- Email: [email protected]
Production of viable cultures of Flavobacterium psychrophilum: approach and control
Res. Microbiol. 150 (1999) 351−358 © Elsevier, Paris
Production of viable cultures of Flavobacterium psychrophilum: approach and control Christian Michela*, Dolorès Antoniob, Ronald P. Hedrickb a
Unité de virologie et immunologie moléculaires, E´quipe de pathologie infectieuse et immunité des poissons, Inra-CRJJ, Domaine de Vilvert, Centre de recherches de Jouy-en-Josas, 78352 Jouy-En-Josas cedex, France b Department of Medicine and Epidemiology, School of Veterinary Medicine, University of California at Davis, Davis, CA 95616, USA (Submitted 12 September 1998; accepted 10 March 1999)
Abstract — Although the fish pathogen Flavobacterium psychrophilum is a major source of concern in salmonid hatcheries, few studies have been conducted on its pathogenicity. Difficulties are often experienced when trying to control or quantify standard procedures for in vitro culture of the bacterium. Plate enumeration and counting chamber enumeration combined with epifluorescent microscopy with fluorescent dyes determined that no more than 25% of the bacterial cells present in the cultures were able to produce colonies on agar media. This was strongly dependent upon different medium components. Tryptone-enriched Anacker and Ordal medium proved more suitable than tryptone-yeast extract-salts with skimmed milk. Adding horse serum and trace elements in controlled proportions offered the most reproducible results. Viable but nonculturable forms were apparently not responsible for the difficulties in production of F. psychrophilum, but the cells were highly susceptible to osmotic conditions. Improvements in the media and careful handling of the bacteria in isotonic suspension media resulted in predictable production of viable bacteria and allowed an absorbance/colony-forming-units relation curve to be established. © Elsevier, Paris Flavobacterium psychrophilum / bacterial growth / microbial nutrition / viability
1. Introduction Fish pathogenic Flavobacterium species, most of which were formerly ascribed to the genera Flexibacter and Cytophaga [4], have become an increasing source of concern for world aquaculture in the last decade. This development results from a growing interest in taxonomy, the need * Correspondence and reprints Tel.: +33 1 34 65 25 86 Fax: +33 1 34 65 25 91 [email protected] Abbreviations: AO, Anacker and Ordal (medium); CFU, colony-forming unit; CTC, 5-cyano-2,3-di-4-tolyl-tetrazolium chloride; DiBAC4 [3], bis-(1,3-dibutylbarbituric acid) trimethine oxonol; EAO, enriched Anacker and Ordal (medium); EAOA, EAO agar; EAOB, EAO broth; FCS, foetal calf serum; HS, horse serum; NCIMB, National Collection of Industrial and Marine Bacteria; OD, optical density; PBS, phosphate buffer saline; RTFS, rainbow trout fry syndrome; TYESA, TYES agar; TYESB, TYES broth; TYES, tryptone-yeast extract salts; VNC, viable but non culturable (form).
for accurate diagnosis of fish pathogenic bacteria, and a better understanding of the dynamics of the diseases themselves. The development of intensive production systems, as in the marine environment, has probably caused a wider geographical dissemination of Flavobacterium, prompted by large-scale trade exchanges. This is illustrated by Flavobacterium psychrophilum (formerly Flexibacter psychrophilus or Cytophaga psychrophila), which suddenly spread from northwestern America, where it had long been known [9], to western Europe [22], Japan [23], Chile [6], and Australia [20]. The bacterium has now become one of the most significant salmonid pathogens worldwide. Its very wide impact in Europe originates from a severe systemic clinical form found in young farmed salmonid fry (rainbow trout fry syndrome, or RTFS), which not only causes death, but also requires near-permanent antimicrobial drug administration to be reasonably controlled.
352
Michel et al.
Several culture media (table I) have been described for the cultivation of Flavobacterium species, the most commonly used being that of Anacker and Ordal (AO) [2], also frequently referred to as ‘Cytophaga medium’ [9]. F. psychrophilum was first isolated in 1948 by Borg [5] on a medium containing tryptone 0.5% and agar 0.9%. When produced on AO, the bacterium usually exhibits a relatively slow and fastidious growth. Therefore, reinforcement of tryptone content (enriched AO, or EAO) [3] or use of tryptone-yeast extract-salts (TYES) [10] with added skimmed milk are generally preferred. Although these media appear convenient for isolation and cultivation of the bacterium, several authors who attempted to produce bacteria for experimental purposes had to propose cultural devices or additional enrichment procedures, such as incubation at 15 °C [9], supplementation with foetal calf serum (FCS) [19], or use of 0.7% agar instead of 1% [12]. These repeated modifications might suggest that the performance of EAO and TYES was not optimal each time. Indeed, when trying to control bacterial suspensions of low density through classical colony enumeration procedures, failure or inconsistencies in bacterial viability were generally observed. Accurate viable bacterial counts are necessary for followup studies, such as experimental infection or antimicrobial activity testing. The difficulty with F. psychrophilum may partly explain why most studies have focused on tax-
onomy, diagnosis, or detecting the bacteria and why, up to now, little has been published on the disease pathogenesis and the relationship between the agent and its host [7]. We therefore decided to re-examine the requirements of the bacterium to improve the medium formulation and obtain as constant a production of viable bacteria as possible. The different stages and the resulting improvements described in this paper made it possible to suggest an explanation for the variability in viability counts. We then established a relationship curve between optical density and the number of culturable bacteria in freshly yielded suspensions.
2. Materials and methods 2.1. Culture
The F. psychrophilum strains used in this study all came from rainbow trout but were collected from different places: JIP 14–85, 14–86, and 02–97 were isolated in France, UCD 004–95 in Idaho (USA), and SVS 910611–1 was a Danish strain supplied by H. Lorenzen (National Veterinary Laboratory, Århus, Denmark). JIP 14–85 was used in all tests and considered as a reference; the other strains served to confirm and generalize trends in the results. All stocks were stored for maintenance at 4 °C in EAO containing tryptone 0.5% and agar 0.4%, as advised by Anacker and Ordal [2]. Cultures were revived in EAO broth (EAOB) at 18 °C for 48 to 72 h
Table I. Composition of the media recommended for culture and production of F. psychrophilum, with some of the modifications proposed for culture improvement (pH adjusted to 7.2–7.4). AO Tryptone Yeast extract Beef extract NaCH3CO2, 3H2O (Agar) Water
TYES + skimmed milk 0.5 g 0.5 g 0.2 g 0.2 g 10–12 g 1L
Tryptone Yeast extract CaCl2, 2 H2O MgSO4, 7 H2O (Agar) Water + Skim milk 20 %
Proposed modifications to the basic AO medium Tryptone 5 g / L Foetal calf serum added at 10 % agar concentration 0.7 %
Bernardet and Kerouault [3] Obach and Baudin-Laurencin [19] Lorenzen [12]
4g 0.4 g 0.2 g 0.5 g 10–12 g 1L 1%
Flavobacterium psychrophilum viable culture production
before seeding Erlenmeyer flasks containing 50 mL of EAOB or TYES, which were modified according to the different factors to be tested. These flasks were placed into a cooled orbital incubator (Gallenkampf, France), adjusted to 18°C, and stirred for 24 to 48 h at 80 rev/min before being harvested in medium or late exponential phase. 2.2. Plate enumeration
Absorbance of the bacterial suspensions was measured at 525 nm on a spectrophotometer Camspec M 330 (Kontron, France). Tenfold serial dilutions were prepared in 10-mL tubes before being dispensed onto the dry surface of enumeration agar plates according to the drop method of Miles and Misra [17], and incubated for three days at 18 °C. As in production steps, different compositions of dilution and solid enumeration dilution medium were tested. Colony enumeration by triplicate drops was performed at a magnification of 20 × under a stereoscopic microscope with a transilluminating device at a 45° angle.
