|Home | About | Journals | Submit | Contact Us | Français|
Overexpression of the lactococcal CsiA protein affects the cell wall integrity of growing cells and leads to leakage of intracellular material. This property was optimized and exploited for the targeted release of biologically active compounds into the extracellular environment, thereby providing a new delivery system for bacterial proteins and peptides. The effects of different levels of CsiA expression on the leakage of endogenous lactate dehydrogenase and nucleic acids were measured and related to the impact of CsiA expression on Lactococcus lactis cell viability and growth. A leakage phenotype was obtained from cells expressing both recombinant and nonrecombinant forms of CsiA. As proof of principle, we demonstrated that CsiA promotes the efficient release of the heterologous Listeria bacteriophage endolysin LM4 in its active form. Under optimized conditions, native and heterologous active-molecule release is possible without affecting cell viability. The ability of CsiA to release intracellular material by controlled lysis without the requirement for an external lytic agent provides a technology for the control of both the extent of lysis and its timing. Taken together, these results demonstrate the potential of this novel approach for applications including product recovery in industrial fermentations, food processing, and medical therapy.
Lactococcus lactis is a GRAS (generally regarded as safe) Gram-positive bacterium with a long history of widespread use in the food industry for the production and preservation of fermented products. Accordingly, it has been the subject of molecular genetic analysis and biotechnological research aimed at enhancing its existing applications and generating new technologies. In these contexts it has been developed as a host for recombinant protein expression. In addition to food applications, L. lactis has been exploited as a vehicle for the delivery of bioactive compounds to the human and animal gastrointestinal (GI) tracts, where its resistance (e.g., to acid) facilitates its own survival and protects the bioactive “payload” that it carries (2, 9, 15, 18, 19, 27). For some applications it is important to facilitate the release of homologous intracellular enzymes or active recombinant proteins in order to harvest the product, release the product into a food matrix, or deliver a bioactive compound into the GI tract. There is industrial interest in the lysis of lactococcal cells in situ during cheese making for the release of cytoplasmic peptidases, because these enzymes accelerate protein and peptide breakdown for optimal cheese maturation (1). Since the ripening of cheese can be both a slow and a costly process, controlling the rate and level of lysis would be extremely beneficial from the point of view of cheese manufacturers (9).
Strain MG1363 of L. lactis contains a conjugative element called the sex factor integrated into its chromosome (7, 8, 26). The csiA gene of the sex factor encodes a protein whose function is essential for DNA conjugal transfer (22). It has the capacity to inhibit the last stages of cell wall biosynthesis and is predicted to enable the assembly of the DNA secretion machinery through the cell envelope. CsiA is an 870-amino-acid protein that has homologues in the related genera Enterococcus and Streptococcus. In its C-terminal moiety, the protein contains a highly conserved domain (HCD) and a cysteine, histidine-dependent amidohydrolase/peptidase (CHAP) domain located at the C-terminal end. When overexpressed, the protein causes a cell lysis phenotype, and we have shown previously that the HCD is responsible for this (22). The cell-lytic phenotype of CsiA is exploited here for the release of biologically active compounds into the environment of growing cells.
We measured the effects of different levels of CsiA expression on L. lactis cell viability, cell integrity, and the leakage of DNA and endogenous cytoplasmic lactate dehydrogenase (LDH). We optimized CsiA expression to promote the release of intracellular proteins without impairing culture viability and growth. We demonstrate that controlled expression of the csiA gene leads to the release of an active heterologous enzyme, the listerial LM4 bacteriophage lysin, thereby establishing a novel approach for the delivery of intracellularly expressed proteins and peptides to the external environment. This lactococcal technology has potential for the manufacture of recombinant proteins and peptides as well as for the release of bioactive compounds into a food matrix or the human or animal GI tract. This work has been protected by a patent filing and is being commercialized through PBL (Norwich, United Kingdom).
