MGG

Mol Gen Genet (1992) 234:401M11

© Springer-Verlag 1992

Protein export elements from Lactococcus lactis Gaspar Perez-Martinez*, Jan Kok, Gerard Venema, Jan Maarten van Dijl, Hilde Smith**, and Sierd Bron Department of Genetics, Centre of Biological Sciences, Kerklaan 30, 9751 NN Haren, The Netherlands Received November 21, 1991 / Accepted April 14, 1992

Summary. Broad-host-range plasmids carrying a-amylase or ]3-1actamase reporter genes lacking a signal sequence were used to select export elements from Lactococcus lactis chromosomal D N A that could function as signal sequences. Fragments containing such elements were identified by their ability to direct the export of the reporter proteins in Escherichia coli. Several of the selected export elements were also active in Bacillus subtilis and L. lactis, although the efficiencies depended strongly on the host organism and reporter gene used. The export elements AL9 and BL1 were highly efficient in L. lactis in the expression and secretion of at least two heterologous proteins (Bacillus licheniformis a-amylase and E. coli TEM-13-1actamase). AL9 even permitted growth of this organism on starch as the sole carbon source. Nucleotide sequence analysis of five selected fragments indicated that these encode oligopeptides with the major characteristics of typical signal peptides. The putative expression signals had a limited similarity to previously described expression signals for E. coli, B. subtilis and L. lactis. Differences in both expression and export efficiency are likely to underlie the host-specific effects. Key words: Lactococcus protein export - Expression signals - Signal peptide - a-amylase - [3-1actamase

Introduction Knowledge of the genetics and physiology of lactic acid bacteria has rapidly increased during the last few years. In particular, the availability of molecular genetic techniques has contributed in a major way to our knowledge of these industrially important bacteria, which are widely * Present address: Instituto de Agroquimica y Technologia de Alimentos (C.S.I.C.), Jaime Roig 11, 46010 Valencia, Spain ** Present address: Central Veterinary Institute, Edelhertweg 15, 8219 PH Lelystad, The Netherlands Correspondence to: S. Bron

used in the food industry. Potential new applications of these organisms are the production of (heterologous) extracellular proteins and pharmaceuticals. Although, so far, the natural protein export capacity of lactic acid bacteria has scarcely been studied, these bacteria are known to export several proteins, some of which are secreted into the growth medium. Thus, lactococci export a cell wall-associated serine proteinase essential for rapid growth in milk (Kok and Venema 1988), and they secrete Usp45, a protein of unknown function (van Asseldonk et al. 1990). With respect to heterologous proteins, it has been shown that Lactobacillus plantarum is able to secrete Bacillus stearothermophilus a-amylase and Clostridium thermocellum cellulase A, using the native signal peptides of these proteins (Scheilink et al. 1989). The secretion by Lactococcus lactis of Bacillus subtilis neutral protease fused to a lactococcal promoter was demonstrated by van de Guchte et al. (1990). Escherichia coli TEM-[3-1actamase, when joined to lactococcal expression/export signals, can also be secreted by L. laetis (Sibakov et al. 1991). Exported proteins are initially synthesized as precursors with an N-terminal extension, the signal peptide. The signal peptide is required to guide the protein into the export pathway by interacting with components of the cellular export machinery and the membrane (for reviews see Saier et al. 1989; Wickner et al. 1991). Moreover, the presence of the signal peptide retards the folding of the precursor (Laminet and Plfickthun 1989; Park et al. 1988), which appears to be important for export competence (Saier et al. 1989; Wickner et al. 1991). Although signal peptides from different exported proteins have little sequence homology, they have a tripartite structure in common: a short (5-8 amino acids) positively charged amino-terminus (n-region); a central hydrophobic core (h-region; 8-15 amino acids); and a short (5-8 amino acids) more polar c-region that contains the cleavage site for signal peptidase (SPase I; von Heijne 1986, 1990). A comparison of signal peptides from E. eoli and bacilli was made by von Heijne and Abrahms6n (1989). On average, signal peptides from bacilli are

402

slightly longer in the n-, h- and c-regions than those from E. coli, and the n-region in bacilli is generally more highly charged. Analyses of signal peptides from lactic acid bacteria have only recently begun. In addition to signal sequences of L. laetis proteinase genes (Kok et al. 1988), a number of randomly selected lactococcal D N A fragments have recently beeen described that direct the export of proteins (Sibakov et al. 1991). The primary aim of the studies reported here was to isolate and characterize D N A fragments from the L. lactis chromosome that could function as signal sequences. Such export elements are valuable for the construction of efficient expression/export vectors for L. lactis. A considerable number of export elements with signal sequence function were selected from L. lactis in a similar way to that described previously for Bacillus (Smith et al. 1987, 1988, 1989). For this purpose plasmids were used that encode mature parts of exoproteins but not the signal peptides. Some of these export elements wet highly efficient in L. lactis. The selected fragments permitted a comparison of expression/secretion efficiencies in E. coli, L. lactis and B. subtilis. Materials and methods

