A continuous culture biofilm model of cariogenic responses

Journal of Applied Microbiology 2001, 90, 440±448

A continuous culture bio®lm model of cariogenic responses R.J. Hodgson, R.J.M. Lynch, G.K. Watson, R. Labarbe, R. Treloar and C. Allison Unilever Research Port Sunlight, Bebington, UK 451/7/00: received 14 July 2000, revised 10 November 2000 and accepted 16 November 2000

R . J . H O D G S O N , R . J . M . L Y N C H , G . K . W A T S O N , R . L A B A R B E , R . T R E L O A R A N D C . L L I S O N . 2001.

Aims: To validate an in vitro model for the analysis of physiological and ecological responses to sugar challenge in bacterial populations, and subsequent changes in enamel mineralization. Methods and Results: A seven-organism bacterial consortium was grown in a bio®lm mode on enamel and hydroxyapatite (HA) surfaces in a continuous culture system and exposed to repeated sucrose challenges. This produced `pH-cycling' conditions within the system. Populations on HA surfaces were enumerated. Changes in relative proportions of the different populations, and in the total viable count, were observed, between different treatments. Microradiography of the enamel sections showed increasing demineralization with increasing sucrose concentration. The lesions formed were similar to `white-spot' lesions found in vivo. Differences in the quality of bio®lms formed were also observed using Confocal Laser Scanning Microscopy. Conclusions: An in vitro model has been validated for the analysis of both physiological and ecological responses to sucrose challenges in bacterial populations, and subsequent changes in enamel mineralization. Signi®cance and Impact of the Study: This model should facilitate the study of changes in bacterial populations in response to application of putative anticaries agents and concomitant changes in enamel mineralization. INTRODUCTION Several studies have demonstrated the complex multispecies nature of dental plaque which grows as a bio®lm on the hard tissue of the oral cavity. The importance of the bio®lm mode of growth for the expression of speci®c physiological characteristics has been established in a number of studies (Hardie 1992; Brown and Gilbert 1993; Kolenbrander 1993). Growth of plaque bio®lms has been shown to occur via a sequence of colonization events in which initial adhesion to the conditioned enamel surface is followed by further bacterial-enamel binding, bacteria±bacteria interaction and growth. (Scheie 1994; Marsh and Bradshaw 1997; Moller et al. 1998). The environmental in¯uences on the plaque bio®lm are diverse and vary according to the speci®c location within the bio®lm with respect to pH, pO2, redox and substrate availability (Brown and Gilbert 1993; Bradshaw et al. 1998a, b). Oral bio®lm models which allow for the complexity of the ¯ora but provide the opportunity to Correspondence to: Mr Richard Hodgson, Oral Care, Unilever Research Port Sunlight, Quarry Road East, Bebington, Wirral CH63 3JW, UK (e:mail: [email protected]).

measure population densities of the constituent organisms with con®dence have been developed using continuous culture techniques (Bradshaw et al. 1989, 1996). The utility and validity of this approach has been demonstrated in studies of the effects of antimicrobials (Bradshaw et al. 1993), sugar alcohols (Bradshaw and Marsh 1994) and preconditioning of surfaces with saliva or bacterial products on bio®lm development (Bradshaw et al. 1997). In vitro models described previously have indicated changes in enamel integrity as a result of exposure to acid generated by acidogenic bacteria (Gilmour et al. 1993; Fontana et al. 1996). However, these models have used either single or binary species ecologies and hence, have not attempted to model interactions within diverse bacterial populations. Recently, a four-organism system which utilizes an entirely cariogenic ¯ora to develop sub-surface lesions on smooth and root surfaces has been reported (Shu et al. 2000). The model developed in this work has used bovine enamel surfaces and hydroxyapatite (HA) surfaces in a continuous culture bio®lm system in which a complex de®ned consortium was grown. This approach allowed the study of changes in enamel mineralization and pH in conjunction with concurrent changes in a bio®lm population, grown on the HA surface, representative of the ¯ora found in the oral ã 2001 The Society for Applied Microbiology

