Lactic acid bacteria (LAB) were isolated from human colon biopsies on LAMVAB by enrichment with different substrates such as lactose derivatives, rye arabinoxylo-oligosaccharides and rye fractions. The selected isolates were tested for their ability to adhere to Caco-2 cells. Only Lactobacillus species were enriched under these conditions. From 161 isolates screened, 28% were identified by ribotyping as Lac­tobacillus rhamnosus, 29% as L. salivarius, 14% as L. cellobiosus, 13% as L. paracasei and the rest remained unidentified. L. rhamnosus was preferentially enriched by lactulose, L. salivarius by lactobio­nic acid, L. cellobiosus by lactitol and L. paracasei by arabinoxylo-oligosaccharides. The biopsy-de­rived strains L. rhamnosus E-97948 and L. paracasei E-97949 have potential for further evaluations in their probiotic and technological properties. Lactulose may have prebiotic effects on colonic LAB by favouring their growth.'

 

 

Introduction

 

Lactic acid bacteria (LAB) have been used as starters in fermented plant or animal based foods for centuries. Their fermentation end­products enhance the shelf-life of the product, mainly by decreasing the pH and by production of lactic acid and thus preventing the growth of spoilage bacteria. An additional, and in indus­trial field nowadays more important, reason for wide use of LAB is their ability to develop a pleasant aroma in the product. Although their health improving effects were already sug­gested by Metchnikoff (reviewed by Ballongue 1998) in the beginning of the century, scientific interest in the possible probiotic properties of LAB was stimulated as late as in the 1970s (Kalantzopoulos 1997). Nowadays, the probio­tic health effects are investigated on a world­wide basis (Salminen et al. 1996). Specifically, interest is focused on the isolation and develop­ment of new probiotic strains using improved methodologies (Salminen et al. 1996). In addi­tion, it is suggested that probiotic LAB should be isolated from the host's intestine, because their colonization may be improved by host­specific adherence properties (Tannock 1990). Colonization has been shown to be important for the survival of probiotic strains in competi­tion with other intestinal microbes (Saxelin 1991).

Gastrointestinal (GI) microbiota consist of approximately 400 identified anaerobic or aerobic species, of which LAB constitute ap­proximately 10% (Tannock 1990). However, in the small intestine LAB may form the main po­pulation. The diversity and quantity of GI mi­crobiota have made it difficult to evaluate the dynamic GI microbial population using tradi­tional methods. Recently, DNA based methods such as automated ribotyping have been devel­oped to detect and study the ecology of intest­inal LAB (McCartney and Tannock 1995, McCartney et al. 1996, Kimura et al. 1997).

It is already known that diet may affect the intestinal microecology (Tannock 1990). Thus, many commercial prebiotics (fructo-oligosac­charides, xylo-oligosaccharides) have recently been introduced to specifically improve the bi­fidobacterial flora in the intestine (Crittenden 1996). However, prebiotics that may improve the intestinal LAB flora have been less studied. The properties of prebiotic substrates and pro­biotics (synbiotics) may be combined in the functional foods (Gibson and Roberfroid 1995) which mainly consists of dairy and cereal based products in the present food markets.

The aim of this study was to isolate lactic acid bacteria from human colonic biopsies by enrichment with lactose derivatives or rye frac­tions as possible prebiotic substrates. Special emphasis was focused on the ability of the iso­lates to grow on the proposed prebiotic carbo­hydrates and on their adhesion properties in vitro using the Caco-2 cell line. The results of this study show that potential probiotic strains can be selected on the basis of their specific af­finity to certain substrates. This information can be used in the design of new functional foods as substrates in new combinations.

 

 

Materials and Methods

 

Lactose derivatives and preparation of arabinoxylo-oligosaccharides and rye fractions

 

The lactose derivatives used were lactulose [(4-0-0-D-galacto-pyranosyl)-D-fructose] (Sigma Chemical Co., St. Louis, Missouri, USA), lactitol [(4-0-0-D-galactopyranosyl)-D-glucitol] (Xyrofin, Kotka, Finland) and lactobionic acid [(4-0-0-D-galactopyranosyl-D-gluconic acid] (ICN, Eshwege, Germany). Lactitol was obtained from Valio Ltd., Helsinki, Finland.

