five cells with the baculovirus containing kbPAP cDNA resulted in TRAP activity in the culture medium, which reached maximum levels at day 5 postinfection. Coexpression of GFP as a proof for successful baculoviral transfection and infection process did not affect TRAP activity (data not shown). Western blot analysis revealed kbPAP protein in the medium, but not in the lysate of the insect cells (data not shown).
Under reducing conditions of SDS–PAGE analysis the protein migrated to about 55 kDa, under nonreducing conditions, to 110 kDa representative for disulfide bond formation between two identical monomeric units as observed for native kbPAP. All this indicated functionality of the native signal sequence for the secretion of the protein into the medium. Moreover, signal peptide cleavage and glycosylation (see below) are also in support of a secretory pathway of the recombinant protein as opposed to protein liberation by cell lysis.
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
169
Table 1 Purification of recombinant kbPAP from High Five insect cellsa Purification step
Total protein (mg)
Total activityb (units)
Sp. act. (units/mg)
Purification factor
Yield (%)
Medium (1 liter) S-Sepharose Blue-Sepharose Mono Q
660.0 44.7 5.0 1.5
955c 638 538 398
1.5 14.3 108.6 265.8
1 10 75 184
100 67 56 42
a
Representative example; final yield and specific activity were reproducible. Continuous assay with addition of tartrate, Fe2þ , and ascorbate as described under Materials and Methods. c Background activity of medium from wild-type baculovirus-infected cells was subtracted. b
Purification and characterization of recombinant kbPAP Purification. After SDS–PAGE analysis of protein from culture medium of kbPAP-transfected insect cells a Coomassie-stained band corresponding to kbPAP was not found. Recombinant kbPAP became visible, however, after the first purification step (S-Sepharose) and the specific activity increased about 10-fold (Table 1). Affinity chromatography on Blue-Sepharose, the main step in purification of native kbPAP [27], led to a fraction with the recombinant protein being the main band. The final anion-exchange chromatography step (Mono Q) yielded pure recombinant kbPAP by the criterion of SDS–PAGE. Although a strict comparison of the final purity of the enzyme preparations could not be made, it is worth mentioning that the specific activity of 266 units/mg of the recombinant enzyme exceeded that of native kbPAP (230 units/mg). Taken together, from 1 liter of culture medium 1.5 mg of pure recombinant kbPAP was obtained with an overall yield of 42% (Table 1). Primary structure analysis. After proteolytic cleavage of recombinant and native kbPAP with trypsin and AspN, respectively, several mass peaks could be assigned by MALDI-MS fingerprint analysis to the peptide backbone for both enzymes covering about 70% of the sequence, this confirmed the identity of the heterologously expressed protein (supplementary Tables 1 and 2 and supplementary Fig. 1). The identity of recombinant kbPAP was further supported by PSD analysis of selected peptides from the tryptic digest. Residual regions not unambiguously identified corresponded to glycopeptides or to high-mass fragments. Mass peaks corresponding to the predicted C-terminal peptide were detected, indicating no proteolytic processing at this region. In both samples peaks were not obtained that could be attributed to a peptide starting with Phe1 as N-terminal amino acid, but mass peaks corresponding to peptides Ser3 –Arg3 (observed mass 709.3 Da/calculated mass 709.4 Da) were detected after the tryptic digest. Moreover, after the tryptic digestion of recombinant kbPAP a mass peak could be assigned to peptide Gly5 –Lys7 (1365.5/1365.8) which would correspond to the cleavage site predicted by von Heijne rules. The detection of Pheþ1 as N-terminal amino acid
