- Email: [email protected]
Synthesis of [14C]pyrroloquinoline quinone (PQQ) in E. coli using genes for PQQ synthesis from K. pneumoniae
Biochimica et Biophysica Acta 1524 (2000) 247^252
www.elsevier.com/locate/bba
Synthesis of [14 C]pyrroloquinoline quinone (PQQ) in E. coli using genes for PQQ synthesis from K. pneumoniae Tracy E. Stites, Tracey R. Sih, Robert B. Rucker * Department of Nutrition, One Shields Avenue, University of California, Davis, Davis, CA 95616, USA Received 1 August 2000; accepted 19 October 2000
Abstract Radiochemical forms of pyrroloquinoline quinone (PQQ) are of utility in studies to determine the metabolic role and fate of PQQ in biological systems. Accordingly, we have synthesized [14 C]PQQ using a tyrosine auxotrophic strain of Escherichia coli (AT2471). A construct containing the six genes required for PQQ synthesis from Klebsiella pneumoniae was used to transform the auxotrophic strain of E. coli. E. coli were then grown in minimal M9 medium containing 3.7U109 Bq/mmol [14 C]tyrosine. At confluence, the medium was collected and applied to a DEAE A-25 anionic exchange column; [14 C]PQQ was eluted using a KCl gradient (0^2 M in 0.1 M potassium phosphate buffer, pH 7.0). Radioactivity co-eluting as PQQ was next pooled, acidified and passed through a C-18 column; [14 C]PQQ was eluted with a phosphate buffer (0.1 M, pH 7.0). Reverse phase HPLC (C-18) using either the ion-pairing agent trifluoroacetic acid (0.1%) and an acetonitrile gradient or phosphoric acid and a methanol gradient were used to isolate [14 C]PQQ. Fractions were collected and analyzed by liquid scintillation counting. 14 C-labelled compounds isolated from the medium eluted corresponding to the elution of various tyrosinederived products or PQQ. The radioactive compound corresponding to PQQ was also reacted with acetone to form 5-acetonyl-PQQ, which co-eluted with a 5-acetonyl-PQQ standard, as a validation of [14 C]PQQ synthesis. The specific activity of synthesized [14 C]PQQ was 3.7U109 Bq/mmol [14 C]PQQ, equal to that of [U-14 C]tyrosine initially added to the medium. ß 2000 Elsevier Science B.V. All rights reserved. Keywords : Pyrroloquinoline quinone; Tyrosine auxotroph; Imidazolopyrroloquinoline; Klebsiella pneumoniae
1. Introduction From studies over the past 20 years, it has become clear that Gram-negative bacteria contain a novel class of dehydrogenase enzymes, which utilize pyrroloquinoline quinone (PQQ) as a cofactor [1^3]. Many bacteria that require PQQ as a cofactor also synthesize PQQ [4,5]. However, certain organisms, e.g. Escherichia coli and Salmonella typhimurium, do not synthesize PQQ even though they contain enzymes, such as glucose dehydrogenase, that require PQQ [6]. The complete cassette of genes encoding proteins for the synthesis of PQQ has been partially characterized in several bacteria. The number of genes required for bacterial synthesis of PQQ ranges from as few as four genes in
* Corresponding author. Fax: +1-530-752-8966; E-mail : [email protected]
Acinetobacter calcoaceticus [7] to seven genes in Methylobacterium extorquens [8]. In A. calcoaceticus, three of the four genes required for PQQ synthesis encode proteins with molecular weights of 29 700, 10 800, and 43 600 Da, and these proteins are involved in enzymatic processes and also the transport of PQQ into the periplasmic space [9]. The fourth gene that is required for PQQ biosynthesis encodes a polypeptide of 24 amino acids. Klebsiella pneumoniae [10], M. extorquens, [11] and Pseudomonas £uorescens [12] also contain a similar polypeptide that is required for PQQ synthesis. A conserved motif, Glu-Val-Thr-X-Tyr, is contained in all bacterial forms of this peptide. Tyrosine and glutamate contained in the short polypeptide are used as substrates for PQQ biosynthesis (Fig. 1). In the current study, the synthesis of [14 C]PQQ in E. coli (AT2471) was accomplished by utilizing a plasmid containing the complete cassette of PQQ synthesis genes (six genes) from K. pneumoniae. Since tyrosine is a substrate for PQQ synthesis (Fig. 1), by providing radiolabeled tyrosine, and preventing the endogenous synthesis of tyro-
0304-4165 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 0 0 ) 0 0 1 6 6 - 5
BBAGEN 25121 30-11-00
248
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
sine by using an E. coli tyrosine auxotroph, the resultant PQQ should incorporate the radiolabel without dilution of speci¢c activity. This approach was used in preparing [14 C]PQQ with a relatively high speci¢c activity. 2. Materials and methods 2.1. Chemicals and reagents All chemicals were obtained from Fisher Scienti¢c (Pittsburgh, PA) or Sigma Chemical Corporation (St. Louis, MO) unless stated and were of the highest purity available. PQQ was a gift from Dr. James Mah at the University of Southern California School of Dentistry, Los Angeles, CA. Reagents required for the growth and maintenance of bacteria were obtained from Difco Laboratories (Detroit, MI). E. coli strain AT2471 was obtained from the E. coli Genetic Stock Center located in the department of biology at Yale University in New Haven, CT. The E. coli bacterial strain JA221, containing pBCP165, was a gift from Dr. Pieter Postma at E.C. Slater Institute for Biochemical Research, and Biotechnology Center, University of Amsterdam, Amsterdam, The Netherlands [10]. The plasmid pgp492 containing the genes for the synthesis of a glucose dehydrogenase enzyme (sGDH) from A. calcoaceticus was a gift from Dr. Nora Goosen (Department of Molecular Genetics, University of Leiden, The Netherlands) [13]. 2.2. Transforming bacteria For the isolation of plasmid pBCP165 from the E. coli strain JA221, the bacteria were grown to con£uence then lysed and the plasmid was extracted and puri¢ed by a phenol:chloroform protocol [14]. The puri¢ed plasmid was restricted with BamHI and XhoI (Roche, Indianapolis, IN) and compared to the known restriction digest map of pBCP165. As a control, the genes for PQQ synthesis were removed from pBCP165 to regenerate the original cloning vector pCB104 [15]. Following agarose gel electrophoresis (1% agarose), the band corresponding to pCB104 was isolated (Qiagen, Valencia, CA), and the vector was then religated using standard procedures [14]. 2.3. [14 C]PQQ synthesis An E. coli tyrosine auxotroph (AT2471) was used to synthesize PQQ speci¢cally labeled with [U-14 C]tyrosine (Fig. 1). AT2471 was made electrocompetent, and the bacteria were transformed with either pBCP165 or the vector pCB104 using a standard electroporation technique [14]. The chloramphenicol antibiotic resistance gene present in both plasmids was used for selectivity of plasmid incorporation. Bacteria were grown initially in LB (1% tryptone, 0.5% yeast extract and 1% NaCl) or in minimal medium
