Detection and Localisation of the Abalone Probiotic Vibrio midae SY9 and Its Extracellular Protease, VmproA, within the Digestive Tract of the South African Abalone, Haliotis midae

Probiotics have been widely reported to increase the growth rate of commercially important fish and shellfish by enhancing the digestion of ingested feed through the production of extracellular enzymes such as proteases and alginases. In order to investigate this further, the objective of this study was to localise the bacterial probiont Vibrio midae SY9 and one of the extracellular proteases it produces in the digestive tract of the South African abalone Haliotis midae. This was accomplished by inserting a promotorless gfp gene into the chromosome of the bacterium which was incorporated in an artificial, fishmeal-based abalone feed. In situ histological comparison of abalone fed either a basal diet or the basal diet supplemented with V. midae SY9::Tn10.52 using a cocktail of DNA probes to the gfp gene localised the probiont to the crop/stomach and intestinal regions of the H. midae digestive tract. Generally, the ingested probiotic bacterium occurred in association with feed and particulate matter within the crop/stomach and intestinal regions, as well as adhered to the wall of the crop/stomach. Histological immunohistochemical examination using polyclonal anti-VmproA antibodies localised an extracellular protease produced by V. midae SY9 to the H. midae crop/stomach and intestine where it appeared to be associated with feed and/or other particulate matter in the abalone gut. Thus the data suggests that V. midae SY9 colonises and/or adheres to the mucous lining of the abalone gut. Furthermore, the close association observed between the bacterium, its extracellular protease and ingested feed particles supports the theory that V. midae SY9 elevates in situ digestive enzyme levels and thus enhances feed digestion in farmed abalone.


Introduction
South Africa has a rapidly developing abalone aquaculture industry, based on the cultivation of Haliotis midae [1]. However, the relatively slow growth rate of abalone represents a major constraint on the aquaculture industry. The use of probiotic microorganisms is becoming increasingly accepted as a means of improving the health and growth of aquacultured species [2]. Macey and Coyne [3,4] demonstrated that H. midae fed a high protein diet supplemented with the probiotic Vibrio midae SY9 had increased digestive tract protease levels, enhanced protein digestion and increased growth rates in comparison to animals fed an un-supplemented diet.
Several possible modes of action have been proposed for probiotic effects observed within aquaculture environments [5,6], including the production and secretion of extracellular hydrolytic enzymes that contribute to, and improve, the digestion efficiency of the host. Several studies have demonstrated the effect of probiotic supplementation on abalone digestive enzyme activity levels and/or growth, and have suggested a possible role for 'nutritional probiotics' in abalone aquaculture [7,8].
Abalone possess a unique microbiota that is capable of producing extracellular enzymes which degrade the major constituents of abalone feeds [9]. However, less than 10% of the microorganisms associated with the abalone digestive tract can be cultured in the laboratory [10]. Consequently, culture-independent methodologies are necessary for investigating gut microorganisms within their natural habitat [11]. In situ hybridization (ISH) using specific 16S rDNA oligonucleotide probes is a cultureindependent method used for investigating bacterial population diversity [11], and is an ideal method for investigating microorganisms in vivo [12]. ISH techniques have been successfully used to investigate the microbiota of goldfish [13], abalone [4], Artemia nauplii [14] and salmon [15], and to specifically localise intracellular prokaryotes in abalone tissue sections [16]. Rengpipat et al. [17] successfully tagged the shrimp probiotiont Bacillus S11 with GFP and then monitored the presence of this probiotic within the digestive tract of the Black Tiger shrimp Penaeus monodon following dietary supplementation. Histological analysis of intestinal samples revealed that the GFP-tagged probiotic bacterium was viable and localised to the surface of the shrimp's intestine. Macey and Coyne [3] observed significantly increased growth rates in abalone fed a probiotic supplemented feed, as well as increased protease activity, protein digestion and protein absorption within the intestinal region of these abalone. This finding supports the view that feeding aquacultured species with probiotic microorganism(s) capable of producing and secreting hydrolytic extracellular enzymes may improve digestion efficiency of the host animal, resulting in enhanced host growth rates [18]. Detection of the V. midae SY9 extracellular protease VmproA within the digestive tract of H. midae fed ABFEEDH S34 supplemented with the probiont may indicate that a similar process is responsible for the increased growth rate reported in abalone fed ABFEEDH containing the bacterium [3]. Thus the aim of this study was to utilize immunohistochemistry, ISH and standard histological staining techniques to investigate the spatial distribution of V. midae SY9 and VmproA within the digestive tract of H. midae.

