Introduction

Lactic acid bacteria (LAB) are widely used in food biotechnology and efficient control of these microbiological processes requires an increase in our knowledge about their ecology and biochemical activities. Certain species of LAB are associated with the manufacture of fermented beverages like wines (grapes) and ciders (apples). They perform malolactic fermentation (MLF), which is a suitable process for correct acidity. An adequate MLF process is fundamental to guarantee a high standard quality cider. In cider, as in wine, Lactobacillus and Oenococcus were described as the predominant genera during MLF, finding Pediococcus in low proportion (Sánchez et al. 2010). Although LAB are beneficial for fermented beverages quality, certain species can deteriorate the products (del Campo et al. 2008). Thus, is important to be able to distinguish the interest species from those that are prejudicial. On the other hand, it has been reported that LAB can multiply in fresh apple juice causing formation of slime, gas, off flavor, turbidity and changes in acidity (Siegmund and Pollinger-Zierler 2006).

It is known that the rate of growth of LAB depends on their ability to utilize the substrates available in the medium. LAB have numerous nutritional requirements for growth especially with regard to nitrogen sources (Amoroso et al. 1993; Elli et al. 1999). Saguir and Manca de Nadra (2007) reported that two strains of Lactobacillus plantarum from oranges had an absolute requirement for a minimum of six common amino acids and that they mainly consumed the essential amino acids and the stimulatory amino acid arginine. l-Arginine, is usually considered as major amino acid of fruits and juices. Two routes for its bacterial degradation have been described: the arginine-urease pathway, which involves the enzyme arginase with the formation of ornithine and urea, and, more commonly, the arginine deiminase (ADI). The ADI pathway involves three enzymes: arginine deiminase, catabolic ornithine transcarbamoylase and carbamate kinase. This pathway leads to the production of ammonia, ornithine, ATP, carbamyl phosphate (carbamyl-P) and some citrulline is also excreted fulfilling various roles: to provide energy by substrate phosphorylation, to supply carbamyl phosphate for citrulline or pyrimidines biosynthesis; and to protect bacteria against acid damage (Tonon and Lonvaud-Funel 2002). Therefore, the expression of the ADI pathway in industrial microorganisms, such as LAB can be of great significance since it can be considered as a mechanism of energy production and pH regulation. However, if arginine degradation is beneficial for the bacteria, improving its growing ability and its adaptability, this increases the risk of degradation of the organoleptic and hygienic properties of fermented beverages such as cider. For example arginine may also be responsible for the formation of citrulline, a direct precursor of carcinogenic ethyl carbamate (EC) (Tonon and Lonvaud-Funel 2002).

The ADI pathway has been mainly studied in LAB isolated from dairy sources or wines (Manca de Nadra et al. 1988; Arena et al. 2008). However, to date little information is known on potential arginine utilization by LAB that constitutes the microflora from apples, in spite of its significance in cider-making.

As a consequence, the aim of this study was to identify the LAB isolated from two apples variety with potentialities to utilize arginine through ADI pathway, by investigating the citrulline and ammonia production in a basal medium added with arginine at different initial pH values. The use of a basal medium poor in sources of amino acids and adjusted at low pH was necessary to clarify the effect of arginine on the growth of LAB in conditions of nutritional and acidic stress, such as apple juice.

Materials and methods

Samples

Red Delicious and Golden Delicious apples were obtained from farms of Tucumán and Jujuy states, Argentina, respectively. All tree-harvested apples manually were placed in sterile bags and immediately transported to the laboratory under refrigeration, being analyzed on arrival. The products were all of agreeable sensory quality.

Processing, enumeration and isolation of LAB from apples

For enumeration and isolation of LAB, skins of each apple variety were washed three times with sterile distilled water. Each water washing was collected under sterile conditions. From water washing samples, 0.1 ml-aliquots (diluted if necessary) were plated on MRS agar (Oxoid Ltd., London, England) supplemented with 15% (v/v) of tomato juice, pH 6.0 containing 1.7 μg/ml of pimaricin (MRS-P) (Sigma Chemical Co., St. Louis, MO) to suppress yeasts growth. MRS-P plates were incubated anaerobically for 72 h at 30°C before enumeration.

