The Influence of pH, NaCl, and the Deacidifying Yeasts Debaryomyces hansenii and Kluyveromyces marxianus on the Production of Pigments by the Cheese-Ripening Bacteria Arthrobacter arilaitensis

Arthrobacter arilaitensis is a food-related bacterial species under investigation for its involvement in the coloration of surface-ripened cheeses. Presently, information about this species in association with the development of appropriate cheese coloration is still lacking. This study was performed in order to investigate—with the use of spectrocolorimetry—the influence of pH, NaCl, and deacidifying yeasts on the pigmentation of Arthrobacter arilaitensis biofilms. Three types of cheese-based (curd) solid media were prepared by using different deacidification methods: (i) chemical deacidification by NaOH (CMNaOH); (ii) biological deacidification by the yeast strain Debaryomyces hansenii 304 (CMDh304); and (iii) biological deacidification by the yeast strain Kluyveromyces marxianus 44 (CMKm44). Each medium was prepared with initial pH values of 5.8, 7.0, and 7.5. After pasteurization, agar was incorporated and NaCl was added in varying concentrations (0%, 2%, 4%, and 8% (w/v)). A. arilaitensis Po102 was then inoculated on the so prepared “solid-curd” media, and incubated at 12 °C under light conditions for 28 days. According to the data obtained by spectrocolorimetry in the Compagnie Internationale de l’Eclairage (CIE) L*a*b* color system, all controlled factors appeared to affect the pigments produced by the A. arilaitensis strain. NaCl content in the media showed distinct inhibitory effects on the development of color by this strain when the initial pH was at 5.8. By contrast, when the initial pH of the media was higher (7.0, 7.5), only the highest concentration of NaCl (8%) had this effect, while the coloring capacity of this bacterial species was always higher when D. hansenii 304 was used for deacidification compared to K. marxianus 44.


Introduction
The color of smear-ripened cheese rind is one of the major attributes determining a consumer's purchasing decision, and it is linked with several cheese qualities, e.g., maturity, flavor, and cleanliness [1]. A large number of smear-ripened cheeses are recognized by consumers through their rind's characteristics, especially the presence of color and flavor, which depend on the relevant properties of each cheese. The surface of smear-ripened cheeses is the result of pigment synthesis by a complex surface microflora, consisting of various species, both prokaryotes and eukaryotes [2]. Due to its presence at different stages of the ripening process, the genus Arthrobacter has long been identified

Yeast and Bacterial Strains
The bacterial and yeast strains used in this study, namely, Arthrobacter arilaitensis Po102, Debaryomyces hansenii 304, and Kluyveromyces marxianus 44, were kindly provided by the Institut National de la Recherche Agronomique (INRA), France. They were maintained during this study on medium depending on their nutrient requirements (i.e., milk ingredient-based medium for bacteria or potato dextrose agar for yeasts, as described below), stored at 4 • C, and subcultured monthly.

Cultures
Potato dextrose broth (BD Biosciences, Le Pont de Claix, France) was used as a subculture medium for yeasts while the milk ingredient-based medium was prepared and used for bacteria. A milk ingredient-based liquid medium containing 5 g casamino acids (BD Biosciences, France), 1 g yeast extract (BD Biosciences, France), 5 g NaCl (Fisher Scientific, Illkirch-Graffenstaden, France), 20 g D-glucose (Fisher Scientific, France), and 1 g KH 2 PO 4 (Fisher Scientific, France), per liter of deionized water, was prepared. Before sterilizing at 121 • C for 15 min, the pH of the medium was adjusted to 7.0 ± 0.2.
To prepare a microbial suspension, each of D. hansenii 304 and K. marxianus 44 was inoculated into 250 mL Erlenmeyer flask containing 50 mL of PDB, whilst the milk ingredient-based liquid medium was used instead of PDB for A. arilaitensis Po102. These flasks were incubated at 25 • C with agitation at 150 rpm for 2 days. After the cultivation period, the cells were harvested by centrifugation at 6000× g for 10 min, and resuspended in peptone saline diluent (1 g casein peptone (Sigma-Aldrich, Darmstadt, Germany) and 8.5 g NaCl, pH adjusted to 7.0 ± 0.2 at 25 • C)). A suspension containing 10 7 cells mL −1 was prepared to be used in the next step.

Deacidified Cheese-Based Media and Factors Controlled
The procedure for preparing the experimental media employed in this study was adapted according to Leclercq-Perlat et al. [17], Rea et al. [23], and Masoud and Jakobsen [21]. The cheese-based (curd) media were prepared in two steps: (i) the pH adjustment by chemical or biological deacidification, depending on the observation of coloration influenced by yeasts; and (ii) the inoculation of bacterial strains on the media surface.
The cheese curd used throughout the experiment was kindly supplied by Fromagerie de Bourbon, Reunion Island, France.
After receiving it, the cheese curd was frozen at −80 • C to prevent enzymatic reaction [24,25].
Before preparing the cheese-based media, the frozen cheese curd was thawed at 4 • C for roughly 24 h. The defrosted cheese curd was mixed with distilled water (60% (w/v)); then the mixture was homogenized using a blender and placed in a water bath (100 • C) for 30 min. After cooling down to 60 • C, the mixture was blended, again, until the texture was homogeneous. This cheese slurry was divided depending on the volume required for each treatment, and then poured into sterilized flasks for further steps.

pH Adjustment
Chemical Deacidification A flask containing cheese slurry was pH adjusted to 5.8, 7.0, or 7.5, with 1 M NaOH (Fisher Scientific), followed by the addition of 2 g granulated agar (BD Biosciences, France) and NaCl (0%, 2%, 4%, or 8% (w/v)). The final mixture was pasteurized by heating at 80 • C for 10 min. After cooling down to 45-50 • C, the pour plate technique was applied to prepare the cheese-based (curd) solid medium contained in Petri dishes, which were left to solidify for 2 h before inoculation with A. arilaitensis Po102.

