Gastrointestinal microbial community shifts ... - Wiley Online Library

FEMS Microbiology Ecology 43 (2003) 133^140

www.fems-microbiology.org

Gastrointestinal microbial community shifts observed following oral administration of a Lactobacillus fermentum strain to mice Laura Plant, Chi Lam, Patricia L. Conway , Katie O’Riordan School of Microbiology and Immunology, The University of New South Wales, Sydney, NSW 2052, Australia Received 16 July 2002; received in revised form 20 August 2002; accepted 23 August 2002 First published online 17 September 2002

Abstract The indigenous gastrointestinal microbiota acts as an integral defense against the colonisation of orally introduced microbes. Whilst this can be important in host protection, some introduced species, including lactobacilli, can have a positive impact on existing microbial communities. The interaction of a candidate probiotic strain of Lactobacillus fermentum within the gastrointestinal tract was monitored in a mouse model and its effect on the indigenous microbiota observed. L. fermentum KLD was administered via oro-gastric doses to mice with both a specific pathogen-free (SPF) and an ampicillin-depleted gut microbiota, containing no detectable lactobacilli. Its persistence was monitored by detection in faecal homogenates using culturing methods and polymerase chain reaction with L. fermentum specific primers. Microbial population shifts were observed using denaturing gradient gel electrophoresis (DGGE). L. fermentum KLD was detected within the gastrointestinal tract of SPF mice for up to 36 h, and for greater than 11 days in the ampicillin-treated mice. The administration resulted in substantial changes within the host Lactobacillus levels, determined by DGGE of 16S rDNA from faecal samples. Denaturing gradient profiles, from faecal samples collected at a range of pre- and post-dose intervals of groups of 10 SPF mice, indicated that several other constituents of the gastrointestinal community also fluctuated following dosing. These included Bifidobacterium and Eubacterium, which increased following KLD administration. The indigenous microbiota affected the persistence of L. fermentum KLD and in SPF mice the administration of this strain induced significant shifts in the indigenous microbial community. 6 2002 Published by Elsevier Science B.V. on behalf of the Federation of European Microbiological Societies. Keywords : Gastrointestinal microbial community ; 16S rDNA ; DGGE ; Probiotic

1. Introduction In the past 30 years, with the advent of animals with an absent (germ-free) or manipulated microbiota, the importance of the indigenous microbiota has been substantiated [1^3]. The intestinal tract harbours a complex microbial community that plays a key role in nutrition and health [4]. Certain indigenous organisms are reported to have important e¡ects within the gastrointestinal tract, including the capacity to increase detoxicity and immunostimulation [5], and provide resistance to infectious agents. The indigenous microbiota has also been reported to inhibit the colonisation of introduced organisms [6,7] by mechanisms including space occupation [8], competition for sub-

* Corresponding author. Tel. : +61 (2) 9385 1593; Fax : +61 (2) 9385 1591. E-mail address : [email protected] (P.L. Conway).

strates [9] and receptors [10] at mucosal surfaces, and secretion of bacteriocins [11] and other regulatory factors such as short-chain fatty acids[12]. It has been estimated that between 60 and 80% of bacteria in environmental samples cannot be cultured [13,14]. The use of evolutionary marker molecules, such as rRNA genes and polymerase chain reaction (PCR) techniques, has greatly helped microbial ecologists in surveying species diversity [15]. One of the pioneering phylogenetic inventories of the human faecal microbiota described 113 di¡erent species from 20 faecal samples, and more recent studies employing molecular methodologies have demonstrated that biodiversity within the gut is even greater than anticipated [16,17]. Although a number of studies have utilised PCR^denaturing gradient gel electrophoresis (DGGE)/ TGGE techniques to study faecal communities [17^20], only recently have these methodologies been exploited for monitoring shifts that occur following administration of allochthonous strains. Combining 16S rDNA analysis with the electrophoresis techniques of DGGE/TGGE has

0168-6496 / 02 / $22.00 6 2002 Published by Elsevier Science B.V. on behalf of the Federation of European Microbiological Societies. PII : S 0 1 6 8 - 6 4 9 6 ( 0 2 ) 0 0 3 7 4 - 4

FEMSEC 1428 5-2-03

134

L. Plant et al. / FEMS Microbiology Ecology 43 (2003) 133^140

been increasingly employed in the study of complex aquatic and terrestrial ecosystems [21,22] and has been demonstrated to be a powerful tool for monitoring community shifts following environmental changes [23,24]. Lactobacillus fermentum KLD has previously been used as both a prophylactic and therapeutic agent in the treatment of gastrointestinal disturbances [25,26] and is an interesting candidate probiotic strain. Using the murine system as model for the complex bacterial community of the gastrointestinal tract, we sought to assess the impact of administration of this strain and to examine the biodiversity within faecal samples from L. fermentum-treated mice using conventional techniques, PCR and DGGE.