A total of 68 colonies (35 from Red Delicious apples and 33 from Golden Delicious apples) that exhibited distinct morphological differences in color, shape and size and presumed to be LAB were randomly picked up from MRS-P agar plates and purified by making streak plates on MRS (Oxoid Ltd., England) and incubated at 30°C for 48 h. Pure cultures that exhibited gram positive and catalase negative reactions were considered as LAB and selected for further investigations. They were also characterized for cell morphology and motility using a phase contrast microscope (Olympus CX41, Japan). Bacterial cultures were maintained at −20°C in MRS medium with tomato juice (15%, v/v) and glycerol (20%, v/v). Working cultures were prepared from the frozen cultures by two consecutive transfers in MRS broth with tomato juice (15%, v/v) at 30°C, pH 6.0.

Phenotypic characterization of LAB isolates at species level

Selected isolates considered as LAB on the basis of their results for Gram staining, cell morphology, motility, spore formation and catalase reaction were tested for fermentative catabolism of glucose, gas and D- or L-lactic acid isomers production from glucose in Gibson medium (Gibson and Abdel-Malek 1945) and by using an enzymatic method, Boehringer Kit (Mannheim, Germany), respectively and nitrate reduction. Ability of growth, under microaerophilic conditions in BBL GasPak jars (Becton–Dickinson, Argentina) in which the content of oxygen was reduced by use of a lighted candle, was determined on MRS agar plates incubated at 15, 30, 37 and 45°C. Ability of growth at different NaCl concentrations (2, 4, 6 and 8%, w/v) and to ferment carbohydrates in MRS broth without glucose containing bromocresol purple (0.04 g/l), inverted Durham tube and supplemented with 1% (v/v) of carbohydrates and related compounds were carried out. The API 50 CHL fermentation test was performed in some isolates (BioMérieux, Marcy-l’Etoile, France).

Genotypic identification

Lactobacillus sp. identified during this study were further characterized by sequence analysis of segments of the 16S ribosomal DNA. Essentially, pure cultures were grown and the DNA was extracted as described by Reguant and Bordons (2003). Fragments of the 16S rRNA gene sequence were amplified by PCR using conserved primers close to the 3′and 5′ends of the gene or a set of universal primers (Endo et al. 2010). The amplicons were confirmed by agarose gel electrophoresis and then purified using a Prep-A-Gene kit (Biorad, USA) according to the manufacturer’s instructions. They were sequenced using an Applied Biosystems 3730 XL sequencer (Applied Biosystems, USA) by Macrogen Inc. (908 World Meridian Venture Center, #60-24, Gasan-dong, Geumchun-gu, Seoul 153-781, Korea). The sequences determined in this study were compared with those available in the GenBank database and similitude percentages were calculated after alignment. Lactobacillus plantarum, Lactobacillus pentosus and Lactobacillus paraplantarum strains were further distinguished by means of partial amplification product comparison of the recA gene, according to the method of Torriani et al. (2001). Preparation of bacteria for this PCR reaction involved growing of a bacterial colony on MRS agar plates at 30°C until colonies could be clearly distinguished. The 20 μl PCR reaction mix consisted of 0.25 μM of each paraF, pentF, and pREV primers and 0.12 μM of plan F primer (Torriani et al. 2001), 0.2 mM of each dNTP (Amersham Pharmacia Biotech, Inc., Piscataway, NJ, USA), 1.5 mM of MgCl2, 5 μl of 10× reaction buffer, 0.5 U of Taq DNA polymerase (Amersham Pharmacia Biotech, Inc.) and a single colony. The profile of amplification was: one cycle of 94°C for 3 min, 30 cycles of 94°C for 30 s, 56°C for 10 s and 72°C for 30 s. A final extension of 72°C for 10 min was also included. The amplification products were visualized by ethidium bromide (5 μg/ml) staining after gel electrophoresis. A Biometra TRIO-Thermoblock machine was used for all PCR reactions.

