Peroxide-inducible catalase in Aeromonas salmonicida subsp ...

Article available online at http://www.idealibrary.com on

Microbial Pathogenesis 1999; 26: 149–158 Article No. mpat.1998.0260

MICROBIAL PATHOGENESIS

Peroxide-inducible catalase in Aeromonas salmonicida subsp. salmonicida protects against exogenous hydrogen peroxide and killing by activated rainbow trout, Oncorhynchus mykiss L., macrophages Andrew C. Barnesa,∗, Timothy J. Bowdena, Michael T. Horneb & Anthony E. Ellisa a

FRS Marine Laboratory, PO Box 101, Victoria Road, Aberdeen AB11 9DB, Scotland, and bAqua Health (Europe) Ltd., Unit 31, Enterprise House, Springkerse Business Park, Stirling FK7 7UF, Scotland Received August 31, 1998; accepted in revised form November 21, 1998)

Aeromonas salmonicida subsp. salmonicida expresses a single cytoplasmically located catalase which was found to be inducible by exposure to 20 lM hydrogen peroxide in mid-exponential phase resulting in a 4 fold increase in activity. Subsequent exposure to 2 mM peroxide in lateexponential/early-stationary phase resulted in further induction of catalase activity which increased to 20 fold higher levels than those found in uninduced cultures. Exponentially induced cultures were protected against subsequent exposure to 10 mM peroxide which was lethal to non-induced cultures. Bacteria subjected to induction in mid-exponential and early-stationary phase were resistant to 100 mM peroxide, although viability was greatly reduced. Growth of the bacterium under iron-restricted conditions had no effect on the peroxide induction of catalase. As current evidence indicates, the latter is an iron-co-factored heme catalase, this result suggests that catalase induction has a high priority in the metabolism of iron. Furthermore, exposure to peroxide also induces expression of periplasmic MnSOD. A. salmonicida MT423 was resistant to normal rainbow trout macrophages, but was susceptible to killing by activated macrophages. However, if catalase was induced by prior exposure to 20 lM peroxide during mid-exponential phase, A. salmonicida was resistant to killing by activated macrophages. The ability of A. salmonicida to upregulate periplasmic MnSOD and cytoplasmic catalase production under iron restricted conditions and low level peroxide (conditions expected to exist during the early stages of an infection) may be vital for its ability to withstand attack by phagocytic cells in vivo.  1999 Academic Press Key words: Catalase, A. Salmonicida, macrophages, peroxide, superoxide dismutase, iron metabolism. ∗ Author for correspondence: FRS Marine Laboratory, PO Box 101, Victoria Road, Aberdeen AB11 9DB, Scotland. 0882–4010/99/030149+10 $30.00/0

 1999 Academic Press

150

A. C. Barnes et al.

Introduction The killing of most extracellular pathogens by macrophages is largely dependent on the respiratory burst following phagocytosis and is an important defence mechanism in mammals [1] and fish [2]. During the respiratory burst, reactive oxygen species including superoxide anion [2], hydrogen peroxide [3], hydroxyl radical [4] and nitric oxide [5–7] are produced. These agents are variously toxic to bacteria and thus bacteria have evolved defences against these agents. Detoxifying enzymes include superoxide dismutases (SODs), catalase, glucose-6-phosphate dehydrogenase, glutathione synthetase and glutathione reductase [8]. Aeromonas salmonicida subsp salmonicida is a facultatively anaerobic pathogen of fish. Virulent isolates have been reported to vary in their resistance to oxidative killing by salmonid macrophages [9–11] and superoxide anion generated in cell free systems [12, 13]. In a previous study, we demonstrated the presence of two SODs in Aeromonas salmonicida, regulated in response to iron, and showed that these may contribute to defences against the superoxide anion [13]. However, no data has yet been published concerning catalases in this species. In other species, such as Shigella flexneri [14]. Haemophilus influenzae [15] and Staphylococcus aureus [16], catalases have been associated with virulence and resistance to phagocytic attack. Regulation of expression of catalases is complex and has been studied exhaustively in Escherichia coli [17–19] which expresses two catalases, HPI which is peroxide inducible and HPII which is under stationary phase control. Peroxide inducible catalases have also been reported in other species such as Pseudomonas

syringae [20], Pseudomonas aeruginosa [21] and Xanthomonas oryzae [22]. Furthermore, inducible catalase in Salmonella typhimurium has been implicated as an important antigen in cell-mediated immunity in mice [23]. Although Aeromonas salmonicida is known to be catalase positive [24], little is known of the biochemical and regulatory properties of the enzyme or its role in pathogenicity. In this study we identify a single peroxide-inducible catalase which appears to be cytoplasmically located and protects against exogenous H2O2 and killing by activated rainbow trout macrophages.

