INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The information necessary for a protein to direct its export from the cell cytoplasm across a membrane has been extensively studied in both prokaryotes and eukaryotes (26, 34, 41, 46). Export generally requires a particular N-terminal sequence (called a signal peptide) which directs the precursor to the secretion machinery and is cleaved during successful export of the mature protein (58). While the signal peptide is necessary for export, it is not sufficient, as not all proteins are secreted even if they do bear this sequence (3, 6, 35, 60). This suggests that information in the mature region of a secreted protein is also important for export. Indeed, alterations in the N-terminal end of the mature protein can drastically impair maturation (18, 21, 50, 59). It was also shown that positive charges block export when introduced directly after the signal peptide of exported proteins, while alterations which maintain a neutral or negative charge have little effect on maturation. In all these studies, the wild-type proteins were efficiently exported; only modifications which impair secretion were identified.

Some secreted bacterial proteins are translocated as preproproteins (5, 44, 46). The propeptide, which is processed after translocation, may improve translocation efficiency. Long propeptides (60 to 200 residues), found in most bacterial exoproteases, are autocatalytically cleaved and have an intramolecular chaperone activity (5, 13, 61). Short propeptides (with fewer than 60 residues) are found in different secreted enzymes from gram-positive bacteria, including Bacillus subtilis -amylase (51), Bacillus cereus -lactamase (Bla) (22), and Staphylococcus aureus nuclease (Nuc) (7). These propeptides are cleaved by unknown proteases, and their roles in secretion, folding, or stability of the mature protein are uncertain (39). Recent results show that the Nuc propeptide enhances Nuc secretion efficiency in Escherichia coli; it was proposed that the propeptide has Nuc-specific chaperone activity (49).

Here we examine the secretion capacity of Lactococcus lactis, a well-characterized lactic acid bacterium. Several properties of L. lactis make it an attractive host for the secretion of biologically useful proteins in fermentors or directly in food: L. lactis is a food-grade organism and is extensively used in dairy fermentations. Furthermore, a plasmidless strain of L. lactis does not secrete proteases, nor almost anything else, in quantity. A single secreted protein, Usp45, of unknown function (53), can be systematically identified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), while other secreted proteins are present in trace amounts. This feature could simplify analyses of secreted proteins.

To date, numerous heterologous proteins have been secreted in L. lactis using host-specific or nonnative signal peptides (29, 31, 40, 43, 46, 48, 54; for a review, see reference 8). However, in most of these studies, secretion efficiency (i.e., the proportion of total protein which is present in mature secreted form) was not determined.

In this study, we use staphylococcal Nuc as a model protein to determine whether secretion of heterologous proteins can be efficient in L. lactis. Among the desirable features of Nuc are that it is small, stable, and resistant to denaturation and that its activity can be detected on petri plates as well as in PAGE zymograms. The pre-Nuc reportedly contains an unusually long, 60-residue signal peptide (15). Two active forms of secreted Nuc are detected; the B form includes a 19-residue N-terminal propeptide (Fig. 1). In vivo, this propeptide is removed by proteolytic cleavage, resulting in the A form (7). The nuc gene has already been cloned (42), and Nuc is secreted from numerous gram-positive bacteria, including L. lactis subsp. cremoris, L. lactis subsp. lactis, Streptococcus salivarius subsp. thermophilus (17), B. subtilis (15), Corynebacterium glutamicum (19), and Lactobacillus sake (25). In addition, fusions to the N-terminal end of the mature protein do not abolish enzymatic activity (17, 24, 28, 32, 33, 37). In this study we examine different parameters which affect the secretion efficiency of Nuc in L. lactis. We have observed that the absence of the Nuc propeptide results in a marked decrease of secretion efficiency. However, this can be overcome by introducing a 9-residue synthetic propeptide, LEISSTCDA, just after the signal peptide cleavage site. LEISSTCDA also improved the secretion efficiencies of native Nuc and of the -amylase of B. stearothermophilus (AmyS) (27). While most previous studies on secretion have introduced modifications which impaired export, in this study we show that N-terminal insertion of a specific synthetic peptide can improve export.

