lable at ScienceDirect

Anaerobe 30 (2014) 161e168

Contents lists avai

Anaerobe

journal homepage: www.elsevier .com/locate/anaerobe

Molecular biology, genetics and biotechnology

The effect of moonlighting proteins on the adhesion and aggregationability of Lactobacillus helveticus

Adam Wa�sko a, b, *, Magdalena Polak-Berecka a, Roman Paduch c, Krzysztof J�o�zwiak d

a Department of Biotechnology, Human Nutrition and Food Commodity Science, University of Life Sciences in Lublin, Skromna 8, 20-704 Lublin, Polandb Institute of Agrophysics, Polish Academy of Sciences, Do�swiadczalna 4, 20-290 Lublin, Polandc Department of Virology and Immunology, Institute of Microbiology and Biotechnology, Maria Curie-Skłodowska University, Polandd Laboratory of Medicinal Chemistry and Neuroengineering, Medical University of Lublin, Poland

a r t i c l e i n f o

Article history:Received 13 June 2014Received in revised form24 September 2014Accepted 3 October 2014Available online 13 October 2014

Keywords:Lactobacillus helveticusMoonlighting proteinsLabel-free methodAdhesionSurface properties

* Corresponding author. Department of BiotechnoScience of Food Commodities, University of Life Scien704 Lublin, Poland. Tel.: þ48 81 46 23 368; fax: þ48

E-mail address: awasko1@tlen.pl (A. Wa�sko).

http://dx.doi.org/10.1016/j.anaerobe.2014.10.0021075-9964/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

The goal of this study was to identify moonlighting proteins in Lactobacillus helveticus that play animportant role in adhesion and aggregation. The label-free method was used for identification andanalysis of expression of cellular proteins. The analysis revealed the presence of eight moonlightingproteins in the cell envelope of Lb. helveticus. The tested strains mainly differed with respect to thepresence of S-layer proteins and the level of expression of moonlighting proteins in Lb. helveticus strainT159. These surface proteins give the cell a hydrophobic character and play a role in specific interactionswith intestinal epithelium cells and with other bacteria. In Lb. helveticus T159, the S-layer associated withmoonlighting proteins could act as adherence factors, which was evidenced by the high capability ofadhesion, auto- and coaggregation. The hydrophobicity, adhesion and aggregation abilities providebiological activities in food products and they are regarded as an important criterion for probioticselection.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Lactobacillus helveticus strains are extensively used in foodfermentation, in particular as lactic acid starter cultures. They alsoplay an important role as probiotic gastrointestinal microflora [1].One of the most significant criteria for probiotic selection is thecapability to adhere to the host intestinal epithelium. In addition,the ability of lactobacilli to form multicellular aggregates is animportant property for colonization of the human gut [2]. More-over, the coaggregation abilities of Lactobacillus species may form abarrier to prevent colonization by pathogens [3]. It was found thatthe adherence of bacterial strains is related to the cell surfacecharacteristics [4]. The surface proteins are expected to haveappreciable effects on the properties of the cell wall of manyLactobacillus strains. Indigenous lactobacilli often showed thepresence of specific, highly basic, hydrophobic cell surface proteins,such as surface layer (S-layer) proteins which are involved in the

logy, Human Nutrition andces in Lublin, Skromna 8, 20-81 46 23 400.

adhesion and aggregation process [5,6]. S-layer proteins have beenidentified in several strains of Lb. helveticus [7,8]. They constitutethe outermost structure of the cell envelope, which is an array ofsingle non-covalently bound proteins. The biological functions of S-layer proteins are cell protection, determination of cell shape,molecular and ion trapping, and adhesion to surfaces [9,10].Recently, microbiologists have been exploring proteomics as a toolin research on adaptation of microorganisms to their environment.In lactobacilli, several multifunctional proteins (moonlightingproteins) have been identified as associated with the cell surfaceand/or in the extracellular space. Cell surface glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as well as elongation factorTu (EF-Tu), triosephosphate isomerase, and enolase were identifiedon the surface of Lactobacillus plantarum as molecules mediatingadhesion to intestinal epithelial cells [11e13]. In today's quantita-tive bacterial proteomics, novel non-gel-based, mass spectrometricmethods known as label-free quantitation of proteins are used [14].The rapid development of label-free quantitative proteomic tech-niques has provided fast and low-cost measurement of proteinexpression levels in complex biological samples.

This study was performed on two strains of Lb. helveticus: T103and T159, which had been previously characterized in respect of theability to produce S-layer proteins and ferulic acid esterase activity[15]. This work was aimed at more detailed proteomic analysis of

A. Wa�sko et al. / Anaerobe 30 (2014) 161e168162

tested strains. For this purpose, the label-free method was used foridentification and analysis of expression of cellular proteins. Inaddition, comparative analysis of the surface properties andadhesion and aggregation ability of both strains was performed.The objective of this study was to identify moonlighting proteinswhich play an important role in the physicochemical character andbehavior of Lb. helveticus cells.

