Cloning and expression of cellulase and xylanase genes in Lactobacillus plantarum

AppI Microbiol Biotechnol (1990)33:534-541

Applied Microbiology Biotechnology © Springer-Verlag 1990

Cloning and expression of cellulase and xylanase genes in Lactobacillus plantarum Trees Scheirlinck 1, Jan De Meutter 1, Greta Arnaut 1, Henk Joos 1, Marc Claeyssens 2, and Frank Michiels ~ 1 Plant Genetic Systems NV, Plateaustraat 22, B-9000 Ghent, Belgium z Laboratory for Biochemistry, State University Ghent, Ledeganckstraat 35, B-9000 Ghent, Belgium Received 27 December 1989/Accepted 23 April 1990

Summary. Eleven cellulase genes from Gram-positive bacteria were cloned in a Lactobacillus plantarum silage inoculum. Eight o f these genes were expressed as active enzymes from their original promotors and translation signals. Where tested, the enzymes produced by transformed L. plantarum had the same temperature and p H optimum as enzymes produced in the original host, or in transformed Escherichia coli. Using chloramphenicol acetyltransferase as a cell-internal marker enzyme, it could be demonstrated that at least endoglucanase D from Clostridium thermocellum was actively secreted by transformed L. plantarum. In growing L. plantarum cultures, most o f the enzymes were irreversibly inactivated when the p H decreased below 4.5. If the transformed strains were to be applied as an inoculum in silage, this p H inactivation might be useful in preventing overdigestion o f the crop fibre.

Introduction In ensiled crops, lactic acid bacteria convert low molecular weight carbohydrates into lactic acid, which is the main preservative in ensilage. In some crops it is important to add lactic acid bacteria as an inoculum during ensiling, in order to boost fermentation in the initial stages (Woolford 1984). Lactobacillus plantarum is generally preferred for this purpose (Seale 1986). In crops such as grass or lucerne, the low amount of fermentable carbohydrate limits lactic fermentation, which leads to unreliable preservation of the silage. This can be artificially improved by adding sources o f fermentable carbohydrates, or enzymes which release the latter from, for example, plant fibre. Addition o f cellulase to grass or lucerne silage significantly improves its preservation (Seale 1987). Although cellulase is sold and used as a commercial silage additive, its application is rather expensive. Therefore, the introduc-

Offprint requests to: F. Michiels

tion of cellulase activity in L. plantarum could be an interesting and practical application o f genetic engineering. Genetic manipulation of L. plantarum becomes possible through the establishment of transformation techniques (Josson et al. 1989; Luchansky et al. 1988). As shown previously, the celA and celE endoglucanase genes from Clostridium thermocellum can be functionally expressed in L. plantarum (Scheirlinck et al. 1989; Bates et al. 1989). The cloning and expression of several other cellulase and xylanase genes in L. plantarum is presently reported. Most of these genes produce active enzymes, which are secreted into the culture supernatant. We have analysed the stability of different genetic constructs for expressing heterologous genes in Lactobacilli, and the properties and yield of the enzymes produced.

Materials and methods Genetic constructions, transformations and 9rowth conditions The bacterial strains and plasmids used are described in Table 1. Plasmid constructions were made using standard cloning techniques (Maniatis et al. 1982), transformed in Escherichia coli MC1061, and selected on Luria-Bertani (LB) agar (Maniatis et al. 1982) containing 100 Ixg/ml erythromycin (Em), 25 ~tg/ml tetracycline (Tc) or 25 ~tg/ml chloramphenicol (Cm). Plasmids expressing the desired enzyme activity in E. coli were electroporated into L. plantarum 80 (LpSO)as described previously (Josson et al. 1989). Transformed Lp80 was grown on MRS agar (Oxoid, Basingstoke, Hampshire, UK) containing 10 ~tg/ ml Em and 100 txg/ml lincomycin, or 10 Jxg/ml Cm. For determining cellulase or xylanase activity, Lp80 transformants were grown selectively on MRS broth supplemented with 20 g/1 of CaCO3 (initial pH 6.8).

