Detection Causal Agent of Furunculosis SalmonidFish ...clinical furunculosis canoccurin fish populationswhichappar-ently have not come into contact with the pathogen. It also lends

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, OCt. 1994, p. 3874-38770099-2240/94/$04.00+0

Detection of Aeromonas salmonicida, Causal Agent ofFurunculosis in Salmonid Fish, from the Tank Effluent

of Hatchery-Reared Atlantic Salmon SmoltsDAMIEN O'BRIEN,1 JENNY MOONEY,1 DERVILLA RYAN,1 EITHNE POWELL,1

MAURA HINEY,2 PETER R. SMITH,2 AND RICHARD POWELL'*Recombinant DNA Group' and Fish Disease Group, 2 Department of

Microbiology, University College Galway, Galway, Ireland

Received 20 December 1993/Accepted 4 August 1994

The fish pathogen, Aeromonas salmonicida, could be detected only by bacteriological culture from the kidneyof dead or moribund fish in one tank in a hatchery rearing Atlantic salmon (Salmo salar L.) smolts. However,by using a DNA probe specific for this species, allied to a PCR assay, the pathogen could be detected in water,feces and effluent samples taken from this fish tank. Also, the presence of the pathogen was found in effluentsamples from two fish tanks containing apparently healthy fish. Subsequently, the presence of pathogen inthese tanks was confirmed by an increase in the daily mortality rate and by plate culture from moribund fish.

Aeromonas salmonicida is the causal agent of furunculosis(19), a disease of major significance in the culture of salmonidfish (2). The disease represents a serious problem to farming ofAtlantic salmon and causes extensive economic losses tofreshwater hatcheries and sea farms. The absence of anefficient selective medium and the poor plating efficiency of theorganism in mixed cultures (14) have hampered the develop-ment of an efficient diagnostic test for A. salmonicida and,consequently, the control of furunculosis in salmonid culture.The application of PCR (21) for identification and detection

of microorganisms is now well established in the field ofdiagnostic bacteriology (4, 10, 12, 23). The application of thistechnology is particularly useful when an organism of interestproves difficult to culture by normal bacteriological techniques.This is the case when attempting to detect A. salmonicidaoutside the host. Like many gram-negative bacteria, A. salmo-nicida has been associated with a physiological state in partic-ular environments where the organisms become nonculturablebut viable (1, 18, 20). This may help explain how outbreaks ofclinical furunculosis can occur in fish populations which appar-ently have not come into contact with the pathogen. It alsolends controversy to the present obligate pathogen status ofthis bacterium and the culture-based studies showing limitedsurvival of the organism in freshwater and seawater (1, 5,15-17, 22). More directly, the limited ability to culture A.salmonicida from environmental samples has hindered diseasemanagement in salmonid aquaculture.

Previous reports of DNA probes specific for A. salmonicidahave stated their potential as a valuable aid in furunculosiscontrol (3, 6). We described a specific DNA probe for A.salmonicida (7) and, coupled with PCR, a sensitive diagnostictest for this organism. Morgan et al. (18) found this DNAprobe ideal for the detection of A. salmonicida in mixed lakewater populations and commented on its potential as a rapidmethod of detecting nonculturable A. salmonicida cells inother environments. Here, we report the use of this probe forthe detection of A. salmonicida in effluent, water, fecal, and

* Corresponding author. Mailing address: Department of Microbi-ology, University College Galway, Galway, Ireland. Phone: 353 9124411, ext. 2404. Fax: 353 91 25700. Electronic mail address: [email protected].

sediment samples taken from a commercial freshwater hatch-ery rearing Atlantic salmon (Salmo salar L.) smolts. Specifi-cally, this analysis was performed after notification of increasedfish mortalities due to furunculosis in one tank, tank 5, in theweeks preceding the expected movement of these fish to sea.

