Blackwell Science, LtdOxford, UKMMIMolecular Microbiology0950-382X© 2005 The Authors; Journal compilation © 2005 Blackwell Publishing Ltd? 200558410251038Original ArticleAn anti-phagocyte exotoxin of Photobacterium damselae subsp. piscicidaA. do Vale et al.

Molecular Microbiology (2005) 58(4), 1025–1038

doi:10.1111/j.1365-2958.2005.04893.x First published online 29 September 2005

AIP56, a novel plasmid-encoded virulence factor of Photobacterium damselae subsp. piscicida with apoptogenic activity against sea bass macrophages and neutrophils Ana do Vale,1* Manuel T. Silva,1 Nuno M. S. dos Santos,1 Diana S. Nascimento,1 Pedro Reis-Rodrigues,1 Carolina Costa-Ramos,1 Anthony E. Ellis2 and Jorge E. Azevedo1,3 1 Institute for Molecular and Cell Biology, Rua do Campo Alegre, 823; 4150–180 Porto, Portugal. 2 FRS Marine Laboratory, PO Box 101, Victoria Road, Aberdeen, Scotland, UK. 3 Instituto de Ciências Biomédicas Abel Salazar, Largo do Prof Abel Salazar, 2; 4050 Porto, Portugal. Summary A strategy used by extracellular pathogens to evade phagocytosis is the utilization of exotoxins that kill host phagocytes. We have recently shown that one major pathogenicity strategy of Photobacterium damselae subsp. piscicida (Phdp), the agent of the widespread fish pasteurellosis, is the induction of extensive apoptosis of sea bass macrophages and neutrophils that results in lysis of these phagocytes by post-apoptotic secondary necrosis. Here we show that this unique process is mediated by a novel plasmid-encoded apoptosis inducing protein of 56 kDa (AIP56), an exotoxin abundantly secreted by all virulent, but not avirulent, Phdp strains tested. AIP56 is related to an unknown protein of the enterohemorrhagic Escherichia coli O157:H7 and NleC, a Citrobacter rodentium type III secreted effector of unknown function. Passive immunization of sea bass with a rabbit anti-AIP56 serum conferred protection against Phdp challenge, indicating that AIP56 represents a key virulence factor of that pathogen and is a candidate for the design of an anti-pasteurellosis vaccine. Introduction To establish successful infections, pathogens modulate host functions for their own advantage. One crucial, first Accepted 1 September, 2005. *For correspondence. E-mail [email protected]; Tel. (+351) 22 6074900; Fax (+351) 22 6099157.

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd

line, antibacterial host defence mechanism against extracellular pathogens is phagocytosis (Nahm et al., 1999). It is now well recognized that one of the strategies used by pathogens to evade phagocytosis is the utilization of molecules that efficiently kill host phagocytes (DeLeo, 2004). Many of these molecules kill these leukocytes by triggering their endogenous cell death machinery leading to an apoptotic process of cell elimination (DeLeo, 2004). This anti-phagocytic process is effected either through direct contact between the pathogen and the host cell target (Weinrauch and Zychlinsky, 1999) or at distance through the secretion of exotoxins, many of them belonging to a well-conserved, homogeneous family of proteins named RTX (for repeats in toxin) toxins (Narayanan et al., 2002). Photobacterium damselae subsp. piscicida (Phdp), is a Gram-negative, extracellular pathogen isolated for the first time from a massive fish kill in Chesapeake Bay (Snieszko et al., 1964). The disease was named pasteurellosis after the initial classification of this agent as Pasteurella piscicida (Janssen and Surgalla, 1968). The pathogen was later reassigned to the genus Photobacterium as Photobacterium damsela subsp. piscicida (Gauthier et al., 1995), name afterwards corrected to Photobacterium damselae subsp. piscicida (Trüper and de’Clari, 1997). Phdp has been found to affect more than 20 warm water fish species worldwide, including sea bass, sole, yellowtail, gilt-head sea bream and turbot both in natural environments and in aquacultures (Thune et al., 1993; Romalde, 2002). Pasteurellosis is now recognized as one of the most threatening bacterial diseases in mariculture worldwide due to its wide host range, massive mortality, ubiquitous geographical distribution, widespread antibiotic resistance and lack of efficient vaccines (Barnes et al., 2005). In its acute form, this disease has a septicaemic nature and has a short course (Thune et al., 1993; Romalde, 2002). As a successful extracellular pathogen, Phdp must be able to overcome the host phagocytic mechanisms. The ability of Phdp to avoid phagocytosis and thus to cause disease, may be explained by our recent results showing that, in experimental pasteurellosis of sea bass, virulent Phdp strains induce extensive apoptosis of macrophages and neutrophils present in high numbers in Phdp-infected foci (do Vale et al., 2002), resulting in lysis

1026 A. do Vale et al. of these leukocytes by post-apoptotic secondary necrosis (do Vale et al., 2003). Considering that apoptosis of macrophages and neutrophils was observed after injection of virulent bacterial culture supernatants but not of UV-killed virulent bacteria, and because the apoptogenic activity of culture supernatants was abolished by heat-treatment, it was proposed that one (or several) protein(s) secreted by virulent Phdp strains could be responsible for the apoptogenic activity of these bacteria (do Vale et al., 2003). Here we identified and characterized AIP56, a protein abundantly secreted by virulent strains of Phdp, as the apoptogenic factor responsible for the induction by Phdp of apoptosis of sea bass macrophages and neutrophils. Evidence showing that AIP56 represents a novel, key virulence molecule of Phdp is presented.

Results The apoptogenic activity of Phdp towards sea bass macrophages and neutrophils correlates with the secretion of a major 56 kDa protein SDS-PAGE analysis of mid-exponential phase culture supernatants from Phdp virulent strain MT1415 and avirulent strain ATCC29690, revealed that the patterns of secreted proteins are quite different (Fig. 1A). A major protein band of approximately 56 kDa, representing approximately 65%, by weight, of total protein, was present only in the culture supernatant from the virulent strain (Fig. 1A, lane 1). As expected from these results, native-PAGE analysis of MT1415 culture supernatants also revealed a simple protein profile (Fig. 1B). As a first step to identify the apoptogenic factor secreted by Phdp, concentrated culture supernatants from strain MT1415 were resolved by native-PAGE, fractionated, and protein fractions were tested in vivo for the capacity to induce apoptosis of sea bass macrophages and neutrophils. Fractions I (FI), II (FII) and III (FIII) (see Fig. 1B) were injected intraperitoneally (i.p.) in sea bass and apoptosis of peritoneal macrophages and neutrophils assessed 6 h after injection. Apoptosis of the two phagocytes, as detected by the occurrence of cell shrinkage, chromatin condensation, nuclear fragmentation, TUNEL-positive nuclei and formation of apoptotic bodies, was observed in peritoneal exudates collected from fish injected with FI, but not with FII or FIII (not shown). SDS-PAGE analysis of FI fraction revealed a single protein band of approximately 56 kDa (data not shown). This observation suggested that this protein could be the apoptogenic factor produced by Phdp virulent strains. The protein was named AIP56 (apoptosis inducing protein of 56 kDa). However, the biological activity observed when using fraction FI from the native gel could be due, not to the 56 kDa protein,