(M9) with supplemented amino acids and thiamine as required (AT2471 requires thiamine). For [14 C]PQQ synthesis AT2471 containing plasmid pBCP165 was plated on agar plates containing M9 minimal medium with 100 Wg/ml chloramphenicol, 100 WM tyrosine and 1 mM thiamine. An isolated colony was harvested and added to 5 ml of M9 minimal sterile medium containing 10 WM tyrosine (estimated to be the minimal requirement), 1 mM thiamine, and 100 Wg/ml chloramphenicol. The bacteria were grown for 24 h at 37³C in a shaking incubator (250 rpm). A 2 ml aliquot was then transferred to 500 ml of M9 minimal sterile medium containing 1 mM thiamine, 100 Wg/ml chloramphenicol and 3.7U109 Bq/mmol [U-14 C-tyrosine] (3.86 Wmol tyrosine) plus 1.14 Wmol [U-14 C-tyrosine], speci¢c activity, 1.62U1010 Bq (438 mCi)/mmol. The bacteria were grown to con£uence (72 h), and were then pelleted by centrifugation at 1000Ug. The medium (supernatant fraction) was saved for isolation of [14 C]PQQ. 2.4. PQQ puri¢cation The medium was applied to a DEAE Sephadex A-25 (1U3 cm) column at a £ow rate of 1 ml/min. The column was washed with 0.1 M sodium phosphate (pH 7.0) until the eluent radioactivity was at background. PQQ was eluted from the column using a KCl gradient (0^2 M) in 0.1 M sodium phosphate (pH 7.0). All fractions from the DEAE A-25 column were analyzed by the redox cycling assay as described by Gallop and colleagues [16,17]. The fractions containing redox cycling activity corresponding to the known elution pro¢le for authentic PQQ (0.6^0.8 M KCl) were pooled for further puri¢cation. The pH of the pooled fractions was adjusted to pH 1.5 with HCl, and applied to a C-18 low pressure Bond Elut column (Varian, Walnut Creek, CA). PQQ was eluted with a 0.1 M potassium phosphate bu¡er (pH 7.0) and saved for HPLC analysis. Reverse phase HPLC was performed using a HewlettPackard series 1100 HPLC system (Agilent Technologies, Palo Alto, CA), with an in-series diode array and £uorescence detector. Absorbance was monitored at 249, 280, and 422 nm and £uorescence at ex = 360 and em = 480 nm. A Phenomenex 5 W Prodigy (250 cmU4.6 mm I.D., 10 nm pore size) C-18 column was used for analytical separation. Fractions (0.33 ml) were collected and counted by a liquid scintillation counter (Wallac, Gaithersburg, MD) to determine 14 C radioactivity. Two separate HPLC protocols were utilized. The initial mobile phase for the ¢rst HPLC protocol consisted of 5% acetonitrile and 95% water (containing 0.1% tri£uoroacetic acid (v/v)). A linear gradient was applied from 5 to 30 min with a ¢nal concentration of 75% acetonitrile and 25% water (containing 0.1% tri£uoroacetic acid (v/v)). The initial mobile phase for the second HPLC protocol consisted of 30% methanol and 70% 0.06 M phosphoric acid. A
BBAGEN 25121 30-11-00
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
249
glucose dehydrogenase enzyme assay [19]. Soluble glucose dehydrogenase enzyme (sGDH, soluble type; EC 1.1.99.17) was produced in E. coli transformed with pgp492, a plasmid containing the GDH synthesis genes from A. calcoaceticus. To isolate GDH, transformed E. coli were grown to con£uence and centrifuged at 1000Ug. The bacteria were lysed twice in a French press (16000 psi) and were then treated with DNase (1 mg DNase/g bacteria (2000 units DNase/mg)) for 48 h at 20³C. The bacterial debris was pelleted by centrifugation at 48 000Ug and the supernatant fraction was applied to a CM-Sepharose cation exchange column (1.5U20 cm). The sGDH enzyme was eluted from the column with a KCl gradient (0^0.6 M). Fig. 1. The synthesis of PQQ from tyrosine and glutamic acid. [U14 C]Tyrosine was added to cultures of a tyrosine auxotrophic strain of E. coli (AT2471) to synthesize [14 C]PQQ labeled in nine of the PQQ's 14 carbons.
linear gradient was applied from 5 to 30 min with a ¢nal concentration of 70% methanol and 30% 0.06 M phosphoric acid.
3. Results 3.1. Synthesis of PQQ In preliminary studies, the bacterial strain AT2471 transformed with either pBCP165, containing the PQQ
2.5. PQQ derivatization PQQ was derivatized with acetone to form the acetone adduct (5-acetonyl-PQQ) to aid in identity and validation [18]. PQQ (200 nmol, 0.2 ml) was derivatized in 0.1 M sodium carbonate (pH 9.2, 0.1 ml) with the addition of 16% acetone (v/v, 0.1 ml) at 37³C for 30 min. Samples containing radioactivity were derivatized using the same procedure. 2.6. PQQ analysis The PQQ concentration of samples was determined by a
Fig. 2. Tyrosine-derived products. Radiochemically labeled tyrosine-derived products from the tyrosine auxotrophic strain AT2471 transformed with pBCP165 from K. pneumoniae. AT2471 was grown in minimal M9 medium containing 3.7U109 Bq/mmol tyrosine. The bacteria were pelleted and the supernatant fraction was applied to a DEAE-A25 (1U3 cm) column. PQQ was eluted from the column with a 0^2 M gradient of KCl. Fractions (5 ml) were collected and analyzed by liquid scintillation counting, and by a redox cycling assay [17]. The peak of redox cycling activity corresponding to a PQQ standard was pooled and used for further analysis.