The arabinoxylo-oligosaccharides were pre­pared from rye arabinoxylan (Megazyme, Bray, Ireland). Solutions of xylan (2% w/v) dissolved in 5.0 mm ammonium acetate pH 5^.0 were pre­pared, after which 5000 nkat g^-' purified xyla­nase (Tenkanen et al. 1992) from Trichoderma reesei was added to the solutions. The hydrolysis was carried out for 24 h at 40°C in a water bath after which the enzyme was inhibited by heat­ing for 20 min at 80°C. The solution was freeze­dried for 72 h (Epsilon 2-25 DS) to obtain a mix­ture of dry oligosaccharides. The composition of reducing sugars in the preparation was ana­lysed by the HPLC (Dionex, Sunnyvale, Califor­nia, USA) method of Tenkanen et al. (1997). The linear xylo-oligosaccharides (Megazyme) and arabinose substituted xylo-oligosaccharides (Tenkanen et al.1996) were used as standards.

The rye fraction (pentosans) was prepared from 1 kg rye flour, which was refluxed by 90% ethanol to inactivate endogenous enzymes and to extract lipids. The solution was dispersed in water with termamyl to hydrolyze starch and the residual precipitate, isolated by centrifuga­tion, was extracted with Ca(OH)₂+KOH. The solution was neutralized to pH 5^.0 and the solu­bilized polysaccharides were precipitated with 50% ethanol to obtain the alkali-soluble pento­san fraction. The residue was an insoluble pen­tosan fraction. The soluble fraction was used for isolation of LAB.

 

 

Media

 

The composition of basic MRS-broth was, per litre: 10 g peptone from casein (Difco, Detroit, Michigan, USA), 5^.0 g yeast nitrogen base w/o amino acids (Difco), 5^.0 g Na-acetate, 2^.0 g K₂HPO₄ x 3H₂O, 2^.0 g (NH₄)₃C₆H₅O₇ x 2H₂O, 0^.2 g MgSO₄ x 7H₂O, 0^.05 g MnSO4 x 4H₂O and 1^.0 ml Tween 80 (Fluka, Buchs, Switzer­land). The basic MRS-broth was supplemented with 2% (w/v) of the carbohydrate to be tested. The pH of the broth was adjusted to 5^.0, 5^.5 or 6^.2 with HCl. The basic medium was sterilized in an autoclave at 121°C for 20 min and supple­mented carbohydrates were sterilized using 0^.22 μm filters (Millipore SA, Saint-Quentin, France) and rye fractions by radiation.

 

 

Isolation of LAB and testing the growth

 

Biopsy-derived bacteria were isolated from the samples obtained from voluntary patients hos­pitalized for gastrointestinal disorders, which made it possible to get biopsies during the rou­tine colonoscopy. The patients had their normal diet, which was not monitored. The biopsy sam­ples (3 x 3 mm) were taken from the healthy part of the descending colon and they were transported in 12 ml Na-thioglycollate (Difco) medium at 4°C to VTT, Finland. Samples were homogenized in 38 ml thioglycollate using Sto­macher 400 (Seward, GWB, Finland), at normal speed for 30 s. The homogenates were added (10% v/v of final broth) to basic-MRS broth sup­plemented with 2% lactose derivatives, arabi­noxylo-oligosaccharides or rye fraction, and incubated at 37°C for 24 h in anaerobic jars. The enriched cultures were plated on LAM­VAB agar (Hartemink et al. 1997) and incu­bated aerobically at 37°C for 72 h. Aerobic incubation was used in order to isolate more aerotolerant species for possible industrial use. Colonies were selected on the basis of vi­sually different morphology for further purifi­cation using MRS agar. Identified isolates were tested for their growth (Jaskari et al. 1998) in basic-MRS broth supplemented with carbohydrates using automatic turbidometer Bioscreen C system (Labsystems, Helsinki, Fin­land) and by plating on MRS agar after 48 h in­cubation. Results are presented as mean values of two tests.

 

 

Selection and identification

 

All isolates were handled in ambient atmo­sphere and were tested for Gram reaction (EBC Analytica Microbiologica, Nurnberg, Ger­many), catalase activity (Smibert and Krieg 1981) and cell morphology (light microscopy). Carbohydrate fermentation tests of selected iso­lates were carried out using the relevant API strips according to the manufacturer's instruc­tions (bioMerieux SA, Marcy-1'Etoile, France). Incubations were carried out at 37°C in anaero­bic conditions for up to 6 days. Identifications were performed by comparing the fermentation profiles with the databases contained in version 1.7.6 of ATB plus (bioMerieux SA).

