Nisin-Controlled Extracellular Production of Interleukin-2 in Lactococcus lactis Strains, without the Requirement for a Signal Peptide Sequence

Secretion of the cytokine interleukin-2 (IL-2) was investigatedin Lactococcus lactis using the secretory machinery of the bacteriocinlactococcin A. Surprisingly, the lcnCD transport genes werenot essential for mouse IL-2 secretion. Furthermore, expressionof a mature mouse IL-2 gene resulted in interleukin secretionwithout the requirement for a leader sequence.

INTRODUCTION

Top

INTRODUCTION

In most cases secretion of heterologous proteins and peptidesin Lactococcus lactis has involved fusion to a signal peptiderecognized by the general secretory (Sec) pathway (6, 8, 21).Also, the dedicated secretion systems (10) associated with bacteriocinsproduced by lactic acid bacteria have been exploited (19). Inthis case leader peptides serve as recognition signals for thecorresponding ABC transporters and are cleaved during exportat a specific processing site adjacent to two conserved glycineresidues located at positions –1 and –2 (11). Theflexibility of ABC transporters for secretion and processingof class II bacteriocins has been reported previously (1, 14,15, 16, 22). In previous studies we exploited the lactococcinA secretion pathway for lactococcal production of pediocin PA-1(14, 15) and the Escherichia coli bacteriocin colicin V (16).Here we describe lactococcal expression and secretion of theeukaryotic cytokine interleukin-2 (IL-2) using the lactococcinA secretion system under control of the nisin promoter P_nisA(17) and the unexpected observation that mouse IL-2 (mIL-2)can be secreted without the requirement for a signal peptide.

Construction of pFI2398, a chimeric gene (L-mIL-2) consisting of an in-frame fusion of sequences encoding the lactococin A leader and mature mIL-2.

Top

Construction of pFI2398, a...

Primers pAD1 (5'-CCTGAATAATATAGAGATAGGTT-3') and pIL2a (5'-CTTGAAGTGGGTGCTCCTCCGTTAGCTTC-3')were used to amplify a fragment containing P_nisA and the partof lcnA gene encoding the lactococcin A leader. Plasmid pFI2396(7) was used as the template. Primers pIL2c (5'-GAAGCTAACGGAGGAGCACCCACTTCAAG-3')and pIL2f (5'-GGATCCTTATTGAGGGCTTGTTGAG-3') were used to amplifya fragment comprising part of the mIL-2 structural gene. PlasmidpT1-IL2 (20) was used as the template.

Plasmids pFI2398 and pFI2148 were introduced into L. lactisstrains FI5876 (4, 13), FI7847 (5), and UKLc10 (23) (Table 1).The growth conditions were the growth conditions used previously(4, 5, 13, 23). Plasmid pFI2148 contains the lactococcin A-dedicatedtranslocatory function genes (lcnCD). Strains FI7847 and UKLc10were induced with nisin (at the optimal concentration, 50 ngml^–1; added at an optical density at 600 nm of 0.5), whereasFI5876 was autoinduced due to the presence of a complete nisinbiosynthesis gene cluster.

View this table:
[in this window]
[in a new window]

TABLE 1. Lactococcal strains used in this study

Western blotting of culture supernatants from all three constructsrevealed two intense bands consistent with unprocessed L-mIL-2(19.6 kDa) and processed mIL-2 (17.2 kDa) (Fig. 1A) (L-mIL-2is a chimeric protein containing the lactococcin A leader andmIL-2). Culture supernatants were assayed for the presence ofmIL-2 by slot blot analysis (Fig. 1B), and the concentrationof the cytokine was measured by a sandwich enzyme-linked immunosorbentassay (ELISA) (Table 1). For each strain, the intensities ofthe slot blot bands correlated well with the values obtainedby the sandwich ELISA. Interestingly, extracellular productionof mIL-2 was also detected in the supernatants from the transformantscarrying only pFI2398 (L-mIL-2), suggesting that extracellularproduction was independent of the lactococcin A lcnCD genes.However, in this case only unprocessed L-mIL-2 was observed,and the presence of pFI2148 was associated with a higher concentrationof mIL-2 in the culture supernatants than in the equivalentstrains lacking this plasmid (Table 1 and Fig. 1B).