Detection of eellulase and xylanase activities In E. coli transformants. Endoglucanase (EG) and xylanase activity was tested as described by Cornet et al. (1983) using an overlay

535 Table 1. Strains and plasmids used in this study Strain or plasmid

Genetic markers or description

Origin or reference

Escherichia coli MC1061

Casadaban and Cohen (1980)

Lactobacillus plantarum 80 (LpSO) (grass silage starter)

Josson et al. (1989)

A: Plasmids used in genetic constructions pWS11 contains Clostridium thermocellum celA gene encoding endoglucanase EGA pCT303 contains C. thermocellum celC gene encoding endoglucanase EGC pCT510 contains C. thermocellum celF gene encoding endoglucanase EGF pCT600 contains C. thermocellum celD gene encoding endoglucanase EGD pCT700 contains C. thermocellum celG gene encoding endoglucanase EGG pCT1202 contains C. thermocellum xynZ gene encoding xylanase Z pCT404 contains C. thermocellum ~-glucosidase gene pHZll7 contains C. acetobutylicurn endoglucanase gene pHZ302 contains C. acetobutylicurn xylanase gene pES420 contains Butyrivibrio fibrisolvens end1 gene encoding an endoglucanase pES510 contains B. fibrisolvens end2 gene encoding an endoglucanase pSA3 10.2-kb shuttle vector carrying Em r pAM401 10.4-kb shuttle vector carrying C m r, Tc r pWP37 4.2-kb shuttle vector carrying Em r pERM3.2 7.6-kb shuttle vector carrying Em r

Schwarz et al. (1986) Gift from Dr. P. Brguin Gift from Dr. P. Brguin Hazlewood et al. (1988) Millet et al. (1985) Gift from Dr. P. Brguin Gift from Dr. P. Brguin Zappe (1988) Zappe et al. (1987) Berger et al. (1989) Gift from Mrs. E. Berger Dao and Ferretti (1985) Wirth et al. (1987) MahiUon et al. (1989) Josson et al. (1989)

B: Recombinant plasmids described in this work

pST11 pST12 pWT303 pWT510 pWT512 pWT600 pWT602 pAT600 pWT700 pWT702 pWT1202 pWT404 pWZ117 pWZ302 pWES420 pWES510

3.3-kb 3.3-kb 2.6-kb 3.3-kb 3.3-kb 6.3-kb 6.3-kb 6.3-kb 4.0-kb 4.0-kb 5.7-kb 4.0-kb 1.7-kb 3.0-kb 2.8-kb 3.7-kb

HindlII/Klenow fragment of pWSll in pSA3-EcoRV HindlII/Klenow fragment of pWSll in pSA3-EcoRV EcoRI-PstI fragment of pCT303 in pWP37-EcoRI-PstI EcoRV fragment of pCT510 in pWP37-EcoRV EcoRV fragment of pCT510 in pWP37-EcoRV EcoRI fragment of pCT600 in pWP37-EcoRI EcoRI fragment of pCT600 in pWP37-EcoRI EcoRI fragment of pCT600 in pAM401-EcoRI HindlII fragment of pCT700 in pWP37-HindlII HindlII fragment of pCT700 in pWP37-HindlII EcoRI fragment of pCT1202 in pWP37-EcoRI EcoRI fragment of pCT404 in pWP37-EcoRI EcoRI-PstI fragment of p H Z l l 7 in pWP37-EcoRI-PstI EcoRI fragment of pHZ302 in pWP37-EcoRI HindlII fragment of pES420 in pWP37-HindlII AccI-SspI/Klenow fragment of pES510 in pWP37-EcoRV

Orientation +

+ + + 0 0 + 0 + + + +

+, having the inserted gene in the same orientation as the Gram-positive selectable marker gene; - , having the inserted gene in the opposite orientation as the Gram-positive selectable marker gene; 0, the orientation of the inserted gene is not known; Era, erythromycin; Cm, chloramphenicol; Tc, tetracycline

agar containing 0.1% carboxymethylcellulose (CMC) (Fluka, Buchs, Switzerland) and 0.01% xylan, respectively. For detection offl-glucosidase activity, p-nitrophenyl-fl-D-glucoside (pNPG) was used at a final concentration of 1 mM in the overlay agar. After 2 h incubation at 60 ° C pNPGase-producing E. coli transformants were surrounded by a yellow halo.

In L. plantarum transformants. EG, xylanase and pNPGase phenotypes were determined by plating the Lp80 transformants on buffered MRS plates (0.2M 2-[N-morpholino]ethanesulphonic acid, MES, pH 6.8) containing 0.1% CMC, 0.01% xylan, or 1 mM pNPG respectively. The plates were incubated and stained as described for E. coli.

E n z y m e assays Well-plate test. Both EG and xylanase activities were tested in wells made in buffered H~O-agar plates (0.2 M Na phosphate, pH 6.0) containing 0.1% CMC or 0.01% xylan. Concentrated supernatant (100 ~tl) was added per well and the plates were left overnight at room temperature to allow diffusion. Enzyme activities were

revealed after incubation for 1 h at 37 ° C or 60° C and Congo red staining.