Fig. 3. Chromatography of PQQ-related products. Reverse phase HPLC analysis of tyrosine, PQQ, and IPQ standards. The initial mobile phase consisted of 5% acetonitrile and 95% water (containing 0.1% tri£uoroacetic acid (v/v)). A linear gradient was applied from 5 to 30 min with a ¢nal concentration of 75% acetonitrile and 25% water (containing 0.1% tri£uoroacetic acid (v/v)). Peak 1, 1 mM tyrosine (280 nm); peak 2, 0.1 mM PQQ (ex = 360 nm and em = 480 nm); peak 3, 0.4 mM IPQ (422 nm). Also indicated is an HPLC analysis of 14 C-labeled products synthesized by the bacterial strain AT2471. Medium from AT2471 was applied to a DEAE-A25 column and [14 C]PQQ was eluted with a KCl gradient. A redox cycling peak corresponding to a PQQ standard was collected, and further puri¢ed by addition to a low pressure C-18 column. [14 C]PQQ was eluted with 0.1 M phosphate bu¡er (pH 7), and this fraction was analyzed by reverse phase HPLC. HPLC fractions (20 s, 0.33 ml) were collected and radioactivity was determined by liquid scintillation counting.
BBAGEN 25121 30-11-00
250
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
synthesis genes, or pCB104, vector alone, were analyzed for PQQ production. The PQQ concentration in the medium (LB), when AT2471 was transformed with pBCP165, was 165 nmol/l and was below detection limits (0.5 nmol/l) when AT2471 was grown with vector pCB104. For the synthesis of [14 C]PQQ, AT2471 was grown to con£uence in the presence of 3.7U109 Bq/mmol [U-14 C]tyrosine. The bacterial medium was applied to a DEAE A-25 column and PQQ was eluted with a KCl gradient (Fig. 2). Fractions with a KCl concentration range of 0.6^0.8 M KCl contained detectable redox cycling activity and were pooled for reverse phase HPLC analysis (Figs. 3 and 4).
Fig. 5. CM-Sepharose chromatography of the soluble glucose dehydrogenase enzyme (sGDH). E. coli transformed with pgp492 containing the synthesis genes for sGDH from A. calcoaceticus was used as a source of the enzyme. Cells were grown to con£uence, lysed with a French press (two times) and pelleted to remove bacterial debris. The suspension was treated for 2 days with DNase and centrifuged at 48 000Ug. Fractions were then applied to a CM-Sepharose cationic exchange column. A linear gradient (0^0.6 M KCl) was applied from fraction number 30 to 80. The majority ( s 95%) of the protein applied to the column eluted in fractions 1^29. Enzyme activity eluted in the area indicated by the line. A typical standard curve is shown in the inset. The enzyme was assayed as described by Olsthoorn and Duine [19] using phenazine methosulfate as the primary electron acceptor, dichlorophenolindophenol as the secondary electron acceptor and colorimetric agent. Glucose was used as the substrate and the detection range was from 0.05 to 0.5 pmol of PQQ.
3.2. Puri¢cation and analysis of [14 C]PQQ synthesized by AT2471
Fig. 4. Chromatography of tyrosine, PQQ, 5-acetonyl-PQQ (APQQ) and IPQ. The initial mobile phase consisted of 30% methanol and 70% 0.06 M phosphoric acid. A linear gradient was applied from 5 to 30 min with a ¢nal concentration of 70% methanol and 30% 0.06 M phosphoric acid. Peak 1, 1 mM tyrosine (280 nm); peak 2; 0.05 mM 5-acetonyl-PQQ (ex = 360 nm and em = 480 nm); peak 3, 0.1 mM PQQ (ex = 360 nm and em = 480 nm); peak 4, 0.4 mM IPQ (422 nm). Also indicated is the reverse phase HPLC analysis of acetone derivatized [14 C]PQQ synthesized by the bacterial strain AT2471. Medium from AT2471 was applied to a DEAE-A25 column and [14 C]PQQ was eluted with a KCl gradient. A redox cycling peak corresponding to a PQQ standard was collected, and further puri¢ed by addition to a low pressure C-18 column. [14 C]PQQ was eluted with 0.1 M phosphate bu¡er (pH 7). The puri¢ed [14 C]PQQ sample was reacted with 16% acetone in sodium bicarbonate bu¡er (pH 9.2), and this fraction was analyzed by reverse phase HPLC. HPLC fractions (20 s, 0.33 ml) were collected and the radioactivity was determined by a liquid scintillation counting instrument. The principal radiochemically labeled peaks corresponded to [U-14 C]tyrosine and [14 C]5-acetonyl-PQQ (APQQ).
Two solvent systems were used for the reverse phase HPLC puri¢cation of PQQ and related derivatives. Fig. 3 shows the elution pro¢le for PQQ, tyrosine, and imidazolopyrroloquinoline (IPQ) when the ion-pairing agent tri£uoroacetic acid (0.1%) and an acetonitrile gradient (5^ 75% from 5 to 30 min) were used. Separation was adequate for quantitation of tyrosine, PQQ, and IPQ (eluting at 10.7, 12.8 and 14.4 min, respectively). Using this protocol, 14 C-labelled compounds from the E. coli medium eluted as a broad peak between 11 and 15 min (Fig. 3). Although ambiguous, the radioactivity eluted in the region corresponding to tyrosine, PQQ, and or IPQ. Consequently, the second HPLC protocol was used to provide better separation of tyrosine, PQQ, and IPQ (Fig. 4). Using 0.06 M phosphoric acid and a methanol gradient (30^70% from 5 to 30 min) the separation of the three standards was greatly improved. The elution of tyrosine, PQQ, and IPQ was 3, 14.4 and 24.4 min, respectively (Fig. 4). Using this protocol both the putative [14 C]PQQ and a PQQ standard eluted as a broad peak due possibly to the formation of methyl ketals, or interactions with the ODS column (data not shown). To further con¢rm the identity of the putative [14 C]PQQ the puri¢ed compound and a PQQ standard were reacted with acetone to form 5-acetonyl-PQQ [18]. 5-Acetonyl-PQQ eluted as a sharp peak
BBAGEN 25121 30-11-00
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
251
Table 1 Comparison of the speci¢c activity of [14 C]PQQ synthesized by M. extorquens grown with [U-14 C]tyrosine [23] with an E. coli tyrosine auxotroph (AT2471) transformed with pBCP165 containing PQQ synthesis genes from K. pneumoniae Bacterial source [U-14 C]Tyrosine speci¢c activity Synthesized [14 C]PQQ Dilution