Biological Deacidification
When the temperature of the cheese slurry decreased to approximately 30 • C, it was inoculated with a suspension containing 10 7 cells mL −1 of either D. hansenii 304 or K. Marxianus 44, with a suspension-to-slurry ratio of 1:100 (v/v). After the inoculation, the slurry was homogenized for a while, and then the flask was closed with a cork cap, wrapped in an aluminum foil, and incubated in the darkness at 25 • C while agitating at 150 rpm until the pH values increased to 5.8, 7.0, or 7.5. In order to stop the yeast from affecting the pH value, the cheese slurry was heated at 80 • C for 10 min. Agar with varying NaCl contents, as described before, was added to the mixture. Afterwards, the same procedures as mentioned above in "Chemical deacidification" (after adjusting the pH by a chemical reagent) were applied.

Inoculation of Arthrobacter arilaitensis and Cultivation Conditions
After solidifying, one mL of a 10 7 cell suspension of A. arilaitensis Po102 was spread over the surface of the deacidified media. All Petri dishes were then incubated at 12 • C under daylight fluorescent illumination. Three independent replicates for each treatment were collected at 7, 14, 21, and 28 days, and measured using spectrocolorimetry for the values of color coordinates.

Color Measurements
The cheese-based (curd) solid medium inoculated with A. arilaitensis Po102 was measured by using a CM-3500d spectrocolorimeter (Minolta Co., Ltd., Hino-shi, Tokyo, Japan) driven with SpectraMagic NX Pro. Software (Minolta). The reference illuminant was D65 (standard daylight). According to the Compagnie Internationale de l'Eclairage (CIE, 1978) the data were reported in the L*a*b* colorimetric system. A single-pieced disk of agar, large enough to cover the entire light spot, was cut in each Petri dish, from an area of homogeneous colony development. The sample was then held culture-down, and layered in the bottom of a 45 mm diameter glass Petri dish (CM-A128, Konica Minolta, Hino-shi, Tokyo, Japan) for incident light color measurements.
To characterize a color in the CIE L*a*b* color system, 3 colorimetric coordinates are obtained from the spectrocolorimeter. L* defines lightness (ranges from 0% to 100%, dark to light), a* indicates the red/green value (from −60 to 60, green to red), and b* denotes the blue/yellow value (from −60 to 60, blue to yellow). The attributes of color, C* and h • , describe the chroma (vividness or dullness) and the hue angle or tone of the color, respectively. The value of chroma C* is 0 at the center, and increases according to the distance from the center. Hue angle h • is defined as starting at the +a* axis and is expressed in degrees: 0 • would be +a* (red), 90 • would be +b* (yellow), 180 • would be −a* (green), and 270 • would be −b* (blue). Hue values correspond to the angle of the a*/b* coordinated of the points.

Statistical Analysis
The data were analyzed using SigmaPlot software (Systat Software, Inc., San Jose, CA, USA). To determine the color coordinate values of the biofilms produced by A. arilaitensis affected by initial pH, additional NaCl concentration, and deacidifying yeasts, one-way analysis of variance (one-way ANOVA) was performed in order to compare mean values among the whole data for each factor, while multivariate analysis of variance (MANOVA) was applied when determining the combined effects of controlled factors on the color of A. arilaitensis Po102 biofilms. The differences of considered variable were estimated by Tukey and Holm-Sidak tests according to an α risk of 5% and 1% for one-way ANOVA and MANOVA, respectively.

Diversity of Coloration by Arthrobacter arilaitensis Po102 on Different Deacidified Cheese-Based Media
After 28 days of simultaneous incubation at 12 • C under light conditions, the cheese-based (curd) solid media inoculated with A. arilaitensis Po102 were measured for color attributes using spectrocolorimetry in the CIE L*a*b* color system. Among the three types of media, depending on the method of deacidification, i.e., chemical deacidification by NaOH (CM NaOH ), biological deacidification by the yeast strain D. hansenii 304 (CM Dh304 ), and biological deacidification by the yeast strain K. marxianus 44 (CM Km44 ), the coloration of the cultures on CM NaOH showed hue alignments closer to the 90 • , with a large number of the a* negative value, expanding from −2.50 to 3.30, while the b* value varied between 6.00 and 29.30 (data not shown). Most of cultures on CM NaOH displayed cream or pale colors. By contrast, those on CM Dh304 and CM Km44 presented yellow colors, due to their hue alignments, which were set onto the color plane between 80.50 and 105.85 • . The results indicated a diversity of colors among these three media, particularly between CM NaOH and CM Dh304 , whose hues were statistically significantly different (p < 0.05).
The color position details of biofilms produced by A. arilaitensis Po102 on CM NaOH , CM Dh304 , and CM Km44 , coordinating with different initial pH and NaCl concentrations, are shown in Figure 1. The color projected onto a (a*, b*) plane, among the same deacidified media, showed a significant difference between the cultures in media with initial pH at 5.8, and those of 7.0 and 7.5 (p < 0.05), while there was no significant difference between pH 7.0 and 7.5 (p > 0.05). A narrow range of color coordinate values was obvious, and demonstrated at all concentrations of NaCl. However, the colors among the entire set with an initial pH 7.0 was wider distributed on (a*, b*) than those with other pH values in all types of media. variable were estimated by Tukey and Holm-Sidak tests according to an α risk of 5% and 1% for oneway ANOVA and MANOVA, respectively.