Ribotyping of the isolates was carried out using the RiboPrinter^® Microbial Characteri­zation System (Qualicon^TM, Wilmington, De­troit, USA) according to the manufacturer's instructions. The automated system includes five stages: (1) DNA preparation and restriction by EcoR1, (2) separation by gel electrophoresis directly linked to a membrane transfer, (3) hy­bridization with an rRNA universal probe (Es­cherichia coli region encoding the rRNA 165-23S genes) for detection, (4) extraction and vi­sualization of the pattern (RiboPrint) and (5) characterization (RiboGroups), and where pos­sible identification, by computerized compari­son with the existing RiboPrint databases. During the screening step, each isolate was analysed only once. The isolates deposited in the VTT Culture Collection (Espoo, Finland) were ribotyped three times.

 

 

Adhesion properties

 

The adhesion properties of the tested strains were studied after growth on glucose. The hu­man colonic tumour cell line Caco-2 ATCC HTB 37 (American Type Culture Collection, Manassas, Virginia, USA) was used to indicate the bacterial colonization ability of the human gastrointestinal tract. Caco-2 cells were cultured in RPMI-HEPES medium (RPMI, Gibco BRL, Paisley, UK) supplemented with 20% fetal calf serum (YA Kemia, Helsinki, Finland), 2 mm L-glutamine (Sigma), 1 % non­essential amino acids and 100 IUml^-' penicil­lin-streptomycin solution (Gibco BRL) at 37°C in a 5% CO₂/95% air atmosphere of 5% CO₂ and 95% air. Caco-2 cells were seeded at a con­centration of 3^.2 x 10⁴ cells ml^-' to obtain con­fluence. The cell cultures were maintained for 14 days on a Chamber Slide^TM (Nunc, Naper­ville, Illinois, USA) monolayer. The culture medium was replaced every second day. Before the adhesion test the cells were gently washed with 300 µl PBS (PBS, per liter: 13^.8 g of NaH₂PO₄ x H₂O, 17^.9 g of Na₂HPO₄ x 211₂O, 9 g of NaCl. Prepare 7 mm phosphate-buffer (pH 7^.1) and add 140 mm NaCl solution) and they were overlapped with 300 µl of different dilutions of bacterial cell suspension (cell concentrations varying between 5 x 10⁵ and 1 x 10⁸ cfu ml^-1 were used in RPMI-HEPES medium without supplements. Bacterial cells were labelled using 5 µl ml^-' [methyl-1,2-³H]-thymidine (113 Ci/mmol, Amersham, Buckin­ghamshire, UK). After incubation for 1 h at 37°C the Chamber Slide (area of one cuvette: 0^.36 cm2) was gently washed with 6 x 300 µl of PBS and fixed with methanol for 10 min. The radioactivity was measured by liquid scintilla­tion (Wallac 1410, Liquid Scintillation Counter; Wallac, Espoo, Finland). Results are mean va­lues of two tests.

 

 

Results

 

Composition of arabinoxylo-oligosaccharides and rye fractions

 

Arabinoxylo-oligosaccharides consisted of (w/v, %) arabinose (0^.1), xylose (1^.5), disacchar­ide (1^.6), trisaccharides (0^.2), tetrasaccharides (0^.2), pentasaccharides (0^.1), arabinose-xylose­trisaccharides (c. 30^.7), arabinose-xylose-tetra­saccharides (c. 53^.9) and arabinose-xylose­trisaccharides (c. 11^.7) (Fig. 1). Rye fractions consisted of 60% pentosans and 40% possible protein residues.

 

Figure 1. HPLC chromatography of arabinoxylo-oligomers. X, xylose; A, arabinose; IS, internal standard; figures indicate the number of monomers in the compound.

 

 

 

Identification and characterization of the isolates

 

The numbers of identified isolates obtained using different enrichment substrates are listed in Table 1. It can be seen that the major­ity of strains isolated using lactulose were L. rhamnosus. On lactitol, the typical isolates were L. cellobiosus, on lactobionic acid L. sali­varius and on arabinoxylo-oligosaccharides L. cellobiosus and L. paracasei. Rye fractions were favoured by L. rhamnosus, L. salivarius, and to a lesser extent by L. paracasei, while only one isolate of L. cellobiosus belonged to this group.

