ABSTRACT
Aeromonas salmonicida causes furunculosis in salmonids and is a threat to Atlantic salmon aquaculture. The epithelial surfaces that the pathogen colonizes are covered by a mucus layer predominantly comprised of secreted mucins. By using mass spectrometry to identify mucin glycan structures with and without enzymatic removal of glycan residues, coupled to measurements of bacterial growth, we show here that the complex Atlantic salmon intestinal mucin glycans enhance A. salmonicida growth, whereas the more simple skin mucin glycans do not. Of the glycan residues present terminally on the salmon mucins, only N-acetylglucosamine (GlcNAc) enhances growth. Sialic acids, which have an abundance of 75% among terminal glycans from skin and of <50% among intestinal glycans, cannot be removed or used by A. salmonicida for growth-enhancing purposes, and they shield internal GlcNAc from utilization. A Ca2+ concentration above 0.1 mM is needed for A. salmonicida to be able to utilize mucins for growth-promoting purposes, and 10 mM further enhances both A. salmonicida growth in response to mucins and binding of the bacterium to mucins. In conclusion, GlcNAc and sialic acids are important determinants of the A. salmonicida interaction with its host at the mucosal surface. Furthermore, since the mucin glycan repertoire affects pathogen growth, the glycan repertoire may be a factor to take into account during breeding and selection of strains for aquaculture.
INTRODUCTION
The Atlantic salmon is an anadromous salmonid with a life cycle that includes both freshwater (FW) and seawater (SW) stages. This is reflected in current husbandry protocols, where juveniles (parr) are grown in FW until they have undergone a developmental stage, i.e., the parr-smolt transformation or smoltification, preparing the fish for a life in the sea. After successful smoltification, wild fish, called smolts, normally migrate from their native FW streams into the sea. Farmed smolts are instead usually transferred to large net pens in the sea for growth until slaughter (this growth phase is termed “on-growth”). The global yearly production rate of Atlantic salmon, Salmo salar L., exceeded 2.3 million tons in the year 2014, and the yearly production is expected to continue to grow even more (2).
Fish primary barriers, the skin, gills, and gut, are in direct contact with the environment and capable of nurturing both bacteria and viruses (3). Diseases are easily spread among fish under farming conditions, and the health of the fish is dependent on functional, intact primary barriers. Historically, one of the most important fish diseases is furunculosis, which can cause high morbidity and high mortality in salmonid aquaculture (4). The causative agent of this disease is Aeromonas salmonicida subsp. salmonicida (here referred to as A. salmonicida), a Gram-negative, facultative anaerobic bacterium that persists in both FW and SW (5, 6). The intestinal tract and skin of the Atlantic salmon have been demonstrated to be major routes for A. salmonicida infection (7–10). In aquaculture, vaccines are effective in preventing the disease, at the cost of side effects that cause reduced fish health and welfare (11). Despite vaccinations, disease outbreaks occur that require use of antibiotics (with the possible consequent emergence of antibiotic-resistant pathogens [12, 13]), and occasionally the entire fish stock is sacrificed to prevent reinfection (14). Thus, alternative ways to prevent and treat furunculosis would be advantageous for salmonid aquaculture.
One alternative way for improving fish health and welfare in aquaculture could be to strengthen the innate mucosal defense of fish. The first point of interaction between pathogens and fish occurs at the mucus layer covering the epithelium, which constitutes the first part of the innate immune system. The main components of this mucus layer are secreted mucins. Mucins consist of a protein backbone covered by a vast array of O-glycan structures linked via serines and threonines. In mammals, a range of mucins, named MUC1 to -20, have been identified (15). Next-generation sequencing of Atlantic salmon skin has revealed partial mucin sequences with homology to the human MUC2 and MUC5 mucins (16).
Mucins can bind pathogens and can modulate bacterial growth, e.g., via stimulatory or inhibitory glycan motifs (17, 18) and/or by being food sources for bacteria through released glycans (19). The ability of the mucin glycan structures to bind bacteria plays an important role in the mucosal defense against infection (15, 20). In Atlantic salmon, we found 109 different mucin O-glycan structures, and both the distribution of these glycans and the level of A. salmonicida binding to mucins differed between epithelial sites (21, 22). Within each epithelium site, Atlantic salmon had very similar O-glycan profiles, but only 12% of the structures were detected in all four epithelial sites investigated (i.e., skin, pyloric ceca, proximal intestine, and distal intestine). Of the 109 O-glycans identified, most were sialylated; three types of sialic acids (N-acetylneuraminic acid [Neu5Ac], N-glycolylneuraminic acid [Neu5Gc], and deaminoneuraminic acid [Kdn]) were detected on skin mucins, whereas mainly Neu5Ac was found on the gastrointestinal mucins (21). Skin O-glycans were also shorter and less diverse than intestinal O-glycans (21, 22).
Fish maintain their blood pH to a fairly constant level, between 7.7 and 7.9 (23). In the environment, on the other hand, the pH is more variable, depending on geographic location as well as on spatial and temporal fluctuations within each location. In smolt production facilities in Norway and Chile, pHs between 5 and 8 have been reported (24). The pH to which the fish are exposed in the holding tanks depends on the specific water flow (in liters per kilogram per minute); a low specific water flow leads to CO2 accumulation, which in turn reduces pH (25). Also, calcium levels can vary from below 0.1 to 0.5 mM, depending on the source of intake water (24). During on-growth of the fish, levels up to 10 mM Ca2+ are expected in full-strength SW.
Although Ca2+ levels can affect the surface characteristics of A. salmonicida (26), effects of changes in Ca2+ levels on the bacterium-mucin interaction in fish epithelia or on A. salmonicida growth per se have not been studied. In humans, pH has been shown to affect the mucin interaction with Helicobacter pylori (27), whereas the impact of pH on mucin-pathogen interactions in fish is currently unexplored. Therefore, the main objectives of our project were to study the effect of purified Atlantic salmon mucins on A. salmonicida growth as well as to elucidate how environmental factors, i.e., Ca2+ concentration and pH, affect A. salmonicida growth per se and the bacterium's interaction with purified mucins.
RESULTS
Intestinal mucins enhance A. salmonicida growth.The effect of mucins on A. salmonicida growth was analyzed in defined medium (DM) with and without the addition of 2 g/liter glucose (DM+ versus DM-), since A. salmonicida may be able to utilize mucins as a carbon source, similarly to some other microbes (28–31). We verified that alamarBlue (a redox indicator that yields a signal in response to metabolic activity), used to analyze the growth of A. salmonicida, closely followed the growth curves (measured in CFU per milliliter) and thus is an accurate measure of A. salmonicida growth within the ranges of growth and bacterial density at which we performed all experiments presented in this study (r2 = 0.99) (Fig. 1B). The change in alamarBlue signal followed approximately a 1:1 relationship with corresponding changes in CFU per milliliter in the range in which the experiments were performed. In DM+, mucins from pyloric ceca and proximal and distal intestine enhanced the A. salmonicida growth rate up to 20% (P < 0.05), whereas no effect was observed with skin mucins (n = 5) (Fig. 1C). In DM-, the pathogen grew slower and the effect of mucins was more pronounced; mucins from proximal and distal intestine enhanced growth by 20 to 50% (P < 0.01) and mucins from the pyloric ceca tended to enhance growth (3 to 10%; P = not significant), whereas skin mucins had no effect (n = 5) (Fig. 1D). Thus, mucins from different epithelial sites of the fish differentially affected growth. This effect was more pronounced in the absence of glucose, indicating that the pathogen may be able to utilize the mucins as a nutrient source.