Diversity of Coloration by Arthrobacter arilaitensis Po102 on Different Deacidified Cheese-Based Media
After 28 days of simultaneous incubation at 12 °C under light conditions, the cheese-based (curd) solid media inoculated with A. arilaitensis Po102 were measured for color attributes using spectrocolorimetry in the CIE L*a*b* color system. Among the three types of media, depending on the method of deacidification, i.e., chemical deacidification by NaOH (CMNaOH), biological deacidification by the yeast strain D. hansenii 304 (CMDh304), and biological deacidification by the yeast strain K. marxianus 44 (CMKm44), the coloration of the cultures on CMNaOH showed hue alignments closer to the 90°, with a large number of the a* negative value, expanding from −2.50 to 3.30, while the b* value varied between 6.00 and 29.30 (data not shown). Most of cultures on CMNaOH displayed cream or pale colors. By contrast, those on CMDh304 and CMKm44 presented yellow colors, due to their hue alignments, which were set onto the color plane between 80.50 and 105.85°. The results indicated a diversity of colors among these three media, particularly between CMNaOH and CMDh304, whose hues were statistically significantly different (p < 0.05).
The color position details of biofilms produced by A. arilaitensis Po102 on CMNaOH, CMDh304, and CMKm44, coordinating with different initial pH and NaCl concentrations, are shown in Figure 1. The color projected onto a (a*, b*) plane, among the same deacidified media, showed a significant difference between the cultures in media with initial pH at 5.8, and those of 7.0 and 7.5 (p < 0.05), while there was no significant difference between pH 7.0 and 7.5 (p > 0.05). A narrow range of color coordinate values was obvious, and demonstrated at all concentrations of NaCl. However, the colors among the entire set with an initial pH 7.0 was wider distributed on (a*, b*) than those with other pH values in all types of media.

Changes in Color Development of Arthrobacter arilaitensis Po102
The color characteristics in terms of hue angle (h°) and chroma (saturation, C*), under the influences of pH, NaCl, and deacidification, are shown in Table 1. According to the analysis of