2. Materials and methods 2.1. Bacterial strains and growth conditions L. fermentum KLD was provided by Lantma«nnen, Sweden. The following Lactobacillus strains were obtained from various culture collections: L. fermentum DSMZ 20391, 20052 (DSMZ; Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Germany), L. fermentum LMG 8896 (LMG; Laboratorium voor Microbiologie, Universiteit Gent, Belgium), L. fermentum BMF 30025, 30026 (BMF; Bundesanstalt fu«r Milchforschung, Kiel, Germany), Lactobacillus reuteri DSMZ 20016, Lactobacillus oris DSMZ 4864, Lactobacillus casei ATCC 393 (ATCC, American Type Culture Collection, Rockville, USA) and Lactobacillus isolates from mouse faeces, Lactobacillus FII 532600, 532700, 532800, 532900, 533000, 533100, 533200, 532300 (FII; Food Industry Innovation Collection, The University of New South Wales (UNSW), Sydney, Australia). Primary cultures of all strains were grown anaerobically at 37‡C in Mann Rogosa Sharpe (MRS) broth (Difco), harvested by centrifugation at 5000Ug, washed with phosphate-bu¡ered saline (PBS) (0.1 M, pH 7.2) and adjusted to the optical density (V600nm ) required to yield appropriate concentrations. 2.2. Animals Six-week old female SPF C57BL/6 mice weighing 18^22 g were obtained from the Biological Resource Centre, Sydney, Australia. Ethical approval for animal experiments was obtained from the UNSW Animal Ethics Committee. SPF mice (n = 10) were housed together for 10 days prior to initiation of the experiment. Four mice with a depleted Lactobacillus community were prepared by breeding from 6-week old SPF C57BL/6 parent mice, which were given ampicillin in drinking water at 4.0 g l31 for 1^2 weeks until no lactobacilli were detected in faecal homogenates. All mice with a depleted Lactobacillus community were housed in ¢lter-top cages, and given sterile food and water. The microbial pro¢les of these were monitored

for the duration of the experiment, and the experiments conducted when progeny were 6 weeks old. 2.3. Colonisation of the murine gastrointestinal tract by L. fermentum KLD L. fermentum KLD was administered to SPF mice and freshly void faecal pellets collected at the following sampling times: two samples prior to the ¢rst dose (pre-dose samples), a sample 2 h prior to the second dose (46 h after ¢rst dose), and 4, 8, 11, 24, 36, 54, 78, 102, 126 and 148 h after the second dose. Two doses were administered to enhance the persistence of the strain within the gastrointestinal tract. Each dose consisted of approximately 1U108 CFU mouse31 . A non-dosed control group was monitored for community changes resulting from environmental factors. Similarly, viable lactobacilli detected in the faeces of ampicillin-treated mice were monitored in freshly void faeces collected at 24 h and 2, 3, 4, 5, 6, 10, 11 and 14 days following a single oro-gastric dose (5U107 CFU mouse31 ). For enumeration, samples of the faecal pellets (10 mg) were homogenised in PBS, diluted, and enumerated on Rogosa agar after 72 h of anaerobic incubation. Faecal samples collected from SPF mice were used in PCR and DGGE analysis. 2.4. Determination of 16S rRNA gene sequence from L. fermentum KLD An aliquot (10 Wl) of crude cell lysate, prepared from L. fermentum KLD through cell disruption with 0.1 mm silica beads (Biospec, USA), was used as template in PCR ampli¢cation employing 25 pmol each of the universal eubacterial 16S rRNA primers 27f and 1492r [27]. DNA ampli¢cation was performed using 0.2 mmol deoxyribonucleotide mix, 0.7 mM MgCl2 and 1 U of Taq polymerase (all reagents supplied by Fisher Biotech Int., Australia), in a total volume of 50 Wl in a Hybaid express cycler programmed for an initial denaturation of 95‡C for 3 min followed by 25 cycles of (94‡CU30 s)+(50‡CU30 s)+ (72‡CU1 min) and including a ¢nal extension of 72‡C for 7 min. The PCR products obtained were puri¢ed using the QIAquick kit (Qiagen, NSW, Australia). The product (1540 bp) was sequenced in its entirety on both strands, designing primers from existing sequences when necessary and checking ambiguous nucleotides until a de¢nitive sequence was obtained. 2.5. Extraction and puri¢cation of DNA from faecal samples Samples (10 mg) of faecal material were resuspended in 0.3 ml lysis bu¡er (200 mmol Tris, 50 mmol EDTA, 2 mmol sodium citrate, 10 mmol CaCl2 and 200 mmol NaCl, pH 8.0) containing 3 Wg lysozyme, and approximately 0.3 g acid-washed silica beads (0.1 mm, Biospec,