Results All six isolates of Aeromonas salmonicida examined, produced a single catalase with identical mobility in native PAGE gels. Isolate MT423, previously characterised [13], was selected for further study. There was no apparent stationary phase induction of catalase, as activity detected in exponential and stationary phase cultures were similar (Table I). However, pulsing the culture with H2O2 in mid-exponential phase resulted in a fourfold increase in catalase activity (Table I) compared to unstressed controls. This appeared to result from induction of the single catalase enzyme as no additional bands could be visualised on the gels (Fig. 1). Interestingly, if incubation was continued beyond the midexponential pulse, and a subsequent higher dose of 2 mM peroxide was applied late in exponential or at the onset of stationary phase, a twentyfold increase in catalase activity was noted (Table 1). This appeared to result from a secondary induction of the same catalase enzyme. If 2 mM peroxide was applied to the

Table 1. Catalase activity detected in A. salmonicida under various culture conditions. Culture conditions Early-stationary phase Mid-exponential phase Induced 20 lM peroxide in mid-exponential phase Induced 2 mM peroxide in early stationary phase Induced 20 lM mid-exponential followed by 2 mM in early stationary phase

Catalase activity (units/109 cells)

Standard deviation

Percent activity relative to uninduced stationary phase

0.40 0.33 1.92

±0.090 ±0.020 ±0.400

100 83 480

0.06

±0.001

15

8.02

±0.600

2005

Inducible catalase in Aeromonas salmonicida 1 2 3 4

151

1 2 3 4

1

2

1

2

Catalase FeSOD MnSOD

(a)

(b)

Figure 1. Native polyacrylamide gel (10%) stained for (a) superoxide dismutase activity and (b) catalase activity in lysates from A. salmonicida (MT423) cells cultured as follows: Lane 1: TSB with exposure to peroxide in mid-exponential and early stationary phase; lane 2: TSB; lane 3: TSB+100 lM 2,2-dipyridyl, with exposure to peroxide in both mid-exponential and early stationary phase; lane 4: TSB+100 lM 2,2dipyridyl.

culture in stationary phase without prior exposure to 20 lM peroxide, no catalase activity was detected at all, and the culture lost viability (Table 1). The uninduced exponential phase level of catalase expressed was about one third of that expressed by S. typhimurium under similar conditions [23], as was the level expressed following exposure to 20 lM peroxide. The level expressed by A. salmonicida following double exposure to peroxide was similar to that expressed following single exposure of S. typhimurium [23]. The enzyme would appear to be an iron-cofactored catalase as activity could not be detected on gels following heating at 80°C for 1 h, exposure to 100 mM potassium cyanide or 1 mM mercuric chloride. Reduced activity was detected in the presence of 12.5 mM sodium azide, 50 mM potassium cyanide or 0.5 mM mercuric chloride. In contrast, Mn-containing pseudocatalases are resistant to these compounds [28, 29]. Furthermore, Mn catalase retains 95% of its activity following heating at 80°C for 1 h [29]. In spite of the probable ironcofactor, limiting the iron available in the medium had no effect on catalase induction by peroxide (Fig. 1). However, manganese SOD, normally only expressed under iron limitation [13], was induced by peroxide under iron replete conditions (Fig. 1). Catalase induction would appear to require de novo protein synthesis, as pulsing the culture with 35S-methionine concurrently with peroxide stress resulted in uptake

(a)

(b)

Figure 2. (a) Native acrylamide gel (10%) stained for catalase and (b) autoradiograph of the same gel showing uptake of 35S-methionine in lysates from MT423 cells cultured in M9-MEM; control cells following exposure to 20 lM peroxide in midexponential phase (lane 1) or without peroxide exposure (lane 2).

of radiolabel into the catalase enzyme (Fig. 2). This was confirmed as no induction of catalase by peroxide in either exponential or stationary phase was detected in the presence of 40.0 mg/l chloramphenicol (not shown). Furthermore, at the higher induction dose in late exponential phase the culture lost viability in the presence of chloramphenicol. The catalase appeared to be located in the cytoplasm as no activity was detected in periplasmic extracts. Purity of periplasmic extracts was confirmed by assaying glutamine synthetase activity in sonicates and periplasmic fractions. In sonicates, glutamine synthetase activity was detected at a level of 5085 units/mg protein, whilst in periplasmic fractions, activity was too low to detect. The success of isolation of active

152

A. C. Barnes et al. 1000

Percent survival

100

10

1

0.1

0.01 0

0.1

1

10

100

Peroxide concentration (mM)

Figure 3. Protection of A. salmonicida MT423 against exogenous peroxide following induction of catalase. Mid exponential phase control (Ε). Early stationary phase control (∨). Induced mid-exponential with 20 lM peroxide (Φ). Induced mid-exponential with 20 lM peroxide followed by 2 mM peroxide in early stationary phase (∆). Results show means of triplicate values.

