Infectious Salmon Anemia Virus Specifically Binds to and Hydrolyzes 4-O-Acetylated Sialic Acids

Viruses and cells.

The virus strain ISAV Glesvaer/2/90 (5) was cultivated on ASK cells (7) or SHK cells (5). Cells were grown in Leibowitz L-15 medium with 10% fetal calf serum. Virus from approximately 600 ml of cell culture supernatant was polyethylene glycol precipitated, layered on a continuous sucrose gradient, and centrifuged at 150,000 × g (35,000 rpm in a Beckman SW41 rotor) overnight. Banded virus was collected by side puncture of the tube wall, pelleted, and resuspended in PBS.

Hemagglutination and hemagglutination inhibition assays.

Hemagglutination in microtiter plates was carried out as previously described (10). The hemagglutination inhibition test was performed by first mixing 25 μl of inhibitor, diluted serially in PBS, with 25 μl of virus suspension containing 4 hemagglutination units, followed by incubation for 1 h. Fifty microliters of 0.5% Atlantic salmon erythrocytes or 0.75% rabbit or horse erythrocytes was then added, and the agglutination end point was read following 1 h of incubation on ice. The hemagglutination inhibitors tested included bovine submaxillary mucin (BSM), rat serum, guinea pig serum, horse serum, and rabbit serum. Control preparations were made by saponification of O-acetyl esters; i.e., the inhibitors were incubated with 0.1 N NaOH as described earlier (58). After incubation for 30 min at room temperature, the samples were brought to neutral pH by addition of 1 N HCl. The hemagglutinating activity (hemagglutination units) is expressed as the reciprocal of the highest dilution showing complete agglutination of erythrocytes.

Enzyme assays.

The acetylesterase activity of ISAV was determined with p-nitrophenyl acetate as described previously (19, 53). One milliunit of viral esterase activity was defined as the amount of enzymatic activity resulting in the hydrolysis of 1 nmol of p-nitrophenyl acetate per minute.

For control reactions, we used a chimeric recombinant soluble influenza C virus HE fused to green fluorescent protein (3) and a recombinant soluble esterase derived from the cloned HE gene of sialodacryoadenitis virus (SDAV) fused to GFP. These recombinant esterases specifically hydrolyze glycosidically bound Neu5,9Ac₂ or Neu4,5Ac₂, respectively. Details on the cDNA constructs of the recombinant O-acetylesterases and their expression by recombinant baculoviruses will be described elsewhere (47a).

Fluorimetric HPLC analysis.

Reverse-phase high-pressure liquid chromatography (HPLC) analysis of sialic acids was performed essentially as described previously (36). Free sialic acids were prepared from BSM or horse serum glycoproteins by acid hydrolysis (28) and purified as described previously (36). The precise position of O-acetylation at carbon 4 or 9 was determined by the elution characteristics on HPLC and by susceptibility to the acetylesterases derived from influenza C virus and MHV S, which specifically hydrolyze either Neu5,9Ac₂ (13) or Neu4,5Ac₂ (36).

For assays involving glycosidically bound sialic acids, virus was incubated at 16 or 37°C with either BSM (Sigma type I-S) or horse serum glycoproteins. Heat-inactivated virus was used as a control.

Samples containing glycosidically bound sialic acids were first hydrolyzed with 2 M propionic acid for 4 h at 80°C. The hydrolyzed mixtures were centrifuged for 10 min, and the supernatants were lyophilized. The released sialic acids were then incubated with 1,2-diamino-4,5-methylene-dioxybenzene (DMB) reagent for 1 h at 56°C. After centrifugation for 10 min, aliquots of the supernatant were injected into an RP-18 column and eluted isocratically with water-methanol-acetonitrile (86:7:9 by volume) at a flow rate of 1 ml per min. DMB-derivatized sialic acids were detected at an excitation wavelength of 373 nm and an emission wavelength of 448 nm.

Solid-phase binding assay.

