J OURNAL OF B IOSCIENCE AND B IOENGINEERING
V ol. 98, No. 1, 48–56. 2004
Fluorescent In Situ Hybridization Analysis of Open Lactic Acid Fermentation of Kitchen Refuse Using rRNA-Targeted
Oligonucleotide Probes
KENJI SAKAI,1*MASATSUGU MORI,1AKIRA FUJII,1YUKO IWAMI,1
EKACHAI CHUKEATIROTE,1,2AND YOSHIHITO SHIRAI3
Department of Applied Chemistry, Faculty of Engineering, Oita University, 700 Dannoharu, Oita 870-1192, Japan,1Department of Biotechnology, School of Science, Mae Fah Luang University, Tasud Muang,
Chiang Rai 57100, Thailand,2and Department of Biochemical Engineering, Graduate School
of Life Science and Systems Engineering, Kyushu Institute of Technology,
2-4 Hibikino, Wakamatsu, Kitakyushu, Fukuoka 808-0196, Japan3
Received 16 January 2004/Accepted 8 May 2004
Reproducible amounts of lactic acid accumulate in minced kitchen refuse under open condi-
tions with intermittent pH neutralization [Sakai et al., Food Sci. Technol. Res., 6, 140 (2000)].
Here, we showed that such pH-controlled open fermentation of kitchen refuse reproducibly re-
sulted a selective proliferation of a major lactic acid bacterial (LAB) species. In one experiment,
the predominant microorganisms isolated during the early phase (6h) were Gammaproteobac-
teria. In contrast, those that predominated during the late phase (48h) were always Lactobacillus
plantarum in three independent experiments. To further quantify the microbial community within
open lactic acid fermentation, we performed fluorescent in situ hybridization (FISH) analysis tar-
geting 16S (23S) rRNA. We designed two new group-specific DNA probes: LAC722(L) was active
for most LAB including the genera Lactobacillus,Pediococcus,Leuconostoc and Weisella, whereas
Lplan477 was specific for L. plantarum and its related species. We then optimized sample prepa-
ration using lysozyme and hybridization conditions including temperature, as well as the forma-
mide concentration and the salt concentration in the washing buffer. We succeeded in quantifica-
tion of microorganisms in semi-solid, complex biological materials such as minced kitchen refuse
by taking color microphotographs in modified RGB balance on pre-coated slides. FISH analysis
of the fermentation of kitchen refuse indicated that control of the pH swing leads to domination
by the LAB population in minced kitchen refuse under open conditions. We also confirmed that L.
plantarum, which generates lactic acid in high quantities but with low optical activity, became the
dominant microorganism in kitchen refuse during the late phase of open fermentation.
[Key words:kitchen refuse, food waste, lactic acid, open fermentation, lactic acid bacteria, fluorescent in situ
hybridization (FISH), 16S rRNA, microbial analysis]
Poly-L-lactic acid (PLLA), a biodegradable and recycla-ble plastic is now being produced from cornstarch via lactic acid fermentation (1) and the potential of PLLA as a green
plastic has been discussed (2). Such materials should be-come globally applicable, but producing PLLA and other plant-derived plastics is costly, which prevents widespread
application. Municipal food waste constitutes a potential bio-mass resource in Japan, since around 20% of the 50 million tonnes or so of waste that are annually generated consist of
refuse from kitchens and food industry. We recently pro-posed a new system for recycling municipal food waste, from which high quality PLLA plastics can be produced (3). The system would be effective not only for treating munici-pal food waste, but also for avoiding a resource conflict, since corn starch feedstock is also a source of foodstuffs for humans and other animals.Our proposal consists of four steps: removal of endogenous D,L-lactic acids from collected food waste by a propionibacterium, semi-solid L-lactic acid fermentation, purification of L-lactic acid by butyl-esterifica-tion and L-lactic acid polymerization via LL-lactide. Lacto-bacillus rhamnosus fermented L-lactic acid in minced and autoclaved food waste. However, the production of PLLA from food waste requires relatively more energy than that from cornstarch (3).
During the investigation, we found that lactic acid was selectively produced from non-autoclaved minced kitchen refuse under open conditions, regardless of inoculated lactic acid bacteria (LAB) (4). Intermittent pH neutralization of minced kitchen refuse enhanced lactic acid production and stabilized the accumulated lactic acid. Such non-sterilized open fermentation can proceed in simpler facilities than those required for sterilized closed fermentation, and does not need energy for sterilization. Thus, non-sterilized open
* Corresponding author.e-mail: sakai@cc.oita-u.ac.jp
phone/fax: +81-(0)97-554-7892
48
FISH ANALYSIS OF OPEN LACTIC FERMENTATIONS
V OL. 98, 200449
fermentation of kitchen refuse could be applied to the on-
site treatment of scattered municipal food waste. However,
the optical activity of lactic acid produced by the non-steril-ized open fermentation was low. Polylactic acid with lower
optical activity is less crystalline than PLLA, and its use
might be restricted, for example, to molding applications, when such crystallinity is not required. However, a more
precise analysis of microbial constituents is required to im-
prove and control open lactic acid fermentation. Here, we conventionally isolated and identified the microbes in kitchen refuse during open fermentation. Although we initially cul-
tivated, purified and characterized several microbial isolates, these processes are laborious and time-consuming, and only
predominant cultivable species can be counted and identi-
fied. Therefore, we applied 16S (23S) rRNA-targeted fluo-rescent in situ hybridization (FISH) to analyze the microbial population during open lactic acid fermentation. The FISH procedure has altered the nature of microbial ecology stud-ies (5, 6) as the microbial community can be rapidly and directly analyzed. Furthermore, ribosomal RNA is a useful target because of its unique organization, highly conserved structure, high copy number, and the on-line availability of an organized database with phylogenetic information (7). However, few FISH analyses of lactic acid bacteria have been reported (8). We designed new probes for our target LAB and determined the optimal conditions for their spe-cific detection.
