The effects of agar concentration on the growth and morphology of submerged colonies of motile and nonmotile

Transcription

1 Journal of Applied Microbiology 1997, 83, 76 8 The effects of agar concentration on the growth and morphology of submerged colonies of motile and nonmotile bacteria A.J. Mitchell and J.W.T. Wimpenny School of Pure and Applied Biology, University of Wales College of Cardiff, Cardiff, UK 58/6/96: received 19 June 1996, revised and accepted 5 November 1996 A.J. MITCHELL AND J.W.T. WIMPENNY The growth and morphology of submerged bacterial colonies was investigated. Five separate colonial forms were recognized depending both on species and on agar concentration. These were (i) branched, dendritic structures seen only with Bacillus cereus; (ii) lenticular colonies for all other species at high agar concentrations; (iii) small lobed to spherical colonies for non-motile organisms at low agar concentrations; (iv) and (v) large diffuse spherical colonies which can be further subdivided into snowball or wispy types for motile bacteria growing at agar concentrations below about 65% w/v. Viable count determinations suggested that agar concentration had little effect in the early stages of growth but that motile cells at low agar concentrations achieved higher cell numbers than did those in concentrations greater than 65% w/v. Transmission electron microscopy indicated that bacteria in lenticular colonies were tightly packed within lens-shaped splits in the agar whilst at low agar concentrations motile cells were well separated and appeared to move through the agar matrix. INTRODUCTION Correspondence to: Dr J. W. T. Wimpenny, School of Pure and Applied Biology, University of Wales College of Cardiff, Museum Avenue, Cardiff CF1 3TL, UK ( sabjw@cardiff.ac.uk). Bacterial growth takes place in a homogeneous well-mixed growth medium largely as dispersed cells. The same organisms can grow on and inside solid media where they form aggregates known as colonies. Surface colonies generally possess a characteristic mor- phology and it is often possible to use this to recognize particular species. The shape of a colony is determined by a combination of intrinsic genotypic factors and extrinsic physicochemical factors, including specifically solute diffusion gradients. Much less is known about the growth of subsurface colonies. These differ in their topology since oxygen and substrates generally enter from opposite sides of the surface colony, whilst they enter the subsurface colony from its outer surface only. Many early observations on subsurface bacterial growth derived from the studies in shake cultures. A particularly good example of this was the growth of lactic acid bacteria (Whittenbury 1963). Recently there has been a growing interest in understanding what happens to bacterial populations developing within a heterogeneous food matrix. In the present investigation agar was chosen to form the physical matrix, because it is commonly used both by micro- biologists and in the food industry. In view of the importance of food poisoning and food spoilage bacteria, a selection of both types of bacteria was included in this work. MATERIALS AND METHODS Organisms Staphylococcus aureus (CRA 1), Bacillus cereus (CRA 666), Salmonella typhimurium (CRA 518), Listeria monocytogenes (CRA 33, NCTC 1199) were acquired from the Campden and Chorleywood Food Research Association, Chipping Campden, Gloucestershire, UK. A highly motile variant of the salmonella was selected by passage through Craigie tubes. It was designated Salm. typhimurium CRA 518-M. Escherichia coli E6188 serotype O:H5 (E. coli H5) and E61883 serotype O:H (E. coli H ) were obtained from Dr B. Rowe, Division of Enteric Pathogens, Central Public Health 1997 The Society for Applied Bacteriology