353
Table II. Trace element solution for F. psychrophilum (after Lewin and Lounsbery, in [18]. Recommended use at 1 p.1000). Composition for 1 L: H3BO3 MnCl2, 4 H2O FeSO4 Na2C4H4O6, 2H2O CuCl2, 2H2O ZnCl2 CoCl2, 6H2O Na2MoO4, 2H2O
2.85 g 1.80 g 1.36 g 1.77 g 26.9 mg 20.8 mg 40.4 mg 25.2 mg
extemporarily, just before inoculating the media. All the media and additives were usually stored at 4 °C. After harvesting, dilutions steps were conducted in saline water 0.9% (pH 6.65). In some specific cases, phosphate buffer saline (PBS) 0.05 mol/L pH 7.4 and sterile distilled water of standard (pH 6.4) or ultra-pure (pH 7.2) quality were also used to test the survival of bacterial cells. In fact, F. psychrophilum does not use carbohydrates and the pH of 48 h AO cultures never exceeds 7.6. 2.4. Direct enumeration of total and viable cells
2.3. Factors and parameters tested
Two basic media (EAO and TYES) were compared as production medium and enumeration support. Two different concentrations of agar (1% and 0.7%) were tested in EAOA. Two agar qualities, ‘bacteriological’ (Gibco BRL) and ‘washed’ (Sigma), were also tested to eliminate any possible effect noted with other environmental microorganisms, like cyanobacteria [24]. Different combinations of factors were examined to enrich the culture media. These included sterile FCS (Boehringer-Mannheim, Germany) 10%, horse defibrinated serum (HS; SanofiDiagnostic Pasteur, France) 5 or 10%, and several available formulas of mineral element solutions that are normally used in synthetic media for bacteria (i.e., ‘Metals 70’ of Véron [21]). These solutions were tested at different dilutions of between 10–1 and 10–3. One of these, the ‘trace element solution’ (table II), was especially designed for freshwater ‘Flexibacter’ [11]. These additions were carried out
Direct enumeration was performed in PetroffHausser counting chambers (Hausser, Horsham, USA). Bacterial cells were observed through immersion in phase contrast, at a magnification of 1000 ×. To differentiate viable and dead bacteria, different fluorescent products, including bis-(1,3-dibutylbarbituric acid)trimethine oxonol (DiBAC4) (3), Molecular Probes, Oregon), 5-cyano-2,3-di-4-tolyl-tetrazolium chloride (CTC; Polysciences, PA), and the twocolour fluorescence kit Live/Dead BacLight (Molecular Probes Europe, the Netherlands), which combines SYTO-9® and propidium iodide, were tested on fresh cultures diluted in different suspension media. The cells were treated and incubated according to supplier specifications: 20 µL/mL 10 min at room temperature for DiBAC4 (3), 2.5 mmol/L 60 min for CTC, 1.5 µL/mL of each reagent for 15 min at room temperature for Live/Dead BacLight test. Microscopic observation was performed on wet mounts using a 50 × oil-immersion objective
354
Michel et al.
and a Leitz Orthoplan microscope equipped with an H long pass filter (530 µm). To achieve an acceptable level of accuracy, enumeration was performed on 12 to 15 fields chosen randomly on the slide. This usually represented 250–350 cells per slide when the culture was diluted to 10–1.
3. Results 3.1. Oxygenation and serum enrichment
Figure 1 gives details of subsequent trials performed in nonstirred glass tubes tested without or with FCS enrichment, as well as the results obtained from stirring cultures in Erlenmeyer flasks. The basic medium was always EAO and, except in the very early assays (72 h), the bacteria were harvested at 48 h. The extent to which oxygenation and enrichment with FCS can improve the results of F. psychrophilum plate enumeration is clearly illustrated here. 3.2. Medium
The real interest of adding serum was confirmed by further attempts (table III), in which the most commonly used media, EAO and TYES, were compared as both convenient liquid media for mass-producing bacteria and suitable solid media for detecting typical isolated colonies. Counting on EAOA enriched with serum always provided better results and testing the cultures harvested from different liquid media in similar conditions confirmed that serumsupplemented EAO medium yielded better results.
Figure 1. Enumeration of 21 F. psychrophilum cultures grown under different culture conditions, expressed as a percentage of values expected after OD measurement. Culture conditions: 1–5 without serum or stirring; 6–10 with FCS, no stirring; 11–21 with FCS, and stirring in Erlenmeyer flask
3.3. Other factors
Even so, strong variations could still occur, as can be seen from samples 15, 16, and 20 in figure 1. Several additional parameters, namely agar quality and concentration, the type of serum, the concentration of trace elements, and the medium used for dilution were also systematically examined. Table IV summarizes the results from repeated trials. Among the mineral element solutions tested, all but the trace elements solution negatively affected bacterial growth and were discarded. None of these parameters taken separately appeared to be greatly significant. Moderate improvements, however, could often be detected. For reasons set out in the discussion, it was eventually decided to retain horse serum 5% and low levels
Table III. Average numbers of F. psychrophilum colonies enumerated on different plating media (droplets 50 µL, dilution 10-5) after production in corresponding liquid media. TYESA, TYESB: tryptone-yeast extract-salts + skimmed milk (agar and broth). EAOA, EAOB: EAO agar and broth. Enumeration EAOA EAOA + FCS TYEA TYEA + FCS
Production (stirred flasks) EAOB
EAOB + FCS
TYEB
TYEB + FCS
5 48.6 0 0
0.66 84 0 0.33
3.3 21.3 0 0
10.3 16.3 0 0
Flavobacterium psychrophilum viable culture production
355
Table IV. Some parameters affecting F. psychrophilum colony enumeration (expressed as the percentage of CFUs obtained using the less effective formulation, compared to the best one). Medium components Agar quality: concentration: Serum: Trace elements: Dilution Saline 0.9 %/PBS
washed agar/bacteriological agar 0.7 %/1 % FCS/HS no addition/enrichment 0.05 %
78–82 % irregular 84.5 % 84.5 % 87 %
of trace elements as the essential factors to be added to EAO medium in all further studies. 3.4. Direct viable count
Results obtained using fluorescent dyes for viability assessment were irregular. Preliminary trials were conducted with DiBAC4 [3], which did not appear appropriate, as almost all bacteria were stained, and CTC, which provided clear red fluorescent images, but seemed to stain only a very small fraction of the bacterial population. Further use of the Live/Dead BacLight kit provided better results and enabled us to understand that culture was not the only critical step and that F. psychrophilum could be readily inactivated during harvesting, depending on the suspension medium. Suspension in distilled water resulted in red or intermediate fluorescence on almost all the cells. Saline or PBS gave distinctly separate populations of dead red-stained and dominant living greenstained bacteria (table V). Subsequent experiments were greatly facilitated when bacteria were gently handled in isotonic conditions to preserve optimal viability. 3.5. Predictive value of optical density curves
Repeated enumerations on properly enriched EAOA made it possible to draw a theoretical
relation curve (figure 2) between the optical density (OD) of the culture and the number of viable cells expressed as colony-forming units (CFUs), with figures of gradually decreasing deviation. All of our F. psychrophilum isolates were used to carry out this last step and confirmed similar behaviour. The only possible inconvenience was limited cell aggregation the extent of which could vary according to the cultures. Results of direct counting in PetroffHausser chambers are also shown in figure 2, which reveals substantial differences between the culturable cells (CFUs) and the estimated total number of bacteria.