2. Materials and methods

2.1. Bacterial strains and culture conditions

Lb. helveticus strains were kindly provided by Prof. ŁucjaŁaniewska-Trokenheim (University of Warmia and Mazury inOlsztyn). Strains had been isolated from Polish fermented milkproducts and deposited in the Polish Collection of Microorganisms(Wroclaw, Poland). Lactobacilli were stored and cultured in the deMan, Rogosa, and Sharpe (MRS) (BTL, Poland) broth supplementedwith 0.5 g/l L-cysteine. Bacteria were incubated in anaerobic con-ditions at 42 �C for 48 h.

2.2. Isolation of surface proteins

Cell surface proteins were isolated from the bacterial biomassfrom 500-ml cultures grown in MRS medium for 2 days at 42 �C inanaerobic conditions. Bacterial cultures were adjusted to the sameoptical density OD600 of 1.8. Isolation of surface proteins anddetection was performed according to the method described pre-viously [16].

2.3. Sample preparation and mass spectrometry

Protein solutions were subjected to a standard procedure ofreduction, alkylation, and trypsin digestion. The resulting peptidemixture was applied to a nanoACQUITY UPLC Trapping Column(Waters) using water containing 0.1% formic acid as the mobilephase and then transferred to the nanoACQUITY UPLC BEH C18Column (Waters, 75 mm inner diameter; 250-mm long) using anacetonitrile gradient (5e35% acetonitrile over 160 min) in thepresence of 0.1% formic acid with a flow rate of 250 nl/min. Thecolumn outlet was directly coupled to the ion source of the LTQ-Orbitrap Velos mass spectrometer (Thermo Scientific Waltham,MA, USA) working in the regime of data-dependent MS to MS/MSswitch. Other Orbitrap parameters were as follows: one MS scanwas followed by a max. 5 of MS/MS scans, capillary voltage was1.5 kV, data were acquired in positive polarity mode. For eachsample, an additional profile LC-MS spectrum was obtained forquantitative information.

2.4. Qualitative MS/MS data processing

The acquired MS/MS data were pre-processed with MascotDistiller software (v. 2.3.2, MatrixScience) and the search was per-formed with the Mascot Search Engine (MatrixScience, MascotServer 2.4) against the database composed of all Lb. helveticusproteins deposited in the NCBInr database, and their randomizedversions for FDR computation (42236 sequences; 11382912 resi-dues). To reduce mass errors, the peptide and fragment masstolerance settingswere established separately for individual LC-MS/MS runs after a measured mass recalibration as described in Ref.[17]. Other Mascot search settings were as follows: enzyme e

semiTrypsin, missed cleavages e 1, fixed modifications: Carbami-domethyl (C), variable modifications: Oxidation (M). The statisticalassessment of peptide assignments was based on the concatenatedtarget/decoy database search strategy (merged target/decoy

databases generated with in-house developed software). This pro-cedure provided q-value estimates for each peptide spectrummatch (PSM) in the dataset. All PSMs with q-values >0.01 wereremoved from further analysis. A protein was qualified for quanti-tative analysis when at least two peptides of this protein werefound in the whole identification dataset. Proteins identified by asubset of peptides from another protein were excluded from anal-ysis. Proteins that exactly matched the same set of peptides werecombined into a single group (cluster). The mass calibration anddata filtering described above were carried out with in-housedeveloped MScan software (http://proteom.ibb.waw.pl/mscan/)[17].

2.5. Quantitative MS data processing

The lists of peptides that matched the acceptance criteria fromthe LC-MS/MS runs were merged into one common list. Thiscommon list was overlaid onto 2-D heat maps generated from theLC-MS profile datasets by tagging the peptide-related isotopic en-velopes with corresponding peptide sequence tags on the basis ofthe measured/theoretical mass difference, the deviation from thepredicted elution time, and the match between the theoretical andobserved isotopic envelopes. A more detailed description of thequantitative extraction procedure implemented by our in-housesoftware is available in Ref. [18]. The abundance of each peptidewas determined as the height of a 2-D fit to the monoisotopic peakof the tagged isotopic envelope. Quantitative values were nextexported into text files, along with peptide/protein identifications,for statistical analysis performed by in-house Diffprot software[19].

2.6. Analysis of identified proteins using bioinformatics programs

The subcellular locations of the proteins were predicted using asubcellular location prediction program, Cello version 2.0 (http://cello.life.nctu.edu.tw/). Transmembrane helices in membrane pro-teins were predicted using the TMHMMServer version 2.0 program(http://www.cbs.dtu.dk/services/TMHMM-2.0/). The signal peptidewas predicted using SignalP (www.cbs.dtu.dk/services/SignalP).Physicochemical properties were predicted using ProtParam(web.expasy.org/protparam/). Moonlighting proteins wereanalyzed as previously described by Hern�andez et al. [20].

2.7. Surface analysis by FTeIR

FTeIR spectra were collected on a Nicolet 8700 FTeIR (ThermoScientific, Waltham, MA, USA) equipped with a KBr beam-splitterand Mercury Cadmium Telluride MCT/A detector in trans-reflecting mode. The spectra were recorded over the range of4000e650 cm�1. Each spectrum represents an average of 120 scansand is appodized with a Happ-Genzel function; the number of scanpoints was 8480. The spectral resolution was 8 cm-1. Each samplewas measured five times, each spectrum was baseline correctedand then the spectra were normalized and averaged. The averagedspectrawere normalized to minimize differences due to the sampleamount and the baseline was corrected in all the spectral regions.Assignment of the functional groups of FT-IR spectra was made onthe basis of Naumann et al. [21].