in the plant enzyme [27], obtained by sequencing peptides after BrCN cleavage, most likely resulted from unspecific proteolysis during enzyme preparation or BrCN cleavage. Interestingly, N-terminal sequencing of recombinant as well as of native protein by Edman degradation was not possible; blocking by acetylation or myristoylation could be an explanation, but corresponding mass peaks were not found by MALDI-MS. In conclusion, the MALDI-MS analysis does not support the N-terminal amino acid reported previously [27], rather indicates a signal peptide cleavage following von Heijne rules. Moreover, an N-terminal modification, yet to be defined, is very likely. Analysis of sugar composition. It was shown that native kbPAP contains five N-linked glycosylation sites at Asn 81, 109, 143, 211, and 396. The glycans have a complex-type xylose-containing structure [35]. The glyco content was reported to be 10% of the protein mass which was confirmed here by MALDI-MS after reduction of the protein samples with b-mercaptoethanol. The Table 2 Assignment of mass peaks obtained by MALDI-MS fingerprint analysis of glycopeptides derived from recombinant and native kbPAP by trypsin and AspN digestion Peptide
MHþ observed (Da)
MHþ calculated (Da)
Carbohydrate compositiona; b
Recombinant 72–94 79–93 79–93 79–93 79–93 109–112 141–149
kbPAP 3670.6 2703.3 2849.4 2865.4 3011.5 1853.9 2383.5
3670.7 2703.2 2849.3 2865.3 3011.3 1853.7 2383.9
GlcNAc2 Man3 GlcNAc2 Man2 Fuc GlcNAc2 Man2 Fuc2 GlcNAc2 Man3 Fuc GlcNAc2 Man3 Fuc2 GlcNAc2 Man5 Fuc GlcNAc2 Man4 Fuc2
Native kbPAP 79–93 2997.3 79–93 3200.3 109–112 1718.7 109–113 2386.1 393–408 3194.0
2997.3 3200.4 1718.7 2386.0 3194.4
GlcNAc2 Man3 FucXyl GlcNAc3 Man3 FucXyl GlcNAc3 Man3 Xyl GlcNAc4 Man3 FucXylGal GlcNAc2 Man3 FucXyl
a Derived from molecular mass with possible carbohydrate compositions GlcNAc2 Man2–9 Fuc0–2 and those detected by Stahl et al. [35] for native kbPAP. b GlcNAc, N-acetylglucosamine; Man, mannose; Fuc, fucose; Xyl, xylose; Gal, galactose.
170
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
Fig. 2. Digestion of kbPAP by endoglycosidases. Coomassie-stained SDS–PAGE (10%) under reducing conditions of kbPAP samples (5 lg). Recombinant and native kbPAP were analyzed after incubation with endoglycosidases under denaturing conditions. Recombinant kbPAP expressed as His-tag fusion protein and purified from E. coli inclusion bodies with a molecular mass of 52 kDa was applied for comparison.
latter analysis revealed glycosylation of the recombinant protein to a similar extent. Whereas the recombinant enzyme showed one peak with a molecular mass of 54,838 Da, the spectrum of native kbPAP showed two peaks with molecular masses of 54,442 and 55,847 Da (data not shown). The heterogeneous composition of the native enzyme became also apparent by the double band seen after SDS–PAGE (Fig. 2), and can be attributed to residue Asn81 , which is only partially glycosylated [35]. In order to compare the carbohydrate content of recombinant and native kbPAP, samples were analyzed by SDS–PAGE following digestion with endoglycosidases. A shift in molecular mass for recombinant kbPAP was observed after PNGase F digestion, which confirms glycosylation of recombinant kbPAP (Fig. 2). The band of the resulting protein migrated almost identical as nonglycosylated kbPAP, which had been heterologously expressed in E. coli as His-tag fusion protein. Digestion of recombinant kbPAP with endo H, which is specific for high-mannose glycans with more than six mannose residues, resulted in a small shift only, whereas digestion of native kbPAP had no effect on protein migration. However, PNGase F digestion of native kbPAP resulted in a slight shift in molecular mass. This is in agreement with the glycan structure analysis by Stahl et al. [35], which identified four of the five sugar side chains containing an a1-3 fucose attached to the asparagine-linked N-acetylglucosamine, preventing removal of these glycans by PNGase F. For three of the five potential glycosylation sites in recombinant kbPAP MALDI-MS fingerprints revealed glycan modification (Table 2). In agreement with the PNGase F experiment, the carbohydrates that could be assigned for the recombinant enzyme were all of the high mannose type with up to two additional fucose residues. Five different sugar compositions were derived for position Asn81 , indicating heterogeneous glycosylation at this site. Mass peaks for glycans linked to the native protein were not found and vice versa. As a proof for the quality of the glycan analysis carried out here, we could reproduce six glycan structures for the native protein which have been described earlier by Stahl et al. [35].