The hatchery contained 24 indoor tanks, each populated byapproximately 5,500 S1 presmolt Atlantic salmon and had no

previous history of furunculosis. The increase in mortalities intank 5 (0.06% compared with the daily average of 0.01 to0.03% in other tanks) suggested a possible clinical outbreak offurunculosis. Agar (tryptone soy agar) cultures were madefrom the kidneys of fish that had died in this tank. Brown-pigmented colonies were noticed and confirmed as A. salmo-nicida by microscopic examination and sensitized latex agglu-tination (13). In order to investigate the presence of thepathogen in other fish tanks, we decided to analyze samplesfrom a selected number of tanks and from the overall farminfluent and effluent in a manner that might provide usefulinformation on the application of a DNA probe-based diag-nostic test for A. salmonicida in the salmon culture industry.

Samples were collected from the hatchery as follows. Indi-vidual tank effluents (tanks 1 through 9) were filtered throughcotton netting until 2 to 3 g of particulate matter had beentrapped. Total farm influent and effluent samples were alsocollected in this manner. A second effluent sample was alsotaken from the retentate of the hatchery effluent filtrationsystem (which represents a 100-fold concentration) and depos-ited in a sterile container. Two feces samples (approximately 1g), from tank 5, were collected from the tank bottom anddeposited in sterile containers. Water (10 ml) was collectedfrom tank 5. Aseptic procedure was used throughout, and allsamples were stored in the laboratory at 4°C for not more than2 days prior to analysis.Crude preparations of DNA from all samples were obtained

as follows. Approximately 0.5 g of particulate material or fecalmatter was resuspended in 500 ,Iu of sterile H20 in a sterilemicrocentrifuge tube. The 10-ml tank water sample was cen-

trifuged at 2,000 x g for 20 min, and the resulting pellet wasresuspended in 500 Rl of sterile H20 and transferred into a

sterile 1.5-ml microcentrifuge tube. Each sample was leftshaking overnight at ambient temperature to aid resuspensionbefore low-speed centrifugation at 320 x g for 5 min to remove

3874

Vol. 60, No. 10

on July 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

NOTES 3875

large particulate matter. The supernatant was recentrifuged. A250-,u portion of the supernatant was removed, and its cellswere centrifuged at 8,000 x g for 10 min. The resultant cellpellet was resuspended in 200 RI of sterile H20. Each samplewas then heated to 95°C for 10 min to lyse the cells. Tenmicroliters of proteinase K (10 mg/ml; Boehringer Mannheim,GmBH, Germany) was added before incubation at 55°C for 10min. The proteinase K was subsequently denatured by heatingthe samples to 95°C for a further 10 min. After cooling, the celldebris was pelleted by centrifugation at 10,000 x g for 30 s, andthe DNA-containing supernatant was collected.

Usually, the crude DNA extracts obtained from effluent,feces, sediment, and H20 samples are unsuitable as a templatefor PCR. This is thought to be due to the presence of PCRinhibitors. Tsai and Olson (24, 25) reported the presence ofPCR-inhibitory humic substances in crude DNA preparationsfrom environmental samples and a methodology for theirselective removal by column chromatography. In a similarfashion, our crude DNA preparations were purified withSephadex G-200 (Pharmacia Fine Chemicals, Uppsala, Swe-den) spun columns. One-half gram of Sephadex G-200 wasequilibrated in 20 ml of TEN buffer (10 mM Tris-Cl [pH 8.0],1 mM EDTA [pH 8.0], 100 mM NaCl). One milliliter ofSephadex solution was added to a glass wool-plugged 1-mlsterile syringe. The column was then packed by centrifugationat 300 x g. Following this, the column was washed twice by theaddition of 100 [lI of TEN buffer and centrifugation at 300 xg. Finally, 200 ,u of crude DNA extract was divided into two100-pI aliquots which were purified separately. The twoeluents were pooled, and 50 pI of the mixture was used as thetemplate for PCR amplification.An initial eubacterial-specific PCR of a 16S rRNA gene