View larger version (10K): [in this window] [in a new window]   FIG. 1.   Signal peptidase and secondary cleavage sites used by S. aureus, B. subtilis, L. lactis, and C. glutamicum for processing of Nuc. Arrows represent main processing sites, and dashed arrows represent minor processing sites as described in references 7, 19, and 23. The N termini of NucB and NucA of S. aureus Foggie are indicated.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, and media. E. coli TG1 (11) and L. lactis MG1363 (10) were used as hosts. Plasmids used are described in Fig. 2 and listed in Table 1. E. coli was grown on Luria-Bertani (LB) medium (38) and incubated at 37°C. L. lactis was grown on M17 medium (52) in which lactose was replaced by 0.5% glucose (M17-Glu) or on brain heart infusion (Difco) and incubated at 30°C. Chemically defined medium (CDM) (36) was used to grow L. lactis for pulse-chase experiments. The following antibiotics were added at the indicated concentrations: erythromycin, 5 µg/ml for L. lactis or 150 µg/ml for E. coli; and ampicillin, 100 µg/ml for E. coli.

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Schematic structures of proteins encoded by the indicated plasmids. For details of plasmid construction, see the text and Table 1. The amino acid sequence of the region deleted from NucT is shown in Fig. 4. Black arrowhead, promoter of the native nuc gene (Pstaf) or lactococcal strong promoter (P59); RBS, ribosome binding site of the nuc gene; SP-Nuc, Nuc signal peptide coding region; gray bar, Nuc propeptide coding region; black bar, sequence encoding the LEISSTCDA synthetic propeptide; open bar, NucA or mature AmyS coding sequence (not to scale).

Nuc and AmyS plate activity assays. Nuc plate assays were performed as previously described (17). AmyS activity was detected by growing cells on LB or M17-Glu agar plates containing 0.2% starch and, after 48 h, covering the colonies with Lugol solution (Sigma).

DNA manipulations. Plasmid DNA was isolated essentially as described previously (4) except that for L. lactis, TES buffer (25% sucrose, 1 mM EDTA, 50 mM Tris-HCl; pH 8) containing lysozyme (10 mg/ml) was used for 10 min at 37°C to prepare protoplasts. Enzymes were used as recommended by the suppliers. General procedures for DNA manipulations were performed as described elsewhere (38). Electroporation of L. lactis was performed as described elsewhere (16), and transformants were plated on M17-Glu agar plates containing the required antibiotic.

Amplification by PCR and oligonucleotide synthesis. PCRs were performed with a Perkin-Elmer Cetus (Norwalk, Conn.) apparatus using Taq DNA polymerase (Promega) as recommended by the manufacturer. Oligonucleotides were synthesized with a DNA synthesizer (Applied Biosystems, San Jose, Calif.). A 762-bp fragment was PCR amplified from the pBS:P_stafnuc matrix (17) (Table 1), which contains the nuc gene devoid of the promoter. The oligonucleotides used were 5'-GGAATTCAAAAGAAAGAGGTGTTAGTTATG-3' (oligo 1) for the coding strand and 5'-GGAATTCCGATCTAAAAATTATAAAAGT-3' (oligo 2) for the complementary strand. This DNA fragment was then cloned on pBluescript (pBS) vector in E. coli TG1, resulting in pBS:nuc.

Plasmid constructions. Plasmid constructions are summarized in Fig. 2. Plasmids pNuc6, pNuc7, pNuc9, and pNuc10 were obtained directly in L. lactis. To construct pNuc6, a BamHI/SalI nuc fragment isolated from pBS:nuc was inserted into BamHI/SalI-cut pJDC9:P₅₉, downstream of the strong lactococcal promoter P₅₉. pJDC9:P₅₉ nuc was established in E. coli TG1. The P₅₉ nuc cassette cut by KpnI/BspXI and treated with mung bean nuclease was inserted into XbaI-filled-in pVE3556, resulting in pNuc6.

Protein analysis, immunoblotting, and zymograms. SDS-PAGE, electroblotting onto polyvinylidene difluoride membranes (Millipore), and immunoblotting were performed as described previously (38) or according to manufacturer recommendations. Antibodies against Nuc, Usp45 and AmyS, raised in rabbits, were kindly provided by J. R. Miller (anti-Nuc) and W. M. de Vos (anti-Usp45 and anti-AmyS). Immunodetection was performed with protein G-horseradish peroxidase conjugate (Bio-Rad) and an enhanced chemiluminescence kit (Dupont-NEN) as recommended by the suppliers. To compare Nuc distribution or to quantitate Nuc secretion, three to six independent samples were prepared (see below). Samples to be compared were prepared at the same time and loaded on the same gel. After enhanced chemiluminescence detection, different nonsaturated film exposures were scanned with a Scanjet II scanner (Hewlett-Packard) using Deskscan II and ImageQuant programs to get average values. For quantitation, signals were compared to those of known amounts of a commercial NucA control. B and A forms of Nuc were taken into account in these estimations. Nuc enzyme activity was evaluated on zymograms of SDS-PAGE, after removal of the SDS, as described previously (19).

