Enhanced phagocytosis and complement-mediated killing of Mannheimia haemolytica serotype 1 following in-frame CMP-sialic acid synthetase (neuA) gene deletion

ABSTRACT Mannheimia haemolytica is the most significant bacterial pathogen associated with the bovine respiratory disease complex. Although sialic acid is a known virulence factor in other members of Pasteurellaceae, such as Histophilus somni and Pasteurella multocida, the significance of sialic acid to the virulence of M. haemolytica is currently unknown. Therefore, the role of sialic acid as a virulence determinant of M. haemolytica was investigated by constructing an in-frame neuA [CMP-N-acetylneuraminic acid (Neu5Ac/sialic) synthetase] mutant, which was shown by high-performance anion exchange chromatographic analysis (HPAEC) to be devoid of sialic acid on the lipopolysaccharide (LPS). Both the neuA mutant and wild-type parent strains exhibited similar growth rates in the growth curve assay. Real-time qPCR and ELISA evaluation showed no differences in proinflammatory cytokine expressions (IL-1β, IL-6, and IL-8) between the neuA mutant and parent strain when peripheral blood mononuclear cells were incubated with LPS. Interestingly, the neuA mutant was three to four logs more sensitive to a whole-blood bacterial killing assay than the parent strain. Similar results were also observed in plasma and serum bacterial killing assays. Flow cytometry analyses showed higher uptake of neuA mutant by phagocytes, compared to the parent strain, in the whole-blood phagocytosis assay; however, no difference in reactive oxygen species production in neutrophils or monocytes was detected for either strain. Taken together, these results indicate that sialylation of M. haemolytica LPS plays a vital role in reducing complement-mediated and phagocytic killing. IMPORTANCE The Gram-negative coccobacillus Mannheimia haemolytica is a natural inhabitant of the upper respiratory tract in ruminants and the most common bacterial agent involved in bovine respiratory disease complex development. Key virulence factors harbored by M. haemolytica are leukotoxin, lipopolysaccharide, capsule, adhesins, and neuraminidase which are involved in evading innate and adaptive immune responses. In this study, we have shown that CMP-sialic acid synthetase (neuA) is necessary for the incorporation of sialic acid onto the membrane, and inactivation of neuA results in increased phagocytosis and complement-mediated killing of M. haemolytica, thus demonstrating that sialylation contributes to the virulence of M. haemolytica.

M. haemolytica harbors several virulence factors such as leukotoxin, lipopolysacchar ide (LPS), capsule, and various adhesins which can enhance bacterial colonization and survival in the host (6).When host defense mechanisms become suppressed follow ing environmental stress or viral infection, M. haemolytica (replication) of serotypes 1 and/or 6 is greatly increased in the upper respiratory tract which increases the odds of subsequent inhalation of M. haemolytica to the lower respiratory tract and establishment of fibrinous pneumonia (6).
Phagocytosis plays a crucial role in the host innate defense system.Ingestion of invading microorganisms by phagocytic cells, such as neutrophils, monocytes, and macrophages, results in bacterial killing by various mechanisms which include exposure to reactive oxygen species (ROS) (7,8).Consequently, infection frequently results if this balance is disrupted, such as in immunocompromised cattle or those with pre-existing health conditions (9).Alterations in local immunity likely contribute to the development of the BRDC.
Sialic acids are acidic amino sugars that share a common nine-carbon backbone and are found on the cell membranes of most vertebrate organisms.Frequently, sialic acid is a terminal sugar on cell surface molecules or serum glycoconjugates in mammals and birds (10,11).Many pathogenic bacteria, including Neisseria gonorrhoeae, Haemophilus ducreyi, Haemophilus influenzae, Histophilus somni (12)(13)(14)(15), and Pasteurella multocida (16), are known to incorporate sialic acid into the termini of their LPS or lipooligosaccharide (LOS) molecules, mimicking host glycoconjugates and the appearance of self, thereby evading host innate defense mechanisms (17,18).
The role of sialic acid as a virulence determinant in M. haemolytica is unknown, so this study was undertaken to investigate this question.To do so, an in-frame neuA [CMP-Nacetylneuraminic acid (Neu5Ac/sialic) synthetase] gene deletion mutant was constructed and compared to its parent strain using cytokine transcripts, and protein expression, phagocytosis, ROS production, and complement-mediated killing (plasma and serum) assays.