(SSU rDNA) fragment was carried out on each sample. Thisclearly and rapidly evaluates the removal of PCR inhibitorsfrom the DNA preparations. The two primers used were 27f(5'-AGAGTlTTGATCMTGGCTCAG-3') and 685r2 (5'-TCTACGCATIT'CACYGCTAC-3') SSU rRNA sequencing prim-ers, which specifically amplify this 658-bp fragment from allbeta and gamma proteobacteria (11). Amplification of the423-bp fragment specific toA. salmonicida was carried out withthe primers, PAAS 1 (5'-CGTTGGATATGGCTC'TICCT-3')and PAAS 2 (5'-CTCAAAACGGCTGCGTACCA-3'), asdone previously in our laboratory (7) (PAAS signifies probeassay for A. salmonicida). All four oligonucleotides weresynthesized at the Oswel DNA Service Laboratory (Depart-ment of Chemistry, University of Edinburgh, Edinburgh, Scot-land). PCRs specific for both the (i) SSU rDNA and (ii) A.salmonicida were performed under the following conditions.Fifty-microliter reaction cocktails containing 10 pI of 1oXPCR buffer (200 mM Tris-Cl [pH 8.0], 500 mM KCl), 6 pI of25 mM MgCl2, 16 plI of 1.25 pRM deoxynucleoside triphosphatemix (Promega Corp., Madison, Wis.), 200 pmol of each primer(27f and 685r2 or PAAS 1 and PAAS 2), and 1 U of Taqpolymerase (Promega Corp.) were prepared. Then, 50 [lI ofthe purified sample DNA preparation was added to thesereaction cocktails. DNA was amplified in sterile 500-pA PCRreaction tubes (Perkin-Elmer Cetus, Norwalk, Conn.) in aprogrammable thermocycler (Omnigene TR3 CM220; HybaidLimited, Teddington, Middlesex, United Kingdom). Cyclingconditions (i) for the SSU rDNA fragment were 35 cycles of95°C for 1 min, 45°C for 1 min, and 72°C for 1 min; and (ii) forthe A. salmonicida-specific reaction, 35 cycles of 95°C for 1min, 55°C for 1 min, and 72°C for 1 min were performed. Bothsets of cycling conditions were followed by 5 min at 72°C toensure complete extension. Ten microliters of each PCRproduct was analyzed by gel electrophoresis on ethidium

B 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

FIG. 1. (A) PCR analysis of the SSU rDNA fragment. Lanes 1 and20, X HindIII- and pBR322 HaeIII-digested molecular weight markers(0.1 ,ug each). Lane 2, PCR positive control with 1 ng ofA. salmonicidaDNA as the template. Lane 3, PCR negative control with no DNAtemplate. Lanes 4 and 5, fecal samples from the infected tank (tank 5).Lane 6, total farm influent. Lane 7, total farm effluent. Lane 8, totalfarm effluent filter retentate. Lanes 9 to 17, particulate effluent samplesfrom tanks 1 to 9, respectively. Lane 18, H20 from the infected tank(tank 5). Lane 19, blank. (B) PCR analysis of the A. salmonicida-specific DNA fragment. Lanes 1 and 20, X HindlIl- and pBR322HaeIII-digested molecular weight markers (0.1 pug each). Lane 2, PCRpositive control with 1 ng ofA. salmonicida DNA as the template. Lane3, PCR negative control with no DNA template. Lanes 4 and 5, fecalsamples from tank 5. Lane 6, total farm influent. Lane 7, total farmeffluent. Lane 8, total farm effluent filter retentate. Lanes 9 to 17,particulate effluent samples from tanks 1 to 9, respectively. Lane 18,H20 from tank 5. Lane 19, blank.

bromide-stained 1% agarose gels (Sigma Chemical Co., St.Louis, Mo.).