FEMSEC 1428 5-2-03


USA) was added. The mixture was incubated for 40 min at 37‡C. To initiate lysis, 20% (w/v) sodium dodecyl sulfate (SDS; 10 Wl) and proteinase K (1.2 mg) were added and the suspension incubated at 50‡C for 30 min. Cetyl tri-methyl ammonium bromide (CTAB) was incorporated into the reaction (80 Wl of a 10% (w/v) CTAB solution in 0.7 M NaCl) and incubation was continued at 65‡C for 10 min. To complete lysis, 200 Wl SDS (20% (w/v)) and 400 Wl of phenol:chloroform:isoamyl alcohol (24:24:1) was added and the mixture homogenised in a mini-bead beater for 30 s using the highest setting. Nucleic acid was extracted from these preparations as previously described [28]. The concentration of the nucleic acid was determined by measuring the absorbance at 260 and 320 nm using a DU 640 spectrophotometer (Beckman, USA). When required for PCR and DGGE, crude lysates were prepared from reference Lactobacillus cultures by mechanical disruption of the cells with silica (0.1 mm) beads using a mini-bead beater. 2.6. PCR conditions PCR primers 341f-GC (GCclamp CCT ACG GGA GGC AGC AG) and 534r (ATT ACC GCG GCT GCT GG) were used to amplify the variable V3 region of eubacterial 16S rDNA for DGGE analysis [29]. A 40-nucleotide GC clamp (CGC CCG CCG CGC GCG GCG GGC GGG GCG GGG GCA CGG GGG G) [30] was incorporated onto the 5P end of the 341f primer to ensure the DNA remained partially double-stranded during DGGE. PCR reactions were prepared in a total volume of 50 Wl using reagents as described above. The ¢nal DNA concentration was adjusted to 30 ng and 0.3^0.4 Wmol of each primer was added. Ampli¢cation of DNA was performed in a Hybaid express thermal cycler programmed for an initial denaturation of 94‡C for 5 min followed by 27 cycles (94‡C U5 min)+(62‡CU30 s)+(72‡CU1 min) and a ¢nal extension of 72‡C for 7 min. PCR products were analysed on 2% agarose gels, stained with ethidium bromide, visualised and photographed using a Bio-Rad, Gel Doc imaging system. 2.7. DGGE analysis DGGE was performed with the D code system (BioRad) using the protocol outlined in the manufacturer’s manual. The only modi¢cation involved the addition of 20% glycerol to enhance settlement of the gel. PCR products (42 Wl of ampli¢ed products from faecal samples or 10 Wl of products ampli¢ed from pure culture extracts) were separated by DGGE with a linear gradient of urea and formamide (20^60% (v/v)) in an 8% acrylamide gel at 200 V and 60‡C for 5 h. Gels were stained with ethidium bromide and analysed under UV illumination using the Gel Doc system. Gel slices containing the DNA bands of interest were excised and DNA recovered using the ‘crush and soak’ technique described by Sambrook [31].

135

DNA concentration was determined using a spectrophotometer. Depending on the concentration, this DNA was used directly in sequencing reactions or was subjected to a second ampli¢cation using previous conditions and primers (341f and 534r) if the DNA concentration was low ( 6 30 ng). Re-ampli¢ed products were column-puri¢ed prior to sequencing. 2.8. DNA sequence analysis DNA sequencing reactions were performed using an ABI prism dye terminator cycle sequencing ready reaction kit with Amplitaq DNA polymerase (Perkin Elmer). Sequence determination was performed using an Applied Biosystems 377 automated DNA sequencer (Applied Biosystems, Foster City, CA, USA) employing synthetic oligonucleotides (Life Technologies, NSW, Australia). Sequences were assembled using AutoAssembler package (Applied Biosystems). Database searches [32] were performed using the on-line server maintained at the National Center for Biotechnology Information (NCBI), Bethesda, MD, USA. 2.9. Statistical analysis Signi¢cant di¡erences in the Lactobacillus numbers and the total amount of DNA extracted from faecal samples were obtained using analysis of variance (ANOVA) and Student’s t-test. Duncan’s multiple range test (95%) was used to determine di¡erences between the amount of nucleic acid extracted and either sampling time or individual mouse.

3. Results 3.1. E⁄ciency of nucleic acid extraction, speci¢city and sensitivity of L. fermentum KLD strain-speci¢c primer The nucleotide sequence of the KLD 16S rRNA gene determined in this study was used to design a L. fermentum KLD strain-speci¢c primer P1 (5P-AAGTCGAGCGGTTGGCC-3P). This primer was used in conjunction with the universal primer 519r [(5P-GWATTACCGCGGCKGC-3P) with K = G:T (1:1), W = A:T (1:1)][27] and was assessed for speci¢city against a large number of strains including six L. fermentum strains, L. reuteri, L. oris, L. casei, eight unspeciated Lactobacillus isolates and 17 other bacterial strains. The primers could amplify products from three closely related L. fermentum strains (DSMZ 20391, DSMZ 20052, LMG 8896) in addition to the KLD strain, which it could detect at levels of 1 ng or greater. No product was obtained from any of the other lactobacilli or bacterial strains tested. The extraction was most e⁄cient at levels of 104 cells ml31 , at which approximately 7 ng of nucleic acid could be obtained from a

FEMSEC 1428 5-2-03

136


Table 1 L. fermentum KLD detected in faecal samples of SPF C57BL/6 mice using conventional culturing methods and PCR with L. fermentum-speci¢c primers (P1 and 519r) following a duplicate oro-gastric dose of 1U108 viable organisms per mouse (n = 10 mice) Time following administration (h)

Detection using viable countsa

Number of positive samples detected using PCR

Pre-treatment 4 8 11 24 36 54 78 102 126 148

N.D. 3.33 V 3.29 6.05 V 5.68 5.54 V 4.83 5.42 V 5.49 N.D.c N.D. N.D. N.D. N.D. N.D.