MATERIALS AND METHODS
Fermentation of kitchen refuse Model kitchen refuse (MKR) consisted of the following (4): 14% (w/w) fish residues, 40% (w/w) vegetables (carrot, potato and Chinese radish peel), 30% (w/w) fruits (banana, apple and orange peel), 10% (w/w) cooked rice and 6% (w/w) green tea residues. The ingredients and proportions were established from a survey of Japanese kitchen refuse (9). All the ingredients were minced thoroughly in an equal amount of tap water, then the resulting paste was adjusted to pH7.0 (initial pH, about 5.9) and stored at -20°C. Static fermenta-tion initiated by adding an inoculant (3ml of non-autoclaved MKR or commercial Japanese pickles) to the non-autoclaved MKR paste (30ml in 50ml conical tubes; Falcon 2070, Becton Dickinson, Franklin Lakes, NJ, USA),proceeded at 37°C. To test the effect of the inoculant, the MKR was autoclaved and then a starter culture of isolated strains, or non-autoclaved kitchen refuse (1ml) was added to the autoclaved MKR, which was then statically incubated at 37°C. The fermentation process was periodically monitored (every 6 or 12h) by adjusting the culture pH to 7.0 with 10M so-dium hydroxide, and by manually mixing with a glass rod. All pro-cedures were undertaken without sterilization (open fermentation). Other than standard MKR, refuse samples collected from kitchens at Oita-city, Iizuka-city, Japan, and Kuala Lumpur, Malaysia, were used as seed cultures, and sources for bacterial isolation, after be-ing minced in an equal amount of tap water. Commercial Japanese pickles (Onimaru Corp., Fukuoka) were also used as a seed culture for MKR fermentation.
Viable cell counts and isolation of bacteria The fermented refuse samples were collected every 12h after the first 6h and the microbial population was analyzed. To investigate the microbial population using a culture-dependent method, the samples were serially diluted tenfold in 0.85% saline (103–108). Viable cells (cfu/ml) in the diluted samples (0.1ml) were estimated by plate count method using three series each of various agar media. The total number of bacteria was determined using Standard Agar (per liter: 2.5g yeast extract, 5.0g peptone, 1.0g glucose, 15.0g agar). The number of viable LAB was estimated in DeMan Rogosa Sharpe (MRS) agar. The cell number of viable Escherichia coli and re-lated species (coliform bacteria) was determined on MacConkey agar. All media were purchased from Nissui Pharmaceutical (Tokyo) and samples were anaerobically cultured at 37°C for 1–2d (BBL GasPak Anaerobic System; Becton Dickinson) unless otherwise stated.
Microorganisms were screened from the fermented refuse sam-ples on Standard Agar.Fifty colonies were randomly selected and initially characterized based on morphology and basic physiologi-cal properties. These included cell shape and size, spore formation, Gram staining, motility and catalase activity. All bacterial strains were further tested using API-20E and API-50CHL identification kits (Bio-Merieux, Marcy, I’Etoile, France). Partial 16S rDNA se-quences were analyzed as described (10), after amplifying the gene using the bacterial conserved primers70F(5￠-TAACACATGCA AGTCGA-3￠) and 1387R (5￠-GGGAACTTATTCACCG-3￠) in a PCR. Strains were routinely cultured under static conditions in standard nutrient broth for 18–24h at 37°C.
Organic acid analysis Fermented samples (1ml) were cen-trifuged to remove waste residues (6500′g, 15min, 4°C), and we used HPLC to identify organic acids (formic, acetic, propionic, lactic and butyric acids) in the supernatant. The analytical con-ditions were as follows: solvent delivery system, LC-10 AD; col-umn, Shim-Pak SCR-102H; detector, CDD-6A; mobile phase, 5 mM p-toluenesulfonic acid; reaction mixture, 20mM Bis–Tris con-taining 5mM p-toluenesulfonic acid and 100m M EDTA; tempera-ture, 40°C. Selective accumulation of lactic acid (LA selectivity) was defined as the ratio of the lactic acid concentration to that of total organic acids (%). Optical isomers of lactic acid were charac-terized using D- and L-specific lactate dehydrogenase (Boehringer Mannheim, Indianapolis, IA, USA), according to the manufac-turer’s instructions. The optical activity of lactic acid (%) was de-fined as (L-D)/(L+D)′100.
Design of the probes for LAB detection The new probe se-quences complementary to the 16S rDNA of most LAB and of L. plantarum were designed after visual inspection of a sequence alignment of the LAB. 16S rDNA sequence of 21 LAB strains in-cluding those of Lactobacillus,Weisella,Lactococcus, and several reference microorganisms including E. coli and Bacillus subtilis (Fig. 1) were retrieved from the RDP II (7) and their multiple alignments were performed using Clustal W (11). Specificity of se-lected oligonucleotide sites were tested for further checked.Other than probes for LAB, we also used EUB338 for domain bacteria (6) and GAM42a (23S) for Gammaproteobacteria (12).Oligonu-cleotides and their fluorescent derivatives (5￠-labelled with either FITC or rhodamine) were purchased from Hokkaido System Sci-ence (Sapporo).