Figure 1A shows the successful amplification of SSU rDNAfragments from the 15 samples collected from the fish hatchery(lanes 4 to 18). With some modification to the volume of thecolumn, the use of Sephadex G-200 columns as described byTsai and Olson (25) proved very successful in removinginhibitors from the samples. This indication of the removal ofPCR inhibitors from the processed samples also promotesconfidence that false-negative results can be controlled duringthe pathogen-specific PCR. Figure 1B shows the A. salmoni-cida-specific PCR results from the 15 farm samples. Along withthe positive control (lane 2), amplification of a 423-bp DNAfragment can be seen for all the samples taken from tank 5 (2xfeces [lanes 4 and 5], water [lane 18], and effluent filterretentate [lane 8]). Quantification of pathogen numbers inthese samples is difficult because of to the inherent nonlinearkinetics of PCR and the possible presence of PCR retardantsin the sample. However, by comparison the quantities of PCR

VOL. 60, 1994

on July 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

products generated in the farm samples appear similar to theamount generated when 1 pg of purified A. salmonicida DNAis used as template in a sensitivity assay. This quantity repre-sents approximately 200 A. salmonicida genome equivalents.Tryptone soy agar plate cultures made from all samples typesdid not identify A. salmonicida. Presumably, this reflects thepoor plating efficiency of this bacterium amongst mixed bacte-rial populations and, possibly, a nonculturable but viable state.Bacteriological examinations of kidneys from moribund fish inthis tank were positive for A. salmonicida, reflecting thepresence of pathogen in the organs of clinically infected fish.One other sample (lane 8), particulate matter from the reten-tate of the hatchery effluent, was also positive upon gelelectrophoresis.To confirm that the A. salmonicida-specific PCR products

were correct, and to increase detection sensitivity, 90 ,ul of eachPCR product was analyzed by slot blot and probed with aradiolabelled internally coded oligodeoxynucleotide. Follow-ing standard denaturation with 1 M NaOH and neutralizationwith 1 M HCl, the denatured PCR products were transferredonto a Nytran membrane (Schleicher & Schuell, Keene, N.H.)with a Schleicher & Schuell Minifold I slot blot apparatus. Ahybridization probe was prepared by end labelling 100 ng of aspecific internal oligodeoxynucleotide, PAAS 3 (5'-GCTAGC-CAACTCTCTTYTCCA-3'), with [-y-32P]dATP by using T4polynucleotide kinase (Promega Corp.). Hybridization be-tween 50 ng of the gel-purified probe and the membrane-bound target was achieved as follows. The components wereleft in 5x SSC (lx SSC is 0.15 M NaCl plus 0.015 M sodiumcitrate)-lX Denhardt's solution-0.1% sodium dodecyl sulfate(SDS) overnight at 50°C. The washes used were one 5-minwash in 5x SSC-0.1% SDS at room temperature and one5-min wash in lx SSC-0.1% SDS at room temperature.Subsequent autoradiography (Fig. 2) confirmed that correctamplification from all the sample types of tank 5 (slot Bi,particulate effluent; slots Cl and C2, 2x feces; and slot C6,water) were positive, as was the effluent filtration retentate(slot C5). The sensitivity of detection (Fig. 2, slot A4; 100 fg ofA. salmonicida DNA or approximately 20 A. salmonicidagenome equivalents) was approximately 10-fold greater thanthat of gel electrophoresis. Two tank effluent samples, whichshowed negative results upon gel electrophoresis (Fig. 1B,lanes 16 and 17), proved positive after hybridization (slots B8and B9), suggesting the presence of the pathogen in apparentlyhealthy fish. One important aspect of furunculosis is theexistence of asymptomaticA. salmonicida infection (14). Thesefish show no clinical signs of disease but are assumed to becapable of shedding the pathogen (14). Few data are availableon the number of pathogen cells present in asymptomaticallyinfected fish or shed by asymptomatic fish or on the progres-sion of clinical infection from this asymptomatic state. Withfarm management consent, these fish were kept without ther-apeutic treatment for a further week. An increase in dailymortalities was noticed in one of these tanks, and the presenceof pathogen was confirmed by bacteriological examination ofkidney from moribund fish. The subsequent history of the farmoutbreak involved the removal of fish from tank 5 and atherapeutic treatment for the remaining fish. Following trans-fer of these fish to sea, a clinical outbreak of furunculosisoccurred, which resulted in losses of approximately 7% of thetotal number of fish (8).