Preparation of cellular and supernatant fractions of L. lactis and N-terminal microsequencing. For cell fractionation, 2 ml of L. lactis cultures at a given optical density at 600 nm (OD₆₀₀) were harvested by a 5-min centrifugation at 4°C and 8,000 rpm (Sigma 1K15). The supernatant and cells were processed separately. To compare the amounts of secreted and cell-associated proteins, both cell and supernatant fractions were concentrated. Sample concentration was calculated as follows. The equivalent of 1 ml of 1 OD₆₀₀ unit of culture (cell or supernatant) was concentrated in a 100-µl final volume as described below, and 10 µl was loaded for SDS-PAGE. Supernatants were filtered on 0.2-µm-pore-size filters (low protein retention; Millisar NML Sartorius, Göttingen, Germany) and trichloroacetic acid (TCA) (15% final concentration) was added to the filtrate. The resulting pellet was dissolved in 1/20 volume of 50 mM NaOH. Cell pellets were washed once with 1 ml of ice-cold TES, resuspended in TES, and precipitated with TCA (10% final concentration). Cell pellets were then washed once with 1 ml of cold acetone, dried, and resuspended in 70 µl of TES containing lysozyme (1 mg/ml). After 30 min of incubation at 37°C, cells were lysed with 30 µl of 20% SDS. Equal volumes of 2× loading buffer were added to all samples. Bands on the polyvinylidene difluoride membrane corresponding to NucA, NucB, and LEISSTCDA-Nuc forms were cut and subjected to N-terminal microsequencing performed on a gas-phase sequencer (model 477A/HPLC 120A; Perkin-Elmer).

Pulse-chase conditions. An overnight culture of the appropriate L. lactis strain grown on CDM was used at 2% to inoculate 20 ml of CDM. The culture was incubated at 30°C to an OD₆₀₀ of 0.45, and 10 ml of this culture was centrifuged. The cell pellet was washed in CDM without methionine, resuspended in CDM without Met, and incubated at 30°C for 5 min. Cultures were pulse labelled for 2 min (a 1-min pulse gave similar results) by the addition of 8 µl of [³⁵S]Met (10 mCi/ml) to 8 ml of Met-depleted culture. A total of 700 µl of 5% Met (2,500-fold excess) was added (chase), and 1-ml samples were taken at given intervals. Samples were precipitated with TCA (20% final concentration) and washed with ice-cold acetone. The pellet was resuspended in 120 µl of NET (150 mM NaCl, 2 mM EDTA, 20 mM Tris; pH 7.8) plus 0.2% lysozyme and incubated at 37°C for 30 min. The samples were treated with 1.2 µl of 10% SDS for cell lysis, vortexed, and incubated at 95°C for 5 min. Samples were diluted twofold by the addition of NET plus 2% Triton X-100. Immunoprecipitation was performed as described previously (23).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

nuc gene expression and Nuc secretion efficiency in L. lactis. The nuc gene with natural expression signals was previously cloned (17) on pIL253, a high-copy-number plasmid (45), resulting in plasmid pNuc3 (Table 1 and Fig. 2). To determine secretion efficiency, cell supernatants and lysates were analyzed by Western blot experiments after SDS-PAGE, using polyclonal antiserum for Nuc detection: secretion efficiency was about 70%, and the remaining 30% was present as a cell-associated precursor (pre-Nuc) (Fig. 3A). This low secretion efficiency is in contrast with that observed for the native protein Usp45 (over 95% secreted) (data not shown) and could be due to the atypical structure of the Nuc signal peptide. When dilutions of commercially purified Nuc were used as a concentration standard, the yield of secreted Nuc was determined to be about 0.5 mg per liter in overnight cultures. To raise expression levels, plasmid pNuc6, in which the nuc promoter was replaced by P₅₉, a strong lactococcal promoter (55), was constructed (Table 1 and Fig. 2). Secreted Nuc levels were about 5 mg per liter of overnight culture, a 10-fold increase compared to expression from pNuc3 (Fig. 3A). In this experiment, secretion efficiency was ~60%. The amount of secreted Nuc was slightly greater in stationary-phase than in exponential-phase cultures (Fig. 3B), as expected if Nuc accumulates during growth. Taken together, these results show that Nuc is secreted in substantial amounts but inefficiently from L. lactis. Since (i) secretion efficiency of Nuc was not markedly affected by changing promoter strength (Fig. 3A) and (ii) secretion of Usp45 was not affected in any Nuc-secreting L. lactis strain (data not shown), we think that high Nuc production by pNuc6 does not saturate the secretion capacity of the strain but rather that the pre-Nuc is poorly recognized by L. lactis secretion machinery.