0 6 9 8 5 1 0 0 0 0 0

(5)b (7) (10) (5)

a

Results expressed as log10 CFU g31 (wet weight) of faeces (mean V S.E.M.). Numbers in parentheses represent the number of positive samples. c N.D. = not detected. b

single cell. This became less e⁄cient with each exponential increase in initial cell numbers (data not shown). As L. fermentum KLD and the genetically related L. fermentum strains used in this study are not natural constituents of the mouse gastrointestinal Lactobacillus community, the primer pair P1^519r was su⁄ciently sensitive for use in the mouse model. PCR amplicates were only observed from faecal samples from mice that had been dosed, or faecal samples from non-dosed mice to which L. fermentum KLD had been intentionally added post-collection to serve as PCR controls.

tract of Lactobacillus-depleted C57BL/6 mice by L. fermentum KLD was initially variable, but was maintained in high numbers until 10 days post-dose and was last detected at 11 days post-dose. At 14 days following orogastric administration L. fermentum KLD could not be detected in the faeces. Although it is not known which other populations of the gut microbiota were a¡ected by the ampicillin treatment, the absence of lactobacilli in this system may be the important factor promoting colonisation of the KLD strain. No other strain of Lactobacillus was detected in the faeces for up to 2 months (data not shown), at which time the experiment was terminated.

3.2. In vivo colonisation of SPF and ampicillin-treated mice Culturing and PCR using the designed primers were employed to detect L. fermentum KLD in the faeces of SPThisF mice following oro-gastric administration. This organism, which produced large, shiny colonies that are easily distinguishable from the indigenous lactobacilli by their described appearance and the production of a viscous polysaccharide, could not be detected in the pre-dose samples using either technique. It was detected in the postdose faecal samples for 24 h using culturing, and 36 h by PCR with L. fermentum-speci¢c primers (Table 1). High numbers of L. fermentum KLD were detected in the faeces at between 4 and 8 h post-dose, and remained high until 24 h following dosage. The analysis of faecal material using L. fermentumspeci¢c primers revealed that this organism could not be detected in all animals con¢rmed to contain this organism using conventional techniques. However, using PCR, L. fermentum KLD could be detected at 36 h following oro-gastric administration, albeit in only one sample, compared to ¢nal detection at 24 h using culturing techniques. Following administration of L. fermentum KLD to mice containing a Lactobacillus-depleted microbiota, the presence of this organism was monitored on Rogosa agar as seen in Table 2. The detection within the gastrointestinal

3.3. Enumeration of Lactobacillus species from faecal samples and DGGE resolution of PCR amplicates from lactobacilli Colonies representing all of the ¢ve di¡erent morphotypes of lactobacilli from SPF faecal material were chosen for further analysis. Morphotypes 1 and 2 were isolated

Table 2 Presence of L. fermentum KLD in the faeces of Lactobacillus-free C57BL/6 mice following a single oro-gastric dose of 5U107 viable organisms per mouse (n = 4 mice) Time following administration (days)

Detection using viable counts (mean V S.E.M.)a

Pre-treatment 1 3 4 5 6 10 11 14

N.D. 3.03 V 2.90 5.99 V 5.96 6.34 V 6.10 6.00 V 5.82 6.39 V 6.21 5.51 V 4.75 4.38 V 4.20 N.D.

N.D. = not detected. a Results expressed as log10 CFU g31 (wet weight) of faeces (mean V S.E.M.).