All L. lactis strains used in this work are described in Table Table1.1. Recombinant strains expressing LM4 endolysin were obtained as follows. First, the LM4 lysin gene, integrated into the lacG gene of the lactose operon (17), was transferred by conjugation from strain FI7800 to the sex factor-negative recipient strain FI9012. The resulting transconjugant was transformed with plasmid pUK200, pFI2640, or pFI2641 (Table (Table1).1). The nisin-sucrose conjugative transposon Tn5307, containing an inactivated nisA gene (3), was introduced into the three resulting transformants by conjugative transposition using strain FI9979 as a donor. The presence of transposon Tn5307 allows controlled gene expression from the PnisA promoter in trans with externally added nisin. All conjugation experiments were performed as described previously (21), with donor and recipient mixtures grown on nonselective medium for 16 h prior to the selection of transconjugants on McKay's indicator plates (13) containing the appropriate source of carbon and antibiotics. The resulting strains, FI10717, FI10718, and FI10719, were tested for their abilities to release the LM4 endolysin following nisin induction. To test LM4 lysin activity, the target strain Listeria monocytogenes F6868 (11) was used.
L. lactis strains were grown at 30°C in M17 medium (Oxoid) with either 0.5% glucose (GM17), 0.5% sucrose (SM17), or 0.5% lactose (LM17). Antibiotic resistance markers in L. lactis were selected using chloramphenicol at 5 μg/ml, streptomycin at 200 μg/ml, or rifampin at 200 μg/ml. L. lactis electrocompetent cells were prepared and transformed by the methods of Holo and Nes (10). For LM4 lysin activity assays, Listeria monocytogenes was grown in brain heart infusion (BHI; Oxoid) broth at 30°C.
The cells were grown to an optical density at 600 nm (OD600) of 0.5, and nisin was added to the culture. After 2 h of incubation at 30°C, cell viability and detergent-mediated lysis (which reflects cell wall damage) were assessed. Cell viability was determined via flow cytometry and via CFU counts of serial dilutions in phosphate-buffered saline (PBS) plated on GM17 agar. Cell lysis was measured after the addition of 0.5% sodium dodecyl sulfate (SDS) as described previously (22). To measure LDH release, a 10-ml culture was centrifuged at 4,000 rpm for 10 min, and the supernatant was collected. A colorimetric LDH assay (Bioassay Systems, Hayward, CA) was used to assess the release of the enzyme into the supernatant. The intracellular LDH activity present in the cytoplasmic fraction versus that in the supernatant was measured in 5-ml samples. The culture was centrifuged, and the cell pellet was resuspended in Tris buffer (pH 7.2) and was mechanically homogenized (with a FastPrep FP120 instrument). LDH measurements were performed on the soluble fraction and on a filter-sterilized culture supernatant.
Bacterial samples were stained with Sytox Green (Invitrogen) at 2.5 μM and were analyzed with an influx flow cytometer (Cytopeia). Nucleic acid staining was measured by quantifying the median Sytox Green G fluorescence by using a photomultiplier (PMT) with a 528/538-nm band-pass filter. Morphological information was obtained by quantifying forward and side-scatter using PMTs. In all cases, 100,000 events were acquired. Data were analyzed using FlowJo software (Tree Star).
Total RNA from bacterial cells was extracted as follows. A mixture of phenol and ethanol (1:20, vol/vol) was added to the cell cultures to a final concentration of 20% in order to stop further transcription and prevent RNA degradation. The suspension was maintained on ice for 30 min. Cultures were centrifuged at 5,000 rpm for 10 min. Cells were resuspended in lysis buffer (5 mg/ml lysosyme and 10 U/ml mutanolysin in Tris-EDTA buffer [pH 8.0]) and were incubated for 30 min at 37°C. RNA was prepared using the Promega SV total RNA purification kit. RNA samples were treated with Ambion Turbo DNase to eliminate DNA contamination. The csiA transcript was quantified using the RNA UltraSense One-Step Quantitative RT-PCR system (Invitrogen) under the conditions described by the manufacturer. Two primers generating a 1.2-kb internal csiA fragment were used. An appropriate number of 20 to 30 cycles was chosen to ensure that the reverse transcription-PCR (RT-PCR) amplification was in the linear range. RT-PCR products were analyzed by agarose gel electrophoresis. The signals detected were quantified using TotalLab software (Nonlinear Dynamics).