Important questions remain frustratingly unanswered re-garding the ecology of A. salmonicida and the mode ofdevelopment of furunculosis disease, with a major limitationbeing the lack of methods to identify the pathogen in environ-mental or nonhost samples. Gene amplification technology

A B c

1

2

3

4

5

6

7

8

9

FIG. 2. Slot blot hybridization autoradiogram of samples probedwith the [_y-32P]dATP-radiolabelled PAAS 3 internal oligodeoxynucle-otide. Slots Al to A5, positive controls consisting of serial dilutions ofA. salmonicida genomic DNA (100, 10, and 1 pg and 100 and 10 fg,respectively). Slot A6, negative control with no DNA template. SlotsBi to B9, particulate effluents from tanks 5, 1 to 4, and 6 to 9,respectively. Slots Cl and C2 2x feces samples from tank 5. Slot C3,total farm influent. Slot C4, total farm effluent. Slot C5, total farmeffluent filter retentate. Slot C6, H20 sample from tank 5.

should aid this problem. In the analysis reported here, al-though culture detection was negative, it was possible to detectpathogen in water and particulate matter from fish tankscontaining clinically infected and apparently healthy fish byPCR. The level of sensitivity at approximately 200 A. salmoni-cida genome equivalents per g of sample is a useful limit.However, problems remain for DNA probe-based diagnosticsystems. By design, they will detect free pathogen DNA ornonviable pathogen cells (9). Therefore, care must be taken inthe interpretation of results. Our data show a positive corre-lation between pathogen detection and subsequent clinicaldisease. Most of the samples analyzed were pathogen negative,and the background level for this bacterium in freshwaterappears to be below the limit of detection by PCR. Theseresults provide supporting evidence and demonstrate a poten-tial role for DNA probe technology in disease control ofcultured fish. The sensitivity, time length of analysis, and thenondestructive manner of sampling offer significant advan-tages. This technology may fulfill the need for a suitable tool tomonitor and evaluate management strategies to control dis-ease.

We thank all our colleagues in the Recombinant DNA Group andthe Fish Disease Group for their helpful discussion. We also thank thefish farm management and staff for their assistance.

This work was supported by the European Community Agricultureand Agro-Industry Research Programme and BioResearch Ireland.

3876 NOTES

on July 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

NOTES 3877

REFERENCES1. Allen-Austin, D., B. Austin, and R. R Colwell. 1984. Survival of

Aeromonas salmonicida in river water. FEMS Microbiol. Lett.21:143-146.

2. Austin, B., and D. A. Austin. 1987. Bacterial fish pathogens:disease in farmed and wild fish, p. 112-117. Ellis Horwood Ltd.,Chichester, United Kingdom.

3. Barry, T., R Powell, and F. Gannon. 1990. A general method toobtain DNA probes for microorganisms. Bio/Technology 8:233-236.

4. Bej, A. K., J. L. DiCesare, L. Haff, and R M. Atlas. 1991.Detection of Escherichia coli and Shigella spp. in water by using thepolymerase chain reaction and gene probes. Appl. Environ. Mi-crobiol. 57:1013-1017.

5. Dubois-Darnaudpeys, A. 1977. Epidemiologie de furunculose dessalmonides. I. Etude experimentale des conditions de survie deAeromonas salmonicida dans un environment abiotique. Bull. Fr.Piscic. 49:121-127.

6. Gustafson, C. E., C. J. Thomas, and T. J. Trust. 1992. Detection ofAeromonas salmonicida from fish by using polymerase chainreaction amplification of the virulence surface array protein gene.Appl. Environ. Microbiol. 58:3816-3825.

7. Hiney, M., M. T. Dawson, D. M. Heery, P. R Smith, F. Gannon,and R Powell. 1992. DNA probe for Aeromonas salmonicida.Appl. Environ. Microbiol. 58:1039-1042.