Fixation and permeabilization of MKR samples Cells were fixed and hybridized using the protocol reported by Amann (13) with some modification. The MKR samples were fixed in 3% paraformaldehyde/PBS (PBS: 130mM NaCl, 10mM sodium phos-phate buffer, pH7.0) for 1–3h at 4°C, pelleted by centrifugation (3500′g, 15min, 4°C) and then stored in a 1:1 mixture of ethanol and PBS. Fixed cell suspensions diluted using an ethanol/PBS mixture (8m l) were spotted on coated glass slides (12 wells per slide, Cel-Line; Erie Scientific, Portsmouth, New Hampshire, UK), dried at 46°C for 30min and immersed for 3min each in 50%, 80%, and 99% (v/v) ethanol. Cell smears were covered with 20m l of lysozyme (50mg/ml PBS; 37°C, 30min). Enzymatic digestion was terminated by thoroughly rinsing the slides with distilled water followed by air-drying at room temperature. Permeabiliza-tion activity was tested using lysozyme (chicken egg Type IV, 60,000units/mg; Wako, Tokyo), labiase (10munits/mg; Seikagaku-
SAKAI ET AL.J. B IOSCI . B IOENG .,
50Kogyo, Tokyo), mutanolysin (6000units/mg; Sigma-Aldrich Japan,Tokyo), and proteinase K (23units/mg; Wako).
Whole cell fluorescent in situ hybridization (FISH)Sam-ples in 8m l of hybridization buffer (0.9M NaCl, 20mM Tris–HCl [pH 7.2], 0.01% SDS, 60–120ng of probe, 0–30% formamide) were applied to the wells on the slides and incubated at an appropriate temperature for 2h in an equilibrated humid chamber. Free probes were removed from the slide by rinsing with 1ml of washing so-lution (20mM Tris–HCl, 0.01% SDS, 0–450mM NaCl, 5mM EDTA). The slides were then incubated at 48°C for 20min in 50ml of washing solution, rinsed briefly with distilled water, air-dried and mounted with SlowFade antifading reagent (Molecular Probes, Eugene, OR, USA). Fluorescence was observed using an epifluorescence microscope (BX50; Olympus, Tokyo) and photo-micrographs were taken using a chilled 3-CCD color camera (640′483pixels, C5810; Hamamatsu Photonics, Shizuoka).
Fluoro-stained cells were observed and counted in duplicate wells for appropriately diluted samples, and those in five randomly selected fields (1000-fold magnification) were recorded in a com-puter. Thereafter, fluoro-stained cells on the computer monitor (17inches) were counted. The concentration of fluoro-stained cells (cells/ml) was calculated as follows:
(C ′D ′A m )/(V ′A w ′F )
where C is the cell number observed on a monitor (cells), D is the dilution rate (–), A m was the monitor area calculated using an ob-jective micrometer (0.0024mm 2),V was the sample volume spot-ted in the well (8.0′10–3ml), A w was the well area on the glass slide (19.6mm 2) and F was the fixation efficiency (0.82). The cell concentration was generally expressed as a logarithm. The relative standard deviation was 19%, and the lower limit for the calculation was 4.7 log (cells/ml) (5′104cells/ml).
RESULTS
Identification of microorganisms in open fermentation of kitchen refuse We showed that high levels of lactic acid (27–45g/l ) with high LA selectivity (over 94%) were generated during MKR fermentation by intermittent pH neu -tralization, even without inoculation under open conditions (4). Such fermentation was reproducible regardless of the source of kitchen refuse. We examined the microbial struc-ture of collected refuse samples by random isolation from anaerobic nutrient agar plates (Table 1). The most abundant cultivable organisms during the early phase of run A (6h)were gram-negative bacilli, which were classified into two groups according to their physiological properties in the API20E test. The profile of the largest group (76% of iso-lates) was the same as that of Klebsiella pneumoniae (posi-tive, b -galactosidase, lysine decarboxylase, assimilation of citric acid, urease, acetoin formation, fermentation of D -glu-cose, D -mannitol, inositol, D -sorbitol, L -rhamnose, sucrose,melibiose, amygdalin, L -arabinose, NO 3– reduction to NO 2–;negative, arginine dihydrolase, ornithine decarboxylase, H 2S production, tryptophan deaminase, and gelatinase, oxidase test, NO 3– reduction to N 2) and the partial sequence of one sample of strain OME1-3 (638bp, accession no. AB178484)was 98% homologous to K. pneumoniae ATCC13884 (ac-cession no. AB004753). The API20E profile of the other group (7% isolates) was the same as that of Enterobacter cloaceae (positive, b -galactosidase, arginine dihydrolase,ornithine decarboxylase, assimilation of citric acid, acetoin formation, fermentation of D -glucose, D -mannitol, D -sorbi-tol, L -rhamnose, sucrose, melibiose, amygdalin, L -arabinose and NO 3– reduction to NO 2–; negative, lysine decarboxylase,urease, tryptophan deaminase, gelatinase, fermentation of inositol, oxidase test, NO 3– reduction to N 2). The partial se -quence of strain OME1-13 (578bp, accession no. AB178485)was 100% homologous to Enterobacter agglomerans str.AH16 (accession no. AJ010096). E. agglomerans is differ-ent from E. cloaceae only in activities of arginine hydrolase and ornithine decarboxylase. Only 10% of the total iso-lates were gram-positive cocci and these were subsequently identified as Lactococcus lactis from their profiles in