Overnight cultures of the different strains were diluted 100 times in fresh M17 buffered with potassium phosphate (0.2 M; pH 7.0). The cells were grown to an OD600 of 0.6, and the culture medium was split into two sets of parallel samples: a control culture and a culture to which 10 ng/ml of nisin was added to induce csiA expression. The cells were grown for 16 h and were then centrifuged at 4,000 rpm for 10 min. The supernatants of these cultures were filter sterilized and were then either used neat or concentrated five times by filtration using a 6-ml Vivaspin concentrator column (molecular weight [MW], 5,000; Sartorius). Ten microliters of the different supernatants was loaded into preformed wells in an agar layer (1% agar in 0.2 M potassium phosphate buffer [pH 7.0]) inoculated with a 1:100 dilution of autoclaved Listeria monocytogenes FI6868 cells (11) from a 100-ml overnight culture. The agar plate was incubated for 24 h at 30°C, and lytic zones reflecting the activity of LM4 on Listeria target cells were monitored.
Overnight cultures of strains FI10703 and FI10704 were diluted 100 times in SM17 containing 5 μg/ml chloramphenicol, and the cells were grown to an OD of 0.5. Nisin at a final concentration of 10 ng/ml was added to the culture, and the cells were grown for a further 16 h. A 1-ml aliquot of the culture was removed; the cells were harvested by centrifugation (4,000 rpm); and the supernatant was discarded. Cells were washed twice with isotonic sucrose buffer and were concentrated 10 times in isotonic sucrose. The samples were then incubated on cleaned glass slides for 40 min. Excess liquid was wicked off the slides using filter paper, and the slide (with the now-adherent cells) was washed by dipping into 250 ml of water and then air dried. Imaging was carried out in contact mode in air, on a combined atomic force microscope (AFM)-inverted optical microscope (Lumina; Veeco Instruments) using 200-μm-long V-shaped silicon nitride cantilever tips (NP; Veeco Instruments). Scan rates of 1 to 2 Hz and forces of around 3 to 5 nN were employed throughout.
The impact of CsiA on bacterial cell wall integrity makes it possible to develop a system for the controlled delivery of bioactive molecules. However, such an application depends on first defining the impact of csiA expression levels both on cell viability and on the release of intracellular contents. To evaluate cell viability, two different expression levels of CsiA were obtained following induction of csiA by exogenous nisin in different Lactococcus lactis strains: a control strain that does not express CsiA, a strain expressing the CsiA protein, and a strain expressing CsiA lacking the C-terminal CHAP domain (CsiAΔCHAP). The latter form of the protein defines a deletion variant of CsiA that retains the effect on cell wall integrity observed with the whole protein. Each strain contains a chromosomal copy of the complete nisin operon with the structural gene nisA inactivated (3), enabling good control of the activity of the nisin promoter, PnisA, through the addition of varying amounts of exogenous nisin. The growth rate of each strain induced with 1 ng/ml or 10 ng/ml nisin was monitored (Fig. (Fig.1).1). The growth of cells expressing CsiA or CsiAΔCHAP was affected, with the most dramatic effect observed for cells induced with 10 ng/ml nisin. The growth rate of the control cells was not affected by the presence of nisin, showing that the effect observed for CsiA- or CsiAΔCHAP-expressing cells is due to the production of that protein. We know from previous studies that the cell walls of cells overexpressing CsiA do not require enzymatic digestion prior to SDS-mediated cell lysis because of the capacity of CsiA to cause damage to growing cells by altering the cell wall structure (22). Figure Figure11 presents the results of cell damage measurements based on SDS-mediated lysis, and these correlate with the growth rates, since significant detergent-mediated lysis was observed for a 1-ng/ml nisin induction and the most dramatic effect was observed for cells in which CsiA and CsiAΔCHAP expression was induced with 10 ng/ml nisin. No cell damage was observed for the control cells. The combination of these results indicates that controlled csiA expression can be used to create a cell damage phenotype without affecting cell viability. Significant cell damage (17%) could be observed after 2 h of induction with 1 ng/ml of nisin added to CsiA-expressing cells without affecting their culturability. The same phenomenon could be observed for cells expressing CsiA with its CHAP domain deleted, confirming that the HCD of CsiA is solely responsible for the cell damage phenotype (22). However, in both cases, the growth rate observed under these conditions was reduced. Flow cytometry analysis indicated that the integrity of 10% of the CsiA-expressing cells was already affected with 0.5 ng/ml nisin, compared to only 0.3% of the control cells (Fig. (Fig.2);2); this could explain the reduction in growth observed for cells induced with 1 ng/ml nisin.