Fig. 1. A. SDS-PAGE profiles of culture supernatants from Photobacterium damselae subsp. piscicida (Phdp) strains MT1415 (virulent) and ATCC29690 (non-virulent). Total proteins from mid-exponential culture supernatants were prepared as described in Experimental procedures, separated by 10% SDS-PAGE and stained with Coomasie-blue. Lanes: 1, strain MT1415; 2, strain ATCC29690. B. Native-PAGE analysis of culture supernatant proteins from Phdp virulent strain MT1415. Concentrated supernatants were separated by native-PAGE and stained with Coomassie-blue. The regions of the gel containing protein fractions tested for apoptogenic activity in vivo are indicated (FI, FII and FIII).

but rather to a minor (undetectable) protein present in this fraction. To clarify this issue, the gene encoding AIP56 was cloned, expressed in Escherichia coli and the recombinant protein was tested for apoptogenic activity, as described below. Cloning of the gene encoding AIP56 As a first step to clone the gene encoding AIP56, the purified protein was subjected to Edman degradation. Sequencing of two HPLC-purified tryptic fragments yielded the following sequences: NNDKPDASDDKY ADYVVR and YTAAATEYTVIDALFHSPTFR (regions in bold were used to design degenerate primers A1upper and Blower respectively). Total DNA from the MT1415 (virulent) and ATCC29690 (avirulent) strains of Phdp were then subjected to polymerase chain reaction (PCR) amplifications using the primers A1upper and Blower. A 200 bp DNA fragment was amplified only when total DNA from the virulent strain was used as a template (data not shown). Using the PCR-derived 200 bp DNA fragment described above as the starting point in a multistep cloning strategy, the complete DNA sequence encoding AIP56 was

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

An anti-phagocyte exotoxin of Photobacterium damselae subsp. piscicida obtained (GenBank Accession number DQ066884; see Experimental procedures for details). AIP56 is 513 amino acid residues long and displays a hydropathic profile typical of a non-membrane protein (data not shown). Analysis of the primary structure of AIP56 using the SignalP-NN, version 3.0 computer program (http://www.cbs.dtu.dk/ services/SignalP/) suggested the existence (although with a poor score) of a putative signal peptide with a cleavage site between amino-acid residues 16 and 17. Fortuitously, one of the sequences obtained by Edman degradation of tryptic peptides from AIP56 starts at amino-acid residue 17, an asparagine. Considering that asparagine 17 is preceded by an alanine in the primary structure of AIP56 and that trypsin cleaves peptide bonds on the carboxyl side of lysine and arginine residues only, it can be concluded that the N-terminus of this peptide did not originate from the action of trypsin. Thus, these data strongly suggest that asparagine 17 represents the N-terminus of the secreted protein (additional data supporting this conclusion are presented below). The predicted molecular mass of the mature form of AIP56 is 56.185 kDa, which is in perfect agreement with the value of 56 kDa estimated by SDSPAGE (see Fig. 1A). Database searches using the primary structure of AIP56 showed that the fragment corresponding to the first 324 amino-acid residues of AIP56 has 30% identity with an unknown protein of the enterohemorrhagic E. coli O157:H7 encoded by cryptic prophage CP-933K (NCBI Accession number NP_286533) and 31% identity with NleC (NCBI Accession number AAS47019), a Citrobacter rodentium type III secreted effector of unknown function. (The latter two proteins are 90% identical). Interestingly, the most conserved region of the three bacterial proteins contains the zinc-binding region signature HEXXH typical of most zinc metallopeptidases [see http://us.expasy.org/ prosite; (Jongeneel et al., 1989; Hase and Finkelstein, 1993)]. Whether or not these proteins display peptidase activity remains to be elucidated. No homologies to other proteins were detected. Recombinant AIP56 expressed in E. coli retains the apoptogenic activity of Phdp AIP56 Plasmid pETAIP56H+ was used to express His-tagged AIP56 (AIP56H+) in BL21 E. coli cells. SDS-PAGE analysis of BL-21 E. coli cells induced at 37°C revealed a robust expression of a non-soluble (present in the inclusion bodies fraction) 58 kDa protein (Fig. 2, lanes 6 and 7). The apparent molecular mass of this protein is 2 kDa higher than the one displayed by authentic AIP56 (compare lanes 7 and 8 in Fig. 2), indicating that the signal sequence of the precursor form of AIP56 was not cleaved. No apoptotic effect was observed when the insoluble 58 kDa protein was i.p. injected into sea bass (not shown). SDS-PAGE

1027

Fig. 2. Heterologous expression of AIP56. SDS-PAGE analysis of total lysates from BL21 Escherichia coli carrying the pETAIP56H+ collected before (lanes 1 and 5) and after (lanes 2 and 6) induction with IPTG at 17°C (lanes 1 and 2) or 37°C (lanes 5 and 6); lane 3, His-tagged AIP56 (AIP56H+) purified from the soluble fraction of BL21 cells induced at 17°C; lane 7, insoluble fraction from BL21 cells induced at 37°C; lanes 4 and 8, concentrated culture supernatants from Photobacterium damselae subsp. piscicida MT1415. The precursor (pre-AIP56) and mature (AIP56) forms of AIP56 are indicated. The numbers on the left indicate the position of the molecular weight standards (in kilodaltons).

analysis of BL-21 E. coli cells harbouring pETAIP56H+ induced at 17°C also revealed a strong expression of AIP56H+ (Fig. 2, lane 2). However, AIP56H+ produced by these cells at this temperature displayed the apparent molecular mass of authentic AIP56 (compare lanes 3 and 4 in Fig. 2) suggesting that the signal peptide of the precursor protein was correctly processed. Furthermore, the heterologously expressed AIP56H+ was found in the soluble fraction obtained after centrifugation of sonicated cells suggesting that AIP56H+ expressed at 17°C by BL21 E. coli cells was correctly folded and located at the periplasmic compartment. Most importantly, when purified AIP56H+ (Fig. 2, lane 3) was injected i.p. into sea bass, high numbers of apoptotic neutrophils and macrophages in the peritoneal cavities were observed 4 h after injection (Fig. 3). The apoptotic nature of the alterations seen in these leukocytes was established by the early occurrence of a set of indicators (Wyllie et al., 1980; Pallardy et al., 1999; Willingham, 1999; Vermes et al., 2000): cell shrinkage, chromatin condensation, nuclear fragmentation and production of apoptotic bodies (Fig. 3A), internucleosomal DNA degradation revealed by nuclear TUNEL positivity (Fig. 3B) and DNA electrophoresis (Fig. 3C), and activation of the apoptosis executioner caspase-3 (Fig. 3D). No increase in the very low number of apoptotic cells found in resting peritoneal cavities was detected by microscopy (Fig. 3A), TUNEL staining (Fig. 3B) and DNA electro-

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

1028 A. do Vale et al.

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

An anti-phagocyte exotoxin of Photobacterium damselae subsp. piscicida

1029

Fig. 3. Heterologously expressed AIP56 displays apoptogenic activity. A. Wright-stained cytospins of sea bass peritoneal leukocytes 4 h after i.p. injection of 2 µg of His-tagged AIP56 (AIP56H+) or the same amount of heat-inactivated AIP56H+ (control). Neutrophils were labelled by peroxidase detection (brown granules). Notice in leukocytes exposed to AIP56H+ the occurrence of extensive nuclear fragmentation and chromatin condensation in macrophages (M) and neutrophils (N), cell blebbing in neutrophils, and the presence of apoptotic bodies (AB). The eosinophilic granular cells (E) look normal in both samples. B. Fluorescence microscopy of sea bass peritoneal exudates processed for the detection of DNA fragmentation by TUNEL staining. Sea bass peritoneal leukocytes were collected 4 h after injection of 2 µg of purified AIP56H+ or of soluble extracts of BL21 cells (Control) and processed for TUNEL staining. Notice the extensive nuclear condensation and fragmentation of TUNEL-positive nuclei. C. Agarose gel electrophoresis of DNA extracted from peritoneal leukocytes of sea bass collected 4 h after injection of 2 µg of AIP56H+ (lanes 1–3) or of soluble extracts of BL21 cells (lanes 4–6). Each lane contains DNA from an individual fish. Numbers on the left indicate the position of the molecular weight standards (in base pairs). D. Caspase-3 activity, determined using the fluorescent substrate Ac-DEVD-AMC, in lysates from sea bass peritoneal cells collected 4 h post injection of 2 µg of His-tagged AIP56 (AIP56H+) or the same amount of heat-inactivated AIP56H+ (Control). Lysates incubated with the caspase3 inhibitor Ac-DEVD-CHO were used as specificity controls.