BIOFILM MODEL OF CARIOGENIC RESPONSES

cavity. Furthermore, by application of sucrose pulses, a `pHcycling' regime was established for the assessment of the cariogenic response at both the microbiological and enamel levels. This study presents data to support the utility and reproducibility of the new oral bio®lm model system to measure ecological and mineral density effects in multiple oral bio®lms and enamel surfaces in parallel. MATERIALS AND METHODS Bacterial strains The bacteria grown in the continuous culture consortium were Streptococcus gordonii (formerly Strep. sanguis) SB179 (dental plaque isolate, Unilever Research Port Sunlight, UK), Strep. salivarius 196 (dental plaque isolate, Unilever Research Port Sunlight, UK), Strep. mutans R9 (dental plaque isolate, London Hospital Medical College (LHMC), Whitechapel, UK), Actinomyces naeslundii (formerly Act. viscosus) WVU 627 (dental plaque isolate, West Virginia University, USA), Veillonella parvula (formerly V. dispar) ATCC 17745 (Rockville, MD, USA), Fusobacterium nucleatum ATCC 10953 (Rockville, MD, USA) and Prevotella nigrescens (formerly P. intermedia) T588 (dental plaque isolate, LHMC, Whitechapel, UK). Primary continuous culture growth conditions The mixed bacterial culture was grown in 250 or 500 ml vessels (Soham Scienti®c, Cambridge, UK), operated at 37°C under a gas phase of 5% (v/v) CO2 in nitrogen, at a dilution rate of 0á1 h)1. The pH was maintained at 7á0 by the automated addition of 0á5 mol l)1 NaOH. The growth medium contained 2á5 g l)1 hog gastric mucin Type III (Sigma-Aldrich), 2 g l)1 proteose peptone (Difco), 1 g l)1 trypticase peptone (BBL), 1 g l)1 yeast extract (Oxoid), 0á5 g l)1 glucose, 2á5 g l)1 KCl, 0á1 g l)1 cysteine-HCl (Sigma-Aldrich) and 1 mg l)1 haemin (Sigma-Aldrich). The medium was adjusted to pH 7á5 prior to being autoclaved (Bradshaw et al. 1989). The continuous culture was inoculated from a pooled mixture of the organisms, stored at ) 80°C in BHI broth containing 30% (v/v) glycerol. After inoculation, the continuous culture was allowed to establish for at least 7 days prior to the start of bio®lm experiments. Bio®lm growth conditions Glass continuous culture second stages vessels were constructed with glass holders containing longitudinal grooves to allow insertion of HA discs (Calcitek Inc., Carlsbad, CA, USA) or pieces of enamel ®xed into the grooves with Para®lm before autoclaving (Fig. 1). Multiple second stages were inoculated simultaneously with a mixture of planktonic