Reducing sugar test. A mixture of 1 ml of 1.0% CMC or 0.01% xylan in 0.2 M Na phosphate, pH 6.0, was supplemented with 50 lxl concentrated supernatant and, after incubation for 2 h at 37 ° C (the EG and xylanase of C. acetobutylicum ; EG 1 and 2 of Butyrivibriofibrisolvens) or 60°C (the enzymes from C. thermocellum), the concentration of reducing sugars (as glucose equivalent) in 500 Ixl aliquots was determined by the Somogyi-Nelson method (Nelson 1952). 2'-Chloro, 4'-nitrophenyl-fl-D-cellobioside (CNPC) as a chromogenic substrate. Culture supernatant, concentrated if necessary, was tested for CNPCase activity by mixing 100 lxl of 0.5 mM CNPC (Claeyssens 1988) in 0.2 M Na phosphate (pH 6.0) with 20 lxl enzyme solution in a microtitre plate cup and incubating for 1 h at 37°C or 60° C. The reaction was stopped by adding 100 Ixl of 1 M TRIS base (pH 9.8) and the optical density was measured at 405 nm (OD405). One unit (U) of CNPCase activity of the purified C. acetobutylicum EG was defined as the amount that releases 1 Ixmol 2-chloro, 4-nitrophenol/min at 37 ° C.

536

Preparation of crude L. plantarum transformant cellulase Supernatant (2.0 ml) of transformed Lp80grown on MRS medium supplemented with 20 g/1 CaCO3 was concentrated tenfold (Centricon-10, Amicon, Danvers, Mass., USA). The concentrate (200 lxl) was washed twice with 50 mM Na-phosphate (pH 7.0) and the 200-~tl retentate was designated the concentrated supernatant.

Purification of the C. acetobutylicum endoolucanase Lp80 (pWZ117) was grown to saturation on MRS medium supplemented with 20 g/1 CaCO3. Supernatant (1 1) was concentrated by ultrafiltration (PM-10 membrane, Amicon) to a final volume of 8 ml. The protein fraction precipitating between 30% and 50% ammonium sulphate and representing > 90% of the EG activity, was resuspended in 4 ml Na phosphate (0.2 M; pH 6.0). After desalting and buffer change to 10 mM TRIS/MES, pH 7.5, (PD-10 column, Pharmacia, Uppsala, Sweden), the protein was loaded on a DE52 column (Whatman, Springfield, USA) and a KC1 gradient (0-150 mM) was applied. The EG eluted in two peaks around 100 mM KC1. Fractions showing EG activity were pooled, dialysed against 10 mM TRIS/HC1, pH 8.5, loaded on a Mono Q column (Pharmacia, Uppsala, Sweden) and eluted with a gradient of 0-500 mM KC1. Two EG peaks (eluting at 85 and 107 mM KC1) were again recovered and the fraction containing the bulk of enzyme activity (eluting at 85 mM KC1) was retained as purified endoglucanase.

Lysis of L. plantarum Lp80 was harvested in the early exponential growth phase (OD60o 0.5-0.7). Culture fluid (1.0 ml) was centrifuged and cells washed once with water and resuspended in 180 Ixl of 250 mM TRIS/HC150 mM ethylenediaminetetraacetate (EDTA), pH 8.2. Lysozyme and mutanolysine were added to a final concentration of 1.25 rag/ ml and 125 U/ml respectively, and the mixture incubated for 45 min at 37° C. Lysis of samples was completed, as checked microscopically, by twice freezing ( - 70° C) and thawing (37° C) followed by the addition of 800 lxl of 250 mM TRIS/HC1-0.5 mM Phenylmethylsulfonyl fluoride (PMSF), pH 8.0. After centrifugation, the supematant was retained as the lysate and the resuspended pellet (1 ml of 200 mM Na phosphate, pH 6.0) was recovered as the cell-associated fraction.

Chloramphenicol acetyl transferase (CA T) assay Culture supernatant or lysate (50 lxl) was mixed with 150 lxl Cm (1 mM), and 10 ~1 labelled butyryl coenzyme A (NEC-801; 10 lxCi/ml [New England Nuclear, NEN, Boston, Mass, USA). The mixture was overlaid with 5 ml Econfluor (NEF-941, New England Nuclear, Boston, Mass, USA). CAT activity was expressed as the increase in cpm over a period of 30 min (Neumann et al. 1987). Results