The Fluorescence in situ hybridization (FISH) method was used for Leuconostoc sp. identification. The 16S rRNA sequences used in this study were obtained from EMBL and GenBank databases by Blasco et al. (2003). The eubacterial Eub 338 probe 5′ end labeled with fluorescein by MWG biotech was used as positive control and Non 338 5′-fluorescein labeled, complementary to 338, as the negative control for non-specific binding (Amann et al. 1990). LU2 oligonucleotide probe 5′end-labelled with fluorescein was specific for Leuconostoc sp. FISH experiments were performed according to the method described by Blasco et al. (2003). Fluorescence was detected with a Leica DMRB microscope fitted for epifluorescence microscopy with a 100-W mercury lamp high-pressure bulb and Leitz Wlter blocks A (UV light exciter BP 340–380 nm, beamsplitter RKP 400 nm, emitter LP 430 nm), I3 (blue light exciter BP 450–490 nm, beamsplitter RKP 510 nm, emitter LP 520 nm), and N2.1 (light exciter BP 515–560 nm, beamsplitter RKP 580 nm, emitter LP 580 nm).

Media culture and growth conditions

The LAB isolates were routinely propaged in MRS broth with tomato juice (15%, v/v). The strains were screened for its ability of growing in a basal medium (BM) containing in g/l: peptone, 5; yeast extract, 3; glucose, 0.5; tomato juice, 2% (v/v) and supplemented with L-arginine (SIGMA, Louis, MO, USA) at a concentration of 1 g/l (BMA) adjusted at pH 4.5. Microorganisms that gave positive results were selected for further studies. BM and BMA were adjusted to pH 4.5 or 5.2 before sterilization at 121°C for 20 min.

Cells grown in MRS broth with tomato juice (15%, v/v), pH 6.0, incubated without agitation at 30°C were precultured in BM under the same conditions, harvested by centrifugation, washed twice with sterile distilled water and resuspended in sterile distilled water to an optical density (OD) at 560 nm of ca. 1.0. The bacterial suspensions were used to inoculate the experimental media at a rate of 2% (v/v). All cultures were incubated statically at 30°C for 10 days.

Growth and pH changes

Bacterial growth was monitored by periodic spectrophotometric measurements (optical density, OD560nm) using a Bausch and Lomb Spectronic-20 spectrophotometer. Samples were taken every 4 h for pH measurements and analytical determinations. The pH value of each sample was measured with a microprocessor pHmeter (HANNA Instruments, Milan, Italy).

Analytical method

Ammonium was detected qualitatively with Nessler reagent, positive reaction appeared as orange color. Ammonia and urea were quantified by enzymatic method (Kit from Wiener Lab., Rosario, Argentina). Citrulline concentration was analyzed according to the procedure of Spector and Jones (1963).

Statistical analysis

All experiments were performed in triplicate. The means of the data and standard deviations are presented.

Results

Microbial counts and LAB isolation from apples

Apples surface contained total average levels of bacteria ranging from log 2.49 ± 0.53 to log 3.73 ± 0.48 cfu/ml for Red Delicious and Golden Delicious varieties, respectively.

From 68 colonies randomly picked up from MRS-P agar plates, a total of 48 strains (22 from Red Delicious and 26 from Golden Delicious) were identified as belonging to LAB group because they were Gram positive, catalase negative, non-spore forming and non-motile bacteria. Morphologically the most of cells from Red Delicious apples were spherical and occurred in pairs or tetrads accounting for 86.4% of the total of strains whilst ones from Golden Delicious variety were rods in pairs and short chains accounting for 73% of the total of isolates.

A total of 31 strains of LAB (15 from Red Delicious and 16 from Golden Delicious apples) were able to grow in BM supplemented with 1 g/l arginine, pH 4.5 in which the concentration of tomato juice was increased to 15% (v/v). They reached maximum population levels of OD560nm comprised among 0.35 and 0.50 and were selected for phenotypical and genotypical characterization at species level.

Phenotypic identification of LAB at species level

The morphological and biochemical characteristics of selected strains were determined as shown in Table 1. All LAB isolates exhibited a fermentative metabolism from carbohydrates. The majority of isolates were homofermentative while only a few isolates were obligatory heterofermentaive. Isolates were divided in 6 groups according to sample origin and their morphological and physiological features:

Table 1 Grouping physiological and biochemical characteristics of lactic acid bacteria isolates from apples
Full size table

Group I

Consisted of 10 homofermentative isolates from Red Delicious apples identified as Pediococcus and clearly differentiated from other LAB by the typical cell morphology of tetrad-forming cocci and formation of DL lactate isomers from glucose. Of these bacteria, three strains fitted the description of Pediococcus pentosaceus and the remaining as Pediococcus acidilactici according to their carbohydrate fermentation patterns. Pediococcus acidilactici produced acid from arabinose, mannose and D-xylose and grew at 45°C, while members of Pediococcus pentosaceus fermented arabinose, mannose, maltose, raffinose and sucrose.