View larger version (49K): [in this window] [in a new window]   FIG. 3.   (A) Nuc distribution at low and high expression levels in L. lactis. Proportions of secreted Nuc and accumulated pre-Nuc (prec) were analyzed in overnight cultures in M17-Glu of MG1363 containing pNuc3 or pNuc6 (which produce low and high levels of Nuc, respectively). Western blot experiments of TCA-treated samples are shown. Ten microliters of cells, the equivalent of 0.1 OD600 unit of cells, or culture supernatants thereof was deposited per well. Left panel, low Nuc expression (from pNuc3); center panel, high Nuc expression (from pNuc6); right panel, commercially purified NucA (5 µg/ml; Sigma). (B) Nuc distribution as a function of growth phase. Culture samples of strain MG1363 containing plasmid pNuc6 grown in M17-Glu were taken in exponential phase (EXP) (OD600 = 0.6), early stationary phase (STAT) (OD600 = 2.3), and after overnight growth (ON) (OD600 = 1.9). Western blot experiments were performed on cell lysates and filtered supernatants of each sample, after treatment with 10 and 15% TCA, respectively. Migration positions of precursor (prec) or mature forms of both NucA and NucB (abbreviated as A and B, respectively) are indicated by arrows. C, cell lysates; S, supernatant fractions.

Pre-Nuc is enzymatically inactive. The strain producing high levels of Nuc (from pNuc6) and a nonproducing strain (containing cloning vector only) have similar growth characteristics, indicating that intracellular accumulation of pre-Nuc is not toxic. This observation suggests that pre-Nuc is inactive in the cytoplasm. Note that mature Nuc is devoid of cysteines and thus should not be affected by intracellular reducing conditions. Pre-Nuc inactivity could be due either to nonoptimal intracellular conditions for enzyme activity, to aggregation, or to the antifolding activity of the Nuc signal peptide (20, 33). SDS-PAGE of cell lysate and supernatant extracts of the L. lactis strain containing pNuc6 was analyzed by zymogram. Both NucA (detected in both cell lysate and supernatant) and NucB (in supernatant) were enzymatically active. Despite the accumulation of ample amounts of pre-Nuc as shown by Western blot experiments, no enzymatic activity was detected at the expected migration position of pre-Nuc on the zymogram. Subsequent analyses of Nuc distribution were therefore performed by immunological assays rather than activity tests. This result indicates that inactivity of the pre-Nuc form is due at least in part to the antifolding activity of the signal peptide.

N-terminal microsequencing of secreted Nuc forms. Nuc has two mature active forms, B and A (7). The A form results from proteolytic cleavage of B (predominantly at position +19) in S. aureus (7). The predominant mature form in L. lactis, unlike S. aureus, is NucB (Fig. 1). Inefficient cleavage to the NucA form may be due to a scarcity of L. lactis proteases. Both NucB and NucA appear to be stable, as no degradation products were detected by Western blot experiments. N-terminal microsequencing of the secreted forms B and A was performed to determine the cleavage sites of pre-Nuc in L. lactis (Fig. 1). Signal peptide cleavage is conserved and occurred just after the ANA motif, as predicted from von Heijne rules (56). The A form, however, initiated 2 amino acids downstream (+21) of the N-terminal end reported for S. aureus and is identical to the cleavage site found in B. subtilis (23). Cleavage to the NucA form thus appears to involve host proteases. These results show that processing of pre-Nuc in L. lactis is accurate and can result in the export of significant amounts of Nuc which remain stable.

