Optimization of microbial transglutaminase production using experimental designs

Appl Microbiol Biotechnol (1997) 48: 730±734

Ó Springer-Verlag 1997

SHORT CONTRIBUTION

M. Junqua á R. Duran á C. Gancet á P. Goulas

Optimization of microbial transglutaminase production using experimental designs

Received: 23 July 1997 / Accepted: 25 August 1997

Abstract In prokaryotes, transglutaminase (TGase) has been found only in actinomycetes from the genus Streptoverticillium. The role of this TGase, as well as the mechanism regulating the enzyme expression, are still unknown. In order to improve TGase production by Streptoverticillium cinnamoneum CBS 683.68 and simultaneously elucidate the relationship between growth and TGase activity, we decided to study these two responses using di€erent designs of statistical analysis. Among the ®ve factors tested, casein, glycerol, peptones, yeast extract and oligoelements, only oligoelements were found to have no e€ect either on growth or on TGase production in a complete factorial design. The two factors casein and glycerol were found to have a highly signi®cant e€ect on both dry weights and TGase activity in a Box-Behnken design used to improve the model. Finally, the TGase activity was increased three times to reach 0:331  0:038 U/ml with optimum concentrations of casein (38.4 g/l) and glycerol (31.2 g/l) calculated with the help of a composite design. In the course of these experiments, the two responses varied in the same way, demonstrating that growth and TGase production were tightly correlated under the conditions described. However, TGase was produced during the stationary phase of growth in optimized medium, indicating that the enzyme production could be induced.

M. Junqua á R. Duran á P. Goulas (&) Laboratoire d'Ecologie MoleÂculaire, I.B.E.A.S., Universite de Pau et des Pays de l'Adour, Avenue de l'Universite F-64000 Pau, France Tel.: +33 5 59 92 31 45 Fax: +33 5 59 80 83 11 e-mail: [email protected] C. Gancet Groupement de Recherches de Lacq, POBOX 34, Lacq, F-64170, Artix, France

Introduction Transglutaminases (TGase, R-glutaminyl peptide: amine c-glutaminyltransferase, EC 2.3.2.13) are thiol enzymes that catalyse protein reticulation by introducing isopeptide bonds [e-(c-glutaminyl)lysine crosslinking] (Folk and Finnlayson 1977). TGases have been found in a wide variety of eukaryotic organisms. These eukaryote transglutaminases, Ca2+-dependent, are involved in numerous biological functions ranging from blood clotting to cell di€erentiation (for reviews see Friedrich and Aszodi 1992; Greenberg et al. 1991; Ichinose et al. 1990). In addition, TGases have been proposed in industrial processes for the production of modi®ed proteins by incorporation of essential amino acids or glycosyl groups (Bercovici et al. 1987; Colas et al. 1993) and formation of thermally stable gels (Motoki et al. 1987; Nonaka et al. 1989). In prokaryotes, TGase activity has been found only in actinomycetes from the genus Streptoverticillium and a transglutaminase has been isolated from Streptoverticillium S-8112 a variant of Streptoverticillium mobaraense (Ando et al. 1989; Kanaji et al. 1993). The two important features of this TGase are to be extracellular and especially Ca2+-independent. These properties increased the industrial interest for this type of enzyme. However the physiological role and the potential inducers of this bacterial transglutaminase are still unknown. The aims of this work were to elucidate the relationship between growth and TGase activity and simultaneously to improve TGase production by the use of experimental designs.

