Increased production of recombinant pyrroloquinoline quinone (PQQ) glucose dehydrogenase by metabolically engineered Escherichia coli strain capable of PQQ biosynthesis

Biotechnolo Journal of Biotechnology 49 (1996) 239-243

Short note

Increased production of recombinant pyrroloquinoline quinone (PQQ) glucose dehydrogenase by metabolically engineered Escherichia coli strain capable of PQQ biosynthesis Koji Sode”,*, Koji Ito”, Arief Budi Witartoa, Kazumoto Hiromi Yoshida”, Pieter Postmab

Watanabea,

“Department of’ Biotechnology, Faculty of Technology, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei-shi, Tokyo 184, Japan bE.C. Slater Institute, BioCentrum Amsterdam, University of Amsterdam, 1018 TV Amsterdam, The Netherlands Received 18 October 1995; revised 20 April 1996; accepted 23 April 1996

Abstract We have previously shown that the production of recombinant Escherichia coli PQQGDH was greatly improved by using a medium supplemented with the cofactor PQQ, which is not synthesized in E. coli. We show here that the increase in the accumulated PQQGDH is due to the increased stability of the holo-enzyme over apo-enzyme, using recombinant Acinetobacter calcoaceticus PQQGDH. In order to achieve cost-effective PQQGDH production, we incorporated the genes for PQQ biosynthetic pathway from Klebsiella pneumoniae into E. coli, which as a result allowed E. roli to produce PQQ. Using this metabolically engineered E. coli strain as a host, a IO-fold increase in the production of recombinant A. calcoaceticus PQQGDH was achieved, compared to the condition without PQQ and MgCI,. Keywords: Pyrroloquinoline cus

quinone (PQQ); Glucose dehydrogenase;

Metabolic engineering; Acinetobacter

calcoaceti-

1. Introduction

* Corresponding author.

Various types of glucose oxidoreductases have been used as a biosensor component in the determination of glucose. Among them, glucose dehy-

0168-1656/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved PIZ SO168-1656(96)01540-4

240

K. So&

et ul.

: Journul

of Biotcchno1og.v

drogenases (GDHs) possessing pyrroloquinoline quinone (PQQ) as the prosthetic group are recognized as ideal enzymes for mediator-type glucose sensors since they are very insensitive to the presence of oxygen in the samples (Turner et al., 1987; Yokoyama et al., 1989; Sode et al., 1993; Ye et al., 1993). The authors have been carrying out protein engineering of PQQGDHs with the aim to improve their enzymatic characteristics (Sode and Sano, 1994; Sode et al., 1994, 1995). In order to advance the protein engineering of PQQGDHs, development of efficient production of recombinant PQQGDHs is essential. We previously reported the production of recombinant E. coli PQQGDH using an E. co/i host strain defective in PQQGDH biosynthesis (Sode et al., 1994). We found that the presence of PQQ and MgCl, in the medium during cultivation greatly enhanced the amount of active PQQGDH. Since E. coli cannot synthesize PQQ, this might be due to the formation of stable holo-PQQGDH enzyme in the presence of PQQ, compared to unstable apoPQQGDH enzyme which was synthesized in the absence of PQQ. However, the high cost of PQQ prevents its use as a medium component for the large scale preparation of recombinant PQQGDHs. With the recent advance in the molecular biology of PQQ biosynthesis, gene clusters (pqq operon) containing the genes involved in PQQ biosynthesis have been cloned and characterized (Biville et al., 1989; Goosen et al., 1989; Meulenberg et al., 1990, 1992; Morris et al., 1994). This has resulted in E. coli strains that harbor heterologous pqq operons and can synthesize PQQ. Therefore, by using such metabolically engineered E. coli strains, efficient PQQGDH production might be possible. In this study, we report the production of recombinant PQQGDH derived from A. culcorzceticus, by using a metabolically engineered E. coli strain, capable of PQQ biosynthesis. 2. Materials 2. I. Bacteriul