8. Hiney, M., 0. Samuelson, R Coyne, and P. R Smith. Submittedfor publication.

9. Josephson, K. L., C. P. Gerba, and I. L. Pepper. 1993. Polymerasechain reaction detection of nonviable pathogens. Appl. Environ.Microbiol. 59:3513-3515.

10. Lampel, K. A., J. A. Jagow, M. Trucksess, and W. E. Hill. 1990.Polymerase chain reaction for detection of invasive Shigellaflexneriin food. Appl. Environ. Microbiol. 56:1536-1540.

11. Lane, D. J. 1991. 16S/23S rRNA sequencing, p. 115-147. In E.Stackebrandt and M. Goodfellow (ed.), Nucleic acid techniques inbacterial systematics. John Wiley & Sons, Ltd., Chichester, UnitedKingdom.

12. Mahbubani, M. H., A. K. Bej, R Miller, L. Haff, J. DiCesare, andR M. Atlas. 1990. Detection of Legionella with polymerase chainreaction and gene probe methods. Mol. Cell. Probes 4:175-187.

13. McCarthy, D. H. 1975. Detection of Aeromonas salmonicida

antigen in diseased fish tissue. J. Gen. Microbiol. 88:384-386.14. McCarthy, D. H. 1977. Some ecological aspects of the bacterial fish

pathogen, Aeromonas salmonicida. Soc. Appl. Bacteriol. Symp.Ser. 6:229.

15. McCarthy, D. H., and R J. Roberts. 1980. Furunculosis offish-the present state of our knowledge. Adv. Aquat. Microbiol.2:293-340.

16. Michel, C., and A. Dubois-Darnaudpeys. 1980. Persistence of thevirulence of Aeromonas salmonicida strains kept in river sedi-ments. Ann. Rech. Vet. 11:375-380.

17. Morgan, J. A. W., P. A. Cranwell, and R. W. Pickup. 1991. SurvivalofAeromonas salmonicida in lake water. Appl. Environ. Microbiol.57:1777-1782.

18. Morgan, J. A. W., G. Rhodes, and R W. Pickup. 1993. Survival ofnonculturable Aeromonas salmonicida in lake water. Appl. Envi-ron. Microbiol. 59:874-880.

19. Popoff, M. 1984. Genus III. Aeromonas Kluyver and Van Niel1936, 398AL, p. 545-548. In N. R. Krieg and J. G. Holt (ed.),Bergey's manual of systematic bacteriology, vol. 1. Williams &Wilkins Co., Baltimore.

20. Rose, A. S., A. E. Ellis, and A. L. S. Munro. 1990. Evidence againstdormancy in the bacterial fish pathogen, Aeromonas salmonicidasubsp. salmonicida. FEMS Microbiol. Lett. 68:105-108.

21. Saiki, R K., D. H. Gelfand, S. Stoffel, S. F. Scharf, R Higuchi,G. T. Horn, K. B. Mullis, and H. A. Erlich. 1988. Primer directedenzymatic amplification of DNA with a thermostable DNA poly-merase. Science 239:487-491.

22. Sakai, D. K. 1986. Electrostatic mechanism of survival of virulentAeromonas salmonicida strains in river water. Appl. Environ.Microbiol. 51:1343-1349.

23. Samadpour, M., J. Liston, J. E. Ongerth, and P. I. Tarr. 1990.Evaluation of DNA probes for detection of Shiga-like-toxin-producing Escherichia coli in food and calf fecal samples. Appl.Environ. Microbiol. 56:1212-1215.

24. Tsai, Y.-L., and B. H. Olson. 1992. Detection of low numbers ofbacterial cells in soils and sediments by polymerase chain reaction.Appl. Environ. Microbiol. 58:754-757.

25. Tsai, Y.-L., and B. H. Olson. 1992. Rapid method for separation ofbacterial DNA from humic substances for polymerase chainreaction. Appl. Environ. Microbiol. 58:2292-2295.

VOL. 60, 1994

on July 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from