Ethics Statement
No ethics permits were required for the described study, which complied with all relevant regulations. Permission to work in the Department of Agriculture, Forestry and Fisheries Research Aquarium was not required since it is a research facility that is available to my research group on the basis that my research is government funded. The facility is not attached to a national park. This study did not use any endangered or protected species.

Screening Transconjugants
V. midae SY9 strains putatively harbouring the chromosomally integrated mini-Tn10-gfp-kan transposable element were inoculated into 5 ml MB containing 120 mg/ml Sm and 400 mg/ml Kan, and grown for approximately 16 hours at 22uC on a rotary shaker. Chromosomal DNA was isolated according to the method described by Ausubel et al. [20]. PCR amplification of the gfp gene using the oligonucleotide primers gfp-F (59-GATTTCTA-GATTTAAGAAGC-39) and gfp-R (59-TCATATTTGTA-TAGTTCATCC-39) was used to screen for the presence of the integrated gfp gene in order to identify gfp-tagged V. midae SY9::Tn10.52 strains for further analysis.

Southern Hybridization
The plasmid pLOFKmgfp was purified from E. coli SM10lpir using the Qiagen Midi Prep kit according the manufacturer's instructions. A 717 bp DNA fragment of the gfp gene was PCR amplified from pLOFKmgfp using the synthetic oligonucleotide primers gfp-F and gfp-R. The 717 bp PCR amplicon was resolved on a 1% (wt/vol) TAE agarose gel and purified using the BioSpin Gel Extraction Kit (BioFlux) according to the manufacturer's instructions. The purified 717 bp DNA fragment was radioactively labelled using a random-primed DNA labelling kit (Roche) according to the manufacturer's instructions and hybridised to equal amounts of V. midae SY9 and V. midae SY9::Tn10.52 chromosomal DNA. The Southern hybridization procedure was performed according to Church and Gilbert [21].
Preparation of V. midae SY9::Tn10.52-supplemented Feed The basal diet consisted of ABFEEDH S34 weaning chips as supplied by Marifeed (Pty.) Ltd., Hermanus, South Africa. The gfp chromosomally-tagged probiotic strain V. midae SY9::Tn10.52 was incorporated into ABFEEDH S34 weaning chips by vacuum infusion, to a final concentration of at least 1.0610 8 culturable cells g -1 feed. Briefly, V. midae SY9::Tn10.52 was cultivated for 24 hrs in 3 litres of P-MBM supplemented with 120 mg/ml Sm and 400 mg/ml Kan on an orbital shaker (100 rpm) at 22uC. Thereafter, the bacterial cells were harvested by centrifugation (8,0006g for 15 minutes at 4uC), washed with one volume of artificial seawater (ASW) [(wt/vol): 3% NaCl, 0.23% MgCl 2 .6H 2 O, 0.03% KCl] and resuspended in 100 ml of ASW. Approximately 200 g of the ABFEEDH S34 weaning chips were sealed inside a glass vacuum jar and a vacuum drawn within the jar to approximately 80 kPa. Approximately 20 ml of the V. midae SY9::Tn10.52 suspension was drawn into the vacuum jar and thoroughly mixed with the feed. During the process of drawing the bacterial suspension into the chamber, the vacuum was decreased to 50 kPa and maintained at this pressure for 5 minutes. Thereafter, the vacuum was slowly released, the impregnated ABFEEDH S34 weaning chips removed from the jar and dried at 22uC overnight, before being sealed into clean plastic bags and stored at 4uC. Fresh batches of feed were prepared every 7 days and each batch was analyzed by determining the total number of culturable bacterial cells to ensure that there was at least 1610 6 culturable V. midae SY9::Tn10.52 cells g/feed.  During acclimatization the abalone were fed ABFEEDH S34 weaning chips to satiation. All uneaten food was removed from the baskets and the tanks and baskets were thoroughly cleaned every 2 days before the addition of fresh feed.
Abalone in both tanks were starved for a period of 24 hours prior to the beginning of the experiment. At the start of the experiment (Day 0), six randomly selected animals were removed from each tank and immediately sacrificed. The remaining abalone in one tank were fed the V. midae SY9::Tn10.52supplemented diet, while those in the other tank were fed the basal diet. The animals were fed to satiation on the respective diets for the duration of the experiment. Six randomly selected animals were sacrificed from each treatment group on days 2 and 14.