Cloninff of cellulase and xylanase genes into L. plantarum The celA of C. thermocellum has previously been cloned in the shuttle vector pSA3, and the recombinant vectors p S T l l and pST12 have been electroporated into Lp80 (Scheirlinck et al. 1989). For cloning o f the endogluca-

nase genes ceIC, celD, celF, ceIG, the xylanase gene xynZ and a fl-glucosidase gene of C. thermocellum, as well as an endoglucanase and a xylanase gene of C. acetobutylicum, and two endoglucanase genes (end1 and end2) from B. fibrisolvens, the shuttle vector pWP37 was used (Table 1). pWP37 was chosen because of its small size (4.2 kb), the presence of a polylinker, its ability to replicate in E. coli, and its high efficiency in Lp80 electroporation (105 transformants/Ixg DNA). Details of the cloning procedures are available on request. Electroporation efficiencies of the r e c o m b i n a n t plasmids into Lp80 were c o m p a r a b l e to those found for pWP37 (data not shown), except for pWT600 and pWT602 (containing celD), which failed to produce any E m r Lp80 plantarum transformants. The 6.3-kb fragment carrying the celD gene was subsequently cloned in the shuttle vectors pERM3.2, pVA838 (Macrina et al., 1982), and pAM401. From the cloning in E. coli, the ceID gene was obtained in two orientations in p E R M 3 . 2 and pVA838, and in one orientation in pAM401. All five celD constructions exhibited C M C a s e activity in E. coli. However, Lp80 E m r colonies were only obtained after electroporation with pAT600 containing celD in pAM401. As in B. subtilis (Gruss and Ehrlich 1989), large D N A fragments m a y frequently undergo deletions and recombinations in Lp80 when they are cloned on plasmids replicating via a single strand intermediate, such as pWP37, pERM3.2, or pVA838. The use of pAM401, which does not replicate via a single-stranded intermediate, m a y obviate this problem.

Expression of heterologous cellulases and xylanases produced in L. plantarum 80 The eleven cellulase and xylanase genes were tested for enzyme expression in transformed Lp80. Plate tests for C M C a s e were positive in colonies of Lp80 t r a n s f o r m e d with celA, celD, ceIF, ceIG, for the endoglucanase of C. acetobutylicum, and endl of B. fibrisolvens (Fig. 1). Xylanase activity was observed in Lp80 transformed with xynZ as well as with the xylanase gene of C. aeetobutylicum (not shown). In Lp80 transformed with ceIC, or end2, or with the C. thermocellum fl-glucosidase gene, C M C a s e (or p N P G a s e ) expression could not be detected (Fig. 1). This could be due to the heterologous gene itself: D N A stability, transcription, translation, processing or secretion level, or expression in a specific growth stage o f the cell. However, even if no activity is detected, it does not necessarily m e a n that no active enzyme is produced. Because of the strict substrate specificity of the enzymes, activity m a y only be detected using the appropriate substrates under their specific assay conditions (pH, temperature), as demonstrated below. Both p S T l l and pST12, carrying the celA gene in different cloning orientations in pSA3, gave rise to eellulase-positive Lp80 transformants. Although both cloning orientations of ceIF (pWT510 and pWT512) express C M C a s e in E. coli, and both transform Lp80 to

537

However, in a quantitative test measuring cellulase and xylanase activities as the release of reducing sugars from CMC or xylan, activity was only detected in the concentrated supernatant of Lp80 transformed with pSTll, pAT600 and pWZll7 (respectively 1200, 550, and 2200 ~tg glucose equivalents/ml supernatant per hour). For the other cellulases and the two xylanases the expression level was probably too low to measure any activity by means of the reducing sugar test. Activities of C. thermocellum EGC and EGD can be determined with the chromogenic substrate CNPC (Claeyssens 1988). As shown in the present study, activity of the C. acetobutylicum EG can also be followed with the same substrate, thus representing an elegant alternative for CMC. The easy dosage of the reaction product 2-chloro, 4-nitrophenol (pK = 5.5, •405 n m = 9 0 0 0 M -1 cm -a at pH 6.0) allows for rapid and reliable activity measurements as exemplified below. The EG of C. acetobutylicum produced by Lp80 (pWZ117) was purified from culture supernatant as described in Materials and methods. In the last fractionation step (Mono Q ion exchange), two separate fractions containing endoglucanase activity were eluted (Fig. 2). The fraction eluting at the lowest salt concentration (85 mM KC1) and containing the bulk of activity was considered to be pure enzyme.

pH and temperature optima of some cellulases produced by L. plantarum transformants In order to detect possible abnormalities in the enzymes produced by transformed LpSO, the properties of three enzymes (the EG of C. acetobutylicum, EGA and EGD of C. thermocellum) expressed either in E. coli or Lp80 were compared. Only activities on the same substrate can be compared, since the pH optimum and pH Fig. 1. Carboxymethylcellulase (CMCase) and xylanase activity of Lactobacillus plantarum 80 transformed with different plasmids: 1, pWP37; 2, p S T l l ; 3, pST12; 4, pWT303; 5, p W Z l l 7 ; 6, pWES420; 7, pWES510; 8, pWT510; 9, pWT512; 10, pWT700; 11, pWT702. Enzyme activity is apparent as a halo in the Congo red staining around the colonies

