A method for the identification of proteins secreted by lactic acid bacteria grown in complex media


  • Editor: Wolfgang Kneifel

Correspondence: Borja Sánchez, Laboratoire de Microbiologie et Biochimie Appliquée, Université de Bordeaux, UMR 5248 CBMN, UBX1-ENITAB, ENITAB, 1 cours du Général de Gaulle, 33175 Gradignan Cedex, France. Tel.: +33 5 57 35 59 92; fax: +33 5 57 35 07 39; e-mail: b-sanchez@enitab.fr


Lactic acid bacteria (LAB) are known for their special nutritional requirements, being usually cultured in complex media to achieve optimal growth. In this paper, a protocol based on trichloroacetic acid precipitation of peptides and proteins is presented. The method has been tested on four probiotic LAB strains grown in De Man Rogosa Sharpe (MRS) broth, a complex medium that is often used for the culture of such bacteria. This protocol allowed the detection of 19 proteins after sodium dodecyl sulfate-polyacrylamide gel electrophoresis, 10 of them being successfully identified by tandem MS. Thereafter, the 10 were found to be secreted or surface associated by bioinformatic means. In conclusion, this work supplies a method for the identification of proteins secreted by LAB, allowing discrimination between the proteins present in the MRS and those produced by probiotic LAB.


Secreted proteins are thought to play essential roles in the molecular intercommunication between host–bacteria, and in the monitoring of the bacterial environment (van Pijkeren et al., 2006). In commensal and probiotic lactic acid bacteria (LAB), these proteins could be responsible for bacterial–intestinal cell intercommunication, and for certain probiotic traits such as pathogen inhibition and immunomodulation (Buck et al., 2005). Thus, identification of proteins secreted by probiotic LAB is crucial for elucidating their mechanism of action.

Secreted proteins are transported from the bacterial cytoplasm to the bacterial environment. This is usually achieved by the presence of a signal peptide in the N-terminal part of the protein, which directs the protein toward the secretion machinery (van Wely et al., 2001). Included in this group are some surface-associated proteins that are released into the external medium due to the physiological turnover of the cell wall (Turner et al., 2004).

Although several scientific papers describe the precipitation of secreted proteins in probiotic bacteria, they usually start from chemically defined media, in order to avoid the presence of proteins originating in the protein extracts that are part of such media (Trost et al., 2005; Sánchez et al., 2008). In contrast, some LAB strains are not able to grow or grow deficiently in defined media, making difficult the identification of the secreted proteins following those methods.

In the present work, a trichloroacetic acid (TCA)-based protocol for the precipitation and identification of proteins secreted by LAB is presented. This method allowed the identification of the proteins secreted by four probiotic LAB, which were grown in De Man Rogosa Sharpe (MRS) medium.

Materials and methods

Bacterial strains used and growth conditions

Lactobacillus gasseri B3, Lactobacillus reuteri Protectis and Lactobacillus rhamnosus R-11 were isolated from the probiotic products Bion®3, Stimulobiotic and Biotravel, all manufactured in France. Lactococcus lactis n35 was isolated from a sheep artisanal cheese.

Strains were identified at the species level by partial 16S rRNA gene sequencing using the universal 20F (5′-AGAGTTTGATCATGGCTCAG-3′) and 1500R (5′-GGTTACCTTGTTACGACTT-3′) primers (Weisburg et al., 1991). Sequences were used to query the GenBank database, the strain identification score being shown in Table 1. Strains were routinely grown aerobically without shaking at 37 °C in MRS broth (Becton Dickinson France SAS, Le Pont-De-Claix, France). Lactococcus lactis n35, able to grow in MRS, was used as positive control.

Table 1.   Secreted proteins identified in the supernatant of several LAB strains
Bands*Putative functionMicroorganismAccession no.MMplMS/MSMWE§SPpsortb
  • *

    Codes refer to bands marked by arrows in Fig. 1.

  • Protein sequence GI number.

  • Fragmented MS/MS peptides allowing the identification of the protein.

  • §

    MOWSE score resulting from the ion MS/MS search against the nonredundant NCBI protein database. All scores are statistically significant (P<0.05).

  • Signal peptides were predicted using the psortb package (Gardy et al., 2005).

  • Final subcellular localization was predicted using the psortb package (Gardy et al., 2005).

  • MM, molecular mass.