Changes in Color Development of Arthrobacter arilaitensis Po102
The color characteristics in terms of hue angle (h°) and chroma (saturation, C*), under the influences of pH, NaCl, and deacidification, are shown in Table 1. According to the analysis of  The color characteristics in terms of hue angle (h • ) and chroma (saturation, C*), under the influences of pH, NaCl, and deacidification, are shown in Table 1. According to the analysis of variance, there was a significant interaction between these variables on hue and chroma of the cultures of A. arilaitensis Po102 (p < 0.001). The hue angles extended from 80.56 to 105.74, with a large number of data showing responses close to h • value obtained from the blank agar, which corresponded to a cream or pale color (poorly colored). Among the same type of media, h • values decreased when cultivated on agars with initiated pH values of 7.0 and 7.5, leading to displaying more colored cultures than those with initial pH at 5.8, especially on media deacidified by both yeast strains, D. hansenii 304 and K. marxianus 44, which presented intense yellows at all levels of NaCl concentration. When 8% (w/v) NaCl was added into these media in coordination with initial pH at 5.8, the A. arilaitensis strain was likely not capable of forming color. By contrast, in the absence of NaCl, the cultures of all types of media showed yellow colors, with h • values of 87.04, 80.85, and 82.20 • for CM NaOH , CM Dh304 , and CM Km44 , respectively.
Similar to hue, pH variations, NaCl concentration, and deacidification affected the saturation, C*, of A. arilaitensis Po102 biofilms. The C* responses of these biofilms widely ranged from 6.56 to 38.54, whereby a large diversity of different groups obtained significantly differing results (p < 0.001). At pH value of 5.8, for all 3 kinds of media, the C* values of biofilms were higher at 0% of NaCl while, at other NaCl concentrations, the saturation was much lower. However, the lowest saturations of color were measured for all types of media, at all pH levels, for 8% (w/v) of additional NaCl content. The color attributes of A. arilaitensis Po102 cultures versus time were observed during the cultivation period on day (d) 7, 14, 21, and 28. Detailed positions of cultures of the A. arilaitensis strain, when projected in the CIE L*a*b* colorimetric scale after d 28 incubation, are shown in Figure 2. Among the same category, i.e., initial pH and NaCl concentration, the overall changes during the cultivation on CM Dh304 and CM Km44 appeared to be dependent on these variables. With the exception of the set at 0% NaCl, the color coordinate values obtained from all types of media at pH value 5.8, evolved very slightly throughout the cultivation period. On the other hand, when considering the cultures on these media with initial pH value at 5.8 without addition of NaCl, the development of color saturation of CMDh304 and CMKm44 were similar. There were, however, strong differences between these two media and CMNaOH. The difference of C* values between CMNaOH and the 2 others was obvious after d 7, and this remained until d 28 ( Figure  3). In contrast to CMDh304 and CMKm44, the C* responses of the A. arilaitensis strain cultures on CMNaOH slightly increased from the beginning until d 21, whose values were not significantly different from On the other hand, when considering the cultures on these media with initial pH value at 5.8 without addition of NaCl, the development of color saturation of CMDh304 and CMKm44 were similar. There were, however, strong differences between these two media and CMNaOH. The difference of C* values between CMNaOH and the 2 others was obvious after d 7, and this remained until d 28 ( Figure  3). In contrast to CMDh304 and CMKm44, the C* responses of the A. arilaitensis strain cultures on CMNaOH slightly increased from the beginning until d 21, whose values were not significantly different from On the other hand, when considering the cultures on these media with initial pH value at 5.8 without addition of NaCl, the development of color saturation of CMDh304 and CMKm44 were similar. There were, however, strong differences between these two media and CMNaOH. The difference of C* values between CMNaOH and the 2 others was obvious after d 7, and this remained until d 28 ( Figure  3). In contrast to CMDh304 and CMKm44, the C* responses of the A. arilaitensis strain cultures on CMNaOH slightly increased from the beginning until d 21, whose values were not significantly different from  On the other hand, when considering the cul without addition of NaCl, the development of col There were, however, strong differences between t values between CMNaOH and the 2 others was obvi 3). In contrast to CMDh304 and CMKm44, the C* respon slightly increased from the beginning until d 21, w ), and 8% (  On the other hand, when considering the cul without addition of NaCl, the development of colo There were, however, strong differences between t values between CMNaOH and the 2 others was obvio 3). In contrast to CMDh304 and CMKm44, the C* respon slightly increased from the beginning until d 21, w ).
On the other hand, when considering the cultures on these media with initial pH value at 5.8 without addition of NaCl, the development of color saturation of CM Dh304 and CM Km44 were similar. There were, however, strong differences between these two media and CM NaOH . The difference of C* values between CM NaOH and the 2 others was obvious after d 7, and this remained until d 28 ( Figure 3). In contrast to CM Dh304 and CM Km44 , the C* responses of the A. arilaitensis strain cultures on CM NaOH slightly increased from the beginning until d 21, whose values were not significantly different from those of the blank agar (p > 0.05). Then, in between d 21 and d 28, the changes in their saturation augmented rapidly, ranging from 8.99 to 24.53. Unlike the color saturation, the hue angles of the bacterial cultures were decreased throughout the incubation period. The h • was slightly changed during 14 days on CM Dh304 and CM Km44 , and 21 days on CM NaOH and, as is visible in the results, reduced at the end of cultivation period.  However, the changes in these responses were different from the cultures found on the biologically deacidified media coordinated with initial pH values higher than 5.8. The color development of A. arilaitensis Po102 on CMDh304 with initial pH values of 7.0 and 7.5, and on CMKm44 at pH 7.5, containing between 0-4% NaCl (w/v), increased extensively in the first 7 days of cultivation. The color of these cultures was changed from white-cream to yellow. Afterwards, from d 7 to d 28, the increase of saturation and hue slowed down. Although almost all color development patterns of the A. arilaitensis strain on CMKm44 were similar to those of CMDh304, the different forms of color evolution were apparently visible when coordinated with initial pH at 7.0. The changes of saturation C* provided by this strain on the two biologically deacidified media, in relation to initial pH and concentration of NaCl, versus time, are shown in Figure 4. L* values of the A. arilaitensis cultures on d 28 varied between 60.34 and 78.52, over all data observed in the present study (data not shown). Regardless of NaCl concentration, the highest lightness of the cultures from among all media was found at pH value of 7.0. Like the changes in color saturation, L* values appeared to depend on the controlled factors; in addition, the pattern of the change in lightness, in terms of period, was similar to those of chroma changes in CMDh304 and CMKm44 conditions with initial pH values of 7.0 and 7.5. However, the changes in these responses were different from the cultures found on the biologically deacidified media coordinated with initial pH values higher than 5.8. The color development of A. arilaitensis Po102 on CM Dh304 with initial pH values of 7.0 and 7.5, and on CM Km44 at pH 7.5, containing between 0-4% NaCl (w/v), increased extensively in the first 7 days of cultivation. The color of these cultures was changed from white-cream to yellow. Afterwards, from d 7 to d 28, the increase of saturation and hue slowed down. Although almost all color development patterns of the A. arilaitensis strain on CM Km44 were similar to those of CM Dh304 , the different forms of color evolution were apparently visible when coordinated with initial pH at 7.0. The changes of saturation C* provided by this strain on the two biologically deacidified media, in relation to initial pH and concentration of NaCl, versus time, are shown in Figure 4. L* values of the A. arilaitensis cultures on d 28 varied between 60.34 and 78.52, over all data observed in the present study (data not shown). Regardless of NaCl concentration, the highest lightness of the cultures from among all media was found at pH value of 7.0. Like the changes in color saturation, L* values appeared to depend on the controlled factors; in addition, the pattern of the change in lightness, in terms of period, was similar to those of chroma changes in CM Dh304 and CM Km44 conditions with initial pH values of 7.0 and 7.5. Apart from the yellow pigments synthesized by A. arilaitensis Po102, it was likely that this strain also produced other pigment(s) diffusing inside the cultures. Indeed, a large number of cheese curdbased agar plates inoculated with this bacterial strain displayed a pink/red-brown color ( Figure 5). The pink appeared slightly on certain plates of cultures on the three kinds of deacidified media during the first week of the incubation time. The A. arilaitensis strain was found to produce pink idering the cultures on these media with initial pH value at 5.8 lopment of color saturation of CMDh304 and CMKm44 were similar. nces between these two media and CMNaOH. The difference of C* thers was obvious after d 7, and this remained until d 28 ( Figure   44, the C* responses of the A. arilaitensis strain cultures on CMNaOH ng until d 21, whose values were not significantly different from idering the cultures on these media with initial pH value at 5.8 lopment of color saturation of CMDh304 and CMKm44 were similar. nces between these two media and CMNaOH. The difference of C* thers was obvious after d 7, and this remained until d 28 ( Figure   44, the C* responses of the A. arilaitensis strain cultures on CMNaOH ng until d 21, whose values were not significantly different from  On the other hand, when considering the cultures on these media with initial pH value at without addition of NaCl, the development of color saturation of CMDh304 and CMKm44 were simi There were, however, strong differences between these two media and CMNaOH. The difference of values between CMNaOH and the 2 others was obvious after d 7, and this remained until d 28 (Figu 3). In contrast to CMDh304 and CMKm44, the C* responses of the A. arilaitensis strain cultures on CMN slightly increased from the beginning until d 21, whose values were not significantly different fr  On the other hand, when considering the cultures on these media with initial pH value at without addition of NaCl, the development of color saturation of CMDh304 and CMKm44 were simi There were, however, strong differences between these two media and CMNaOH. The difference of values between CMNaOH and the 2 others was obvious after d 7, and this remained until d 28 (Fig  3). In contrast to CMDh304 and CMKm44, the C* responses of the A. arilaitensis strain cultures on CMN slightly increased from the beginning until d 21, whose values were not significantly different fr ).
Apart from the yellow pigments synthesized by A. arilaitensis Po102, it was likely that this strain also produced other pigment(s) diffusing inside the cultures. Indeed, a large number of cheese curd-based agar plates inoculated with this bacterial strain displayed a pink/red-brown color ( Figure 5). The pink appeared slightly on certain plates of cultures on the three kinds of deacidified media during the first week of the incubation time. The A. arilaitensis strain was found to produce pink pigments on CM NaOH and CM Km44 with the initial pH value of 7.5, while the pink pigmentation exhibited on CMDh304 when the initial pH of the media were 7.0 and 7.5. However, only with 0-2% NaCl in CM NaOH and less than 8% NaCl in CM Dh304 and CM Km44 , did the cultures display the pink coloration. By contrast, the other cultures that were treated differently, showed the red-brown color on d 21 and 28, respectively, including (i) CM Dh304 and CM Km44 , in combination with initial pH value of 7.0 and 7.5, with added 8% (w/v) NaCl; and (ii) all types of deacidified media at pH value of 5.8, without additional NaCl. pigments on CMNaOH and CMKm44 with the initial pH value of 7.5, while the pink pigmentation exhibited on CMDh304 when the initial pH of the media were 7.0 and 7.5. However, only with 0-2% NaCl in CMNaOH and less than 8% NaCl in CMDh304 and CMKm44, did the cultures display the pink coloration. By contrast, the other cultures that were treated differently, showed the red-brown color on d 21 and 28, respectively, including (i) CMDh304 and CMKm44, in combination with initial pH value of 7.0 and 7.5, with added 8% (w/v) NaCl; and (ii) all types of deacidified media at pH value of 5.8, without additional NaCl. A short experiment was then performed to determine these pigments, in order to obtain the first information for further investigation on the role of pink pigmentation by A. arilaitensis in smearripened cheeses. The results showed that the pink pigments excreted by A. arilaitensis Po102 were water-soluble, and the extracts exhibited fine structure with wavelengths of maximum absorption at 394, 497, 530, 567, and 620 nm; in addition, the extract fluoresced under UV light. These basic occurrences were similar to data obtained from the red-brown pigments produced by A. nicotianae, which were not precisely identified in the study of Bockelmann et al. [26]. Based on the review of pigments produced by the genus Arthrobacter [27], most bacteria in this genus could produce pigments in a broad range of hues, e.g., orange and yellow (riboflavin, carotenoids), blue and green (indigoidine, indochrome, and derived salts), and red (porphyrins, carotenoids). However, in comparison with the general porphyrin characteristics, including absorption spectra and fluorescence property, as well as the recent information reported by Cleary et al. [28], the pink/redbrown observed in the present study would probably be pigments in a porphyrin cluster.