Bacterial strains, plasmids and media. Table 1 lists the bacterial strains and plasmids used. TY medium was used for culturing E. coli and B. subtilis (Bron and Luxen 1985), and M17 medium supplemented with 0.5% glucose for culture ofL. lactis (Kok et al. 1984). If required, 1% starch was added to the media. Erythromycin was used at 5 gg/ml for L. lactis and B. subtilis, and at 200 gg/ml for E. coli. To score for J3-1actamase export in E. coli, 5 gg/ml of ampicillin (Amp) (plus 100 gg/ml erythromycin to select for the plasmid) were used in the initial screening. E. coli was transformed using standard procedures (Maniatis et al. 1982), B. subtilis by the competent cell procedure (Bron and Luxen 1985), and

L. lactis by electrotransformation (van der Lelie et al. 1988).

DNA techniques. General procedures for the manipulation of D N A were carried out essentially as described by Maniatis et al. (1982). Restriction enzymes and T4 D N A ligase, obtained from Boehringer (Mannheim, FRG), or New England BioLabs (Beverly, Mass.), were used as specified by the suppliers. Nucleotide sequences of both strands of D N A were determined by the dideoxy chaintermination method (Sanger et al. 1977) using the T7 D N A polymerase sequencing kit (Pharmacia LKB, Uppsala, Sweden). Either M13mp10/mpll subclones or pGA14/pGB14 derivatives were used in sequencing. Mapping of the 5' termini of m R N A by primer extension with reverse transcriptase was performed as described by Calzone et al. (1987). m R N A was isolated from L. lactis M G 1363 carrying the appropriate plasmids as previously described (van der Vossen et al. 1987).

Enzymatic assays. All enzymatic activities were determined in exponentially growing cells at an optical density of 0.6 at 600 nm. 13-Lactamase and isocitrate dehydrogenase activities were measured as described by Smith et al. (1987). [3-Galactosidase was determined by the method of Miller (1972), phosphoglucose isomerase by that of Michal (1984) and a-amylase activity was determined using the procedure of Spiro (1966). Spheroplasting. The contens of the periplasmic space were released from E. coli cells by treatment for 30 min at 37 ° C with 5 mg/ml of lysozyme in 0.1 M potassium phosphate buffer, pH 7.5 plus 10 mM EDTA and 0.3 M sucrose. Under these conditions no [3-galactosidase activity was released into the medium, indicating that the E. coli cells remained intact during spheroplasting.

Table l. Bacterial strains and plasmids

Strain or plasmid

Relevant characteristics

Origin or reference

Escherichia coli BHB2600 E. coli MRE600 E. coli C600 Lactococcus lactis MG1363 Bacillus subtilis DB104 B. subtilis DB104 (amy)

803supE supF met rk-m k RNAse I thr leu thi lacy tonA phx supE vtr lac prtP his nprR2 nprE18 aprA3 a-Amylase negative derivative of strain DB104

K. Murray de Vrije et al. (1987) Raleigh et al. (1988) Laboratory collection Kawamura and Doi (1984) Laboratory collection

a-Amylase-based vector for selection of export elements; pWV01 replicon; contains SPO2 promoter; Em r [3-Lactamase-based vector for selection of export elements; pWV01 replicon; Em r pGB14 containing the SPO2 promoter pGA14 derivatives containing the AL9 (AL39) inserts directing export of a-amylase pGB14 derivatives containing the BL1 (BL11, BL13) inserts directing export of ]3-1actamase

This work

Plasmids pGA14 pGB14 pGB18 pGAL(9, 39) pGBL(1, 11, 13)

This work This work This work This work

403

Results

We define stages in protein export according to Pugsley and Schwartz (1985): export, the translocation of precursor proteins to an extracytoplasmic location, including the cytoplasmic membrane; processing, the removal of the signal peptide by signal peptidase; secretion, the special case of export that results in the extracellular accumulation of exported proteins. In the context of the present studies, we define export elements as signal sequence-like functions enabling the export (translocation) of reporter exoproteins across the cytoplasmic membrane. As discussed in the following section, the presence of a processing site for signal peptidase was not required for all of the export elements described here. Selection of protein export elements from L. lactis DNA