Group II and IV

Consisted of one and three heterofermentative isolates from Red Delicious and Golden Delicious varieties, respectively identified as Leuconostoc on the basis that they exhibited production of D-lactic acid isomer from glucose, growth at 10°C but no growth at 45°C and showed the typical Leuconostoc-like ovoid cell shape. Of these bacteria, one strain shared features with Leuconostoc dextranicum since it exhibited slime production, and did not form acid from arabinose, galactose and xylose while the remaining strains with Leuconostoc mesenteroides (Group IV).

Group III and V

Six heterofermentative isolates that produced D (-) lactate isomer from glucose, were able to grow at 20, 37°C but not at 45°C and formed acid from glucose, arabinose, xylose, maltose, raffinose, sucrose and esculine were classified as Lactobacillus brevis.

Group VI

Consisted of eleven strains isolated from Golden Delicious apples that produced DL-lactate isomers from glucose by homolactic fermentation, grew in presence of 2 or 4% NaCl, at 20 or 37°C and they fermented glucose, sorbose, arabinose, mannose, rafinose, xylose, esculin and sucrose. They were assigned to Lactobacillus plantarum specie.

Genotypic identification of LAB isolates

Lactobacillus sp. strains when identified by sequencing of partial 16S rDNA supported the division between the two groups. Thus, all isolates of group VI showed high percentages of identity (100%) to kwon sequences of Lactobacillus plantarum in Gen Bank. Similar result was observed for those phenotyphically identified as Lactobacillus brevis with percentages of similarity of about 98%. The GenBank/EMBL/DDBJ accession numbers for the 16S rRNA gene sequences of Lactobacillus brevis M15, Lactobacillus plantarum M41, Lactobacillus plantarum M42, Lactobacillus plantarum M60, Lactobacillus plantarum M66, Lactobacillus plantarum M67, Lactobacillus plantarum M68, Lactobacillus plantarum M74 are JN247626, JN247627, JN247628, JN247629, JN247630, JN247631, JN247632 and JN247633, respectively. On the other hand the multiplex PCR assay using species specific primers for Lactobacillus plantarum, Lactobacillus paraplantarum and Lactobacillus pentosus gave a single product of 318 bp, the expected size of the PCR fragment for Lactobacillus plantarum, only when one colony from Lactobacillus plantarum isolates and the reference strain were used as target for specific PCR reaction. This fact confirmed the morphological, biochemical and physiological characterization of Lactobacillus plantarum isolates.

The identification of strains classified as Leuconostoc sp. was confirmed by FISH technique using the LU2 specific probe (Blasco et al. 2003). Whole-cell hybridization results of Leuconostoc sp., Lactobacillus brevis and Lactobacillus plantarum strains isolated from apples during this study showed no autofluorescence and yielded strong hybridization signals with the Eub338 oligonucleotide probe. The LU2 probe hybridized exclusively with the 16S rRNA of three strains of Leuconostoc sp. giving negative reaction with the other bacterium analyzed. Thus, the strain identified phenotypically as Leuconostoc mesenteroides ssp. that did not show this positive result was excluded from this study.