Methylotrophic bacteria [23]
E. coli AT2471+pBCP165
s 1.85U1010 Bq/mmol s 500 mCi/mmol 9.25U106 Bq/mmol (0.42 mCi/mmol) V1200 fold
3.7U109 Bq/mmol 100 mCi/mmol 3.7U109 Bq/mmol (100 mCi/mmol) 0 fold
with enhanced £uorescence at 11.0 min (Fig. 4). The putative [14 C]PQQ also reacted with acetone and eluted as a sharp peak at 11.0 min. These observations were taken as evidence that [14 C]PQQ was synthesized by AT2471 transformed with pBCP165. The speci¢c activity of synthesized [14 C]PQQ was 3.7U109 Bq/mmol [14 C]PQQ, equivalent to the speci¢c activity of [U-14 C]tyrosine, added to the medium. 3.3. PQQ production E. coli auxotrophs, such as the tyrosine-requiring strain AT2471, can synthesize PQQ when transformed with plasmids containing PQQ synthesis genes. Yields ranged from 125 to 175 nmol PQQ/l (see Fig. 5 for methodological details). 4. Discussion Bacteria that are capable of PQQ synthesis typically secrete PQQ directly into the periplasmic space and eventually into the medium [20^22]. Meulenberg et al. have demonstrated that E. coli containing pBCP165 from K. pneumoniae synthesize 250^300 nmol free PQQ/l medium [10]. The data reported herein con¢rm the observations of Meulenberg et al. and also demonstrate that E. coli auxotrophs, such as the tyrosine-requiring bacterial strain AT2471, can synthesize compatable amounts of PQQ when transformed with the same plasmid containing these genes. Using this approach facilitated the incorporation of radiochemical isotopes into PQQ without dilution of the speci¢c activity. Previously, this laboratory described the synthesis of [14 C]PQQ by the methylotrophic bacterium M. extorquens AM1 (Pseudomonas AM1, American Type Culture Collection 14718) [23]. [U-14 C]Tyrosine (V85U1010 Bq/mmol) was used as the radiochemical source. PQQ labeled with 14 C was generated, however the speci¢c activity was only 9.25U106 Bq/mmol PQQ (Table 1). The dilution of speci¢c activity was V1200 fold, due to de novo tyrosine synthesis. Although the generated [14 C]PQQ was absorbed in mice, the speci¢c activity was not su¤cient to assess PQQ cellular uptake or distribution [23]. Therefore, the approach outlined here was designed to increase the speci¢c activity of synthesized PQQ by using a
tyrosine auxotrophic bacterial strain of E. coli, which should aid in the preparation of PQQ with su¤cient label for pharmacokinetic and nutritional studies. In the current study, the speci¢c activity of [14 C]PQQ synthesized by the bacteria was 3.7U109 Bq/mmol PQQ (Table 1), the same as that added to the medium. That the speci¢c activity of synthesized [14 C]PQQ was not diluted also provides additional proof that tyrosine is a precursor for PQQ synthesis (cf. [24]). This system should be useful in future studies that focus on identi¢cation of intermediates in the bacterial synthesis of PQQ. Additionally, the use of 14 C-labeled PQQ will aid in eukaryotic studies focusing on the determination of dietary absorption, whole body turnover, tissue localization, metabolism, and cellular compartmentalization of PQQ. Although the yield of PQQ in this study was low (V100^200 nmol/l), the use of a more e¤cient promoter, a di¡erent plasmid, or modi¢cation of growth conditions are approaches that could increase PQQ production. Acknowledgements Funded in part by NIH Grant RO1 DK 56031 and generous support from M.pM. Mars, Inc., Hacketstown, NJ.
References [1] W.S. McIntire, Essays Biochem. 27 (1992) 119^134. [2] C.M. Anthony, Progr. Biophys. Mol. Biol. 69 (1998) 1^21. [3] S.A. Salisbury, H.S. Forrest, W.B. Cruse, O. Kennard, Nature 280 (1979) 843^844. [4] P.M. Goodwin, C. Anthony, Adv. Microb. Physiol. 40 (1998) 1^80. [5] T.E. Stites, A.E. Mitchell, R.B. Rucker, J. Nutr. 130 (2000) 719^727. [6] K. Matsushita, J.C. Arents, R. Bader, M. Yamada, O. Adachi, P.W. Postma, Microbiology 143 (1997) 3149^3156. [7] N. Goosen, H.P. Horsman, R.G. Huinen, A. de Groot, P. van de Putte, Antonie van Leeuwenhoek 56 (1989) 85^91. [8] H. Toyama, L. Chistoserdova, M.E. Lidstrom, Microbiology 143 (1997) 595^602. [9] N. Goosen, D.A. Vermaas, P. van de Putte, J. Bacteriol. 169 (1987) 303^307. [10] J.J. Meulenberg, E. Sellink, W.A. Loenen, N.H. Riegman, M. van Kleef, P.W. Postma, FEMS Microbiol. Lett. 59 (1990) 337^343. [11] C.J. Morris, F. Biville, E. Turlin, E. Lee, K. Ellermann, W.H. Fan, R. Ramamoorthi, A.L. Springer, M.E. Lidstrom, J. Bacteriol. 176 (1994) 1746^1755.
BBAGEN 25121 30-11-00
252
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
[12] U. Schnider, C. Keel, C. Voisard, G. Defago, D. Haas, Appl. Environ. Microbiol. 61 (1995) 3856^3864. [13] A.M. Cleton-Jansen, N. Goosen, K. Vink, P. van de Putte, Mol. Gen. Genet. 217 (1989) 430^436. [14] F.M. Ausubel, Short Protocols in Molecular Biology: A Compendium of Methods from Current Protocols in Molecular Biology, Greene/Wiley, New York, 1992. [15] J.J. Meulenberg, E. Sellink, N.H. Riegman, P.W. Postma, Mol. Gen. Genet. 232 (1992) 284^294. [16] M.A. Paz, P. Martin, R. Fluckiger, J. Mah, P.M. Gallop, Anal. Biochem. 238 (1996) 145^149. [17] R. Fluckiger, M.A. Paz, P.M. Gallop, Methods Enzymol. 258 (1995) 140^149. [18] K. Kano, B. Uno, C. Kawasaki, K. Horiki, S. Kawai, Anal. Sci. 7 (1991) 737^739.
[19] A.J. Olsthoorn, J.A. Duine, Arch. Biochem. Biophys. 336 (1996) 42^ 48. [20] M. Ameyama, K. Matsushita, E. Shinagawa, M. Hayashi, O. Adachi, Biofactors 1 (1988) 51^53. [21] M.A. van Kleef, J.A. Duine, Appl. Environ. Microbiol. 55 (1989) 1209^1213. [22] T. Urakami, K. Yashima, H. Kobayashi, A. Yoshida, C. Itoyoshida, Appl. Environ. Microbiol. 58 (1992) 3970^3976. [23] C.R. Smidt, C.J. Unkefer, D.R. Houck, R.B. Rucker, Proc. Soc. Exp. Biol. Med. 197 (1991) 27^31. [24] C.J. Unkefer, D.R. Houck, B.M. Britt, T.R. Sosnick, J.L. Hanners, Methods Enzymol. 258 (1995) 227^235.