After the preliminary screening, 161 isolates were characterized and identified (lowest simi­larity 85%), byAPI tests or by ribotyping using the commercial (DUP) and VTT databases. From the isolates, 45 were identified as Lactoba­cillus rhamnosus (28^.0%), 47 as L. salivarius or closely related to it (29^.2%), 23 as L. cellobiosus (14^.3%), 21 as L. paracasei (13^.0%) (Table 1) and 25 remained unidentified (15^.5%). However, most of the unidentified isolates were presum­ably members of L. salivarius, because their closest similarities were to the identified

 

 

 

Table 1. Colonic lactic acid bacterial isolates according to their carbohydrate source in enrichment broth

 

 

Table 2. Identification of the selected isolates deposited to the VTT Culture Collection

 

 

 

L. salivarius isolates. In addition, the digestion and/or restriction with these isolates were probably only partial and at this step the ana­lyses were carried out only once. The represen­tatives of each group were deposited in the VTT Culture Collection (Espoo; Table 2) and they were used in further studies.

L. rhamnosus isolates were detected from the samples of the patients B1 and B2 and matched three different ribogroups with good similarity (Table 2, Fig. 2). The similarity of these ri­bogroups to that of the type strain E-96031 (isolation source unknown) was 71-90%. The VTT database, which includes fingerprints of isolates from the dairy industry, identified the isolates reliably, but the profiles of API identifi­cations, with one exception, were only doubtful for these biopsy isolates.

L. salivarius species was detected from the samples of the patients B2 and B4. These iso­lates matched two ribogroups and the identifi­cation with the existing databases was poor. The similarity of these groups to that of the type strain E-97853, isolated from rumen, were 82% and 88%, respectively. The profiles of API identifications ranged from doubtful to excellent.

L. cellobiosus isolates were detected only from samples of the patient B2. All of these iso­lates matched the same ribogroup, which was not identified by the existing databases. The si­milarity of this group to that of the type strain E-82167, isolated from saliva, was only 67%. The profiles of API identifications ranged from good to excellent.

L. paracasesi isolates were detected only from the samples of the patient B1. All of these isolates matched the same ribogroup. The simi­larity of this group to that of the type strain E-93490y, isolated from dental caries, was 87% and it was identified by the VTT database. The

profiles of API identifications were good or very good.

 

 

 

Figure 2. RiboPrinter^® fingerprints of the identified biopsy isolates and of the relevant type strains.

 

 

 

Growth of selected isolates

 

The growth of bacteria (5^.0 x 10³-1^.1 x 10⁹ cfu ml^-') was dependent on the carbon source (Table 3). The best growth occurred with the strains L. rhamnosus E-97948 and E-97951, L. paracasei E-97949 and L. salivarius E-97950. In addition, lactobionic acid seemed to en­hance the growth of L. cellobiosus E-97957 and E-97958 when compared to their growth on lac­tulose, lactitol or arabinoxylo-oligosacchar­ides. With the exception of lactulose, the best substrate for growth was not the same as used in enrichment. For example the strains from E-97948 to E-97951, which were enriched with a rye fraction, showed better growth on the other carbohydrates studied, but not on arabinoxylo­oligosaccharides, The rye fraction was not tested because of the quantity of protein resi­dues it contained and because it was insoluble in water, which caused false results in the Bioscreen analysis. According to API carbohy­drate tests L. rhamnosus utilized D-arabinose, but none of the strains were able to utilize xylose.

 

 

 

Table l. Growth of isolates on different carbohydrates in 24 h measurement using Bioscreen

 

 

 

However, L. rhamnosus E-97948 and L. paracasei E-97949 showed moderate growth on arabinoxylo-oligosaccharides (Table 3). The growth of the four selected bacteria was also enumerated by plating, whereby the growth was enhanced 1-3 log_lo cfu ml^-' on all supple­mented carbohydrates (data not shown).

 

 

Adhesion properties of selected isolates

 

Over 75% adhesion was observed with L. sali­varius E-98999. This result is, however, an arte­fact of the very strong cell aggregation observed with this strain. Good adhesion (> 20%) was detected with the strains L. rham­nosus E-97948, E-97951, E-97959, E-97960, E-971000 and L. paracasei E-97949 and E-971004 (Fig. 3). Adhesion of 4-6% was detected with L. salivarius E-97955, L. cellobiosus E-97957 and E-98997 and L. rhamnosus E-97952 (negative control). Lactobacillus GG E-96522 and L. rham­nosus E-97800 (positive controls) showed 15% and 30% adhesion, respectively. The optimal amount of bacterial cells for adhesion was strain dependent and generally between 7 x 10⁶ and 8 x 10⁸ cfu ml^-1.

 

 

 

 

Figure 3. Adhesion of the selected isolates (%) on the Caco-2 cell line. Adhesion lg cfu ml-1(¨).