Mucin isolation and effects of mucins on A. salmonicida growth rate. (A) Mucins were isolated by isopycnic density gradient centrifugation, and fractions were collected from the bottom of the tube. Fractions were analyzed for glycan content (europium count), DNA content (in nanograms per microliter), and density (grams per milliliter). The fractions for all samples were pooled based on the glycan peak (fractions were pooled as indicated by the vertical lines). (B) Time course of A. salmonicida growth in DM-, measured in parallel by alamarBlue reduction and CFU counts per milliliter. The vertical lines indicate the log phase of growth. (C and D) The effects of mucins (100 μg/ml) on A. salmonicida growth, expressed as the percent difference from the growth rate of bacteria without mucin treatment and measured by alamarBlue reduction during the A. salmonicida log phase of growth. Data show the effect of mucins in the presence of glucose (DM+) for 30 min (C) and the effect in the absence of glucose (DM-) after 160 min (D); the longer duration was chosen to achieve total growth similar to the growth in DM+. The differences induced by mucins are expressed as the percent difference from the dialysis control (mucin isolation buffer dialyzed in parallel with the mucins), and the assay was performed at pH 7.4 with a Ca2+ concentration of 0.43 mM. Data points (n = 20) represent mucin samples isolated from four organs of five individuals (biological replicates, not pooled). The results were reproduced twice and are plotted as means ± SEM. Statistics used included a one-way ANOVA with Dunnett's post hoc test to analyze differences between growth in the absence of mucins (control) and in the presence of mucins (n = 5) *, P < 0.05; **, P < 0.01; ***, P < 0.001. Abbreviations: AB reduction, reduction of alamarBlue; pyloric, pyloric ceca; proximal, proximal intestine; distal, distal intestine.
A. salmonicida is not able to cleave Neu5Ac, but Neu5Ac on Atlantic salmon mucins modulate A. salmonicida growth.Each of our Atlantic salmon mucin isolates carried 50 to 70 O-glycan structures differing in size, linkages, and monosaccharide composition (21, 22). These structures were either linear or branched, the former having one terminal residue and the latter having two to four terminal residues (Fig. 2A). The majority of the skin mucin glycans had linear O-glycans (one terminal residue), 20% of the structures had two terminal residues, and less than 1% had three terminal residues (Fig. 2A). A total of 60 to 70% of mucin O-glycans from the intestinal sites had two terminal residues, 5 to 15% had linear structures, 22 to 26% had three terminal residues, and less than 1% had four terminal residues (Fig. 2A). The terminal residues were galactose, N-acetylglucosamine (GlcNAc), N-acetylgalactosamine (GalNAc), fucose, Neu5Ac, NeuGc, and Kdn. Mucins from all organs were highly sialylated; the dominant terminal sialic acid on mucins from all sites was Neu5Ac, and skin mucins also terminated in Neu5Gc and Kdn (Fig. 2B). N-Acetylhexosamines (including GlcNAc, GalNAc, and residues for which we could not determine if they were GlcNAc or GalNAc and were therefore labeled HexNAc) were the second most abundant terminal residues in the proximal and distal intestine, and the relative amounts of these at glycan terminals were lower in skin (Fig. 2B). In the pyloric ceca, Gal was the second most abundant terminal residue (35%) (Fig. 2B), whereas Gal was less abundant at the other sites. Fucose constituted less than 10% of the terminal residues at all tissue sites, with the highest relative abundance on the skin mucins.
Terminal residues present on Atlantic salmon mucins. (A) The mean relative abundance of structures with 1 to 4 terminal residues (branches) on the O-glycans of mucins from skin, pyloric ceca (pyloric), proximal intestine (proximal), and distal intestine (distal) (n = 5). (B) Weighted average ratios of terminal residues present on the mucin O-glycans. Abbreviations: HexNAc, N-acetylhexosamine (GlcNAc or GalNAc); GlcNAc, N-acetylglucosamine; GalNAc, N-acetylgalactosamine; Fuc, fucose; NeuAc, N-acetylneuraminic acid; NeuGc, N-glycolylneuraminic acid; Kdn, deaminated neuraminic acid.
To unravel the molecular basis for the differences in growth with mucins from different epithelial cell types, we performed correlation analysis with the level of binding and glycosylation. We found correlations between the size (number of residues in the structure) of the O-glycan structures and A. salmonicida growth, in both DM+ (Pearson's r18 = 0.64, P ≤ 0.01) and DM- (Pearson's r18 = 0.57, P ≤ 0.01). Furthermore, growth correlated with the Neu5Ac quantity of those mucins, both in DM+ (Pearson's r18 = 0.53, P ≤ 0.05) and DM- (Pearson's r18 = 0.63, P ≤ 0.01). Therefore, we hypothesized that A. salmonicida can cleave Neu5Ac for subsequent nutrient utilization. Aeromonas hydrophila, a species closely related to A. salmonicida, has sialidase/neuraminidase activity (32). The presence of neuraminidase in A. salmonicida has not been confirmed in the literature, nor did we find homology for the A. hydrophila neuraminidase in the sequenced A. salmonicida A449 strain. Furthermore, in the present study, A. salmonicida whole culture (including supernatant) did not cleave the 2′-(4-methylumbelliferyl)-α-d-N-acetylneuraminic acid sodium salt (4MU-Neu5Ac) substrate during a 12-h time interval, whereas sialidase A [used as a positive control; it cleaves α(2-3)-, α(2-6)-, α(2-8), and α(2-9)-linked Neu5Ac from complex carbohydrates] cleaved most of the α-bonds of the 4MU-Neu5Ac substrate within 5 h (Fig. 3A). This result suggests that A. salmonicida cannot release terminal Neu5Ac from Atlantic salmon mucins and that the increase in growth after culture with mucins was not due to the uptake of Neu5Ac by A. salmonicida.
Effects on A. salmonicida growth and β-N-acetylhexosaminidase expression of glycans and monosaccharides present on terminals of Atlantic salmon mucins. (A) Sialidase activity was measured by analyzing the fluorescence of liberated 4MU from the 4MU-Neu5Ac substrate. Sialidase A (positive control) released fluorescent product, whereas the negative control (A. salmonicida growth medium), as with the A. salmonicida whole-cell culture, did not cleave 4MU-Neu5Ac regardless of Ca2+ concentration. RFU, relative fluorescence units. The results are plotted as means (n = 4). (B) A. salmonicida growth in response to 100 μg/ml core 1 (Galβ1-3GalNAcα), Gd1a (Neu5Acα2-3Galβ1-3GalNAcβ1-4(Neu5Acα2-3)Galβ1-4Glcβ- ganglioside), GlcNAcβ1-3Gal, or GalNAcβ1-3Gal. Results are pooled from two independent experiments and are means ± SEM (n = 9). (C) A. salmonicida growth in response to monosaccharides (100 μg/ml) that were present terminally on the mucin glycans. Values denote the percent change in alamarBlue reduction compared to growth without carbohydrates (control). Results were reproduced three times and are plotted as means ± SEM. (D) A. salmonicida β-N-acetylhexosaminidase gene (ASA_RS14095) mRNA expression after culture with glycans and mucins. Data represent normalized fold expression levels compared to the control, as determined by quantitative PCR analysis (n = 2, 3 technical replicates each). The mean of the control CT values was 27.5. The results were reproduced twice and are plotted as means ± SEM. Statistical analyses entailed one-way ANOVA and Dunnet's test (for panels B andC). *, P < 0.05; **, P < 0.01; ***, P < 0.001. HexNAc, N-acetylhexosamine (GlcNAc or GalNAc); Glc, glucose; Gal, galactose; GlcNAc, N-acetylglucosamine; GalNAc, N-acetylgalactosamine; Fuc, fucose; NeuAc, N-acetylneuraminic acid; NeuGc, N-glycolylneuraminic acid.