BBAGEN 25121 30-11-00
www.elsevier.com/locate/bba
Synthesis of [14 C]pyrroloquinoline quinone (PQQ) in E. coli using genes for PQQ synthesis from K. pneumoniae Tracy E. Stites, Tracey R. Sih, Robert B. Rucker * Department of Nutrition, One Shields Avenue, University of California, Davis, Davis, CA 95616, USA Received 1 August 2000; accepted 19 October 2000
Abstract Radiochemical forms of pyrroloquinoline quinone (PQQ) are of utility in studies to determine the metabolic role and fate of PQQ in biological systems. Accordingly, we have synthesized [14 C]PQQ using a tyrosine auxotrophic strain of Escherichia coli (AT2471). A construct containing the six genes required for PQQ synthesis from Klebsiella pneumoniae was used to transform the auxotrophic strain of E. coli. E. coli were then grown in minimal M9 medium containing 3.7U109 Bq/mmol [14 C]tyrosine. At confluence, the medium was collected and applied to a DEAE A-25 anionic exchange column; [14 C]PQQ was eluted using a KCl gradient (0^2 M in 0.1 M potassium phosphate buffer, pH 7.0). Radioactivity co-eluting as PQQ was next pooled, acidified and passed through a C-18 column; [14 C]PQQ was eluted with a phosphate buffer (0.1 M, pH 7.0). Reverse phase HPLC (C-18) using either the ion-pairing agent trifluoroacetic acid (0.1%) and an acetonitrile gradient or phosphoric acid and a methanol gradient were used to isolate [14 C]PQQ. Fractions were collected and analyzed by liquid scintillation counting. 14 C-labelled compounds isolated from the medium eluted corresponding to the elution of various tyrosinederived products or PQQ. The radioactive compound corresponding to PQQ was also reacted with acetone to form 5-acetonyl-PQQ, which co-eluted with a 5-acetonyl-PQQ standard, as a validation of [14 C]PQQ synthesis. The specific activity of synthesized [14 C]PQQ was 3.7U109 Bq/mmol [14 C]PQQ, equal to that of [U-14 C]tyrosine initially added to the medium. ß 2000 Elsevier Science B.V. All rights reserved. Keywords : Pyrroloquinoline quinone; Tyrosine auxotroph; Imidazolopyrroloquinoline; Klebsiella pneumoniae
1. Introduction From studies over the past 20 years, it has become clear that Gram-negative bacteria contain a novel class of dehydrogenase enzymes, which utilize pyrroloquinoline quinone (PQQ) as a cofactor [1^3]. Many bacteria that require PQQ as a cofactor also synthesize PQQ [4,5]. However, certain organisms, e.g. Escherichia coli and Salmonella typhimurium, do not synthesize PQQ even though they contain enzymes, such as glucose dehydrogenase, that require PQQ [6]. The complete cassette of genes encoding proteins for the synthesis of PQQ has been partially characterized in several bacteria. The number of genes required for bacterial synthesis of PQQ ranges from as few as four genes in
* Corresponding author. Fax: +1-530-752-8966; E-mail : [email protected]
Acinetobacter calcoaceticus [7] to seven genes in Methylobacterium extorquens [8]. In A. calcoaceticus, three of the four genes required for PQQ synthesis encode proteins with molecular weights of 29 700, 10 800, and 43 600 Da, and these proteins are involved in enzymatic processes and also the transport of PQQ into the periplasmic space [9]. The fourth gene that is required for PQQ biosynthesis encodes a polypeptide of 24 amino acids. Klebsiella pneumoniae [10], M. extorquens, [11] and Pseudomonas £uorescens [12] also contain a similar polypeptide that is required for PQQ synthesis. A conserved motif, Glu-Val-Thr-X-Tyr, is contained in all bacterial forms of this peptide. Tyrosine and glutamate contained in the short polypeptide are used as substrates for PQQ biosynthesis (Fig. 1). In the current study, the synthesis of [14 C]PQQ in E. coli (AT2471) was accomplished by utilizing a plasmid containing the complete cassette of PQQ synthesis genes (six genes) from K. pneumoniae. Since tyrosine is a substrate for PQQ synthesis (Fig. 1), by providing radiolabeled tyrosine, and preventing the endogenous synthesis of tyro-
0304-4165 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 0 0 ) 0 0 1 6 6 - 5
BBAGEN 25121 30-11-00
248
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
sine by using an E. coli tyrosine auxotroph, the resultant PQQ should incorporate the radiolabel without dilution of speci¢c activity. This approach was used in preparing [14 C]PQQ with a relatively high speci¢c activity. 2. Materials and methods 2.1. Chemicals and reagents All chemicals were obtained from Fisher Scienti¢c (Pittsburgh, PA) or Sigma Chemical Corporation (St. Louis, MO) unless stated and were of the highest purity available. PQQ was a gift from Dr. James Mah at the University of Southern California School of Dentistry, Los Angeles, CA. Reagents required for the growth and maintenance of bacteria were obtained from Difco Laboratories (Detroit, MI). E. coli strain AT2471 was obtained from the E. coli Genetic Stock Center located in the department of biology at Yale University in New Haven, CT. The E. coli bacterial strain JA221, containing pBCP165, was a gift from Dr. Pieter Postma at E.C. Slater Institute for Biochemical Research, and Biotechnology Center, University of Amsterdam, Amsterdam, The Netherlands [10]. The plasmid pgp492 containing the genes for the synthesis of a glucose dehydrogenase enzyme (sGDH) from A. calcoaceticus was a gift from Dr. Nora Goosen (Department of Molecular Genetics, University of Leiden, The Netherlands) [13]. 2.2. Transforming bacteria For the isolation of plasmid pBCP165 from the E. coli strain JA221, the bacteria were grown to con£uence then lysed and the plasmid was extracted and puri¢ed by a phenol:chloroform protocol [14]. The puri¢ed plasmid was restricted with BamHI and XhoI (Roche, Indianapolis, IN) and compared to the known restriction digest map of pBCP165. As a control, the genes for PQQ synthesis were removed from pBCP165 to regenerate the original cloning vector pCB104 [15]. Following agarose gel electrophoresis (1% agarose), the band corresponding to pCB104 was isolated (Qiagen, Valencia, CA), and the vector was then religated using standard procedures [14]. 2.3. [14 C]PQQ synthesis An E. coli tyrosine auxotroph (AT2471) was used to synthesize PQQ speci¢cally labeled with [U-14 C]tyrosine (Fig. 1). AT2471 was made electrocompetent, and the bacteria were transformed with either pBCP165 or the vector pCB104 using a standard electroporation technique [14]. The chloramphenicol antibiotic resistance gene present in both plasmids was used for selectivity of plasmid incorporation. Bacteria were grown initially in LB (1% tryptone, 0.5% yeast extract and 1% NaCl) or in minimal medium