periplasmic components was confirmed by detection of b-lactamase activity known to be present in this isolate. Aeromonas salmonicida MT423 cultures with catalase induced in mid-exponential phase were resistant to exogenous peroxide at 10 mM with 22% of the inoculum surviving 60 min at this concentration whilst controls were killed (Fig. 3). However, bacteria subjected to induction in mid-exponential followed by subsequent induction at the onset of stationary phase were resistant to 100 mM peroxide, although viability was reduced to 0.1% (Fig. 3). There was a slight increase in resistance to peroxide in stationary phase cultures compared to exponential phase cultures, even though there was no notable difference in the amount of catalase expressed (Fig. 3, Table 1). To determine the relevance of catalase in resistance to host phagocytes, macrophages isolated from a single fish were subdivided into two groups. The first group was incubated with lipopolysaccharide and TNF-alpha to activate the respiratory burst cascade. The second group were incubated in the same medium but without LPS and TNF-alpha as non-activated controls, and the killing activity of the two populations was studied simultaneously to identify differences. The ratio of macrophages to bacteria

was also investigated as bactericidal activity of rainbow trout macrophages against A. salmonicida has been reported to be influenced by bacterial numbers [11]. When approximately 5×105 bacteria/well were incubated with approximately 106 macrophages/well [multiplicity of infection (MOI)=0.5] there was no bactericidal activity, regardless of the bacterial culture conditions or the activation state of the macrophages [Fig. 4(a)]. When the experiment was conducted using a lower MOI (0.025), no bactericidal activity was detected in wells containing non-activated macrophages [Fig. 4(b)]. However, activated macrophages were able to kill A. salmonicida cultured under either iron replete or iron limited conditions, provided catalase had not previously been induced by prior exposure to hydrogen peroxide [Fig. 4(b)]. When catalase was induced by a single exposure to 20 lM hydrogen peroxide in mid-exponential phase, A. salmonicida was resistant to killing by activated macrophages. No appreciable difference was recorded between cells cultured under iron replete or iron limited conditions [Fig. 4(b)].

Discussion In E. coli, intracellular peroxide levels are maintained within a narrow range (approx. 0.2 lM), predominantly through transcriptional regulation of peroxide-inducible HPI catalase, at least during exponential growth [25]. However, in E. coli, a second catalase (HPII) is present and is upregulated at the onset of stationary phase [26]. Aeromonas salmonicida only produces a single catalase with no apparent stationary phasedependent induction. However, after mid-exponential phase induction with sublethal concentrations of peroxide, subsequent exposure to high concentrations of peroxide, results in a further increase in catalase activity. The regulatory mechanisms governing catalase expression in A. salmonicida at the molecular level are unknown, however, oligonucleotide probes based on sequences derived from E. coli KatG hybridised with DNA from an A. salmonicida library [27]. This suggests that A. salmonicida catalase bears at least some homology to E. coli HP-I. It may be, therefore, that the response to peroxide is regulated in a similar manner and further work is required to determine if this is the case.

Inducible catalase in Aeromonas salmonicida 160

(a)

140

Percent survival

120 100 80 60 40 20 0 Non-activated macrophages

Activated macrophages

Non-activated macrophages

Activated macrophages

(b) 100

Percent survival

80

60

40

20

0

Figure 4. Effect of catalase induction on killing of A. salmonicida MT423 by rainbow trout macrophages. (a) multiplicity of infection (MOI)=0.5. (b) MOI= 0.025. Bacterial cells cultured in M9-MEM, iron replete (Γ); M9-MEM iron replete, with catalase induced by exposure to 20 lM peroxide in midexponential phase (Ε); M9-MEM iron limited (∧); M9-MEM iron limited, catalase induced in mid-exponential phase with 20 lM peroxide (Φ).

Induced catalase was able to protect against levels of peroxide which were lethal to noninduced cultures. Furthermore, stationary phase cultures were more resistant to peroxide than exponential phase cultures, in spite of there being no notable difference in catalase levels.

153

This may reflect the involvement of other enzymes such as endonuclease IV or glutathione synthetase which may be under stationary phase control. Alternatively, it may be that the low metabolic activity in stationary phase cultures reduces the generation of reactive oxygen intermediates internally thus, even in the presence of exogenous peroxide, basal levels of catalase may suffice to maintain intracellular peroxide levels within acceptable tolerances. Olivier et al. [30] reported that virulent A. salmonicida are cytotoxic to Atlantic salmon peritoneal macrophages when the ratio of bacteria to macrophages was 1:1 (MOI=1.0) or greater. In the present study we demonstrated that rainbow trout head kidney macrophages were unable to kill virulent A. salmonicida when MOI=0.5, even if the macrophages were activated with TNF alpha and LPS prior to exposure. Only when the number of macrophages greatly exceeded the number of bacteria (MOI=0.025), and the macrophages were activated, was any bactericidal activity noted. However, bacteria were resistant to this activity if they had previously been exposed to 20 lM H2O2 during exponential growth. Interactions between A. salmonicida and rainbow trout macrophages have also been reported previously [11]. Avirulent A. salmonicida was killed in a concentration-dependent fashion by rainbow trout head kidney macrophages. In contrast, virulent A. salmonicida was resistant to these macrophages unless the catalase inhibitor aminotriazole was included in the assay whilst presence of a SOD inhibitor did not reduce the bactericidal activity. However, in this case the macrophages had not been previously activated [11]. In the present study, virulent A. salmonicida were only killed by activated macrophages, and only when macrophages greatly outnumbered the bacteria. These results also corroborate the observations of Sharp and Secombes [11] that catalase activity is important in the resistance of A. salmonicida to the bactericidal activity of rainbow trout head kidney macrophages. Indeed, a single induction of catalase by prior exposure to micromolar quantities of peroxide is capable of protecting against activated macrophages. The significance of other detoxifying enzymes however is less clear. In a previous study, we reported that A. salmonicida expresses one of two SOD enzymes depending upon the availability of iron [13]. In iron replete conditions, a cytoplasmic FeSOD was expressed but under iron limitation, this was replaced by a