On one hand, we found a positive correlation between the abundance of the salmon mucin Neu5Ac and A. salmonicida growth, suggesting that Neu5Ac enhances A. salmonicida growth. However, on the other hand, A. salmonicida was not able to cleave Neu5Ac. To reveal the role of Neu5Ac in this mucin-pathogen interaction, we carried out growth experiments with salmon mucin samples with and without enzymatic removal of the sialic acids, achieved by using sialidase A. Approximately 60% of the terminal Neu5Ac residues were removed by the enzymatic treatment, as shown using mass spectrometry (Table 1). The A. salmonicida growth rate was higher with mucins treated with sialidase A than in cultures with mock-treated mucins from any of the four epithelial sites (Fig. 4A to D; P ≤ 0.05). This opposes that Neu5Ac on mucins is growth promoting for A. salmonicida. Furthermore, since we have previously shown that A. salmonicida binding to these mucins decrease after sialidase treatment (22), the increase in growth after sialidase treatment (Fig. 4) indicates that A. salmonicida growth and mucin binding are not causally related. It is likely that the increased growth after desialylation was due to terminal Neu5Ac shielding other mucin O-glycans from enzymatic digestion and subsequent utilization by A. salmonicida.
Glycan structures present on the mucins before and after enzymatic treatmenta
Effects of sialidase A treatment on A. salmonicida growth in response to mucins. Sialidase A treatment of mucins from different organs resulted in higher growth rates of A. salmonicida versus rates for mucins that had undergone the same treatment in the absence of enzyme (mock). The data points represent technical replicates of the mucin samples pooled from five individual Atlantic salmon. The assay was performed at pH 7.4 with a Ca2+ concentration of 0.43 mM. The results are plotted as means ± SEM. Statistical analysis entailed Student's t test (n = 5). *, P < 0.05; **, P < 0.01; ***, P < 0.001. The results were reproduced twice.
N-Acetylglucosamines present on mucin glycans increase the growth rate of A. salmonicida.The abundance pattern of terminal residues that best matched the growth response to mucins from the different regions was that of terminal GlcNAc combined with HexNAc (compare Fig. 1C and D with 2B). To investigate which of the terminal residues could have contributed to the increased growth of A. salmonicida in response to mucins from the different epithelial regions (Fig. 1 C and D), we cultured the bacteria in the presence of the different oligosaccharides and monosaccharides occurring on the mucins. Among the studied oligosaccharides, only GlcNAcβ1,3Gal increased the growth of A. salmonicida (P ≤ 0.05) (Fig. 3B). The only monosaccharide among those found in our mucin samples that increased A. salmonicida growth was GlcNAc (+43%; P ≤ 0.01). Adding glucose, which is not present on salmon mucins, increased the growth rate of A. salmonicida by 65% (P ≤ 0.001) (Fig. 3C).
Removal of terminal HexNAc from mucins reduces growth of A. salmonicida, and sialidase treatment exposes HexNAc for bacterial access.β-N-Acetylhexosaminidase treatment of skin and distal intestinal mucins partially removed GlcNAc, GalNAc, and HexNAc (by 15 to 40%, determined using mass spectrometry) (Table 1). The growth response of A. salmonicida to the enzymatically treated mucins was reduced compared to the mock-treated mucins (P < 0.05 and 0.001, respectively; n = 10) (Fig. 5A and B). Furthermore, the level of terminal GlcNAc and HexNAc that were made accessible for interactions with bacteria by enzymatic removal of sialic acid (which was shown in Fig. 4 to increase A. salmonicida growth) increased after sialidase treatment (Table 1). Neu5Ac is both bulky (C11H19NO9) and negatively charged, and therefore sialic acids removed from branches terminating adjacent to GlcNAc and HexNAc also increase accessibility (Table 1). Together with the results in Fig. 3C, these results support that the terminal GlcNAc determines A. salmonicida growth in response to mucins.
Effect of decreasing the levels of accessible GlcNAc on mucins for A. salmonicida growth. (A and B) A. salmonicida growth in response to mucins with and without prior β-N-acetylhexosaminidase treatment. The data points represent technical replicates of the mucin samples pooled from five individual Atlantic salmon. The experiment was carried out twice, and the data points are pooled from the experiments. The assay was performed at pH 7.4 with a Ca2+ concentration of 0.43 mM; results are plotted as means ± SEM. Statistical analysis entailed Student's t test (n = 10). Mock, mucins treated with enzyme buffer with no enzyme present. Enzymatic removal of GlcNAc and HexNAc was confirmed via mass spectrometry (see Table 1).
Genomic analysis of transporters and β-N-acetylhexosaminidase.The β-N-acetylhexosaminidase gene is part of a cluster of 10 genes in strain A449 and possibly part of the same operon. This cluster includes five genes annotated as oligopeptide transporters, an endoglucanase, an N-acetylglucosamine kinase, and a hypothetical methyl-accepting chemotaxis protein. (GenBank identifiers for the genes in strain A449 are ABO91093.1 , ABO91094.1 , ABO91095.1 , ABO91096.1 , ABO91097.1 , ABO91098.1 , ABO91099.1 , ABO91100.1 , ABO91101.1 , and ABO91102.1 ). We searched 27 published A. salmonicida genomes and identified the same 10 genes and gene cluster in all strains.
The genome of the strain used in this study has not been sequenced. To investigate if the strain used in our study also carried this gene, we analyzed gene expression using primers designed against this enzyme in strain A449, and a threshold cycle (CT) value of 27.5 in unstimulated A. salmonicida indicated that this is the case. Neither mucins nor glycans affected expression of this enzyme (Fig. 3D). This gene is thus constitutively expressed. However, since no functional studies have been performed on this enzyme, we cannot be certain it is responsible for GlcNAc utilization.