Bacterial growth curves
For growth measurements, overnight cultures of M. haemolytica WT and ΔneuA strains at a final 5 × 10 8 CFU/mL concentration were serially diluted 1:10 in BHI broth.Two hundred microliters of each serially diluted culture were transferred into Honeycomb 100-well plates (Growth Curves USA, Piscataway, NJ).The plates were covered with lids and placed in an automated growth curve reader (Bioscreen C; Growth Curves USA, Piscataway, NJ), which was programmed for continuous shaking at 37°C, and optical density readings at 600 nm (OD 600nm ) were recorded every hour for 24 h.BHI medium (without bacteria) was used as a negative control for the assay.The optical density data were analyzed, and growth curve graphs for each isolate were generated using MS Excel.

Construction of in-frame neuA [CMP-N-acetylneuraminic acid (Neu5Ac/sialic) synthetase] gene deletion mutant
Construction of the M. haemolytica D153 neuA MT was accomplished using the temperature-sensitive plasmid pCT109GA189, and the genetic manipulations to produce the in-frame mutant are described in detail elsewhere (19,20).The M. haemolytica D153 neuA MT was shown by sequencing to possess an in-frame deletion devoid of amino acids 42-414 (Fig. 1).To confirm the chromosomal deletion of neuA, PCR assay from whole cells was performed with the neuA primer pair: neuA F 5′-cgacaccttatgaagatgtgaat-3′ and neuA R 5′-attgcggatccaaatgca-3′ as previously described (19).

Purification of lipopolysaccharides
M. haemolytica WT and MT LPS were extracted from bacterial pellets prepared from 4 L of cultures (BHI broth without sialic acid supplement) by the hot phenol-water method described previously (21).Briefly, 4 L of culture was centrifuged at 10,000 × g for 15 min.The pellets were washed with PBS (pH 7.4) and resuspended in 25 mL/L of culture or 50 mL/5 g dry weight of cells in 50 mM sodium phosphate buffer (pH 7.0) containing 5 mM EDTA and 0.05% sodium azide.Hen egg white lysozyme (0.1 g, 15,500 units/mg) was added and stirred at 4°C for 16 h.The suspension was placed at 37°C for 20 min, then stirred vigorously for 3 min.The suspension volume was adjusted to 100 mL with 20 mM MgCl 2 , and bovine pancreas ribonuclease and deoxyribonuclease were added at final concentrations of 1-2 μg/mL.Then the suspension was incubated at 37°C for 10 min followed by another incubation at 60°C for 10 min.Then proteinase K (25 µg/mL) was added and incubated at 60°C for 60 min.An equivalent volume of 95% phenol was added and rapidly stirred for 15 min at 65-70°C.The extract was immediately cooled on ice to 15°C and centrifuged at up to 18,000 × g for 15 min to separate the phases.The upper aqueous phase was transferred to a clean tube, and extraction was repeated on the lower phase with the same volume of water.Aqueous phase was pooled, and residual phenol was removed by dialysis against 4 L of distilled.The crude-lyophilized LPS was diluted in distilled water (up to 5 mg/mL) and heated at 65°C to disperse aggregates.Samples were centrifugated at 105,000 × g for 3 h, and LPS pellet was dissolved in water and stored at −80°C.

Analysis of sialic acids in LPS by high-performance anion exchange chroma tography
To analyze the sialic acid content in the extracted LPS, samples were submitted to the complex carbohydrate research center, at the University of Georgia.To release sialic acids from samples, freeze-dried aliquots of LPS samples were hydrolyzed with 2 N acetic acid at 80°C for 3 h.Released sialic acids were identified by high-performance anion exchange chromatography/pulsed amperometry detection using a Dionex ICS 3000 system as described previously (22,23).

Blood sampling and CBC analyses
Four healthy female Holstein dairy cows, aged between 26 and 30 months, maintained at the National Animal Disease Center in Ames, Iowa, were selected for the study.Jugular venous blood was drawn into K2 EDTA vacutainer tubes (BD Vacutainer, Franklin, NJ, USA).Complete blood count (CBC) of each blood sample (collected into EDTA tubes) was performed using an automated hematology systems analyzer (Sysmex America, Mundelein, IL).