154

periplasmic MnSOD and the switch was associated with increased resistance to exogenous superoxide anion [13]. However, in the present study, the iron available in the culture medium did not greatly affect resistance to activated rainbow trout macrophages. What may be significant is that peroxide is also capable of inducing manganese SOD under iron replete conditions. Thus following exposure to subinhibitory peroxide concentrations in mid-exponential phase, and subsequent exposure to millimolar quantities later in the growth phase, the bacteria expressed both FeSOD and MnSOD under iron replete conditions. This would seem paradoxical as extra SOD would surely only serve to increase intracellular peroxide concentrations. However, since the same exposure to peroxide induces increased production of catalase, the bacterium would still be protected as indeed the present data concerning increased resistance to exogenous peroxide concentrations demonstrate. Furthermore, millimolar quantities of peroxide may cause inactivation of FeSOD by bleaching of the active site, indeed there would appear to be reduced FeSOD activity following exposure to H2O2 [Fig. 1(a), lane 1]. One may be tempted to speculate, that the induction of MnSOD results from increased intracellular superoxide anion concentration. However, in a previous study, we demonstrated that MnSOD was not induced by increased superoxide anion [13]. Under iron-restricted conditions expression of the FeSOD was terminated and expression of MnSOD was induced. This confirms previous work which also demonstrated that FeSOD was cytoplasmic, whilst MnSOD was located in the periplasm [13]. Moreover, induction of increased catalase production by peroxide exposure still occurred under iron-restricted conditions. Like FeSOD, the data presented here suggest that catalase is also iron-cofactored. Thus, it would appear that under iron-restricted conditions and exposure to peroxide, the available cytoplasmic iron may be conserved by stopping production of FeSOD and diverting the iron into catalase production. The present data suggest that the effects of iron and peroxide on expression of SOD-type and catalase could be highly adaptive for in vivo survival of A. salmonicida. From the data, it can be hypothesised that on entering the host A. salmonicida will encounter an iron-restricted environment. This will induce the switch over from production of cytoplasmic FeSOD to periplasmic

A. C. Barnes et al.

MnSOD. With the influx of phagocytes to the site of infection O2− and H2O2 will be produced as a result of respiratory burst activity. Aeromonas salmonicida cells which are not immediately killed may respond to the low levels of peroxide by further enhancing MnSOD production as well as greatly increasing catalase production. Such a strategy may, as suggested by the present in vitro data, provide an effective defence against killing by phagocytes and may contribute to the extremely low LD50 of virulent isolates of this organism reported in previous studies [32]. Furthermore, as infection progresses with the bacterium obtaining iron from its high affinity uptake systems [33, 34], iron would become available for production of cytoplasmic FeSOD. With the availability of nutrients allowing multiplication of the bacterium, this enzyme may be important for controlling O2− levels within the matrix generated by increased metabolic activity, and countering the additional threat presented by these radicals resulting from ironcatalysed Fenton reactions [35]. Nevertheless, if attack by phagocytes persists, the exogenous peroxide they produce would continue to stimulate production of MnSOD and catalase providing protection against reactive oxygen intermediates. As the present data show, as well as other reports [9], a MOI above 1.0 results in A. salmonicida being capable of overwhelming the phagocytes which would enable the bacterium to spread from the infection site and become systemic. Thus the regulatory effects of peroxide and iron on the expression of SODs and catalase by A. salmonicida described herein, may account for the ability of A. salmonicida to resist elimination by salmonid macrophages and partially account for the extreme virulence of this pathogen once it has penetrated the host’s integument.

Materials and methods Bacterial strains and culture conditions Aeromonas salmonicida subspecies salmonicida MT423, MT1526, MT1560, MT1662, MT1712 and MT1759 were isolated from geographically disparate outbreaks of furunculosis at Scottish salmon farm sites and were maintained as freeze dried stocks in the collection at FRS Marine Laboratory, Aberdeen. Bacteria were cultured on Tryptone soya agar for 48 h. Broth cultures

Inducible catalase in Aeromonas salmonicida

were carried out in TSB or TSB+100 lM 2,2dipyridyl for iron limitation, or defined medium (M9-MEM), comprising M9 salts[36]+MEM amino acids (Gibco, U.K.)+100 lM ferric chloride for iron replete conditions, or 60 lM 2,2dipyridyl (iron limited conditions) [37].