441

phase from the continuous culture and fresh medium (in a 1 : 9 ratio) with maintenance of a constant liquid level by means of a weir system (20 ml working volume). The growth medium was supplemented with 5 mmol l)1 phosphate buffer, pH 7á5. A total ¯ow rate of 15 ml h)1 of combined inoculum/fresh medium was used to obtain a dilution rate of 0á75 h)1. The contents of the second stage vessels were re-circulated (60 ml min)1) to achieve constant mixing. Liquid feeds to the second stage vessel, recirculation of contents and removal of waste were controlled by peristaltic pumps (Watson-Marlow 505 DU, WatsonMarlow, Falmouth, UK). In experiments to assess the effects of sucrose pulses, the second stage vessels were pulsed every 12 h with sucrose solution (100 mmol l)1 and 500 mmol l)1 in water) to produce initial concentrations of 10 mmol l)1 and 50 mmol l)1 in the bulk liquid phase. The pH of the planktonic phase was monitored over the course of the experiments using semi-micro electrodes (Russell, Fife, UK); these were sterilized in sodium hypochlorite and washed with sterile dH2O before use. Enumeration of organisms Bio®lms and planktonic phases of the continuous culture or second stage were disrupted (sonication, 20 s and vortex mixing, 30 s), serially diluted in phosphate-buffered saline containing 0á00001% (w/v) cetyl triammonium bromide and 0á1% (w/v) sodium thioglycollate, and spread using a spiral plater (Don Whitley Scienti®c, UK) onto a range of selective and non-selective media. Total aerobic (15% CO2 in air atmosphere) and anaerobic counts were performed on supplemented BHI agar (Holdeman et al. 1977) containing 5% (v/v) de®brinated horse blood (BHIS). Streptococcus gordonii and Strep. salivarius were enumerated on TYC agar (IDG) and Strep. mutans on TCYSB agar (van Palenstein Helderman et al. 1983). Anaerobic bacteria were enumerated on the BHIS agar containing 2á5 mg l)1 vancomycin (Sigma-Aldrich) after anaerobic growth for 4±7 days. Identi®cations were based on colony morphology and Gramstaining reaction (Holdeman et al. 1977). Enamel preparation and analysis Enamel sections were prepared using sound bovine incisors. The labial enamel surface was coarsely ground using a water-cooled, diamond sintered abrading wheel (Malvern Instruments, Malvern, UK), and the ground surface was polished on a glass plate using a 1200-grade alumina slurry (Electro Minerals Co., Manchester, UK). Enamel sections were cut from this polished surface using a diamondstudded cutting wire (Well model 3242, Le Locle, Switzerland). Four sections, of approximate area 6 ´ 1á7 mm, were cut from each tooth, the major axes running corono-

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 440±448

442

R.J. HODGSON ET AL.

Fig. 1 A schematic diagram of the multicell continuous culture bio®lm system

cervically. After treatment in the ¯ow cell and swabbing for microbial population analysis, a thin section, approximately 250 lm thick, was removed from each for radiographic analysis using a method developed from that described by Schafer (1989). The thin sections were polished to a ®nal measured thickness of about 100 lm and a radiograph taken (Type-1 A Plates, Kodak, Rochester, NY, USA). Samples were exposed to Cu K-a radiation, generated at 40 mA and 40 kV, for 50 s. The distance from sample to source was 300 mm. The developed radiograph was analysed using a computerized image analysis system (IBAS, Kontron Bildanalyse, Munich, Germany). Integrated mineral loss (IML) values were calculated for each lesion. The IML was the product of the depth of the lesion (lm) and the mean mineral density over that depth, relative to 100% (sound) mineral beneath the lesion; hence, each lesion had its own internal control. Arbitrary units were percentage mineral.lm, and the larger the IML value the greater the degree of demineralization. Qualitative data were obtained in the form of mean mineral density pro®les. These described mineral density within the lesion as a function of lesion depth.

Bio®lm imaging Fluorescence imaging of bio®lm structure (up to 50 lm thickness) on HA discs was performed using an Odyssey Confocal Laser Scanning Microscope (Noran Instruments, Bicester, UK) after staining with BacLightTM (Molecular Probes, Eugene, OR, USA). Staining procedures were as described previously (Singleton et al. 1997). A Leitz NPL Fluotar 50´ water immersion objective (Wetzlar, Germany) was used and optical sections taken at 0á5 lm and 1 lm steps through the bio®lm. Images were processed using PV WAVE Advantage (Visual Numerics, Houston, TX, USA) and Voxel View (Vital Images, Fair®eld, TA, USA). Statistical analyses Statistical analysis of comparisons between bio®lm populations on HA discs was achieved using a two-tailed Student's t-test. Statistical signi®cance of results calculated as log10 colony-forming units per disc (log cfu disc)1) was assumed at the P < 0á05 level. Calculated values of IML were also statistically analysed using the Student's t-test.