49 (1996) 139- 24.7

which the PQQGDH structural gene (gcd) was disrupted by insertion mutagenesis, was used as the host strain for the production of recombinant PQQGDH. For the production of A. calcoaceticus PQQGDH, a plasmid, pGAc1 (Sode et al.. 1995) was used. This plasmid was constructed by inserting the structural gene of A. culcouceticus IF012552, the encoding membrane-bound PQQGDH. under control of the trc promoter. present in the expression vector pTrc99A (Pharmacia, Sweden). For the biosynthesis of PQQ, a plasmid, pBCP168Km was used (Fig. 1). This plasmid was derived from pBCP168 (Velterop et al., 1995) containing the pqq operon from Kkhsirllu pncwnoniac~ by inserting a kanamycin resistance gene block (Pharmacia, Sweden) in the EcoRI site. 2.2. C’u1fur.e condition crnulysis

und plumid

.ytuhilitJ

E. coli PP2418/pGAcI and PP2418ipGAcIi pBCPl68Km were cultured in Luria broth in 500-ml culture flask with 150 ml of medium supplemented with 25 jig ml ’ chloramphenicol, 25 ’ ampicillin for PP2418/‘pGAcI and also /ig ml 25 jcg ml ’ kanamycin for PP24 18/pGAcI,

and methods Fig. I. Map of plasmid pBCPl68Km. P,~qq, promoter of pqq operon: Cm, chlordmphenicol resistance gene: Km, kanamycin resistance gene; pqqA F, pyq operon from Kkchsidla prwumo-

strain and plusmids

E. coli PP2418

(Cleton-Jansen

et al., 1990)

in

/liar.

K. Sode et al. /Journal

of Biotechnology

pBCP168Km. To form holo-enzyme, 600 nM PQQ and 10 mM MgCl, were added to the growth medium. The cells were cultured on a shaker aerobically at 37°C. When the optical density at 660 nm reached 0.8, 0.3 mM of IPTG was added to induce the expression of the gcd gene encoding the A. culcoaceficus GDH enzyme. Plasmid stability of pGAc1 and pBCP168Km in E. coli PP2418 was analysed as follows. E. coli PP2418/pGAcI/pBCP168Km was cultured as described above, and 50 ,~l of culture was sampled after 10 h of cultivation. After appropriate dilution (lO~h~lO~X), samples were spread over a Luria agar plate containing either (a) 25 pug ml ’ chloramphenicol or (b) 25 pug ml-’ of ampicillin and each of chloramphenicol, kanamycin. Colonies appearing on the plates containing all three antibiotics were considered to be cells harbouring both pGAc1 and pBCP168Km, whereas colonies appearing on the plate containing only chloramphenicol were considered to represent the total number of viable cells. Percentage of cells harboring both pGAc1 and pBCP168Km was calculated by dividing the number of colonies appeared on (b) agar plate by that of (a) agar plate. 2.3. Enzyme

assay condition

PQQGDH activity was determined as previously reported (Sode and Sano, 1994). The time course of thermo-inactivation was measured by incubating solubilized membrane fraction of PP2418/pGAcI at 37°C. Samples were taken every 2 or 5 min and were stored at 4°C for 1 min. The remaining PQQGDH activity was assayed as described previously (Sode and Sano, 1994). 2.4. PQQ

bioassay

The amount of PQQ produced by metabolically engineered E. coli was biochemically determined according to Ameyama et al. (Ameyama et al., 1985; van der Meer et al., 1990) using recombinant E. coli apo-PQQGDH produced by PP2418/pGEcI as reported previously (Sode et

49 (1996) 239-243

-2 ’ -0

241

I

I

I

20

40

60

Heating time (min) Fig. 2. Inactivation plot of apo-form and holo-form of PQQGDH of A. cnlcoaceticus at 37°C. Solubilized membrane was incubated in a 10 mM potassium phosphate buffer, pH 7.0, containing 0.2 (w/v) ‘% Triton X-100. Enzyme activity was measured at room temperature. C, holo-form; l , apo-form.

al., 1994). Based on this method, of PQQ detection was 0.1 nM.

the lower limit

3. Results and discussion 3.1. Thermal stability holo-PQQGDH

oj’ upo - and

The time course of thermo-inactivation at 37°C for both apo- and holo-PQQGDHs from A. culcoaceticus is shown in Fig. 2. Thermo-inactivation could be adequately described by first-order kinetics, from linear regression of ln(A/AO) against time. Holo-PQQGDH was not inactivated at this temperature over a period of at least 60 min. The activity of apo-PQQGDH gradually decreased, however, at 37°C with a half-life of 1500 s. Considering this time course, recombinant PQQGDH in apo-form will easily lose its activity during the cultivation. These results clearly indicated that thermal stability of holo-PQQGDH of A. calcoaceticus was much higher than that of apo-PQQGDH. Recently, Matsushita et al. (1995) reported the conformational change of A. culcoaceticus PQQGDH during holo-enzyme formation. They reported that holo-enzyme retained