To prepare the sample, the freeze-dried bacteria were smashedon an aluminum coated microscopic glass. The background spec-trum was recorded at every sample. Atlus microscopy Software forOMNIC-8 was used for measuring and analysis of the FTIR spectra.

A. Wa�sko et al. / Anaerobe 30 (2014) 161e168 163

2.8. Scanning electron microscopy

Bacteria were transferred directly from the culture ontomicroscopic stubs, frozen in liquid nitrogen, and transferred to thecryo system (Polaron Range, PP 7480). The samples were sublimed(15 min, �85 �C), sputter-coated with platinum and transferredinto the microscope chamber. The bacteria were observed with theuse of an in-lens detector, EHT 3 kV, SEM (Zeiss, ULTRA PLUS).

2.9. Determination of bacterial hydrophobicity

The cell surface hydrophobicity was determined by microbialadherence to solvents (MATS) according to Bellon-Fontaine et al.[22] with slight modifications. Briefly, untreated and LiCl treatedbacterial cells were washed and suspended with PUM buffer(22.2 g/l of potassium phosphate trihydrate, 7.26 g/l of monobasicpotassium phosphate, 1.8 g/l of urea, and 0.2 g/l of magnesiumsulfate heptahydrate, pH 7.1) to ca. 108 CFU ml�1. The cell suspen-sion (5 ml) was mixed with 1 ml of a solvent. The mixtures wereincubated at 30 �C for 10 min, subsequently vortexed vigorously for2 min, and then allowed to stand for 15min at room temperature toensure complete separation of the organic and aqueous phases. Theabsorbance of the aqueous layer was measured at 600 nm. Theaffinities to the solvent were expressed using the formula: (1 e A/A0) � 100.

2.10. Bacterial adhesion capacity

Human colon adenocarcinoma cell line HT29 (ATCC no. HTB-38)was used to assess the bacterial adhesion capacity. Cells werecultured in RPMI 1640 medium supplemented with 10% fetal calfserum (FCS; Gibco™, Paisley, UK) and antibiotics (100 U/ml peni-cillin and 100 mg/ml streptomycin; Sigma, St. Louis, MO) at 37 �C ina humidified atmosphere with 5% CO2. Cells were seeded onto a 24-well tissue culture plate (Nunc, Roskilde, Denmark) at a concen-tration of 5 � 105 cells ml�1. After 24 h of incubation, a monolayerwas obtained. The bacterial strains were resuspended in the HT29growth medium to a final concentration of 5 � 107 cells ml�1 and1 ml of each suspension was added to appropriate wells of theculture plate. After 2 h of incubation, the monolayers were washedthree times with phosphate-buffered saline (PBS with Ca2þ andMg2þ ions, pH 7.4) to remove bacteria that did not attach to theHT29 cells. Thereafter, the cells were lysed using 0.1% (v/v) Triton-X100 (Sigma, St. Louis, MO) and the number of viable adherentbacteria was determined by plating serial dilutions on MRS agarplates. The results of adhesion assays are expressed as the adhesionindex for each strain. The adhesion index is defined as the numberof bacterial cells adhering per 100 epithelial cells [23].

Fig. 1. Location of the LC-MS/MS identified proteins in Lb. helveticus T103 and T159 (a).Location of proteins with a high level of expression in Lb. helveticus T159 (b).

2.11. Auto- and coaggregation assays

The autoaggregation assay was done according to the methodsdeveloped by Golowczyc et al. [24] with slight modification. Briefly,the lactobacilli were harvested at the stationary phase, collected bycentrifugation (10000 � g for 10 min), washed twice, and resus-pended in PBS (pH 7.2). In all the experiments, the bacterial sus-pension was standardized to OD600 ¼ 1.0 (2 � 108 CFU ml�1).Optical density was measured in a spectrophotometer (Biorad,USA) at regular intervals (2, 3, 4, and 5 h) without disturbing themicrobial suspension and the kinetics of sedimentation was ob-tained. The autoaggregation coefficient (ACt) was calculated at ttime as:

Act ¼�1� ODt

ODi

�� 100

where ODi is the initial optical density at 600 nm of the microbialsuspension and ODt is the optical density at t time.

In the coaggregation assay, lactobacilli suspensions were ob-tained as described previously. Pathogenic bacteria were harvestedin the stationary phase by centrifugation during 4 min at 5000 � gand resuspended in PBS (pH 7.2). One milliliter of Lactobacillussuspension and 1ml of pathogenic bacterial suspension at the sameoptical density (OD600 ¼ 1.0) were mixed. Optical density wasmeasured at regular intervals (2, 3, 4, and 5 h) in order to obtain thekinetics of sedimentation. The coaggregation coefficient (CCt) wascalculated as:

CCt ¼��Ax þ Ay

��2�� Atðxþ yÞ

Ax þ Ay�2

� 100

2.12. Statistical analysis

Values from all the tests performed are the means of threeseparate experiments ±standard deviation. The data were analyzedusing the Excel statistical package. Statistical significance wasdetermined by the Student's t-test and set at P < 0.01.