Ca2+ and pH affect the growth of A. salmonicida, and Ca2+ modulates the effect Atlantic salmon mucins have on A. salmonicida growth.The environmental pH and calcium concentration may influence susceptibility of fish to infection, affecting both the host (33, 34) and pathogen (35). Furthermore, these levels may differ between different growth and assay buffer media used for in vitro studies. To reveal the effects of these factors on A. salmonicida growth, the pathogen was cultured in four different media based on DM-. Media with pHs of 7.2 and 7.8 were chosen, as the pH of salmon blood is 7.8, whereas the pH of many assay media is 7.2. The pH in the fish gut lumen and water surrounding the fish can vary more widely. Ca2+ concentrations of 0.1 mM (corresponding to very soft FW) and 10 mM Ca2+ (the concentration in SW) were compared to 0.43 mM (a standard concentration of the in vitro DM medium and within the range present in normal FW) in the experiments described above. In the absence of mucins, 10 mM CaCl2 resulted in 3-fold-higher growth of A. salmonicida than 0.1 mM CaCl2 at both pHs (P < 0.001; n = 4) (Fig. 6A and B). Furthermore, growth was 20% higher at pH 7.2 than at 7.8, both with 0.1 mM CaCl2 (P < 0.05; n = 4) and 10 mM CaCl2 (P < 0.001; n = 4) (Fig. 6A and B). At 10 mM CaCl2, the intestinal mucins enhanced A. salmonicida growth by 30 to 40% (Fig. 6C and D) (P < 0.001), similar to the results obtained at 0.43 mM (Fig. 2B), whereas at 0.1 mM CaCl2 mucins had no statistically significant effect on A. salmonicida growth. The effect of skin mucins on A. salmonicida growth ranged from −6% (0.1 mM CaCl2, pH 7.8) to +10% (0.1 mM CaCl2, pH 7.2) and was not statistically different from growth without mucins, although A. salmonicida cultured with skin mucins under a high CaCl2 concentration grew faster than under the low concentration (P ≤ 0.05, n = 12) (Fig. 6C). The effects on both skin and intestinal mucins on A. salmonicida growth tended to be higher at pH 7.2 compared to pH 7.8 (P = not significant). Thus, a moderate drop in pH, increase in Ca2+ concentration, and distal intestinal mucins all enhance A. salmonicida growth (Fig. 6E). Furthermore, as intestinal mucins are growth promoting at 10 mM Ca2+ and 0.43 mM (Fig. 2), but not at 0.1 mM, the shift between intestinal mucins promoting growth or not occurs between 0.1 mM and 0.43 mM Ca2+. For the results shown in Fig. 6A to E, CaCl2 was used to manipulate the Ca2+ concentration; however, similar results were obtained in parallel growth assays with equimolar Ca(NO3)2 (data not shown), demonstrating that the change in Cl− was not the cause of the effect. A. salmonicida binding to skin and distal intestinal mucins was similar in 0.43 mM Ca2+ at pH 7.8 as described in the previously published results, from an experiment performed at pH 7.2 in the absence of Ca2+ (22). Increasing the concentration to 10 mM Ca2+ increased binding to both mucins by 25% (P < 0.05) (Fig. 6F). Thus, the Ca2+ concentration affects A. salmonicida growth per se and also the effects mucins have on A. salmonicida growth and mucin binding.
Effect of Ca2+ and pH on A. salmonicida growth and mucin binding. (A) Growth curves of A. salmonicida, measured with alamarBlue, at pH 7.2 or 7.8 and with 0.1 mM Ca2+ versus 10 mM Ca2+ (n = 12). (B) Bar graph showing the area under the curve values from panel A. The results are plotted as means ± SEM (n = 12). (C) Effect of Ca2+ on Atlantic salmon mucins and their alterations of growth of A. salmonicida in DM- at pH 7.8. *, P < 0.05; ***, P < 0.001. The distal intestinal mucins increased A. salmonicida growth compared to the nonmucin control at 10 mM Ca2+. ##, P < 0.01. (D) Effect of pH on Atlantic salmon mucins and their alterations of growth of A. salmonicida in DM- with 10 mM CaCl2 (not significant [ns]). The distal intestinal mucins increased A. salmonicida growth, compared to the nonmucin control, at pH 7.2 and pH 7.8. ##, P < 0.01. Data are presented as the percent difference in alamarBlue reduction versus that with the nonmucin control under each condition (i.e., the data presented in panels A and B; n = 12). Statistical analyses entailed Student's t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001, based on one-way ANOVA with Dunnet′s post hoc test. ##, P < 0.01. The results were plotted as means ± SEM of technical replicates (mucins were pooled) and were reproduced three times. (E) Overview of the absolute growth, presented as the fold difference in alamarBlue reduction, normalized to that with DM- medium at pH 7.8 with 0.1 mM CaCl2 (y = 1). (F) Adhesion of A. salmonicida to mucins after culture in DM (pH 7.8) containing 0.43 mM Ca(NO3)2 versus 10 mM Ca(NO3)2. The assay was performed with 10 mM Ca(NO3)2 or in the absence of Ca(NO3)2 (control). A. salmonicida binding to mucins was similar after culture in the presence of 0.43 mM Ca(NO3)2, as reported in a previous study performed at pH 7.2 after culture in BHI broth (22).
DISCUSSION
In this study, we demonstrated that mucin glycans, pH, and Ca2+ affect A. salmonicida growth and that the Ca2+ concentration determines if mucins have an effect on A. salmonicida growth. The mucin glycosylation differs between epithelial sites, and the effects of mucins on A. salmonicida growth depend on the epithelial site from which the mucins were isolated. Sialic acids, which are the most abundant terminal residues on Atlantic salmon glycans, shield the remaining glycan from being utilized as a growth-promoting nutrient and cannot be utilized as a nutrient source by A. salmonicida. Of the terminal residues present on the salmon mucins, only GlcNAc promotes A. salmonicida growth. Since removal of GlcNAc and HexNAc decreased the growth in response to the mucins, whereas removal of sialic acids, which expose more HexNAc and GlcNAc, increased growth, accessible GlcNAc appears to be the factor that causes increased A. salmonicida growth in response to intestinal mucins.
Information on the effect of fish mucins on pathogenic bacteria is scarce. Analyzing growth of A. salmonicida in vitro with purified mucins is a feasible approach to estimate how this pathogen reacts to the mucin environment. Previous studies suggested both slower (36) and higher (37) bacterial growth rates in vivo versus in vitro. The in vitro growth rate of a bacterial pathogen thus cannot be directly translated to the in vivo situation, as factors such as host immune defense and nutrient availability will modulate the growth pattern. A. salmonicida needs to adapt to the in vivo environment before establishing disease and increase in numbers (38). On the other hand, in vitro growth experiments can give reliable indications about certain aspects of the mucin defense, separately from other factors that might influence an in vivo infection. Performing in vivo studies where one removes the mucins/mucus by physical means can damage the epithelial surface, whereas removing them long term via genetic knockout, for example, leads to a vastly changed microbiota. In vitro experiments therefore represent a reductionist approach, as the effects of these confounding factors are removed.
Atlantic salmon skin mucins are less glycosylated, carry a lower diversity of structures than intestinal mucins (21, 22), and linear structures dominate the skin glycan repertoire whereas intestinal mucin O-glycans are predominantly branched. Intestinal mucins enhance A. salmonicida growth, whereas skin mucins have no effect, and intestinal mucins also provide more adhesion sites for A. salmonicida than skin mucins. Although we see no benefit for the host in having glycans that enhance pathogen growth, this may be a side effect of an evolutionary benefit of providing diverse mucin glycans for retaining a beneficial commensal flora. In external sites such as the skin and gills, where there is a constant convection of water washing possible pathogens away, a more inert glycan repertoire may be beneficial, providing both fewer sites for adhesion and fewer grazing opportunities for both pathogenic and beneficial bacteria. Although the ability of other salmon-associated bacteria to utilize mucin glycans is unknown, the skin of the fish has indeed a different flora and considerably lower bacterial counts than the intestine (39).