Catalase induction An aliquot (10 ll) of overnight culture in 10 ml broth (TSB or M9-MEM) was used to inoculate 50 ml broth in 250 ml Erlenmeyer flasks. Cells were incubated with shaking at 140 rpm at 22°C until cell density reached A540=0.4–0.6 (midexponential phase) or A540=1.0–1.2 (late exponential/early stationary phase). Catalase was induced in mid-exponential phase using 20 lM H2O2, and subsequently induced in late exponential/early stationary phase using 2 mM H2O2. Cells were harvested 60 min post-induction.

Cell fractionation For whole cell extracts, A. salmonicida MT1712 (b-lactamase positive) cells were harvested by centrifugation at 3500 g, 4°C for 20 min, and washed in 50 mM sodium phosphate buffer pH 7.5. Cells were resuspended in the same buffer to a density of 109 cfu/ml (A540=1.00) and equal volumes (5.0 ml) were sonicated on ice for 4 bursts of 30 s with 30 s cooling between using a Soniprep 150 ultrasonic disruptor (Sanyo Gallenkamp, Loughborough, U.K.). Lysates were clarified by centrifugation at 15,000 g for 60 min, stored on ice and analysed immediately without freezing. Supernatants were diluted 1:1 in native PAGE sample buffer and analysed by native gel electrophoresis (30 ll/lane). Periplasmic proteins were extracted using chloroform essentially as described by Ames et al. [38], a method previously described as successful for this species [13]. To detect cytoplasmic contamination, glutamine synthetase (GS), a cytoplasmic marker [38], was assayed according to the method of Canovas et al. [39]. GS was selected in favour of malate dehydrogenase as the latter is inhibited by chloroform, whilst GS and catalase are stable to this treatment [40]. Presence of active periplasmic components was determined by assay of b-lactamase, known to be present in this isolate (Barnes and Ratnayaka, unpublished data), using the chromogenic cephalosporin

155

nitrocefin (Oxoid, U.K.) as described previously [41].

Gel electrophoresis Lysates were fractionated on discontinuous acrylamide gels (10%) using a Hoeffer SE260 minigel system at 250 volts for 1 h (Pharmacia Biotech, St Albans, U.K.). Gels were washed extensively with distilled water (4×15 minutes) and soaked in 0.015% H2O2 in distilled water for 10 min. Catalase activity was visualised by staining in 1% ferric chloride, 1% potassium ferricyanide until yellow bands appeared on a dark blue/green background [20]. SOD activity was visualised as described previously [42].

Catalase assay Catalase activity was quantified based on the evolution of oxygen, which was assayed polarographically using a Clark-type oxygen electrode (Rank Instruments, Cambridge, U.K.) [43]. Bacterial cultures (MT423) were washed twice in 50 mM potassium phosphate, pH 7.0 and resuspended to a cell density of 109 cfu/ml (A540= 1.0) in the same buffer at 25°C. Bacterial suspension (2.15 ml) was pipetted into the chamber of the instrument and allowed to equilibrate and establish any baseline. Hydrogen peroxide was added to a final concentration of 2.0 mM and oxygen evolution recorded at 5 s intervals for up to 2 min. Assays were replicated a minimum of four times. The assay was calibrated using a standard solution of catalase from bovine erythrocytes (Sigma, Poole, U.K.).

Effects of inhibitors Catalase from a lysate of MT423 was incubated for 1 h with either potassium cyanide (100 mM, 50 mM), mercuric chloride (1 mM, 0.5 mM), sodium azide (25 mM, 12.5 mM), 50 mM phosphate buffer (control), or heated at 80°C, for 1 h. Equal volumes were then applied to a 10% native acrylamide gel and separated by electrophoresis as described above. The gel was stained for catalase activity.

Pulse/chase of catalase induction To determine whether catalase was synthesised de novo, overnight cultures of A. salmonicida

156

MT423 in M9-MEM were used to inoculate 50 ml M9-MEM in a 250 ml Erlenmeyer flask. When the cell density reached A540=0.4, the culture was divided into two, cells harvested and resuspended in equal volumes of fresh M9-MEM containing 10 lCi/ml 35S-methionine (Amersham, Bucks., U.K.). To one sample, H2O2 was added to a concentration of 20 lM. Incubation was continued and samples were taken after 1 h. The cells were harvested and washed in sodium phosphate buffer, then resuspended in 200 ll native PAGE loading buffer in a microcentrifuge tube. Glass beads (150–212 lm) (Sigma, Poole, U.K.) were added and the tubes vortexed for 2 min to lyse the cells. The resulting suspension was centrifuged at 19,800 g for 10 min at 4°C in a microcentrifuge and the supernatant was loaded onto 10% native acrylamide gels. After electrophoresis the gel was stained for catalase activity then dried between cellophane in a drying frame (Promega, Southampton, U.K.). The dried gel was then incubated in a lightproof autorad case with autoradiograph film (Kodak) for 7 days at −80°C. Autorad and gel were compared to identify catalase bands.