Em r, celF expressed CMCase in Lp80 only when the gene was cloned in the same transcriptional orientation (pWT510) as the Em ~gene of the vector. The absence of expression in the opposite cloning direction may be due to inhibition of the cellulase transcription by the Em r gene transcript. Absence of expression of ceIC and the C. thermoeellum fl-glucosidase gene, and of one cloning orientation of eelG (pWT700) may be caused by the same mechanism. The test for cellulase and xylanase activity on plates is more sensitive than the reducing sugar test on liquid cultures, but unfortunately, it does not allow quantitative measurements. In a well-plate test, cellulase and xylanase activity was found in the concentrated supernatant of Lp80 transformed with pSTll, pST12, pAT600, pWES420, pWZll7, pWZ302 and pWT1202.

o 0:3 od

.06

300

.04

200

J <

iO0 .<

J 0 0

J

\

J

J ,

t0

i

20 FRACTION NUM8ER

i

3O

0

Fig. 2. Absorbance at 280 nm of the eluate of a Mono Q column loaded with fractionated supernatant of L. plantarum (pWZ117). Fractions exhibiting enzyme activity are indicated with an arrow: dashed line, 0-500 m M KC1 gradient

538 ftlO

lfLO

A

100 90

90

80

80

70

70

60

60

50

50

40

40

30

30

20

20

10

:tO

0 3,0

A

:tOO

i

3.5

4.0

4.5

5.0

5.5

6.0 PR

6.5

7.0

7.5

8.0

3.0

8.5

I~0

4.0

"4.5

5.0

i

r

r

5.5

6.0 ,OH

50

55

6.5

7.0

7.5

8.0

8.5

7

80

85

ftlO

B

100:

3.5

B

100

90

90

80

80

70

70

6O

~

s0

~ .<

5o

40

~

4o

a~ 30 20

2o

t0 0 15

1o r

r

i

I

r

I

20

25

30

35

40

45

I

r

I

I

I

I

r

I

50

55

60

G5

70

75

80

85

90

TEMPERATURE

20

25

30

35

40

45

60

65

7

TEMPERATURE

Fig. 3. A pH and B temperature activity curves of cellulase prepared from L. plantarum transformed with pAT600 (A), pWZll7 (O), and pST11 (*). Enzyme produced from pAT600 and pWZ117 is measured as 2'-chloro,4'-nitrophenyl-fl-cellobiosidase (CNPCase), enzyme from pST11 is measured as CMCase

Fig. 4. A pH activity curves of endoglucanase D propared from L. plantarum (pWZll7) using carboxymethylcellulose (CMC) (~) and 2'-chloro,4'-nitrophenyl-fl-D-cellobioside (CNPC) ( 0 ) as substrates. B Temperature activity curves of cellulase prepared from L. flantarum (pAT600) using CMC (ZX) and CNPC (A) as substrates

stability may be different on different substrates, as reported for EGC (P6tr6 et al. 1986). The CMCase activity of EGA and E G D produced by Lp80 transformants (pST11 in Fig. 3A, pAT600 in Fig. 4B), had the same pH and temperature profile as reported for the enzymes produced in E. coli and B. subtilis transformants (Schwarz et al. 1986; Joliff et al. 1986; Soutschek-Bauer and Staudenbauer 1987). Also the CMCase activity of the EG of C. acetobutylicum produced by Lp80 (pWZ117 in Fig. 4A) had the same pH profile as reported for the enzyme produced in E. coli transformants (Zappe 1988). Since at least for EGA the enzyme produced in E. coli has the same properties as the enzyme produced by C. therrnocellum (B6guin et al. 1988), endoglucanase processing in Lp80 is most probably the same as in E. coli and C. thermocellum. When C N P C was used as substrate instead of CMC, the pH range of activity of the E G D preparation from LpSO (pAT600) (Fig. 3A) was more narrow than the one on CMC (not shown), and its temperature optimum was 5°C lower (Fig. 4B). The C. acetobutylicum EG produced by Lp80 (pWZ117) showed a CNPCase

temperature optimum at 45 ° C and more than 75% of its maximum activity was retained between 3 5 ° C and 55°C (Fig. 3B). These temperatures are about 5°C lower than reported for the CMCase activity produced in E. coli (Zappe 1988). Comparison of activity on C N P C versus CMC indicates that C N P C is less efficiently converted at low pH (Fig. 4A) and high temperature (Fig. 4B). The enhanced temperature stability for CMC can at least partially be due to a stabilization of enzyme conformation when this high molecular weight substrate is present.

pH inactivation of some cellulases produced by L. plantarum transformants Initially, the yield of active enzyme from cultures of transformed Lp80 was highly variable. Since enzyme production was reproducible in pH-buffered cultures, this was probably due to the irreversible inactivation of the cellulases by the low pH in saturated L. plantarum cultures.