View larger version (45K):
[in this window]
[in a new window]

FIG. 1. (A) Detection of mIL-2 in culture supernatant of L. lactis FI9957 by Western blot analysis. Lane 1, FI9957; lane L, MagicMark Western protein standard. (B) Slot blot detection of mIL-2 in the culture supernatants of L. lactis strains. Row I, nisin induced (50 ng ml^–1); row U, not induced with exogenous nisin. Lane 1, FI7847; lane 2, FI9947; lane 3, FI9949; lane 4, murine IL-2 (2 ng ml^–1); lane 5, murine IL-2 (0.2 ng ml^–1); lane 6, FI5876; lane 7, FI9955; lane 8, FI9957.

Construction of pFI2584, a plasmid containing the mature mIL-2 sequence under control of P_nisA.

Top

Construction of pFI2584, a...

Cytokine secretion was evaluated in the absence of a leaderpeptide. Primers pAD1 and pIL2nl3' (5'-AGCTTGAAGTGGGTGCCATTTTGAGTGCCTCC-3')were used to amplify a fragment encoding P_nisA and a part ofmature mIL-2. Plasmid pFI2398 was used as the template. PrimerspIL2nl5' (5'-GGAGGCACTCAAAATGGCACCCACTTCAAGC-3') and pIL2f wereused to amplify a fragment comprising a part of the sequenceencoding mIL-2. Western blot analysis of culture supernatantsand intracellular content samples from FI5876, FI7847, and UKLc10derivatives transformed with pFI2584 revealed that the anti-mIL-2polyclonal antibodies specifically recognized a 17.2-kDa band,which is the expected size for mIL-2 (Fig. 2A). The ELISA confirmedthat there was extracellular production of mIL-2 (Table 1).

View larger version (54K):
[in this window]
[in a new window]

FIG. 2. Western blot detection of mIL-2 in culture supernatants. Lane 1, mIL-2; lane 2, L. lactis FI10586; lane 3, L. lactis FI10537; lane 4, intracellular contents of L. lactis FI10608; lane 5, culture supernatant of L. lactis FI10608; lane L, MagicMark Western protein standard.

Previously, Steidler et al. (20) reported mIL-2 secretion usingthe usp45 signal peptide, but they observed the presence ofa single 20.1-kDa band. A pTG262 derivative plasmid (pFI2593)carrying the lactococcal usp45 signal peptide fused with thesequence encoding mature mIL-2 under the control of the nisinpromoter was engineered and introduced into L. lactis FI5876(strain FI10608). Western blot analysis of concentrated culturesupernatants of L. lactis FI10608 revealed two distinct bandscorresponding to processed and unprocessed mIL-2 (Fig. 2B),as observed when the lcnA leader was used (Fig. 1A).

The CTLL-2 assay described by Gillis et al. (9) and modifiedby Hopkins and Failla (12) was used to evaluate mIL-2 activity.The values obtained after addition of supernatants from L. lactisderivative strains to the CTLL-2 cells agreed with the resultsobtained with the sandwich ELISA. In addition, we demonstratedthat L. lactis strains with complete nisin biosynthesis geneclusters (FI9955, FI9957, and FI10585) coproduced nisin A (6.5to 7.9 µg ml^–1) and mIL-2.