Histology
The abalone shell was gently removed by severing the adductor muscle as close to the shell as possible using a thin metal spatula, without rupturing the digestive tract. The animals were placed adductor muscle side down in labeled embedding cassettes and fixed in Davidson's solution [24] (per litre (vol/vol): 330 ml 95% ethanol (Merck), 220 ml 100% formalin (Merck), 115 ml glacial acetic acid (Merck) and 335 ml distilled water) for 36 hours at 4uC, before being transferred to 70% (vol/vol) ethanol at 4uC. Following fixation, the abalone samples were dehydrated through a graded ethanol series to 100% xylene (Saarchem) in a Tissue Trek II tissue processor. The dehydrated tissue samples were embedded in paraffin wax, sectioned at 5 mm, adhered onto positively charged microscope slides (SuperFrostH Plus, Menzel-Glä ser) and stored in slide boxes at room temperature.
The H. midae sections were deparaffinised and stained using standard Harris' H&E stain according to Hayat [25] and mounted with phosphate-buffered glycerin jelly [50% (vol/vol) Glycerol (Merck), 0.5 M phosphate buffer (pH 7.0) and 7.5% (wt/vol) gelatine (Merck)]. The sections were viewed using a Nikon Eclipse 50 i Compound Microscope equipped with a Nikon DS Camera Control Unit DS-U2 and DS-5M Camera head with Nikon Software (NIS Elements Documentation and Digital 3D Imaging). In situ Hybridization H. midae sections were processed for ISH after being deparaffinized and rehydrated through an ethanol series [99.9, 95, 80, 70, and 50% (v/v) ethanol]. Slides were covered with prewarmed ISH buffer [50% (vol/vol) formamide (Sigma), 4x SSC, 1x Denhardt's solution (Sigma), 0.2 mg yeast tRNA (Sigma), 0.5 mg denatured Herring sperm DNA (Sigma)] and prehybridized at 42uC for 60 minutes in a humid chamber to reduce non-specific hybridization of the probe(s). Thereafter, the tissue sections were heat denatured on a heating block (98uC for 10 minutes) and subsequently cooled rapidly on ice. The sections were probed with the cocktail of gfp-specific probes (GFP001, GFP002 and GFP003). Each probe was added at a concentration of approximately 6.66 pmol/ml of ISH buffer, resulting in a total probe concentration of 20 pmol/ml. The positive (EUB338) and the negative (ECJ109) control oligonucleotide probes were added at a concentration of 20 pmol/ml of ISH buffer.
Pre-heated ISH buffer containing 20 pmol/ml of the DIGlabeled probes was evenly layered onto the tissue sections and incubated for approximately 16 hours at 40uC in a humid chamber. Any unbound DIG-labeled probes were removed from the tissue sections by washing twice for 5 minutes in 2x SSC at 22uC with gentle agitation, twice for 5 minutes in 1x SSC at 22uC with gentle agitation, and twice for 10 minutes in 0.5x SSC at 40uC with gentle agitation. The tissue sections were then equilibrated in buffer 1 [150 mM NaCl, 100 mM Tris-HCl (pH 7.5)] for 2 minutes at 22uC. The sections were washed in blocking buffer [100 mM Tris-HCl (pH 7.5), 150 mM NaCl, 2% (vol/vol) feotal calf serum (Invitrogen) and 0.3% (vol/vol) Triton X-100 (Sigma)] for 60 minutes at 22uC with gentle agitation to reduce non-specific antibody binding. Anti-digoxigenin-AP (Roche) signalling molecule was applied to the sections which were incubated for 3 hrs at 22uC in a humid chamber.
The tissue sections were rinsed briefly with buffers 1 and 2 [100 mM Tris-HCl (pH 9.5), 100 mM NaCl, 50 mM MgCl 2 ], respectively, before being overlaid with a NBT/BCIP (Roche) colour development solution, prepared in buffer 2 according to the manufacturer's instructions, and incubated at 22uC in a dark humid chamber until a signal was visible. The colour development reaction was stopped by washing the slides with TE buffer [10 mM Tris-HCl (pH 7.6), 2 mM EDTA] and rinsing with distilled water. The sections were counter-stained with 0.05% (w/ v) aqueous Methyl Green (Sigma) for 2 minutes to enhance the background tissue morphology, rinsed with dH 2 O, dried and mounted as described above. Hybridization signals indicating the presence of the target microorganism(s) were visible as a purple precipitate where the DIG-labelled probe had bound to homologous target DNA within the tissue sections. The stained and mounted sections were viewed and photographed as described above.
Unbound secondary antibodies were removed by sequential washes in PBT at 22uC with gentle agitation.
The tissue sections were rinsed briefly with buffer 1 and buffer 2, respectively, overlaid with NBT/BCIP (Roche) colour development solution and incubated at 22uC in a dark humid chamber until a signal was clearly visible. The colour development reaction was stopped by washing the slides with TE buffer and rinsing with distilled water, and the slides treated and mounted as described above. AP-conjugated anti-rabbit secondary antibodies bound to anti-[VmproA] polyclonal antibodies that had in turn bound to VmproA within the tissue sections were visible as areas of purple precipitate (NBT/BCIP). The stained and mounted sections were viewed and photographed as described above.