Inhibition of protein synthesis with chloramphenicol Inhibition of protein synthesis was performed essentially as described by Barnes [44]. Briefly, for exponential induction, a mid-exponential phase (A540=0.6) culture of MT423 in TSB (200 ml) was split 4×50 ml into sterile 250 ml shake flasks. Chloramphenicol was added to two flasks at the minimum inhibitory concentration (MIC) (40 mg/l in this case). Incubation was continued for 20 min to allow complete inhibition of protein synthesis then peroxide (20 lM) was added to one flask containing chloramphenicol and one without as a positive control for catalase induction. Incubation was continued with vigorous agitation (140 rpm) for 40 min. Aliquots (100 ll) were serially diluted and checked for viability on TSA. Cells from the remaining volume were harvested by centrifugation, washed in cold 50 mM potassium phosphate buffer pH 7.2 and resuspended to equal cell densities in the same buffer. Cells from 10 ml of each suspension were pelleted, resuspended in 300 ll native sample buffer and lysed by vortexing for 2 min with 40 mg glass beads (115–212 lM, Sigma, U.K.). Lysates were cleared by centrifugation at 19,800 g for 10 min,

A. C. Barnes et al.

the supernatants analysed by native gel electrophoresis and stained for catalase activity as described above. For high level induction a 200 ml TSB culture, pre-exposed to 20 lM peroxide at A540=0.6 was incubated to early stationary phase (A540=1.2–1.4). The culture was split as described above and the above experiment duplicated except that 2 mM peroxide was used as the inducing dose.

Resistance to exogenous hydrogen peroxide The degree of protection against hydrogen peroxide-mediated killing afforded by the induction of catalase was determined using a bactericidal assay: A. salmonicida MT423 was grown in M9MEM as described above. Induced cells were pulsed with either 20 lM H2O2 at A540=0.4–0.6 (mid-exponential) or 2 mM H2O2 at A540=1.2–1.4 (early stationary) or pulsed both in mid-exponential and then in early stationary phase. Cells were harvested 1 h post-induction, washed in M9 salts solution and resuspended to a density of 109 cfu/ml (A540=1.00). Aliquots (100 ll) of these suspensions were then inoculated into 9.9 ml M9 containing H2O2 at concentrations of 0, 0.1, 1, 10 and 100 mM. Assays were incubated for 1 hour at 22°C, serially diluted and bacteria were enumerated by viable counting in triplicate on TSA plates.

Bactericidal activity of rainbow trout macrophages Rainbow trout, Oncorhynchus mykiss (800–1350 g) were held in 1 m circular tanks containing fresh water at 12–14°C at a flow of 100 litres per h. Macrophages were isolated essentially as described by Secombes [45]. Briefly, fish were killed by overdose of anaesthetic (MS222, Sigma, Poole, U.K.) and exsanguinated by removal of the gills. Head kidney was asceptically removed and pushed through a sterile metal gauze (100 lm) with L-15 containing 2% foetal calf serum (FCS), penicillin/ streptomycin (P/S) (100 units/ml) and heparin (10 units/ml). The resulting suspension was washed twice in the same medium and layered onto 34%/51% Percoll gradients. The gradients were centrifuged at 400 g for 25 min at 4°C in a Beckman swing out rotor with the brake off. The band at the gradient interface was collected

Inducible catalase in Aeromonas salmonicida

and washed in L15 with 0.5% FCS, 100 U/ml P/ S; 10 U/ml; heparin (L15 wash). Viable cells were counted in 4% trypan blue and resuspended to approximately 107 cells per ml in L15 wash. Aliquots (100 ll) were added to each well of a 96 well tissue culture (TC) plate (Life Technologies, Paisley, U.K.) and cells were allowed to attach for 3 h in a humid box at 18°C. Non-adherent cells were then removed using three washes of L15 wash buffer. Attached macrophages were then fed with either L15, 5% FCS, P/S (100 U/ ml), heparin (10 U/ml), or, for activation, with the same buffer containing 25 units/ml human recombinant TNF alpha expressed in yeast (Sigma, Poole, U.K.) and 25 lg/ml E. coli 055: B5 lipopolysaccharide (LPS). Macrophages were incubated for 40 h at 18°C to allow activation [46]. Activated and non-activated macrophages in the same TC plate were washed three times in L-15 wash without P/S. A number of control wells were lysed using 0.1 M citric acid, 1% Tween 20, 0.05% crystal violet and the nuclei counted in a haemocytometer to determine the number of attached macrophages. A. salmonicida MT423 grown in M9-MEM iron replete and iron limited medium with and without peroxide induction of catalase in mid-exponential phase (A540=0.4–0.6) with 20 lM peroxide were washed twice in L-15, 5% FCS, 10 U/ml heparin (i.e. without P/S) and counted in a haemocytometer. Bacteria were added to quadruplicate wells in a microtitre plate such that the MOI was approximately 0.5 or 0.025 (approximately 106 macrophages per well incubated with approximately 5×105 or 2.5×104 bacteria). Control wells comprised bacteria added to macrophage monolayers lysed with distilled water. Microplates were spun at 400 g for 5 min to bring bacteria into contact with the macrophage monolayer and the plates were incubated for 5 h at 18°C. After incubation the monolayers were lysed with distilled water. Aliquots (100 ll) were serially diluted in M9 salts solution and bacteria enumerated by viable count on TSA.