the APICHL test (fermentable sugars: D -ribose, D -xylose, D -ga -lactose, D -glucose, D -fructose, D -mannose, mannitol, sorbi-tol, N -acetyl-D -glucosamine, amygdalin, arbutin, esculin,salicin, cellobiose, maltose, lactose, sucrose, trehalose, me-lezitose, starch, gentiobiose, turanose and gluconate. Non-fermentable sugars: L -arabinose, melibiose, inulin, raffinose,glycogen, xylitol, D -lyxose, D -tagatose, D -fucose, L -fucose,D -arabitol, L -arabitol, 2-ketogluconate, 5-ketogluconate). In contrast, all bacteria isolated after fermenting the three batches for 48h (stationary conditions of runs A, B and C)were gram-positive rods or cocci that did not form spores.After run A, the sugar assimilation pattern of 63% of these was the same as that of L. plantarum in the APICHL test.They fermented L -arabinose, D -ribose, D -galactose, D -glu-cose, D -fructose, D -mannose, mannitol, sorbitol, N -acetyl-D -glucosamine, amygdalin, arbutin, esculin, salicin, cellobi-ose, maltose, lactose, melibiose, sucrose, trehalose, inulin,melezitose, raffinose, gentiobiose, turanose and gluconate.The size of KY-1 cells was 1.4–3.8′0.7m m, and this prolif-
FIG.1.Phylogenetic neighbor-joining tree of 16S rDNA shows target groups of the probes, LAC722(L), LAC722(H), and Lplan447.Horizontal length reflects phylogenetic depth. Bar represents 10%sequence divergence. Accession numbers in DDBJ for 16S rDNA se-
quence data for each strain used are included in the figure.
FISH ANALYSIS OF OPEN LACTIC FERMENTATIONS V OL . 98, 200451
erated at 15°C but not at 45°C. Strain KY-1 produced L - and D -lactic acid at a molar ratio of 63:37 in MRS medium, with an 88% yield from glucose. The partial 16S ribosomal DNA sequence of srain KY-1 (529bp, accession no.AB178487)isolated from run A was 99.0% homologous to that of L.plantarum ATCC8014 (accession no. M58827). The partial 1262bp sequence of L. lactis OML1-10 (accession no.AB178486) isolated from the stationary phase of run A was 99.5% homologous to strain IL1403 (accession no.L200142). Fifty isolates in run B were classified as L. plan-tarum (48% of total isolates), L. brevis (38%), L. fermen-tum ,L. paracasei ,L. delbrueckii or L. lactis according to the APICHL test (data not shown). In addition, the assimila-tion pattern of over 85% of isolates in run C was quite simi-lar to that of L. plantarum .
The amount of lactic acid that accumulated in autoclaved MKR varied according to the isolated bacterial species (Table 1). High concentrations of lactic acid accumulated (over 18.0g/l ) in most LAB except for L. brevis in run B.Very small amounts of lactic acid were detected in non-LAB. The LA selectivity was over 95% in most of the tested LAB. The most effective producer of lactic acid was L.plantarum , with over 95% selective accumulation of lactic acid in all batches. Nevertheless, the maximal amount of lac-tic acid produced by mixed cultures (small amount of non-autoclaved MKR) as opposed to monocultures was about 28g/l . Lactic acid accumulated by L. plantarum or in the mixed culture system was almost totally conglomerate (op-tical activity, 4–24%), whereas L. lactis produced optically active L -lactic acid (98–100%).
Design of group- and species-specific probes for lactic acid bacteria in FISH To directly analyze bacterial pop-ulations in a mixed culture system, we examined the appli-cability of FISH using 16S (23S) rDNA probes. We evalu-ated several probes for most LAB, as well as for L. plan-tarum . Figure 1 shows the relationship among the major genera and groups of lactic acid bacteria, as well as the
target organisms of the LAC722, and Lplan447 https://www.doczj.com/doc/3d6343155.html,C722 (systematic name [14], S-G-Lacb-0722-a-A-25)has one mismatch sequence compared with L. lactis , and four compared with the Gammaproteobacteria (Table 2).Although Lplan447 (S-S-Lplan-0447-a-A-18) was designed to selectively detect L. plantarum (Table 3), it could not differentiate this species from Lactobacillus pentosus .
TABLE 1.Dominant microorganisms isolated from various refuse sources
Run a Phase Microorganism
Population (%)
Lactate produced b
(g/l )
LA selectivity c
(%)
A
Early log d
K. pneumoniae 760.20.3E. aglomerans 78.764L. lactis
1018.192Stationary e
L. plantarum 6318.596L. lactis
3519.095B
Stationary e
L. plantarum 4822.598L. brevis 388.881L. paracasei 7––L. delbrueckii 3––L. fermentum
3
–
–
C
Stationary e
L. plantarum 8823.298
a
Refuse from various kitchens. Run A, Oita, Japan; run B, Kitakyushu, Japan; run C, Kuala Lumpur, Malaysia.b
To test the effect of the inoculant, the MKR was autoclaved and starter culture of each isolated strain (1ml) was added to the autoclaved MKR,which was then statically incubated at 37°C.c
Lactic acid selectivity. See Materials and Methods.d
Refuse sample after 6h incubation.e
Refuse sample after 48h incubation.
TABLE 2.Difference alignments for probes LAC722
16S rRNA sequence at the target sites of the probes are displayed
for representative reference organism.
TABLE 3.Difference alignments for probes Lplan447
16S rRNA sequence at the target sites of the probes are dis-played for representative reference organism.