Deletion of 17 residues of the Nuc propeptide results in impaired secretion efficiency in L. lactis. In L. lactis, the 21-residue Nuc propeptide interferes with neither activity nor cell localization since both B and A forms are active and located in the supernatant. As in C. glutamicum, only trace amounts of NucA are found to be cell associated (Fig. 4) (19). To test the involvement of the Nuc propeptide in the secretion process, we fused the Nuc signal peptide to NucT, an N-terminal truncation of Nuc. In NucT, residues +3 through +19 of the 21-residue propeptide are deleted (Fig. 4A). The first 10 amino acid residues of this mature protein, SQATSTKKLH, have a net charge of +3 (Table 2). As would be expected for a protein with a positively charged mature N-terminal end (57), the secretion efficiency of NucT was very low (30%) (Fig. 4B).

View larger version (21K): [in this window] [in a new window]   FIG. 4.   Deletion of 17 amino acid residues of the Nuc propeptide results in impaired secretion efficiency of Nuc in L. lactis. (A) The three last amino acid residues of Nuc signal peptide and the N termini of native NucB and NucT are given. The amino acids deleted in NucT are underlined. Signal peptide cleavage sites (resulting in NucB and NucT) and secondary processing site found in L. lactis (resulting in NucA) are indicated by arrows. (B) MG1363 containing either pNuc6 (encoding Nuc) or pNuc9 (encoding NucT) were grown overnight. Western blot experiments were performed on cell lysates and filtered supernatants of each sample, after treatment with 10 and 15% TCA, respectively. Migration positions of the precursor (prec) or mature (B, A, and T) forms of Nuc are indicated by arrows. C, cell lysates; S, supernatant fractions.

Nuc and NucT secretion efficiency is improved by insertion of a synthetic propeptide (LEISSTCDA) at the N terminus of the mature sequence. Studies of E. coli indicate that the presence of positively charged amino acids just after the signal peptide cleavage site can result in precursor accumulation (18). We considered Nuc and NucT to be good models to test the converse, i.e., whether the introduction of negative charges at the N terminus of the mature moiety could improve secretion efficiency. For this experiment, we used an oligonucleotide linker which was originally designed to introduce restriction sites for cloning immediately downstream of the sequence encoding the Nuc signal peptide (17) (Table 2 and Fig. 2). The oligonucleotide was designed to avoid codons for positively charged amino acids (17). The synthetic propeptide, LEISSTCDA, adds two negative charges at positions +2 (commonly found in the mature moiety of secreted proteins [57]) and +8. The LEISSTCDA-Nuc fusion is enzymatically active, as judged from PAGE zymograms of culture supernatants (data not shown). The secretion efficiencies of Nuc and LEISSTCDA-Nuc (encoded by pNuc6 and pNuc7, respectively) and those of NucT and LEISSTCDA-NucT (encoded by pNuc9 and pNuc10, respectively) were examined (Fig. 5A and B).

View larger version (31K): [in this window] [in a new window]   FIG. 5.   (A) Nuc distribution is altered by insertion of the LEISSTCDA peptide between the signal peptide and mature Nuc. MG1363 containing either pNuc6 (encoding Nuc) or pNuc7 (encoding LEISSTCDA-Nuc) was grown overnight. Western blotting was performed on cell lysates (C) and filtered supernatants (S) of each sample, after treatment with 10 and 15% TCA, respectively. Migration positions of precursor (prec) or mature forms of both NucA and NucB (abbreviated A and B, respectively) and LEISSTCDA-Nuc (LEISSTCDA-B) are indicated by arrows. (B) NucT secretion efficiency is improved by insertion of the LEISSTCDA peptide just after the Nuc signal peptide. MG1363 containing either pNuc9 (encoding NucT), at left, or pNuc10 (encoding LEISSTCDA-NucT), at right, were grown overnight. Western blotting was performed on cell lysates and filtered supernatants of each culture after treatment with 10 and 15% TCA, respectively. Prec, precursor; T, NucT; LEISSTCDA-T, LEISSTCDA-NucT; C, cell lysates; S, supernatant fractions. (C) Quantitation of Nuc. Supernatants of overnight cultures of MG1363 containing either pNuc6 (encoding Nuc) (center panel) or pNuc7 (encoding LEISSTCDA-Nuc) (right panel) were examined by Western blotting without any TCA precipitation. Amounts of secreted Nuc were determined by scanning blots and comparing signals with those of known amounts of commercially supplied NucA (left panel). Ten microliters was loaded per well. Values given above the lanes indicate NucA concentrations per milliliter. LEISSTCDA-B, LEISSTCDA-Nuc; B, NucB; A, NucA.