Changes in pH of Arthrobacter arilaitensis Po102 Cultures
Since the color development appeared not only on the surface but, also, inside the media inoculated with A. arilaitensis Po102, the cultures were blended and were then measured for their pH. When considering the whole experimental pH data of the present study, all controlled variables seemed to potentially influence the changes of pH values of the cultures. Differences in changes of pH values appeared between each type of deacidified media, especially between the chemical deacidified medium and the biologically deacidified media by yeast strains. The comparison of changes in pH of the cultures, categorized by the biological deacidification methods, could help distinguish some configuration of these changes. With an initial pH value of 5.8, two major patterns among CMDh304 and CMKm44 were categorized, i.e., (i) slightly increased or decreased throughout the A short experiment was then performed to determine these pigments, in order to obtain the first information for further investigation on the role of pink pigmentation by A. arilaitensis in smear-ripened cheeses. The results showed that the pink pigments excreted by A. arilaitensis Po102 were water-soluble, and the extracts exhibited fine structure with wavelengths of maximum absorption at 394, 497, 530, 567, and 620 nm; in addition, the extract fluoresced under UV light. These basic occurrences were similar to data obtained from the red-brown pigments produced by A. nicotianae, which were not precisely identified in the study of Bockelmann et al. [26]. Based on the review of pigments produced by the genus Arthrobacter [27], most bacteria in this genus could produce pigments in a broad range of hues, e.g., orange and yellow (riboflavin, carotenoids), blue and green (indigoidine, indochrome, and derived salts), and red (porphyrins, carotenoids). However, in comparison with the general porphyrin characteristics, including absorption spectra and fluorescence property, as well as the recent information reported by Cleary et al. [28], the pink/red-brown observed in the present study would probably be pigments in a porphyrin cluster.