Two plasmids (Fig. 1) were constructed that carried either the Bacillus licheniformis a-amylase gene (pGA 14), or the E. coli TEM-[Mactamase gene (pGB14). The reporter genes were Y-terminally truncated and lacked their transcription/translation initiation signals and the signal sequence, which were replaced by a multiple cloning site (Smith et al. 1987). Consequently, the reporter proteins are not expressed from these plasmids. The selection vectors are based on the broad-host-range lactococcal plasmid pWV01 (Kok et al. 1984), which enabled the comparison of the effects of cloned D N A fragments in different bacteria without subcloning. The copy numbers of this plasmid per chromosome equivalent are about 5 for B. subtilis and L. lactis and about 50 for E. coli (Kok et al. 1984). Cloned DNA fragments

that restore the export of the reporter proteins can be selected with these plasmids, provided that they are placed in frame with the reporter genes. In addition, the selection of export elements with pGB14 requires the cloning of expression signals. With pGA14, carrying the SPO2 promoter, which is functional in all three bacteria used in these studies, the cloning of a promoter is not essential. The initial selection of export elements was carried out in E. coll. Since resistance to Amp can only develop when [Mactamase has been translocated into the periplasm (Kadonaga et al. 1984), this provides a positive assay for pGB14 clones containing export elements. We have shown before (van Dijl et al. 1991b) that translocated [3-1actamase that is not processed also renders E. coli cells resistant to Amp. As a consequence, the presence of a signal peptidase cleavage site on the selected export elements is not required in this assay. With pGA14, halo formation around E. coli colonies on starch-containing plates was used to monitor a-amylase export. Since this assay requires the diffusion of a-amylase into the agar, processing of the s-amylase precursor is required in this case. Chromosomal D N A from L. lactis MG1363 was digested with the restriction enzymes RsaI plus HaeIII, or AluI, to obtain fragments of about 0.3 to 1.5 kb. The digested D N A was ligated into the Sinai site of pGA14 and pGB14 and the ligation mixtures were used to transform E. coli BHB2600. Among approximately 50000 erythromycin-resistant E. coli transformants obtained with pGB14, 12 were resistant to > 5 gg/ml Amp. The selected fragments were designated BL. From approximately 50000 erythromycin-resistant transformants obtained with pGA14, 48 exported a-amylase. The selected fragments were designated AL.

'¢¢-omy

So

P ....

,~co.t "..

\ Sinai \ S~l

BomH! Smo!

EcoR I

or-amylase

~-lactamasc +1

G D P L E S T A A A A GAATTCGAGC TC GC C C ~ T C .CT C T A G A G T C G A C C G C A G C C,C~ G G C A EcoRI Sacl SmaI BamMl XbaI Sall

Fig. 1. Plasmids used for the selection of protein export elements from Lactococcus lactis. Plasmids pGA14 and pGB14 contain the 5'-terminally truncated Bacillus licheniformis a-amylase and Escherichia coli TEM-13-1actamase genes, respectively. The truncated reporter genes were described by Smith et al. (1987). pGA14 contains the bacteriophage SPO2 promoter upstream of the

+1 +2 P P ~AATTC GAGC TCGC CCGGC2GATCC T_CTAGAGTCGAC C GC e CAAGC T T G C CCC C C A EcoRI SacI SmaI BamHI XbaI SalI HindIII G

D

P

L

E

S

T

A

Q

A

C

truncated a-amylase gene. Both plasmids contain the replication functions of the broad-host-range plasmid pWV01 (Kok et al. 1984). The plasmids contain the erythromycin resistance marker from plasmid pE194. The sequences below the plasmids show the regions of the fusions between the multiple cloning sites and the reporter genes

404 Table 2. [3-Lactamase activities obtained with export elements from Lactococcus lactis Export function

Escherichia coli Ampr Activitya (gg/ml) (U/ml)

BL1 BL4 BL5 BL10 BL11 BL12 BL13 BL14 AL9 None (pGB14)

10 10 5 10 75 30 500 10 10 1

40 30 130 200 1580 210 1785 420 nd < 10

Exportedb (%)

Bacillus subtilis Activitya Exportedb (U/ml) (%)

Lactococcuslact~ Activitya Exportedb (U/ml) (%)

100 35 53 48 26 45 18 16 nd nd

880 700 150 0 3480 620 1660 730 nd < 10

18 700 160 570 80 1270 160 3 660 90 38 850 < 10

100 100 100 0 69 100 100 100 nd nd

92 100 100 100 24 100 65 100 98 nd

Size insert (bp)