PBMC isolation and stimulation with LPS
Peripheral blood mononuclear cells (PBMCs) were isolated from acid citrate dextrose (ACD) blood samples (collected from four cows) by density-gradient centrifugation using Percoll as described previously (24).PBMCs were resuspended (2 × 10 6 /mL) in a complete RPMI 1640 growth medium (cRPMI).One hundred microliters of PBMC were transferred to a round-bottom 96-well plate followed by the addition of 100 µL cRPMI medium containing M. haemolytica WT and neuA MT LPS (final LPS concentrations = 0.1 and 1 µg/mL).Cells incubated with medium alone and E. coli LPS (1 µg/mL) were used as negative and positive controls, respectively.Cells were incubated for 4, 8, and 24 h at 37°C in a humidified atmosphere of 5% CO 2 incubator.Cells were harvested by centrifugation (1,000 × g for 5 min), and supernatants were pooled in each treatment group and stored at −80°C.Cell pellets were used for RNA purification.

RNA purification
Total cellular RNAs from untreated and LPS-treated PBMCs were purified using Qiagen RNeasy Mini Kit as described by the manufacturer (Qiagen Inc., Valencia, CA).Purified RNA was aliquoted and stored at −80°C until needed.cDNA synthesis cDNA was synthesized using purified total cellular RNA primed with random primers along with Super Script III first-strand synthesis system as described by the manufacturer (ThermoFisher, Carlsbad, CA).cDNA was diluted 1:10 with RNase-free water and stored at −20°C in aliquots until needed.

RT-qPCR and analyses
The proinflammatory cytokines and housekeeping gene-specific primers used in this study are listed in Table 1.Primers for three cytokines (IL-1β, IL-6, and IL-8) and house keeping gene (GAPDH) were synthesized by Integrated DNA Technologies (Coralville, IA).PCR reactions were conducted in 20 µL volume containing 2 µL cDNA template, 2.5 µL each primer (10 µM), and 10 µL of 2× SYBR Green PCR Master mix (Applied Biosystem, Foster City, CA) in nuclease-free water.Samples were loaded into 96-well plates (in duplicate wells) and sealed with optical Adhesive Cover (Applied Biosystems).A reaction mixture containing water, but no cDNA, was used as negative control.cDNA prepared from cells incubated with E. coli LPS (0.1 and 1.0 ug/mL) was used as a positive control for cytokines.PCR was run on an Applied Biosystems Quant Studio 5 Real-Time PCR System.Reaction conditions were as follows: 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s, 60°C for 1 min, 72°C for 30 s, and an extended elongation phase at 72°C for 10 min.At least two technical replicates were used to obtain each average Ct (AvgCt) value.Relative quantities (RQ) of cytokine transcripts for each reaction were calculated using Quant Studio Design and Analysis Software v1.5.1 [RQ = 2(−ΔΔCt) method] as described previously (25).

ELISA
Bovine IL-1β and IL-8 were purchased from Kingfisher Biotech (Saint Paul, MN, USA), and bovine IL-6 was purchased from ThermoFisher Scientific.ELISA was performed according to the manufacturer's instructions.

Whole-blood bacterial killing assay
A whole-blood bacterial killing assay was performed with freshly collected blood (four cows) and bacterial cultures at multiplicity of infection (MOI) at a 1:10 ratio (neutro phils/monocytes: bacteria).Briefly, 100 µL of heparin-anticoagulated blood (~5 × 10 5 neutrophils and monocytes) was transferred to a 1.5-mL microcentrifuge tube, and then 5 µL of bacteria [M.haemolytica WT or neuA MT (~5 × 10 6 CFU)] in PBS was added.Tubes were incubated at 37°C for 1 h in a CO 2 incubator.The number of bacterial CFU in the inocula and blood was determined by serially diluting in PBS (10-fold) plated on sheep TSA II blood agar plates and incubated overnight at 37°C in a CO 2 incubator.

Complement-mediated bacterial killing assay
The sensitivity of M. haemolytica WT and neuA MT to complement-mediated killing was assessed in both plasma and serum.Briefly, plasma and serum were separated from heparin whole blood or serum tubes collected from four cows by centrifugation.To inactivate complement, plasma and serum samples were incubated at 56°C for 30 min.Ninety microliters of plasma, serum, heat-inactivated plasma, and heat-inactivated serum were incubated with 5 µL M. haemolytica WT and MT (~5 × 10 6 CFU) and 5 µL PBS (total volume = 100 µL) at 37°C for 1 h in a CO 2 incubator.To enumerate CFU, samples were serially diluted in PBS (10-fold) and plated on sheep TSA II blood agar plates and incubated overnight at 37°C in a CO 2 incubator.