(M9) with supplemented amino acids and thiamine as required (AT2471 requires thiamine). For [14 C]PQQ synthesis AT2471 containing plasmid pBCP165 was plated on agar plates containing M9 minimal medium with 100 Wg/ml chloramphenicol, 100 WM tyrosine and 1 mM thiamine. An isolated colony was harvested and added to 5 ml of M9 minimal sterile medium containing 10 WM tyrosine (estimated to be the minimal requirement), 1 mM thiamine, and 100 Wg/ml chloramphenicol. The bacteria were grown for 24 h at 37³C in a shaking incubator (250 rpm). A 2 ml aliquot was then transferred to 500 ml of M9 minimal sterile medium containing 1 mM thiamine, 100 Wg/ml chloramphenicol and 3.7U109 Bq/mmol [U-14 C-tyrosine] (3.86 Wmol tyrosine) plus 1.14 Wmol [U-14 C-tyrosine], speci¢c activity, 1.62U1010 Bq (438 mCi)/mmol. The bacteria were grown to con£uence (72 h), and were then pelleted by centrifugation at 1000Ug. The medium (supernatant fraction) was saved for isolation of [14 C]PQQ. 2.4. PQQ puri¢cation The medium was applied to a DEAE Sephadex A-25 (1U3 cm) column at a £ow rate of 1 ml/min. The column was washed with 0.1 M sodium phosphate (pH 7.0) until the eluent radioactivity was at background. PQQ was eluted from the column using a KCl gradient (0^2 M) in 0.1 M sodium phosphate (pH 7.0). All fractions from the DEAE A-25 column were analyzed by the redox cycling assay as described by Gallop and colleagues [16,17]. The fractions containing redox cycling activity corresponding to the known elution pro¢le for authentic PQQ (0.6^0.8 M KCl) were pooled for further puri¢cation. The pH of the pooled fractions was adjusted to pH 1.5 with HCl, and applied to a C-18 low pressure Bond Elut column (Varian, Walnut Creek, CA). PQQ was eluted with a 0.1 M potassium phosphate bu¡er (pH 7.0) and saved for HPLC analysis. Reverse phase HPLC was performed using a HewlettPackard series 1100 HPLC system (Agilent Technologies, Palo Alto, CA), with an in-series diode array and £uorescence detector. Absorbance was monitored at 249, 280, and 422 nm and £uorescence at ex = 360 and em = 480 nm. A Phenomenex 5 W Prodigy (250 cmU4.6 mm I.D., 10 nm pore size) C-18 column was used for analytical separation. Fractions (0.33 ml) were collected and counted by a liquid scintillation counter (Wallac, Gaithersburg, MD) to determine 14 C radioactivity. Two separate HPLC protocols were utilized. The initial mobile phase for the ¢rst HPLC protocol consisted of 5% acetonitrile and 95% water (containing 0.1% tri£uoroacetic acid (v/v)). A linear gradient was applied from 5 to 30 min with a ¢nal concentration of 75% acetonitrile and 25% water (containing 0.1% tri£uoroacetic acid (v/v)). The initial mobile phase for the second HPLC protocol consisted of 30% methanol and 70% 0.06 M phosphoric acid. A
BBAGEN 25121 30-11-00
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
249
glucose dehydrogenase enzyme assay [19]. Soluble glucose dehydrogenase enzyme (sGDH, soluble type; EC 1.1.99.17) was produced in E. coli transformed with pgp492, a plasmid containing the GDH synthesis genes from A. calcoaceticus. To isolate GDH, transformed E. coli were grown to con£uence and centrifuged at 1000Ug. The bacteria were lysed twice in a French press (16000 psi) and were then treated with DNase (1 mg DNase/g bacteria (2000 units DNase/mg)) for 48 h at 20³C. The bacterial debris was pelleted by centrifugation at 48 000Ug and the supernatant fraction was applied to a CM-Sepharose cation exchange column (1.5U20 cm). The sGDH enzyme was eluted from the column with a KCl gradient (0^0.6 M). Fig. 1. The synthesis of PQQ from tyrosine and glutamic acid. [U14 C]Tyrosine was added to cultures of a tyrosine auxotrophic strain of E. coli (AT2471) to synthesize [14 C]PQQ labeled in nine of the PQQ's 14 carbons.
linear gradient was applied from 5 to 30 min with a ¢nal concentration of 70% methanol and 30% 0.06 M phosphoric acid.
3. Results 3.1. Synthesis of PQQ In preliminary studies, the bacterial strain AT2471 transformed with either pBCP165, containing the PQQ
2.5. PQQ derivatization PQQ was derivatized with acetone to form the acetone adduct (5-acetonyl-PQQ) to aid in identity and validation [18]. PQQ (200 nmol, 0.2 ml) was derivatized in 0.1 M sodium carbonate (pH 9.2, 0.1 ml) with the addition of 16% acetone (v/v, 0.1 ml) at 37³C for 30 min. Samples containing radioactivity were derivatized using the same procedure. 2.6. PQQ analysis The PQQ concentration of samples was determined by a
Fig. 2. Tyrosine-derived products. Radiochemically labeled tyrosine-derived products from the tyrosine auxotrophic strain AT2471 transformed with pBCP165 from K. pneumoniae. AT2471 was grown in minimal M9 medium containing 3.7U109 Bq/mmol tyrosine. The bacteria were pelleted and the supernatant fraction was applied to a DEAE-A25 (1U3 cm) column. PQQ was eluted from the column with a 0^2 M gradient of KCl. Fractions (5 ml) were collected and analyzed by liquid scintillation counting, and by a redox cycling assay [17]. The peak of redox cycling activity corresponding to a PQQ standard was pooled and used for further analysis.
Fig. 3. Chromatography of PQQ-related products. Reverse phase HPLC analysis of tyrosine, PQQ, and IPQ standards. The initial mobile phase consisted of 5% acetonitrile and 95% water (containing 0.1% tri£uoroacetic acid (v/v)). A linear gradient was applied from 5 to 30 min with a ¢nal concentration of 75% acetonitrile and 25% water (containing 0.1% tri£uoroacetic acid (v/v)). Peak 1, 1 mM tyrosine (280 nm); peak 2, 0.1 mM PQQ (ex = 360 nm and em = 480 nm); peak 3, 0.4 mM IPQ (422 nm). Also indicated is an HPLC analysis of 14 C-labeled products synthesized by the bacterial strain AT2471. Medium from AT2471 was applied to a DEAE-A25 column and [14 C]PQQ was eluted with a KCl gradient. A redox cycling peak corresponding to a PQQ standard was collected, and further puri¢ed by addition to a low pressure C-18 column. [14 C]PQQ was eluted with 0.1 M phosphate bu¡er (pH 7), and this fraction was analyzed by reverse phase HPLC. HPLC fractions (20 s, 0.33 ml) were collected and radioactivity was determined by liquid scintillation counting.