4. Discussion The very first difficulties encountered when trying to produce viable F. psychrophilum suspensions with known physiological characteristics (namely a controlled number of infective units) were caused by unpredictable variations in CFUs. This could result from inadequate medium composition, culture conditions, or intrinsic unstable characteristics of the bacterial species in artificial conditions. We therefore set out to examine these hypotheses. The importance of the medium composition and the possibility of improving standard for-
Table V. Suspension medium and survival of F. psychrophilum: examples of direct enumeration of total and viable cells performed on a 24-h culture of the strain JIP 14-86, diluted to 10-1 in different media and double-stained with the Live/Dead BacLight kit. PBS Number of fields Means per field: total cells viable cells % dead cells %
Saline 0.9 %
Distilled water
15
12
11
14.7 10.6 72.1 4.1 27.9
19.7 13.8 70 5.8 30
21.2 0.3 1.4 20.9 98.6
356
Michel et al.
Figure 2. F. psychrophilum: OD/ CFU(full line) and OD/total cell number (dotted line) correspondences (a = slope; R2 = coefficient of determination).
mulations had already been assumed and reported by several authors. In dish cultures, neither the concentration of agar nor its industrial quality were found to significantly affect quantitative results. Our attempts, however, clearly confirmed the value of serum in medium enrichment, as stated by Obach and BaudinLaurencin [19] and Lorenzen [12]. After several tests, we decided to use defibrinated HS rather than FCS, in view of lower costs, a slightly better performance in the comparative enumeration of cultures, and differences in surface properties conferred to the agar. On FCSenriched media, calibrated droplets had a marked tendency to spread out and fuse, making enumeration more difficult. Conversely, the droplets were always seen to be more homogenous on HS-enriched agar and were more rapidly asborbed. Finally, subsequent assays have shown that using HS at 5% instead of 10% did not alter the results at all. Stirring promoted proper oxygenation, which apparently improved the culturability and culture density. None of the other parameters produced major improvements in culture enumeration. Several factors, however, such as incorporation of mineral element traces, the use of bacteriological agar at standard concentration, and the use of buffered medium to harvest and process the cultures, combined to produce optimal results. This was instrumental in trying to
adjust culture conditions and obtain reproducible measurements. Supplementation with trace elements solution needs further comment. Its composition was found in the NCIMB catalogue [18], based on a descriptive paper of Lewin and Lounsbery [11], but the exact nature of the salts to be used was not specified in this primary paper. In our trials, enrichment of the liquid production media or agar enumeration dishes with trace solution seemed to constantly improve the number of recorded CFUs when used at low levels, i.e., below 0.05%. Above that threshold, the effect was less constant and could even appear negative. When both production and enumeration medium were enriched, the results were considerably weaker than those of controls. Presumably, one or several components in the formula bordered on ratios inhibitory for F. psychrophilum. In the absence of further data, and to avoid any problems, mineral enrichment was only provided for production media. The bacterial cells were considered to incorporate enough essential factors to secure their needs for the subsequent enumeration step, yet this subtle balance of mineral requirements is likely to create concern after several subcultures. Perhaps this is an alternative explanation for failures experienced after successful primary isolation that were attributed to various intrinsic properties of the strains [12]. The existence of viable but nonculturable (VNC) forms in bacterial populations has been suspected for about 15 years. These forms have been well documented in Gram-negative species [14] and have even been reported among fish pathogens such as Aeromonas salmonicida [1] and Photobacterium damselae subsp. piscicida [15]. Although the exact interpretation is still controversial, the most common view is that the phenomenon would represent an adaptative property particularly expressed when the bacteria are submitted to suboptimal conditions of survival, as is often the case in the natural water environment. In some examples, a relationship between the appearance of VNC forms and varying stress endured by the bacteria could be assessed unequivocally [8]. Whether such dor-
Flavobacterium psychrophilum viable culture production
mant cells are likely to be revived when better conditions are again met or have simply entered the first stages of an irreversible deterioration process is a matter for disagreement that could centre on the species and the experimental context [13]. The formation of such VNC cells in F. psychrophilum might be assumed from the inconstant production of CFUs from one culture to another. Later observations confirmed that VNC cells were likely to occur, namely when the right proportions of nonculturable (about 75%; figure 2) and dead cells (about 30%; table V) in cultures produced under controlled conditions were compared. Here, however, neither the culturable nor the nonculturable fraction varied dramatically, and it appeared from our assays with the Live/Dead BacLight kit that the high susceptibility of the bacteria to stressful conditions, rather than the presence of VNC cells, could account for the large initial variability in CFU enumeration. The properties of the three tested fluorescent dyes are quite different. DiBAC4 (3) is an indicator of membrane potential perturbation, CTC specifically reveals respiratory activity, and the results obtained with the Live/Dead BacLight kit depend on the membrane permeability. In our tests, only distilled water proved stressful, which suggests that in a hypotonic environment, the cell membrane was readily impaired, enabling the penetration of the propidium iodide and its linking to DNA. The suspension of bacteria in distilled water is often specified in commercial kits or reagent notices, as the cell wall of these organisms is generally supposed to preserve them from harmful environments, at least temporarily. No special suspension medium was recommended for Live/Dead Bac Light kit, but the instructions prescribed several washings and centrifugations in ‘water’. Whatever the medium used, following this procedure always resulted in uniformly red-stained bacterial cells, as did direct dilution in distilled water. Conversely, dilution in saline or PBS and immediate observation made it possible to observe a large majority of cells stained by the green SYTO-9 dye. We think that using isotonic sus-
357
pension media and gentle handling, while avoiding strong stirring, vortexing, and centrifugations as much as possible, are critical for obtaining reliable results from F. psychrophilum viable cell enumerations. By taking all these different factors and parameters into account, we could control the conditions in which bacterial suspensions were produced and obtain more reproducible cell counts on agar medium. We could also establish a clear linear relationship between culture density and the number of culturable bacteria, even though these cells, when compared to direct microscopic counts, were well below the total number of cells actually present. This ratio, approximately 25%, falls into the order of magnitude reported for other environmentassociated bacteria [16]. The most significant feature is that the relative proportion of culturable cells may persist, as long as the culture does not suffer nutrient depletion. This makes possible a quick estimation of the bacterial inoculum after spectrophotometric measures. Although some degree of cell aggregation may result in larger variations when compared to more classical bacteria such as Vibrionaceae or Enterobacteriaceae, our subsequent assays were consistent. In summary, we i) confirmed the usefulness of serum as a major enrichment factor in F. psychrophilum culture, ii) indicated possible requirements in mineral salts, and iii) identifed practical conditions for optimal culture performance, free of fear of interference by VNC forms. We trust that the present observations and suggestions are valuable for further improvements in F. psychrophilum production and that they may be useful for devising, testing, and standardizing experimental procedures that require well-adjusted doses of bacteria. These recommendations should thus help to optimise experimental models of infection and tests for assessing the in vitro susceptibility of the bacterium to drugs or fish defence factors. Résumé — Production de cultures viables de Flavobacterium psychrophilum : approche et contrôle.