Materials and methods Bacterial strain and culture conditions Streptoverticillium cinnamoneum CBS 683.68 (Centraalbureau voor Schimmelculture, Baarn, Netherlands) was grown for about 120 h

731 at 28 °C in agitated ¯asks (140 rpm) of 250 ml containing 50 ml culture medium. Culture stock aliquots were conserved in glycerol 20% at )80 °C. The medium of culture was inoculated at 2% with a preculture of 48 h. All chemicals were purchased from Sigma. Medium A Medium A was as described by Shirling and Gottlieb (1966) and modi®ed as follows: 1 g/l K2HPO4, 1 g/l MgSO4 á 7H2O, 1 g/l NaCl, 1 g/l nitrogen source [(NH4)2SO4, agmatine, asparagine, diaminopentane, ethanolamine, gelatine, D(+)-glucosamine, L-glutamine, histidine, hydroxylamine, lysine, spermidine, spermine, urea], 10 g/l carbon source (acetic acid, citric acid, cellobiose, glucose, glycerol, lactose, maltose, mannitol, saccharose, ammonium tartrate, xylose), 1 mg/l FeSO4 á 7H2O, 1 mg/l MnSO4 á H2O, 1 mg/l ZnSO4 á 7H2O; pH 7.0±7.4. (NH4)2SO4 at 1 g/l and 10 g/l glycerol were used to test carbon sources and nitrogen sources respectively. Medium B Medium B described by Takagi et al. (1992), was modi®ed as follows: 1 g/l soya peptones, 4 g/l glycerol, 0.5 g/l MgSO4, 2 g/l KH2PO4, 5 g/l Na2HPO4, 4 g/l yeast extract, and oligoelements FeSO4 á 7H2O, ZnSO4 á 7H2O, MnSO4 á H2O: 1 mg/l each. Enzyme assay Cell-free culture medium after centrifugation was used for determination of TGase activity. Hydroxamate formation from Ncarbobenzoxy-L-glutaminylglycine (ZQG) was measured by a colorimetric procedure (Grossowicz et al. 1950) optimized for the strain used. Final concentrations of the substrates ZQG and hydroxylamine in the assay were 0.015 M and 0.1 M respectively, in 0.2 M citrate bu€er at pH 6.0 and 37 °C. As the microbial enzyme was found to be calcium-independent, assays were done in the absence of calcium. One unit was de®ned as the amount that causes the formation of 1 lmol hydroxamate/min at 37 °C. L-Glutamic acid c-monohydroxamate was used for the calibration curve. Biomass production Dry weights were determined when transglutaminase activity began to decrease (120 h): cultures were ®ltered and dried overnight at 105 °C. Glycerol assay Cell-free culture medium after centrifugation was used for determination of glycerol concentrations with a glycerol kit supplied by Boehringer. Experimental designs Factor analysis by statistical treatments was conducted with an SAS program, 6th version (Cary N.C., SAS Institute Inc., 1988). Complete factorial design The ®ve factors casein (X1), glycerol (X2), peptones (X3), yeast extract (X4) and oligoelements (X5) were studied at two levels of concentration noted )1 and +1. Level 0 (the mean concentration) was used to determine experimental error. The levels were as follows: casein [)1:5 g/l], [0:12.5 g/l], [+1:20 g/l]; glycerol [)1:0 g/l], [0:2 g/l], [+1:4 g/l]; peptones [)1:0 g/l], [0:0.5 g/l], [+1:1 g/l]; yeast extract [)1:0 g/l], [0:2 g/l], [+1:4 g/l]; oligoelements [)1:0 mg/l],