We undertook an analysis of the impact of CsiA expression on the release of the intracellular cytoplasmic lactate dehydrogenase (LDH) enzyme in the absence of in situ cell lysis. This involved subjecting a growing lactococcal culture (optical density, 0.15) to different levels of nisin induction and subsequently monitoring the release of LDH into the culture medium. It is remarkable that, as shown in Table Table2,2, cells expressing CsiA at levels that did not affect growth (0.5 ng/ml nisin) released nearly 6 times more LDH than control cells, with approximately 76% of the activity detected in the supernatant. At higher levels of CsiA expression that affected growth but not survival (1 ng/ml nisin), 8 times more LDH was released into the medium (representing 98% of the total activity), and as much as 11.3 times more LDH activity was detected when the number of cells measured in the growth medium (expressed as the OD600) was taken into account. We show here that controlled csiA expression can be used to create a cell leakage phenotype. The lower activity of LDH released from cells grown in the presence of 10 ng/ml nisin (rather than 1 or 2.5 ng/ml) could result from negative regulation of LDH activity in cells whose metabolic activity is dramatically reduced. In conclusion, we defined the optimum growth conditions of CsiA-expressing cells for maximum efficiency of release of intracellular proteins without affecting cell viability. This information is of value for the future development of a low-cost production technology and for the delivery of enzymes or other bioactive molecules.
To confirm that intracellular release results from cell leakage and not from in situ cell lysis, the release of macromolecular nucleic acid into the growth medium was measured under different conditions of CsiA expression. At a low nisin concentration (0.5 ng/ml), no significant difference was observed in the amounts of nucleic acid released by the CsiA-expressing strain FI10704 and the control strain FI10703 (Fig. (Fig.2).2). This result indicates that the 6-fold-increased LDH release measured under these conditions (Table (Table2)2) is mainly a consequence of cell leakage rather than cell lysis. This also explains why cell growth is not affected. At a higher nisin concentration (5.0 ng/ml), the CsiA-expressing strain released as much as 10-fold more nucleic acid than the control strain. In this case, the increase in LDH release measured in the supernatant (Table (Table2)2) is likely to result from a combination of cell lysis and cell leakage, a hypothesis supported by the flow cytometry data, which show >45% of the cells stained with Sytox Green, indicating altered cell integrity (Fig. (Fig.2D).2D). In addition, all cells showed an increased scatter, suggesting morphological changes related to cell integrity. We also confirmed that cell contents are not released from CsiA-expressing cells during the centrifugation process used to separate the culture medium from the cells. Indeed, no differences in DNA release were observed before and after centrifugation (data not shown), indicating that although the cell structure became fragile following the action of CsiA, cell integrity was maintained and centrifugation did not affect intracellular protein release.
To visualize the effect of high expression of CsiA (induced with 5 ng/ml nisin) on cell morphology, the cells were incubated for an additional 14 h and were observed by AFM. These observations showed that the cell wall structure was no longer maintained in cells overexpressing CsiA (Fig. 3B and C), whereas control cells grown under the same conditions exhibited an intact morphology (Fig. (Fig.3A3A).