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 440±448

BIOFILM MODEL OF CARIOGENIC RESPONSES

RESULTS Reproducibility of continuous cultures and bio®lms Data on reproducibility between continuous cultures are shown in Table 1. There were generally no signi®cant differences between cultures at steady state after 1 week. A slow development of the population densities of obligate anaerobes occurred during the ®rst few days of the continuous culture operation, and the low P. nigrescens counts were correspondingly variable with high standard deviation (Table 1). Continuous cultures were always operated for at least 7 days before the second stage was connected and experiments were initiated. The reproducibility of bio®lm populations in replicate discs from the same cell, and variation between experimental runs, are shown in Table 2. The numerically predominant Table 1 Reproducibility of continuous culture populations Log10 cfu ml)1 of organism in planktonic phase Period of operation (days) 3

6±7

14

Organism

Mean

S.D.

Mean

S .D.

Mean

S.D.

Strep. gordonii Strep. salivarius Strep. mutans Act. Naeslundii Fus. Nucleatum V. parvula P. nigrescens

8á08 7á30 7á25 6á66 6á62 7á00 5á75

0á14 0á73 0á91 0á37 1á12 0á63 2á23

8á24 6á61 6á83 6á69 6á75 7á09 7á52

0á23 0á56 0á92 0á22 0á99 0á49 0á15

8á50 6á20 5á86 6á48 6á11 7á42 7á45

0á02 0á20 0á45 0á43 1á49 0á47 0á11

The data are from three continuous cultures, operated using the same inoculum, over a three month period.

Table 2 Reproducibility of bio®lm populations in the continuous culture second stages Log cfu disc)1 bio®lm population

Strep. gordonii Strep. salivarius Strep. mutans Act. naeslundii Fus. nucleatum V. parvula P. nigescens Total viable count

Experiment 1

Experiment 2

Mean

S . D.

Mean

S . D.

6á55 5á75 4á49 4á93 4á99 6á42 3á30 6á91

0á656 0á140 0á287 0á747 0á778 0á832 0 0á621

6á37 4á69 5á09 4á82 4á89 7á47 4á01 7á52

0á088 0á103 0á258 0á231 1á128 0á083 0á999 0á077

The data are from three replicate discs in two experiments.

443

organisms present were Strep. gordonii, Strep. salivarius and V. parvula, with lower Strep. mutans and A. naeslundii and low levels of the obligate anaerobes Fus. nucleatum and P. nigrescens. Generally, no signi®cant difference was observed between numbers of organisms obtained in separate experiments (Table 2). Response of bio®lm populations to sucrose pulses The response of the bacterial population to six sucrose pulses was followed over a period of 3 days (Fig. 2). An increase in total bio®lm population on HA discs occurred, with signi®cant increases of Strep. salivarius and Strep. mutans in bio®lms pulsed with sucrose concentrations. Streptococcus salivarius eventually represented more than 85% of the bio®lm population in the second stage vessel pulsed with 50 mmol l)1 sucrose. Concomitantly, decreases of Act. naeslundii and V. parvula were observed under 50 mmol l)1 sucrose-pulsed conditions. In contrast, Act. naeslundii, V. parvula and Fus. nucleatum populations increased in bio®lms pulsed with 10 mmol l)1 sucrose. Prevotella nigrescens was consistently detected at low numbers in bio®lms generated in the presence and absence of sucrose (Fig. 2). pH responses to sucrose pulses The planktonic phase of the second stage was allowed to attain a steady-state of pH 6á2 during an initial phase of bio®lm growth (18 h). Subsequently, pulsing with 50 mmol l)1 sucrose resulted in pH decreases (Fig. 3). The speed and magnitude of the pH fall increased with successive pulses until responses were observed to occur within 1 h, with a decrease to about pH 4á0. The pH remained stable at about pH 6á2 in the water-pulsed control. Effect of sucrose pulsing on enamel demineralization Mean mineral density pro®les are shown in Fig. 4. The areas between the curves and y ˆ 100% (i.e., sound enamel) correspond to IML values of 356, 861á5 and 2572á16% min.lm, while mean depths of 0, 96 and 160 lm were observed in the water-, 10 mmol l)1 sucrose- and 50 mmol l)1 sucrose-pulsed vessels, respectively. When analysed by Student's t-test, the values for the 50 mmol l)1 sucrose treatment were found to be signi®cantly different, at the P < 0á05 level, to the water and 10 mmol l)1 sucrose treatment groups. The small, positive IML value for the water-pulsed sections is an artefact caused by the data analysis programme and is not representative of the

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 440±448

444

R.J. HODGSON ET AL.