[0:0.5 mg/l], [+1:1 mg/l]. Responses were expressed as an equation of ®rst degree: Y ˆ b0 ‡ bi X i ‡ bij X ij ‡ e (with 1 £ i < j £ 5); where Y ˆ response, b0 ˆ mean value of response, bi ˆ principal effect of a factor Xi , bij ˆ interaction between two factors Xi and Xj, and e ˆ experimental error. Box-Behnken design (Box and Draper 1987) The four factors casein (X1), glycerol (X2), peptones (X3) and yeast extract (X4) were studied at the following levels of concentration: casein [)1:5 g/l], [0:17.5 g/l], [+1:30 g/l]; glycerol [)1:2 g/l], [0:16 g/l], [+1:30 g/l]; peptones [)1:1 g/l], [0:10.5 g/l], [+1:20 g/l]; yeast extract [)1:0 g/l], [0:2.5 g/l], [+1:5 g/l]. To obtain a linear model, responses were transformed with the equation: Yt ˆ …Y k ÿ 1†=…kY †kÿ1 where Yt ˆ transformed response, Y ˆ crude response, Y ˆ 1=NR log…Y †, (N ˆ number of responses), k ˆ 0:720 for TGase activity and k ˆ 0:220 for dry weights, respectively (constants chosen to obtain the greatest R2 for the considered response). Composite design (Box et al. 1954) The ®ve levels used for casein and glycerol were as follows: casein [)Ö2:26 g/l], [)1:30 g/l], [0:40 g/l], [+1:50 g/l], [+Ö2:54 g/l]; glycerol [)Ö2:6 g/l], [)1:10 g/l], [0:20 g/l], [+1:30 g/l], [+Ö2:34 g/l].

Results Choice of a medium of production S. cinnamoneum CBS 683.68 produced 0.1 U/ml TGase in medium B and the biomass rose to 0.3 g dry weight/ 50 ml culture after 5 days of culture. To improve TGase production we have studied the in¯uence of di€erent nutrients and potential inducers in the clearly de®ned medium A. Among the carbon sources tested, only glycerol and glucose allowed the strain to grow with dry weights that corresponded to 58% and 52%, respectively, of the biomass obtained in the complete medium B. Other carbon sources are much less metabolised by the bacteria; biomass did not exceed 11% (corresponding to maltose) of the biomass obtain in medium B. TGase activity was only detectable with glycerol (0.019 U/ml) and with glucose (0.015 U/ml). The nitrogen sources tested were amino acids and proteins usually employed in culture media, or compounds known to be natural (spermine, spermidine; Klein et al. 1992) or synthetic (lysine, agmatine, diaminopentane; Bercovici et al. 1987) substrates for TGase. Although dry weights reached respectively 100%, 81% and 89% of those obtained in medium B in the presence of histidine, urea and agmatine, TGase activity was only detected with ammonium sulphate (0.021 U/ml). Optimization of TGase production with the help of experimental designs Since TGase activity was very low in the de®ned medium A, we chose the complete medium B to optimize the

732

production. We tested the in¯uence of salts and observed that MgSO4, KH2PO4 and Na2HPO4 were necessary for the growth of the micro-organism but had no e€ect on TGase activity. Moreover, we noticed that there was no TGase production in the absence of peptones or yeast extract. As TGases reticulate proteins, the e€ect of di€erent proteins or extracts (bovine serum albumin, gelatine, casein, malt extract, bacto beef extract, tryptone) used in place of peptones has been tested. Medium B complemented with 20 g/l casein allowed the TGase activity to increase by 40% (0.14 U/ml). To study e€ects and interactions of the factors casein, glycerol, peptones and yeast extract on growth and TGase activity simultaneously, di€erent experimental designs were analysed in which a ®fth factor, oligoelements, was taken into account. Determination of signi®cant factors: complete factorial design Table 1 is the correspondence matrix giving all the possible combinations between the ®ve factors and the two responses obtained for each assay: TGase activity (Y1) and dry weights (Y2). Comparing the results obtained with the assays 2, 3, 4 and 32 we could deduce that casein and glycerol had a positive e€ect on both growth and TGase activity. The signi®cance of the main e€ects and interactions was analysed by means of an F-test. We could express TGase activity (Y1) and growth (Y2) as ®rst-degree polynomials: Y1 ˆ 0:068 ‡ 0:029X1 ‡ 0:047X2 ‡ 0:003X4 ‡ 0:014X1 X2 ‡ 0:003 Y2 ˆ 0:222 ‡0:045X1 ‡ 0:054X2 ‡ 0:005X4 ‡ 0:007X1 X2 ÿ 0:008X4 X1 ‡ 0:002: The statistical analysis con®rmed that casein (X1) and glycerol (X2) were highly signi®cant (at the 1% level) for both of the responses. Moreover it showed that the interaction casein/glycerol (X1X2) had a highly signi®cant e€ect on TGase activity (at the 1% level) but was less signi®cant for growth (P ˆ 0:018). These experiments showed that oligoelements (X5) were not signi®cant for TGase activity or for growth, indicating that the bacteria found enough oligoelements elsewhere. The peptones (X3) also were found not to be signi®cant in the experimental domain considered, but concentrations usually employed in actinomycete cultures can be 20 times greater than those tested here (1 g/l). To calculate the optimum concentrations that might be relevant for other experimental domains, higher concentrations than those used for level +1 were tested in a Box-Behnken design.