Over 90% of both skin and intestinal glycans contain sialic acids, and Neu5Ac is the most abundant sialic acid on all sites, although the skin also contains approximately 20% Neu5Gc and a small percentage of Kdn (21). The large sizes of sialic acids, their negative charges, and their terminal location support that they can play important roles in interactions with bacteria. We previously demonstrated that Neu5Ac is an important part of the epitope to which A. salmonicida adheres (22), and we have demonstrated here that Neu5Ac shields mucins from promoting growth of this pathogen. Numerous bacteria are able to use sialic acids as a carbon and energy source (40). The literature provides no evidence that A. salmonicida possesses any genes for sialidase enzymes, and the results showing that A. salmonicida growth increased after removal of sialic acids from the mucins, as well as A. salmonicida's inability to cleave 4MU-Neu5Ac, support that A. salmonicida lacks the ability to cleave and subsequently utilize mucin-linked Neu5Ac. Sialic acids thus have a complex role: Neu5Ac promotes a closer contact between mucins and A. salmonicida through adhesion, but they also protect growth-enhancing O-glycans underneath the terminal Neu5Ac from A. salmonicida access. In vivo, a microflora containing bacteria that are able to cleave Neu5Ac may aid A. salmonicida utilization of mucin glycans, and the composition of the microbiota may therefore impact the severity of the infection.
GlcNAc appears responsible for the increased growth observed when A. salmonicida is cultured with Atlantic salmon mucins. The role of GlcNAc in the synthesis of the peptidoglycan cell wall may be a reason that GlcNAc positively affects A. salmonicida growth. In Escherichia coli, Neu5Ac has been shown to promote type I fimbria production, while GlcNAc has an inhibitory role (41). As type I fimbriae are important adhesion and virulence factors in A. salmonicida (42), one can speculate that Atlantic salmon mucins exert similar changes in fimbria production. The presence of GlcNAc may signal to the pathogen that it is in the mucus niche, and downregulating such factors may in turn free up more resources for growth. Indeed, in the human stomach, evidence suggests that downregulating pathogen binding to the gastric mucins is beneficial for the pathogen, as its removal from the gastric niche is decreased (43), and the pathogen decreases the level of adhesins after culture with mucins carrying the ligand to which it adheres (44).
Many of the Atlantic salmon O-glycan structures are different from those described for mammals (21) and are not commercially available. Therefore, we deciphered the mucin glycan effect by measuring the growth rate in the presence of the monosaccharide building blocks of salmon mucins, and we combined those findings with the results of enzymatic modification of the salmon mucins. Theoretically, A. salmonicida has the enzyme repertoire to cleave and metabolize all sugars on Atlantic salmon mucins except for sialic acids (http://www.cazy.org/b534.html ). We showed that GlcNAc and GlcNAcβ,3Gal increased the growth of A. salmonicida, but the other monosaccharides or glycans had no effect. Elimination of Neu5Ac increased both the terminal availability of HexNAc and GlcNAc and the growth rate of A. salmonicida in response to mucins, whereas enzymatic removal of GlcNac reduced the growth rate of A. salmonicida. Furthermore, of the residues present terminally on mucins, the pattern of abundance that best matched the proliferative response to mucins from the different regions was that of terminal GlcNAc and HexNAc. The only teleost O-glycome published so far is that of Atlantic salmon from one river (21). It is possible that Atlantic salmon from different regions and breeds differ in mucin glycan repertoire, and since the mucin O-glycans affect the pathogen, certain breeds may be more or less hospitable to A. salmonicida infection. The Atlantic salmon glycan repertoire may thus be a factor to take into account during breeding and selection of strains, in order to select a fish with robust innate defenses against infection.
Ca2+ concentrations and pH vary substantially between (and within) aquaculture sites, and both factors can also vary between bacterial culture media and buffers used for in vitro assays. The pH in the milieu adjacent to the epithelial surface under the mucus is most likely close to the pH of the blood, whereas further out in the mucus niche it is more affected by factors such as the pH of the surrounding water (skin) and digestive secretions (intestine); in parallel to that, such a gradient has been demonstrated in rat gastric mucus (45). The range investigated here thus represents moderate changes within the pH range that may occur in the mucus niche. Even though we found a 20% higher growth rate of A. salmonicida at pH 7.2 than at pH 7.8, the effect of mucins on A. salmonicida growth was very similar for these pH levels. Most bacterial growth media, assay buffers, and reporter agents for microbiological experiments have a pH of 7.2 or 7.4, due to their having been developed for mammalian research, whereas some methods that have been specifically developed for fish research are performed at a pH closer to 7.8. Thus, differences in pH between most growth media and assay buffers used for A. salmonicida research may have slight effects on the amplitude of the results, but not on the conclusions. Furthermore, the inverse relationship between partial CO2 pressure and pH results in reduced water pH if the CO2 levels are not properly managed. Thus, our results indicate that management of pH close to 7.8 may be beneficial from a pathogen control perspective, by lowering the bacterial growth rate.
Also, Ca2+ concentrations vary with water source and geographical area, and we studied 0.1 mM (a concentration corresponding to that in very soft FW), 0.43 mM (within the range present in normal FW and also that present in a standardized defined bacterial growth medium, DM), and 10 mM Ca2+ (corresponding to the concentration that can be found in very hard FW or in SW). Our results demonstrated that the Ca2+ concentration strongly affects A. salmonicida growth per se. In addition, the Ca2+ concentration determines if A. salmonicida can utilize intestinal mucins in a growth-promoting manner and affects the amplitude of the A. salmonicida adhesion to mucins. The greater growth-changing effects of distal intestinal mucins in our assays after the addition of CaCl2 suggest that intestinal mucins have higher plasticity for altering bacterial growth with changes in Ca2+ concentration than skin mucins. Furthermore, Ebanks et al. demonstrated that Ca2+ is necessary for the development of the A. salmonicida type III secretion system, the major virulence factor of A. salmonicida (35), suggesting that the environmental Ca2+ level is not only instrumental in determining pathogen attachment and growth but is also one important factor that determines the virulence of A. salmonicida. Although the mechanism for how Ca2+ contributes to the A. salmonicida ability to utilize mucins for growth-promoting purposes is unknown, the Clostridium perfringensN-acetylhexosaminidase has been shown to coordinate a calcium ion suggested to stabilize the enzyme (46), and a calcium ionophore inhibits secretion of a murine hexosaminidase, pointing to the possibility that an enzyme directly involved in the digestion of the glycan may be regulated by calcium (47). For A. salmonicida to be able to utilize mucins for growth-promoting purposes, the concentration needs to be above 0.1 mM, whereas we obtained similar results with 0.43 mM and 10 mM. It may therefore be that environmental Ca2+ levels are especially important for farming of Atlantic salmon in smolt production facilities with hard water, brackish water, or SW which contains Ca2+ levels above 0.1 mM. Environmental Ca2+ levels would affect the host-pathogen interaction not only on skin and gills but also the intestine, especially in brackish water and SW. It is well established that fish in a hyperosmotic environment continuously drink water and this affords ion-coupled fluid uptake (48, 49). The uptake of Ca2+ in the gut is low (50), but the alkaline luminal environment as a result of secondary active HCO3− secretion causes the ionized Ca2+ to precipitate into CaCO3 (51). Thus, the ionized form decreases along the intestine, with levels typically between 2 and 6 mM in rectal fluids (23, 49, 50). Thus, these levels are in the range of growth-promoting Ca2+ levels.