Acknowledgements We thank Juan Campos-Perez and Anne Langston for advice and expertise in working with rainbow trout macrophages. Andrew Barnes is an employee of Aqua Health (Europe) Ltd.

157

References 1 Hurst JK, Barretter Jr WC. Leucocytic oxygen activation and microbicidal oxidative toxins. Crit Rev Biochem Mol Biol 1989; 24: 271–328. 2 Secombes CJ, Fletcher TC. The role of phagocytes in the protective mechanisms of fish. Ann Rev Fish Dis 1992; 2: 53–71. 3 Babior BM. The respiratory burst oxidase. Adv Enzymol Rel Areas Mol Biol 1992; 65: 49–65. 4 Fridovich I. The biology of oxygen radicals. Science 1978; 201: 875–80. 5 Iyengar R, Stuehr DJ, Marietta MA. Macrophage synthesis of nitrite, nitrate, and N-nitrosamines: precursors and role of the respiratory burst. Proc Natl Acad Sci USA 1987; 84: 6369–73. 6 Nunoshiba T, DeRojas-Walker T, Tannenbaum SR, Demple B. Roles of nitric oxide in inducible resistance of Escherichia coli to activated murine macrophages. Infect Immun 1995; 63: 794–8. 7 Laing KJ, Grabowski PS, Belosevic M, Secombes CJ. A partial sequence for nitric oxide synthetase from a goldfish (Carassius auratus) macrophage cell line. Immunol Cell Biol 1996; 74: 374–9. 8 Farr SB, Kogoma T. Oxidative stress responses in Escherichia coli. Microbiol Rev 1991; 55: 561–85. 9 Olivier G, Eaton CA, Campbell N. Interaction between Aeromonas salmonicida and peritoneal macrophages of brook trout (Salvelinus fontinalis). Vet Immunol Immunopathol 1986; 12: 223–34. 10 Graham S, Jefferies AH, Secombes CJ. A novel assay to detect macrophage bactericidal activity in fish: Factors influencing the killing of Aeromonas salmonicida. J Fish Dis 1988; 11: 389–96. 11 Sharp GJE, Secombes CJ. The role of reactive oxygen species in the killing of the bacterial fish pathogen Aeromonas salmonicida by rainbow trout macrophages. Fish Shell Immunol 1993; 3: 119–29. 12 Karczewski JM, Sharp GJE, Secombes CJ. Susceptibility of strains of Aeromonas salmonicida to killing by cell-free generated superoxide anion. J Fish Dis 1991; 14: 367–73. 13 Barnes AC, Horne MT, Ellis AE. Effect of iron on expression of superoxide dismutase by Aeromonas salmonicida and associated resistance to superoxide anion. FEMS Microbiol Lett 1996; 142: 19–26. 14 Franzon VL, Arondel I, Sansonetti PI. Contribution of superoxide dismutase and catalase activities to Shigella flexneri pathogenesis. Infect Immun 1990; 58: 529–34. 15 Bishai WR, Smith HO, Barcak GJ. A peroxide/ascorbate inducible catalase from Haemophilus influenzae is homologous to the Escherichia coli katE gene product. J Bacteriol 1994; 176: 2914–21. 16 Mandell GL. Catalase, superoxide dismutase and virulence of Staphylococcus aureus: In vitro and in vivo studies with emphasis on staphylococcal–leucocyte interaction. J Clin Invest 1975; 55: 561–6. 17 Storz G, Tartaglia LA, Ames BN. Transcriptional regulation of oxidative stress inducible genes: Direct activation by oxidation. Science 1990; 248: 189–94. 18 Schellhorn HE. Regulation of hydroperoxidase (catalase) expression in Escherichia coli. FEMS Microbiol Lett 1994; 131: 113–19. 19 Loewen PC. Bacterial catalases. In: Scandalios JG, Ed.