358
Michel et al.
Flavobacterium psychrophilum est un agent infectieux majeur des élevages de Salmonidés dont le pouvoir pathogène reste peu étudié en raison des difficultés rencontrées dans la production de cultures standardisées. Les dénombrements comparés, par titrage sur milieu gélosé ou par comptage direct en microscopie (éventuellement associée à l’épifluorescence), permettent rarement de détecter plus de 25 % de bactéries vivantes et capables de former des colonies dans les cultures réalisées en conditions usuelles. Cette proportion dépend en fait de la composition des milieux. Les meilleurs résultats ont été obtenus avec le milieu d’Anacker et Ordal enrichi en tryptone plutôt qu’avec le milieu «tryptone/extrait de levure/sels minéraux» additionné de lait écrémé ; et l’ajout de sérum de cheval et d’éléments minéraux en faible proportion a permis d’améliorer nettement leur reproductibilité. Plutôt qu’à l’intervention de formes viables mais non cultivables les difficultés de culture sont imputables à une sensibilité très marquée de F. psychrophilum aux conditions osmotiques. L’optimisation des cultures, jointe à des manipulations précautionneuses en milieu isotonique, a conduit à des résultats reproductibles et prédictibles, matérialisés par l’établissement d’une courbe de relation entre bactéries cultivables et densité optique. © Elsevier, Paris Flavobacterium psychrophilum / croissance bactérienne / nutrition microbienne / viabilité
Acknowledgments The authors are indebted to Dr J.-F. Bernardet (Jouy-en-Josas), whose experience and remarks helped to complete this work, and who provided valuable criticisms on the manuscript, and to Dr Rosemarie Rippka (Pasteur Institute, Paris) for fruitful discussions about environmental bacteria requirements. A large part of this study was conducted at the Fish Disease Laboratory, UC Davis, during a study leave generously granted by the Direction des Relations internationales of INRA.
References [1] Allen-Austin D., Austin B., Collwell R.R., Survival of Aeromonas salmonicida in river water, FEMS Microbiol. Lett. 21 (1984) 143–146. [2] Anacker R.L., Ordal E.J., Study of a bacteriophage infecting the myxobacterium Chondrococcus columnaris, J. Bacteriol. 70 (1955) 738–741.
[3] Bernardet J.F., Kerouault B., Phenotypic and genomic studies of ‘Cytophaga psychrophila’ isolated from diseased rainbow trout (Oncorhynchus mykiss) in France, Appl. Environ. Microbiol. 55 (1989) 1796–1800. [4] Bernardet J.F., Segers P., Vancanneyt M., Berthe F., Kersters K., Vandamme P., Cutting a Gordian knot: emended classification and description of the genus Flavobacterium, emended description of the family Flavobacteriaceae, and proposal of Flavobacterium hydatis nom. nov. (basonym, Cytophaga aquatilis Strohl and Tait 1978, Int. J. Syst. Bacteriol. 46 (1996) 128–148. [5] Borg A.F., Studies on myxobacteria associated with diseases in salmonid fishes, J. Wild. Dis. (microfiche) 8 (1960) 1–85 (2 microcards). [6] Bustos P., Calbuyahue A., Montaña J., Opazo B., Entrala P., Solervicens R., First isolation of Flexibacter psychrophilus, as causative agent of rainbow trout fry syndrome (RTFS), producing trout mortality in Chile, Bull. Eur. Ass. Fish Pathol. 15 (1995) 162–164. [7] Dalsgaard I., Virulence mechanisms in Cytophaga psychrophila and other Cytophaga-like bacteria pathogenic for fish, Ann. Rev. Fish Dis. (1993) 127–144. [8] Herson D.S., Marthi B., Payer M.A., Baker K.H., Enumeration of chlorine-stressed organisms with acridine orange 2-(piodophenyl)-3-(p-nitrophenyl)-5-phenyltetrazolium chloride (AOINT), Curr. Microbiol. 13 (1986) 77–80. [9] Holt R.A., Cytophaga psychrophila the causative agent of bacterial coldwater disease in salmonid fish, Ph.D. Thesis, Oregon State University, Corvallis, Oregon, 1987. [10] Holt R.A., Rohovec J.S., Fryer J.L., Bacterial coldwater disease, in: Inglis V., Roberts R.J., Bromage N.R. (Eds.) Bacterial Diseases of Fish Blackwell Scientific Publications, Oxford, , 1993, pp. 3–22. [11] Lewin R.A., Lounsbery D.M., Isolation, cultivation and characterization of flexibacteria, J. Gen. Microbiol. 58 (1969) 145–170. [12] Lorenzen E., The importance of the brand of the beef extract in relation to the growth of Flexibacter psychrophilus in Anacker & Ordal medium, Bull. Eur. Ass. Fish Pathol. 13 (1993) 64–65. [13] McDoulagd D., Rice S.A., Weichart D., Kjellerberg S., Nonculturability: adaptation or debilitation? FEMS Microbiol. Ecol. 25 (1998) 1–9. [14] McKay A.M., Viable but non-culturable forms of potentially pathogenic bacteria in water, Appl. Microbiol. 14 (1992) 129–135. [15] Magariños B., Romalde J.L., Barja J.L., Toranzo A.E., Evidence of a dormant but infective state of the fish pathogen Pasteurella piscicida in seawater and sediment, Appl. Environ. Microbiol. 60 (1994) 180–186. [16] Mason C.A., Hamer G., Bryers, J.D., The death and lysis of microorganisms in environmental processes, FEMS Microbiol. Rev. 39 (1986) 373–401. [17] Miles A.A., Misra S.S., The estimation of the bacteriocidal power of blood, J. Hyg. 38 (1938) 732–749. [18] Dando T.R., Young J.E.P. (Eds.), Flexibacter medium, National Collection of Industrial and Marine Bacteria, Catalogue of strains, ref. 281, 1990, pp. 177. [19] Obach A., Baudin-Laurencin F., Vaccination of rainbow trout (Oncorhynchus mykiss) against the visceral form of cold-water disease, Dis. Aquat. Org. 12 (1991) 13–15. [20] Schmidtke L.M., Carson, J., Characteristics of Flexibacter psychrophilus isolated from Atlantic salmon in Australia, Dis. Aquat. Org. 21 (1995) 157–161. [21] Véron M., Nutrition et taxonomie des Enterobacteriaceae et bactéries voisines. I. Méthode d’étude des auxanogrammes, Ann. Inst. Pasteur/Microbiol. 126A (1975) 267–274. [22] von Weis J., Über das Vorkommen einer Kaltwasserkrankheit bei Regenbogenforellen Salmo gairdneri, Tierärzl. Umschau 42 (1987) 575–577. [23] Wakabayashi H., Horiuchi M., Bunya T., Hoshiai G., Outbreaks of cold-water disease in coho salmon in Japan, Fish Pathol. 26 (1991) 211–212. [24] Waterbury J.B., Willey J.M., Isolation and growth of marine planktonic Cyanobacteria, Methods Enzymol. 167 (1988) 100–105.