phoresis (Fig. 3C) after injection of soluble extracts from E. coli BL21 cells carrying the pET28a(+) empty vector or heat-inactivated AIP56H+. Accordingly, no significant activation of caspase-3 was detected in control cell suspensions (Fig. 3D). The apoptotic effect detected in this assay was morphologically indistinguishable from the one obtained after injection of AIP56 secreted by Phdp, or of live virulent bacteria or their culture supernatants (see do Vale et al., 2003). The apoptogenic activity of AIP56 was also detected when the toxin was incubated ex vivo with sea bass phagocytes from resting (containing 68% phagocytes) or Incomplete Freund’s Adjuvant injected (containing 85% phagocytes) peritoneal cavities. This effect was also seen with washed peritoneal leukocyte suspensions. These results suggest that the apoptogenic activity of AIP56 is effective in the absence of peritoneal mesothelial cells and fish serum factors. These in vivo and ex vivo results demonstrate that AIP56 is the factor responsible for the apoptogenic activity of virulent Phdp strain MT1415. AIP56 and AIP56-related proteins are encoded in multicopy plasmids in all tested virulent strains of Phdp During the Southern blotting experiments carried out to

clone the gene encoding AIP56 from the virulent MT1415 Phdp strain, we noticed that probe-reactive DNA fragments were always superimposable with sharply defined and abundant DNA bands on the ethidium bromide stained gels. These observations suggested that multiple copies of the AIP56 gene could be present in this strain, perhaps as part of a multicopy plasmid. In order to test this hypothesis, total plasmid DNA was isolated from the MT1415 Phdp strain using the alkaline lysis method (Sambrook and Russell, 2001). Agarose-gel electrophoresis of this DNA preparation revealed the existence of a major plasmid (see Fig. 4A, upper panel, lane 3). Southern-blotting analysis of the restriction enzyme cleaved plasmid revealed the presence of the AIP56 gene (data not shown). The entire plasmid (hereafter referred to as pPHDP10) was sequenced. pPHDP10 comprises 9631 bp and contains several open reading frames (ORFs), as predicted by the GeneMark.hmm 2.1 computer program (http://opal.biology.gatech.edu/GeneMark/). In addition to the AIP56 encoding region, ORFs coding for a putative chorismate mutase, several putative transposases, a putative resolvase and four hypothetical proteins were found in pPHDP10 plasmid. Previously, we reported that several virulent strains of Phdp isolated from different fish species at different geo-

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

1030 A. do Vale et al.

Fig. 4. A. AIP56 is a plasmid-encoded protein. Plasmid DNA extracted from non-virulent (lanes 1 and 2) and virulent (lanes 3–10) Photobacterium damselae subsp. piscicida (Phdp) strains were separated by 1% agarose gel electrophoresis and either stained with ethidium bromide (upper panel) or subjected to Southern blotting with an AIP56 gene fragment as a probe (lower panel). Lanes 1–10: strains ATCC29690, EPOY 8803II, MT1415, B51, PP3, MT1375, DI21, MT1588, MT1594 and PTAVSA95 respectively. The numbers on the left indicate the position of the molecular weight standards (in kilobases). B. AIP56 gene is present in all virulent strains tested and absent in the two non-virulent isolates. Total genomic DNA extracted from Phdp nonvirulent (lanes 1 and 2) and virulent (lanes 3–10) was subjected to PCR with AIP56-specific primers. Lanes 1–10: strains ATCC29690, EPOY 8803-II, MT1415, B51, PP3, MT1375, DI21, MT1588, MT1594 and PTAVSA95 respectively. M, molecular weight standard. The numbers on the left indicate the position of the molecular weight standards (in kilobases). C. AIP56-related proteins are secreted by several virulent strains of Phdp. Total proteins from exponential phase culture supernatants of virulent (lanes 1–8) and non-virulent (lanes 9 and 10) Phdp strains were separated by 12% SDS-PAGE and either stained with Coomasie-blue (upper panel) or subjected to Western blotting analysis with the anti-AIP56 rabbit antiserum (lower panel). Lanes 1–10: strains MT1415, B51, PTAVSA95, PP3, MT1375, MT1588, MT1594, DI21, ATCC 29690 and EPOY 8803-II respectively. The numbers on the left indicate the position of the molecular weight standards (in kilodaltons).

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

An anti-phagocyte exotoxin of Photobacterium damselae subsp. piscicida graphical locations display apoptogenic activity towards sea bass macrophages and neutrophils (do Vale et al., 2003). In order to determine if these strains also contain plasmid(s) encoding AIP56, we extended the Southern blotting analysis to these isolates. The results of this experiment are presented in Fig. 4A. Most virulent strains contain a plasmid displaying the electrophoretic mobility of pPHDP10 (compare lanes 3 and 4, 7, 8, 9, 10 of the upper panel). In addition, similar DNA patterns were obtained when these plasmids were digested with EcoRI, HindIII and BamHI restriction enzymes (data not shown). Importantly, all these plasmids hybridized with a labelled DNA fragment containing nucleotides 435–1269 of the AIP56 gene (Fig. 4A, lower panel). Thus, all these strains contain pPHDP10 or a pPHDP10-like plasmid. The two avirulent strains ATCC29690 and EPOY 8803-II and the virulent isolates MT1375 and PP3 showed a more complex plasmid profile (Fig. 4A, upper panel, lanes 1, 2, 5 and 6). However, Southern blotting analysis revealed the presence of plasmids containing AIP56-coding sequences in the two virulent strains (Fig. 4A, lower panel, lanes 5 and 6) but not in the two avirulent strains (Fig. 4A, lower panel, lanes 1 and 2). The findings regarding the presence of AIP56-coding sequences in all virulent strains tested and their absence in the avirulent strains were also confirmed by PCR, this time using total Phdp DNA. As shown in Fig. 4B, DNA fragments of the expected size were obtained only when DNA isolated from the virulent strains was used in this assay. Finally, the presence of AIP56 in the culture supernatants of the strains described above was also determined (see Fig. 4C). In contrast to the results obtained with the two avirulent strains (upper panel, lanes 9 and 10), a major protein band comigrating with AIP56 from the MT1415 strain was detected in five of the virulent strains analysed (upper panel, lanes 2–4, 7 and 8); virulent strains PP3 and MT1375 secrete high amounts of a slightly smaller (about 54 kDa) protein (upper panel, lanes 5 and 6). Most importantly, all these abundantly secreted proteins were recognized by the anti-AIP56 rabbit serum (Fig. 4C, lower panel). The reason for the smaller size of the AIP56 proteins in strains PP3 and MT1375 is presently unknown. Such discrepancy may arise from subtle differences in the AIP56-encoding genes present in these strains or from some different proteolytic cleavage of the AIP56 proteins. Passive immunization with anti-AIP56 rabbit serum protects sea bass against Phdp infection To further evaluate the importance of AIP56 as a virulence factor of Phdp in vivo, the rabbit antibody directed to AIP56 was used in passive immunization tests. In one