Flow cytometry for reactive oxygen species analysis in whole blood
M. haemolytica WT and ΔneuA strains were transformed with green fluorescent protein (GFP) expression plasmid for this assay.One hundred microliters of heparinized whole blood were incubated with LPS (10 µg/mL), 5 × 10 6 CFU of M. haemolytica WT or ΔneuA at 37°C for 1 h in a CO 2 incubator.After 30 min of incubation, CellROX Deep Red (ThermoFisher) was added to a final concentration of 2.5 µM.After 45 min of incubation, antihuman CD14 antibody clone M5E2 conjugated with Brilliant Violet 421 (Biolegend, San Diego, CA) was added.Red blood cells were lysed, and white blood cells were fixed with BD FACS lysing solution as per manufacturer's recommendation.Two hundred microliters of the lysed solution containing polymorphonuclear leukocytes

Gene name
Primer sequence (5'−3') Amplicon size (bp) (PMNs) and PBMCs from each microcentrifuge tube were transferred into individual wells in the round-bottom 96-well plate before flow cytometry evaluation.Cells were assessed using a BD FACSymphony A5 Flow cytometer (Research Cell Analyzer, BD Biosciences).PMNs and PBMCs were visualized in forward scatter vs side scatter followed by singlecell discrimination.Eosinophil cells were then identified by autofluorescence in a side scatter area (SSA) vs fluorescein isothiocyanate (FITC) plot.The neutrophil, monocyte (CD14 + ), and lymphocyte profiles were determined using an SSA vs BV421 plot.CellROX Deep Red stained cells were identified in an SSA vs APC for each population.Analysis gate positioning was done using Fluorescence minus one control for all fluorescence channels.Data were analyzed using FlowJo V10 software (BD Biosciences).

Statistical analyses
The data were analyzed using the general linear models (GLM) procedure from SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA).For the ROS and phagocytosis assays, the statistical model included the effect of treatment (MT vs WT).For the analysis of secreted cytokines (IL-1β, IL-6, and IL-8) by ELISA, the statistical model included the effect of treatment [WT, MT, positive (E. coli LPS), and negative control].The RT-qPCR for the cytokines (IL-1β, IL-6, and IL-8) included the effect of treatment (WT, MT, positive, and negative control), time (4 and 8 h), and their interaction.For the plasma and serum bacterial killing assays, the model included the effect of treatment [WT, MT, and heat-inactivation of plasma and serum (yes or no)].Mean comparisons were done using the predicted differences option when significant.Significant differences were established at P < 0.05.

Construction of M. haemolytica neuA mutant
Detailed examination of the amino sugar and nucleotide sugar metabolism pathway of M. haemolytica D153 with Kegg site (https://www.genome.jp/pathway/mhae00520+F382_04085) and the whole-genome sequence analysis (GenBank Accession no: CP005972.1)revealed that D153 strain appears to contain an intact de novo sialic acid synthesis pathway but not sialic acid uptake genes (nanPU).Therefore, we targeted neuA (CMP-N-acetylneuraminic acid synthetase) to prevent the final stage of LPS sialylation.An M. haemolytica ΔneuA MT (Δ43-414 amino acids) was generated using the temperature-sensitive plasmid pCT109GA189-Kan.The precise neuA deletion is depicted in (Fig. 1A).Deletion of neuA in M. haemolytica MT strain was confirmed by PCR assay (Fig. 1B).

Growth patterns of M. haemolytica WT and neuA MT in broth culture
To determine whether deletion of M. haemolytica neuA affected bacterial growth rate, both WT and ΔneuA strains were grown in BHI broth at 37°C for 24 h in a growth curve reader, and OD 600nm readings were recorded at hourly intervals.Similar growth kinetics were observed for both M. haemolytica WT and MT strains, indicating that deletion of ΔneuA did not impact on the growth rate but did slightly reduce peak absorbance of this mutant (Fig. 2).