539 The p H inactivation curve of the E G of C. acetobutylicum and E G D of C. thermocellum was determined between p H 3.5 and 6.5 and measured as the residual C N P C a s e activity at optimal pH, after incubation at a given p H for several time periods. At p H 4.0, the C. acetobutylicum E G was immediately and totally inactivated, whereas at p H 4.5 a decrease of activity was only observed after 5 h incubation (Table 2A). This corresponds to the p H inactivation curve of the enzyme prepared from E. coli (Zappe et al. 1986). For EGD, incubation at p H 4.0 resulted in a slow decrease o f enzyme activity, however maximum activity was retained after incubtion for 5 h at p H 4.5 (Table 2A). In the supernatant of L. plantarum transformed with celA, xynZ, end1 and the xylanase of C. acetobutylicum, enzyme activity could not be measured quantitatively by the reducing sugar test (see above). Therefore the well-plate assay was used to test the p H inactivation of these enzymes (Table 2B).

Table 2. Irreversible inactivation of cellulase and xylanase by low pH A Incubation pH Time (min) 30

60

Time (min) 120 600 30

Lp80 (pAT600) 6.0 5.5 5.0 4.5 4.0 3.5

100 100 100 100 60 0

100 100 100 100 40 0

100 100 100 100 25 0

60

120

600

Lp80 (pWZ117) 100 100 100 100 0 0

100 100 100 100 0 0

100 100 100 100 0 0

100 100 100 100 0 0

100 100 100 60 0 0

A. Cellulase activity measured on 2'-chloro,4'-nitrophenyl-fl-Dcellobioside (CNPC). Residual CNPCase activity is expressed as the percentage of the untreated control (assay conditions are described in Materials and methods) B

Incubation pH Time (min) 5

30

Time (min) 60

120 5

Lp80 (pST11) 6.5 5.0 4.0 3.5

+ + + .

+ + + .

+ + . .

+ + +

+ + + .

+ + .

. .

60

120

Lp80 (pWES420) + + +/.

Lp80 (pWT1202) 6.5 5.0 4.0 3.5

30

+ + .

.

+ + -

+ +/-

.

Lp80 (pWZ302) + +

.

+ + +/-

+ + + .

+ + + .

+ + +/-

+ + +/-

.

B. Cellulase and xylanase activity determined in wells on carboxymethylcellulose (CMC) and xylan plates, respectively (see Materials and methods): +, no detectable decrease of activity compared to the untreated control; + / - , decreased residual enzyme activity; - , no detectable residual activity

Localization of the cloned cellulolytic enzymes produced by the L. plantarum transformants For most of the transformed L. plantarum strains that produced cellulase, enzyme activity was detected in the (concentrated) culture supernatant. To test whether this is due to a specific secretion rather than cell leakage or lysis, we investigated whether CAT (encoded, for example, on the vector part of pAT600) can be used as an intracellular marker enzyme in L. plantarum: Lp80 (pAT600) cells in the exponential growth phase were lysed, and supernatant, lysate and the cell-associated fraction were tested separately for C N P C a s e and CAT activities. The lysate of Lp80 (pAT600) contained 97% o f the C A T activity, vs 3% in the supernatant. The cell b o u n d fraction did not show any CAT activity. Nevertheless, in a control experiment, CAT activity from the lysate was completely maintained after incubation in the pHbuffered supernatant for 3 h, demonstrating its stability. The E G activity in Lp80 (pAT600) was 90% extracellular; 10% was retained in the cell debris fraction and no intracellular cellulase activity could be detected. This proves that the presence of E G D in the supernatant of Lp80 (pAT600) cultures is due to active secretion, rather than cell lysis. Since the other cellulases and xylanases were cloned into pWP37 (which does not express CAT), we had no cell-internal marker enzyme to prove secretion in the Lp80 transformants. Nevertheless, one can presume that these enzymes are secreted as well, since they are found in the supernatant of the L. plantarum transformants. Since C A T activity can be measured easily and quantitatively, we believe that C A T can be generally used as an intracellular marker enzyme for L. plantarum, and other bacteria. The main limiting factor may be the lysis of the host without the use of a detergent in order to avoid denaturation of the relevant enzymes. The lysis method used in this study was only successful on Lp80 cultures in the early exponential growth phase. Thus it cannot be used for localization of enzymes which are not or are poorly expressed in that growing stage.