1031

experiment, two groups of eight fish were i.p. injected with either 100 µl or 300 µl of first bleed rabbit serum per fish; a third group of eight fish were injected with 300 µl of control serum per fish. In another experiment, one group of eight fish was i.p. injected with 300 µl of first bleed rabbit serum per fish, another group of eight fish with 300 µl of second bleed rabbit serum per fish and the third group with 300 µl of control serum per fish. Fish were challenged with 2.1 ± 0.2 × 107 colony-forming units (cfu) of Phdp virulent strain PTAVSA95 immediately following passive immunization. The results of these tests are shown in Fig. 5. The first deaths occurred in day 1 post challenge, while the last death occurred in day 6 post challenge. As no further mortalities occurred for eight consecutive days, the trial was terminated 15 days after immunization and challenge. A significant protection against Phdp infection was obtained when anti-AIP56specific rabbit serum was used to passively immunize sea bass. This protection was dose-dependent, as lower percentage mortalities were obtained when 300 µl of immune serum were used instead of 100 µl (Fig. 5A). Second bleed anti-AIP56 antiserum, with an eightfold higher titre of anti-AIP56-specific antibodies than first bleed serum (see Experimental procedures), was more effective in protecting fish against Phdp infection (Fig. 5B). These results indicate that AIP56 is a key virulence molecule of Phdp. Discussion In this work we report the isolation and characterization of AIP56, a protein exotoxin abundantly secreted by virulent Phdp, as the factor responsible for the recently reported apoptogenic activity of that pathogen towards sea bass macrophages and neutrophils (do Vale et al., 2003). This apoptogenic activity destroys sea bass neutrophils and macrophages by a highly efficient mechanism involving the extensive lysis of the phagocytes by postapoptotic secondary necrosis (do Vale et al., 2003). Since all virulent, but not avirulent, strains of Phdp analysed secrete high amounts of AIP56 or AIP56-related proteins, secretion of this exotoxin appears as a common pathogenicity mechanism used by virulent strains of Phdp. The observation that passive immunization with anti-AIP56 antibodies protects sea bass against Phdp infection indicates that AIP56 is a key virulence molecule of this pathogen. The importance of this novel pathogenicity factor of Phdp is emphasized by the observation that AIP56induced phagocyte destruction also occurs in sea bass and sole with natural pasteurellosis (A. do Vale, unpubl. results). AIP56 appears as a major protein band in the electrophoretic profiles of exponential culture supernatants of virulent Phdp. It should be noted that the low complexity of the electrophoretic profiles of these supernatants is in

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

1032 A. do Vale et al. Fig. 5. Passive immunization of sea bass with anti-AIP56 antibodies protects against Photobacterium damselae subsp. piscicida (Phdp) infection. Cumulative mortality of sea bass passively immunized with rabbit anti-AIP56 serum and challenged with virulent Phdp strain PTAVSA95. A. Dose effect and protection using rabbit antiAIP56 serum from first bleed. B. Titre effect and protection using rabbit antiAIP56 serum from first and second bleed. (
) 100 µl of rabbit anti-AIP56 serum from first bleed per fish; () 300 µl of rabbit anti-AIP56 serum from first bleed per fish; () 300 µl of rabbit anti-AIP56 serum from second bleed per fish; () 300 µl of rabbit preimmune serum per fish.

contrast to previous studies suggesting the secretion of a highly complex mixture of different proteins by Phdp (Bakopoulos et al., 1997; Mazzolini et al., 1998). The larger protein complexity of the culture supernatants observed by those researchers could be explained by the fact that stationary-phase cultures were used in those studies. In contrast, in the present study we used midexponential cultures. Indeed, we found that the electrophoretic profiles of Phdp culture supernatants become more complex as the cultures approached the stationary phase (not shown). This is probably due not only to the accumulation in the culture supernatant of proteins secreted during growth but also to the proteolytic degradation of those proteins. As observed for many bacterial toxin-specific genes, which are often located on mobile genetic elements (reviewed in Hacker et al., 1997), AIP56 was also found to be encoded in a plasmid. The location of AIP56 gene in a high-copy plasmid is probably the reason why Phdp

virulent strains are capable of producing such large amounts of exotoxin. Protein and nucleotide sequence data revealed that AIP56 is synthesized as a precursor protein 513 aminoacid residues long that possesses an N-terminal cleavable signal peptide of 16 residues. Heterologous expression of AIP56 in E. coli cells confirmed this interpretation: insoluble AIP56 expressed at 37°C in E. coli BL21 cells behaved as a 58 kDa protein upon SDS-PAGE; soluble, biologically active AIP56 expressed at 17°C by the same E. coli cells displayed an apparent molecular mass of 56 kDa implying that the signal peptide was correctly cleaved. Thus, taken together, these data suggest that AIP56 is secreted by Phdp by a two step process which probably begins with Sec-dependent, signal peptidemediated translocation across the cytoplasmic membrane to the periplasmic space (reviewed in Hueck, 1998). Whether translocation of AIP56 across the outer membrane involves the terminal branch of the general secre-

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

An anti-phagocyte exotoxin of Photobacterium damselae subsp. piscicida tion pathway (type II secretion), a type IV pathway or even a two-partner secretion system (reviewed in Hueck, 1998; Burns, 2003; Sandkvist, 2001; Ma et al., 2003) awaits further elucidation. Several pathogenic bacteria secrete cytocidal exotoxins active against host immune cells as an efficient virulence mechanism (Narayanan et al., 2002). AIP56 exotoxin has no homology with any of those molecules. However, the domain comprising the first 324 amino-acid residues of AIP56 is homologous to NleC, a C. rodentium type III secreted effector of unknown function (Deng et al., 2004) and to an enterohemorrhagic E. coli O157:H7 unknown protein encoded by cryptic prophage CP-933K. Although this finding per se does not provide a clue about the function of these homologous proteins, the fact that a zincbinding region signature typical of most zinc metallopeptidases is present in the most conserved regions of the three proteins strongly suggests that this domain is important for their functions. It is noteworthy that many bacterial toxins possessing different biological activities are members of the metalopeptidase family (reviewed in Jongeneel et al., 1989; Hase and Finkelstein, 1993). In the context of our work, the lethal factor LeTx from Bacillus anthracis is a particularly interesting example: this zinc-dependent protease cleaves mitogen-activated proteinase kinase kinase leading to apoptotic cell death of murine macrophage cell lines and human peripheral blood mononuclear cells (Popov et al., 2002a,b). The pathogenicity of Phdp certainly is multifactorial. However, the AIP56-dependent pathogenicity mechanism most likely is central in the etiopathogenesis of fish pasteurellosis. Experimental (do Vale et al., 2003) and natural (Thune et al., 1993; Romalde, 2002) acute pasteurelloses are rapid, severe infections with a very high mortality. This may be related to the characteristics of this AIP56-dependent Phdp pathogenicity mechanism that indicate that it is a highly potent one, as follows. First, using an exotoxin secreted in high amounts, this mechanism can operate at distance without requiring contact between the pathogen and the target cells. Second, this pathogenicity mechanism uses a weapon with multiple effects that converge to simultaneously promote evasion of the pathogen from host defences and produce tissue injury. Indeed, by inducing apoptosis followed by secondary necrotic lysis of both macrophages and neutrophils, Phdp not only impairs the participation of these leukocytes in phagocytosis and killing of Phdp, but also releases intraphagocytic bacteria (do Vale et al., 2003), two effects that promote survival and extracellular multiplication of Phdp. Additionally, destruction of macrophages and neutrophils would affect cytokine production by these leukocytes, further impairing the host immune response. Concomitantly, this destruction leads to tissue-damage with deleterious consequence for the host. In fact, destruction of macrophages, the cells with