Fig. 2 The effect of sucrose pulsing on bio®lm populations in the continuous culture system. Second stages were pulsed every 12 h for 90 h with (j) control (h) 10 mmol l)1 sucrose ( ) 50 mmol l)1 sucrose. (n ˆ 3). *Signi®cant difference from water control P < 0á05

Fig. 3 The effect of sucrose pulsing in the continuous culture second stage. The vessels were pulsed with sucrose (d) after 18 h and then every 12 h (indicated by arrows) or with water (j). The pH of the planktonic phase was monitored every 60 min

condition of the sections. Microscopic analysis of lesions generated in the model typically showed the presence of a mineralized surface zone (Fig. 5). Bio®lm microstructure Three-dimensional visualization by confocal laser scanning microscopy of bio®lms generated under 50 mmol l)1 sucrose-pulsed conditions shows marked differences from water-pulsed control bio®lms. A larger overall colonization of the HA surface was observed when sucrose pulsing was carried out, with a tendency to generate large distinct regions of dense biomass, typically up to 50 lm in diameter and 25±30 lm in depth (Fig. 6a). The water-pulsed control bio®lms showed a more open distribution of bacteria over the HA surface, forming small colonies of typically 5 lm diameter and 10 lm depth (Fig. 6b).

DISCUSSION Dietary stress has been attributed as the main antagonist of the caries process, with several studies demonstrating the development of a `more cariogenic' micro¯ora in vivo following repeated intake of fermentable carbohydrates (de Stoppelaar et al. 1970; Dennis et al. 1975; Straat et al. 1975; Minah et al. 1985). These effects have also been demonstrated in animal models (Van der Hoevan et al. 1985; Beighton and Hayday 1986). Human and animal studies, however, have limitations for studying physiological and ecological responses relating to enamel lesion formation, due to the extremely complex micro¯ora and issues regarding low throughput. As a result they are unsuitable as a primary screen for potential anticaries agents. The use of an in vitro model capable of demonstrating a direct change in a representative ¯ora, together with reproducible analysis of mineral, is therefore valuable in understanding the caries

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 440±448

BIOFILM MODEL OF CARIOGENIC RESPONSES

445

Fig. 4 The effect of sucrose pulsing on enamel mineral pro®les. (j) control, (s) 10 mmol l)1 sucrose, (+) 50 mmol l)1 sucrose. (n ˆ 4)

Fig. 5 Photograph of a typical sub-surface lesion generated in the continuous culture model (´100)

Fig. 6 Confocal 3-D rendered images of typical areas of bio®lm structure generated under (a) sucrose- and (b) water-pulsed conditions. Dimensions 90 ´ 90 ´ 33 lm ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 440±448

446

R.J. HODGSON ET AL.