300 800 750 900 720 800 1300 360 720 -

One unit of activity (see Materials and methods) was defined as the amount of J3-1actamasethat in 1 ml samples raised the absorbance at 486 nm by 0.001/rain at room temperature. Resistance to ampicillum (Amp) was expressed as the maximal concentration of Amp (in gg/ml) at which growth to colonies occurred. AL9, which had been selected with plasmid pGA14, was inserted into plasmid pGB14

a Total amount of ~-lactamase activity. In the case of E. coli this represents the total activity in pelleted cells; in the case of B. subtilis and L. lactis this represents the activities in cells plus culture supernatants b Fraction (in %) of the 13-1actamaseactivity present in the periplasmic space (E. coli), or in the culture supernatant (B. subtilis and L. lactis) ;nd, not determined

Two assays were used to obtain information a b o u t the relative efficiencies of the selected export elements. The first was the measurement of the level of resistance to A m p in E. coli. Since only exported (periplasmic) fi-lactamase causes resistance to A m p ( K a d o n a g a et al. 1984), this is a measure of export efficiency in E. coli. This assay cannot be used for the Gram-positive bacteria, because with TEM-13-1actamase, these do not develop resistance to Amp. The second assay involved measurement of export and secretion of a-amylase and 13-1actamase in the three organisms tested. We have shown previously (Smith et al. 1988) that not only mature, processed forms, but also precursor forms of the c~-amylase reporter protein are enzymatically active. Therefore, determination o f the activities in cell-associated and extracellular fractions (B. subtilis and L. Iactis), or those of cell-associated and periplasmic fractions (export in E. coli), allows a comparison o f the efficiencies of selected export elements, and a comparison o f effects associated with the host bacteria. Although quantitative data are lacking, we assume that the same holds for ~-lactamase. Since pre-[3lactamase (Laminet and Plfickthun 1989) and hybrid pre-~-lactamases are active (van Dijl et al. 1991a), activity measurements give information a b o u t amounts of protein. A potential problem with this assay is instability of the hybrid proteins in the three hosts. This would reduce the enzymatic activities, as has been shown for hybrid [3-1actamase in B. subtilis (Smith et al. 1987). This problem was avoided by using exponentially growing cultures. Prolonged incubation o f cell extracts indicated that no losses of a-amylase and J3-1actamase activity occurred as a function o f time (all three bacterial species were tested; results not shown).

in E. coli (Table 2). BL13 and B L l l were exceptions: c o m p a r e d with export elements isolated from B. subtilis, the first conferred high levels and the second moderate levels o f resistance. Plasmids containing the BL inserts were transferred to B. subtilis DB104 and L. lactis MG1363, and the total ~-lactamase activities were determined. In addition, the fractions of the activities exported into the periplasm (E. coli) or culture supernatant (B. subtilis and L. lactis) were determined. Controls with the intracellular enzymes [3-galactosidase (E. coli), isocitrate dehydrogenase (B. subtilis), or phosphoglucose isomerase (L. laetis) indicated that in none of the species were detectable amounts o f intracellular proteins released by cell lysis (data not shown). In E. coli the total activities of J3-1actamase varied greatly, indicating that different export elements had been selected. In B. subtilis and L. lactis all selected elements (except BL10 in B. subtilis) directed the expression of J3-1actamase. However, the relative levels o f the activities were different in the three bacterial species. F o r instance, BL11 was the most efficient in B. subtilis, BL13 in E. coli, and BL1 in L. lactis. B L l l and BL13 were moderately active in L. lactis, and BL1 had only a low activity in B. subtilis and E. eoli. These and other data f r o m Table 2 indicate that the host had a drastic effect on the total ~-lactamase activity. Depending on the selected fragment, the fraction of exported 13-1actamase also varied considerably. In E. coli, BL1 was efficient in directing export: all [3-1actamase activity was located in the periplasm. With all other elements 50 % or m o r e o f the [3-1actamase remained associated with the cells. In spite of this, some o f these elements were so highly efficient in expression (BLl 1 and BL13) that, as c o m p a r e d with BL1, large amounts of f3-1actamase accumulated in the periplasm. A remarkable observation was that with BL13 a very high level o f resistance to A m p (500 gg/ml) was obtained. Since only periplasmically localised [Mactamase can cause resis-

Export elements selected with the ~lactamase 9ene M o s t of the D N A fragments selected with pGB14 conferred rather low levels of A m p resistance ( < 30 gg/ml)