1033

the crucial role of eliminating apoptotic cells (Maderna and Godson, 2003), results in deficient clearance of AIP56induced abundant apoptotic neutrophils which leads to their lysis by secondary necrosis (do Vale et al., 2003), with the severe consequences of the release of highly cytotoxic neutrophil molecules (Henson and Johnston, 1987). Finally, relying on a protein exotoxin, this pathogenicity mechanism requires the production of anti-AIP56 antibodies as a defensive response from the host, a response that cannot be achieved in the short time lapse of the disease. Bacterial induction of phagocyte cell death is now a well recognized pathogenicity mechanism (Weinrauch and Zychlinsky, 1999; Narayanan et al., 2002; DeLeo, 2004). However, the AIP56-dependent phagocyte destruction by Phdp seems to represent a unique modality since no other instance of in vivo induction of extensive, simultaneous lysis of both professional phagocytes has been reported so far (DeLeo, 2004). The present identification, characterization and heterologous expression of AIP56 will allow to unravel the molecular mechanism of the unique virulence strategy of Phdp depending on AIP56-induced apoptosis, and may lead to a better understanding of the EHEC O157:H7 pathogenicity. Our results indicating that neutralization of AIP56 activity by transfer of rabbit antibodies against this protein is effective in protecting sea bass against experimental Phdp infection not only confirm AIP56 as a key virulence molecule of Phdp but also open very promising prospects towards the development of an effective anti-pasteurellosis vaccine based on AIP56.

Experimental procedures Experimental fish Sea bass (Dicentrarchus labrax) were purchased from a commercial fish farm and were maintained in recirculating aerated seawater, at 18–19°C. Fish used for passive vaccination and challenge experiments were previously acclimatized to 23 ± 1°C. Water quality was maintained with mechanical and biological filtration and UV-disinfection and fish were fed ad libitum on commercial pellets.

Bacteria The origin and virulence of the Phdp strains used in this study are listed in Table 1. Strains DI21, B51, and EPOY 8803-II were provided by Prof. Alicia E. Toranzo (Departamento de Microbiología y Parasitología, Facultad de Biologia, University of Santiago de Compostela, Spain), strains MT1415, PP3, MT1375, MT1588 and MT1594 were provided by Dr Andrew C. Barnes (Marine Laboratory, Aberdeen, UK). Strain PTAVSA95 is from our collection. Strain ATCC 29690 was obtained from the American Type Culture Collection, USA. The virulence of each isolate was previously determined (do

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

1034 A. do Vale et al. Table 1. Origin and virulence of the Photobacterium damselae subsp. piscicida strains used in this study.

Strain

Origin

Virulence for sea bass

MT1415 PP3 MT1375 MT1588 MT1594 PTAVSA95 DI21 B51 EPOY 8803-II ATCC29690

Dicentrarchus labrax, Italy Seriola quiqueradiata, Japan Dicentrarchus labrax, Italy Dicentrarchus labrax, Greece Sparus aurata, Greece Sparus aurata, Portugal Sparus aurata, Spain Dicentrarchus labrax, Spain Epinephelus akaara, Japan Seriola quiqueradiata, Japan

+ + + + + + + + – –

Vale et al., 2003). Bacteria were routinely cultured at 22°C in tryptic soy broth (TSB) or tryptic soy agar (TSA) (both from Difco) supplemented with NaCl to a final concentration of 1% (w/v) (TSB-1 and TSA-1 respectively). Stocks of these strains were maintained at −80°C in TSB-1 supplemented with 15% (v/v) glycerol.

Fractionation of culture supernatant proteins by native-PAGE Concentrated culture supernatants from Phdp virulent strain MT1415 were diluted 1:1 in 2 × native-PAGE sample buffer (the same composition of SDS-PAGE sample buffer, except that no SDS is included and the concentration of β-mercapthoethanol is reduced to 5 mM) and separated in a 10% native-PAGE (as for SDS-PAGE, except that SDS was omitted from all solutions). The lanes on the extremities of the gels were cut, stained with Comassie-blue and used to locate the position of the main protein bands. Proteins were extracted from minced gel slices by diffusion (overnight at 4°C with gentle agitation) using 20 mM Tris-HCl (pH 8.0) as the elution buffer. Protein fractions were tested in vivo for the ability to induce apoptosis of sea bass peritoneal phagocytes, as previously described (do Vale et al., 2003). Briefly, groups of 10 sea bass weighing 126.4 ± 16.6 g were injected i.p. with 200 µl of the eluted native protein fractions, leukocytes were collected from peritoneal cavities 6 h after injection, by a procedure described in detail elsewhere (Afonso et al., 1997; do Vale et al., 2002) and occurrence of apoptosis assessed as described below.

Preparation and analysis of Phdp culture supernatants Bacterial strains were grown in TSB-1 at 22°C with shaking (100 r.p.m.) to an optical density (OD) at 600 nm of approximately 0.6 (mid-exponential phase). Bacterial cells were removed by centrifugation (3000 g, 30 min, 4°C) and the culture supernatants subsequently filtered through 0.22 µmpore-size filter (Schleicher and Shuell). Cell-free supernatants from strains MT1415 (virulent) and ATCC29690 (avirulent), prepared as described above, were concentrated 100fold using a Vivaflow 200 concentrator (Sartorius), dialysed against 20 mM Tris-HCl (pH 8.0) and stored at −80°C. The sterility of the culture filtrates was confirmed by the absence of colonies after plating on TSA-1 plates. Direct SDS-PAGE analysis of filtered culture supernatants from Phdp was performed after trichloroacetic acid (TCA) precipitation, as follows. Proteins from 1.5 ml aliquots of cellfree culture supernatants were precipitated with 10% (w/v) TCA for 30 min on ice and recovered by centrifugation. Protein pellets were washed in 10% (w/v) TCA, recovered by centrifugation, washed with acetone, allowed to dry, solubilized in SDS-sample buffer and subjected to SDS-PAGE. The latter was performed in 10% or 12% polyacrylamide gels using the Laemmli discontinuous buffer system (Laemmli, 1970). To determine the percentage of the total protein of the bacterial culture supernatants corresponding to AIP56, Coomassie-blue stained SDS-PAGE gels were analysed by densitometry using the UN-SCAN-IT automated digitizing system. Western blotting onto nitrocellulose membranes (Schleicher and Schuell) was performed according to the manufacturer’s instructions. Immunostaining was performed with a 1:15000 (v/v) dilution of a rabbit antiserum against AIP56 from strain MT1415 (see below). Rabbit IgGs on westernblots were detected with a donkey anti-rabbit IgG horseradish peroxidase-linked secondary antibody (Amersham Biosciences) using an enhanced chemiluminescence kit (Pierce Biotechnology).