Identification of sialic acid in LPS preparations by high-performance anion exchange chromatography
LPS is the major component of outer membrane of Gram-negative bacteria and consists of lipid A, oligosaccharide core, and polysaccharide O-antigen.Sialic acid (or neu5Ac) residues are typically located within O-antigen of LPS, whereas sialic acid residues are located at the terminus of O-antigen in most LOS (26).Since M. haemolytica does not appear to have a sialic acid uptake system (nanPU), we did not add exogenous sialic acid into growth medium.LPS was extracted from overnight BHI cultures as described in the Materials and Methods.To define the sialic acid retention time by HPAEC chromato grams, sialic acid standards were used to generate a standard curve.Sialic acid retention time was ~7 min, the concentration and peak were linear (data not shown).M. haemoly tica WT LPS preparation contained a peak with a retention time close to sialic acid standard (~7 min) with an area of 0.2720 (Fig. 3A).However, no such visible peak indicative of sialic acid was observed with ΔneuA LPS preparation (Fig. 3C).To further confirm the presence (or absence) of sialic acid, each sample was spiked with sialic acid standard.Sialic acid peak area of WT LPS and added sialic acid was 0.8653 (Fig. 3B), close to the combined peak area of sialic acid from the LPS sample (0.2720) and the added standard (0.6558).However, sialic acid peak area in neuA MT LPS preparation and added sialic acid was 0.6229 (Fig. 3D), close to the added sialic acid standard value.Taken together, these HPAEC findings support that LPS of M. haemolytica WT (but not ΔneuA) is sialylated.

Quantification of cytokine transcripts by RT-PCR
Previous reports define the induction of inflammatory cytokines in bovine alveolar macrophages when stimulated with M. haemolytica LPS (27,28).Since HPAEC analysis confirmed that M. haemolytica WT LPS is sialylated, the goal of RT-qPCR assay was to determine whether there is a difference in proinflammatory cytokine transcripts (mRNA) expression following stimulation of PBMCs (prepared from four cows) with non-sialylated LPS compared to sialylated LPS.Cells stimulated with LPS from E. coli were used as a positive control (for cytokines), while unstimulated cells were used as a negative control.We tested two concentrations of LPS (0.1 and 1.0 µg/mL) at two time points (4 and 8 h) for cytokine transcripts expression analyses.Significantly, high IL-1β (P < 0.0269) and IL-8 (P < 0.0001) transcript expression levels were observed with PBMCs at 4 h stimulated with both M. haemolytica WT and neuA MT LPS preparations as well as with E. coli LPS (Fig. 4A and C).In contrast, significant IL-6 transcript expression at 4-h stimulation was observed with M. haemolytica WT LPS (Fig. 4B; P < 0.0171) and E. coli LPS (Fig. 4B; P < 0.0002) but not M. haemolytica neuA MT.Significant transcript expression levels for all three cytokines were observed at 8-h incubation for both LPS concentrations with all three LPS preparations (Fig. 4D through F; P < 0.0006).It is important to highlight that except IL-6 expression at 4-h stimulation with M. haemolytica MT LPS preparation, we did not observe any significant difference in cytokine transcript expression levels between M. haemolytica sialylated (WT) and non-sialylated (neuA MT) LPS as well as with E. coli LPS.

Detection of secreted cytokines by ELISA
In addition to evaluating proinflammatory cytokine transcripts expression, we also measured secreted cytokines (at protein level) following stimulation of PBMCs with sialylated (WT) and non-sialylated (neuA MT) LPS prepared from M. haemolytica strains by ELISA.Cells stimulated with LPS from E. coli were used as a positive control, while unstimulated cells were used as a negative control.PBMCs prepared from four animals were stimulated with LPS (0.1 and 1.0 µg/mL), and culture supernatants were collected at 24 h post-incubation.We observed significantly higher IL-1β and IL-8 production when PBMCs were incubated with both concentrations of M. haemolytica WT, neuA MT, and E. coli LPS compared to untreated control (Fig. 5A and B; P < 0.0093); however, there were no differences between LPS preparations (P = 0.7952).Although significantly higher IL-6 transcript expression levels were observed with all three LPS preparations (Fig. 4E), significant IL-6 protein production was only observed with E. coli LPS (Fig. 5C, P < 0.0001).