Strains constructed using conjugation are not subject to genetic modification (GM) regulations, and, as a natural process, conjugation does not cause the same consumer concerns that have impeded the introduction of GM industrial strains, especially in Europe. The csiA gene is part of a large operon involved in sex factor conjugation, and it is expressed at low levels under normal circumstances (22). Our previous studies described the cointegration of the sex factor with an autonomously replicating lactose plasmid (8) and showed that this molecular rearrangement can cause elevated expression of another gene (cluA) that is part of the same operon (20). These DNA rearrangements involve nonrecombinant and naturally occurring phenomena (transposition involving an insertion element [IS element]), and here we investigate their potential to elevate csiA expression and thereby enhance intracellular enzyme release. Quantitative RT-PCR experiments showed that the amount of csiA RNA transcript was 10 times higher in MG1827, which contains one of the cointegrate plasmids, than in the parental strain MG1363, which carries the chromosomally located sex factor (data not shown). This confirms the previously observed trend for cluA (20). The release of LDH promoted by MG1827 was compared with that for MG1629, a derivative of the MG1363 strain containing the lactose plasmid pLP712. This ensured that the two strains carried lactose plasmid DNA and thus were genetically similar. Following growth on lactose as a carbon source, the LDH activity measured in the medium of MG1827 (17.9 ± 2.7 IU/liter) was 8-fold higher than that for strain MG1629 (2.2 ± 0.6 IU/liter); this result is in agreement with the levels of LDH release measured when csiA was induced with 1 ng/ml of nisin (Table (Table2).2). The activity of the LDH released into the medium of MG1827 represented 72% of the total LDH activity. Nonrecombinant strains expressing CsiA at levels allowing cell content leakage without affecting cell viability would permit the utilization of non-GM strains in diverse applications. Another major benefit of this approach is that, although the sex factor is found only in strains related to MG1363, the conjugation machinery that it encodes facilitates efficient transfer to any strain of Lactococcus lactis (23). This enhances the potential for industrial strain development.
In order to demonstrate the potential of CsiA for the release of a heterologously expressed protein, we investigated the Listeria bacteriophage LM4 endolysin, which has commercial potential as a biological control agent (6, 6a, 6b, 6c, 6d, 6e). We hypothesized that the expression of csiA in an L. lactis strain producing the heterologous protein LM4 intracellularly would lead to the release of the active form of the heterologous protein into the external medium. We constructed three strains, each containing a control plasmid, a plasmid harboring the entire csiA gene, or a plasmid encoding CsiAΔCHAP, the shorter version of CsiA. All strains expressed the LM4 gene under the control of the lac promoter (17). The activity of the released protein was tested as described in Materials and Methods and is reflected by the formation of lytic zones on target Listeria cells (Fig. (Fig.4).4). No lytic zone corresponding to LM4 activity could be observed for L. monocytogenes cells exposed to the extracellular culture medium of a control strain, FI10717, that lacks the csiA gene (Fig. (Fig.4).4). On the other hand, clear lytic zones could be obtained with strains FI10718 and FI10719, both of which express CsiA, showing that CsiA expression facilitates Listeria cell lysis. The absence of lytic zones observed for FI10717 grown in the presence of sublethal concentrations of the inducer nisin confirmed that nisin had no effect on the release of active protein into the medium. To demonstrate that the observed lytic zones are the result of the LM4 lysin activity released into the L. lactis culture medium and are not caused by CsiA activity, strain FI10704, which expresses CsiA but not LM4, was grown under the same conditions used for FI10717, FI10718, and FI10719. The neat culture supernatant and the 5-times-concentrated supernatant were loaded onto the same plate (Fig. (Fig.4).4). No lytic zone was formed, indicating that LM4 alone was responsible for the lysis of Listeria cells. It is noticeable that in the absence of the inducer nisin, FI10718 and FI10719 still produced a small lytic zone, likely the result of the basal activity of the PnisA promoter (14, 22).