SAKAI ET AL.J. B IOSCI . B IOENG .,
52Lactis445 (S-S-Lctis-0445-a-A-18) has been designated as a species-specific probe for L. lactis , but it did not work well under our experimental conditions (data not shown). Other than these, the ability of LGC354 probes (15) to detect gram-positive bacteria with a low DNA G +C content, par-ticularly those of the Bacillus species, has been tested.However, in vitro results showed that the probe LGC354B also stained L. plantarum and L. lactis with intense and weak signals, respectively. Thus, the application of the LGC354B probe to examinations of kitchen refuse is re-stricted due to the abundance of LAB. Use of the LGC354A and LGC354C probes is limited due to their specificity for LAB. In conclusion, LAC722 was the most effective gen-eral probe for most lactobacilli .
Effect of lytic enzyme digestion on FISH Since gram-positive bacteria including LAB are resistant to whole cell hybridization with FITC-labeled oligonucleotides (16),we investigated the sample preparation procedures for fixed cells using lytic enzymes and FITC-EUB338. As a result,incubation with 50mg/ml of lysozyme at 30°C for 20min significantly improved the permeability of all fixed lactic acid bacteria tested including L. plantarum and L. lactis . On the other hand, digestion with labiase, mutanolysin and pro-teinase K rather decreased the cell number even at 0.5, 0.5,and 0.1mg/m1, respectively. Fixed L. lactis was particularly susceptible to digestion, whereas 50mg/ml of lysozyme did not affect the morphology or numbers of this strain, or of K.pneumoniae ,Escherichia coli , or E. agglomerans .Conditions for selective hybridization and probe spec-ificity We confirmed probe specificity using bacteria iso-lated from MKR fermentation (Fig. 2). Analysis using FISH showed that Lplan447 specifically worked with L. plan-tarum at 46°C with no formamide. LAC722 stained all lacto-bacilli except L. lactis ,K. pneumoniae and E. agglomerans under high stringency [51°C, 30% formamide, wash with 250mM NaCl, designated as LAC722(H)], but under low stringency [46°C, 0% formamide, wash without NaCl, des-ignated as LAC722(L)] this probe also detected L. lactis .Figure 2 shows that these probes distinguished target and non-target organisms. We also confirmed the selective affin-ity of these probes by observing mixed cell-suspensions of each bacterium (data not shown). The optimal hybridiza-tion conditions for EUB338 and GAM42a were 46°C in the presence of 20% formamide, followed by a wash with buffer containing 180mM NaCl (6, 12).
Analysis of open MKR fermentation by FISH We openly fermented MKR in the absence of inoculant (Fig. 3),and in the presence of Japanese pickles as the inoculant (Fig. 4). Less lactic acid accumulated without inoculation (25g/l , Fig. 3) than with seed cultures from the pickles (46g/l , Fig. 4). The optical activities of accumulated lactic acid in each fermentation after 120h were 58% and 54%, re-spectively. We then enumerated the microbial population by conventional colony counting. The number of total bacteria in both MKR batches counted using anaerobic Standard Agar plates was initially 5.3±0.2log(cfu/ml). This rapidly
FIG.2.Differential staining of lactic acid bacteria by Lplan447 and LAC722. Cultured L. plantarum KY-1, L. rhamnosus KY-3, L. lactis OLM1-10, K. pneumoniae OEM 1-3, and E. agglomerans OEM 1-13 cells were stained with corresponding FITC-labeled oligonucleotide probes.Hybridization conditions: LAC722(H), 54°C, 35% formamide, 250mM washing salt; LAC722(L), 46°C, 0% formamide, 0% washing salt;GAM42a, 46°C, 20% formamide, 180mM washing salt; EUB338, 46°C, 20% formamide, 180
mM washing salt.
FISH ANALYSIS OF OPEN LACTIC FERMENTATIONS V OL . 98, 200453
increased after 12h (8.6±0.2log(cfu/ml)) and maintained a level of more than 108cfu/ml throughout the fermentation.Changes in the number of the LAB cells appeared similar to those of total bacteria. The numbers of the coliform bac-teria significantly differed over time. The number of coli-
form bacteria in the control MKR batch (Fig. 3) was ini-tially 4.3±0.2log(cfu/ml). This peaked at 12h (7.9±0.2log(cfu/ml)) and remained above 106cfu/ml for the next 24–72h. On the other hand, Fig. 4 shows that the number of coliform bacteria, 4.3±0.3log(cfu/ml) initially, slightly in-creased to 5.0±0.3log(cfu/ml) at 12h and decreased over the incubation period to slightly less than 4.0±0.2log(cfu/ml). Among 50 colonies randomly isolated from MKR fermented for 6h with the inoculant, 10 strains were gram-positive, catalase negative cocci that showed homo-lactic acid fermentation. The sugar fermentation of seven of them was similar to that of L. lactis , but those of other three isolates varied and thus the isolates could not be relegated to any type culture strains in the APICHL test. In contrast, all isolates from MKR fermented for 72h with the inoculant were gram-positive, catalase negative bacilli. The APICHL test tentatively classified 10 isolates as L. plantarum (5/10),L. pentosus (2/10), L. brevis (2/10), and L. collinoides (1/10).Samples collected at various times during fermentation were analyzed by FISH. Under optimal hybridization condi-tions, microorganisms were specifically visualized and de-tected with using the corresponding probes. Figure 5 shows representative micrographs of the fluorescent bacterial cells in the KR samples double-stained by rhodamine-EUB338 and FITC-LAC722(L). Solid and complex materials in kitchen
FIG.3.Microbial analysis of pH swing fermentation of standard kitchen refuse without inoculation under open conditions. Symbols:closed circles, lactic acid; open circles, acetic acid; closed squares, pH;open squares, viable cell numbers (log(cfu/ml)) of total bacteria;closed triangles, viable cell numbers of lactic acid bacteria; open trian-gles, viable cell numbers of coliform bacteria.