LEISSTCDA-Nuc precursor is more efficiently processed than pre-Nuc in L. lactis. Processing of the pre-Nuc and of the pre-LEISSTCDA-Nuc was analyzed by pulse-chase labelling experiments using [³⁵S]Met (Fig. 6). The effect of LEISSTCDA on secretion efficiency was found to be comparable in CDM, medium used for pulse-chase labelling, with that observed in rich medium (data not shown). Pulse-labelled pre-Nuc expressed from pNuc6 was present for at least 1 min after the chase. In contrast, no pre-LEISSTCDA-Nuc expressed from pNuc7 was detected, even at time zero (just after the pulse). These results are consistent with the conclusion that pre-LEISSTCDA-Nuc is processed more efficiently than pre-Nuc in L. lactis, without altering the native cleavage site.

View larger version (23K): [in this window] [in a new window]   FIG. 6.   Comparison of kinetics of Nuc and LEISSTCDA-Nuc maturation by pulse-chase experiments. MG1363 containing pNuc6 (encoding Nuc) (lanes on the left side) or pNuc7 (encoding LEISSTCDA-Nuc) (lanes on the right side) was grown in CDM and pulse-labelled with [35S]Met for 2 min. Samples were taken at different times after the pulse as indicated (in minutes). The time 0 min corresponds to a sample taken just at the end of the pulse. Positions of migration of different Nuc species were determined by parallel Western blotting performed on labelled samples. A faint contaminating band is present in all lanes at the expected position for LEISSTCDA-Nuc precursor (stippled arrow). As this band is equally present in samples obtained from Nuc and LEISSTCDA-Nuc producers, we suppose that it does not correspond to LEISSTCDA-Nuc precursor. A, NucA; B, NucB; LEISSTCDA-B, LEISSTCDA-Nuc; prec, precursor.

Insertion of LEISSTCDA peptide improves secretion of AmyS in L. lactis. We asked whether LEISSTCDA could improve the secretion efficiency of a protein not derived from Nuc. We chose the -amylase of Bacillus stearothermophilus (AmyS) (27), which has already been expressed in L. lactis (54). The mature region of AmyS was fused to the Nuc signal peptide, followed or not by LEISSTCDA. The first 10 amino acid residues of this mature region, SAPFNGTMMQ, have a neutral net charge, and the insertion of LEISSTCDA adds two negative charges (Table 2). To compare the yields of secretion of AmyS and LEISSTCDA-AmyS, two plasmids containing these fusions, pAmy4 (no LEISSTCDA) and pAmy5 (encoding the LEISSTCDA-AmyS fusion), were introduced in L. lactis. Activity tests on petri plates showed that significantly more LEISSTCDA-AmyS than AmyS was secreted in L. lactis (Fig. 7). Secretion efficiencies of AmyS and LEISSTCDA-AmyS were also compared by Western blots of supernatants and cell fractions of L. lactis strains containing pAmy4 and pAmy5, respectively, with polyclonal antiserum used for AmyS detection (not shown). Cell lysate fractions could not be precisely evaluated due to a contaminating band which obscured the precursor, even after preincubation of AmyS antiserum. However, consistent with activity test results, mature LEISSTCDA-AmyS was readily detected in supernatants, whereas only trace amounts of mature AmyS were detected in overexposed blots. These results indicate that LEISSTCDA-AmyS precursor is more efficiently processed in L. lactis than AmyS. The ensemble of these results indicates that the secretion efficiencies of proteins fused to the Nuc signal peptide are improved by the presence of the LEISSTCDA propeptide.