BBAGEN 25121 30-11-00
250
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
synthesis genes, or pCB104, vector alone, were analyzed for PQQ production. The PQQ concentration in the medium (LB), when AT2471 was transformed with pBCP165, was 165 nmol/l and was below detection limits (0.5 nmol/l) when AT2471 was grown with vector pCB104. For the synthesis of [14 C]PQQ, AT2471 was grown to con£uence in the presence of 3.7U109 Bq/mmol [U-14 C]tyrosine. The bacterial medium was applied to a DEAE A-25 column and PQQ was eluted with a KCl gradient (Fig. 2). Fractions with a KCl concentration range of 0.6^0.8 M KCl contained detectable redox cycling activity and were pooled for reverse phase HPLC analysis (Figs. 3 and 4).
Fig. 5. CM-Sepharose chromatography of the soluble glucose dehydrogenase enzyme (sGDH). E. coli transformed with pgp492 containing the synthesis genes for sGDH from A. calcoaceticus was used as a source of the enzyme. Cells were grown to con£uence, lysed with a French press (two times) and pelleted to remove bacterial debris. The suspension was treated for 2 days with DNase and centrifuged at 48 000Ug. Fractions were then applied to a CM-Sepharose cationic exchange column. A linear gradient (0^0.6 M KCl) was applied from fraction number 30 to 80. The majority ( s 95%) of the protein applied to the column eluted in fractions 1^29. Enzyme activity eluted in the area indicated by the line. A typical standard curve is shown in the inset. The enzyme was assayed as described by Olsthoorn and Duine [19] using phenazine methosulfate as the primary electron acceptor, dichlorophenolindophenol as the secondary electron acceptor and colorimetric agent. Glucose was used as the substrate and the detection range was from 0.05 to 0.5 pmol of PQQ.
3.2. Puri¢cation and analysis of [14 C]PQQ synthesized by AT2471
Fig. 4. Chromatography of tyrosine, PQQ, 5-acetonyl-PQQ (APQQ) and IPQ. The initial mobile phase consisted of 30% methanol and 70% 0.06 M phosphoric acid. A linear gradient was applied from 5 to 30 min with a ¢nal concentration of 70% methanol and 30% 0.06 M phosphoric acid. Peak 1, 1 mM tyrosine (280 nm); peak 2; 0.05 mM 5-acetonyl-PQQ (ex = 360 nm and em = 480 nm); peak 3, 0.1 mM PQQ (ex = 360 nm and em = 480 nm); peak 4, 0.4 mM IPQ (422 nm). Also indicated is the reverse phase HPLC analysis of acetone derivatized [14 C]PQQ synthesized by the bacterial strain AT2471. Medium from AT2471 was applied to a DEAE-A25 column and [14 C]PQQ was eluted with a KCl gradient. A redox cycling peak corresponding to a PQQ standard was collected, and further puri¢ed by addition to a low pressure C-18 column. [14 C]PQQ was eluted with 0.1 M phosphate bu¡er (pH 7). The puri¢ed [14 C]PQQ sample was reacted with 16% acetone in sodium bicarbonate bu¡er (pH 9.2), and this fraction was analyzed by reverse phase HPLC. HPLC fractions (20 s, 0.33 ml) were collected and the radioactivity was determined by a liquid scintillation counting instrument. The principal radiochemically labeled peaks corresponded to [U-14 C]tyrosine and [14 C]5-acetonyl-PQQ (APQQ).
Two solvent systems were used for the reverse phase HPLC puri¢cation of PQQ and related derivatives. Fig. 3 shows the elution pro¢le for PQQ, tyrosine, and imidazolopyrroloquinoline (IPQ) when the ion-pairing agent tri£uoroacetic acid (0.1%) and an acetonitrile gradient (5^ 75% from 5 to 30 min) were used. Separation was adequate for quantitation of tyrosine, PQQ, and IPQ (eluting at 10.7, 12.8 and 14.4 min, respectively). Using this protocol, 14 C-labelled compounds from the E. coli medium eluted as a broad peak between 11 and 15 min (Fig. 3). Although ambiguous, the radioactivity eluted in the region corresponding to tyrosine, PQQ, and or IPQ. Consequently, the second HPLC protocol was used to provide better separation of tyrosine, PQQ, and IPQ (Fig. 4). Using 0.06 M phosphoric acid and a methanol gradient (30^70% from 5 to 30 min) the separation of the three standards was greatly improved. The elution of tyrosine, PQQ, and IPQ was 3, 14.4 and 24.4 min, respectively (Fig. 4). Using this protocol both the putative [14 C]PQQ and a PQQ standard eluted as a broad peak due possibly to the formation of methyl ketals, or interactions with the ODS column (data not shown). To further con¢rm the identity of the putative [14 C]PQQ the puri¢ed compound and a PQQ standard were reacted with acetone to form 5-acetonyl-PQQ [18]. 5-Acetonyl-PQQ eluted as a sharp peak
BBAGEN 25121 30-11-00
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
251
Table 1 Comparison of the speci¢c activity of [14 C]PQQ synthesized by M. extorquens grown with [U-14 C]tyrosine [23] with an E. coli tyrosine auxotroph (AT2471) transformed with pBCP165 containing PQQ synthesis genes from K. pneumoniae Bacterial source [U-14 C]Tyrosine speci¢c activity Synthesized [14 C]PQQ Dilution
Methylotrophic bacteria [23]
E. coli AT2471+pBCP165
s 1.85U1010 Bq/mmol s 500 mCi/mmol 9.25U106 Bq/mmol (0.42 mCi/mmol) V1200 fold
3.7U109 Bq/mmol 100 mCi/mmol 3.7U109 Bq/mmol (100 mCi/mmol) 0 fold