Changes in pH of Arthrobacter arilaitensis Po102 Cultures
Since the color development appeared not only on the surface but, also, inside the media inoculated with A. arilaitensis Po102, the cultures were blended and were then measured for their pH. When considering the whole experimental pH data of the present study, all controlled variables seemed to potentially influence the changes of pH values of the cultures. Differences in changes of pH values appeared between each type of deacidified media, especially between the chemical deacidified medium and the biologically deacidified media by yeast strains. The comparison of changes in pH of the cultures, categorized by the biological deacidification methods, could help distinguish some configuration of these changes. With an initial pH value of 5.8, two major patterns among CM Dh304 and CM Km44 were categorized, i.e., (i) slightly increased or decreased throughout the entire cultivation period, and (ii) slightly changed during the first 14 days and, then, highly increased afterward. entire cultivation period, and (ii) slightly changed during the first 14 days and, then, highly increased afterward. As shown in Figure 6, pH of the cultures of A. arilaitensis Po102 changed slightly when the bacterial strain grew on the media containing NaCl concentrations, whereas the pH changes were manifest when grown on media without NaCl. However, when the initial pH value was higher than 5.8, patterns of the changes in pH of A. arilaitensis Po102 cultures were dissimilar. Two different patterns were clearly observable on both CMDh304 and CMKm44, with initial pH values of 7.0 and 7.5, i.e., (i) pH augmented rapidly during the first 7 days, and after that, slowly, and (ii) pH changed slightly during 7 or 14 days, and then increased rapidly (Figure 7). On the other hand, when considering the cultures on these media with initial pH value at 5.8 without addition of NaCl, the development of color saturation of CMDh304 and CMKm44 were similar. There were, however, strong differences between these two media and CMNaOH. The difference of C* values between CMNaOH and the 2 others was obvious after d 7, and this remained until d 28 ( Figure  3). In contrast to CMDh304 and CMKm44, the C* responses of the A. arilaitensis strain cultures on CMNaOH slightly increased from the beginning until d 21, whose values were not significantly different from ), 2% ( On the other hand, when considering the cultures on these media with initial pH value at 5.8 without addition of NaCl, the development of color saturation of CMDh304 and CMKm44 were similar. There were, however, strong differences between these two media and CMNaOH. The difference of C* values between CMNaOH and the 2 others was obvious after d 7, and this remained until d 28 ( Figure  3). In contrast to CMDh304 and CMKm44, the C* responses of the A. arilaitensis strain cultures on CMNaOH slightly increased from the beginning until d 21, whose values were not significantly different from ), 4% ( On the other hand, when considering the cultures without addition of NaCl, the development of color sat There were, however, strong differences between these values between CMNaOH and the 2 others was obvious a 3). In contrast to CMDh304 and CMKm44, the C* responses o slightly increased from the beginning until d 21, whose ), and 8% ( On the other hand, when considering the cultures without addition of NaCl, the development of color sat There were, however, strong differences between these values between CMNaOH and the 2 others was obvious a 3). In contrast to CMDh304 and CMKm44, the C* responses o slightly increased from the beginning until d 21, whose ).
As shown in Figure 6, pH of the cultures of A. arilaitensis Po102 changed slightly when the bacterial strain grew on the media containing NaCl concentrations, whereas the pH changes were manifest when grown on media without NaCl. However, when the initial pH value was higher than 5.8, patterns of the changes in pH of A. arilaitensis Po102 cultures were dissimilar. Two different patterns were clearly observable on both CM Dh304 and CM Km44 , with initial pH values of 7.0 and 7.5, i.e., (i) pH augmented rapidly during the first 7 days, and after that, slowly, and (ii) pH changed slightly during 7 or 14 days, and then increased rapidly (Figure 7).  In the present study, pH changed during the growth, and seemed to be correlated with the development of pink color by A. arilaitensis Po102 on the cheese-based solid media studied. Regardless of the concentration of additional NaCl and the method used for deacidification, all of experimental agars inoculated with this bacterial strain would intensely present a pink color when their pH was increased to higher than 7.5 during the cultivation (data not shown).