158

20 21

22

23

24

25

26 27 28

29 30

31 32

Oxidative Stress and the Molecular Biology of Antioxidant Defenses. Cold Spring Harbor Press. 1997: 273–308. Klotz MG, Hutcheson SW. Multiple periplasmic catalases in pathogenic strains of Pseudomonas syringae. Appl Env Microbiol 1992; 58: 2468–73. Brown SM, Howell ML, Vasil ML, Anderson AJ, Hassett DJ. Cloning and characterisation of the KatB gene of Pseudomonas aeruginosa encoding a hydrogen peroxide inducible catalase: Purification of KatB, cellular localization and a demonstration that it is essential for optimal resistance to hydrogen peroxide. J Bacteriol 1995; 177: 6536–44. Mongkolsuk S, Loprasert S, Vattanaviboon P, Chanvanichayachai C, Chamnongpol S, Supsamran N. Heterologous growth phase- and temperature-dependent expression and peroxide toxicity protection of a superoxide-inducible monofunctional catalase gene from Xanthomonas oryzae pv. oryzae. J Bacteriol 1996; 178: 3578–84. Kagaya K, Miyakawa Y, Watanabe K, Fukazawa Y. Antigenic role of stress-induced catalase of Salmonella typhimurium in cell-mediated immunity. Infect Immun 1992; 60: 1820–5. Austin B, Austin DA. Aeromonadaceae representatives (Aeromonas salmonicida). In: Austin B, Austin DA, Eds. Bacterial Fish Pathogens: Disease in Farmed and Wild Fish, 2nd Edition. Chichester: Ellis Horwood, 1993: 86–170. Gonzalez-Flecha B, Demple B. Homeostatic regulation of intracellular hydrogen peroxide concentration by aerobically growing Escherichia coli. J Bacteriol 1997; 179: 382–8. Loewen PC, Sitwala J, Triggs-Raine BL. Catalases HPI and HP-II in Escherichia coli are induced independently. Arch Biochem Biophys 1985; 243: 144–9. Adams C, (1998) PhD Thesis, Herriot-Watt University, Edinburgh. Kono Y, Fridovich I. Isolation and characterisation of a pseudocatalase of Lactobacillus plantarum: A new manganese containing enzyme. J Biol Chem 1983; 258: 6015– 19. Allgood GS, Perry JJ. Characterisation of a manganese containing catalase from the obligate thermophile Thermophileum album. J Bacteriol 1986; 168: 563–7. Olivier G, Moore AR, Fildes J. Toxicity of Aeromonas salmonicida cells to Atlantic salmon Salmo salar peritoneal macrophages. Devel Comp Immunol 1992; 16: 49– 61. Crapo JD, McCord JM, Fridovich I. Preparation and assay of superoxide dismutase. Methods Enzymol 1978; 53: 382–93. Vipond R, Bricknell IR, Durrant E, Bowden TJ, Ellis AE, Smith M, MacIntyre S. Defined deletion mutants

A. C. Barnes et al.

33 34

35 36 37 38 39

40

41 42 43 44 45

46

demonstrate that major secreted toxins are not essential for the virulence of Aeromonas salmonicida. Infect Immun 1998; 66: 1990–8. Hirst ID, Hastings TS, Ellis AE. Siderophore production by Aeromonas salmonicida subsp. salmonicida. J Gen Microbiol 1991; 137: 1185–92. Hirst ID, Ellis AE. Iron-regulated outer membrane proteins of Aeromonas salmonicida are important protective antigens in Atlantic salmon against furunculosis. Fish Shell Immunol 1994; 4: 29–45. Halliwell B, Gutteridge JG. Oxygen toxicity, oxygen radicals, transition metals and disease. Biochem 1984; 219: 1–14. Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press 1989: Volume 3, A.3. Bowden TJ. (1998) PhD Thesis, University of Aberdeen, Department of Zoology, Aberdeen, Scotland. Ames GF, Prody C, Kustu S. Simple rapid and quantitative release of periplasmic proteins by chloroform. J Bacteriology 1984; 160: 1181–3. Ca´novas FM, Canto´n FR, Gallardo F, Garcı´a-Gutie´rrez A, de Vincente A. Accumulation of glutamine synthetase during the early development of maritime pine (Pinus pinaster) seedlings. Planta 1991; 185: 372–8. Balebona MC, Barnes AC, Horne MT, Ellis AE. Methods for identifying the intracellular location of superoxide dismutase in Photobacterium damsela sbsp. piscicida. In: Barnes AC, Davidson GA, Hiney MP, McIntosh D, Eds. Methodology in Fish Diseases Research. Aberdeen: Fisheries Research Services, 1998: 17–23. Barnes AC, Hastings TS, Amyes SGB. Amoxycillin resistance in Scottish isolates of Aeromonas salmonicida. J Fish Dis 1994; 17: 357–63. Beauchamp C, Fridovich I. Superoxide dismutase: improved assays and an assay applicable to acrylamide gels. Anal Biochem 1971; 44: 267–74. Rørth M, Jensen PK. Determination of catalase activity by means of the Clark oxygen electrode. Biochim Biophys Acta 1967; 139: 171–3. Barnes AC. (1992) PhD Thesis, University of Edinburgh Medical School, Edinburgh, Scotland. Secombes CJ. Isolation of fish macrophages and analysis of their killing activity. In: Stolen JS, Fletcher TC, Anderson DP, Roberson BS, van Muiswinkel WB, Eds. Techniques in Fish Immunology FITCI. New Jersey: SOS Publications, 1990: 137–54. Campos-Perez JJ, Ellis AE, Secombes CJ. Investigation of factors influencing the ability of Renibacterium salmoninarum to stimulate rainbow trout macrophage respiratory burst activity. Fish Shell Immunol 1997; 7: 555–66.