Production of viable cultures of Flavobacterium psychrophilum: approach and control Christian Michela*, Dolorès Antoniob, Ronald P. Hedrickb a
Unité de virologie et immunologie moléculaires, E´quipe de pathologie infectieuse et immunité des poissons, Inra-CRJJ, Domaine de Vilvert, Centre de recherches de Jouy-en-Josas, 78352 Jouy-En-Josas cedex, France b Department of Medicine and Epidemiology, School of Veterinary Medicine, University of California at Davis, Davis, CA 95616, USA (Submitted 12 September 1998; accepted 10 March 1999)
Abstract — Although the fish pathogen Flavobacterium psychrophilum is a major source of concern in salmonid hatcheries, few studies have been conducted on its pathogenicity. Difficulties are often experienced when trying to control or quantify standard procedures for in vitro culture of the bacterium. Plate enumeration and counting chamber enumeration combined with epifluorescent microscopy with fluorescent dyes determined that no more than 25% of the bacterial cells present in the cultures were able to produce colonies on agar media. This was strongly dependent upon different medium components. Tryptone-enriched Anacker and Ordal medium proved more suitable than tryptone-yeast extract-salts with skimmed milk. Adding horse serum and trace elements in controlled proportions offered the most reproducible results. Viable but nonculturable forms were apparently not responsible for the difficulties in production of F. psychrophilum, but the cells were highly susceptible to osmotic conditions. Improvements in the media and careful handling of the bacteria in isotonic suspension media resulted in predictable production of viable bacteria and allowed an absorbance/colony-forming-units relation curve to be established. © Elsevier, Paris Flavobacterium psychrophilum / bacterial growth / microbial nutrition / viability
1. Introduction Fish pathogenic Flavobacterium species, most of which were formerly ascribed to the genera Flexibacter and Cytophaga [4], have become an increasing source of concern for world aquaculture in the last decade. This development results from a growing interest in taxonomy, the need * Correspondence and reprints Tel.: +33 1 34 65 25 86 Fax: +33 1 34 65 25 91 [email protected] Abbreviations: AO, Anacker and Ordal (medium); CFU, colony-forming unit; CTC, 5-cyano-2,3-di-4-tolyl-tetrazolium chloride; DiBAC4 [3], bis-(1,3-dibutylbarbituric acid) trimethine oxonol; EAO, enriched Anacker and Ordal (medium); EAOA, EAO agar; EAOB, EAO broth; FCS, foetal calf serum; HS, horse serum; NCIMB, National Collection of Industrial and Marine Bacteria; OD, optical density; PBS, phosphate buffer saline; RTFS, rainbow trout fry syndrome; TYESA, TYES agar; TYESB, TYES broth; TYES, tryptone-yeast extract salts; VNC, viable but non culturable (form).
for accurate diagnosis of fish pathogenic bacteria, and a better understanding of the dynamics of the diseases themselves. The development of intensive production systems, as in the marine environment, has probably caused a wider geographical dissemination of Flavobacterium, prompted by large-scale trade exchanges. This is illustrated by Flavobacterium psychrophilum (formerly Flexibacter psychrophilus or Cytophaga psychrophila), which suddenly spread from northwestern America, where it had long been known [9], to western Europe [22], Japan [23], Chile [6], and Australia [20]. The bacterium has now become one of the most significant salmonid pathogens worldwide. Its very wide impact in Europe originates from a severe systemic clinical form found in young farmed salmonid fry (rainbow trout fry syndrome, or RTFS), which not only causes death, but also requires near-permanent antimicrobial drug administration to be reasonably controlled.
352
Michel et al.
Several culture media (table I) have been described for the cultivation of Flavobacterium species, the most commonly used being that of Anacker and Ordal (AO) [2], also frequently referred to as ‘Cytophaga medium’ [9]. F. psychrophilum was first isolated in 1948 by Borg [5] on a medium containing tryptone 0.5% and agar 0.9%. When produced on AO, the bacterium usually exhibits a relatively slow and fastidious growth. Therefore, reinforcement of tryptone content (enriched AO, or EAO) [3] or use of tryptone-yeast extract-salts (TYES) [10] with added skimmed milk are generally preferred. Although these media appear convenient for isolation and cultivation of the bacterium, several authors who attempted to produce bacteria for experimental purposes had to propose cultural devices or additional enrichment procedures, such as incubation at 15 °C [9], supplementation with foetal calf serum (FCS) [19], or use of 0.7% agar instead of 1% [12]. These repeated modifications might suggest that the performance of EAO and TYES was not optimal each time. Indeed, when trying to control bacterial suspensions of low density through classical colony enumeration procedures, failure or inconsistencies in bacterial viability were generally observed. Accurate viable bacterial counts are necessary for followup studies, such as experimental infection or antimicrobial activity testing. The difficulty with F. psychrophilum may partly explain why most studies have focused on tax-
onomy, diagnosis, or detecting the bacteria and why, up to now, little has been published on the disease pathogenesis and the relationship between the agent and its host [7]. We therefore decided to re-examine the requirements of the bacterium to improve the medium formulation and obtain as constant a production of viable bacteria as possible. The different stages and the resulting improvements described in this paper made it possible to suggest an explanation for the variability in viability counts. We then established a relationship curve between optical density and the number of culturable bacteria in freshly yielded suspensions.
2. Materials and methods 2.1. Culture
The F. psychrophilum strains used in this study all came from rainbow trout but were collected from different places: JIP 14–85, 14–86, and 02–97 were isolated in France, UCD 004–95 in Idaho (USA), and SVS 910611–1 was a Danish strain supplied by H. Lorenzen (National Veterinary Laboratory, Århus, Denmark). JIP 14–85 was used in all tests and considered as a reference; the other strains served to confirm and generalize trends in the results. All stocks were stored for maintenance at 4 °C in EAO containing tryptone 0.5% and agar 0.4%, as advised by Anacker and Ordal [2]. Cultures were revived in EAO broth (EAOB) at 18 °C for 48 to 72 h
Table I. Composition of the media recommended for culture and production of F. psychrophilum, with some of the modifications proposed for culture improvement (pH adjusted to 7.2–7.4). AO Tryptone Yeast extract Beef extract NaCH3CO2, 3H2O (Agar) Water
TYES + skimmed milk 0.5 g 0.5 g 0.2 g 0.2 g 10–12 g 1L
Tryptone Yeast extract CaCl2, 2 H2O MgSO4, 7 H2O (Agar) Water + Skim milk 20 %
Proposed modifications to the basic AO medium Tryptone 5 g / L Foetal calf serum added at 10 % agar concentration 0.7 %
Bernardet and Kerouault [3] Obach and Baudin-Laurencin [19] Lorenzen [12]
4g 0.4 g 0.2 g 0.5 g 10–12 g 1L 1%
Flavobacterium psychrophilum viable culture production
before seeding Erlenmeyer flasks containing 50 mL of EAOB or TYES, which were modified according to the different factors to be tested. These flasks were placed into a cooled orbital incubator (Gallenkampf, France), adjusted to 18°C, and stirred for 24 to 48 h at 80 rev/min before being harvested in medium or late exponential phase. 2.2. Plate enumeration
Absorbance of the bacterial suspensions was measured at 525 nm on a spectrophotometer Camspec M 330 (Kontron, France). Tenfold serial dilutions were prepared in 10-mL tubes before being dispensed onto the dry surface of enumeration agar plates according to the drop method of Miles and Misra [17], and incubated for three days at 18 °C. As in production steps, different compositions of dilution and solid enumeration dilution medium were tested. Colony enumeration by triplicate drops was performed at a magnification of 20 × under a stereoscopic microscope with a transilluminating device at a 45° angle.