Determination of quadratic e€ects: Box-Behnken design The factors casein, glycerol, peptones and yeast extract were combined in pairs (Table 2). To analyse signi®cant e€ects and interactions, responses could be modelled as follows (Yt1 ˆ TGase activity; Yt2 ˆ dry weights): Yt1 ˆ ÿ 0:546 ‡ 0:061X1 ‡ 0:069X2 ÿ 0:019X3 ‡ 0:016X4 ÿ 0:04X12 ‡ 0:044X1 X2 ‡ 0:034X2 X3 ÿ 0:035X1 X4 : Yt2 ˆ ÿ 0:391 ‡ 0:104X1 ‡ 0:297X2 ÿ 0:004X3 ÿ 0:004X3 ÿ 0:044X12 ‡ 0:07X1 X2 ‡ 0:150X2 X3 : Besides the fact that casein and glycerol were con®rmed in their signi®cant e€ect, peptones and yeast extract were only found to have a signi®cant e€ect on TGase activity at the 5% level. Their concentrations were then ®xed at their mean level, 10.5 g/l for peptones Table 1 Experimental matrix and responses obtained for the complete factorial design. X1 ˆ casein, X2 ˆ glycerol, X3 ˆ peptones, X4 ˆ yeast extract, X5 ˆ oligoelements, Y1 ˆ transglutaminase activity (U/ml), Y2 ˆ dry weights (g/50 ml), ‡ ˆ ‡1 level, ÿ ˆ ÿ1 level, 0 ˆ 0 level Assay

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

Factors

Responses

X1

X2

X3

X4

X5

Y1

Y2

) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ 0 0 0 0 0

) ) ‡ ‡ ) ) ‡ ‡ ) ) ‡ ‡ ) ) ‡ ‡ ) ) ‡ ‡ ) ) ‡ ‡ ) ) ‡ ‡ ) ) ‡ ‡ 0 0 0 0 0

) ) ) ) ‡ ‡ ‡ ‡ ) ) ) ) ‡ ‡ ‡ ‡ ) ) ) ) ‡ ‡ ‡ ‡ ) ) ) ) ‡ ‡ ‡ ‡ 0 0 0 0 0

) ) ) ) ) ) ) ) ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ) ) ) ) ) ) ) ) ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ 0 0 0 0 0

) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ 0 0 0 0 0

0.018 0.022 0.056 0.141 0.006 0.020 0.009 0.152 0.010 0.024 0.090 0.156 0.006 0.051 0.087 0.176 0.011 0.023 0.069 0.143 0.000 0.040 0.083 0.193 0.007 0.054 0.084 0.156 0.000 0.052 0.096 0.160 0.100 0.087 0.094 0.087 0.085

0.088 0.138 0.158 0.315 0.072 0.157 0.185 0.294 0.102 0.157 0.215 0.275 0.101 0.185 0.222 0.264 0.079 0.171 0.145 0.298 0.089 0.163 0.170 0.316 0.095 0.178 0.185 0.268 0.096 0.179 0.198 0.291 0.233 0.219 0.241 0.212 0.208

733 Table 2 Experimental matrix and responses obtained for the BoxBehnken design. X1 ˆ casein, X2 ˆ glycerol, X3 ˆ peptones, X4 ˆ yeast extract, Y1 ˆ transglutaminase activity (U/ml), Y2 ˆ dry weights (g/50 ml) Assay