Whole blood, plasma, and serum bacterial killing assay
The effect of sialic acid on complement-mediated killing of H. somni and P. multocida has been previously reported (16,29).Although sialylated (LOS) H. somni was highly resistant to phagocytosis and complement-mediated killing ( 29), no such resistance was observed with sialylated (LPS) P. multocida for complement-mediated killing (16).Since these two closely related bacteria showed differential sensitivity to phagocytosis and complementmediated killing, it was of interest to compare the sensitivity of sialylated (WT) and non-sialylated (neuA MT) M. haemolytica by both killing mechanisms.It has recently been reported that whole blood is better for opsonophagocytic bacterial killing which mimics the natural killing mechanism compared to incubation of complement-opson ized bacteria with purified neutrophils and monocytes (30).Therefore, we performed M. haemolytica killing assay with whole blood (heparinized) collected from four cows.CBC was initially performed with all four blood samples before the phagocytosis assay to determine the neutrophil and monocyte numbers.Approximately, 3.7 × 10 5 neutrophils and 1.5 × 10 5 monocytes were present in 100 µL of blood sample.Therefore, in order to have 1:10 MOI, 100 µL of blood sample (~5 × 10 5 neutrophils and monocytes) was incubated with 5 µL WT or neuA MT inocula (~5 × 10 6 CFU).The CFU of M. haemolytica WT and neuA MT, which were incubated for 1 h with PBS control treatment, were similar to the original inocula (data not shown).However, significant reduction in M. haemolytica MT numbers (1,000-10,000-fold) compared to WT was observed following 1-h incuba tion with whole blood (Fig. 6A, P < 0.0001).We further performed a bacterial killing assay with normal and heat-inactivated (56°C for 30 min) plasma and serum samples to assess the involvement of complement during bacterial killing.Significant reductions in bacterial numbers (100-1,000-fold) were also observed when M. haemolytica MT was incubated with plasma or serum samples as compared to heat-inactivated plasma or serum samples (Fig. 6B and C; P < 0.0001).Although magnitude of bacterial killing was different (compared to neuA MT), significant reduction in M. haemolytica WT numbers was also observed with normal plasma and serum as compared to heat-inactivated plasma and serum (Fig. 6B and C; P < 0.0001).These findings strongly suggest that M. haemolytica with sialylated LPS is significantly more resistant to both phagocytic-and complement-mediated killing.

Detection of phagocytized bacteria and ROS production by flow cytometry
It is well known that the ROS in phagocytes such as neutrophils, monocytes, and macrophages can directly kill ingested bacteria by causing oxidative damage to biocompounds and stimulating other pathogen elimination mechanisms.CellROX fluorogenic reagents have been widely used for ROS detection and quantification in living cells (31).Therefore, to assess the role of ROS in both neutrophils and monocytes in terms of bacterial killing of M. haemolytica, flow cytometry analyses were performed with whole blood which was incubated with GFP expressing M. haemolytica WT and MT strains, CD14 mAb, and CellROX Deep Red as described in the Materials and Methods.Flowcytometry analyses revealed significantly higher numbers of M. haemolytica neuA MT [as per high mean fluorescent (GFP) intensity] were in neutrophils as compared to WT bacteria (Fig. 7A; P < 0.0004).Although higher numbers of M. haemolytica MT compared to WT bacteria were also found in monocytes [again, as per high mean fluorescent GFP) intensity], bacterial numbers were not significantly different (Fig. 7B; P = 0.0558).Despite having higher numbers of M. haemolytica neuA MT as compared to WT in the neutrophils and monocytes, the amount of ROS in both cell types was not significantly different (Fig. 7C; P < 0.6042 and Fig. 7D; P < 0.1381).