Results
Transposon Mutagenesis of V. midae SY9 Fifty eight V. midae SY9Sm r cells growing on VNSS agar supplemented with Sm and Kan were screened by PCR amplification in order to confirm chromosomal integration of the mini-Tn10-gfp-kan cassette (data not shown). Three strains were identified as being gfp chromosomally-tagged V. midae SY9

Southern Hybridization
Mini-Tn10-gfp-kan integration into the V. midae SY9 genome was confirmed by a Southern hybridisation experiment using a 0.7 kb fragment of the gfp gene as a probe against chromosomal DNA isolated from V. midae SY9 and V. midae SY9::Tn10.52 (Fig. 2). The presence of a single hybridisation band of 1.8 and 13.3 kb following hybridisation of the gfp fragment to HindIIIand EcoRI-digested V. midae SY9::Tn10.52 chromosomal DNA indicated that the transposable element integrated at a single site in the V. midae SY9 genome.

Histology
Morphological differences between H. midae fed either the ABFEEDH S34 basal diet or the V. midae SY9::Tn10.52 supplemented feed were not observed by histological examination of H&E stained whole-animal sections over the course of the 14 day experimental period (data not shown). The oesophageal region of the alimentary canal was devoid of any food in both groups of animals, while the crop/stomach and intestinal regions contained mucous and food particles.
Localization of V. midae SY9 within the Digestive Tract ISH was used to detect the presence of V. midae SY9 in the digestive tract of H. midae fed ABFEEDH S34 supplemented with V. midae SY9::Tn10.52 over a fourteen day period. Strong hybridization signals were observed within the crop/stomach and intestinal regions of H. midae fed either the basal or the V. midae SY9::Tn10.52 -supplemented feeds when whole-animal H. midae tissue sections were hybridized with the universal eubacterial probe EUB338 (Figures 3 and 4). The signals appear to be associated with food and particulate matter in the crop/stomach and intestine. Hybridization signals were not detected within the oesophageal or digestive gland regions of the H. midae digestive tract (data not shown). Hybridization signals were not detected in whole-animal sections prepared from abalone fed the basal ABFEEDH S34 diet for 0, 2 or 14 days probed with the gfpspecific oligonucleotide cocktail (data not shown). Similarly, the gfp-specific probes did not produce detectable hybridization signals in sections prepared from abalone fed the V. midae SY9::Tn10.52 supplemented diet sampled at day 0 (Fig. 5). However, hybridization signals were detected in the crop/stomach and intestine of whole-animal sections of H. midae fed V. midae SY9::Tn10.52 supplemented ABFEEDH S34 for 2 and 14 days (Fig. 5). The hybridization signals appear to be associated with feed and/or particulate matter in these regions of the abalone digestive tract, as well as along the lining of the crop/stomach. The absence of hybridization signals in whole-animal H. midae sections hybridized with the E. coli JM109-specific DIG-labelled oligonucleotide probe ECJ109 confirmed the specificity of the signals obtained with the universal eubacterial probe and the cocktail of gfp-specific oligonucleotide probes (data not shown).

Immunohistochemical Localization of VmproA within the H. midae Digestive Tract
Immunohistochemical signals corresponding to the presence of anti-VmproA polyclonal antibodies were detected in the crop/ stomach and intestinal regions of whole-animal sections prepared from H. midae fed ABFEEDH S34 supplemented with V. midae SY9::Tn10.52 (Fig. 6). VmproA appears to be associated with food and/or other particulate matter within both the crop/stomach and intestine. Immunostaining signals were not evident in wholeanimal sections prepared from H. midae fed unsupplemented ABFEEDH S34 (Fig. 6). Macey and Coyne (2005) investigated in situ protease activity and protein digestion in H. midae fed a high-protein artificial diet supplemented with a mixture of three probiotic strains that included V. midae SY9. They found that H. midae fed ABFEEDH S34 supplemented with the probiotic strains had enhanced intestinal alkaline protease activity in comparison to animals fed a basal ABFEEDH S34 diet. They also observed a significant improvement in intestinal protein digestion and absorption in the abalone fed the probiotic-supplemented diet [3].