405 t a n c e ( K a d o n a g a et al. 1984), this s u g g e s t e d t h a t , in E. coli, BL13 p e r m i t s efficient t r a n s l o c a t i o n o f [3-1actam a s e into the p e r i p l a s m . T h e l o w p e r c e n t a g e o f free [ M a c t a m a s e in the p e r i p l a s m (18%) m a y i n d i c a t e t h a t p r e - ( B L 1 3 ) - ~ - l a c t a m a s e was p o o r l y processed b y SPase I. M o s t o f the B L e l e m e n t s c a u s e d secretion o f n e a r l y all ~ - l a c t a m a s e i n t o the e x t r a c e l l u l a r m e d i u m o f B. subtilis a n d L. lactis. BL11 a n d BL13 were e x c e p t i o n s : w i t h these e x p o r t e l e m e n t s c o n s i d e r a b l e a m o u n t s o f the ~ - l a c t a m a s e a c t i v i t y were a s s o c i a t e d w i t h the cell f r a c t i o n . F o r c o m p a r i s o n , the results o b t a i n e d w i t h A L 9 (see f o l l o w i n g p a r a g r a p h ) a r e also s h o w n in T a b l e 2. W h e n p l a c e d in p l a s m i d p G B 1 4 , this f r a g m e n t a p p e a r e d to b e the m o s t efficient o f all in L. lactis (38 850 U / m l ~ - l a c t a m a s e ; 98 % secreted).

Export elements selected with the a-amylase 9ene T w e n t y - t h r e e o f the A L e x p o r t e l e m e n t s selected w i t h p G A 1 4 in E. coli were t r a n s f e r r e d to B. subtilis D B 104 (amy) a n d L. laetis. T h e t o t a l a - a m y l a s e activities were d e t e r m i n e d , as well as the f r a c t i o n s o f the activity t h a t were e x p o r t e d ( T a b l e 3). W i t h these e l e m e n t also, a c o n s i d e r a b l e v a r i a t i o n in t o t a l a m o u n t o f activity was o b s e r v e d in E. coli, i n d i c a t i n g that, again, m a n y different elements h a d b e e n selected. T h e m a j o r i t y o f the A L

elements d i r e c t e d low a - a m y l a s e p r o d u c t i o n in the G r a m - p o s i t i v e b a c t e r i a : in B. subtilis, a - a m y l a s e activity c o u l d be m e a s u r e d o n l y w i t h A L 9 , A L l 4 , A L 3 1 , A L 3 9 , a n d A L 4 0 , while in L. lactis o n l y A L 9 was active. This d e m o n s t r a t e s t h a t t h e r e is a d r a s t i c host-specific effect o n a - a m y l a s e p r o d u c t i o n . T h e f r a c t i o n o f the a - a m y l a s e released i n t o the p e r i p l a s m o f E. coli v a r i e d f r o m 22 % to 85%. I n B. subtilis, A L 9 a n d A L 3 9 a p p e a r e d to be m o r e efficient in d i r e c t i n g e x p o r t o f a - a m y l a s e ( > 8 8 % secreted) t h a n in E. coli ( a b o u t 60% e x p o r t e d ) . A L l 4 , AL31 a n d A L 4 0 were m o d e r a t e l y efficient in b o t h E. coli a n d B. subtilis ( 5 0 % - 7 0 % e x p o r t e d ) . A s in B. subtilis, A L 9 was also efficient in the secretion o f a - a m y l a s e b y L. lactis (90% in the s u p e r n a t a n t ) . T h e host-specific effects o n the a m o u n t s o f e x p o r t e d a - a m y l a s e a r e illustrated in Fig. 2, w h i c h shows h a l o f o r m a t i o n o n starchc o n t a i n i n g p l a t e s b y v a r i o u s E. coli, B. subtilis, a n d L. Iactis clones. Several o f the D N A f r a g m e n t s selected w i t h the aa m y l a s e gene were t r a n s f e r r e d as SaeI-XbaI f r a g m e n t s into the c o r r e s p o n d i n g sites in p G B 1 8 . This p l a c e d the e x p o r t elements d o w n s t r e a m o f the S P O 2 p r o m o t e r , in f r a m e w i t h the ~ - l a c t a m a s e r e p o r t e r gene. T h e levels o f resistance to A m p were m e a s u r e d in E. coli ( T a b l e 3 c o l u m n 4). A l t h o u g h all A L elements c a u s e d resistance to A m p , little c o r r e l a t i o n was o b s e r v e d b e t w e e n a - a m y lase a c c u m u l a t i o n in the p e r i p l a s m a d the level o f A m p

Table 3. a-Amylase activities obtained with export elements from Lactococcus lactis Export function

Escherichia coli Activitya (U/ml)