FIG.4.Microbial analysis of pH swing fermentation of standard kitchen refuse inoculated with pickles under open conditions. Symbols as in Fig. 3
.
SAKAI ET AL.J. B IOSCI . B IOENG .,
54refuse have weak self-fluorescence that hampered observa-tions of bacteria by confocal microscopy (data not shown).On the other hand, pictures taken by a chilled CCD color camera with a modified RGB color balance helped distin-guish the fluorescence of FITC-probes from the self-fluo-rescence of the refuse materials. The self-fluorescence of refuse materials was slightly yellowish in the fermented sample. In contrast to natural seed culture (Fig. 5A , B),most EUB338-positive cells in the fermentation with the in-oculate were also stained by LAC722(L) (Fig. 5C , D).
Table 4 shows the results of FISH counting. Since the lower limit of this protocol is about 4.7 log(cells/ml), the initial cell numbers could not be determined. The numbers of EUB-positive cells after 12h in the MKR sample with and without the inoculate were 9.0±0.1 and 9.2±0.0log(cells/ml), respectively, which were about 1.5-fold
higher than those determined by colony counting (see Figs.3 and 4). The number of LAC722-positive cells after 12h was as low as the lower detection limit, but increased rapidly in the fermented kitchen refuse without inoculation.In contrast, the number of GAM42a-positive cells reached 8.5±0.1 log(cells/ml) after 12h and rather decreased there-after. Profiles of the bacterial community, expressed in rela-tive data with the total number of bacteria being 100%, are presented in Fig. 6. The major group in the control MKR batch was the Gammaproteobacteria after a 12h incuba-tion, which accounted for about one-third of the microbial population, and decreased to 12.5% at 72h. The population of LAB722-positive cells that was initially detected at 24h,gradually increased and peaked at 72h (ca. 62.5%). A large proportion of the microbial community remained unidenti-fied, especially at the early state of fermentation (about 60% and 50% at 12 and 24h, respectively). In contrast,LAB722-positive strains predominated in the MKR batch inoculated with Japanese pickles. These bacteria accounted for 80% of the population at the start of fermentation, and remained above 75% throughout the study. At the end of fermentation, the ratio of lactic acid bacteria population was almost 100%. We then applied the species-specific probe, Lplan447 to detect L. plantarum (Fig. 7). Without the inoculate, Lplan447-positive cells, which would be L.plantarum or a neighbor species, were undetectable after 12h, but became dominant at the middle stage of the fermentation without inoculation. At the midpoint of fer-mentation with pickle inoculates, 78–96% of cells were also Lplan447-positive.
DISCUSSION
This report describes two important results. Firstly, the open fermentation of any kitchen refuse regardless of origin with pH swing control, leads to the reproducible succes-sion of the major bacterial species, L. plantarum . Secondly,quantitative FISH using group and species specific rDNA-targeted probes was applicable to the direct analysis of semi-solid open lactic fermentation. Each of these results is discussed below.
Several preparations tested, including kitchen refuse from various sources, were fermented with reproducible lactic acid accumulation and the selective proliferation of specific LAB species. We showed that intermittent pH adjustment helps the stable accumulation of lactic acid during the non-sterilized fermentation of kitchen refuse (4).Due to the
TABLE 4.FISH analysis of open fermentation of kitchen refuse Cell count at each incubation time (log(count/ml))12h
24h 36h 48h 60h 72h With no inoculant EUB3389.0±0.19.3±0.19.4±0.29.4±0.29.4±0.19.3±0.2LAC722(L)ND 8.9±0.28.9±0.39.3±0.29.4±0.29.2±0.1GAM42a
8.5±0.18.4±0.28.5±0.18.5±0.18.2±0.17.8±0.1With Japanese pickles EUB3389.2±0.09.3±0.19.4±0.19.3±0.29.2±0.19.3±0.1LAC722(L)9.1±0.19.2±0.19.3±0.19.2±0.19.4±0.19.5±0.1GAM42a
ND
ND
ND
ND
ND
ND
ND, Not detected (<4.7).
FIG.6.Group-specific FISH analysis of pH swing fermentation of standard kitchen refuse without (A) or with (B) seed culture inocula-tion. Relative cell numbers of GAM42a- (gray zone) and LAC722(L)-positive (black zone) were shown in 100% EUB338-positive cells.White zone, EUB338-positive, but GAM42a- and LAC722(L)-nega-
tive cells.
FISH ANALYSIS OF OPEN LACTIC FERMENTATIONS V OL . 98, 200455
open conditions, several microorganisms might contribute to this transformation, although the LAB community would undoubtedly play a key role. Conventional laboratory meth-ods showed that the predominant groups of bacteria differed between the early and the late phases of refuse fermentation.The Gammaproteobacteria predominated at the beginning of one fermentation experiment, whereas LAB became the major species during the stationary phase, among which over 60% of isolates were L. plantarum (Table 1). In addi-tion, this species was also isolated as a major LAB in sev-eral experimental runs using refuse from different origins as well as in another run using Japanese pickles inoculant. In addition, we further investigated the abilities of each bac-terial isolate to produce lactic acid, and found that less lactic acid was produced from each isolate than from natural seed cultures (mixed cultures of microorganisms). This finding emphasizes the importance of the mixed culture system.Most LAB are fastidious and usually cannot utilize most polysaccharides (16). Therefore, other microbial groups might be involved in a cooperative interaction. For exam-ple, they might hydrolyze complex substrates and thus pro-vide an appropriate environment for LAB colonization at the late stage of fermentation.