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Factors

Responses

X1

X2

X3

X4

Y1

Y2

) ‡ ) ‡ ) ‡ ) ‡ ) ‡ ) ‡ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

) ) ‡ ‡ 0 0 0 0 0 0 0 0 ) ‡ ) ‡ ) ‡ ) ‡ 0 0 0 0 0 0 0 0 0 0

0 0 0 0 ) ) ‡ ‡ 0 0 0 0 ) ) ‡ ‡ 0 0 0 0 ) ‡ ) ‡ 0 0 0 0 0 0

0 0 0 0 0 0 0 0 ) ) ‡ ‡ 0 0 0 0 ) ) ‡ ‡ ) ) ‡ ‡ 0 0 0 0 0 0

0.030 0.073 0.068 0.336 0.085 0.232 0.024 0.111 0.007 0.157 0.119 0.155 0.121 0.157 0.048 0.227 0.060 0.221 0.048 0.262 0.133 0.125 0.165 0.144 0.156 0.160 0.155 0.157 0.163 0.174

0.165 0.237 0.420 1.042 0.265 0.548 0.343 0.461 0.304 0.483 0.315 0.441 0.284 0.627 0.098 1.063 0.210 1.014 0.187 1.033 0.418 0.463 0.466 0.451 0.413 0.387 0.423 0.383 0.454 0.438

and 2.5 g/l for yeast extract. Under these conditions, plots of isoresponse curves, showing Yt as a function of casein and glycerol, revealed that optimum concentrations for casein and glycerol were higher than level +1 concentrations (results not shown). A composite design was then analysed to determine these optima. Optimization: composite design To determine the experimental domain, we studied the in¯uence of each factor independently. When concentrations were tested up to 50 g/l, a maximum of TGase activity was obtained at 40 g/l for casein and 20 g/l for glycerol. The experimental domains were chosen around these two values. The matrix and responses are listed in Table 3. In the experimental domain tested, casein was no longer a signi®cant factor. However an optimum for TGase activity of 0.294 U/ml was calculated for a glycerol concentration of 31.2 g/l (level +1.12) and a casein concentration of 38.4 g/l (level )0.16). These values were de®ned by the equation: Y ˆ 0:230 ÿ 0:006X1 ‡ 0:079X2 ÿ 0:011X12 ÿ 0:025X22 : In control experiments, in the optimum medium also containing peptones (10.5 g/l), yeast extract (2.5 g/l),

Table 3 Experimental matrix and responses obtained for the composite design. X1 ˆ casein, X2 ˆ glycerol, Y ˆ transgluta minase activity (U/ml) Assay

1 2 3 4 5 6 7 8 9 10 11 12 13

Factors X1

X2

) + ) +p ) 2 0p 2 0 0 0 0 0 0

) ) + + 0 p ) 2 0  p 2 0 0 0 0 0

Response Y 0.109 0.113 0.264 0.271 0.234 0.070 0.190 0.300 0.238 0.232 0.243 0.212 0.228

Fig. 1 Transglutaminase (TGase) production in the optimized medium. d TGase activity (U/ml), u glycerol (g/l), s log (dry weight) (g/50 ml)

MgSO4 (0.5 g/l), KH2PO4 (2 g/l) and Na2HPO4 (5 g/l), we obtained a TGase activity of 0.331 ‹ 0.038 U/ml. This maximum activity was reached when the medium became de®cient in glycerol (Fig. 1).