Discussion By cloning several cellulase genes, we have demonstrated that L. plantarum can actively produce and secrete heterologous enzymes from a variety of Gram-positive bacteria. It is in this respect comparable to B. subtilis as a host for enzyme production. The amount of enzyme produced by the L. plantarum transformants was determined quantitatively for the endoglucanase (EG) of C. acetobutylicum, and E G A and E G D of C. thermocellum. After purification of the C. acetobutylicum E G out of the L. plantarum transformant, a specific activity of 1800 U / m g C N P C a s e was obtained. From this, we calculated that a saturated cul-

540 ture o f the L. plantarum t r a n s f o r m a n t p r o d u c e d a b o u t 10 ~tg E G / m l supernatant. A s s u m i n g a specific activity o f 140 U / m g for E G A ( C M C a s e activity o f e n z y m e purified f r o m E. coli t r a n s f o r m e d with celA [B6guin et al. 1988]), a saturated culture o f Lp80 (pST11) w o u l d contain 0.8 ktg E G A / m l . F o r E G D , a s s u m i n g a specific activity o f 428 U / m g ( C M C a s e activity o f the e n z y m e f r o m E. coli [B6guin et al. 1988]), an E G p r o d u c t i o n o f 0.12 lxg/ml was o b t a i n e d in the Lp80 (pAT600) culture. This illustrates that the yield for m o s t cellulases is very low. I n o r d e r to p r o d u c e useful a m o u n t s o f cellulases in L. plantarum, the expression level has to be raised, p r o b a b l y b y using specific a n d stronger promoters. T r a n s f o r m a t i o n with a synergistic c o m b i n a t i o n o f cellulase a n d / o r xylanase genes m a y enable L. plantarum to release f e r m e n t a b l e c a r b o h y d r a t e s f r o m ensiled crops, a n d t h e r e b y i m p r o v e silage quality. T h e r e m a r k able irreversible inactivation o f all tested cellulases a n d xylanases a r o u n d p H 4.0 c o u l d be very i m p o r t a n t for their a p p l i c a t i o n in silage. Firstly, the c o m b i n a t i o n o f a cellulase such as E G D having high activity at l o w e r t e m p e r a t u r e s a n d higher p H (conditions o f the early stage in silage), with E G A , being m o r e active a n d stable at lower p H , m a y contribute to a r a p i d acidification a n d a s u p p r e s s i o n o f clostridial g r o w t h in silage. O n the o t h e r h a n d , the p H sensitivity makes it unlikely that cellulase activity will persist a n d overdigest the silage, o n c e a p H l o w e r t h a n 4.0 has b e e n reached. Acknowledgements. We are grateful to Mrs. Eldie Berger, Dr. Harold Zappe and Dr. David Woods (University of Cape Town, South Africa) for providing the cloned genes from C. acetobutylicum and B. fibrisolvens, for communicating unpublished results, for advice in this project and for critically reading the manuscript. We thank Dr. Don Clewell (University of Michigan, Ann Arbor, Mich.) for sending pAM401, and Dr. Pierre B6guin (Institut Pasteur, Paris, France) for sending the C. thermocellum cellulase and xylanase genes. This research was supported by Grant 86 017 from the Flemish Government.

References Bates EEM, Gilbert H J, Hazlewood GP, Huckle J, Laurie JI, Mann SP (1989) Expression of a Clostridium thermocellum endoglucanase gene in Lactobacillus plantarum. Appl Environ Microbiol 55:2095-2097 Bbguin P, Millet J, Gr6pinet O, Navarro A, Jui M, Amit A, Poljak R, Aubert J-P (1988) The cel (cellulose degradation) genes of Clostridium thermocellum. In: Aubert J-P, B6guin P, Millet J (eds) Biochemistry and genetics of cellulose degradation. Academic Press, New York, pp 267-282 Berger E, Jones WA, Jones DT, Woods DR (1989) Cloning and sequencing of an endoglucanase (endl) gene from Butyrivibrio fibrisolvens H17c. Mol Gen Genet 219:193-198 Casabadan M, Cohen SN (1980) Analysis of gene control signals by fusion and cloning in Escherichia coli. J Mol Biol 138:179207 Claeyssens M (1988) The use of chromogenic substrates and specific assays in the study of structure-activity relationships of cellulolytic enzymes. In: Aubert J-P, Bbguin P, Millet J (eds) Biochemistry and genetics of cellulose degradation. Academic Press, New York, pp 393-397 Cornet P, Millet J, B6guin P, Aubert J-P (1983) Characterisation of two cel (cellulose degradation) genes of Clostridium thermocellurn coding for endoglucanases. Bio/Technology 1:589601