disease process. Alternatively, this may offer a stringent screen for novel anticaries agents where a mode of action via the micro¯ora is postulated. The model discussed in this paper is based on a de®ned mixed-species primary continuous culture developed by Marsh and colleagues (Bradshaw et al. 1989, 1996). The main nutrient sources in the growth medium were amino acids, peptides and mucin, which provided the primary carbon source for the bacterial mixed culture similar to that available in vivo (Beighton and Hayday 1986). Pulsed sucrose was used to generate a cariogenic response (Bradshaw et al. 1989). By using multiple second-stage vessels, each receiving identical constant inoculation from a primarystage continuous culture, the production of reproducible bio®lms formed on inserted surfaces is ensured. Thus, a major advantage this system has over other methods for generating bio®lms is that differing in situ treatments of each vessel can occur during the process of bio®lm formation. The small volume second-stage vessels allow operation of high dilution rates, more representative of those found in the oral cavity, with relatively small volumes of fresh medium, while still maintaining a degree of control over the clearance of agents introduced into the system. Previous studies using a nine-membered consortium containing organisms representative of the oral ¯ora (McKee et al. 1985) demonstrated stable primary continuous culture populations. The seven-membered consortium used in the current studies showed similar levels of reproducibility, allowing this primary culture to act as a suitable inoculum for the second stages containing enamel surfaces. Data generated for bio®lms formed on HA surfaces in two experiments, in the absence of sucrose pulsing, demonstrated good reproducibility, Strep. gordonii and V. parvula forming the numerically predominant organisms. These data were consistent with results generated using a 10-membered community, where bio®lm populations after 4 days comprised mainly Strep. gordonii with comparable levels of V. parvula (Bradshaw et al. 1996). However, Fus. nucleatum and P. nigrescens populations were lower in the current study. The smaller numbers of Fus. nucleatum and the blackpigmented anaerobe observed may be due to a selective effect of the higher dilution rate (D ˆ 0á75 h)1 compared with 0á1 h)1 used in this work). The effect of carbohydrate on a de®ned oral micro¯ora in vitro has been shown to be selective and demonstrates cariogenic shifts in the absence of pH control (Bradshaw et al. 1989). The pH measurements made in the planktonic phase of second stages pulsed with sucrose showed `ampli®ed' Stephan responses to sucrose challenge. Clearance of acid was over a 12-h period, with about 4-h at minimum pH; this is compared with in vivo Stephan response times of around 60 min (Nisengard and Newman 1994). This increased intensity of the pH response accelerated the

cariogenic shifts associated with bacterial populations and enamel. The selective increase in numbers of the aciduric organisms, Strep. salivarius and Strep. mutans, with concomitant decreases in populations of the commensal organisms Strep. gordonii and V. parvula in the presence of 50 mmol l)1 sucrose pulses, was consistent with previous studies obtained in planktonic cultures, which showed increases in aciduric Lact. casei and Strep. mutans with decreases in Strep. sanguis (Bradshaw et al. 1989). Pulses of 10 mmol l)1 sucrose were less selective than 50 mmol l)1 sucrose, and all bacterial population densities increased in parallel, due to an insuf®cient period of low pH to induce a species-speci®c ecological shift. Increased frequency of sucrose pulses in the 50 mmol l)1 treatment could result in a failure to clear all the acid from the system before the subsequent pulses. In this case, the pH would most likely progress to a constant low pH, resulting in more severe enamel demineralization and a shift to an exclusively aciduric micro¯ora. Radiographic analysis of the enamel sections revealed subsurface lesions consistent with those associated with `white spot' lesions found in vivo, i.e., a demineralized lesion body and a more highly mineralized surface zone. The formation of this type of lesion in vivo, as opposed to an erosive lesion, has been proposed to be in¯uenced by several factors, including limited diffusion at the enamel surface (Anderson and Elliot 1992). The presence of a salivary pellicle or plaque may therefore in¯uence sub-surface lesion formation. This lesion pro®le was successfully reproduced in this study using the mixed culture bio®lm. Bio®lm structure was clearly affected by sucrose pulsing, producing a more compact bio®lm structure compared with the water control. A lower level of colonization of the HA surface was observed under carbohydrate limitation, and this was re¯ected in both the total viable count (Fig. 2) and bio®lm imaging (Fig. 6). These ®ndings con®rm previous work using a similar bio®lm system (Singleton et al. 1997). Increased density of bio®lm structure has been implicated as a barrier to diffusion of antimicrobials through bio®lms to reach their site of action (Anwar et al. 1992; Brown and Gilbert 1993). Furthermore, extracellular polysaccharides produced as a result of sucrose metabolism may also play a role in enhancing the cariogenic challenge, possibly through in¯uence on sugar and acid diffusion (McNee et al. 1982; Dibdin et al. 1983), or through the facilitation of fermentable substrate penetration to deeper regions of the bio®lm, where the reduced ®xed buffer effect from lower numbers of bacteria may allow a more pronounced pH fall at the plaque enamel interface (Zero et al. 1986; Dibdin and Shellis 1988). The ecological plaque hypothesis for dental caries (Marsh and Bradshaw 1997) states that generation of a low pH