353
Table II. Trace element solution for F. psychrophilum (after Lewin and Lounsbery, in [18]. Recommended use at 1 p.1000). Composition for 1 L: H3BO3 MnCl2, 4 H2O FeSO4 Na2C4H4O6, 2H2O CuCl2, 2H2O ZnCl2 CoCl2, 6H2O Na2MoO4, 2H2O
2.85 g 1.80 g 1.36 g 1.77 g 26.9 mg 20.8 mg 40.4 mg 25.2 mg
extemporarily, just before inoculating the media. All the media and additives were usually stored at 4 °C. After harvesting, dilutions steps were conducted in saline water 0.9% (pH 6.65). In some specific cases, phosphate buffer saline (PBS) 0.05 mol/L pH 7.4 and sterile distilled water of standard (pH 6.4) or ultra-pure (pH 7.2) quality were also used to test the survival of bacterial cells. In fact, F. psychrophilum does not use carbohydrates and the pH of 48 h AO cultures never exceeds 7.6. 2.4. Direct enumeration of total and viable cells
2.3. Factors and parameters tested
Two basic media (EAO and TYES) were compared as production medium and enumeration support. Two different concentrations of agar (1% and 0.7%) were tested in EAOA. Two agar qualities, ‘bacteriological’ (Gibco BRL) and ‘washed’ (Sigma), were also tested to eliminate any possible effect noted with other environmental microorganisms, like cyanobacteria [24]. Different combinations of factors were examined to enrich the culture media. These included sterile FCS (Boehringer-Mannheim, Germany) 10%, horse defibrinated serum (HS; SanofiDiagnostic Pasteur, France) 5 or 10%, and several available formulas of mineral element solutions that are normally used in synthetic media for bacteria (i.e., ‘Metals 70’ of Véron [21]). These solutions were tested at different dilutions of between 10–1 and 10–3. One of these, the ‘trace element solution’ (table II), was especially designed for freshwater ‘Flexibacter’ [11]. These additions were carried out
Direct enumeration was performed in PetroffHausser counting chambers (Hausser, Horsham, USA). Bacterial cells were observed through immersion in phase contrast, at a magnification of 1000 ×. To differentiate viable and dead bacteria, different fluorescent products, including bis-(1,3-dibutylbarbituric acid)trimethine oxonol (DiBAC4) (3), Molecular Probes, Oregon), 5-cyano-2,3-di-4-tolyl-tetrazolium chloride (CTC; Polysciences, PA), and the twocolour fluorescence kit Live/Dead BacLight (Molecular Probes Europe, the Netherlands), which combines SYTO-9® and propidium iodide, were tested on fresh cultures diluted in different suspension media. The cells were treated and incubated according to supplier specifications: 20 µL/mL 10 min at room temperature for DiBAC4 (3), 2.5 mmol/L 60 min for CTC, 1.5 µL/mL of each reagent for 15 min at room temperature for Live/Dead BacLight test. Microscopic observation was performed on wet mounts using a 50 × oil-immersion objective
354
Michel et al.
and a Leitz Orthoplan microscope equipped with an H long pass filter (530 µm). To achieve an acceptable level of accuracy, enumeration was performed on 12 to 15 fields chosen randomly on the slide. This usually represented 250–350 cells per slide when the culture was diluted to 10–1.
3. Results 3.1. Oxygenation and serum enrichment
Figure 1 gives details of subsequent trials performed in nonstirred glass tubes tested without or with FCS enrichment, as well as the results obtained from stirring cultures in Erlenmeyer flasks. The basic medium was always EAO and, except in the very early assays (72 h), the bacteria were harvested at 48 h. The extent to which oxygenation and enrichment with FCS can improve the results of F. psychrophilum plate enumeration is clearly illustrated here. 3.2. Medium
The real interest of adding serum was confirmed by further attempts (table III), in which the most commonly used media, EAO and TYES, were compared as both convenient liquid media for mass-producing bacteria and suitable solid media for detecting typical isolated colonies. Counting on EAOA enriched with serum always provided better results and testing the cultures harvested from different liquid media in similar conditions confirmed that serumsupplemented EAO medium yielded better results.
Figure 1. Enumeration of 21 F. psychrophilum cultures grown under different culture conditions, expressed as a percentage of values expected after OD measurement. Culture conditions: 1–5 without serum or stirring; 6–10 with FCS, no stirring; 11–21 with FCS, and stirring in Erlenmeyer flask
3.3. Other factors
Even so, strong variations could still occur, as can be seen from samples 15, 16, and 20 in figure 1. Several additional parameters, namely agar quality and concentration, the type of serum, the concentration of trace elements, and the medium used for dilution were also systematically examined. Table IV summarizes the results from repeated trials. Among the mineral element solutions tested, all but the trace elements solution negatively affected bacterial growth and were discarded. None of these parameters taken separately appeared to be greatly significant. Moderate improvements, however, could often be detected. For reasons set out in the discussion, it was eventually decided to retain horse serum 5% and low levels
Table III. Average numbers of F. psychrophilum colonies enumerated on different plating media (droplets 50 µL, dilution 10-5) after production in corresponding liquid media. TYESA, TYESB: tryptone-yeast extract-salts + skimmed milk (agar and broth). EAOA, EAOB: EAO agar and broth. Enumeration EAOA EAOA + FCS TYEA TYEA + FCS
Production (stirred flasks) EAOB
EAOB + FCS
TYEB
TYEB + FCS
5 48.6 0 0
0.66 84 0 0.33
3.3 21.3 0 0
10.3 16.3 0 0
Flavobacterium psychrophilum viable culture production
355
Table IV. Some parameters affecting F. psychrophilum colony enumeration (expressed as the percentage of CFUs obtained using the less effective formulation, compared to the best one). Medium components Agar quality: concentration: Serum: Trace elements: Dilution Saline 0.9 %/PBS
washed agar/bacteriological agar 0.7 %/1 % FCS/HS no addition/enrichment 0.05 %
78–82 % irregular 84.5 % 84.5 % 87 %
of trace elements as the essential factors to be added to EAO medium in all further studies. 3.4. Direct viable count
Results obtained using fluorescent dyes for viability assessment were irregular. Preliminary trials were conducted with DiBAC4 [3], which did not appear appropriate, as almost all bacteria were stained, and CTC, which provided clear red fluorescent images, but seemed to stain only a very small fraction of the bacterial population. Further use of the Live/Dead BacLight kit provided better results and enabled us to understand that culture was not the only critical step and that F. psychrophilum could be readily inactivated during harvesting, depending on the suspension medium. Suspension in distilled water resulted in red or intermediate fluorescence on almost all the cells. Saline or PBS gave distinctly separate populations of dead red-stained and dominant living greenstained bacteria (table V). Subsequent experiments were greatly facilitated when bacteria were gently handled in isotonic conditions to preserve optimal viability. 3.5. Predictive value of optical density curves
Repeated enumerations on properly enriched EAOA made it possible to draw a theoretical
relation curve (figure 2) between the optical density (OD) of the culture and the number of viable cells expressed as colony-forming units (CFUs), with figures of gradually decreasing deviation. All of our F. psychrophilum isolates were used to carry out this last step and confirmed similar behaviour. The only possible inconvenience was limited cell aggregation the extent of which could vary according to the cultures. Results of direct counting in PetroffHausser chambers are also shown in figure 2, which reveals substantial differences between the culturable cells (CFUs) and the estimated total number of bacteria.