K. Sode et al. I Journal of Biotechnology 49 (1996) 239-243

242 Table 1 Bacterial strains Strain

or plasmid

Escherichia coli PP24 18 Plasmids pTrc99A pGEc1 pGAc1 pBCPl68 pBCPl68Km

and plasmids Genotype

Source

or phenotype

or reference

pts I, thi, galP, gcd::cat (Cmr)

Cleton-Jansen

Ap’; pBR322(CoIEl) derivative, trcP vector, laclq Ap’; pTrc99A derivative containing E. coli gc’d gene Ap’; pTrc99A derivative containing A. calcoacrticus gcd gene Cm’; pACYCl84(pl5A) derivative containing the complete ~44 operon Km’; pBCPl68 derivative with Km resistance gene inserted in Cm resistance gene

Phannacia Sode and Sano (1994) Sode et al. (1995) Velterop et al. (1995)

higher resistance toward tryptic digestion than apo-enzyme. According to their observation, the increase in the thermal stability of PQQGDH by holo-enzyme formation may be due to the conformational change. Therefore, for the production of recombinant A. calcoaceticus PQQGDH, conditions that favor the formation of holo-enzyme are preferred.

This study

P

3.2. PQQGDH production by metabolically engineered E. coli

O7

In order to achieve a cost-effective PQQGDH production, we constructed an E. coli strain which was capable of both overexpression of recombinant PQQGDH and of PQQ biosynthesis. Since the pGAc1 and pBCP168Km are compatible (Table l), both plasmids were transformed into E. coli PP2418 to construct the target strain. The time course of PQQGDH production using this strain is shown in Fig. 3. This strain secreted lo-20 nM PQQ into the growth medium and contained a high level of active PQQGDH. This amount of PQQ synthesized was lower than that observed in the previous studies (230 nM; Meulenberg et al., 1990, 180 nM; Velterop et al., 1995). Since we used PP2418 as E. coli mutant strain, the expression level of pqq operon may be dependent upon the host strain. It was also possible that the simultaneous introduction of an expression vector to achieve the over-production of A. calcoaceticus PQQGDH might result in the decrease in the expression level of PQQ biosynthe-

et al. (1990)

3

14001200 -

g 8

1000-

$j

800-

k

600

-

!j

400

-

z

200 0

- I= 0

I 2

4

q

I

6 8 Time (hour)

10

I 12’

Fig. 3. Time course of PQQGDH production. The cell growth, PQQGDH production (for both PP2418ipGAcI and PP2418i pGAcIipBCPl68Km) and extra cellular PQQ concentration (PP2418/pGAcI~pBCP168Km) were presented. Cells were cultured in the presence of 10 mM MgCl,. The arrow indicates the time when IPTG was added. The amount of active PQQGDH produced was expressed as observed enzyme activity, [U], per culture volume, [liter], in each condition. A, PQQ concentration; n , PQQGDH produced from PP241R/pGAcl; 0, PQQGDH produced from PP2418,/pGAcl/pBCP168Km; q , Cell growth of PP2418/pGAcl; Cl; Cell growth of PP2418/ pGAclipBCPl68Km.

K. Sode et al. /Journal

of Biotechnology 49 (1996) 239-243

sis pathway. The maximum level of PQQGDH synthesized was 1500 U 1-l (2.5 U mg- ’ protein), which was more than lo-fold higher than under conditions using a medium without PQQ and additional MgCl,. A similar level of PQQGDH production was also achieved by adding 600 nM PQQ together with 10 mM MgCl, to the medium (1100 U 1-l; 2.0 U mg-’ protein; results were not shown), as had been observed in recombinant E. coli PQQGDH production (Sode et al., 1994). The plasmid stability of both pGAc1 and pBCP168Km was investigated. After 10 h of cultivation, the population of E. coli cells harboring both plasmids was 90%, and therefore, these plasmids were quite stably maintained in the host strain. We conclude that the use of a metabolically engineered E. coli strain, capable of PQQ biosynthesis, allowed the production of a high level of active PQQGDH.