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 440±448

BIOFILM MODEL OF CARIOGENIC RESPONSES

environment from sugar fermentation results in an ecological shift to a more cariogenic micro¯ora and an upset of the enamel demineralization/remineralization balance. The results presented here using the continuous culture system con®rm the ecological responses consistent with this hypothesis. Furthermore, this work has shown that the relationship between ecology, physiology (acid production) and enamel structure may be modelled in a single system. REFERENCES Anderson, P. and Elliot, J.C. (1992) Subsurface demineralization in dental enamel and other permeable solids during acid dissolution. Journal of Dental Research 71, 1473±1481. Anwar, R.I., Strap, J.L. and Costerton, J.W. (1992) Establishment of ageing bio®lms : possible mechanism of bacterial resistance in antibiotic therapy. Antimicrobial Agents and Chemotherapy 36, 1347± 1351. Beighton, D. and Hayday, H. (1986) The in¯uence of the availability of dietary food on the growth of Streptococci on the molar teeth of monkeys (Macaca fascicularis). Archives of Oral Biology 31, 449±454. Bradshaw, D.J. and Marsh, P.D. (1994) Effect of sugar alcohols on the composition and metabolism of mixed culture of oral bacteria grown in a chemostat. Caries Research 28, 251±256. Bradshaw, D.J., Marsh, P.D., Gerritsen, H., Vroom, J., Watson, G.K. and Allison, C. (1998b) Detection of pH gradients using 2-photon excitation microscopy. Abstract. Journal of Dental Research 77, 988. Bradshaw, D.J., Marsh, P.D., Watson, G.K. and Allison, C. (1998a) Role of Fusobacterium nucleatum and coaggregation in anaerobe survival in planktonic and bio®lm oral microbial communities during aeration. Infection and Immunity 66, 4729±4732. Bradshaw, D.J., Marsh, P.D., Schilling, K.M. and Cummins, D. (1996) A modi®ed chemostat system to study the ecology of oral bio®lms. Journal of Applied Bacteriology 80, 124±130. Bradshaw, D.J., Marsh, P.D., Watson, G.K. and Cummins, D. (1993) The effect of Triclosan and zinc citrate, alone and in combination, on community of oral bacteria grown in vitro. Journal of Dental Research 72, 25±30. Bradshaw, D.J., Marsh, P.D., Watson, G.K. and Schilling, K.M. (1997) The effect of conditioning ®lms on oral microbial bio®lm development. Biofouling 11, 217±226. Bradshaw, D.J., McKee, A.S. and Marsh, P.D. (1989) Effects of carbohydrate pulses and pH on population shifts within oral communities in vitro. Journal of Dental Research 68, 1298±1302. Brown, M.R.W. and Gilbert, P. (1993) Sensitivity of bio®lms to antimicrobial agents. Journal of Applied Bacteriology 74, 87S±97S. Dennis, D.A., Gawronski, T.H., Sudo, S.Z., Harris, R.S. and Folke, L.E.A. (1975) Variations in microbial and biochemical components of four day plaque during a four week controlled diet period. Journal of Dental Research 54, 716±722. De Stoppelaar, J.D., Van Houte, J. and Backer Dirks, O. (1970) The effect of carbohydrate restriction on the presence of Streptococcus mutans, Streptococcus sanguis and indophilic polysaacharide producing bacteria in human dental plaque. Caries Research 4, 114±123. Dibdin, G.H. and Shellis, R.P. (1988) Physical and biochemical studies of Streptococcus mutans sediments suggest new factors linking the