View larger version (124K): [in this window] [in a new window]   FIG. 7.   Secretion of AmyS is improved by insertion of the LEISSTCDA peptide just after the signal peptide. Cultures of MG1363 containing pIL252 (negative control) (streak 1), MG1363 containing pAmy4 (Nuc signal peptide followed directly by AmyS) (streak 2), and pAmy5 (Nuc signal peptide followed directly by LEISSTCDA-AmyS fusion) (streak 3) were streaked on M17-Glu agar plates containing 0.2% starch. AmyS secretion was visualized by a Lugol overlay.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Role of native Nuc propeptide for secretion of Nuc. Numerous secreted enzymes, including Nuc, are synthesized with an N-terminal propeptide, which may be subsequently cleaved to generate the mature protein (26, 46). In many cases (but not for Nuc) enzyme activity requires processing of the propeptide to mature protein form. There is evidence that propeptides may also improve secretion efficiency. Recent studies using E. coli, comparing the secretions of NucB and NucA fusions to an E. coli-derived signal peptide, indicate that the 19-residue Nuc propeptide improves secretion efficiency and confers SecA-independent secretion of Nuc (49). Our results show that secretion of the N-terminal truncation of Nuc (NucT) fused to its own signal peptide is also poorly efficient in L. lactis. The first 10 amino acid residues of NucT have a net charge of +3 compared to a neutral net charge for Nuc (Fig. 4 and Table 2). As recently proposed (49), the positive charge at the N-terminal region of NucA may inhibit its secretion efficiency by a possible interaction with the negatively charged heads of the membrane phospholipids. Other studies of E. coli confirm the importance of the mature moiety in secretion efficiency (2, 9, 18, 57, 58). Results with NucT, showing poor secretion efficiency of a protein containing a positively charged N terminus, extend these observations to L. lactis.

Alterations in the Nuc mature moiety improve secretion efficiency of Nuc and recombinant proteins. In our model system, about 60% of native Nuc was secreted, while the rest was detected as cell-associated precursor. This is in contrast to most previous studies, in which the initial proteins were well processed, so any enhancement of secretion efficiency would not have been detected. We exploited pre-Nuc accumulation to show that the peptide LEISSTCDA improved secretion efficiency to ~80% when inserted just after the Nuc signal peptide. This improvement was accompanied by a two- to fourfold increase in the amounts of Nuc-secreted product (up to 20 mg per liter). Similar experiments in which several synthetic peptides were inserted just after the signal peptide in a heterologous hybrid protein in B. subtilis resulted in slightly enhanced secretion; however, processing occurred at an altered cleavage site (14). In contrast to the results of those previous studies, LEISSTCDA-Nuc fusions conserve the original cleavage site. Improved secretion efficiency by LEISSTCDA insertion was even more dramatic for the protein devoid of its propeptide (NucT); in this case, the level of secreted product rose from ~30 to ~90% of the total protein. Improved secretion efficiency was also shown for AmyS. It thus appears that the sequence information in the first 10 amino acid residues of the mature moiety are critical for efficient secretion driven by the Nuc signal peptide in L. lactis. These results show that LEISSTCDA acts as efficiently as the native propeptide in enhancing Nuc secretion. It will be of interest to determine whether, like the Nuc propeptide in E. coli (49), the presence of LEISSTCDA obviates the need for SecA in NucT translocation.

How does LEISSTCDA insertion improve secretion? LEISSTCDA alters the N terminus of the mature protein by introducing two negative charges at the +2 and +8 positions. Insertion of this peptide could affect precursor conformation and thus facilitate its processing by cytoplasmic secretory chaperones, or it might optimize the charge balance around the signal cleavage site to facilitate translocation. Both of these effects could be involved in improved secretion. The presence of a negatively charged amino acid residue at position +2 is particularly common in secreted proteins (57). Studies of E. coli have led von Heijne to propose the positive-inside rule, in which the charge at the N terminus of a precursor protein should be superior to the charge surrounding the cleavage site (including the N terminus of the mature protein) for efficient translocation (57). Our results conform to this rule and may further suggest that negative charges in the mature protein can enhance export. We propose that peptides like LEISSTCDA could be of particular interest for the secretion of recombinant proteins.

Secretion of heterologous proteins in L. lactis. Several gram-positive organisms, including B. subtilis (for reviews, see references 26 and 46), Streptococcus gordonii, Staphylococcus xylosus, Staphylococcus carnosus, Listeria monocytogenes, C. glutamicum, and Lactobacillus plantarum, have been successfully used either for export of recombinant proteins or for antigen display on the cell surface. We selected L. lactis as a potential secretion host, rather than the above-mentioned organisms, because it has the particularity of being one of the major species used in the food industry, which ensures that it is nontoxic. L. lactis has already been successfully tested as a vaccine vector (30). It is also the best characterized of the lactic acid bacteria and will allow us to examine secretion at the genetic level. The parameters identified in the present study that affect secretion efficiency will be helpful in directing a choice of secretion vector. In these studies, export efficiency in L. lactis was improved by altering the N-terminal sequence of the mature protein to be exported. These results should allow us to examine and potentially optimize secretion efficiency of other fusion proteins in L. lactis.