AL6 44.9 AL9 16.0 ALl2 20.2 ALl4 11.9 AL24 81.0 AL25 10.6 AL27 24.6 AL29 46.2 AL30 84.0 AL31 87.4 AL32 42.2 AL33 106.0 AL35 6.8 AL36 86.0 AL37 > 126 AL39 > 126 AL40 63.0 AL41 29.8 AL42 18.1 AL43 31.7 AL46 122 AL47 63.0 AL48 118 None (pGA14) < 2 BL1 nd BL 13 nd

Bacillus subtilis

Lactococcus lactis

Exported b (%)

Amp r (gg/ml)

Activitya (U/ml)

Activity" (U/ml)

33 62 75 52 73 48 49 66 30 50 85 78 56 44 50 56 71 22 32 27 46 62 55 nd nd nd

10 10 nd 200 200 10 200 200 20 10 200 10 nd nd 10 200 10 200 10 l0 75 nd 100 nd nd

+ 13.6 4.2 + 3.2 + 1.4 2.2 + +

Exported b (%)

Exported b (%)

-

95

11.7

62

-

90

-

-

-

62

-

-

88 48

-

-

-

+ + +

Units of activity (see Materials and methods) are given in millimoles of glucose equivalents released from starch per hour and per millilitre culture, nd, not determined; + , halo formation on starch plates, but activity too low to be measurable; -, no activity detectable

-

-

9.6 5.0

77 65

Size insert (bp)

300 720 300 720 300 300 300 < 300 1200 750 < 300 650 600 < 300 < 300 < 300 < 300 1200 < 300 < 300 < 300 490 400 300 1300

a. b See Table 2 (here c~-amylase activities) Levels of resistance to ampicillin (Amp) (in E. coli) obtained with pGB 18 plasmids into which the AL fragments had been introduced

406 D N A sequences and expression signals

With the aim of identifying expression signals and export signal functions, five of the fragments were sequenced: BL1 (efficient in L. lactis); AL39 (efficient in E. coli); BL11 and BL13 (efficient in both E. coli and B. subtilis); and AL9 (efficient in both L. lactis and B. subtilis). The data (Fig. 3A-E) indicated in all sequenced fragments the presence of: (1) an open reading frame (ORF) fused in frame to the reporter gene; and (2) promoter-like sequences upstream of the ORFs. The similarity of the putative promoters to previously described lactococcal promoters (de Vos 1987; Koivula et al. 1991a; van de Guchte et al 1992; van der Vossen et al. 1987) was rather limited, however. None of the putative promoters described here had a perfect match with the consensus - 35 (TTGACA) and - 1 0 sequences (TATAAT). We consider the sequence denoted b in AL9 as the best candidate for a functional promoter (spacing 19 bp). In fragment BL1, the sequences denoted a and al are likely - 3 5 and 10 promoter sequences (spacing 15 bp). Primer extension assays with mRNA extracted from L. lactis were used to map the start sites of transcription (results not shown) in the two most efficient export elements in L. lactis (AL9 and BL1). These start sites fitted well with the promoter regions deduced from the sequence. The 35 region of the putative promoter on BL1 is contained within a 19 bp (imperfect) inverted repeat sequence, which can potentially form a stern-loop structure (AG = - 14.8 kCal/mol). In AL9, BL1 and BL13, which contain the most efficient expression signals for L. lactis (Tables 2 and 3), the region preceding the putative promoters is extremely A/T rich. In addition, in BL1 (position 25 to 50, Fig. 3B) and BL13 (position 24 to 76, Fig. 3D), stretches of three or more As are present with a helical periodicity (about 10 bp). These regions might facilitate DNA bending. In four of the fragments (AL9, AL39, B L l l , and BL13) potential translation signals are present: a start codon (AUG), preceded at a short distance (less than 12 bp) by a Shine-Dalgarno (SD) sequence. Some support for the assignment of the translation start sites was obtained by estimating from SDS-polyacrylamide gel electrophoresis the actual sizes of [3-1actamase precursors synthesized in vitro. These corresponded well with those predicted from the sequences (results not shown). The high level of expression by BLI in L. lactis was surprising, since no SD-like sequence is present in the region preceding the single AUG codon (Fig. 3B; position 117) in frame with the reporter gene. Moreover, this codon is directly adjacent to the transcription start site (position 116; Fig. 3B), which makes the assignment of the translational start at position 117 doubtful. However, no obvious alternative start codon nor SD sequence is present downstream from the transcription start point. -