FEMSEC 1428 5-2-03


from pre-dose samples, whereas morphotypes 3^5 were selected from post-dose groups. Isolates of morphotype 5 were identi¢ed as L. fermentum KLD by colony morphology and PCR. Morphotypes 1^4 did not yield an amplicate with the KLD-speci¢c primers but were shown using API identi¢cation to be L. fermentum (morphotypes 1 and 3) and L. crispatus (morphotypes 2 and 4), respectively. PCR amplicates obtained using primers 341f and 534r and a range of L. fermentum strains (KLD, DSMZ 20391, DSMZ 20052, LMG 8896), L. oris, L. reuteri, representatives of morphotypes 1^5, and nucleic acid extracted from faecal material as templates, were analysed by DGGE. Under the conditions applied, DGGE did not resolve the bands of di¡erent Lactobacillus strains present in the faecal samples. As observed in Fig. 1, bands for all reference L. fermentum strains and for the faecal isolates migrated to the same position as the L. fermentum KLD strain. Only the more divergent species L. reuteri and L. oris could be distinguished. In further DGGE pro¢les of PCR amplicates, the band corresponding to all Rogosaculturable host Lactobacillus species and L. fermentum KLD was referred to as the ‘lactobacilli’ band. 3.4. Detection limit of DGGE analysis The detection sensitivity of DGGE was assessed by comparative analysis of the total host Lactobacillus community determined by plate counting with the intensity of the ‘lactobacilli’ band on the DGGE pro¢le of PCR amplicates. Bands 8^13 (Fig. 2a) corresponded to a faecal lactobacilli community of between 106 and 107 CFU mg31 (Fig. 2b). As bands 3^6 (Fig. 2a) represented both the host Lactobacillus community and the administered L. fermentum KLD strain, the intensity of these bands could not be correlated directly to the host Lactobacillus numbers in Fig. 2b. However, enumeration of L. fermentum KLD demonstrated the presence of 1U102 CFU mg31 of faeces at 24 h in samples from mouse 1. Thus, band 7, that was faint but detectable, corresponded to a total viable lactobacilli community (indigenous species and

Fig. 1. Migration of 16S V3 rDNA amplicons from reference Lactobacillus strains and faecal isolates in an 8% polyacrylamide denaturing gradient (20^60%) gel. Lanes: 1, L. reuteri DSMZ 20016 ; 2, L. oris DSMZ 4864 ; 3, L. fermentum DSMZ 20391 ; 4, L. fermentum DSMZ 20052; 5, L. fermentum LMG 8896 ; 6, faecal isolate group 1; 7, faecal isolate group 2; 8, faecal isolate group 3; 9, faecal isolate group 4; 10, faecal isolate group 5; 11, L. fermentum KLD; 12, faecal sample.

137

Fig. 2. Relationship between intensity of ‘lactobacilli’ band in the DGGE pro¢les generated from resolution of 16S rDNA ampli¢ed from faecal samples and total numbers of viable Lactobacillus spp. determined by enumeration of colony-forming units on Rogosa agar. Faecal samples utilised were collected from mouse 1 following dosing with L. fermentum KLD. a: Changes in intensity of DGGE ‘lactobacilli’ band corresponding to ampli¢ed 16S V3 rDNA regions from faecal samples collected over sampling times. Lanes 1 and 2, pre-dose samples ; lane 3, sample 46 h after ¢rst dose; lanes 4^13, samples taken at 4h (lane 4), 8 h (lane 5), 11 h (lane 6), 24 h (lane 7), 36 h (lane 8), 54 h (lane 9), 78 h (lane 10), 102 h (lane 11), 126 h (lane 12), 148 h (lane 13) after second dose; lane 14, 16S rDNA ampli¢ed from a pure culture of L. fermentum KLD. b: Changes in total host Lactobacillus numbers (CFU mg31 wet weight) in faeces over sampling time. Numbers on the graph correlate to the lanes in panel a and represent samples taken 4 h (4), 8 h (5), 11 h (6), 24 h (7), 36 h (8), 54 h (9), 78 h (10), 102 h (11), 126 h (12), 148 h (13) after the second dose.

L. fermentum KLD) of less than 104 CFU mg31 . Hence, it was reasoned that DGGE pro¢le bands could be visualised in these studies when the corresponding species number in faeces was greater than 104 CFU mg31 . 3.5. Phylogenetic analysis of dominant bands in DGGE pro¢les In order to unravel changes that occurred within the predominant groups in the murine gut, dominant bands from DGGE pro¢les were sequenced and compared to GenBank database entries. Pro¢les from all 10 mice were observed to be extremely similar with bands of equal intensity appearing at the same positions. Mice in the nondosed control group demonstrated unchanged pro¢les over the monitoring period (data not shown). Each mouse in the dosed group also acted as its own control in that it was assessed for random changes prior to dosage and all pre-dose pro¢les did not demonstrate any major variations. The most intense bands were sequenced from four di¡erent mice. The pro¢le obtained from mouse 8 is shown in Fig. 3 and the bands identi¢ed by sequence determination are indicated, and identi¢cation by compari-

FEMSEC 1428 5-2-03

138


Fig. 3. DGGE pro¢les generated using 16S rDNA ampli¢ed from faecal samples (mouse 8) collected prior to and following dosing with L. fermentum KLD. Lanes 1 and 2, pre-dose samples; lane 3, sample 46 h after ¢rst dose; lanes 4^13, samples taken at 4h (lane 4), 8 h (lane 5), 11 h (lane 6), 24 h (lane 7), 36 h (lane 8), 54 h (lane 9), 78 h (lane 10), 102 h (lane 11), 126 h (lane 12), 148 h (lane 13) after second dose; lane 14, 16S rDNA ampli¢ed from a pure culture of L. fermentum KLD. Bands identi¢ed by sequencing are indicated.

son of sequence data to GenBank database entries is shown in Table 3.