Growth and pH variation in BM and BMA

Thirty strains of LAB identified during this study were transferred to BM and BMA and the average maximum populations and final pH values were determined at pH 4.5 or 5.2 (Table 2). In BM, LAB grew at OD560nm ranging from 0.36 ± 0.01 to 0.48 ± 0.01. Pediococcus acidilactici and Leuconostoc mesenteroides strains grew at the lowest OD560nm as compared with the other LAB isolates at both pH values. By contrast Lactobacillus plantarum and Lactobacillus brevis strains showed the highest average final populations of OD560nm at pH 5.2. Arginine added to BM did not produce significant differences (P < 0.05) in the average final populations expressed as OD560nm of LAB strains regards with those obtained in BM without it. However, when analyzed individually each strain, arginine had a stimulatory effect on growth of Lactobacillus brevis M15 of about 13% at the two tested pH values (4.5 or 5.2). Moreover, in both conditions the final pH of the media with added arginine appeared to be significantly higher than of the control media. Figure 1 shows the growth of the M15 strain and pH variations in BM and BMA at pH 4.5. The increase of final pH in BMA was approximately 1.0 unit higher than before incubation. Similar results were observed for Leuconostoc dextranicum M13. In other LAB no significant differences in pH became evident between BM and BMA at the end of growth under any condition (Table 2). Thus, these results suggested that of the total of analyzed isolates Lactobacillus brevis M15 and Leuconostoc dextranicum M13 could be able to degrade arginine and as a consequence ammonia could be produced, which may contribute to the increase in pH.

Table 2 Growth of lactic acid bacteria isolates in BM and BMA media and pH variation
Full size table
Fig. 1
figure 1

Growth (solid line) and variation of pH (dashed line) in basal medium (BM) (filled square) and in the BM added with arginine (open square) by Lactobacillus brevis M15 at pH 4.5. Results are mean values of triplicate determinations

Full size image

Ammonium production by LAB isolates measured at the end of growth

Qualitative detection of ammonia only gave positive reaction in supernatants of cultures from Lactobacillus brevis M15 and Leuonostoc dextranicum M13 at the end of exponential phase under experimental conditions.

When this metabolite was quantified, ammonia production varied from 2.70 to 7.40 mmol/l in BM adjusted at pH 4.5 or 5.2. The addition of arginine did not modify the ammonia production in the majority of tested strains at both pH values (Data not shown). Only in Leuconostoc dextranicum M13, Lactobacillus brevis M15 and five strains of Lactobacillus plantarum the total amount of ammonia formed significantly increased in BMA as compared to BM, which was correlated with arginine utilization. This effect was more pronounced at pH 4.5 than 5.2. Since ammonia was also detected in the BM without added arginine, the net amount of NH3 synthesized in presence of arginine in order of decreasing concentration after 24 h of incubation at pH 4.5 was 13.6, 10.7, 5.2, 4.1, 2.3, 1.7 and 1.3 mmol/l by Lactobacillus brevis M15, Leuconostoc dextranicum M13, Lactobacillus plantarum M67, Lactobacillus plantarum M74, Lactobacillus plantarum M68, Lactobacillus plantarum M42 and Lactobacillus plantarum M66 strains, respectively (Fig. 2a). When pH was adjusted at 5.2 ammonia production from arginine by Lactobacillus brevis M15 and Lactobacillus plantarum M74 strains decreased accounting for 21 and 46%, while no significant differences (P < 0.05) were found in the ammonia concentrations synthesized between BM and BMA by Lactobacillus plantarum M42, Lactobacillus plantarum M66, Lactobacillus plantarum M67 and Lactobacillus plantarum M68 strains (Fig. 2b).

Fig. 2
figure 2

Ammonia produced by lactic acid bacteria strains in basal medium (BM) (filled square) and in the BM added with arginine (open square) at pH 4.5 (a) and 5.2 (b) at the end of exponential growth. Results are mean values of triplicate determinations

Full size image

Citrulline and urea production by LAB isolates measured at the end of growth

At pH 4.5 Leuconostoc dextranicum M13 and Lactobacillus brevis M15 excreted 0.071 and 0.069 mmol/l of citrulline after 24 h incubation in BMA, respectively. At pH 5.2 the M13 strain of Leuconostoc dextranicum produced concentrations significantly lower (0.031 mmol/l) while Lactobacillus brevis M15 released similar amounts as compared with that obtained at pH 4.5. Although at both pH conditions citrulline was produced by these strains in concentrations remarkably higher than those found in BM, lower than 1 mmol/l of citrulline could be formed from arginine. Lactobacillus plantarum M66 and Lactobacillus plantarum M68 when cultured in BMA at pH 4.5 also excreted higher citrulline concentrations than in BM, in general the values found were lower than those obtained for M15 and M13 strains (Fig. 3). No citrulline production was found in supernatants of culture of Lactobacillus plantarum M42 and Lactobacillus plantarum M67 strains under experimental conditions despite that ammonia was formed in a greater extent in BMA than BM.