Discussion
During the last few years, A. arilaitensis has been identified as one important bacterial strain responsible for the coloration of smear-ripened cheeses; however, to our knowledge, only the study by Leclercq-Perlat et al. [29] investigated the coloring capacity of this species on a variety of deacidified media prepared from the curd of several ripened cheeses, including Epoisses, Munster, Livarot, and Reblochon. On the other hand, a number of previous studies had focused on B. linens for its involvement in the cheese coloration in relation with various factors, such as pH, salt concentration, humidity, light, temperature, and yeast used for deacidification, as well as on the interactions between the B. linens strain and other microorganisms isolated from ripened cheeses. The present study is the first investigation which concentrated on the pigmentation of A. arilaitensis under the influence of initial pH, NaCl content, and deacidifying yeasts.
According to the data obtained using spectrocolorimetry in the CIE L*a*b* color system, A. arilaitensis Po102 provided diverse color characteristics on three types of deacidified media. However, most of the biofilms on CMNaOH displayed h o values in the greenish-yellow range, while hues of CMDh302 and CMKm44 were set in the reddish-yellow range. The h° values on CMNaOH in the present study were similar to the previous study investigating the pigments produced by a variety of the A. arilaitensis strains. They showed the color characteristic of yellow A. arilaitensis strains with hues varying from 92.61 to 98.57 [9]. This could be explained by the resemblance of the media used for the 2 studies, as both of them had an initial pH at 7.0, adjusted by using the chemical reagent NaOH. Besides, this bacterial species was reported to be sensitive to the type of media used for its cultivation, as its color attributes were distinctly different when grown on media prepared from various cheese curds [29]. n considering the cultures on these media with initial pH value at 5.8 development of color saturation of CMDh304 and CMKm44 were similar. differences between these two media and CMNaOH. The difference of C* he 2 others was obvious after d 7, and this remained until d 28 ( Figure  CMKm44 n considering the cultures on these media with initial pH value at 5.8 development of color saturation of CMDh304 and CMKm44 were similar. differences between these two media and CMNaOH. The difference of C* he 2 others was obvious after d 7, and this remained until d 28 ( Figure  CMKm44, the C* responses of the A. arilaitensis strain cultures on CMNaOH ginning until d 21, whose values were not significantly different from ), 4% ( On the other hand, when considering the cultures on these media with initial pH val without addition of NaCl, the development of color saturation of CMDh304 and CMKm44 were There were, however, strong differences between these two media and CMNaOH. The differe values between CMNaOH and the 2 others was obvious after d 7, and this remained until d 2 3). In contrast to CMDh304 and CMKm44, the C* responses of the A. arilaitensis strain cultures on slightly increased from the beginning until d 21, whose values were not significantly differ ), and 8% ( On the other hand, when considering the cultures on these media with initial pH va without addition of NaCl, the development of color saturation of CMDh304 and CMKm44 wer There were, however, strong differences between these two media and CMNaOH. The differe values between CMNaOH and the 2 others was obvious after d 7, and this remained until d 2 3). In contrast to CMDh304 and CMKm44, the C* responses of the A. arilaitensis strain cultures on slightly increased from the beginning until d 21, whose values were not significantly differ ).
In the present study, pH changed during the growth, and seemed to be correlated with the development of pink color by A. arilaitensis Po102 on the cheese-based solid media studied. Regardless of the concentration of additional NaCl and the method used for deacidification, all of experimental agars inoculated with this bacterial strain would intensely present a pink color when their pH was increased to higher than 7.5 during the cultivation (data not shown).