 

 

 

 

Discussion

 

LAMVAB agar used in this study had a much higher potential to select lactic acid bacteria from intestinal material than the modified MRS agar used in our earlier study when mainly enterococci and streptococci were en­riched (Kontula et al. 1998). This is also in agreement with the previous study of Harte­mink et al. (1997) who demonstrated the growth of intestinal isolates on LAMVAB agar. However, some lactic acid bacteria of dairy ori­gin, e.g., L. acidophilus, fail to grow on it (this study, data not shown; Hartemink, pers. comm), and this may be due to the different lac­tic acid bacteria content of different indivi­duals (Tannock 1997, Kimura 1997, this study, Table 2), but most likely that L. acidophilus are sensitive to vancomycin used in LAMVAB. Ha­milton-Miller and Shah (1998) found that L. acidophilus strains isolated from probiotic sup­plements or foods were sensitive to vancomy­cin. However, in their study the real origin of the strains remained unclear. For the purpose of isolating LAB, an advantage of LAMVAB medium is that bifidobacteria, estimated as the fourth largest group in the intestinal microbiota (Tannock 1990), do not grow on it (Hartemink 1997, this study, data not shown).

The current starter cultures from 50 Lactoba­cillus species in commercial use are mainly strains of L. acidophilus, L. delbriieckii subsp. bulgaricus, L. casei, L. fermentum, L. helveticus, L. johnsonii, L. kefir, L. paracasei, L. plantarum, L. reuterii and L. rhamnosus with possible antimicrobial or probiotic action (Davidson

and Hoover 1993, Salminen et al. 1993). Tradi­tional dairy starters are L. bulgaricus (yogurt) and L. helveticus (cheese). In our previous study, a potential probiotic strain L. rhamnosus (E-97800) was isolated from a faecal sample (Kontula et al. 1998) and in this second survey the most promising strains (E-97948, E-97949) are members of L. rhamnosus and L. paracasei. These strains showed good in vitro adhesion and growth on lactulose and rye fraction.

The advantages of the automated system are that it has a high discrimination power, and that it is rapid, standardized, labour-saving, easy and versatile. The ability to characterize below the species level can be applied in tra­cing the source of isolates. The disadvantages of the system are that, at present, it is suitable only for bacteria, it is expensive to buy and the running costs are high. Before the method can be used efficiently, many fingerprints have to be collected for different organisms and the custo­mer's own databases for RiboGroups and iden­tification created.

The results indicate that different carbohy­drates favor the enrichment of different bacter­ial species from intestinal samples and it is possible to isolate strains with substrates spe­cificities. However, in most cases there was no correlation between the carbohydrate used for enrichment and the ability to utilize that sub­strate later for growth. The growth of all iso­lates on arabinoxylans was weak, and on the other hand, lactulose was utilized most effi­ciently. Interestingly, the growth of L. cellobio­sus isolated from arabinoxylo-oligosacchar­ides enrichment broth was markedly enhanced by lactobionic acid. Tested strains showed utili­zation of lactose and D-arabinose, but not xylose.

The adhesion of strains tested by radioactive labelling was highly dependent on the concen­tration of the cells. Therefore, we recommend the use of several concentrations in testing the strains to avoid saturation of the Caco-2 cell culture.

Using vancomycin containing LAMVAB medium, this study showed that lactose deriva­tives and rye fraction mixture select L. rhamno­sus, L. paracasei, L. cellobiosus and L. salivarius species from colonic biopsy material. The adhe­sion to Caco-2 cells depended on the isolate.

Many of the strains isolated with the carbohy­drates used showed even better adhesion than the positive control probiotic Lactobacillus GG. This might indicate that this kind of human colonic material shows good promise as a source of potential probiotic isolates. However, results should be verified with in vivo experiments.

 

 

Acknowledgements

 

This work was supported by the Ministry of Education of Finland (ABS graduate school), the Finnish Cultural Foundation and the EU project/^`Enzymatic lactose valorization^'/(FAIR CT 96 -1048). We thank TerttuVilpponen-Varkila and Marja-Leena Kekalainen, Harjula Hospital, Kuopio, Finland for sending colonic mate­rial. We acknowledge Sirpa Karppinen for analysing rye fraction by HPLC, Anu Mietti­nen for technical assistance with isolation, Helena Hakuli for ribotyping, Marja-Liisa Jalovaara for testing adhesion and Ville Saukkonen for preparing arabinoxylo-oligosaccharides.