DISCUSSION
Despite current efforts to control BRDC through vaccination, this syndrome continues to cause significant economic losses for bovine, ovine, and caprine producers associated with medication expenses, mortality, and productivity losses, respectively.Sialic acid is found in the tissues of vertebrate animals, typically detected as a terminal sugar exposed on cell surface molecules (proteins or lipids) or serum glycoconjugates in mammals and birds (10).Contributing to evasion of host defense mechanisms, pathogenic bacteria may incorporate sialic acid into the termini of their LPS or LOS molecules (17,18,32).Several types of bacteria are capable of incorporating sialic acids into LPS/LOS, including N. gonorrhoeae, H. ducreyi, H. influenzae, P. multocida, and H. somni, where it has been shown to be a virulence factor (12)(13)(14)(15).Sialic acid is a known virulence factor in multiple members of the Pasteurellaceae family, such as H. somni (29) and P. multocida (16).Molecular mimicry is an adaptation postulated to help bacterial infections evade and/or neutralize the host's immune responses by presenting a cell surface similar to that of host cells (33).Bacteria typically acquire sialic acid in two ways: (i) by acquiring it from the surrounding environment or (ii) by de novo synthesis.H. somni (15) and P. multocida (11,16) can acquire sialic acid from the environment or growth medium via the nanPU uptake mechanism.Although M. haemolytica apparently does not possess a sialic acid update system (such as nanPU), genomic analysis indicates that it does possess an intact de novo sialic acid synthesis pathway.Whether synthesized de novo or acquired from outside sources, NeuA [CMP-N-acetylneuraminic acid (Neu5Ac/sialic) synthetase] gene product activates sialic acid before transfer of sialic acid to recipient molecules such as LPS.The present study finding showed that, unlike the parent strain, LPS of M. haemolyticaΔneuA was not sialylated.The failure of the M. haemolyticaΔneuA strain to sialylate LPS provides clear evidence that an intact neuA gene is essential for LPS sialylation under the conditions we grew our cells.Interestingly, N. gonorrhoeae scavenges host CMP-Neu5Ac, the activated form of sialic acid, to sialylate its LOS (34).Similar to P. multocida and H. somni, organisms like H. influenzae and H. ducreyi cannot synthesize sialic acid and rely on their respective transporters to supply sialic acid to the LPS sialylation pathways (35).For several diseases, LPS sialylation is a crucial virulence component (10).NeuA in E. coli activates Neu5Ac before its inclusion into the K1 and K92 capsules, whereas the N. meningitidis ortholog performs a similar function in capsule and LPS production.Additionally, linkage-specific sialyltransferases catalyze the sialylation of LPS.
It is reported that stimulation of bovine alveolar macrophages with M. haemolytica LPS induces inflammatory cytokine expression (27,28), and our study supports this finding.Furthermore, we determined that stimulation by M. haemolytica LPS lacking sialic acid elicited proinflammatory cytokine expression of IL-1β, IL-6, and IL-8 by PBMCs, which were similar to wild-type LPS as determined by real-time qPCR and ELISA.These findings suggest that the sialylation status of M. haemolytica LPS does not substantially affect the expression of proinflammatory cytokines.In contrast, Zarc et al. (36) demon strated that Porphyromonas gingivalis LPS isolated from W83 strain is highly sialylated and has a significantly lower inflammatory potential compared to the less sialylated LPS of the ATCC 33277 strain of Porphyromonas gingivalis.
When bacteria are present in blood or lung tissue, they must withstand the bac tericidal activity of the complement system and phagocytes to survive, and to do so, they employ a number of mechanisms that mitigate immune recognition.Our findings highlight the importance of sialylation in the process by which bacterial pathogens subvert hosts' immune systems.Whole blood, plasma, and serum bac terial killing experiments showed a dramatic decrease in bacterial survival in the ΔneuA mutant (compared to wild-type counterpart), demonstrating the involvement of sialylation in the evasion of complement-mediated killing.The ΔneuA mutant was also more efficiently phagocytosed than the parent in whole-blood bacterial killing assay, demonstrating that sialylation is involved in suppressing phagocytosis.Our results demonstrated that M. haemolytica expressing sialylated LPS was significantly more resistant to phagocytic and complement-mediated killing than was the ΔneuA mutant strain.N. gonorrhoeae and non-typable H. influenzae use LPS sialylation to suppress the complement alternative route, albeit in distinct ways (37).Additionally, Bugla-Płoskońska et al. (38) demonstrated that the addition of sialylated LOS from Salmonella O48 strains to human serum reduces complement activation.
We found that higher levels of M. haemolytica ΔneuA in neutrophils and monocytes did not lead to significantly increased ROS production.Neutrophil and monocyte ROS production are not altered in the ΔneuA compared to the parent strain, suggesting that sialylation does not play a significant role in ROS production and that the reduced ΔneuA survival was likely due to complement-mediated killing and increased phagocytosis.During the early stages of infection, phagocytes play a critical role in antimicrobial defense against bacterial pathogens (39).The primary method, which neutrophils and monocytes exert antimicrobial activity, is mediated by phagocytosis and once inside bacteria are eliminated by oxygen-dependent and oxygen-independent pathways (40).Findings in this study are supportive that M. haemolytica, like a number of other bacteria, employs sialylation of LPS to evade host immune intervention, thereby enhancing the capacity of the bacterium to colonize host tissues.The significance of sialylation in vivo requires additional research.