3. Results

3.1. Proteomic analysis

In total, 139 cell wall-associated proteins were identified.Among these, cytoplasmic, membrane, cell wall, and extracellularproteins were identified based on the subcellular location predic-tion program Cello version 2.0 (Fig. 1a). To reduce the complexity,for the quantitative analysis we selected only proteins with a higherexpression ratio in Lb. helveticus T159 (Fig. 1b, Table 1). Within this

Table 1Representative proteins which exhibit a high expression ratio in Lb. helveticus T159.

Gene accessionno. (gi NCBI)

Protein name MW (kDa)/pI Cellular location Signal peptide Transmembrane Ratio Peptides

301350785 Surface layer protein 43.2/9.05 Cell wall 1-30/yes 1/yes 1/2 17403515975 ATP-dependent Clp protease, ATP-binding subunit 78.7/5.33 Cytoplasmic 0 0 1.44 8385813800 5-nucleotidase/2,3-cyclic phosphodiesterase

related esterase91.4/8.43 Extracellular yes/1-30 0 1.86 5

160348160 P-enolpyruvate-protein p-transferase PTS 63.8/4.76 Cytoplasmic 0 0 2.51 6260102334 L-lactate dehydrogenase 1 35.1/5.6 Cytoplasmic 0 0 1.65 5260102904 Chaperone GroELa 57.63/5.15 Cytoplasmic 0 0 1.22 14403079507 Enolasea 47.14/5.16 Cytoplasmic 0 0 1.32 4260103202 Peptidylprolyl isomerase 33.61/9.82 Extracellular 1-30/yes 0 1.74 16160348918 Arginyl-tRNA synthetase 63.25/5.5 Cytoplasmic 0 0 1.53 4403080234 Oligopeptide ABC transporter substrate binding proteina 65.97/9.54 Extracellular 1-21/yes 0 1.16 8403080577 30S ribosomal protein S1a 44.3/5.32 Cytoplasmic 0 0 1.22 5403514657 UDP-glucose 4-epimerase 1.19 6160347772 Phosphoglycerate mutasea 28.1/5.39 Cytoplasmic 0 0 1.29 4260083972 Methionine-tRNA ligase 75.1/5.97 Cytoplasmic 0 0 1.19 7260083356 PTS family mannose porter, IIAB component 36.52/6.33 Periplasmic 0 0 1.33 3260102544 Glucose-1-phosphate adenylyltransferase 26.6/4.68 Cytoplasmic 0 0 1.15 2403080832 Phosphoglycerate kinasea 42.89/5.25 Cytoplasmic 0 0 1.22 30403081293 Ribonucleotide-diphosphate reductase subunit alpha 81.44/8.83 Cytoplasmic/

membrane0 0 1.25 5

403081062 Glycerol-3-phosphate ABC transporter 47.7/9.68 Extracelullar yes/1-32 0 1.28 4385814063 Membrane GTPase involved in stress response 68.6/5.39 Cytoplasmic 0 0 1.28 8417008672 Molecular chaperone DnaKa 65.67/5.0 Cytoplasmic 0 0 1.1 18323466329 Elongation factor Tsa 37.8/5.47 Cytoplasmic 0 0 1.4 6161507712 Dipeptidase 54.91/4.99 Cytoplasmic 0 0 1.26 2161506883 50S ribosomal protein L14 13.21/9.93 Cytoplasmic 0 0 1.21 3403514311 Elongation factor G 76.82/4.98 Cytoplasmic 0 0 1.54 12

a Moonlighting proteins.

A. Wa�sko et al. / Anaerobe 30 (2014) 161e168164

group, 8 proteins were identified as moonlighting and these wereanalyzed for predicting Intrinsically Disordered Proteins/Intrinsi-cally Disordered Regions (IDPs/IDRs) using the ProDos program.The results show that enolase, GroEl, phosphoglycerate mutase,oligopeptide ABC transporter, phosphoglycerate kinase, and elon-gation factor Ts lack disordered regions, whereas DnaK and 30Sribosomal protein S1 had large stretches of predicted IDRs, whichallow inclusion of this protein in the IDP class.

Fig. 2. SEM micrographs of Lb. helveticus T103

3.2. Scanning electron microscopy

The SEM data are presented in Fig. 2. SEM images showed dif-ferences in the cell morphology of the tested strains. Both strainsexhibited characteristic rod-shaped cells. The surface structure ofLb. helveticus T 103 showed locally some bumps on an otherwisesmooth surface. Moreover, the cells of this strain were clumpedtogether. The profile width of this strainwas from 1.0 mm to 1.7 mm.The surface of Lb. helveticus T159 had a heterogeneous and rough

(a, b) and Lb. helveticus T159 (c, d) cells.

Fig. 3. FT-IR spectra of freeze-dried bacteria: the solid line is Lb. helveticus T159 andthe broken line is Lb. helveticus T103. The spectra were averaged, normalized, andbaseline corrected in the 3800e800 cm�1 range.

A. Wa�sko et al. / Anaerobe 30 (2014) 161e168 165

character and there was visible variation in the cell length (from1 mm to 4.9 mm).