Fig. 2. Differencesin relativea-amylaseactivitiesin Escherichia coli, Bacillus subtilis, and Lactoeoccus lactis. Colonies of the three bacterial strains that carried identical pGAL plasmids were toothpicked onto an agar plate containingM 17 mediumplus glucoseand 1%starch. The plates werefirstincubatedfor 18 h at 30° C to permit growth of L. lactis, followedby 24 h at 37° C to provide optimal growth conditions for E. coli and B. subtilis. Halo formation was subsequently visualized by staining with iodine solution (Smith et al. 1988). Upper row, E. coli; middle row, B. subtilis; bottom row, L. lactis. From left to right, AL48, AL39,AL37, AL31, AL24, AL9, and AL6

resistance. For instance, AL37 and AL39 were equally efficient in the export of a-amylase (about 50% in the periplasm), but only AL39 caused high levels of resistance to Amp. This suggested that, in contrast to pre(AL37)-a-amylase, pre(AL37)-f3-1actamase was translocated inefficiently in E. coli. Since in this case identical expression signals were compared, these results indicated that the translocation efficiencies were also affected by the mature part of the reporter proteins.

Growth o f L. lactis on starch

Of the selected export elements, AL9 and BL1 were the most efficient in the production and secretion of both reporter proteins in L. lactis (Tables 2 and 3). BL1 should contain a promoter, since it was selected in the absence of the SPO2 promoter. The removal of the SPO2 promoter from plasmid pG-AL9 did not affect halo formation on starch-containing plates (data not shown), also indicating that AL9 carries transcription/translation signals. AL9 directed such high levels of a-amylase synthesis and secretion (Table 3) that L. lactis could grow well (OD6oo values of at least 1.2 could be obtained) in M17 medium containing 1% starch as the sole carbon source.

-

Characteristics o f the export elements

The hydrophobicity profiles of the five oligopeptides encoded by the selected fragments are shown in Fig. 4.

AL9

A

BLII

C

10 20 30 40 50 60 TCGCCCCTTATTTATAATAAGCAAAAATCCATCTTCTGATTTGTTGGGTATCTTTAAAAC

i0 20 30 40 50 60 GGAAAAGAAATAAAAATTCGTGTTATTGCTGCCAATAAACGAAAAGGAACAGTAGATTTT

70 80 90 I00 ii0 120 CATATTTTTTTGATAAGGAATGATAA'CCAGTTCTGTTGTCAAGATACTCTTTAACAATCT

70 80 90 I00 Ii0 120 GAACAAATCGGTCCTGAAAAAAACTAGCAATTTGCTAGTTTT~'XTCTGTATGTTTTTTAT

130 140 150 160 170 180 TTAGTTTAAATTCAAATGTATATTTTGTCATAAAAAAATTGACCTCCATTAACTAGACTT

130 140 150 160 170 180 TAGCTAATATCGTCTTTACAGTATATTTTACATAACATAAATTATGTTAGCTGTGTTATT

190 200 210 220 230 240 TTTAGTCTAACTTATGGGGGTCATATCATTTTATAGAGGTTTTTTGGTATCTAAAAAAGC

190 200 210 220 230 240 TTGAGTATAAGGGAGCCTTATTTTTGATTACGAATTGATTTAGGCACGGATATCAGTAGA

250 260 270 280 290 300 CAATGGCATTTATTTTTTATATT~CGATTGATAAAAATAATATAAGTGAAAATACAAATT

250 260 270 280 290 300 AAAATGAAATCAAAGTTTGCTTGATAAAAGATTGTTACTATTTAAATTAAGTTATATAGT

a

b il

-~

. . . .

t

a

-ff--

310 320 330 ~ 340 350 360 TATCAATTTTTGTTAGAATAGAGACTATAAATGTCAATACATTGACAATTCAAGGAGC GA

--b--

.

.

.

.

.

310 320.. 330 340 350 360 TATCAAGCGAATATGCTAGCTTAACAAACTATTTATGTAATAATATAGGTATGGTAATCA

~

370 380 390 400 410 420 AAATGG CTACAC TT~AATTAAAGT CTTCAGTGG C GATTGAAGATAAAGAGAT CTCAAAGG

370 380 390 400 410 420 AAAAAAC G CATACATAGATTTAGGGAGGGAACGTATC, ACTAGTAAAATAAAATCATTGAC ..... M T $ K i K S L T

430 440 450 460 470 480 AGTCAACCTAGTCATCAATACCAATGAAATTCATGCAATTATGGGACCTAACGGAACTGG

430 440 450 460 470 480 ATATC TTCTATTATTTATTGTGGGAATC GAAATTATTGGTGG CTTATCTGGTTTCTTTGC Y L L L F I V G I E I I G G L S G F F A