To our knowledge, this is the first report of secretion of aeukaryotic peptide in L. lactis assisted by the secretory machineryof a bacteriocin. The supernatants of the transformed strainscontained a mixture of both processed and unprocessed mIL-2,which has not occurred previously using this system (7, 16).The lactococcin secretory machinery encoded by the lcnCD geneswas not necessary for secretion of unprocessed mIL-2 when thispeptide was fused to the lactococcin A leader. As a result,transformation of the three lactococcal hosts with a plasmid(pFI2584) containing only the sequence of mature mIL-2 withoutany signal peptide under the control of P_nisA led to some extracellularproduction of mIL-2. A number of proteins secreted by bacteriawithout leader sequences have been reported previously (2),but to date such secretion has involved only noneukaryotic proteins.Proteins using this nonclassical secretion pathway can be predictedusing methods such as the method described at the SecretomeP2.0 website (http://www.cbs.dtu.dk/services/SecretomeP). Thismethod gives FASTA protein sequences scores between 0 and 1,and a score greater than 0.5 is considered indicative of nonclassicalsecretion. Using this system to test the mIL-2 protein sequenceresulted in a score of 0.875 when it was tested in the mammaliancategory, which suggests that mIL-2 might be secreted in eukaryoticcells without the requirement for a signal peptide.

There is evidence that several proteins secreted by nonclassicalsystems are also present at high intracellular concentrations(3). This observation suggests that when some proteins are abundantin the cytoplasm, they may actually leak out of the cell. Inthis study, significant mIL-2 levels were detected intracellularlyas well as extracellularly. To evaluate cell leakage, the releaseof cytoplasmic lactate dehydrogenase (LDH) into culture supernatantwas investigated using a modified version of the method of Wittenbergerand Angelo (24). The results for LDH activity measured in culturesupernatants from L. lactis FI10585 did not differ from theresults for the parental strain carrying the empty vector (Fig.3A). The cytokine mIL-2 is smaller than LDH, and the possibilitythat mIL-2 is released from the cell in a passive way cannotbe dismissed.

View larger version (5K):
[in this window]
[in a new window]

FIG. 3. (A) LDH activity in culture supernatants of L. lactis FI10585 ( ${circ}$ ) and L. lactis FI5876(pTG262) ( ${square}$ ) and mIL-2 concentrations in the supernatant of L. lactis FI10585 (•). (B) Flow cytometric analysis of PI-stained bacteria during different phases of growth. PI permeability is expressed as the percentage of PI-positive events compared to the total number. The error bars indicate standard deviations of three biological replicates. Filled bars, L. lactis FI10585; open bars, L. lactis FI5876(pTG262).

In order to answer this question, flow cytometry was used toevaluate intact, permeabilized, and lysed (dead) cells in bacterialcultures of FI10585 and FI5876(pTG262). Intact, nisin-induced,and cetyltrimethylammonium bromide-treated bacteria were labeledwith thiazole orange and/or propidium iodide (PI) and analyzedby flow cytometry to measure the scatter and fluorescence intensityof individual cells. Thiazole orange labeled all cells, facilitatingdiscrimination of bacteria from the background noise, whilecells with permeable membranes were easily distinguished byPI labeling. Samples were taken during different growth phases,and the results showed that in overnight cultures the PI-positivecells accounted for 10 to 15% of the cells of both of the strains(Fig. 3B). After inoculation, the percentage of PI-permeablecells dropped gradually in the exponential growth phase to 1to 3%, which was also observed at 8 h (data not shown).

Therefore, FI10585 and FI5876(pTG262) showed no significantdifferences in membrane PI permeability, suggesting the possibilitythat mIL-2 is secreted without the requirement for a signalpeptide by an alternative ABC transporter, as has been reportedpreviously for other secreted leaderless bacterial peptides(18).

ACKNOWLEDGMENTS

This work was partially supported by grant EX2005-0015 fromthe Ministerio de Educacion y Ciencia, Madrid, Spain. A.F. holdsa fellowship from the Universidad Complutense de Madrid, Spain.

We are grateful to Claire Shearman and especially to Udo Wegmann(IFR, Norwich, United Kingdom) for his interesting and stimulatingcomments on the work and for providing L. lactis UKLc10.

FOOTNOTES

* Corresponding author. Mailing address: Commensals & Microflora, BBSRC Institute of Food Research, Norwich Research Park, Colney, Norwich NR4 7UA, United Kingdom. Phone: 441603-255397. Fax: 441603-255288. E-mail: antonio.dumont{at}bbsrc.ac.uk

Published ahead of print on 28 September 2007.

REFERENCES

Top