Fig. 3
figure 3

Citrulline produced by lactic acid bacteria strains in basal medium (BM) at pH 4.5 (black filled square) and 5.2 (dark gray square) and in BM added with arginine at pH 4.5 (open square) and 5.2 (gray square) at the end of exponential growth. Results are mean values of triplicate determinations

Full size image

Urea was not detected in LAB isolates under any condition, thus excluding the presence of arginine–urease pathway.

Time course of ammonium and citrulline production

Citrulline and ammonia production was measured during growth of Leuconostoc dextranicum M13 and Lactobacillus brevis M15 in BM and BMA at both pH values as shown in Fig. 4. Both strains reached the exponential growth phase at 24 h incubation independently of initial pH. Ammonia liberation began immediately growth began and reached maximum values for 12 h incubation in BMA, pH 4.5 for both strains. Then, it remained almost constant. Similar behavior was observed at pH 5.2, although in the most acidic condition, highest ammonium production by Lactobacillus brevis M15 was detected. Considering this strain, since ca. 10 mM NH3 were produced in the BM without added arginine, the net amount of NH3 synthesized in the presence of 5.74 mM arginine was ca. 12 mM, which as expected from the stoichiometries, would correspond to a ratio of about 1 mM arginine consumed to 2 mM NH3 produced. Similar results were observed for Leuconostoc dextranicum M13. In BM, insignificant citrulline amounts were produced during growth of Leuconostoc dextranicum M13 and Lactobacillus brevis M15. In presence of arginine citrulline liberated to extracellular environment by Lactobacillus brevis M15 increased constantly during incubation up to 0.07 mmol/l at both pH values. However, under any condition concentrations higher than 0.1 mM were obtained. Citrulline production from arginine by Leuconostoc dextranicum M13 was also detected but in a smaller extent than by Lactobacillus brevis M15, especially when cultivated at pH 5.2 (Fig. 4a).

Fig. 4
figure 4

Ammonia (open square) and citrulline (open triangle) formation during incubation of Leuconostoc mesenteroides M13 (a) and Lactobacillus brevis M15 (b) in basal medium (BM) (filled square, filled triangle) and in the BM added with arginine (open square, open triangle) at pH 4.5 (solid line) and 5.2 (dashed line)

Full size image

Discussion

It is well known the beneficial role of LAB in the manufacture of apple cider since they are responsible to carry out MLF (Siegmund and Pollinger-Zierler 2006). However, few data on apple lactic acid microflora and their potentialities to use arginine are found, yet. In our study, significant differences in the bacterial counts on MRS-P agar plates among the varieties of apples studied were found. Keller et al. (2004) reported that lower bacterial counts were associated with apple varieties exhibiting higher Brix values and higher titratable acidity. On the other hand, the bacterial counts on MRS-P agar plates were lower to ones found in other fruits (Sajur et al. 2007). Thirty-one strains of LAB were identified as belonging to Lactobacillus, Pediococcus and Leuconostoc genera accounting for 50, 31 and 19% of the total tested strains, respectively. Martínez-Viedma et al. (2008) reported that LAB species, mainly Lactobacillus collinoides, Lactobacillus dioliovorans and Pediococcus parvulus have been implicated in spoilage of Basque apple cider. In our study Leuconostoc sp. were dominant in Red Delicious apples whilst Lactobacillus plantarum in the Golden Delicious one. Thus, the apples composition, especially Golden Delicious variety could be an important factor in enabling the growth of Lactobacillus sp. requiring growth factors. On the other hand, Lactobacillus brevis was the only specie with potentially to use arginine isolated from both apples variety.

Only two strains belonging to Leuconostoc dextranicum and Lactobacillus brevis species isolated from Red Delicious apples presented the ability to increase the final pH of media supplemented with arginine as compared with control media at pH 4.5 and 5.2, indicating that they could utilize arginine. These differences at level of species and strains could reflect an adaptative response to the natural habitat. Arginine degrading strains from wines included all heterofermentative lactobacilli, Oenococcus oeni, Pediococcus pentosaceus, and some strains of Leuconostoc mesenteroides and Lactobacillus plantarum (Araque et al. 2009).