4. Discussion Our results have shown the interaction between an introduced organism and the indigenous microbiota. The indigenous microbiota of the gastrointestinal tract a¡ected the colonisation of the introduced Lactobacillus strain, with extended colonisation observed in mice with an ampicillin-depleted microbiota. The same strain of Lactobacillus had an impact on the indigenous microbiota of SPF mice, with community changes observed following its administration. Although colonisation resistance can be important in the exclusion of pathogenic organisms from the gut, it can have a negative impact on the introduction of probiotic strains. The most important consideration for the maintenance of a healthy gastrointestinal microbiota is to

resist disturbing the stability of the ecosystem. Whilst probiotics can exert bene¢cial e¡ects within the gut, their introduction to a healthy gastrointestinal tract may a¡ect the already established and protective microbiota. This study showed that the L. fermentum KLD strain had a positive impact within the gastrointestinal tract, stimulating organisms whose genera have been associated with advantageous e¡ects. In SPF mice L. fermentum KLD was retained for 24^ 36 h. However, the persistence of this strain was increased to at least 11 days following administration to mice with an ampicillin-depleted microbiota. This extended detection of the strain could be due to a number of factors, including the availability of niches within the gastrointestinal tract, the lack of inhibitory factors produced by indigenous strains, or the availability of nutrients normally present in lower amounts, all of which may be a result of the absence of the indigenous microbiota. However, regardless of the mechanism, the role of the indigenous microbiota in the resistance of colonisation by introduced strains is apparent. Although the colonisation of L. fermentum KLD in mice with an ampicillin-depleted microbiota was extended to greater than 11 days, there was an absence of ‘contaminating’ lactobacilli following its removal from the system. This suggests that other defense mechanisms of the host may be relevant in preventing extended colonisation of this strain, as well as others. The use of DGGE to resolve 16S rDNA amplicons generated from a number of related Lactobacillus strains was investigated. Indigenous lactobacilli could not be distinguished from reference L. fermentum strains or the KLD strain under the DGGE conditions utilised in this study. However, both L. oris and L. reuteri could be resolved from L. fermentum using these same conditions. This failure to resolve L. fermentum strains was not surprising due to a lack of adequate sequence divergence among these strains in the V3 region. Moreover, the failure of DGGE to separate fragments that di¡ered by one to three nucleotides has been reported previously [33] and Walter and coworkers also demonstrated that several di¡erent Lactobacillus species had identical migration properties in DGGE gels [34]. Although the host lactobacilli and KLD migrated to the same position on the gel, the intensity of the band could be used to approximate the level of a

Table 3 Identi¢cation of isolates from Mouse 8 DGGE pro¢le Band

Species identi¢cation

Percentage similarity

Observed changea

1 2 3 4 5 6 7

Uncultured mouse gastrointestinal isolate S24-10 (AJ400262) Host lactobacilli Eubacterium cylindroides Bacteroides (AF132284) Bi¢dobacterium animalis (AB027536) [14] Mouse GI isolate(AJ400264) Mouse GI isolate(AJ400264)

98%

s variable u s u u u

a

97% 85% 100% 95% 95%

General trend observed in population numbers following administration of L. fermentum KLD.

FEMSEC 1428 5-2-03


particular strain. Thus, detection of a band was estimated to be dependent on the presence of greater than 104 CFU mg31 of faeces, which is considerably lower than the 106 cells mg31 value suggested previously [17]. All 10 mice in the studied group demonstrated a signi¢cant (two log; P 6 0.01) and consistent decrease in indigenous Lactobacillus numbers at 24 h post-dosing as determined by culturing (Fig. 2b shows numbers obtained from mouse 1). The reasons for this decline at this sampling time are unclear. Interestingly, there was a signi¢cant peak in the total nucleic acid extracted at this same sample time, suggesting total bacterial numbers were high (data not shown). Thus, although speculative, these divergent sampling trends at 24 h may re£ect a destabilisation of the host Lactobacillus community with a concomitant stimulation of other bacterial genera, and are worthy of further study. Excising and sequencing bands from the DGGE gel generated a phylogenetic pro¢le of some changes that resulted from dosing. Populations of Bi¢dobacterium animalis and Eubacterium cylindroides were stimulated following dosing with L. fermentum KLD. Both Bi¢dobacterium and Eubacterium are generally regarded as having a positive impact on human health [35]. Other fragments (bands 6 and 7) displayed the greatest homology with uncultured bacteria isolated from the gastrointestinal tract ; thus, these isolates could represent a species of bacteria that have so far eluded classi¢cation by microbial ecologists. Bands that decreased with dosing unfortunately could not be putatively identi¢ed, as they yielded only low or insigni¢cant homologies to known microorganisms. Band 1 (Fig. 3) was homologous to a non-culturable gastrointestinal isolate, highlighting the need for continuing e¡orts in this ¢eld of microbial ecology. A comprehensive understanding of microbial interactions that occur in the intestinal tract in addition to alterations that occur upon exposure to high levels of an exogenous strain is essential for scienti¢c acceptance of the probiotic rationale. Although it has been suggested that L. fermentum KLD has good potential as a probiotic strain [36], other workers reported that the strain exhibited poor pharmacokinetic properties [37] and displayed little probiotic promise when used in treatment of small intestine bacterial overgrowth [26]. This study provides evidence that the introduction of KLD to the murine gastrointestinal tract induces shifts in indigenous microbial communities. This study also reiterates the importance of the indigenous microbiota on the resistance of colonisation by other organisms, including introduced Lactobacillus strains.