The present paper provides insights into Atlantic salmon mucin defense mechanisms against the furunculosis-causing pathogen A. salmonicida from the aspect of bacterial adhesion and growth. Pathogen-mucin interactions are affected by both environmental Ca2+ levels and pH, and these should be considered during in vitro studies. From an in vivo point of view, our findings suggest that ensuring a pH close to 7.8 and low Ca2+ concentrations may limit A. salmonicida growth in tank-based aquaculture. Furthermore, since the mucin glycan repertoire affects growth of the pathogen, the glycan repertoire may be a factor to take into account during breeding and selection of strains for aquaculture.
MATERIALS AND METHODS
Fish and sampling procedures.Atlantic salmon parr (Långhult lax, Långhult, Sweden; the same fish as used in our previously published work [21, 22]) were transported to the Department of Biology and Environmental Sciences and kept in 500-liter tanks. The fish were held in recirculating 10°C FW, supplemented with 10% SW (yielding a salinity of 2 to 3‰), at a flow rate of 8.5 liters/min. The fish were exposed to a simulated natural photoperiod and were hand-fed ad libitum once daily with a commercial dry pellet (Nutra Olympic/3-mm diameter plate; Skretting Averøy, Ltd., Stavanger, Norway).
Five fish (31.06 ± 0.49 cm body length and 280.70 ± 12.78 g body weight; means ± standard errors of the means [SEM]) were randomly netted, anesthetized in metomidate (12.5 mg/liter), and killed with a sharp blow to the head. Mucus from the skin was sampled by gentle scraping of the entire skin surface using microscopy glass slides. The fish were then opened longitudinally and the intestine, from the last pyloric ceca to the anus, was quickly dissected. The intestine was cut open along the mesenteric border, the proximal region was separated from the distal at the ileorectal valve, and the mucus and mucosa were scraped off using microscopy slides. The pyloric ceca were dissected and placed in liquid nitrogen and pulverized using a mortar and a pestle. All samples were placed in 10 mM sodium dihydrogen phosphate containing 0.1 mM phenylmethylsulfonyl fluoride (PMSF) pH 6.5 (sampling buffer) to inhibit proteolytic cleavage.
Isolation and purification of mucins.The scrapings and pulverized tissues in sampling buffer were added to five sample volumes of extraction buffer (6 M guanidine hydrochoride [GuHCl], 5 mM EDTA, 10 mM sodium phosphate buffer [pH 6.5] and containing 0.1 M PMSF), homogenized with a Dounce homogenizer (four strokes with a loose pestle), and stirred slowly at 4°C overnight. The insoluble material was removed by centrifugation at 23,000 × g for 50 min at 4°C (Beckman JA-30 rotor), and the pellet was reextracted twice with 10 ml extraction buffer. The supernatants from these three extractions were pooled and contained the GuHCl-soluble mucins used in the subsequent assays. The samples were dialyzed twice against 10 volumes of extraction buffer and filled to 26 ml with extraction buffer. CsCl was added to the samples by gentle stirring, and the samples were transferred to Quick Seal ultracentrifuge tubes (Beckman Coulter). The tubes were filled with 10 mM NaH2PO4 to give a starting density of 1.35 g/ml, and samples were subjected to density gradient centrifugation at 40,000 × g for 90 h at 15°C. The fractions were collected from the bottom of the tubes with a fraction collector equipped with a drop counter. Density gradient fractions of purified mucin samples were analyzed for carbohydrates as periodate-oxidizable structures in a microtiter-based assay. Fractions were diluted 1:100, 1:500, and 1:1,000 in 4 M GuHCl were added to 96-well plates (PolySorp; Nunc A/S, Roskilde, Denmark) to coat them and incubated overnight at 4°C. The rest of the assay was carried out at 23 to 24°C. After washing three times with washing solution (5 mM Tris-HCl, 0.15 M NaCl, 0.05% Tween 20, 0.02% NaN3; pH 7.75), the carbohydrates were oxidized by adding 25 mM sodium metaperiodate in 0.1 M sodium acetate buffer, pH 5.5, for 20 min. The plates were washed again and the wells were blocked with Delfia blocking solution (50 mM Tris-HCl, 0.15 M NaCl, 90 mM CaCl2, 4 mM EDTA, 0.02% NaN3, 0.1% bovine serum albumin [BSA]; pH 7.75) for 1 h. After further washing steps, the samples were incubated for 1 h with 2.5 mM biotin hydrazide in 0.1 M sodium acetate buffer, pH 5.5, followed by another washing step. Europium-labeled streptavidin was diluted 1:1,000 in Delfia assay buffer (50 mM Tris-HCl, 0.15 M NaCl, 20 mM diethylenetriaminepentaacetic acid, 0.01% Tween 20, 0.02% NaN3, 1.5% BSA; pH 7.75 [PerkinElmer]) and was added to the wells. After 1 h of incubation, the plates were washed six times and then incubated with Delfia enhancement solution (0.05 M NaOH, 0.1 M phthalate, 0.1% Triton X-100, 50 mM tri-n-octylphosphineoxide, 15 mM 2-naphthoyltrifluoroacetone; PerkinElmer) for 5 min at 120 rpm on an orbital shaker. Fluorescence (λexcitation = 340 nm; λemission = 615 nm) was measured using a Wallac 1420 Victor2 plate reader with the europium label protocol (PerkinElmer, Waltham, MA, USA). Density measurements were performed using a Carlsberg pipette as a pycnometer; 300 μl of sample was aspirated into the pipette, weighed, and the density was calculated as grams per milliliter. The DNA concentration was calculated based on the light absorbance at 280 nm.
Preparation of mucin samples.Gradient fractions containing mucins were pooled to obtain one sample for each gradient (Fig. 1A). Mucin concentrations in pooled samples were determined based on their carbohydrate content: serial dilutions of the samples were compared with a standard curve prepared from a fusion protein, constructed from MUC1, 16TR, and IgG2a Fc. The standard curve started at a concentration of 20 mg/ml followed by seven 1:2 serial dilutions. The carbohydrates were detected as periodate-oxidizable structures in a microtiter-based assay described above. This method of concentration determination was chosen because all mucins do not come into solution after freeze-drying, and determining the concentration by freeze-drying therefore can generate large errors as well as selectively remove certain mucin species. Although this is not an exact measure of concentration, it can be used to ensure that the mucins are at the same concentration for comparative assays, and since bacterium-mucin interactions largely occur via the mucin glycans (15), setting the concentration based on the glycan content appears most appropriate.
Aeromonas salmonicida culture conditions. A. salmonicida strain VI-88/09/03175 (culture collection of Central Veterinary Laboratory, Oslo, Norway) was cultured in brain heart infusion broth (BHI) at 19°C, and stocks were stored in BHI-glycerol (1:1) at −80°C. Before the growth experiments, A. salmonicida was cultured overnight in BHI broth on a laboratory shaker at 120 rpm and washed three times with phosphate-buffered saline (PBS; 140 mM NaCl, 2.7 mM KCl, 10 mM phosphate buffer; pH 7.4) at 3,000 × g for 5 min.