The endoglycosidase digest together with the fingerprint analysis indicated that the glycan structures of both enzymes are different and that the glycan structure of the recombinant enzyme is in agreement with that found for glycoproteins that have been heterologously expressed in High Five cells [36]. Enzymatic properties. PAPs with an iron–zinc center show no influence on mild reductants and oxidants, in contrast to di-iron PAPs [3,11]. The latter are activated by a mixture of Fe2þ and ascorbate and inactivated by H2 O2 . In agreement with that, native kbPAP with an iron–zinc center was not activated upon reduction of the enzyme with ferrous ions and ascorbate, but the activity of recombinant kbPAP was increased about fourfold (Table 3). Addition of H2 O2 decreased activity of either enzyme. This could indicate a partial replacement of the iron–zinc active site in the recombinant enzyme by a diiron core. The pH dependence of phosphatase activity was investigated for both enzymes, yielding an almost identical curve under nonreducing conditions with a pH optimum at pH 6.1. The pH optimum for recombinant kbPAP under reducing conditions was a rather sharp maximum at pH 6.2 (data not shown). It is known that tetraoxoanions inhibit PAPs because they mimic the substrate [13], but PAPs are insensitive to inhibition by tartrate, in contrast to other acid phosphatases [20]. EDTA is able to extract zinc from kbPAP [11], although an activation effect at low concentration
Table 3 Effects of redox reagents on phosphatase activity of kbPAPa Additive
Recombinant kbPAP (%)
Native kbPAP (%)
None Fe2þ =ascorbate (0.2/10 mM) H2 O2 (1 mM) H2 O2 (10 mM)
28 4 100 19 22 1 20 4
100 4 64 9 93 10 84 7
a
Discontinuous assay in 0.1 M Mes, pH 6.0, 0.2 M KCl, 50 mM pNPP, and redox reagents in concentrations as indicated. The highest activity obtained was set to 100% (2.3 for recombinant and 3.5 units/ml for native kbPAP). Mean values SD (N ¼ 3).
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
171
Fig. 3. Recombinant and native kbPAP show the same effects on inhibitors. Discontinuous assay for acid phosphatase activity. The reaction buffer was 0.1 M Mes, pH 6.0, 0.2 M KCl, the substrate 50 mM pNPP. In the case of recombinant enzyme Fe2þ (0.2 mM) and ascorbate (10 mM) were added. Inhibitors were added 10 min before starting the assay with substrate addition. 100% activity corresponds to 2.3 and 3.5 units/ml for recombinant and native enzyme, respectively. Values are means SD (N ¼ 3).
was reported [37]. The influence of these inhibitors on activity of recombinant and native kbPAP was investigated and shown to be similar for both enzymes (Fig. 3). In conclusion, expression of kbPAP in baculovirusinfected insect cells has been shown to be the system of choice, as expression in either E. coli or P. pastoris did not result in functional enzyme. The enzyme characteristics of the recombinant protein were comparable to those of the native enzyme, despite differences in the glycan structure and reactivity toward redox agents. Future studies aimed at the elucidation of mechanism and structure–function relationship can now be based on mutant enzymes as well. Moreover, this first description of functional heterologous expression of a plant PAP offers opportunities for other enzymes of this family.
Acknowledgments This work was financially supported by the EC biotechnology program and the Deutsche Forschungsgemeinschaft (SFB 424). This work is part of the Ph.D. thesis of A.V. at the University of M€ unster.
References [1] T. Klabunde, B. Krebs, Struct. Bond. 89 (1997) 177–198. [2] A. Vogel, F. Spener, B. Krebs, in: A. Messerschmidt, R. Huber, T. Poulos, K. Wieghardt (Eds.), Handbook of Metalloproteins, vol. 2, Wiley, Chichester, 2001, pp. 752–767.