We had previously shown that CsiA lacking the C-terminal CHAP domain was not necessary to promote detergent-dependent cell lysis in cells expressing CsiA (22). In accordance with these results, we show here that cells expressing CsiA truncated in its CHAP domain produced lytic zones comparable to those produced by wild-type CsiA. In conclusion, we demonstrate here that the novel CsiA-based system permits the efficient release of biologically active compounds into the external medium and that the shorter version of the csiA gene can be used without affecting the system's capacity.
We have demonstrated that CsiA is able to release intracellular material by controlled lysis without requiring the addition of an external lytic agent. Its controlled expression in L. lactis could be used to develop a GI tract delivery vehicle for a variety of bioactive compounds, including vaccine antigens, immune modulators, and antimicrobials. For delivery to the GI tract, the mode of bioactive-compound release has remained critical until now. Many heterologous prokaryotic, viral, and eukaryotic proteins have been produced and secreted in Lactococcus lactis (16). Nonetheless, secretion efficiency depends on the use of an appropriate secretion signal, and this can lead to small amounts of secreted protein due to the limitations of the bacterial Sec system or of its ability to operate within the GI tract environment. The use of csiA to facilitate timed cell leakage opens up an alternative approach for the delivery of an intracellular “payload” of a bioactive compound.
An additional application that can be envisaged for the CsiA system is its use in accelerated flavor development for matured cheeses in the dairy industry. Of particular value for this application is the potential to produce strains of Lactococcus that express higher levels of CsiA via a process of natural recombination and gene transfer, thus complying with the food industry's current conservative desire to use only nontransgenic materials in its products. Furthermore, the lactose plasmid-sex factor cointegrate used for nonrecombinant CsiA expression contains the lactose plasmid from strain NCDO 712 (5), which encodes functions including lactose metabolism and proteinase production. Both of these activities are required for the acidification and coagulation of milk for cheese production (12). In addition, the cointegrate plasmid expresses CsiA, the activity of which is likely to promote flavor development by facilitating the release of intracellular compounds. These features make this large plasmid a valuable tool for the dairy industry. L. lactis also has potential as a production host for heterologous recombinant proteins, especially where these have a food-related application or require high-volume and/or low-cost production. The cell leakage phenotype generated by CsiA expression is of interest in this regard because protein release can be achieved without the release of significant quantities of nucleic acid.
An approach involving the controlled expression of the holin and lysin genes from a lactococcal prophage to create a “leaky phenotype” was described previously (24). In that system, the expression of chromosomally integrated holin and lysin cassettes is dependent on the presence in the cells of the lactococcal bacteriophage 31 transcriptional activator Tac31A, encoded by a high-copy-number plasmid. In contrast, the CsiA system relies on the expression of a single gene that can be expressed under the control of its native promoter (as shown with the lactose plasmid-sex factor cointegrate), bypassing the use of recombinant DNA techniques and conforming to the GM regulations. The nonrecombinant system described in this study releases 72% of LDH into the supernatant, a level comparable to that of the β-galactosidase activity detected in the supernatant of the recombinant holin-lysin system (approximately 85%).
In earlier work (4), we described the extracellular production of interleukin-2 by L. lactis without the need for a signal peptide sequence. In that work we were unable to provide an explanation for this observation, but it may be relevant to the current work. The interleukin-2 study was undertaken with L. lactis strains derived from MG1363, all of which carry a chromosomally integrated copy of the lactococcal sex factor, which encodes the csiA gene. Given that the sex factor has a very active system for gene transfer by conjugation that depends on csiA expression (albeit tightly controlled), it is conceivable that this allows some leakage of such a small peptide molecule without significant release of larger proteins, such as lactate dehydrogenase.
We thank Martin Stocks for providing helpful comments during the course of this work. We also thank Isabelle Hautefort for reading our manuscript and providing helpful comments.
This work was funded by BBSRC contract BB/FOF/191.
Published ahead of print on 12 March 2010.