447

cariogenicity of plaque with its extracellular polysaccharide content. Journal of Dental Research 67, 890±895. Dibdin, G.H., Wilson, C.M. and Shellis, R.P. (1983) Effect of packing density and polysaccharide to protein ratio of plaque samples cultured in vitro upon their permeability. Caries Research 17, 52±58. Fontana, M., Dunipace, A.J., Gregory, R.L. et al. (1996) An in vitro microbial model for studying secondary caries formation. Caries Research 30, 112±118. Gilmour, A.S.M., Edmunds, D.H., Newcombe, R.G. and Clark, M.F. (1993) An in vitro study into the effects of a bacterial arti®cial caries system on the enamel adjacent to composite and amalgam restorations. Caries Research 27, 169±175. Hardie, J.M. (1992) Oral microbiology : current concepts in the microbiology of dental caries and periodontal disease. British Dental Journal 172, 271±278. Holdeman, L.V., Cato, E.P. and Moore, W.E.C. (eds) (1977) Anaerobe Laboratory Manual 4th edn. Virginia : V.P.I. Anaerobe Laboratory. Kolenbrander, P.E. (1993) Coaggregation of human oral bacteria : potential role in the accretion of dental plaque. Journal of Applied Bacteriology 74, 79S±86S. Marsh, P.D. and Bradshaw, D.J. (1997) Physiological approaches to the control of oral bio®lms. Advances in Dental Research 11, 176±185. McKee, A.S., McDermid, A.S., Ellwood, D.C. and Marsh, P.D. (1985) The establishment of reproducible, complex communities of oral bacteria in the chemostat using de®ned inocula. Journal of Applied Bacteriology 59, 263±275. McNee, S.G., Geddes, D.A.M., Weetman, D.A., Sweeney, D. and Beeley, J.A. (1982) Effect of extracellular polysaccharides on diffusion of NaF and 14C-sucrose in human dental plaque and in sediments of the bacterium Streptococcus sanguis 804 (NCTC 10904). Archives of Oral Biology 27, 981±986. Minah, G.E., Solomon, E.S. and Chu, K. (1985) The association between dietary sucrose consumption and microbial population shifts at six oral sites in man. Archives of Oral Biology 30, 397±401. Moller, S., Sternberg, C., Andersen, J.B. et al. (1998) In situ gene expression in mixed-culture bio®lms : evidence of metabolic interactions between community members. Applied and Environmental Microbiology 64, 721±732. Nisengard, R.J. and Newman, M.G. (1994) Oral Microbiology and Immunology 2nd edn. Philadelpia : W.B. Saunders Company. van Palenstein Helderman, W.H., Ijsseldijk, M. and Huis In't Veld, J.H.J. (1983) A selective medium for the two major subgroups of the bacterium Streptococcus mutans isolated from human dental plaque and saliva. Archives of Oral Biology 28, 599±603. Schafer, F. (1989) Evaluation of the anticaries bene®t of ¯uoride toothpastes using an enamel insert model. Caries Research 23, 81±86. Scheie, A.A. (1994) Mechanisms of dental plaque formation Advances in Dental Research 8, 246±253. Shu, M., Wong, L., Miller, J.H. and Sissons, C.H. (2000) Development of multi-species consortia bio®lms of oral bacteria as an enamel and root caries model system. Archives of Oral Biology 45, 27±40. Singleton, S., Treloar, R., Warren, P., Watson, G.K., Hodgson, R.J. and Allison, C. (1997) Methods for microscopic characterisation of oral bio®lms : analysis of colonization, microstructure, and molecular transport phenomena. Advances in Dental Research 11, 133± 149.

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 440±448

448

R.J. HODGSON ET AL.

Straat, R.H., Gawronski, T.H., Cressey, D.E., Harris, R.S. and Folke, L.E.A. (1975) Effects of dietary sucrose levels on the quantity and microbial composition of human dental plaque. Journal of Dental Research 54, 872±880. Van der Hoevan, J.S., De Jong, M.H. and Rogers, A.H. (1985) Effect of utilisation of substrate on the composition of dental plaque. FEMS Microbial Ecology 31, 129±133.

Zero, D.T., Van Houte, J. and Russo, J. (1986) The intra-oral effect on enamel demineralisation of extracellular matrix material synthesised from sucrose by Streptococcus mutans. Journal of Dental Research 65, 918±923.

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 440±448