References Ameyama, M., Nonobe, M., Shinagawa, E., Matsushita, K. and Adachi, 0. (1985) Method of enzymatic determination of pyrroloquinoline quinone. Anal. Biochem. 151, 2633 267. Biville, F., Turlin, E. and Gasser, F. (1989) Cloning and genetic analysis of six pyrroloquinoline quinone biosynthesis genes in Methylobacterium organophilum DSM760. J. Gen. Microbial. 135, 2917-2929. Cleton-Jansen, A.M., Goosen, N., Fayet, 0. and van de Putte, P. (1990) Cloning, mapping, and sequencing of the gene encoding Escherichia coli quinoprotein glucose dehydrogenase. J. Bacterial. 172, 630886315. Goosen, N., Horsman, H.P.A., Huinen, R.G.M. and van de Putte, P. (1989) Acinetobacter calcoacericus genes involved in biosynthesis of the coenzyme pyrrolo-quinoline-quinone: nucleotide sequence and expression in Escherichia coli K12. J. Bacterial. 171, 447-455. Matsushita, K., Toyama, H., Ameyama, M., Adachi, O., Dewanti, A. and Duine, J.A. (1995) Soluble and membrane-bound quinoprotein D-ghCOSe dehydrogenases of the Acinetobacter calcoaceticus: the binding process of PQQ to the apoenzymes. Biosci. Biotechnol. Biochem. 59, 154881555.

243

Meulenberg, J.J.M., Sellink, E., Loenen, W.A.M., Riegman, N.H., van Kleef, M. and Postma, P.W. (1990) Cloning of Klebsiella pneumoniae pqq genes and PQQ biosynthesis in Escherichia coli. FEMS Microbial. Lett. 71, 337-344. Meulenberg, J.J.M., Sellink, E., Riegman, N.H. and Postma, P.W. (1992) Nucleotide sequence and structure of the Klebsiella pneumoniae pqq operon. Mol. Gen. Genet. 232, 2844294. Morris, C.J., Biville, F., Turlin, E., Lee, E., Ellermann, K., Fan, W., Ramamoorthi, R., Springer, A.L. and Lindstrom, M.E. (1994) Isolation, phenotypic characterization, and complementation analysis of mutants of Methylobacterium extorquens AM 1 unable to synthesize pyrroloquinoline quinone and sequence of pqqD, pqqG, and pqqC. J. Bacteriol. 176, 174661755. Sode, K., Nakasono, S., Tanaka, M. and Matsunaga, T. (1993) Subzero temperature operating biosensor utilizing an organic solvent and quinoprotein glucose dehydrogenase. Biotechnol. Bioeng. 42, 251-254. Sode, K. and Sano, H. (1994) Glu742 substitution to Lys enhances the EDTA tolerance of Escherichia coli PQQ glucose dehydrogenase. Biotechnol. Lett. 16, 455-460. Sode, K., Witarto, A.B., Watanabe, K., Noda, K., Ito, S. and Tsugawa, W. (1994) Over-expression of PQQ glucose dehydrogenase in Escherichia coli under holo-enzyme forming condition. Biotechnol. Lett. 16, 1265- 1268. Sode, K., Yoshida, H., Matsumura, K., Kikuchi, T., Watanabe, M., Yasutake, N., Ito, S. and Sano, H. (1995) Elucidation of the region responsible for EDTA tolerance in PQQ glucose dehydrogenases by constructing Escherichia coli and Acinetobacter calcoaceticus chimeric enzymes. Biochem. Biophys. Res. Commun. 211, 268-273. Turner, A.P.F., D’Costa, E.J. and Higgins, I.J. (1987) Enzymatic analysis using quinoprotein dehydrogenases. Ann. New York Acdd. Sci. Enzyme Eng. 8 (501) 283-287. van der Meer, R.A., Groen, B.W., van Kleef, M.A.G., Frank, J., Jongejan, J.A. and Duine, J.A. (1990) Isolation, preparation, and assay of pyrroloquinoline quinone. Methods Enzymol. 188, 260-283. Velterop, J.S., Sellink, E., Meulenberg, J.J.M., David, S., Bulder, I. and Postma, P.W. (1995) Synthesis of pyrroloquinoline quinone in vivo and in vitro and detection of an intermediate in the biosynthetic pathway. J. Bacterial. 177, 5088-5098. Ye, L., Hammerle, M., Olsthoorn, A.J.J., Schuhmann, W., Schmidt, H.-L., Duine, J.A. and Heller, A. (1993) High current density ‘Wired’ quinoprotein glucose dehydrogenase electrcde. Anal. Chem. 65, 238-241. Yokoyama, K., Sode, K., Tamiya, E. and Karube, 1. (1989) Integrated biosensor for glucose and galactose. Anal. Chim. Acta 218, 1377142.