3.3. Surface analysis by FTeIR

Fig. 3 presents the FTeIR spectra of Lb. helveticus T103 and T159.Among the samples, differences were observed in the fatty acidregions (3000e2800 cm�1). The most significant peak shifts werevisible in the amide region, dominated by amides I and II of variouspeptides (1800e1500 cm�1). Additionally, some differences wereobserved in a mixed region of fatty acid bending vibrations, pro-teins, and phosphate-carrying compounds (1500e1200 cm�1), andin the areas corresponding to the carbohydrate region(1200e950 cm�1) of the spectra.

3.4. Cell surface hydrophobicity

The hydrophobic/hydrophilic and Lewis acidebase properties inthe cell surfaces of Lb. helveticus T103 and T159 were evaluated bythe MATS method. As shown in Table 2, Lb. helveticus T103exhibited low adherence to all the organic solvents used. Incontrast, Lb. helveticus T159 had a strong affinity to chloroform andethyl acetate as well as to hexadecane. After LiCl treatment, celladherence to the solvents decreased significantly. Cell devoid ofsurface proteins did not adhere to hexadecane.

3.5. In vitro adhesion to HT-29 cells

The adhesion potential was determined by the well-establishedplatingmethod. Great variationwas observed in the number of cellsof the tested strains adhering to the HT29 cells (Fig. 4). Highadhesion was observed for Lb. helveticus T159 (the number ofadhering cells 7.9 � 104 CFU ml�1). In turn, Lb. helveticus T103

Table 2Adhesion of untreated and LiCL treated cells of Lb. helveticus T103 and T159 tochloroform, hexadecane, and ethyl acetate.

Bacteria Adhesion (%)

Chloroform Ethyl acetate Hexadecane

Lb. helveticus T103 3.67 ± 0.007 1.91 ± 0.001 1.63 ± 0.005Lb. helveticus T103 LiCl treated 2.84 ± 0.005 1.42 ± 0.006 nda

Lb. helveticus T159 untreated 97.83 ± 0.002 79.15 ± 0.007 38.99 ± 0.008Lb. helveticus T159 LiCl treated 22.39 ± 0.004 9.06 ± 0.001 nda

a nd e not detected.

showed a significantly reduced adhesion capacity (the number ofadhering cells 9.1 � 103 CFU ml�1).

3.6. Auto- and coaggregation assays

The sedimentation rate of Lb. helveticus T103 and T159 cellsuntreated and treated with LiCl was measured over a period of 5 h.The results are shown in Fig. 5a. Under the same growth conditions,Lb. helveticus T159 showed a significantly higher capability of ag-gregation than Lb. helveticus T103. The autoaggregation of bothstrains significantly decreased after LiCl treatment, indicating thatproteins present on the bacterial surface could be associated withthis process. The bacterial mixtures of the tested pathogens andcoaggregating Lb. helveticus strains showed a clear difference in thesedimentation rate (Fig. 5b). A higher capability of coaggregationwith Gramþ and Grame bacteria was exhibited by Lb. helveticusT159. The results indicated a decrease in the auto- and coagger-agtion ability of Lb. helveticus strain T103.

4. Discussion

Bacterial adhesion to solvents was used to investigate the Lewisacidebase properties of the tested Lb. helveticus strains. The resultsindicated a hydrophobic nature of Lb. helveticus T159 and a hy-drophilic character of Lb. helveticus T103. Both strains exhibitedhigher affinity to chloroform (an electron acceptor solvent) than tohexadecane, which indicated that the cells were negativelycharged. The notable difference in the bacterial affinity to hex-adecane and chloroform, i.e. solvents having identical van derWaals forces, showed the importance of the Lewis acid-base in-teractions at the cell surface. These data demonstrated the capacityof the tested Lactobacillus strains to establish some interactionswith the support of other than van der Waals forces [25]. Surfaceconstituents determine the physicochemical properties of the cellwall; particularly, surface proteins confer the hydrophobic prop-erties of the bacterial surface [26]. In our study, the FT-IR analysisrevealed that the tested cells of the Lb. helveticus strains signifi-cantly differed in the region of 1800e1500 cm-1 corresponding tothe amide region of various peptides. Proteomic analysis confirmedthe presence of S-layer proteins and a high level of expression ofmoonlighting proteins only in Lb. helveticus T159. It was expectedthat these proteins had an appreciable effect on cell wall properties.We had previously described that S-layer proteins in Lb. helveticusT159 are highly basic, with an isoelectric point above pH ¼ 9, andthey fully cover the cell wall [16]. The hydrophobic moieties ofsurface proteins are one of the factors that lead to extensiveadhesion and aggregation of bacteria [4]. Hydrophobic interactionsare the strongest long-range non-covalent interactions and areconsidered a determining factor in microbial adhesion to hostepithelial cells. In some lactobacilli, a correlation between theadhesion ability and hydrophobicity has been observed althoughthis property is based on strain-specific attributes [27]. In this work,the hydrophobic strain Lb. helveticus T159 showed a considerablyhigher auto- and coaggregation ability than the hydrophilic Lb.helveticus T103. However, after LiCl treatment, Lb. helveticus T159exhibited a low affinity to the solvents and low aggregation ability.This provided evidence for the role of surface proteins in this strainin the physicochemical properties of bacterial cells. It should benoted that adhesion of lactobacilli to the intestinal surface is aresult of multifactorial specific and non-specific interactions. Theinfluence of other factors such as surface charge and cell surfacesmolecules also contribute to adhesion to host epithelial cells. Thesurface molecules involved in the specific and/or nonspecific bac-teriaehost interactions include S-layer proteins, moonlightingproteins, adhesions, lipoteichoic acid, and exopolysaccharides [26].