Acknowledgements The authors wish to thank Dr Ingela Dahllo«f for expert advice on DGGE. The work was funded by the CRC for Food Industry Innovation and by an ARC-SPIRT grant.

139

References [1] Carter, P. and Pollard, M. (1971) Host responses to ‘normal’ microbial £ora in germ-free mice. J. Reticulol. Soc. 9, 580^587. [2] Tannock, G. (1997) The normal micro£ora of the gastrointestinal tract of rodents. In: Gastrointestinal Microbiology, Vol. 2 (Mackie, R., Isaacson, R. and White, B., Eds.), pp. 187^215. Chapman and Hall, New York. [3] Tannock, G.W., Crichton, C., Welling, G.W., Koopman, J.P. and Midtvedt, T. (1988) Reconstitution of the gastrointestinal micro£ora of Lactobacillus-free mice. Appl. Environ. Microbiol. 54, 2971^2975. [4] Rolfe, R. (2000) The role of probiotic cultures in the control of gastrointestinal health. J. Nutr. 130, s396^402. [5] McCracken, V. and Gaskins, H. (1999) Probiotics and the immune system. In: Probiotics: A Critical Review (Tannock, G.W., Ed.), pp. 85^112. Horizon Scienti¢c Press, Wymondham. [6] Clements, M.L., Levine, M.M., Black, R.E., Robins-Browne, R.M., Cisneros, L.A., Drusano, G.L., Lanata, C.F. and Saah, A.J. (1981) Lactobacillus prophylaxis for diarrhea due to enterotoxigenic Escherichia coli. Antimicrob. Agents Chemother. 20, 104^108. [7] Gorbach, S., Chang, T. and Goldin, B. (1987) Successful treatment of relapsing Clostridium di⁄cle colitis with Lactobacillus GG. Lancet 26, 1519. [8] Hentges, D. (1992) Gut £ora and disease resistance. In: Probiotics: The Scienti¢c Basis (Fuller, R., Ed), pp. 87^108. Chapman and Hall, London. [9] Sweeney, N., Laux, D. and Cohen, P. (1996) Escherichia coli F-18 and E. coli K-12 eda mutants do not colonise the streptomycintreated mouse large intestine. Infect. Immun. 64, 3504^3511. [10] Blomberg, L., Henriksson, A. and Conway, P. (1993) Inhibition of adhesion of Escherichia coli K88 to piglet ileal mucus by Lactobacillus spp.. Appl. Env. Microbiol. 59, 34^39. [11] Klaenhammer, T. (1993) Genetics of bacteriocins produced by lactic acid bacteria. FEMS Microbol. Rev. 12, 39^86. [12] Delneste, Y., Donnet-Hughes, A. and Schi¡rin, E. (1998) Functional foods: Mechanisms of action on immunocompetent cells. Nutr. Rev. 56, s93^s98. [13] Langendijk, P.S., Schut, F., Jansen, G.J., Raangs, G.C., Kamphuis, G.R., Wilkinson, M.H. and Welling, G.W. (1995) Quantitative £uorescence in situ hybridization of Bi¢dobacterium spp. with genus-speci¢c 16S rRNA-targeted probes and its application in fecal samples. Appl. Environ. Microbiol. 61, 3069^3075. [14] Suau, A., Bonnet, R., Sutren, M., Godon, J.-J., Gibson, G.R., Collins, M.D. and Dore¤, J. (1999) Direct analysis of genes encoding 16S rRNA from complex communities reveals many novel molecular species within the human gut. Appl. Environ. Microbiol. 65, 4799^4807. [15] Schneegurt, M. and Kulpa, C. (1998) The application of molecular techniques in environmental biotechnology for monitoring microbial systems. Biotechnol. Appl. Biochem. 27, 73^79. [16] Moore, W. and Holdeman, L. (1974) The normal £ora of 20 Japanese-Hawaiians. Appl. Microbiol. 27, 961^979. [17] Zoetendal, G., Akkermans, A. and de Vos, W. (1998) Temperature gradient gel electrophoresis analysis of 16S rRNA from human fecal samples reveals stable and host-speci¢c communities of active bacteria. Appl. Environ. Microbiol. 64, 3854^3859. [18] Simpson, J.M., McCracken, V.J., White, B.A., Gaskins, H.R. and Mackie, R.I. (1999) Application of denaturant gradient gel electrophoresis for the analysis of the porcine gastrointestinal microbiota. J. Microbiol. Methods 36, 167^179. [19] Millar, M.R., Linton, C.J., Glancy, C.D., Hall, M. and Jalal, H. (1996) Application of 16S rRNA gene PCR to study bowel £ora of preterm infants with and without necrotizing enterocolitis. J. Clin. Microbiol. 34, 2506^2510. [20] Tannock, G.W., Munro, K., Harmsen, H.J.M., Welling, G.W., Smart, J. and Gopal, P.K. (2000) Analysis of the fecal micro£ora of human subjects consuming a probiotic product containing Lactobacillus rhamnosus DR20. Appl. Environ. Microbiol. 66, 2578^2588.