with enhanced £uorescence at 11.0 min (Fig. 4). The putative [14 C]PQQ also reacted with acetone and eluted as a sharp peak at 11.0 min. These observations were taken as evidence that [14 C]PQQ was synthesized by AT2471 transformed with pBCP165. The speci¢c activity of synthesized [14 C]PQQ was 3.7U109 Bq/mmol [14 C]PQQ, equivalent to the speci¢c activity of [U-14 C]tyrosine, added to the medium. 3.3. PQQ production E. coli auxotrophs, such as the tyrosine-requiring strain AT2471, can synthesize PQQ when transformed with plasmids containing PQQ synthesis genes. Yields ranged from 125 to 175 nmol PQQ/l (see Fig. 5 for methodological details). 4. Discussion Bacteria that are capable of PQQ synthesis typically secrete PQQ directly into the periplasmic space and eventually into the medium [20^22]. Meulenberg et al. have demonstrated that E. coli containing pBCP165 from K. pneumoniae synthesize 250^300 nmol free PQQ/l medium [10]. The data reported herein con¢rm the observations of Meulenberg et al. and also demonstrate that E. coli auxotrophs, such as the tyrosine-requiring bacterial strain AT2471, can synthesize compatable amounts of PQQ when transformed with the same plasmid containing these genes. Using this approach facilitated the incorporation of radiochemical isotopes into PQQ without dilution of the speci¢c activity. Previously, this laboratory described the synthesis of [14 C]PQQ by the methylotrophic bacterium M. extorquens AM1 (Pseudomonas AM1, American Type Culture Collection 14718) [23]. [U-14 C]Tyrosine (V85U1010 Bq/mmol) was used as the radiochemical source. PQQ labeled with 14 C was generated, however the speci¢c activity was only 9.25U106 Bq/mmol PQQ (Table 1). The dilution of speci¢c activity was V1200 fold, due to de novo tyrosine synthesis. Although the generated [14 C]PQQ was absorbed in mice, the speci¢c activity was not su¤cient to assess PQQ cellular uptake or distribution [23]. Therefore, the approach outlined here was designed to increase the speci¢c activity of synthesized PQQ by using a
tyrosine auxotrophic bacterial strain of E. coli, which should aid in the preparation of PQQ with su¤cient label for pharmacokinetic and nutritional studies. In the current study, the speci¢c activity of [14 C]PQQ synthesized by the bacteria was 3.7U109 Bq/mmol PQQ (Table 1), the same as that added to the medium. That the speci¢c activity of synthesized [14 C]PQQ was not diluted also provides additional proof that tyrosine is a precursor for PQQ synthesis (cf. [24]). This system should be useful in future studies that focus on identi¢cation of intermediates in the bacterial synthesis of PQQ. Additionally, the use of 14 C-labeled PQQ will aid in eukaryotic studies focusing on the determination of dietary absorption, whole body turnover, tissue localization, metabolism, and cellular compartmentalization of PQQ. Although the yield of PQQ in this study was low (V100^200 nmol/l), the use of a more e¤cient promoter, a di¡erent plasmid, or modi¢cation of growth conditions are approaches that could increase PQQ production. Acknowledgements Funded in part by NIH Grant RO1 DK 56031 and generous support from M.pM. Mars, Inc., Hacketstown, NJ.
References [1] W.S. McIntire, Essays Biochem. 27 (1992) 119^134. [2] C.M. Anthony, Progr. Biophys. Mol. Biol. 69 (1998) 1^21. [3] S.A. Salisbury, H.S. Forrest, W.B. Cruse, O. Kennard, Nature 280 (1979) 843^844. [4] P.M. Goodwin, C. Anthony, Adv. Microb. Physiol. 40 (1998) 1^80. [5] T.E. Stites, A.E. Mitchell, R.B. Rucker, J. Nutr. 130 (2000) 719^727. [6] K. Matsushita, J.C. Arents, R. Bader, M. Yamada, O. Adachi, P.W. Postma, Microbiology 143 (1997) 3149^3156. [7] N. Goosen, H.P. Horsman, R.G. Huinen, A. de Groot, P. van de Putte, Antonie van Leeuwenhoek 56 (1989) 85^91. [8] H. Toyama, L. Chistoserdova, M.E. Lidstrom, Microbiology 143 (1997) 595^602. [9] N. Goosen, D.A. Vermaas, P. van de Putte, J. Bacteriol. 169 (1987) 303^307. [10] J.J. Meulenberg, E. Sellink, W.A. Loenen, N.H. Riegman, M. van Kleef, P.W. Postma, FEMS Microbiol. Lett. 59 (1990) 337^343. [11] C.J. Morris, F. Biville, E. Turlin, E. Lee, K. Ellermann, W.H. Fan, R. Ramamoorthi, A.L. Springer, M.E. Lidstrom, J. Bacteriol. 176 (1994) 1746^1755.
BBAGEN 25121 30-11-00
252
T.E. Stites et al. / Biochimica et Biophysica Acta 1524 (2000) 247^252
[12] U. Schnider, C. Keel, C. Voisard, G. Defago, D. Haas, Appl. Environ. Microbiol. 61 (1995) 3856^3864. [13] A.M. Cleton-Jansen, N. Goosen, K. Vink, P. van de Putte, Mol. Gen. Genet. 217 (1989) 430^436. [14] F.M. Ausubel, Short Protocols in Molecular Biology: A Compendium of Methods from Current Protocols in Molecular Biology, Greene/Wiley, New York, 1992. [15] J.J. Meulenberg, E. Sellink, N.H. Riegman, P.W. Postma, Mol. Gen. Genet. 232 (1992) 284^294. [16] M.A. Paz, P. Martin, R. Fluckiger, J. Mah, P.M. Gallop, Anal. Biochem. 238 (1996) 145^149. [17] R. Fluckiger, M.A. Paz, P.M. Gallop, Methods Enzymol. 258 (1995) 140^149. [18] K. Kano, B. Uno, C. Kawasaki, K. Horiki, S. Kawai, Anal. Sci. 7 (1991) 737^739.
[19] A.J. Olsthoorn, J.A. Duine, Arch. Biochem. Biophys. 336 (1996) 42^ 48. [20] M. Ameyama, K. Matsushita, E. Shinagawa, M. Hayashi, O. Adachi, Biofactors 1 (1988) 51^53. [21] M.A. van Kleef, J.A. Duine, Appl. Environ. Microbiol. 55 (1989) 1209^1213. [22] T. Urakami, K. Yashima, H. Kobayashi, A. Yoshida, C. Itoyoshida, Appl. Environ. Microbiol. 58 (1992) 3970^3976. [23] C.R. Smidt, C.J. Unkefer, D.R. Houck, R.B. Rucker, Proc. Soc. Exp. Biol. Med. 197 (1991) 27^31. [24] C.J. Unkefer, D.R. Houck, B.M. Britt, T.R. Sosnick, J.L. Hanners, Methods Enzymol. 258 (1995) 227^235.
BBAGEN 25121 30-11-00