 

 

References

 

Ballongue, J. (1998) Bifidobacteria and probiotic ac­tion. Lactic acid bacteria, microbiology and func­tional aspects. In Lactic Acid Bacteria (Eds S. Salminen and A. von Wright) pp. 519-587. New York, Marcel Dekker, Inc.

 

Crittenden, R. G. and Playne, M. J. (1996) Produc­tion, properties and applications of food-grade oligosaccharides. Trends Food Sci. Technol. 7, 353-361.

 

Davidson, P. M. and Hoover, D. G. (1993) Antimicro­bial components from lactic acid bacteria. In Lac­tic Acid Bacteria (Eds S. Salminen and A. von Wright) pp. 127-159. New York, Marcel Dekker, Inc.

 

Gibson, G. R. and Roberfroid, M. B. (1995) Dietary modulation of the human colonic microbiota: in­troducing the concept of prebiotics. J. Nutr. 125, 1401-1412.

 

Hamilton-Miller, J. M. T. and Shah, S. (1998) Vanco­mycin susceptibility as an aid to the identification of lactobacilli. Lett. Appl. Microbiol. 26,153-154.

 

Hartemink, R., Domenech,V. R. and Rombouts, F. M. (1997) LAMVAB-A new selective medium for the isolation of lactobacilli from faeces. J. Microb. Methods 29, 77-84.

 

Jaskari, J., Kontula, P., Siitonen, A., Jousimies-Somer, H., Mattila-Sandholm, T. and Poutanen, K. (1998) Oat (3-glucan and xylan hydrolysates as selective substrates for Bifidobacterium and Lac­tobacillus strains. J. Appl. Microbiol. Biotechnol. 49,175-181.

 

Kalantzopoulos, G. (1997) Fermented products with probiotic qualities. Anaerobe 3,185-190.

 

Kimura, K., McCartney, A. L., McConnell, M. A. and Tannock, G. (1997) Analysis of fecal popula­tions of bifidobacteria and lactobacilli and inves­tigation of the immunological responses of their human hosts to the predominant strains. Appl. Enuiron. Microb. 63, 3394-3398.

 

Kontula, P., Suihko, M.-L. von Wright, A and Matti­la-Sandholm, T. (1999) Lactose derivatives as sub­strates for lactic acid bacteria. J. Dairy Sci. 82, 249-256.

 

McCartney, A. L., and Tannock, G. W. (1995) Ribotyp­ing of Bifidobacterium strains using cells em­bedded in agarose plugs and a 16S rDNA probe. Microb. Ecol. Health Dis. 8, 79-84.

 

McCartney, A. L., Wenzhi, W. and Tannock, G. W. (1996) Molecular analysis of the composition of the bifidobacterial and lactobacillus microflora of humans. Appl. Enuiron. Microbiol. 62, 46081613.

 

Salminen, S., Deighton, M. and Gorbach, S. (1993) Lactic acid bacteria in health and disease. In Lac­tic Acid Bacteria (Eds S. Salminen and A. von Wright) pp. 199-225. New York, Marcel Dekker, Inc.

 

Salminen S., Laine, M., von wright, A., Vuopio-Varki­la, J., Korhonen, T. and Mattila-Sandholm, T. (1996) Development of selection criteria for pro­tiotic strains to assess their potential in func­tional foods: a Nordic and European approach. Biosci. Microf.15, 61-67.

 

Saxelin, M. (1991) Dose response colonisation of faeces after oral administration of Lactobacillus casei strain GG. Microb. Ecol. Health Dis. 4, 209-214.

 

Smibert, R. M. and Krieg, N. R. (1981). In Manual of methods for General Bacteriology. pp. 409. Wa­shington DC, American Society for Microbiology.

 

Tannock, G. W. (1990). The microecology of lacto­bacilli inhabiting the gastrointestinal tract. Adv. Microb. Ecol. 11, 147-171.

 

Tenkanen, M., Puls, J. and Poutanen, K. (1992) Two major xylanases of Trichoderma reesei. Enzyme Microb. Technol.14, 566-574.

 

Tenkanen, M., Luonteri, E. and Teleman, A. (1996) Effect of side groups on the action of P-xylosidase from Trichoderma reesei against sub­stituted xylo-oligosaccharides. FEBS Lett. 399, 303-306.

 

Tenkanen, M., Makkonen, M., Perttula, M., Viikari, L. and Teleman, A. (1997) Action of Trichoderma reesei mannase on galactoglucomannan in pine craft pulp. J. Biotechnol. 57,191-204.