Discussion
During the last few years, A. arilaitensis has been identified as one important bacterial strain responsible for the coloration of smear-ripened cheeses; however, to our knowledge, only the study by Leclercq-Perlat et al. [29] investigated the coloring capacity of this species on a variety of deacidified media prepared from the curd of several ripened cheeses, including Epoisses, Munster, Livarot, and Reblochon. On the other hand, a number of previous studies had focused on B. linens for its involvement in the cheese coloration in relation with various factors, such as pH, salt concentration, humidity, light, temperature, and yeast used for deacidification, as well as on the interactions between the B. linens strain and other microorganisms isolated from ripened cheeses. The present study is the first investigation which concentrated on the pigmentation of A. arilaitensis under the influence of initial pH, NaCl content, and deacidifying yeasts.
According to the data obtained using spectrocolorimetry in the CIE L*a*b* color system, A. arilaitensis Po102 provided diverse color characteristics on three types of deacidified media. However, most of the biofilms on CM NaOH displayed h • values in the greenish-yellow range, while hues of CM Dh302 and CM Km44 were set in the reddish-yellow range. The h • values on CM NaOH in the present study were similar to the previous study investigating the pigments produced by a variety of the A. arilaitensis strains. They showed the color characteristic of yellow A. arilaitensis strains with hues varying from 92.61 to 98.57 [9]. This could be explained by the resemblance of the media used for the 2 studies, as both of them had an initial pH at 7.0, adjusted by using the chemical reagent NaOH. Besides, this bacterial species was reported to be sensitive to the type of media used for its cultivation, as its color attributes were distinctly different when grown on media prepared from various cheese curds [29].
All controlled variables i.e., initial pH, NaCl concentration, and deacidifying yeasts, appeared to affect the pigments produced by A. arilaitensis Po102. When the pH value was initiated at 5.8, this bacterial strain was likely to notably produce pigments in the absence of NaCl. As established by the plate counting technique, the viable cell concentration of the A. arilaitensis strain was significantly higher without NaCl compared to cultures grown on media containing 2-8% (w/v) NaCl (data not shown). This incidence could imply that the color intensity of biofilms was directly proportional to the number of bacterial cells. As a result, it also potentially pointed out a direct effect of initial pH on the growth of this A. arilaitensis strain. These results were consistent with the previous studies on the pigmentation of B. linens, which reported a similar relation between pH, viable cell concentration, and color intensity [21,29].
A comparison of the color attributes provided by A. arilaitensis Po102 on either CM Dh304 or CM Km44 , at initial pH values of both 7.0 and 7.5, apparently showed differences between the cultures containing 8% (w/v) NaCl and the other ones. Meanwhile, the color attributes were quite similar among the cultures on media with 0-4% (w/v) NaCl added. In the same way, these differences seemed to depend on the intensity of viable cells as the number of viable A. arilaitensis Po102 on plate count agar was significantly lower when grown on media containing 8% (w/v) NaCl compared to agars with other concentrations (data not shown).
When considering the color of A. arilaitensis Po102 cultures on CM Dh304 and CM Km44 at initial pH values higher than 5.8, with 0-4% (w/v) NaCl, whose conditions were probably more suitable for the growth of the A. arilaitensis strain in the present study, the color attributes, including a*, b*, C*, and h • value, varied greatly among these media. In addition, during the incubation period, the changes in color and pH were diversified. These results could refer to the relation between the yeast strains used for the deacidification, and the pigment production of this bacterial strain. For the media prepared by biological deacidification, yeast cells were eliminated by pasteurization when the pH of the media reached the value required. For this reason, the pigmentation of A. arilaitensis Po102 was possibly impacted by the difference in biochemical composition of CM Dh304 and CM Km44 . Leclercq-Perlat et al. [17] reported that K. marxiamus required more lactose and lactate than D. hansenii for their growth during the deacidification; therefore, the ripening bacteria probably did not have sufficient lactate to produce energy for their growth. This resulted in a variation of the cell biomass, which directly influenced the color. Furthermore, the ability to produce pigments of A. arilaitensis appeared to depend on the type of amino acids and caseinate present in the milk media [30]. According to Roostita and Fleet [30], D. hansenii was able to produce several free amino acids (predominantly proline, alanine, glycine, arginine, and glutamic acid) and free fatty acid, during its growth in milk. This probably affected the growth and, thus, pigment production of the ripening bacteria A. arilaitensis Po102.
In the present study, the development of pink/red-brown colors on all types of deacidified media was evident in cultures produced by A. arilaitensis Po102, as shown in Figure 5. The initial pH of cheese-based solid media seemed to implicate the pink coloration of this bacterium, as there was no development of pink color when the pH was acidic (pH 5.8). Unlike the initial pH, NaCl appeared not to directly affect the pink pigmentation of the A. arilaitensis strain because the red pigments were found in cultures on media which had an initial pH at 7.0 and 7.5, containing all NaCl concentrations. However, after observing the changes in pH during its growth, it could be indicated that the A. arilaitensis strain would produce pink pigments when the pH value of the inoculated media were higher than 7.5. Bockelmann [4] reported that a ripening bacteria, A. nicotianae, a yellow-pigmented bacteria isolated from German cheeses, produced red-brown pigments when co-cultivated with D. hansenii and other bacteria under alkaline growth conditions in a shake liquid milk model system. However, the pure culture of this strain also produced red-brown pigments in liquid growth media containing casein hydrolysate. The same author demonstrated a possibility to use this bacterial species as a ripening culture for the typical red-brown Tilsit cheese. The pink coloration, sometimes identified as pink-brown or dark-brown, has often been reported as a defect in a wide variety of smear-ripened cheeses, which influences the cheese quality, and is rejected by consumers [31,32]. Many microorganisms as primary starters and secondary cultures have been implicated in the pink coloration in cheeses; however, to our knowledge, the pink pigments produced by A. arilaitensis have not yet been clearly and fully identified in many experimental conditions.

Conclusions
The color of cultures provided by A. arilaitensis Po102 on cheese-based (curd) solid media were prepared using different deacidification methods, i.e., (i) chemical deacidification by NaOH (CM NaOH ), (ii) biological deacidification by the yeast strain D. hansenii 304 (CM Dh304 ), and (iii) biological deacidification by the yeast strain K. marxianus 44 (CM Km44 ). In association with varied initial pH values and NaCl in varying concentrations, the cultures on all kinds of media were analyzed by spectrocolorimetry in the CIE L*a*b* color system after 28 days of cultivation at 12 • C under light conditions. According to collected data, all controlled factors appeared to affect the pigments produced by the A. arilaitensis strain, in terms of hue, saturation, and lightness. NaCl content in the media showed distinct inhibitory effects on the development of color by this strain when the initial pH was at 5.8. By contrast, when the initial pH of the media was higher (7.0, 7.5), only the highest concentration of NaCl (8%) had this effect. The coloring capacity of this bacterial species was generally higher when D. hansenii 304 was used for the deacidification, compared to K. marxianus 44. Based on the pH measurement, the yeast strains D. hansenii 304 and K. marxianus 44 appeared to impact, differently, a change in pH of the cultures. These changes were assumed to be associated with the occurrence of pink/red-brown color in the A. arilaitensis cultures, as the pink coloration manifested itself when the pH of the cultures was higher than 7.5. In general, the pink color in smear-ripened cheeses is defined as a defect. Therefore, in order to avoid this defect in the cheese production, the correlation between these factors, particularly, the change in pH during ripening, should be further observed. Different deacidifying yeast strains, mainly found in smear-ripened cheeses, would possibly be applied, and the change in pH throughout the ripening period should be critically focused on. The potential results may indicate the optimum pH that would maintain the growth and color development of A. arilaitensis or other pigmented bacterial strains, but not induce the red pigments. This will be useful for the cheese production to prevent economic losses from the pink discoloration in cheeses.