490 500 510 520 530 540 TAAATCGACTTTATCGGCAGCGATTATGGGGAACCTAATTATACCGTCACTGAGGGAGAA

490 500 510 520 530 540 G G GAATTATCAAGGAAATTTATAATAATCTCATAC TT C C G C C~CTG g CTC CGC CAGATTA G I I K E I Y N N L I L P P L A P P D Y

5~0 560 570 580 590 600 GTTCTTTTTGA~GGTCAAAATGTTTTAGAGATGG~GTTGATGAAAGAGCTGGAGGGAGA

550 560 570 TTTATTTGGAATTGTTTGGGGGGATCCTCTA L F G I V W G D P L

610 620 630 ~40 650 660 AAAT~GAGCATT~TGTTTCTAC~C~CTTTTTAGC~C~T~TTGA~T~TCA~G M R A L M F L Q R F L A T I I D L I I

BLI3

D

I0 20 30 40 50 60 GGTTAAAGGGGTTTTTGG CAACTAAAAGTAAAAAAACAC TGAAAATATGATAAAATATGA

670 680 690 700 TTTA~TACCAGTTTTAC~CTTGTTCAGGGGGATCCTCTA V Y L P V L L L V O G D P L

70 80 90 I00 110 120 TAAAATTAAGAAGAAAGCGTAGTAGTTTTAAGACAAAAAGAACTTAAA~AATAAAGATAG

~L1

B

I0 20 30 ~0 50 __ 60 GTAAAGAAC TTC TAAAGATGAACTAAAACAAAATACAAAA~TAAGAAAGAG C CTATTG

190 200 210 220 GTTTTAACGGTTATTCTTTATCAC CTCTGGGGGGATCCTCTA V L T V I L Y H L W G D P L

a ~o

130 140 150 160 170 180 AAATCCA~AAGGA.AAATGAAGCGTTAC GTCACAGGTTTTAATGGATTAC G~ACTATTGGA ..... M K R Y V T G F N G L R T I G

.

~o

~o

.

.

i~.~i~i~i~i~i~i:i~i~i~i:~i~i~i~i:=i~i~

.

~o

A ~ G ~ C A ~ T ~ T ~ T ~ C ~ T ~ ~ ~ ~ ~ ~ N ~ I I ~ L K T I ~ • ~

~o

~o

CGGTTGAT~T~ ~ L D ~

190 200 210 220 230 240 ~ATCGTTTTCG~G~ATTTCT~G~TGGT~TG~GGATATTC G ~ G G ~AG~G~ V I V F V L F L A S ~ S T I ~ L ~ A T ~ 250 260 270 GTACTGGAGGC~GGG~AGGGGOA~G~TCTA 8 T G A K A E ~ D P L

~::~::~:~::~::~:~:~::~::~::~::~::~

AL39

E

i0 20 30 40 50 60 TTGCTACCGCCAG~GGGTTCTAATCG~CAACTATCCCAAAGTGCTCGGCCTGGGCAAAAC

70 80 90 i00 II0 120 TATCTGTTCACCTGGTTGCGTGATAACGGAATTCGAGGCTCGC CCC CATATACTCACGTT

130 140 150 160 170 180 TAGGAAAT C CAAC CAAC GATI2TTTTTGAAG CAC G CATC GCAGCTCTTGAAGGTGG GAGTG .... M F L K H A S Q L L K V G V 190 200 210 220 CAGCCCTTGGTGTTGGTTCTGGCTCTGCGGGGGGATCCTCTA Q P e V e V L A e R = I) P e

~i~i~i:~ii~iiiii~iiii

Fig. 3A-E. Nucleotide sequence and deduced amino acid sequence of fragments AL9 (A); BL1 (B); BL11 (C); BL13 (D); and AL39 (E). Putative promoter sequences are underlined ( - 35 regions) or indicated with interrupted lines ( - 1 0 regions); Shine-Dalgarno sequences are indicated with dots underneath the nucleotides.

Heavy arrows show the transcription start sitesin AL9 (position 334) and BLI (position ] 16).A n inverted repeat in BLI is indicated by overliningarrows. Amino acids encoded by the multiple cloning site are given in boldface (stippling)

408

40 30 20 10 0 -10 -2(

40 I 30' 2O 10 I 0 --10 --20 -30 i -40-

BL.1

. . . . . .

. . . . . . . . . .

-30 i

-40 L 0

1'o

• • ' HTI IrLXI

2'0

3b

I IIlrllrLlllrU iUrULrLRsrsT

++4--

4'o

s'o

AL.9