Next, the validity of the FISH method that enabled direct observation and quantification of LAB groups and species in mixed turbid samples containing various types of organic refuse materials and microbial species is discussed. We selected the probes, EUB338, GAM42a, LAC722 and Lplan447, based on preliminary studies and achieved the predicted results (Table 4, Figs. 6 and 7). However, few studies using FISH have been reported (17), not only be-cause ‘LAB’ contain a variety of genera, but also because they are members of gram-positive and low G +C bacteria that possess thick cell walls that resist permeation with fluo-rescent probes (13). We tested the ability of several lytic
enzymes to enhance permeability. A high concentration of lysozyme improved the permeability of LAB and main-tained the morphology of gram-negative bacilli. We also confirmed that the FISH protocol was applicable to other high G +C gram-positive organisms such as Bacillus species (data not shown).We examined several candidate L AB-group specific probes for use in whole cell hybridization and found LAC722, which was originally described by Sighr et al . as an RNA dot-blot hybridization probe (18).LAC722was an effective probe for many lactobacilli , but it has only one mismatch sequence against lactococci . This was rather useful for the fermentation system. For wide-range LAB detection, the probe under low stringent hybridization con-ditions can also detect lactococci , and most lactobacilli or lactococci can be differentially detected by varying the strin-gency of hybridization and washing. Lplan447 was designed as a species-specific probe for L. plantarum , but the same complimentary sequence was also found in L. pentosus ,because these are quite close both in their culture properties and phylogenetic position. Their 16S rDNA sequences are over 99% homologous and they are genomically related at the level of 50–60% (19).
After optimization, we quantified each of the target bac-teria. The presence of a large amount of organic solid mate-rial in the fermented samples hampered the estimation of the total numbers of microorganisms by phase-contrast mi-croscopy, and by chromosomal staining using DAPI or ac-ridine orange. We then used EUB338 probes to estimate the total bacteria. The EUB338 probe for detecting all bacteria has recently been improved (20), since the probe does not cover Planctomycetales and Verrucomicrobia . However, these bacteria inhabit fresh and saline waters and they have not been isolated from (fermented) kitchen refuse. As a result,the number of EUB338-positive cells was about 1.5-fold higher than that determined by conventional culture. The numbers of cultural LAB and LAC722-positive microorga-nisms as well as coliform bacteria and GAM42a also slight-ly varied (Fig. 3 and Table 4).The chilled CCD color cam-era with a modified RGB color balance helped to discrimi-nate the probe from the self-fluorescence of refuse mate-rials. One problem that remains to be solved is its high de-tection limit (4.7 log(cells/ml)). Samples cannot be easily concentrated because of contamination by solid materials.Regardless, results from FISH analysis closely agreed with those from the culture method. Such specific and direct de-tection of LAB in semi-solid materials by molecular biolog-ical methods is useful, since this group is key to domestic food fermentation but the dominant microorganisms are sometimes multiple and indeterminate (21).
In the open fermentation of kitchen refuse by natural mi-croorganisms, we confirmed the presence of Gammaproteo-bacteria at the early stage of the fermentation, and the over-whelming growth of LAB, especially of L. plantarum . Yet,over 60% of the total microbial population remained un-identified at the start of the fermentation. These groups are presently being investigated using a series of rRNA-targeted oligonucleotide probes. In addition to the open fermentation of kitchen refuse by naturally occurring microorganisms,we fermented it after inoculation with a mixed seed culture (Japanese pickles), that might be enriched with many LAB
FIG.7.Genus-specific FISH analysis of pH swing fermentation of standard kitchen refuse without (A) or with (B) seed culture inocula-tion. Relative numbers of cells stained with LAC722(L) (shaded bars)and Lplan447 (white bars) against EUB338 (black bars) at various in-
cubation times.
SAKAI ET AL.J. B IOSCI. B IOENG., 56
strains. Under these conditions, a similar LAB consortium formed more rapidly, with more lactic acid production. The dominance of LAC722 and Lplan477 could be traced
throughout the fermentation (Figs. 6 and 7).
Among LAB, L. plantarum has a relatively lower nutri-tional requirement, a wider spectrum of sugar assimilation,
oxygen tolerance and range of growth pH (4). In conclu-sion, species-specific FISH confirmed that L. plantarum be-comes dominant during fermentation with or without an in-
oculant. This resulted in the selective accumulation of opti-cally inactive lactic acid. Although poly lactic acid synthe-sized from optically inactive lactic acid is of interest as a
green plastic, it cannot be used in plastic engineering, be-cause it has lower crystallinity than optically active poly-L-lactic acid, indicating that it has decreased tensile strength. We are currently attempting to produce optically active lac-tic acid under non-sterile conditions, so further studies using FISH are required to regulate economical open lactic acid
fermentation.
ACKNOWLEDGMENTS
This study was supported by Special Coordination Funds, and by Grants-in-Aid for Scientific Research (no. 14360202) from the Ministry of Education, Culture, Sports, Science and Technology of the Japanese Government.