FEMSEC 1428 5-2-03

140


[21] Dahllo«f, I., Ballie, H. and Kjelleberg, S. (2000) rpoB-based microbial community analysis avoids limitations inherent to 16S rDNA intraspecies heterogeneity. Appl. Environ. Microbiol. 66, 3376^3380. [22] Jensen, S., Zvreas, L., Daae, F.L. and Torsvik, V. (1998) Diversity in methane enrichments from agricultural soil revealed by DGGE separation of PCR ampli¢ed 16S rDNA fragments. FEMS Microbiol. Ecol. 26, 17^26. [23] Ferris, M.J., Nold, C., Revsbech, N.P. and Ward, D.M. (1997) Population structure and physiological changes within a hot spring microbial mat community following disturbance. Appl. Environ. Microbiol. 63, 1367^1374. [24] Zvreas, L., Jensen, S., Daae, F.L. and Torsvik, V. (1998) Microbial community changes in a perturbed agricultural soil investigated by molecular and physiological approaches. Appl. Environ. Microbiol. 64, 2739^2742. [25] Katelaris, P., Salam, I. and Farthing, M. (1995) Lactobacilli to prevent traveller’s diarrhea ? N. Engl. J. Med. 333, 1360^1361. [26] Stotzer, P.O., Blomberg, L., Conway, P.L., Henriksson, A. and Abrahamsson, H. (1996) Probiotic treatment of small intestinal bacterial overgrowth by Lactobacillus fermentum KLD. Scand. J. Infect. Dis. 28, 615^619. [27] Lane, D. (1991) 16S/23S rRNA sequencing. In: Nucleic Acid Techniques in Bacterial Systematics (Stackebrandt, E. and Goodfellow, M., Eds.), pp. 115^175. John Wiley, New York. [28] DiMichele, L. and Lewis, M. (1993) Rapid, species-speci¢c detection of lactic acid bacteria from beer using the polymerase chain reaction. J. Am. Soc. Brew. Chem. 51, 63^66. [29] Muyzer, G., De Wall, E. and Uitterlinden, A. (1993) Pro¢ling of complex microbial populations by denaturing gel electrophoresis analysis of polymerase chain reaction-ampli¢ed genes coding for 16S rRNA. Appl. Environ. Microbiol. 59, 695^700.

[30] Teske, A., Wawer, C., Muyzer, G. and Ramsing, N.B. (1996) Distribution of sulfate-reducing bacteria in a strati¢ed fjord (Mariager Fjord, Denmark) as evaluated by most-probable-number counts and denaturing gradient gel electrophoresis of PCR-ampli¢ed ribosomal DNA fragments. Appl. Environ. Microbiol. 62, 1405^1415. [31] Sambrook, J., Fritsch, E. and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd edn. Laboratory Press, Cold Spring Harbor, NY. [32] Altschul, S.F., Gish, W., Miller, W., Myers, E.W. and Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol. 215, 403^ 410. [33] Vallaeys, T., Topp, E., Muyzer, G., Macheret, V., Laguerre, G., Rigaud, A. and Soulas, G. (1997) Evaluation of denaturing gradient gel electrophoresis in the detection of 16S rDNA sequence variation in rhizobia and methanotrophs. FEMS Microbiol. Ecol. 24, 279^ 285. [34] Walter, J., Tannock, G.W., Tilsala-Timisjarvi, A., Rodtong, S., Loach, D.M., Munro, K. and Alatossova, T. (2000) Detection and identi¢cation of gastrointestinal Lactobacillus species by using denaturing gradient gel electrophoresis and species-speci¢c PCR primers. Appl. Environ. Microbiol. 66, 297^303. [35] Bezkorovainy, A. and Miller-Catchpole, R. (1989) Biochemistry and Physiology of Bi¢dobacteria. CRC Press, Boca Raton, FL. [36] Gibson, S. and Conway, P. (1994) Recovery of a probiotic organism from human faeces after oral dosing. In: Human Health, The Contribution of Microorganisms (Gibson, S., Ed.), pp. 119^144. Springer, London. [37] Vesa, T., Pochart, P. and Marteau, P. (2000) Pharmacokinetics of Lactobacillus plantarum NCIMB 8826, Lactobacillus fermentum KLD, and Lactococcus lactis MG 1363 in the human gastrointestinal tract. Aliment. Pharmocol. Ther. 14, 823^828.

FEMSEC 1428 5-2-03