Discussion The successive experimental designs showed that the two factors, casein and glycerol, had a predominant e€ect on both growth and TGase activity denoting a correlation between these two responses. However, biomass production alone was not sucient for TGase production, as in a synthetic medium, where histidine, urea or agmatine was the nitrogen source, S. cinnamoneum grew but did not produce transglutaminase. The presence of casein seemed to be essential to the improvement of TGase production. As casein hydrolysate had no e€ect, crude casein may not have a solely nutritional role. This protein, denatured, may be considered as a good substrate for proteases (Davies 1987). Thus we suggest that

734

casein could protect transglutaminase from degradation by extracellular proteases. This does not exclude the possibility that casein could also be a good substrate for TGase, acting as an inducer. The fact that transglutaminase production could be induced was stressed by the time for the maximum enzyme activity to appear during the stationary growth phase (Fig. 1) when glycerol was absent from the medium. Furthermore, in conditions where transglutaminase production was low, the use of successive experimental designs allowed the level of activity obtained in a complete culture medium to be increased by a factor of three. Acknowledgements Magali Junqua was supported by a grant from the Groupement de Recherche de Lacq (Elf group, France). The authors thank Idir Merchermek (University of Pau, France) for assistance in statistical analysis and Pr. Collombier (University of Pau, France) for helpful advice.

References Ando H, Adashi M, Umeda K, Matsuura A, Nonaka M, Uchio R, Tanaka H, Motoki M (1989) Puri®cation and characteristics of a novel transglutaminase derived from micro-organisms. Agric Biol Chem 53: 2613±2617 Bercovici D, Gaertner HF, Puigserver AJ (1987) Transglutaminasecatalysed incorporation of lysine oligomers into casein. J Agric Food Chem 35: 301±304 Box GEP, Draper NR (1987) Empirical model building and responses surfaces. Wiley, London Box GEP, Connor LR, Cousins WR, Davies OL, Nimsworth FR, Sillitto GP (1954) In: Davis OL (ed) The design analysis of industrial experiments. Oliver & Boyd, London

Colas B, Caer D, Fournier E (1993) Transglutaminase-catalyzed glycosylation of vegetable proteins. E€ect on solubility of pea legumin and wheat gliadins. J Agric Food Chem 41: 1811± 1815 Davies KJA (1987) Protein damage and degradation by oxygen radicals. I. General aspects. J Biol Chem 262: 9895±9901 Folk JE, Finnlayson J (1977) The e-(c-glutamyl)lysine crosslink and the catalytic role of transglutaminases. Adv Protein Chem 31: 1±133 Friedrich P, Aszodi A (1992) Calcium-dependant proteolysis and isopeptide bond formation: calpains and transglutaminases. Pure Appl Chem 64: 1093±1097 Greenberg CS, Birckbichler PJ, Rice RH (1991) Transglutaminase: multifunctional cross-linking enzymes that stabilize tissues, FASEB J 5: 3071±3077 Grossowicz N, Wainfan E, Borek E, Waelsch H (1950) The enzymatic formation of hydroxamic acids from glutamine and asparagine. J Biol Chem 187: 111±125 Ichinose A, Bottenus RE, Davie EW (1990) Structure of transglutaminases. J Biol Chem 265: 13 411±13 414 Kanaji T, Ozaki H, Takao T, Hawajini H, Ide H, Motoki M, Shimoniki Y (1993) Primary structure of microbial transglutaminase from Streptoverticillium sp. strain S-8112. J Biol Chem 268: 11 565±11 572 Klein JD, Guzman E, Kuehn GD (1992) Puri®cation and partial characterization of transglutaminase from Physarum polycephalum. J Bacteriol 174: 2599±2605 Motoki M, Aso H, Seguro K, Nio N (1987) aS1-Casein ®lm prepared using transglutaminase. Agric Biol Chem 51: 993±996 Nonaka M, Tanaka H, Okiyama A, Motoki M, Ando H, Umeda K, Matsuura K (1989) Polymerization of several proteins by Ca2+-independant transglutaminase derived from micro-organisms. Agric Biol Chem 53: 2619±2623 Shirling EB, Gottlieb D (1966) Methods for characterization of Streptomyces species. Int J Syst Bacteriol 16: 313±340 Takagi H, Arafuka S, Matsui H, Washizu K, Ando K, Koikeda S (1992) Recombinant transglutaminase. European patent EP 0 481 504 A1