With respect to ammonia production, Nessler reagent was useful only for detecting strongly ammonia producing strains such as Leuconostoc dextranicum M13 and Lactobacillus brevis M15 in concordance with results previously reported (Pilone et al. 1991). Thus, Lactobacillus plantarum strains that produced lower than 7 mM of ammonia in BMA were not detected. This fact also could explain why the final pH in BMA did not markedly increased as compared with the final pH of BM without arginine at the end of growth of these bacteria.

Strains of Pediococcus sp. were not able to hydrolyze arginine indicating that this characteristic could be used as biochemical tool for their biochemical classification. It have been reported that homofermentative pedicoocci and lactobacilli isolated from the wine environment did not seem to degrade arginine (Spano et al. 2004). On the other hand, ammonia production found in BM without added arginine suggested that this metabolite was also produced from deamination of peptides and/or amino acids from sources such as yeast extract other than arginine. It is interesting to note that among strains that produced ammonia from arginine, only arginine stimulated growth of Lactobacillus brevis M15 and Leuconostoc dextranicum M13 under conditions studied, suggesting effective energy coupling from arginine degradation to growth. Terrade and Mira de Orduna (2008) reported that neither arginine nor citrulline increased growth of two Oenococcus oeni strains in comparison with one Lactobacillus buchneri strain. However, arginine and citrulline were partially degraded in all incubations for both Oenococcus oeni. On the other hand the ammonia production pattern by potentially arginine-positive strains of Lactobacillus plantarum suggested that arginine degradation could mainly contribute to their acid tolerance at the low pH values present in the Golden Delicious apples.

As expected among strains that did not produce ammonia from arginine, no positive results for citrulline were found. However, the positive results obtained for Leuconostoc dextranicum M13, Lactobacillus brevis M15, Lactobacillus plantarum M66 and Lactobacillus plantarum M68 confirmed that these bacteria utilized arginine via ADI pathway and not via arginase-urease. In addition urea was not detected under any condition. An interesting finding was that initial pH affected significantly citrulline production by Leuconostoc dextranicum and Lactobacillus plantarum strains. Vrancken et al. (2009) demonstrated in a strain of Lactobacillus fermentum that the ratio between citrulline and ornithine and the pattern of their formation from arginine degradation showed variation as function of environmental pH. The not detection of citrulline in presence of arginine at low pH by ammonia-positive strains is an important fact considering that citrulline is a direct precursor of carcinogenic EC (Terrade and Mira de Orduna 2008). Citrulline is apparently the main EC precursor produced by Lactobacillus hilgardii strains in spoiled, fortified wine (Azevedo et al. 2002). Thus, this property should be considered as a quality criterion for selecting starter cultures to carry out MLF in the manufacture of fermented beverages.

Analysis of ammonia and citrulline production by Leuconostoc dextranicum M13 and Lactobacillus brevis M15 during incubation revealed at the two assayed pH conditions that ammonia was formed at higher rate than citrulline being the later detected in low concentrations. This fact could be related with a lower rate of citrulline formation rather than removal of citrulline to form ornithine and carbamyl-P, a thermodynamically unfavorable reaction. Similar result was reported by Arena et al. (1999). Thus, the abilities of these strains to derive energy and ammonia from arginine via ADI pathway, could encourage optimal growth responses under non-optimal conditions, such as low pH of fruit juice. So they were selected for further studies in order to analyze its potentialities to be used in food fermentation process such as cider. However, they also could contribute to spoilage development in fresh juice.

In conclusion, the main genera and species of LAB found on surfaces of Golden Delicious and Red Delicious apples with abilities to degrade arginine through the ADI pathway under different pH conditions were reported here at the first time. Majority of LAB strains did not show this property being the arginine positive strains mainly isolated from Golden Delicious apples. Although, the findings suggested that this pathway in Lactobacillus brevis M15 and Leuconostoc dextranicum M13 from Red Delicious apples could be most active in adapting to non-optimal growth conditions. Finally, the ADI pathway in apples LAB might not be mainly relevant for their growth in the acid natural environmental, despite encouraging an extra energy supply for the growth and protection against acid.