2 SUBMERGED COLONY GROWTH 77 Laboratory, London. A highly motile variant of E. coli H5 was isolated. This was designated E. coli H5-M. Enterococcus faecalis and Pseudomonas aeruginosa PU1 were from UWCC, Cardiff. Medium and culture conditions Colonies were removed from the plate and trimmed to remove as much excess agar as possible. The samples were fixed for 1 h in 3% (v/v) gluteraldehyde in phosphate buffer ( 1 mol l 1 ). Samples were washed in 1 mol l 1 phosphate buffer overnight, followed by post-fixation in Millonig s phosphatebuffered osmium tetroxide for 1 h. Samples were then dehy- drated in a graded series of aqueous ethanols (3 1%) over a 8 h period. The alcohol was then replaced with Spurr resin (BioRad). At 1 h intervals the ratio of alcohol to resin was changed as follows: 5:5, 5:75, 1:9 and 1% and mixed overnight. Fresh resin was added and mixed for a further day. All the samples were gently mixed using a rotator (Agar Scientific, UK). The samples were placed in a mould with fresh resin and polymerized at 6 C for 8 h. The embedded colonies were then sectioned with an LKB Ultratome III ultramicrotome. Sections were collected on copper grids and stained for 1 h in saturated uranyl acetate and 3 min in 8 mol l 1 lead citrate and 1 mol l 1 tri-sodium citrate in 16 mol l 1 NaOH before rinsing in mol l 1 sodium hydroxide and double distilled water. Sections were dried on Whatman filter paper and viewed in a Jeol 1S electron microscope at 8 kv. All cultures were maintained on nutrient agar (Oxoid) slopes. The cultures were grown in nutrient broth (Difco) overnight at 3 C. The basal agar medium contained (g l 1 distilled water): Bacto peptone (Difco), 1 ; NaCl (BDH), 5 ; BiTek agar (Difco), 5 ; and was sterilized at 11 C for 15 min. Filter sterilized glucose (BDH) was then added to give a final concentration of 5 g l 1. Pour plates were prepared using ml of medium giving a depth of 7 mm. The empty plates were inoculated with 1 ml of an overnight broth culture appropriately diluted in quarter-strength Ringer s solution (Oxoid). Melted medium ( ml) was then added and mixed with the inoculum followed by a second aliquot of ml. The inoculation procedure was chosen to give ¼ colonies per plate. The plates were then incubated at 3 C. Motility tests and the selection of motile variants The motility of strains used in this work was assessed using Craigie tubes (Craigie 1931; Cowan and Steel 197). Craigie tubes were also used to select for highly motile variants of E. coli and Salm. typhimurium. The motility of each inoculum was checked by phase contrast microscopy before use. Colony diameter measurement For each agar strength 1 well separated colonies from three different plates were marked and their maximum diameters measured at h intervals using a calibrated plate viewing microscope (Olympus) with a 1 objective. Each eye piece division corresponded to 5 mm. This same system was used to photograph the colonies. Viable count determination Viable counts were estimated as follows: a plate was inoculated with 1 ml of an appropriate serial dilution, mixed with ml of agar and incubated at 3 C. Agar plugs with a volume of 35 ml were extracted with a cork borer and dispersed using eight strokes of a hand-held homogenizer in 6 85 ml of quarter-strength Ringer s solution. Nutrient agar spread plates were used to measure the viable count of the homogenized agar suspensions. In the early phase of growth, when colonies were not visible, plugs were removed from an agar plate at random and their viable counts determined as above. This procedure was repeated every h. Duplicate plates were incubated for h and the number of viable colonies per plug determined. After correcting for dilution and the number of colonies per plug, the average number of viable bacteria per colony was calculated. The plates used in the early phase of growth were given a higher inoculum than those in the later phase. In the later phase of growth when individual colonies were visible, plugs were removed containing distinct colonies and their viable count determined as above. Transmission electron microscopy RESULTS Radial growth The growth of a range of motile and non-motile bacteria, expressed as submerged colony diameter, was followed as a function of time and of agar concentration (Fig. 1). The motile species, Ps. aeruginosa, Salm. typhimurium and L. monocytogenes, all showed similar behaviour (Fig. 1a c). At low agar concentrations these organisms produced large colonies whose size was inversely related to the agar concentration. Above a critical value of between 5 and 6% (w/v) agar in the medium, these organisms produced small lenticular colonies that were independent of higher agar con- centrations. The appearance of the colonies of non-motile species, including Ent. faecalis (Fig. 1d) and Staph. aureus (not shown), showed no such relationship with agar concentration. The motile B. cereus was only slightly affected by agar (Fig. 1e). The suggestion that motility was responsible for

3 78 A.J. MITCHELL AND J.W.T. WIMPENNY 8 (a) 8 (e) Colony diameter (mm) (b) 7 6 (f) Colony diameter (mm) Colony diameter (mm) 1 (c) 5 3 (g) Colony diameter (mm) 8 6 (d) Time (h) Time (h) Fig. 1 The effect of agar concentration on colony diameter of (a) Pseudomonas aeruginosa, (b) Listeria monocytogenes, (c) Salmonella typhimurium, (d) Enterococcus faecalis, (e) Bacillus cereus, (f) Escherichia coli motile and non-motile variants and (g) motile and motility impaired Ps. aeruginosa. Motile bacteria:, 5% agar;, 55% agar; e, 6% agar; e, 65% agar;, 7% agar; $, 75% agar; r, 8% agar; t, 1 % agar; u, 1 5% agar; y, % agar. Non-motile bacteria: Ž, 5% agar; E, 6% agar; ž, 7% agar; T, 1 % agar; U, 1 5% agar; Y, % agar 1