Fig. 4. In vitro adhesion of Lactobacillus helveticus T103 and T159 to HT-29 cells. Standard deviations was ±0.1.

A. Wa�sko et al. / Anaerobe 30 (2014) 161e168166

In our study, Lb. helveticus T159 exhibited a significantly highercapability of adhesion to HT29 cells than Lb. helveticus T103 and,undoubtedly, proteinaceous compounds were involved in thisprocess. The label-free analysis revealed the presence of cyto-plasmic housekeeping proteins in the extracellular proteome ofboth strains. Such proteins, defined as moonlighting proteins,display different, seemingly unrelated functions in different celllocations [28]. Most of them lack any extracytoplasmic sortingsequence, and the mechanisms of secretion and cell anchoringremain to be determined [29]. Although the basic functions ofmoonlighting proteins are in essential cellular processes, theirfunctions include also binding to host epithelial cells. Moonlighting

Fig. 5. a. Autoaggregation of Lactobacillus helveticus T103 and T159 untreated (gray bars) anwith Escherichia coli (stripped bars), Salmonella anatum (gray bars), Bacillus subtilis (black bwas in the range 0.01e0.06.

proteins represent an abundant class of bacterial adhesins that arepart of bacterial interactions with the environment [30]. In thiswork among the proteins identified, eight moonlighting proteinswere present in both strains; however, higher expression thereofwas detected in Lb. helveticus T159. We can assume that moon-lighting proteins from three groups were involved in the adhesionprocess in Lb. helveticus T159. These include glycolytic enzymessuch as enolase (ENO), phosphoglycerate kinase, and phospho-glycerate mutase; proteins related to translocation and transcrip-tion such as elongation factor-Ts (EF-Ts), 30S ribosomal protein S1,and oligopeptide ABC transporter substrate binding protein, andstress response and protein folding proteins such as GroEL and

d treated LiCl (white bars). b. Coaggregation of Lactobacillus helveticus T103 and T159ars) and Staphylococcus aureus (white bars). The values are mean ± standard deviation

A. Wa�sko et al. / Anaerobe 30 (2014) 161e168 167

DnaK. To check if moonlighting proteins belong to the IntrinsicallyDisordered Protein (IDP) class, we have predicted IDP from theiramino acid sequences for a number of well-known moonlightingproteins. Our results indicate that most moonlighting proteins donot belong to the IDP class and the disordered amino acid stretcheswere quite short. This observation is similar to results described byHern�andez et al. [20]. Other authors also localized moonlightingproteins on the lactic acid bacterial surface and identified them ascytoplasmic enzymes of the glycolytic pathway or as having othermetabolic functions, or molecular chaperones [30e33]. This articleis the first report on the role of moonlighting proteins in theadhesion and aggregation capability in Lb. helveticus. Moonlightingfunctions described in other bacteria include adhesion to hostepithelia, extracellular matrices (ECMs), and/or secreted mucins aswell as engagement of the host proteolytic plasminogen (Plg) sys-tem and the modulation of host immune responses [30,34]. Glyc-eraldehyde-3-phosphate dehydrogenase (GAPDH) and enolasewere found on the surface of Lb. plantarum, Lb. jensenii, and Lb.crispatus, where they had adhesive functions [13,35]. Other adhe-sive moonlighting proteins detected in lactobacilli include elon-gation factor Tu, heat shock protein GroEL, DnaK, and pyruvatekinase [31,32], which act as adhesion promoting factors for pro-biotics. Phosphoglycerate mutase and kinase were identified asputative plasminogen binding proteins of Bifidobacterium animalissubsp. lactis [36]. Kainulainen et al. [37] observed that releasedmoonlighting proteins were able to reassociate with differentbacterial species, providing a mechanism of bacteriumebacteriuminteractions. In our work, the results of the auto- and coaggregationassay confirm this statement. In Lb. helveticus T159, high expressionof moonlighting proteins was detected and this strain formed ag-gregates with bacteria of the same and other species more easilyand quickly than Lb. helveticus T103. In the light of our study, we canform a hypothesis that S-layer proteins in Lb. helveticus T159 boundmoonlighting proteins and thus increased cell capability of adhe-sion and aggregation. Johnson et al. [38] described in Lb. acidophilusABC sugar transporter and 30S ribosomal protein as S-layer asso-ciated proteins (SLAPs). The authors claimed that SLAPs in Lb. aci-dophilus offered potential in understanding probiotic functionsbetween other adherence factors. In this work, ATP-dependent Clpprotease and dipeptidase were also identified in the Lb. helveticuscell surface proteome. The high level of expression of these proteinsin strain T159 suggested that they might be associated with S-layerproteins. Moreover, interesting observation was the high level ofexpression of technologically important proteins such as L-lactatedehydrogenase, glucose-1-phosphate adenylyltransferase, andglycerol-3-phosphate ABC transporter, which can be also classifiedas SLAPs. Proteinases or proteases and other proteins identified asSLAPs in the cell surface proteome of Lb. helveticus may be impor-tant in the dairy process.