S1Serpin B1Sus scrofagi|41718542.56.02127NoCytoplasmic
S2Hypothetical protein Usp45Lactococcus lactis ssp. lactis Il1403gi|1567421147.08.3182YesExtracellular
S3Cell wall hydrolaseLactobacillus casei ATCC 334gi|11649384949.44.9268YesExtracellular
S4Peptidoglycan-binding LysMLactobacillus reuteri 100-23gi|9208907024.94.81105YesExtracellular
S5Mannosyl-glycoprotein endo-β-N-acetylglucosamidaseLactobacillus reuteri F275gi|14854505860.49.56150YesExtracellular
S6Hypothetical protein LJ0155Lactobacillus johnsonii NCC 533gi|4251824174.59.74103YesExtracellular
S7Hypothetical protein LJ0437Lactobacillus johnsonii NCC 533gi|42518532110.06.2270YesCell wall (LPXTG)
S8Aggregation-promoting factorLactobacillus gasserigi|161959832.09.61139YesExtracellular
S9Muramidase (lysozyme subfamily 2)Lactobacillus gasseri ATCC 33323gi|11662883766.79.82143YesExtracellular
S10Peptidoglycan-binding LysMLactobacillus reuteri 100-23gi|9209014221.67.82163YesExtracellular
S11Mucus adhesion-promoting proteinLactobacillus reuterigi|992926228.69.8152YesExtracellular

Precipitation and identification of secreted proteins

Precipitation of secreted proteins was achieved by adding minor modifications to the method described by Sánchez et al. (2008). Briefly, 50-mL aliquots of fresh MRS broth were inoculated (0.1% v/v) from a 24-h culture and were grown aerobically overnight at 37 °C, at the end of which all cultures were typically at early stationary phase. Protein loadings were standardized on a volume-for-volume basis, which usually corresponded to a protein amount of around 40 μg. Aliquots of 5 mL were harvested by centrifugation (10 min, 3500 g, 4 °C), and the supernatant was filtered (0.45 μm). Ten milligrams of sodium deoxycholate (Sigma-Aldrich Chimie, Saint-Quentin Fallavier, France) were added and mixed – the resulting solution was incubated at 4 °C for 30 min. Chilled TCA (Sigma-Aldrich Chimie) was added at a final concentration of 6% (w/v) and proteins were allowed to precipitate for 2 h at 4 °C. Proteins were recovered by centrifugation (10 min, 9300 g, 4 °C) and pellets were washed twice with 2 mL of chilled acetone (Sigma-Aldrich Chimie). Pellets were allowed to dry at room temperature (RT) and proteins were resolubilized by ultrasonication (10 min, Ultrasonic bath, Deltasonic, Meaux, France) in 40 μL of 1 × Laemmli buffer (Laemmli, 1970). These 40 μL were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis using a final polyacrylamide concentration of 12.5% (w/v) (Laemmli, 1970). Selected bands were excised from gels and digested with trypsin using standard protocols, the resulting peptide mixture being analyzed by tandem MS (MS/MS). Data were acquired using a MALDI Q-Tof Premier mass spectrometer (Waters, Manchester, UK), with α-cyano-4-hydroxy-cinnamic acid (Sigma-Aldrich Chimie) used as a matrix (3.6 mg mL−1 solution in 50% acetonitrile in 0.1% aqueous trifluoroacetic acid). Monoisotopic masses were corrected using the pseudomolecular ion of Glu-Fibrinopeptide as a lock mass (1570.6774 Da).

Proteins were identified using the MS/MS search module from the online version of mascot software (http://www.matrixscience.com) against the nonredundant protein NCBI database, using the monoisotopic masses derived from trypsinolysis. The following parameters were used: peptide charge +1, peptide tolerance ±0.1 Da, MS/MS tolerance ±0.1 Da, and one missed cleavage allowed for trypsin. Gels were repeated three times from independent cultures.

Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) enzymatic assay

In order to test the presence of small amounts of cytoplasmic enzymes in the supernatant medium, GAPDH enzymatic assays were conducted as follows: 50 μL of supernatant media were incubated with glyceraldehyde-3-phosphate 2 mM, 1 mM NAD+ in 950 μL of assay buffer (triethanolamine 40 mM, Na2HPO4 50 mM, EDTA 5 mM and dithiothreitol 0.1 mM; pH 8.6) (all reagents purchased from Sigma-Aldrich Chimie). GAPDH enzymatic activity was determined spectrophotometrically at RT by monitoring NADH apparition at 340 nm. Cytoplasmic extracts were used as positive controls. One unit of GAPDH activity was defined as the amount of protein capable of generating 1 nmol NADH min–1.

Results and discussion

In spite of the interest in the identification of secreted proteins in LAB, few reports are available in the scientific literature (Lee et al., 1997; Turner et al., 1997, 2004; Peant & LaPointe, 2004; Yan et al., 2007; Sánchez et al., 2008). Complex media might contain several peptides/proteins and other molecules that interfere with protein detection techniques, making difficult the identification of proteins secreted by LAB. The absence of a method for the systematic identification of proteins secreted by LAB, grown in complex media, prompted us to undertake this study.