[3] P.G. Debrunner, M.P. Hendrich, J. DeJersey, D.T. Keough, J.T. Sage, B. Zerner, Biochim. Biophys. Acta 745 (1983) 103–106. [4] J.C. Davis, B.A. Averill, Proc. Natl. Acad. Sci. USA 79 (1982) 4623–4627. [5] M. Merkx, B.A. Averill, Biochemistry 37 (1998) 11223–11231. [6] M. Dietrich, D. M€ unstermann, H. Suerbaum, H. Witzel, Eur. J. Biochem. 199 (1991) 105–113. [7] J.L. Beck, L.A. McConachie, A.C. Summors, W.N. Arnold, J. de Jersey, B. Zerner, Biochim. Biophys. Acta 869 (1986) 61–68. [8] G. Schenk, Y. Ge, L.E. Carrington, C.J. Wynne, I.R. Searle, B.J. Carroll, S. Hamilton, J. de Jersey, Arch. Biochem. Biophys. 370 (1999) 183–189. [9] A. Durmus, C. Eicken, B.H. Sift, A. Kratel, R. Kappl, J. H€ uttermann, B. Krebs, Eur. J. Biochem. 260 (1999) 709–716. [10] J.L. Beck, M.J. McArthur, J. de Jersey, B. Zerner, Inorg. Chim. Acta 153 (1988) 39–44. [11] H. Suerbaum, M. K€ orner, H. Witzel, E. Althaus, B.D. Mosel, W. M€ uller-Warmuth, Eur. J. Biochem. 214 (1993) 313–321. [12] D.T. Keough, D.A. Dionysius, J. de Jersey, B. Zerner, Biochem. Biophys. Res. Commun. 94 (1980) 600–605. [13] T. Klabunde, N. Str€ater, R. Fr€ ohlich, H. Witzel, B. Krebs, J. Mol. Biol. 259 (1996) 737–748. [14] N. Str€ater, T. Klabunde, P. Tucker, H. Witzel, B. Krebs, Science 268 (1995) 1489–1492. [15] J. Uppenberg, F. Lindqvist, C. Svensson, B. Ek-Rylander, G. Andersson, J. Mol. Biol. 290 (1999) 201–211. [16] Y. Lindqvist, E. Johansson, H. Kaija, P. Vihko, G. Schneider, J. Mol. Biol. 291 (1999) 135–147. [17] L.W. Guddat, A.S. McAlpine, D. Hume, S. Hamilton, J. de Jersey, J.L. Martin, Struct. Fold. Des. 7 (1999) 757–767. [18] I. Tsyguelnaia, R.F. Doolittle, J. Mol. Evol. 46 (1998) 612–614. [19] J. Ljusberg, B. Ek-Rylander, G. Andersson, Biochem. J. 343 (1999) 63–69. [20] J.B. Vincent, B.A. Averill, FASEB J. 4 (1990) 3009–3014. [21] E.G. Mueller, M.W. Crowder, B.A. Averill, J.R. Knowles, J. Am. Chem. Soc. 115 (1993) 2974–2975. [22] M. Merkx, B.A. Averill, J. Am. Chem. Soc. 121 (1999) 6683– 6689.
172
A. Vogel et al. / Archives of Biochemistry and Biophysics 401 (2002) 164–172
[23] A.R. Hayman, T.M. Cox, J. Biol. Chem. 269 (1994) 1294– 1300. [24] K. Marshall, K. Nash, G. Haussman, I. Cassady, D. Hume, J. de Jersey, S. Hamilton, Arch. Biochem. Biophys. 345 (1997) 230– 236. [25] B. Ek-Rylander, T. Barkhem, J. Ljusberg, L. Ohman, K.K. Andersson, G. Andersson, Biochem. J. 321 (1997) 305–311. [26] T.C. Hall, Y. Ma, B.U. Buchbinder, J.W. Pyne, S.M. Sun, F.A. Bliss, Proc. Natl. Acad. Sci. USA 75 (1978) 3196–3200. [27] T. Klabunde, B. Stahl, H. Suerbaum, S. Hahner, M. Karas, F. Hillenkamp, B. Krebs, H. Witzel, Eur. J. Biochem. 226 (1994) 369–375. [28] M.A. Frohman, Methods Enzymol. 218 (1993) 340–356. [29] V.A. Luckow, S.C. Lee, G.F. Barry, P.O. Olins, J. Virol. 67 (1993) 4566–4579.
[30] D.R. O’Reilly, L.K. Miller, V.A. Luckow, Baculovirus Expression Vectors. A Laboratory Manual, Oxford University Press, New York/Oxford, 1994. [31] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254. [32] J. Gu, C.G. Stephenson, M.J. Iadarola, BioTechniques 17 (1994) 257–262. [33] U.K. Laemmli, Nature 227 (1970) 680–685. [34] G. von Heijne, J. Membr. Biol. 115 (1990) 195–201. [35] B. Stahl, T. Klabunde, H. Witzel, B. Krebs, M. Steup, M. Karas, F. Hillenkamp, Eur. J. Biochem. 220 (1994) 321–330. [36] P.M. Rudd, A.K. Downing, M. Cadene, D.J. Harvey, M.R. Wormald, I. Weir, R.A. Dwek, D.B. Rifkin, P.E. Gleizes, Biochemistry 39 (2000) 1596–1603. [37] A.G. Cashikar, R. Kumaresan, N.M. Rao, Plant Physiol. 114 (1997) 907–915.