Determination of the apoptogenic activity of proteins Apoptosis was detected by the following techniques that, when used together, allow the safe identification of apoptosis by the presence of a set of indicators (Wyllie et al., 1980; Pallardy et al., 1999; Willingham, 1999; Vermes et al., 2000), namely chromatin condensation, nuclear fragmentation, cell shrinkage, cell blebbing with formation of apoptotic bodies, internucleosomal DNA degradation and caspase-3 activation. For routine leukocyte morphologic analysis, cytospin preparations were fixed (10% of 37% formaldehyde in absolute ethanol; 45 s) and stained with Wright’s stain (Haemacolor, Merck) after labelling the neutrophils by peroxidase detection using the Antonow’s technique (Afonso et al., 1998; do Vale et al., 2002). To assess internucleosomal DNA degradation, we used terminal deoxytransferase-mediated dUTP nick end labelling (TUNEL) of DNA strand breaks (In Situ Cell Death Detection Kit, Fluorescein; Roche Applied Sciences) in samples fixed with formaldehyde-ethanol as above. Stained preparations were mounted with the antifading agent Vectashield (Vector Laboratories) containing 4 µg ml−1 propidium iodide as a nuclear stain. Detection of internucleosomal DNA fragmentation by agarose gel electrophoresis was carried out as previously described (do Vale et al., 2003). Caspase-3 activity was determined using a fluorimetric caspase-3 assay kit (Sigma-Aldrich), following the manufacturer’s instruction. Briefly, soluble extracts were prepared from 1.5 × 106 peritoneal leukocytes by incubation for 20 min on ice with 50 µl of lysis buffer. Cell debris were eliminated by centrifugation and 18 µl of the supernatant assayed for caspase-3 activity by incubation with the fluorogenic substrate acetyl-Asp-Glu-Val-Asp-7-amido-4-methylcoumarin (Ac-DEVD-AMC). Cell lysates treated with the caspase-3 inhibitor acetyl-Asp-Glu-Val-Asp-aldehyde (Ac-DEVD-CHO) were used as controls for measuring the non-specific hydrolysis of the substrate. The release of fluorescent amidomethylcoumarin was measured by fluorometry in a Spectra Max

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

An anti-phagocyte exotoxin of Photobacterium damselae subsp. piscicida Gemini XS fluorimeter (Molecular Devices) and results were expressed as relative fluorescence units (RFU). For each experimental situation at least six fish were used and the experiments were repeated twice. When indicated, statistical analysis of the data was performed using the Student t-test and P < 0.05 was considered significant.

Preparation of an antiserum against AIP56 from strain MT1415 Two millilitre of concentrated culture supernatants from strain MT1415 were subjected to Coomassie-blue SDS-PAGE (Schagger et al., 1988). After electrophoretic separation, the 56 kDa band was excised from the gel, minced in elution buffer (0.02% SDS, 10 mM β-mercapthoethanol, 70 µg ml−1 PMSF) and incubated overnight at 4°C with shaking. The acrylamide suspension was then centrifuged at 3000 g for 15 min at 4°C. The supernatant was collected and centrifuged again under the same conditions. The supernatant was collected, frozen at −80°C and lyophilized. The sample was resuspended in 2 ml of distilled water and the protein was then precipitated with acetone (90% v/v) overnight at −20°C. The precipitated protein was recovered by centrifugation at 3000 g for 10 min at 4°C, washed with 90% (v/v) acetone, dried overnight at room temperature and resuspended in PBS. Immunization of a 13 week old female New Zealand rabbit was performed according to established procedures (Harlow and Lane, 1998). The rabbit received three injections of purified 56 kDa protein and was bled after the first boosting (first bleed) and after the second boosting (second bleed). The serum was tested by Western Blotting for its specificity using MT1415 concentrated extracellular products (ECPs) as antigens. The relative titres of the first and second bleed sera were determined by a dot blot assay. MT1415 concentrated supernatants (approximately 10 µg protein) were diluted in Tris-buffered saline (TBS) and were applied to a nitrocellulose membrane using a dot blot manifold. After blocking in 5% (w/ v) low fat powdered milk in TBS containing 0.1% Tween 20 (TTBS), serial twofold dilutions (starting at a dilution of 1:100) of anti-AIP56 rabbit sera from the first and second bleeds as well as of non-immune rabbit serum were applied to the wells. Wells not incubated with rabbit serum were included as controls. Goat anti-rabbit Ig conjugated with alkaline phosphatase was used as the secondary antibody. Detection of the secondary antibody was performed using 5-bromo-4-chloro-3indolyl phosphate/nitro blue tetrazolium (BCIP/NBT).

Partial amino-acid sequence analysis of AIP56 Concentrated supernatants from strain MT1415 were subjected to SDS-PAGE and the 56 kDa protein band was excised from the gel after Coomassie-blue staining. In situ trypsin digestion of the protein and Edman degradation of two HPLC-purified peptides was performed by HHMI/Keck Biotechnology Resource Laboratory (New Haven, CT, USA).

Cloning and sequencing of the gene encoding AIP56 Total bacterial DNA was prepared from strains MT1415 and ATCC29690 as described (Sambrook and Russel, 2001).

1035

Table 2. Primers used in this study. Primer

Primer sequence (5′-3′)

A1 upper

AA(CT)AA(CT)GA(CT)AA(AG)CC(ATCG)GA (CT)GC GA(AG)TA(CT)AC(ATCG)GT(ATCG)AT(ATC) GA(CT)GC TCATTGAACTTTCCTCTGCAGTATC TGAGTATGGTTTAGTGGATGTTCC GTTCACCAGATTAGCTACGATTGC AAGCCATGAGTGACAGTGTG GGTCGAAGCGATACAAGAGC AGATTCATCATGGCCGTACC GGCCATGATGAATCTGAAGG TTATGAAGCGCGAGAGGACC CCATATAGACCGGAATTGAG GCGCCATGGTGAAAAAATACTCAATAAT GCGCTCGAGATTAATGAATTGTGGCGCGT GTAATACGACTCACTATAGGGC

B lower AI56 upper AI56 lower AI56 inc AI56P1 AI56P2 AI56P3 AI56P4 AI56P5 AI56P6 AIP56NcoIFw AIP56XhoIRv T7

Degenerate primers A1upper and Blower (see Table 2) were used in PCR amplifications with total DNA from the MT1415 and ATCC29690 strains. The 200 bp fragment amplified from strain MT1415 total DNA was excised from an agarose gel, purified using the QIAquick Gel Extraction Kit (Qiagen), cloned into the pGEM®-T Easy vector (Promega) following the manufacturer’s instructions and sequenced to confirm that it corresponded to the desired fragment. Total DNA from strain MT1415 was digested with different restriction enzymes (Roche Applied Science, or Gibco BRL), separated in a 1% agarose gel and subjected to Southern Blotting (Ausubel et al., 1999). The PCR-derived 200 bp fragment described above was labelled with AlkPhos Direct (Amersham Biosciences), and used as a probe on Southern Blotting analysis, following the manufacturer’s instructions. A 3100 bp HindIII-HindIII and a 4100 bp NcoI-BamHI reactive fragments were selected for cloning. DNA from agarose slices containing the relevant fragments was extracted using the QIAquick Gel Extraction Kit. The 3100 bp HindIII-HindIII fragment was inserted into pBluescript II KS (Stratagene); the 4100 bp NcoI-BamHI DNA fragment was cloned into the pET-32b vector (Novagen). Ligation mixtures were used to transform XL1 E. coli competent cells. Transformants were selected by PCR using the primers A1upper and Blower, and sequenced. Another DNA probe was generated by PCR using the recombinant plasmid containing the 4000 bp NcoI-BamHI fragment and primers AI56P4 and T7 (see Table 2). This DNA fragment was used as a probe on Southern Blotting analysis of Phdp MT1415 total DNA, following the procedures described above. The region of an agarose gel containing a 1000 bp HindIII-HindIII reactive fragment was excised, the DNA was extracted as above and cloned into the pBluescript II KS vector (Stratagene). Ligation mixtures were used to transform XL1 E. coli competent cells. Transformants with the desired construct were identified by PCR using the primers AI56P4 and T7 and sequenced. The complete DNA sequence reported here was confirmed on both strands, using the Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems) and the sequencing primers AI56upper, AI56lower, AI56Inc and AI56P1-P6 (see Table 2).