4. Discussion The very first difficulties encountered when trying to produce viable F. psychrophilum suspensions with known physiological characteristics (namely a controlled number of infective units) were caused by unpredictable variations in CFUs. This could result from inadequate medium composition, culture conditions, or intrinsic unstable characteristics of the bacterial species in artificial conditions. We therefore set out to examine these hypotheses. The importance of the medium composition and the possibility of improving standard for-
Table V. Suspension medium and survival of F. psychrophilum: examples of direct enumeration of total and viable cells performed on a 24-h culture of the strain JIP 14-86, diluted to 10-1 in different media and double-stained with the Live/Dead BacLight kit. PBS Number of fields Means per field: total cells viable cells % dead cells %
Saline 0.9 %
Distilled water
15
12
11
14.7 10.6 72.1 4.1 27.9
19.7 13.8 70 5.8 30
21.2 0.3 1.4 20.9 98.6
356
Michel et al.
Figure 2. F. psychrophilum: OD/ CFU(full line) and OD/total cell number (dotted line) correspondences (a = slope; R2 = coefficient of determination).
mulations had already been assumed and reported by several authors. In dish cultures, neither the concentration of agar nor its industrial quality were found to significantly affect quantitative results. Our attempts, however, clearly confirmed the value of serum in medium enrichment, as stated by Obach and BaudinLaurencin [19] and Lorenzen [12]. After several tests, we decided to use defibrinated HS rather than FCS, in view of lower costs, a slightly better performance in the comparative enumeration of cultures, and differences in surface properties conferred to the agar. On FCSenriched media, calibrated droplets had a marked tendency to spread out and fuse, making enumeration more difficult. Conversely, the droplets were always seen to be more homogenous on HS-enriched agar and were more rapidly asborbed. Finally, subsequent assays have shown that using HS at 5% instead of 10% did not alter the results at all. Stirring promoted proper oxygenation, which apparently improved the culturability and culture density. None of the other parameters produced major improvements in culture enumeration. Several factors, however, such as incorporation of mineral element traces, the use of bacteriological agar at standard concentration, and the use of buffered medium to harvest and process the cultures, combined to produce optimal results. This was instrumental in trying to
adjust culture conditions and obtain reproducible measurements. Supplementation with trace elements solution needs further comment. Its composition was found in the NCIMB catalogue [18], based on a descriptive paper of Lewin and Lounsbery [11], but the exact nature of the salts to be used was not specified in this primary paper. In our trials, enrichment of the liquid production media or agar enumeration dishes with trace solution seemed to constantly improve the number of recorded CFUs when used at low levels, i.e., below 0.05%. Above that threshold, the effect was less constant and could even appear negative. When both production and enumeration medium were enriched, the results were considerably weaker than those of controls. Presumably, one or several components in the formula bordered on ratios inhibitory for F. psychrophilum. In the absence of further data, and to avoid any problems, mineral enrichment was only provided for production media. The bacterial cells were considered to incorporate enough essential factors to secure their needs for the subsequent enumeration step, yet this subtle balance of mineral requirements is likely to create concern after several subcultures. Perhaps this is an alternative explanation for failures experienced after successful primary isolation that were attributed to various intrinsic properties of the strains [12]. The existence of viable but nonculturable (VNC) forms in bacterial populations has been suspected for about 15 years. These forms have been well documented in Gram-negative species [14] and have even been reported among fish pathogens such as Aeromonas salmonicida [1] and Photobacterium damselae subsp. piscicida [15]. Although the exact interpretation is still controversial, the most common view is that the phenomenon would represent an adaptative property particularly expressed when the bacteria are submitted to suboptimal conditions of survival, as is often the case in the natural water environment. In some examples, a relationship between the appearance of VNC forms and varying stress endured by the bacteria could be assessed unequivocally [8]. Whether such dor-
Flavobacterium psychrophilum viable culture production
mant cells are likely to be revived when better conditions are again met or have simply entered the first stages of an irreversible deterioration process is a matter for disagreement that could centre on the species and the experimental context [13]. The formation of such VNC cells in F. psychrophilum might be assumed from the inconstant production of CFUs from one culture to another. Later observations confirmed that VNC cells were likely to occur, namely when the right proportions of nonculturable (about 75%; figure 2) and dead cells (about 30%; table V) in cultures produced under controlled conditions were compared. Here, however, neither the culturable nor the nonculturable fraction varied dramatically, and it appeared from our assays with the Live/Dead BacLight kit that the high susceptibility of the bacteria to stressful conditions, rather than the presence of VNC cells, could account for the large initial variability in CFU enumeration. The properties of the three tested fluorescent dyes are quite different. DiBAC4 (3) is an indicator of membrane potential perturbation, CTC specifically reveals respiratory activity, and the results obtained with the Live/Dead BacLight kit depend on the membrane permeability. In our tests, only distilled water proved stressful, which suggests that in a hypotonic environment, the cell membrane was readily impaired, enabling the penetration of the propidium iodide and its linking to DNA. The suspension of bacteria in distilled water is often specified in commercial kits or reagent notices, as the cell wall of these organisms is generally supposed to preserve them from harmful environments, at least temporarily. No special suspension medium was recommended for Live/Dead Bac Light kit, but the instructions prescribed several washings and centrifugations in ‘water’. Whatever the medium used, following this procedure always resulted in uniformly red-stained bacterial cells, as did direct dilution in distilled water. Conversely, dilution in saline or PBS and immediate observation made it possible to observe a large majority of cells stained by the green SYTO-9 dye. We think that using isotonic sus-
357
pension media and gentle handling, while avoiding strong stirring, vortexing, and centrifugations as much as possible, are critical for obtaining reliable results from F. psychrophilum viable cell enumerations. By taking all these different factors and parameters into account, we could control the conditions in which bacterial suspensions were produced and obtain more reproducible cell counts on agar medium. We could also establish a clear linear relationship between culture density and the number of culturable bacteria, even though these cells, when compared to direct microscopic counts, were well below the total number of cells actually present. This ratio, approximately 25%, falls into the order of magnitude reported for other environmentassociated bacteria [16]. The most significant feature is that the relative proportion of culturable cells may persist, as long as the culture does not suffer nutrient depletion. This makes possible a quick estimation of the bacterial inoculum after spectrophotometric measures. Although some degree of cell aggregation may result in larger variations when compared to more classical bacteria such as Vibrionaceae or Enterobacteriaceae, our subsequent assays were consistent. In summary, we i) confirmed the usefulness of serum as a major enrichment factor in F. psychrophilum culture, ii) indicated possible requirements in mineral salts, and iii) identifed practical conditions for optimal culture performance, free of fear of interference by VNC forms. We trust that the present observations and suggestions are valuable for further improvements in F. psychrophilum production and that they may be useful for devising, testing, and standardizing experimental procedures that require well-adjusted doses of bacteria. These recommendations should thus help to optimise experimental models of infection and tests for assessing the in vitro susceptibility of the bacterium to drugs or fish defence factors. Résumé — Production de cultures viables de Flavobacterium psychrophilum : approche et contrôle.