REFERENCES
1.Lunt, J.: Large-scale production, properties and commercial
application of polylactic acid polymers. Polym. Degrad. Stab., 59, 145–152 (1998).
2.Gerngross, T.U. and Slater, S.C.:How green are green
plastics? Sci. Am., Aug., 24–29 (2000).
3.Sakai, K., Taniguchi, M., Miura, S., Ohara, H., Matsumoto,
T., and Shirai, Y.:Making plastics from garbage: a novel process for poly-L-lactate production from municipal food waste. J. Ind. Ecol., 7, 63–73 (2003).
4.Sakai, K., Murata, Y., Yamazumi, H., Tau, Y., Mori, M.,
Moriguchi, M., and Shirai, Y.:Selective proliferation of lactic acid bacteria and accumulation of lactic acid during open fermentation of kitchen refuse with intermittent pH ad-justment. Food Sci. Technol. Res., 6, 140–145 (2000).
5.DeLong, E.F., Wickham, G.S., and Pace, N.R.: Phyloge-
netic stains: ribosomal RNA-based probes for the identifica-tion of single microbial cells. Science, 243, 1360–1363 (1989).
6.Amann, R.I., Krumholz, L., and Stahl, D.A.: Fluorescent-
oligonucleotide probing of whole cells for determinative phylo-genetic, and environmental studies in microbiology. J. Bac-teriol., 172, 762–770 (1990).
7.Maidak, B.L., Cole, J.R., Lilburn, T.G., Parker, C.T.,
Saxman, P.R., Farris, R.J., Garrity, G.M., Olsen, G.J., Schmidt, T.M., and Tiedje, J.M.: The RDP-II (Ribosomal Database Project). Nucleic Acids Res., 29, 173–174 (2001).
8.Beimfohr, C., Krause, A., Amann, R., Ludwig, W., and
Schleifer, K.H.:In situ identification of lactococci, entero-
cocci, and streptococci. Syst. Appl. Microbiol., 16, 450–456 (1993).
9.Sankai, T.:Development of recycling system of kitchen
refuse using disposer, p.5–15. In Workshop on Kitchen Refuse Recycling System (ed.), Treatment of kitchen refuse using disposer. The Building Center of Japan, Tokyo (1999).
(in Japanese).
10.Sakai, K., Kudoh, E., Wakayama, M., and Moriguchi, M.:
Molecular analysis of microbial community in an activated sludge enriched by inorganic-nitrite medium. Microbes Envi-ron., 15, 103–112 (2000).
11.Thompson, J.D., Higgins, D.G., and Gibson, T.J.:
CLUSTAL W: improving the sensitivity of progressive multi-ple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res., 22, 4673–4680 (1994).
12.Manz, W., Amann, R., Ludwig, W., Wagner, M., and
Schleifer, K.H.: Phylogenetic oligonucleotide probes for the major subclass of proteobacteria: problems and solutions. Syst.
Appl. Microbiol., 15, 593–600 (1992).
13.Amann, R.:In situ identification of micro-organisms by
whole cell hybridization with rRNA-targeted nucleic acid probes, 3.3.6, p.1–15. In Akkermans, A.D.L., Elsas, J.D., and Bruijn, F.J. (ed.), Molecular microbial ecology manual.
Kluwer Academic Publisher, Dordrecht (1995).
14.Alm, E., Oerther, D.B., Lasten, N., Stahl, D.A., and
Raskin, L.: The oligonucleotide probe database. Appl. Envi-ron. Microbiol., 10, 3557–3559 (1996).
15.Meier, H., Amann, R., Ludwig, W., and Schleifer, K.H.:
Specific oligonucleotide probes for in situ detection of a major group of gram-positive bacteria with low DNA G+C content. Syst. Appl. Microbiol., 22, 186–196 (1999).
16.Hammes, W.P., Weiss, N., and Holzapfel, W.: The genera
Lactobacillus and Carnobacterium, p.1535–1594. In Balows,
A., Truper, H.G., Dworkin, M., Harder, W., and Schleifer,
K.H. (ed.), The Prokaryotes, 2nd ed., vol.2. Springer-V erlag, New York (1992).
17.Lipsky, A., Friedrich, U., and Altendorf, K.: Application of
rRNA-targeted oligonucleotide probes in biotechnology.
Appl. Microbiol. Biotechnol., 56, 40–57 (2001).
18.Sghir, A., Antonopoulos, D., and Mackie, R.I.: Design and
evaluation of a Lactobacillus group-specific ribosomal RNA-targeted hybridization probe and its application to the study of intestinal microecology in pigs. Syst. Appl. Microbiol., 21, 291–296 (1998).
19.Kandler, O. and Weiss, N.: Regular, nonsporing gram-posi-
tive rods, p.1208–1234. In Sneath, P.H.A., Mair, N.S., Sharpe, M.E., and Holt, J.G. (ed.), Bergey’s manual of sys-tematic bacteriology, vol.2. Williams & Wilkins, Baltimore (1989).
20.Daims, H., Bruhl, A., Amann, R., Schleifer, K.H., and
Wagner, M.: The domain-specific probe EUB338 is insuffi-cient for the detection of all Bacteria: development and evalu-ation of a more comprehensive probe set. Syst. Appl. Micro-biol., 22, 434–444 (1999).
21.Ampe, F., Omar, N., and Guyot, J-P.: Culture-independent
quantification of physiologically-active microbial groups in fermented food using rRNA-targeted oligonucleotide probes: application to pozol, a Mexican lactic acid fermented maize dough. J. Appl. Microbiol., 87, 131–140 (1999).