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

1036 A. do Vale et al. Heterologous expression of AIP56 in E. coli To express AIP56 with a C-terminal His-tag in E. coli, total DNA from strain MT1415 was subjected to a PCR amplification using the primers AIP56NcoIFw and AIP56XhoIRv (see Table 2). The PCR-derived DNA fragment was first cloned into the pGEM®-T Easy vector (Promega). After digesting the recombinant plasmid with NcoI and XhoI, the 1.5 kb DNA fragment was purified and cloned into the expression vector pET-28a(+) (Novagen) yielding plasmid pETAIP56H+. Plasmid pETAIP56H+ was used to transform BL21(DE3) E. coli competent cells. BL21 cells transformed with the pET28a(+) empty vector were used as controls (see below). Transformants were grown overnight in Luria Broth (LB) supplemented with 50 µg ml−1 kanamycin at 37°C with shaking. These cultures were diluted 1:50 in fresh LB with 50 µg ml−1 kanamycin and grown at 37°C or 17°C with shaking, until the OD at 600 nm reached 0.6. Isopropyl-β-D-thiogalactopyranoside (IPTG) was then added to a final concentration of 1 mM and growth continued for 4 h and 17 h at 37°C and 17°C respectively. Induced cultures of BL21 cells carrying the pETAIP56H+ or pET28a(+) were pelleted by centrifugation, resuspended in 50 mM phosphate buffer pH 8.0 containing 150 mM NaCl and 1% (v/v) Triton X-100 and sonicated 10 times for 8 s at 42 W on ice (Sonifier B-12, Branson). Soluble and insoluble fractions were obtained by centrifugation of the total lysates at 16 000 g for 10 min at 4°C. Soluble His-tagged AIP56 (AIP56H+) was purified from the soluble fraction of BL21 carrying the pETAIP56H+ induced at 17°C by affinity chromatography using the HIS-select Nickel Affinity Gel (Sigma). Soluble extracts were batch adsorbed to the resin which was then washed twice with 50 mM phosphate buffer (pH 8,0) containing 150 mM NaCl and 5 mM imidazole followed by two washes with the same buffer containing 50 mM imidazole. The AIP56H+ was eluted with 250 mM imidazole and was then dialysed by ultrafiltration against 20 mM tris-HCl pH 8.0 to remove the imidazole using Vivaspin concentrators (Sartorius). Insoluble AIP56H+ was obtained from the insoluble fraction of BL21 cells carrying the pETAIP56H+ induced at 37°C. The apoptogenic activity of the soluble and insoluble AIP56H+ was determined after i.p. injection of groups of six sea bass weighing 37.1 ± 6.0 g with 2 µg of the recombinant proteins in 100 µl of PBS with the osmotic strength adjusted to 355 mOsm. Peritoneal exudates were collected and apoptosis assessed as described above.

Apoptogenic activity of AIP56 towards sea bass phagocytes ex vivo AIP56H+ produced as described above was tested ex vivo for apoptogenic activity against sea bass phagocytes. Peritoneal leukocytes were collected from resting or inflamed peritoneal cavities of fish weighing 37.1 ± 6.0 g (six fish per group). Inflammatory peritoneal exudates were induced with a single i.p. injection of 100 µl of Incomplete Freund’s Adjuvant 10 days before. Peritoneal leukocytes were collected as described above, using L-15 medium (Gibco) containing 10% fetal bovine serum (FBS, Gibco), 1% penicillin/streptomycin (P/S, Gibco) and 10 U ml−1 heparin, instead of PBS. The cell

suspensions were adjusted to 5 × 106 cells ml−1 and were then incubated with 6.7 µg ml−1 of AIP56H+. Aliquots of the same cell suspensions, incubated with the same amount of heat-inactivated (boiling for 30 min) AIP56H+, were used as controls. In some experiments, cells were washed once in L15, 10% FBS, 1%P/S and 10 U ml−1 heparin prior to incubation with the proteins. Apoptosis was assessed 4 h after treatment, as described above.

Plasmid isolation and analysis Plasmid DNA from Phdp strains was isolated using the QIAprep Spin Miniprep Kit (Qiagen) following the manufacturer’s instructions. Sequencing of the MT1415 pPHDP10 plasmid was performed at the MWG Biotech sequencing service (Ebesberg, Germany). Restriction digestions were performed using enzymes from Roche Molecular Biochemicals or Gibco BRL according to the supplier’s instructions and digests analysed by agarose gel electrophoresis. Southern blotting analysis was performed (Ausubel et al., 1999) using approximately 1.5 g of undigested plasmid DNA from each Phdp strain and molecular weight markers separated in a 1% agarose gel. A DNA fragment corresponding to nucleotides 435–1269 of the AIP56 open reading frame (amplified by PCR using primers AI56P5 and AI56P6; see Table 2) was labelled with Alkphos Direct (Amersham Biosciences) and used as a probe.

Passive immunization of sea bass with anti-AIP56 rabbit serum The anti-AIP56 rabbit serum described above was used to passively immunize sea bass with an average weight of 110 g. Pre-immune serum from the same rabbit was used as control. Two independent experiments were carried out. In the first experiment, two groups of eight fish were i.p. injected with either 100 µl or 300 µl of first bleed rabbit serum per fish and a third group of eight fish with 300 µl of control serum per fish. In the second experiment, one group of eight fish was i.p. injected with 300 µl of first bleed rabbit serum per fish, another group of eight fish with 300 µl of second bleed rabbit serum per fish and the third group with 300 µl of control serum per fish. Bacterial challenges were performed by injecting i.p., immediately following passive immunization, 2.1 ± 0.2 × 107 cfu of Phdp virulent strain PTAVSA95. The bacteria were grown overnight on TSA-1 and then resuspended in TSB-1. The bacterial density was measured by spectophotometry at 600 nm and dilutions were made until reaching the expected number of cfu ml−1 as predicted by a curve relating absorbance and cfu. Actual cfu used as a challenge dose were determined by viable counts of dilutions in TSB-1 inoculated on TSA-1 plates, after 48 h at 24°C. A volume of 100 µl of the bacterial suspension was inoculated i.p. into each fish. For the confirmation of the cause of death, Phdp was re-isolated from the head-kidney of moribund and/or dead fish by culturing onto TSA-1. Protection against Phdp infection was evaluated by determining the relative per cent survival {RPS = [1 − (% mortality immunized)/(% mortality controls)] × 100}. Statistical significance of the results was evaluated by applying the χ2-test.

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

An anti-phagocyte exotoxin of Photobacterium damselae subsp. piscicida Nucleotide sequence accession numbers The AIP56 and pPHDP10 nucleotide sequences have been submitted to the GenBank database under Accession numbers DQ066884 and DQ069059 respectively.

Acknowledgements The authors are grateful to Alicia E. Toranzo and Andrew C. Barnes for providing the bacterial strains used in this study and to Paula Sampaio for technical assistance with fluorescence microscopy. Ana do Vale, Diana S. Nascimento and Carolina Costa-Ramos were supported by Grants SFRH/ BPD/11538/2002, SFRH/BD/13054/2003 and SFRH/BPD/ 20881/2004, respectively, from the Portuguese programme POCTI.