Conclusions
By constructing and testing an in-frame M. haemolytica neuA (CMP-N-acetylneuraminic acid synthetase) mutant, we have demonstrated that LPS sialyation reduced phagocyto sis and complement killing in vitro, and the coating of M. haemolytica LPS with sialic acids likely contributes to the survival of this bacterium in the host.

FIG 2
FIG 2 Growth curve assessment of M. haemolytica WT and neuA MT.M. haemolytica WT and neuA MT, from overnight cultures grown on tryptic soy agar plates, supplemented with 5% sheep blood and incubated overnight at 37°C in 7.5% CO 2 were serially diluted 1:10 in BHI broth and incubated in an automated growth curve reader for continuous shaking at 37°C, and optical density readings at 600 nm (OD 600nm ) were recorded every hour for 24 h.One representative graph out of at least three independent experiments is shown.

FIG 3
FIG 3 Identification of sialic acid (Neu5Ac) in LPS.(A) M. haemolytica WT LPS was extracted from overnight culture.The chromatogram of M. haemolytica WT LPS displays a distinct peak with a retention time close to the sialic acid standard, approximately at 7 min.The area of the peak was measured 0.2720, indicated that M. haemolytica WT contains a significant amount of sialic acid.(B) M. haemolytica WT LPS spiked with sialic acid standard shows a sialic acid peak area of 0.8653, close to the combined peak area of sialic acid from the LPS sample (0.2720) and the added standard (0.6558).(C) M. haemolytica neuA MT LPS showed no visible peak indicative of sialic acid.(D) M. haemolytica neuA MT LPS spiked with sialic acid standard exhibits a sialic acid peak area of 0.6229, similar to the added sialic acid standard value.

FIG 5 FIG 6
FIG 5 Cytokine production in PBMCs stimulated with LPS.Measurement of IL-1β, IL-6, and IL-8 production by ELISA in PBMCs stimulated with M. haemolytica sialylated (WT) and non-sialylated (neuA MT) LPS, as well as E. coli LPS.PBMCs from four animals were stimulated with LPS at concentrations of 0.1 and 1.0 µg/mL, and culture supernatants were collected after 24 h of incubation.Significant production of (A) IL-1β and (B) IL-8 was observed when PBMCs were incubated with both concentrations of M. haemolytica WT, MT, and E. coli LPS, compared to the untreated control.(C) Significant IL-6 production was only found with E. coli LPS.The results are presented as mean ± standard deviation of duplicate experiments.*P < 0.0093; **P < 0.0001.

FIG 7
FIG 7 Flow cytometry quantified reactive oxygen species generation in the whole blood.To assess the role of ROS in neutrophils and monocytes regarding bacterial killing, flow cytometry analyses were performed using whole blood incubated with GFP-expressing M. haemolytica WT and neuA MT strains and CellROX Deep Red, as described in the Materials and Methods section.Y-axes indicate relative fluorescence units (RFU) of GFP-expressing M. haemolytica (A and B) and ROS (CellROX Deep Red; C and D).X-axes indicate treatment (MT or WT).(A) Flow cytometry analyses revealed a significantly higher number of M. haemolytica neuA MT bacteria in neutrophils compared to WT bacteria (indicated by a higher mean fluorescent intensity of GFP bacteria).(B) Similarly, a higher number of M. haemolytica MT bacteria were detected in monocytes compared to neuA MT bacteria.However, the difference in bacterial numbers between M. haemolytica WT and MT strains was not significantly different.(C) The amount of ROS in neutrophils, measured by mean fluorescent intensity, was not significantly different between the M. haemolytica WT and MT strains.(D) Likewise, the level of ROS in monocytes showed no significant difference between the M. haemolytica WT and MT strains.*P < 0.0004; **P = 0.0558; ***P < 0.6042; ****P < 0.138.

TABLE 1
List of the primers used for RT qPCR