Discussion
Rengpipat et al. [17] transformed Bacillus S11 with the GFPexpressing plasmid pAD44-12 which they used as a marker for fluorescent in situ localisation of Bacillus S11 in the intestine of the black tiger shrimp Penaeus monodon. Similarly, Macey and Coyne [4] constructed chromosomally tagged strains of V. midae SY9 using a mini-Tn10-gfp-kan transposon mutagenesis system in order to track the persistence of ingested V. midae SY9 within the digestive tract of H. midae. However, Macey [3] observed that relative GFP expression and fluorescence in the V. midae SY9 mini-Tn10-gfp-kan transconjugants was inadequate for in situ detection of V. midae SY9 in the abalone digestive tract. Nevertheless, Macey and Coyne [4] demonstrated that a fragment of the mini-Tn10gfp-kan transposon cassette could be used as a specific DNA probe to detect chromosomally tagged V. midae SY9 in the digestive tract of H. midae. Chromosomal integration of the mini-Tn10-gfp-kan transposon had no significant impact on the growth and protease activity of the gfp chromosomally-tagged V. midae SY9 strains (data not shown). Similarly, Rengpipat et al. [17] demonstrated that the properties of a GFP-expressing mutant strain of a shrimp probiont Bacillus S11-GFP were not significantly different to that of the wildtype strain. Therefore, in situ hybridization (ISH) using specific DNA probes was used in this study to localise ingested V. midae SY9 in the H. midae digestive tract.
Bacteria are known to occur throughout the digestive tract of aquatic invertebrates [27]. Previous studies have identified a variety of bacterial isolates from the abalone gut [3,18,28]. Therefore, it is not surprising that, regardless of diet, strong hybridization signals were detected in both the crop/stomach and intestinal regions when whole-animal sections were probed with the 16S rRNA gene eubacterial probe EUB338. [18] demonstrated that comparatively, the intestine contained the largest number and greatest diversity of culturable enteric bacteria in the H. midae digestive system. Similarly, hybridization signals associated with the intestinal region of the digestive tract appeared to be more intense than those observed in the crop/stomach, indicating a greater load of enteric bacteria within the posterior regions of the abalone digestive tract.
In situ hybridization of whole-animal tissue sections using a cocktail of gfp-specific oligonucleotide probes detected V. midae SY9::Tn10.52 in the crop/stomach and intestinal regions of the H. midae digestive tract. In contrast, hybridization signals were not detected in sections prepared from abalone fed the basal ABFEEDH S34 weaning chips. Similar results have been reported in studies investigating the colonisation potential of probiotics for shrimp [17] and abalone [4]. Macey and Coyne [4] used cell culture and in situ hybridization to detect the chromosomallytagged strain V. midae SY9.8 in both the crop/stomach and intestine of H. midae fed probiotic-supplemented feed.