4 SUBMERGED COLONY GROWTH 79 Radial growth rate (µm h 1 ) Agar concentration (% w/v) Fig. The effect of agar concentration on radial growth rate of subsurface colonies. e, Listeria monocytogenes;, Salmonella typhimurium; r, Bacillus cereus;, Pseudomonas aeruginosa; t, Enterococcus faecalis;, Staphylococcus aureus these agar-related phenomena was given added weight using variants of motile bacteria. A highly motile variant of E. coli O selected for using Craigie tubes generated large colonies in 5% (w/v) agar whilst a non-motile variant did not (Fig. 1f). Colonies of a motility-impaired variant of Ps. aeruginosa were unaffected by agar concentration whilst those of its parent produced large colonies with 5% (w/v) agar in the medium (Fig. 1g). The initial rate of increase in colony diameter, for the species examined, was plotted as a function of agar concentration (Fig. ). The breakpoint at around 6% motile species form smaller spherical structures that were denser and slightly irregular in their surface texture (Fig. 3 b(i)). These were called snowball colonies and are characteristic of Salm. typhimurium. Viable counts and radial growth The relationship between morphology and the recovery of viable cells as colony-forming units was tested in cultures of two motile species, Salm. typhimurium H5-M and Ps. aeruginosa. In the former (Fig. a), agar concentration had little effect on the growth rate of the organism up to 1 h after inoculation. Over this period the increase in numbers was exponential giving specific growth rates of 1 5 and 1 3 h 1 in 5% and 1 % (w/v) agar, respectively. The colony dimensions altered according to the agar concentration. At both concentrations, a linear increase in diameter was observed over the h incubation period, although the diameter of the snowball colony in 5% agar increased at a faster rate than the lenticular structure at the higher agar strength (Fig. a(ii)). After 1 h the number of cfu recovered was higher in the larger diffuse colony than in the compressed lenticular structure (Fig. a(i)). Similar results were obtained with the pseudomonad (Fig. b). Colony structure Sections were taken across a large spherical colony and a compact lens-shaped colony of Ps. aeruginosa in 5% (w/v) and 1 % (w/v) agar, respectively. Examination of the sections by transmission electron microscopy showed that the cells in the large spherical colony were highly dispersed (not shown), in contrast to the densely packed cells seen in the (w/v) agar can be clearly seen. lenticular colony (Fig. 5). It is clear from the data presented in this paper that the agar concentration influenced both the morphology and dimen- sions of colonies, especially of motile bacteria, growing in a three-dimensional gel matrix. For motile species such as Ps. aeruginosa, L. monocytogenes, Salm. typhimurium and E. coli, colonies grew as large diffuse spherical structures whose diameters fell as agar concentration increased. At a critical agar concentration, around 65% (w/v), a distinct mor- phological change from diffuse and spherical to compact and lenticular colonies was also observed. Non-motile strains showed no such clear decrease in colony diameter with increasing agar concentration, although a change from a compact lobed, to a compact lenticular structure was evident. Comparisons of the motile and non-motile strains of E. coli suggested that the characteristic colony shapes observed for each strain were due to the ability of the Colony morphology Colony morphology was examined by light microscopy. The results (Fig. 3) suggest a number of different morphologies. Five different forms were discerned. At agar concentrations above 6% (w/v), colonies of almost all species were small and lenticular (Fig. 3a(ii), b(ii) and c(ii)). The exception was B. cereus (Fig. 3d(ii)). This organism formed dendritic or rhizoid colonies which were almost independent of agar concentration. Below 6% (w/v) agar three further forms were seen. For non-motile species and variants of motile organisms the colonies were small, approximately spherical but with pronounced lobes (Fig. 3c(i)). Motile bacteria generate large diffuse colonies (Fig. 3a(i) and b(i)). These diffuse colonies can be further separated into two sub-forms. The first were the largest in size, showed good spherical symmetry and were very diffuse and hence hard to see (Fig. 3 a(i)). These we refer to as wispy, and were seen with Ps. aeruginosa. Other DISCUSSION

5 8 A.J. MITCHELL AND J.W.T. WIMPENNY Fig. 3 Photographs of colony morphology in (i) 5% and (ii) 1 % w/v agar of (a) Pseudomonas aeruginosa, (b) Salmonella typhimurium, (c) Enterococcus faecalis and (d) Bacillus cereus

6 SUBMERGED COLONY GROWTH 81 Fig. 3 (continued) motile bacteria to actively move through the more dilute agar matrix thus creating the diffuse colony type. Agar levels above the critical concentration produced compact lenticular colonies for both the motile and non-motile species suggesting that the movement of both types was restricted by the agar matrix.