References Afonso, A., Ellis, A.E., and Silva, M.T. (1997) The leukocyte population of the unstimulated peritoneal cavity of rainbow trout (Oncorhynchus mykiss). Fish Shellfish Immunol 7: 335–348. Afonso, A., Lousada, S., Silva, J., Ellis, A.E., and Silva, M.T. (1998) Neutrophil and macrophage responses to inflammation in the peritoneal cavity of rainbow trout Oncorhynchus mykiss. A light and electron microscopic cytochemical study. Dis Aquat Org 34: 27–37. Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., and Smith, J.A. (eds). (1999) Current Protocols in Molecular Biology. New York: John Wiley & Sons. Bakopoulos, V., Adams, A., and Richards, R.H. (1997) The effect of iron limitation growth conditions on the cell and extracellular components of the fish pathogen Pasteurella piscicida. J Fish Dis 20: 297–305. Barnes, A.C., dos Santos, N.M.S., and Ellis, A.E. (2005) Update on bacterial vaccines: Photobacterium damselae subsp. piscicida. In Progress in Fish Immunology, Vol. 121. Mydtlyng, P.J. (ed.). Basel: Karger, pp. 75–84. Burns, D.L. (2003) Type IV transporters of pathogenic bacteria. Curr Opin Microbiol 6: 29–34. DeLeo, F.R. (2004) Modulation of phagocyte apoptosis by bacterial pathogens. Apoptosis 9: 399–413. Deng, W., Puente, J.L., Gruenheid, S., Li, Y., Vallance, B.A., Vazquez, A., et al. (2004) Dissecting virulence: systematic and functional analyses of a pathogenicity island. Proc Natl Acad Sci USA 101: 3597–3602. Gauthier, G., Lafay, B., Ruimy, R., Breittmayer, V., Nicolas, J.L., Gauthier, M., et al. (1995) Small-subunit rRNA sequences and whole DNA relatedness concur for the reassignment of Pasteurella piscicida (Snieszko et al.) Janssen and Surgalla to the genus Photobacterium as Photobacterium damsela subsp. piscicida comb. nov. Int J Syst Bacteriol 45: 139–144. Hacker, J., Blum-Oehler, G., Muhldorfer, I., and Tschape, H. (1997) Pathogenicity islands of virulent bacteria: structure, function and impact on microbial evolution. Mol Microbiol 23: 1089–1097. Harlow, E., and Lane, D. (1998) Antibodies- A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.

1037

Hase, C.C., and Finkelstein, R.A. (1993) Bacterial extracellular zinc-containing metalloproteases. Microbiol Rev 57: 823–837. Henson, P.M., and Johnston, R.B. (1987) Tissue injury in inflammation. Oxidants, proteases and cationic proteins. J Clin Invest 79: 669–674. Hueck, C.J. (1998) Type III protein secretion systems in bacterial pathogens of animals and plants. Microbiol Mol Biol Rev 62: 379–433. Janssen, W.A., and Surgalla, M.J. (1968) Morphology, physiology, and serology of a Pasteurella species pathogenic for white perch (Roccus americanus). J Bacteriol 96: 1606–1610. Jongeneel, C.V., Bouvier, J., and Bairoch, A. (1989) A unique signature identifies a family of zinc-dependent metallopeptidases. FEBS Lett 242: 211–214. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680–685. Ma, Q., Zhai, Y., Schneider, J.C., Ramseier, T.M., and Saier, M.H. (2003) Protein secretion systems of Pseudomonas aeruginosa and P. fluorescens. Biochim Biophys Acta 1611: 223–233. Maderna, P., and Godson, C. (2003) Phagocytosis of apoptotic cells and the resolution of inflammation. Biochim Biophys Acta 1639: 141–151. Mazzolini, E., Fabris, A., Ceschia, G., Vismara, D., Magni, A., Amadei, A., et al. (1998) Pathogenic variability of Pasteurella piscicida during in vitro cultivation as a preliminary study for vaccine production. J Appl Ichthyol 14: 265–268. Nahm, M.H., Apicella, M.A., and Briles, D.E. (1999) Immunity to extracellular bacteria. In Fundamental Immunology. Paul, W.E. (ed.). Philadelphia: Lippincott-Raven Publishers, pp. 1373–1386. Narayanan, S.K., Nagaraja, T.G., Chengappa, M.M., and Stewart, G.C. (2002) Leukotoxins of gram-negative bacteria. Vet Microbiol 84: 337–356. Pallardy, M., Biola, A., Lebrec, H., and Breard, J. (1999) Assessment of apoptosis in xenobiotic-induced immunotoxicity. Methods 19: 36–47. Popov, S., Villasmil, R., Bernardi, J., Grene, E., Cardwell, J., Popova, T., et al. (2002a) Effect of Bacillus anthracis lethal toxin on human peripheral blood mononuclear cells. FEBS Lett 527: 211. Popov, S.G., Villasmil, R., Bernardi, J., Grene, E., Cardwell, J., Wu, A., et al. (2002b) Lethal toxin of Bacillus anthracis causes apoptosis of macrophages. Biochem Biophys Res Commun 293: 349–355. Romalde, J.L. (2002) Photobacterium damselae subsp. piscicida: an integrated view of a bacterial fish pathogen. Int Microbiol 5: 3–9. Sambrook, J., and Russell, D.W. (2001) Molecular Cloning. A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Sandkvist, M. (2001) Type II secretion and pathogenesis. Infect Immun 69: 3523–3535. Schagger, H., Aquila, H., and Von Jagow, G. (1988) Coomassie blue-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for direct visualization of polypeptides during electrophoresis. Anal Biochem 173: 201–205. Snieszko, S.F., Bullock, G.L., Hollis, E., and Boone, J.G.

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038

1038 A. do Vale et al. (1964) Pasteurella sp. from an epizootic of white perch (Roccus Americanus) in Chesapeake Bay tidewater areas. J Bacteriol 88: 1814–1815. Thune, R.L., Stanley, L.A., and Cooper, R.K. (1993) Pathogenesis of Gram-negative bacterial infections in warmwater fish. Annu Rev Fish Dis 3: 37–68. Trüper, H.G., and de’Clari, L. (1997) Taxonomic note: necessary correction of specific epithets formed as substantives (nouns) ‘in apposition’. Int J Syst Bacteriol 47: 908–909. do Vale, A., Afonso, A., and Silva, M.T. (2002) The professional phagocytes of sea bass (Dicentrarchus labrax L.): cytochemical characterisation of neutrophils and macrophages in the normal and inflamed peritoneal cavity. Fish Shellfish Immunol 13: 183–198. do Vale, A., Marques, F., and Silva, M.T. (2003) Apoptosis

of sea bass (Dicentrarchus labrax L.) neutrophils and macrophages induced by experimental infection with Photobacterium damselae subsp piscicida. Fish Shellfish Immunol 15: 129–144. Vermes, I., Haanen, C., and Reutelingsperger, C. (2000) Flow cytometry of apoptotic cell death. J Immunol Methods 243: 167–190. Weinrauch, Y., and Zychlinsky, A. (1999) The induction of apoptosis by bacterial pathogens. Annu Rev Microbiol 53: 155–187. Willingham, M.C. (1999) Cytochemical methods for the detection of apoptosis. J Histochem Cytochem 47: 1101– 1110. Wyllie, A.H., Kerr, J.F., and Currie, A.R. (1980) Cell death: the significance of apoptosis. Int Rev Cytol 68: 251–306.

© 2005 The Authors Journal compilation © 2005 Blackwell Publishing Ltd, Molecular Microbiology, 58, 1025–1038