In conclusion, label-free analysis revealed the presence of eightmoonlighting proteins in the cell envelope of Lb. helveticus. Theywere grouped as glycolytic enzymes, proteins related to trans-location and transcription, and molecular chaperones. The pres-ence of S-layer proteins and the high level of expression ofmoonlighting proteins in strain Lb. helveticus T159 gave the cell ahydrophobic character. These surface proteins play a role innonspecific/specific interactions with intestinal epithelium cellsand other bacteria. In Lb. helveticus T159, the S-layer associatedwith moonlighting proteins could act as adherence factors, whichwas evidenced by the high capability of adhesion, auto- andcoaggregation. Additionally, a high level of expression of techno-logically important proteinases and peptidase, which could beassociated with S-layer proteins, was detected in Lb. helveticusT159. Considering our results, we can assume that Lb. helveticusT159 exhibit probiotic potential because moonlighting proteins in

this strain appear to be part of mechanism in bacteriaehostinteractions.

References

[1] Beganovi�c J, Frece J, Kos B, Pvunc AL, Habjani�c K, �Su�skovi�c J. Functionality ofthe S-layer protein from the probiotic strain Lactobacillus helveticus M92.Antonie van Leeuvenhoek 2011;100:43e53.

[2] Jankovic I, Ventura M, Meylan V, Rouvet M, Elli M, Zink R. Contribution ofaggregation-promoting factor to maintenance of cell shape in Lactobacillusgasseri 4B2. J Bacteriol 2003;185:3288e96.

[3] Schachtsiek M, Hammes WP, Hertel C. Characterization of Lactobacillus cor-nyformis DSM 20001T surface protein Cpf mediating coaggregation with andaggregation among pathogens. Appl Environ Microbiol 2004;70:7078e85.

[4] Collado C, Surono I, Meriluoto J, Salminen S. Indigenous dadih lactic acidbacteria: cell surface properties and interactions with pathogens. J Food Sci2007;72:89e93.

[5] Ambalam P, Kondepudi KK, Nilsson I, Wadstr€om T, Ljungh Å. Bile stimulatescell surface hydrophobicity, Congo red binding and biofilm formation ofLactobacillus strains. FEMS Microbiol Lett 2012;333:10e9.

[6] Nikolic M, Jovcic B, Kojic M, Topisirovic L. Surface properties of Lactobacillusand Leuconostoc isolates from homemade cheeses showing autoaggregationability. Eur Food Res Technol 2010;231:925e31.

[7] Ventura M, Callegari ML, Morelli L. S-layer gene as a molecular marker foridentification of Lactobacillus helveticus. FEMS Microbiol Lett 2000;189:275e9.

[8] Scolari G, Vescovo M, Zacconi C, Vescovi F. Extraction and partial character-ization of proteolytic activities from the cell surface of Lactobacillus helveticusZuc2. J Dairy Sci 2006;89(10):3800e9.

[9] Deepika G, Charlampopoulos D. Surface and adhesion properties of lactoba-cilli. Adv Appl Microbiol 2010;70:127e52.

[10] Martínez MK, Acosta MP, Candurra NA, Ruzal SM. S-layer proteins of Lacto-bacillus acidophilus inhibits JUNV infection. Biochem Biophys Res Commun2012;422:590e5.

[11] Kinoshita H, Uchida H, Kawai Y, Kawasaki T, Wakahara N, Matsuo H, et al. Cellsurface Lactobacillus plantarum LA 318 glyceraldehyde-3-phosphate dehy-drogenase (GAPDH) adheres to human colonic mucin. J Appl Microbiol2008;104:1667e74.

[12] Ramiah K, van Reenen CA, Dicks LM. Surface-bound proteins of Lactobacillusplantarum 423 that contribute to adhesion of caco-2 cells and their role incompetitive exclusion and displacement of Clostridium sporogenes andEnterococcus faecalis. Res Microbiol 2008;159:470e5.

[13] Castaldo C, Vastano V, Siciliano RA, Candela M, Vici M, Muscariello L, et al.Surface displaced alfa-enolase of Lactobacillus plantarum is a fibronectinbinding protein. Microb Cell Fact 2009;8:14.

[14] Van Oudenhove L, Devreese B. A review on recent developments in massspectrometry instrumentation and quantitative tools advancing bacterialproteomics. Appl Microbiol Biotechnol 2013;97:4749e62.

[15] Wa�sko A, Szwajgier D, Polak-Berecka M. The role of ferulic acid esterase in thegrowth of Lactobacillus helveticus in the presence of phenolic acids and theirderivatives. Eur Food Res Technol 2014;238:299e306.

[16] Wa�sko A, Polak-Berecka M, Kuzdrali�nski A, Skrzypek T. Variability of S-layerproteins in Lactobacillus helveticus strains. Anaerobe 2014;25:53e60.

[17] Mikula M, Gaj P, Dzwonek K, Rubel T, Karczmarski J, Paziewska A, et al.Comprehensive analysis of the palindromic motif TCTCGCGAGA: a regulatoryelement of the HNRNPK promoter. DNA Res 2010;17:245e60.

[18] Bakun M, Karczmarski J, Poznanski J, Rubel T, Rozga M, Malinowska A, et al.An integrated LC-ESI-MS platform for quantitation of serum peptide ladders.Application for colon carcinoma study. Prot Clin Appl 2009;3:932e46.