Growth assay. A. salmonicida cells, at a concentration equivalent to an optical density at 600 nm (OD600) of 0.1, were cultured in glucose-containing (DM+) and glucose-lacking (DM-). These two media contained 0.43 mM Ca(NO3)2 and the pH was 7.4. For the analyses on the effects of pH and Ca2+ on the growth of A. salmonicida, the DM- pH was set to 7.2 or 7.8 via HCl. Two Ca2+ concentrations, corresponding to soft FW (0.1 mM) and SW (10 mM), were used. The former had an osmolality of 98 mOsm/liter and for the latter it was 127 mOsm/liter. CaCl2 was used to manipulate the Ca2+ concentration; to certify that the accompanying change in Cl− did not have a major impact on the results, we carried out parallel growth assays with equimolar Ca(NO3)2 or CaCl2 solutions and obtained similar results (data not shown). Purified mucin samples in 4 M GuHCl were dialyzed against PBS eight times and diluted in sterile PBS to a concentration of 100 μg/ml. The glycans were used at the same concentration. The GuHCl-containing isolation buffer was dialyzed and diluted in parallel and used as a reference of normal growth (dialysis control [DC]). PBS was used as a reference for growth in the presence of monosaccharides and oligosaccharides. Monosaccharides were purchased from Sigma-Aldrich. GD1a-oligosaccharide, GlcNAcβ1,3Gal, and GalNacβ1,3Gal were purchased from Elicityl, France. Core 1 glycan (Galβ1,3GalNAc) was purchased from Dextra Laboratories, UK. Bacteria were cultured in 96-well Nunc-Delta surface flat-bottom plates (Nunc A/S, Roskilde, Denmark) in 4 to 8 replicates at 23°C and 120 rpm in the presence of alamarBlue (Invitrogen) to monitor reduction of the dye.
Culture media.(i) DM- medium contained 0.1 g/liter Ca(NO3)2·4H2O, 0.4 g/liter KCl, 0.1 g/liter MgSO4·7H2O, 6 g/liter NaCl, 0.8 g/liter Na2HPO4, 2 g/liter NaHCO3, 2 mg/liter FeSO4, 5 g/liter BSA, 50 mg/liter adenine, 3 mg/liter lipoic acid. For amino acids, DM- medium contained 44.5 mg/liter alanine, 632 mg/liter arginine, 75 mg/liter asparagine, 66.5 mg/liter aspartic acid, 120 mg/liter cysteine, 73.5 mg/liter glutamic acid, 300 mg/liter glutamine, 37.5 mg/liter glycine, 110 mg/liter histidine, 262.5 mg/liter isoleucine, 262 mg/liter leucine, 362.5 mg/liter lysine, 75.5 mg/liter methionine, 165 mg/liter phenylalanine, 57.5 mg/liter proline, 52.5 mg/liter serine, 238 mg/liter threonine, 51 mg/liter tryptophan, 180 mg/liter tyrosine, 234 mg/liter valine. The vitamin content of DM- medium was as follows: 0.2 mg/liter d-biotin, 3 mg/liter choline chloride, 1 mg/liter folic acid, 35 mg/liter myo-inositol, 1 mg/liter niacinamide, 1 mg/liter p-aminobenzoic acid, 1.25 mg/liter d-pantothenic acid, 1 mg/liter pyridoxine hydrochloride, 0.2 mg/liter riboflavin, 1 mg/liter thiamine hydrochloride, 5 μg/liter vitamin B12 (cyanocobalamin). (ii) The DM+ medium was the same as DM- medium, with the addition of 2 g/liter glucose.
Aeromonas salmonicida binding to mucins.The mucin binding assays were carried out as described previously (22). Briefly, mucins were diluted in 4 M GuHCl–PBS and added to 96-well Polysorp plates overnight at 4°C. The plates were washed three times with washing buffer (PBS containing 0.05% Tween 20) and blocked for 1 h with blocking buffer (0.5% BSA in washing buffer). After discarding the blocking buffer, bacteria cultured in the log phase of growth were washed and diluted in blocking buffer to an OD600 of 0.1 and then further diluted 1:10 in blocking buffer. The bacterial suspension was added to the plates, which then were incubated at 120 rpm for 2 h. The plates were washed three times and then incubated for 1 h at room temperature with blocking buffer containing anti-A. salmonicida IgG monoclonal antibody (clone 3B11/G5; Austral Biological, San Ramon, CA) diluted 1:5,000. After washing, horseradish peroxidase (HRP)-conjugated anti-mouse IgG (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) diluted 1:10,000 in blocking buffer was added. After further washing steps, 100 μl tetramethylbenzidine (TMB) substrate (Sigma-Aldrich Co.) was added, and the plates were incubated for 20 min. The reaction was stopped with 100 μl of 0.5 M H2SO4, and the plates were analyzed in a microplate reader at 450 nm after color stabilization. The binding was analyzed at four mucin concentrations (2-fold dilution steps) with results that were similar when the binding signal was expressed per unit of carbohydrate signal for at least three of the concentrations (i.e., the assay was performed within a linear range), including a glycan value of 20,000 europium counts (approximately corresponding to a 4-μg/ml mucin concentration). For each analysis, parallel microtiter plates were coated for glycan detection analysis to ensure that small differences in binding were not due to mucin coating differences. The reported binding values were normalized for a glycan value of 20,000 europium counts.
To study the effect of Ca2+ on A. salmonicida binding, a modified binding assay was used with the following differences from the assay described above. A. salmonicida was cultured overnight in DM- containing 0.43 mM Ca(NO3)2 (control) or 10 mM Ca(NO3)2. The blocking buffer used in each step of the assay was set to pH 7.8, had no calcium in the control wells, and included 10 mM Ca(NO3)2 for the rest of the wells.
Analysis of mucin glycan structures and correlation with bacterial growth.A data set from a previously published study on the O-glycan structures from salmon mucins, based on analysis using liquid chromatography-mass spectrometry and tandem mass spectrometry (MS/MS) data (21), was used to calculate the data shown in Fig. 2. Weighted average glycan chain lengths of carbohydrate structures on each mucin sample were calculated using the relative abundance (percentage) of each glycan structure. The Neu5Ac content used for the correlation analysis with bacterial growth was calculated using the relative quantity of each identified O-glycan structure and the number of Neu5Ac residues on those structures. The ratio of terminal residues (Fig. 2B) on Atlantic salmon mucins was calculated using each terminal residue on all structures (e.g., Neu5Ac, Neu5Ac, and GlcNAc on a branched structure with three terminal residues) and the relative abundance of those structures.
Genome alignment search.The β-N-acetylhexosaminidase gene is part of a cluster of 10 genes in strain A449, and these are possibly part of the same operon (GenBank identifiers of the genes in A449 are ABO91093.1 , ABO91094.1 , ABO91095.1 , ABO91096.1 , ABO91097.1 , ABO91098.1 , ABO91099.1 , ABO91100.1 , ABO91101.1 , ABO91102.1 ). Each of the 10 proteins in the cluster were used as queries in a Basic Local Alignment Search Tool (BLAST) search against all available A. salmonicida genomes (A449, 01-B526, CBA100, 2004-05MF26, JF3224, Y567, JF4097, Y47, Y577, RS534, JF3791, J223, J231, ATCC 33658, J227, 170-68, 01-B522, JF2267, 2009-157 K5, 2010-47 K18, JF2506, JF2507, JF3517, and A. salmonicida subsp. achromogenes AS03, A. salmonicida subsp. pectinolytica, and A. salmonicida subsp. masoucida NBRC). The output from these BLAST searches was used to identify genomic locations of the proteins in all strains. The A. salmonicida subsp. salmonicida A449 genome was also searched for homologies for the Aeromonas hydrophila sialidase.