Dao ML, Ferretti JJ (1985) Streptococcus-Escherichia coli shuttle vector pSA3 and its use in the cloning of Streptococcus genes. Appl Environ Microbiol 49:115-119 Gruss A, Ehrlich SD (1989) The family of highly interrelated single-stranded deoxyribonucleic acid plasmids. Microbiol Rev 53:231-241 Hazlewood GP, Romaniec MPM, Davidson K, Grepinet O, Beguin P, Millet J, Raynaud O, Aubert J-P (1988) A catalogue of Clostridiurn thermocellum endoglucanase, fl-glucanase, and xylanase genes cloned in Escherichia coli. FEMS Microbiol Lett 51 : 231-236 Joliff G, Beguin P, Juy M, Millet J, Ryter A, Poljak R, Aubert J (1986) Isolation, crystallization and properties of a new cellulase of Clostridium thermocellum overproduced in Escherichia coli. Bio/Technology 4:896-900 Josson K, Scheidinck T, Michiels F, Platteeuw C, Stanssens P, Joos H, Dhaes e P, Zabeau M, Mahillon J (1989) Characterisation of a Gram-positive broad host range plasmid isolated from Lactobacillus hilgardii. Plasmid 11:9-20 Luchansky JB, Muriana PM, Klaenhammer TR (1988) Application of electroporation for transfer of plasmid DNA to Lactobacillus, Lactococcus, Leuconostoc, Listeria, Pediococcus, Bacillus, Staphylococcus, Enterococcus, and Propionibacterium. Mol Microbiol 2:637-646 Macrina FL, Tobian JA, Jones KR, Evans RP, Clewell DB (1982) A cloning vector able to replicate in Escherichia coli and Streptococcus sanguis. Gene 19: 345-353 Mahillon J, Chungjatupornchai W, Decock S, Michiels F, Peferoen M, Joos H (1989) Transformation of Bacillus thuringiensis by electroporation. FEMS Microbiol Lett 60:205-210 Maniatis T, Fritsch EF, Sambrook T (1982) Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Millet J, P6tr6 D, Bbguin P, Raynaud O, Aubert J (1985) Cloning of ten distinct DNA fragments of Clostridium thermocellurn coding for cellulases. FEMS Microbiol Lett 29:145-149 Nelson N (1952) A photometric adaptation of the Somogyi method for the determination of glucose. J Biol Chem 153:376-380 Neumann JR, Morency CA, Russian KO (1987) A novel rapid assay for chloramphenicol acetyltransferase gene expression. Biotechniques 5:444-447 P&r~ D, Millet J, Longin R, B~guin P, Girard H, Aubert J-P (1986) Purification and properties of the endoglucanase C of Clostridiurn thermocellurn in Escherichia coil Biochimie 68: 687-695 Scheirlinck T, MahiUon J, Joos H, Dhaese P, Michiels F (1989) Integration and expression of a-amylase and endoglucanase gene in the chromosome of Lactobacillus plantarurn. Appl Environ Microbiol 55:2130-2137 Schwarz WH, Grabnitz F, Staudenbauer WL (1986) Properties of a Clostridium thermocellum endoglucanase produced in Escherichia coil Appl Environ Microbiol 51:1290-1299 Seale DR (1986) Bacterial inoculants as silage additives. J Appl Bacteriol Symp Suppl 1986:9S-26S Seale DR (1987) Bacteria and enzymes as products to improve silage preservation. In: Wilkinson JM, Stark BA (eds) Developments in silage 1987. Chalcombe publications, Marlow, Bucks, UK, pp 47-61 Soutschek-Bauer E, Staudenbauer WL (1987) Synthesis and secretion of a heat-stable carboxymethylcellulase from Clostridium thermocellum in Bacillus subtilis and Bacillus stearothermophilus. Mol Gen Genet 208:537-541 Wirth R, An F, Clewell DB (1987) In: Ferretti JJ, Curtiss R (eds) Streptococcal genetics. American Society for Microbiology, Washington, pp 25-27

541 Woolford MK (1984) The silage fermentation. Dekker, New York Zappe, H (1988) The cloning and characterisation of an endoglucanase and an endoxylanase from Clostridium acetobutylicum in Escherichia coli. Thesis, University of Cape Town, South Africa

Zappe H, Jones DT, Woods DR (1986) Cloning and expression of Clostridium acetobutylicum endoglucanase, cellobiase and amino acid biosynthesis genes in Escherichia coli. J Gen Microbiol 132:1367-1372 Zappe H, Jones DT, Woods DR (1987) Cloning and expression of a Clostridium acetobutylicum xylanase P262 in Escherichia coll. Appl Microbiol Biotechnol 27:57-63