[19] Malinowska A, Kistowski M, Bakun M, Rubel T, Tkaczyk M, Mierzejewska J,et al. Diffprot - software for non-parametric statistical analysis of differentialproteomics data. J Proteom 2012;75(13):4062e73.

[20] Hern�andez S, Amela I, Cedano J, Pi~nol J, Perez-Pons JA, Mozo-Villarias A, et al.Do moonlighting proteins belong to the intrinsically disordered protein class?J Proteomics Bioinform 2012;5(11):262e4.

[21] Naumann D, Helm D, Labischinski H. Microbiological characterizations byFTeIR spectroscopy. Nature 1991;351:81e2.

[22] Bellon-Fontaine MN, Rault J, van Oss CJ. Microbial adhesion to solvents: a novelmethod to determine the electron-donor/electron-acceptor or Lewis acid-baseproperties of microbial cells. Coll Surf B: Biointerfaces 1996;7:47e53.

[23] Gopal PK, Prasad J, Smart J, Gill HS. In vitro adherence properties of Lactoba-cillus rhamnosus DR20 and Bifidobacterium lactis DR10 strains and theirantagonistic activity against an enterotoxigenic Escherichia coli. Int J FoodMicrobiol 2001;67:207e16.

[24] Golowczyc MA, Mobili P, Garrote GL, Abraham AG, De Antoni GL. Protectiveaction of Lactobacillus kefir carrying S-layer protein against Salmonella entericaserovar Enteritidis. Int J Food Microbiol 2007;118:264e73.

[25] Pelletier C, Bouley C, Cayuela C, Bouttier S, Bourlioux P, Bellon-Fontaine M.Cell surface characteristics of Lactobacillus casei subsp. casei, Lactobacillusparacasei subsp. paracasei, and Lactobacillus rhamnosus strains. Appl EnvironMicrobiol 1997;63(5):1725e31.

[26] García-Cayuela T, Korany AM, Bustos I, G�omez de Cadi~nanos LP, Requena T,Pel�aez C, et al. Adhesion abilities of dairy Lactobacillus plantarum strainsshowing an aggregation phenotype. Food Res Int 2014;57:44e50.

A. Wa�sko et al. / Anaerobe 30 (2014) 161e168168

[27] EhrmannMA,KurzakP, Bauer J, Vogel RF. Characterizationof lactobacilli towardstheir use as probiotic adjuncts in poultry. J Appl Microbiol 2002;92:966e75.

[28] Jeffery CJ. Moonlighting proteins: old proteins learning newtricks. TrendsGenet 2003;19:415e7.

[29] S�anchez B, Bressollier P, Urdaci MC. Exported proteins in probiotic bacteria:adhesion to intestinal surfaces, host immunomodulation and molecular cross-talking with the host. FEMS Immunol Med Microbiol 2008;54:1e17.

[30] Kainulainen V, Korhonen TK. Dancing to another tunedAdhesive moon-lighting proteins in bacteria. Biology 2014;3:178e204.

[31] Bergonzelli GE, Granato D, Pridmore RD, Marvin-Guy LF, Donnicola D,Corth�esy-Theulaz IE. GroEL of Lactobacillus johnsonii La1 (NCC 533) is cellsurface associated: potential role in interactions with the host and the gastricpathogen Helicobacter pylori. Infect Immun 2006;74:425e34.

[32] Candela M, Centanni M, Fiori J, Biagi E, Turroni S, Orrico C, et al. DnaK from Bifi-dobacterium animalis subsp. lactis is a surface-exposed human plasminogen re-ceptor upregulated in response to bile salts. Microbiologica 2010;156:1609e18.

[33] Granato D, Bergonzelli GE, Pridmore RD, Marvin L, Rouvet M, Corth�esy-Theulaz IE. Cell surface-associated elongation factor Tu mediates the

attachment of Lactobacillus johnsonii NCC533 (La1) to human intestinal cellsand mucins. Infect Immun 2004;72:2160e9.

[34] Henderson B, Martin A. Bacterial virulence in the moonlight: multitaskingbacterial moonlighting proteins are virulence determinants in infectiousdiseases. Infect Immun 2011;79:3476e91.

[35] Spurbeck RR, Arvidson CG. Lactobacillus jensenii surface associated proteinsinhibit Neisseria gonorrhoeae adherence to epithelial cells. Infect Immun2010;78:3103e11.

[36] Candela M, Bergmann S, Vici M, Vitali B, Turroni S, Eikmanns BJ, et al. Bindingof human plasminogen to Bifidobacterium. J Bacteriol 2007;189:5929e36.

[37] Kainulainen V, Loimaranta V, Pekkala A, Edelman S, Antikainen J, Kylv€aj€a R,et al. Glutamine synthetase and glucose-6-phosphate isomerase are adhesivemoonlighting proteins of Lactobacillus crispatus released by epithelial cath-elicidin LL-37. J Bacteriol 2012;194(10):2509e19.

[38] Johnson B, Selle K, O'Flaherty S, Goh YJ, Klaenhammer T. Identification ofextracellular surface-layer associated proteins in Lactobacillus acidophilusNCFM. Microbiologica 2013;159:2269e82.