358
Michel et al.
Flavobacterium psychrophilum est un agent infectieux majeur des élevages de Salmonidés dont le pouvoir pathogène reste peu étudié en raison des difficultés rencontrées dans la production de cultures standardisées. Les dénombrements comparés, par titrage sur milieu gélosé ou par comptage direct en microscopie (éventuellement associée à l’épifluorescence), permettent rarement de détecter plus de 25 % de bactéries vivantes et capables de former des colonies dans les cultures réalisées en conditions usuelles. Cette proportion dépend en fait de la composition des milieux. Les meilleurs résultats ont été obtenus avec le milieu d’Anacker et Ordal enrichi en tryptone plutôt qu’avec le milieu «tryptone/extrait de levure/sels minéraux» additionné de lait écrémé ; et l’ajout de sérum de cheval et d’éléments minéraux en faible proportion a permis d’améliorer nettement leur reproductibilité. Plutôt qu’à l’intervention de formes viables mais non cultivables les difficultés de culture sont imputables à une sensibilité très marquée de F. psychrophilum aux conditions osmotiques. L’optimisation des cultures, jointe à des manipulations précautionneuses en milieu isotonique, a conduit à des résultats reproductibles et prédictibles, matérialisés par l’établissement d’une courbe de relation entre bactéries cultivables et densité optique. © Elsevier, Paris Flavobacterium psychrophilum / croissance bactérienne / nutrition microbienne / viabilité
Acknowledgments The authors are indebted to Dr J.-F. Bernardet (Jouy-en-Josas), whose experience and remarks helped to complete this work, and who provided valuable criticisms on the manuscript, and to Dr Rosemarie Rippka (Pasteur Institute, Paris) for fruitful discussions about environmental bacteria requirements. A large part of this study was conducted at the Fish Disease Laboratory, UC Davis, during a study leave generously granted by the Direction des Relations internationales of INRA.
References [1] Allen-Austin D., Austin B., Collwell R.R., Survival of Aeromonas salmonicida in river water, FEMS Microbiol. Lett. 21 (1984) 143–146. [2] Anacker R.L., Ordal E.J., Study of a bacteriophage infecting the myxobacterium Chondrococcus columnaris, J. Bacteriol. 70 (1955) 738–741.
[3] Bernardet J.F., Kerouault B., Phenotypic and genomic studies of ‘Cytophaga psychrophila’ isolated from diseased rainbow trout (Oncorhynchus mykiss) in France, Appl. Environ. Microbiol. 55 (1989) 1796–1800. [4] Bernardet J.F., Segers P., Vancanneyt M., Berthe F., Kersters K., Vandamme P., Cutting a Gordian knot: emended classification and description of the genus Flavobacterium, emended description of the family Flavobacteriaceae, and proposal of Flavobacterium hydatis nom. nov. (basonym, Cytophaga aquatilis Strohl and Tait 1978, Int. J. Syst. Bacteriol. 46 (1996) 128–148. [5] Borg A.F., Studies on myxobacteria associated with diseases in salmonid fishes, J. Wild. Dis. (microfiche) 8 (1960) 1–85 (2 microcards). [6] Bustos P., Calbuyahue A., Montaña J., Opazo B., Entrala P., Solervicens R., First isolation of Flexibacter psychrophilus, as causative agent of rainbow trout fry syndrome (RTFS), producing trout mortality in Chile, Bull. Eur. Ass. Fish Pathol. 15 (1995) 162–164. [7] Dalsgaard I., Virulence mechanisms in Cytophaga psychrophila and other Cytophaga-like bacteria pathogenic for fish, Ann. Rev. Fish Dis. (1993) 127–144. [8] Herson D.S., Marthi B., Payer M.A., Baker K.H., Enumeration of chlorine-stressed organisms with acridine orange 2-(piodophenyl)-3-(p-nitrophenyl)-5-phenyltetrazolium chloride (AOINT), Curr. Microbiol. 13 (1986) 77–80. [9] Holt R.A., Cytophaga psychrophila the causative agent of bacterial coldwater disease in salmonid fish, Ph.D. Thesis, Oregon State University, Corvallis, Oregon, 1987. [10] Holt R.A., Rohovec J.S., Fryer J.L., Bacterial coldwater disease, in: Inglis V., Roberts R.J., Bromage N.R. (Eds.) Bacterial Diseases of Fish Blackwell Scientific Publications, Oxford, , 1993, pp. 3–22. [11] Lewin R.A., Lounsbery D.M., Isolation, cultivation and characterization of flexibacteria, J. Gen. Microbiol. 58 (1969) 145–170. [12] Lorenzen E., The importance of the brand of the beef extract in relation to the growth of Flexibacter psychrophilus in Anacker & Ordal medium, Bull. Eur. Ass. Fish Pathol. 13 (1993) 64–65. [13] McDoulagd D., Rice S.A., Weichart D., Kjellerberg S., Nonculturability: adaptation or debilitation? FEMS Microbiol. Ecol. 25 (1998) 1–9. [14] McKay A.M., Viable but non-culturable forms of potentially pathogenic bacteria in water, Appl. Microbiol. 14 (1992) 129–135. [15] Magariños B., Romalde J.L., Barja J.L., Toranzo A.E., Evidence of a dormant but infective state of the fish pathogen Pasteurella piscicida in seawater and sediment, Appl. Environ. Microbiol. 60 (1994) 180–186. [16] Mason C.A., Hamer G., Bryers, J.D., The death and lysis of microorganisms in environmental processes, FEMS Microbiol. Rev. 39 (1986) 373–401. [17] Miles A.A., Misra S.S., The estimation of the bacteriocidal power of blood, J. Hyg. 38 (1938) 732–749. [18] Dando T.R., Young J.E.P. (Eds.), Flexibacter medium, National Collection of Industrial and Marine Bacteria, Catalogue of strains, ref. 281, 1990, pp. 177. [19] Obach A., Baudin-Laurencin F., Vaccination of rainbow trout (Oncorhynchus mykiss) against the visceral form of cold-water disease, Dis. Aquat. Org. 12 (1991) 13–15. [20] Schmidtke L.M., Carson, J., Characteristics of Flexibacter psychrophilus isolated from Atlantic salmon in Australia, Dis. Aquat. Org. 21 (1995) 157–161. [21] Véron M., Nutrition et taxonomie des Enterobacteriaceae et bactéries voisines. I. Méthode d’étude des auxanogrammes, Ann. Inst. Pasteur/Microbiol. 126A (1975) 267–274. [22] von Weis J., Über das Vorkommen einer Kaltwasserkrankheit bei Regenbogenforellen Salmo gairdneri, Tierärzl. Umschau 42 (1987) 575–577. [23] Wakabayashi H., Horiuchi M., Bunya T., Hoshiai G., Outbreaks of cold-water disease in coho salmon in Japan, Fish Pathol. 26 (1991) 211–212. [24] Waterbury J.B., Willey J.M., Isolation and growth of marine planktonic Cyanobacteria, Methods Enzymol. 167 (1988) 100–105.