The method described in the present paper allowed the simple precipitation and identification of the proteins secreted by the four LAB strains used in this study. In total, 19 bands yielded good tryptic profiles after tandem MS analysis (MS/MS). Ten of them were successfully identified (Table 1), and nine not identified (labeled with asterisks in Fig. 1). For unidentified bands, it seems that databases do not yet contain homologs of such proteins. Bands that did not yield good tryptic profiles are not indicated in Fig. 1. Secreted protein profiles were consistent between replicates and between MRS batches; however, variations might appear if a different MRS is used or if changes in any environmental parameters are introduced. In this way, it is known that L. lactis is able to grow and produce several peptides in media derived from molasses, soybean and different source of MRS, the amount of secreted proteins being affected by pH or the presence of surfactants (Todorov & Dicks, 2004; Rodrigues et al., 2006; Xiao et al., 2007).

Figure 1.

 Representative sodium dodecyl sulfate polyacrylamide gel showing the proteins secreted by the four LAB strains used in this study. Lane 1, Lactobacillus rhamnosus R-11 total extract obtained by sonication, which evidences the low complexity of secreted cytoplasmic profiles; lane 2, porcine serpin isolated from fresh MRS; lane 3, Lactobacillus reuteri Protectis secreted proteins; lane 4, Lactococcus lactis ssp. lactis n35 secreted proteins showing protein Usp45; lane 5, L. rhamnosus R-11 secreted proteins; and lane 6, Lactobacillus gasseri B3 secreted proteins. *Proteins not identified. MM, molecular mass.

Following our methodology, the strain L. lactis n35 was shown to mostly produce a single protein of around 55 kDa (S2, Fig. 1, lane 4) identified as protein Usp45. Lactococcus lactis species has been shown to predominantly produce this 45-kDa protein, which is a chromosomally encoded protein of unknown function (van Asseldonk et al., 1990). Remarkably, a shift between the theoretical and experimental molecular masses of Usp45, 45 vs. 55 kDa, respectively, was observed. The aminoacidic analysis of Usp45 showed that the central zone of the protein (between residues 263 and 334) was rich in polar amino acids, notably serine and threonine. For this reason, post-translational modifications of such residues, like O-glycosylations, may be responsible for the mass shift, but this point needs further research.

With regard to the proteins secreted by the other strains, L. rhamnosus R-11 was shown to secrete two proteins, whereas L. reuteri Protectis and L. gasseri B3 secreted several (Fig. 1, lanes 3, 5 and 6). All identified proteins were shown to carry a signal peptide and were identified as extracellular by the psortb 2.0 software (Gardy et al., 2005). The only exception was S7, which carried a C-terminal LPXTG motif, a sequence that may allow the covalent binding of the protein to the cell wall (Siezen et al., 2006). Four proteins secreted by the strain L. gasseri B3 (labeled S8) were identified as aggregation-promoting factor, suggesting that the lower bands might be proteolytic products of the highest band. The rest of the secreted proteins are listed in Table 1 and included cell wall hydrolase (S3), peptidoglycan-binding proteins (S4 and S10), mannosyl-glycoprotein endo-β-N-acetylglucosamidase (S5), muramidase (S9), mucus adhesion-promoting protein (S11) and two hypothetical proteins (S6 and S7). Again, the finding of only proteins carrying export signals (such as signal peptides) is consistent with the aim of our protocol.

Interestingly, our method showed that MRS contained considerable amounts of a protein of about 50 kDa, which was identified as porcine serpin, a leukocyte elastase inhibitor (band S1, Fig. 1). Serpin was degraded to different degrees by the four LAB strains, and might have interfered with the detection of bacterial proteins in this zone of the molecular mass.

Finally, GAPDH enzymatic assays were performed in order to determine the presence of lysis in the supernatant media, and no GAPDH activity was detected (data not shown). Because secreted proteins were precipitated in the stationary phase of culture, cell lysis cannot be excluded, but at least it was minor.

In conclusion, this work provides an efficient, reproducible and simple protocol for the identification of proteins secreted by LAB strains. Further analysis of the function of these proteins, as well as the study of certain environmental effects on synthesis and secretion, will help to elucidate the mechanisms of action of probiotic bacteria in their ecological niches.


B.S. was the recipient of a Clarín postdoctoral contract from the Gobierno del Principado de Asturias funded by the Plan de Ciencia, Tecnología e Innovación de Asturias 2006–2009.