We observed distinct hybridization signals indicating the presence of V. midae SY9::Tn10.52 along the inner surface of the crop/stomach of H. midae fed supplemented ABFEEDH. Similarly, Rengpipat et al. [17] detected GFP-tagged Bacillus S11 (S11-GFP) cells attached to the intestinal mucous lining of the black tiger shrimp P. monodon. Additionally, strong hybridization signals were observed in association with feed or other particulate matter within the crop/stomach and intestine of H. midae fed the V. midae SY9::Tn10.52 supplemented diet.
The extracellular protease VmproA produced by V. midae SY9 was detected immunologically within the crop/stomach and intestinal regions of H. midae fed ABFEEDH S34 supplemented with V. midae SY9::Tn10.52. There was no immunohistochemical evidence of VmproA within the H. midae digestive tract of abalone fed the basal diet over the course of the 14 day experimental period. VmproA appeared to be less abundant in the crop/ stomach than the intestine of abalone fed V. midae SY9::Tn10.52 supplemented feed. Similarly, elevated levels of in situ alkaline protease activity detected in the crop/stomach of abalone fed ABFEEDH supplemented with V. midae SY9 were not as pronounced as the protease activity observed in the intestinal regions of these animals [4]. Since V. midae SY9::Tn10.52 cells detected in the intestine by in situ hybridization with the gfp-specific probes was more abundant in comparison to the crop/stomach, it is likely that the relative abundance of detectable VmproA in the intestine is a function of the elevated V. midae SY9 cell number in this portion of the H. midae digestive tract. Indeed, a significant positive correlation (r = 0.711, P,0.05, n = 17) was found between intestinal protease activity and the number of V. midae SY9.8 cells present, supporting an association between V. midae SY9 and protease activity in the intestine of H. midae [4].
VmproA appeared to be generally associated with the feed and/ or particulate matter within the intestine of H. midae fed the V. midae SY9::Tn10.52 -supplemented ABFEEDH diet. Immunohistochemical localisation of VmproA within the H. midae digestive tract appears to be similar to the in situ localisation of V. midae SY9::Tn10.52 detected by the in situ hybridization analysis of whole-animal tissue sections. Thus, it could be hypothesised that viable V. midae SY9 cells, ingested with the supplemented ABFEEDH pellets, may attach to the surface of the digestive tract but are mostly associated with food particles and/or other particulate matter in the abalone digestive tract. Presumably, proximity to protein-rich ABFEEDH induces V. midae SY9 to secrete proteases such as VmproA within the digestive tract of H. midae.

Conclusion
This study demonstrated that V. midae SY9 and its extracellular protease, VmproA, could be localised within the digestive tract of H. midae fed a V. midae SY9-supplemented diet. The data suggests that V. midae SY9 may colonise and/or adhere to the mucous lining of the abalone gut, as well as associate with the surfaces of ingested food particles passing through the digestive tract. In so doing, V. midae SY9 may elevate in situ enzyme levels and thus enhance feed digestion. To the best of our knowledge this is the first study to localize an extracellular protease produced by a probiotic bacterium in the abalone digestive tract.