7 8 A.J. MITCHELL AND J.W.T. WIMPENNY 9 8 (i) (a) 9 8 (i) (b) Colony diameter (mm) Colony viable count (log 1 cfu) (ii) (ii) Time (h) Time (h) Fig. (i) Viable count and (ii) colony diameters of (a) Salmonella typhimurium and (b) Pseudomonas aeruginosa colonies grown in the basal medium with 5% (w/v) glucose and stabilized with 5% and 1 % (w/v) agar., 5% (w/v) agar; t, 1 % (w/v) agar 3 ited more rapidly than do the motile bacteria which were free to move outwards to higher nutrient concentrations. Agar consists of a mixture of linear galactans. The most extensively studied agar is agarose. In solution the tertiary structure is believed to be expressed as a double helix. Agarose may form a macro-reticular network of several sideways- linked helices whose ends are joined together randomly to form pores. The extent and size of this network is not yet The initial growth rates for Salm. typhimurium and Ps. aeruginosa over 1 h appeared similar, but final viable counts after 8 h for both the large snowball or spherical colonies were higher than for the small, compact lenticular colonies. Transmission electron micrographs demonstrated that cells were well separated and dispersed in the snowball colonies but densely packed in the lenticular colonies. This suggests that the latter might become diffusion lim-

8 SUBMERGED COLONY GROWTH 83 Fig. 5 Transmission electron microscopy section of a h Pseudomonas aeruginosa colony in 1 % agar The results reported here provide a platform for further work on submerged colony growth. A number of phenomena have been ignored; for instance, specific interactions between cell wall components and the agar such as the phase variations seen with Staph. aureus have not been considered. We have also so far not attempted to model submerged colony growth or to fit the data to existing computer models for cell pro- liferation or for fractal structure formation (Markx and Davey 199; Schindler and Rataj 199; Schindler and Rovensky 199). More needs to be done on the effects of nutrient concentration and on changes in the physical chemistry in the immediate vicinity of a submerged colony before any accurate model can be generated. Our main target has been to chart the effects of the matrix material on the way in which a bacterial colony develops within a structure. Such data have implications for growth in microbial ecosystems such as sediments and gelatinous biofilms. Their most interesting applications, however, might be to the food industry where it is important to know how microcolonies develop and spread through a food matrix. ACKNOWLEDGEMENTS This work was supported by the Agricultural and Food Research Council (grant number 1736). The authors would also like to thank the Campden and Chorleywood Food Research Association for its support and Dr Linda Thomas both for her constant interest in this project and for reading this manuscript. Thanks are also due to Mrs Carole Winters for help with electron microscopy. known, but the easiest way of reducing the mesh size of a gel is by increasing its concentration: this also raises its resistance to mechanical deformation (Armisen 1991). In the present investigation it seems possible that the agar, acting as a variable pore size net, restrained motile organisms above a critical value of 65% (w/v) concentration. At lower agar concentrations where the pore size is larger, motile cells may be able to move away from the colony. The diameter of the latter increased in a linear fashion forming either large spherical almost transparent wispy structures or showed a roughening effect leading to the formation of snowball - shaped colonies. Above this critical agar concentration the trapped cells could continue to divide at rates governed by the speed with which nutrients diffused through the agar matrix to the colony. In lenticular colonies the agar split along a particular fault line due to the hydrostatic pressure exerted by the growth. The agar remained compressed in the centre of the structure leading to the formation of the characteristic lens or discusshaped structure. The main anomaly in the work described here is with the highly motile B. cereus. This Gram-positive rod-shaped bacterium generates very large colonies on most media (Wimpenny 1979). Our experiments showed that the colonies were large, branching, dendritic structures and their radial growth was more or less independent of agar concentration up to 1 5% (w/v) (Figs 1e and 3d). It has been reported that B. cereus can degrade agar (Armisen 1991). Although this effect is not obvious on most agar-based media, it is important to realize that only a few chains need to be hydrolysed to increase the pore size significantly, allowing cells to move through the agar.

9 8 A.J. MITCHELL AND J.W.T. WIMPENNY REFERENCES Armisen, R. (1991) Agar and agarose biotechnological applications. Hydrobiologica 1, Cowen, S.T. and Steel, K.J. (197) Manual for the Identification of Medical Bacteria. p. 7. Cambridge University Press. Craigie, J. (1931) Studies on the serological reactions of flagella of B. typhosus. Journal of Immunology 1, 17. Markx, G.H. and Davey, C.L. (199) Applications of fractal geometry. BINARY Computing in Microbiology, Schindler, J. and Rataj, T. (199) Fractal geometry and growth models of a Bacillus colony. BINARY Computing in Microbiology, Schindler, J. and Rovensky, L. (199) A model of intrinsic growth of a Bacillus colony. BINARY Computing in Microbiology 6, Whittenbury, R. (1963) The use of soft agar in the study of conditions affecting the utilisation of fermentable substrates by lactic acid bacteria. Journal of General Microbiology 3, Wimpenny, J.W.T. (1979) The growth and form of bacterial colonies. Journal of General Microbiology 11,