Sialidase activity assay.Sialidase activity of A. salmonicida was analyzed with the modified method of Crost et al. (19). Briefly, 20 μl of fresh A. salmonicida liquid culture was added to a reaction mixture consisting of 500 μM 4MU-Neu5Ac (Sigma-Aldrich Co.) as a substrate in PBS, pH 7.4. The enzymatic reactions were carried out at 23°C for up to 10 h in a plate reader (BMG Labtech, Ortenberg, Germany). The fluorescence of the liberated 4MU was quantified at λexcitation of 340 nm and λemission of 420 nm at 10-min intervals in a white opaque 96-well plate (Falcon). Sialidase A (5 U/ml) diluted 1,000-fold in the final reaction mixture (ProZyme, Hayward, Canada) was used as a positive control, and the negative control was DM-. The assay was carried out at final CaCl2 concentrations of 0.1, 0.43, and 10 mM.
Desialylation and β-N-acetylhexosaminidase treatment of mucins.Sialidase A (ProZyme) or β-N-acetylhexosaminidase (ProZyme) treatment was performed on mucins pooled from five salmon. Samples were dialyzed against sterile 50 mM Na3PO4 buffer four times with a cellulose dialysis membrane with a molecular mass cutoff value of 14 kDa (Sigma-Aldrich Co.). PBS containing 4 M GuHCl was subjected to the same procedure, as the dialysis control (DC). Five microliters of sialidase A or β-N-acetylhexosaminidase enzyme was added to 500 μl of dialyzed mucin sample or DC. Sialidase A or β-N-acetylhexosaminidase enzyme was replaced with dialysis buffer for the mock-treated controls. Samples were incubated at 37°C, 70 rpm overnight for 13 h. After the enzyme treatment, samples were dialyzed against sterile PBS four times in a dialysis membrane with a molecular mass cutoff value of 100 kDa for sialidase A and 300 kDa for β-N-acetylhexosaminidase (Sigma-Aldrich Co.; sialidase A is 88 kDa, and β-N-acetylhexosaminidase is ca. 110 kDa). The dialyzed samples were stored at −80°C.
MS to confirm enzymatic removal of glycan moieties.β-Hexosaminidase, α-sialidase digestedm and mock-treated skin and distal intestinal mucins were dot blotted to a polyvinylidene difluoride membrane (Immobilon P membrane; Millipore, Billerica, MA) and visualized with alcian blue stain (Sigma-Aldrich). Stained spots were cut and subjected to reductive β-elimination with 0.5 M NaBH4 and 50 mM NaOH for 16 h at 50°C to release the glycans. The reduction reaction was quenched by using glacial acetic acid, and the mixture was desalted as the nonretained fraction on strong cation exchange resin packed as a microcolumn on top of hydrophobic C18 material. The solid-phase extraction removed cations and any remaining protein or peptide components. Released O-glycans were analyzed by liquid-chromatography (LC)-MS/MS using a 10-cm by 250-μm (inner diameter) column (in-house) containing 5-μm porous graphitized carbon particles (Thermo Scientific, Waltham, MA, USA) connected to an LTQ mass spectrometer (Thermo Scientific). O-Glycans were eluted using a linear gradient from 0 to 80% acetonitrile in 10 mM ammonium bicarbonate over 40 min at a flow rate of 250 nl/min. Electrospray ionization-MS was performed in negative ion polarity with an electrospray voltage of 3.5 kV, capillary voltage of −33.0 V, and capillary temperature of 300°C. The data were viewed and manually analyzed using Xcalibur software (version 2.2; Thermo Scientific). Due to the dense and complex glycosylation of mucins, complete enzymatic elimination of carbohydrate moieties cannot be expected. MS analysis confirmed success of the sialidase treatment, showing an overall ca. 60% reduction of sialylation, as exemplified by three sialylated structures (m/z 878, 1,081, and 1,372) and a consequent increase in relative abundance of selected nonsialylated structures (m/z 587 and isomers 790a and -b). The structures to which the m/z values correspond are defined in Table 1. After hexosaminidase treatment, we confirmed partial removal (15 to 45%, depending on individual structures) of HexNAc from distal intestinal mucins, as exemplified by structures with m/z 587, 790, and 1,081 and a subsequent increase in the corresponding structures with m/z 384 and 878, lacking the terminal HexNAc. In skin mucins, we found more than a 50% decrease in the abundance of m/z 587 and 790 and an increase in the abundance of the subsequent cleaved structure, m/z 384.
RNA extraction, cDNA synthesis, and real-time PCR.After 7 h of culturing A. salmonicida in DM- medium with 0.43 mM CaCl in the absence (control) or presence of 100 μg/ml glycan or mucin, 200 μl of RNAprotect bacteria reagent (Qiagen GmbH, Hilden, Germany) was added to each well, and 12 wells per sample were pooled. RNA extraction was continued with Qiagen's RNeasy kit, including the DNase treatment step with Qiagen's RNase-free DNase set.
The purity of RNA samples was verified based on the OD260/OD280 ratio. cDNA was synthesized using the Quantitect reverse transcription kit (Qiagen) according to the manufacturer's instructions. In “−RT” reactions, the reverse transcriptase enzyme was replaced with nuclease-free H2O, thereby enabling the detection of contaminating DNA only. We confirmed the absence of contaminating DNA in each sample. Real-time PCR assays were run in 96-well plates (Bio-Rad), in a total volume of 20 μl containing 10 μl Power SYBR green PCR master mix (Thermo Fisher Scientific), 9 μl molecular biology-grade water, and 1 μl of template diluted 10-fold. The primer pair for the A. salmonicida subsp. salmonicida A449 β-N-acetylhexosaminidase (ASA_RS14095) gene was designed using the Primer 3 software (http://bioinfo.ut.ee/primer3-0.4.0/primer3/ ) (forward, TGACCGTACCGACTACACCT; reverse, AAGTGGAAGGTCTTGAGCGG). For the 16S reference gene, we used a previously validated primer pair described previously by Menanteau-Ledouble et al. (forward, ATATTGCACAATGGGGGAAA; reverse, GTTAGCCGGTGCTTCTTCTG) (52). The primer pairs were used at a final concentration of 250 nM. Amplification efficiencies of the selected primers were tested and met general requirements of RT-PCR analysis. Amplification was performed using a CFX96 real-time PCR detection system (Bio-Rad Laboratories) with default settings for the amplification protocol and included a dissociation step in the end of the program for melting temperature (Tm) analysis to confirm amplification specificity. The purity of the template was controlled with the use of nontemplate control wells. The results in Fig. 3D are expressed as the fold change of the β-N-acetylhexosaminidase/16S gene expression ratio, compared to that for the nonglycan/mucin control.
Statistical analyses.Statistical analyses were performed using the Prism 5.0 software package (GraphPad Software Inc.). Student t tests and one-way analysis of variance (ANOVA) followed by Dunnett′s post hoc test were used to compare groups and treatments, as indicated in the figure legends. The level of significance was set at P ≤ 0.05.
ACKNOWLEDGMENTS
The work was supported by the Swedish Research Council Formas (223-2011-1073), the Swedish Research Council (2016-05154_3), Engkvists Foundation, Wilhelm and Martina Lundgrens Foundation, and the Norwegian Research Council (CtrlAQUA SFI 237856/O30). The mass spectrometer (LTQ) was supported by the Knut and Alice Wallenberg Foundation.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
FOOTNOTES
- Received 14 March 2017.
- Accepted 12 May 2017.
- Accepted manuscript posted online 22 May 2017.
- Copyright © 2017 American Society for Microbiology.