Virus binding assays were performed on coated 96-well microtiter plates essentially as described elsewhere (19, 58). Serum was diluted in PBS and allowed to bind at 4°C overnight (100 μl/well). For assays involving complete de-O acetylation, serum was incubated with 0.2 M NaOH for 30 min at room temperature. The wells were then washed with PBS, and the remaining binding sites were blocked with 5% bovine serum albumin in PBS (150 μl/well) for 2 h at room temperature. For assays involving de-O acetylation of 4-O-acetylated sialic acids, incubation with recombinant SDAV esterase (1.5 mU/well) at 37°C overnight was performed. Heat-inactivated recombinant SDAV esterase was incubated at 95°C for 30 min prior to incubation, and then ISAV was allowed to bind for 2 h at 4°C. Unbound virus was removed by washing with PBS. Bound virus was detected by incubation with 100 μM 4-methyl umbelliferyl acetate. Hydrolysis of the substrate was detected at an excitation wavelength of 355 nm and an emission wavelength of 460 nm.

Receptor-destroying enzyme substrate specificity. (i) Analysis with free and glycosidically bound 9-O-acetylated sialic acids.

ISAV has been shown to exhibit an acetylesterase activity (10), but the precise nature of the viral enzyme and its substrate specificity remained unclear. In order to identify potential substrates for the viral esterase, we first determined whether purified ISAV strain Glesvaer could hydrolyze free O-acetylated sialic acids. When we incubated ISAV with free Neu5,9Ac₂, we observed a complete de-O acetylation of the substrate. It should be mentioned that complete de-O-acetylation was observed only when we used 30 mU of ISAV esterase. When we used 3 mU, a partial substrate conversion (40%) was detectable. In contrast, 3 mU of the influenza C virus esterase completely de-O-acetylated this substrate (data not shown). Since incubation with heat-inactivated virus resulted in an unchanged sialic acid profile, we concluded that ISAV specifically hydrolyzed free Neu5,9Ac₂ (Fig. 1) but at a lower reaction rate than for influenza C virus. Previous publications had shown that the ISAV esterase most likely hydrolyzes a substrate which is different from that hydrolyzed by the influenza C virus HE protein (10, 21). Since Neu5,9Ac₂ is the major substrate for influenza C viruses (13), this result was surprising. We then investigated whether the ISAV enzyme was capable of hydrolyzing glycosidically bound Neu5,9Ac₂. As a model substrate we used BSM, which contains relatively large amounts of O-acetylated sialic acids. Besides Neu5,9Ac₂ (37% of the total sialic acids), detectable amounts of 5-N-acetyl-7-O-acetyl neuraminic acid (Neu5,7Ac₂) (4.4%) and 5-N-glycolyl-9-O-acetyl neuraminic acid (Neu5Gc9Ac) (10.8%) were present (Fig. 2a). When we incubated BSM with ISAV at 37°C, we found that the sialic acid pattern was essentially unchanged, indicating that neither glycosidically bound Neu5,9Ac₂ nor Neu5,7Ac₂ or the glycolyl derivative Neu5Gc9Ac was hydrolyzed by ISAV (Fig. 2c). In contrast, the recombinant influenza C virus esterase (3; Strasser et al., submitted) efficiently hydrolyzed Neu5,9Ac₂ and Neu5Gc9Ac (Fig. 2b). To some extent, Neu5,7Ac₂ was also hydrolyzed by the influenza C virus esterase. This is due to the conditions of incubation at pH 7.2, which causes spontaneous migration of the acetyl group from position 7 to position 9 of the glycerol chain of neuraminic acid (16, 52). Since ISAV replicates more efficiently at 15°C than at higher temperatures (10), we investigated whether the incubation temperature of 37°C might influence hydrolysis of O-acetylated sialic acids on glycoproteins. Thus, we also incubated the substrate with ISAV at 16°C. Again, no specific hydrolysis of glycosidically bound O-acetylated sialic acids was detectable (Fig. 2d). Recently, it was speculated that ISAV might utilize O-acetylated 5-N-glycolyl-neuraminic acids for binding to erythrocytes (21). Our results strongly indicate that at least Neu5Gc9Ac can be excluded as a major substrate for the ISAV esterase. In contrast, the influenza C virus esterase hydrolyzed this derivative (Fig. 2b).

(ii) Specific hydrolysis of 4-O-acetylated sialic acids.

We then tested whether Neu4,5Ac₂ might represent a substrate for the ISAV esterase. First, we incubated purified free Neu4,5Ac₂, containing small amounts of Neu5Ac and Neu5Gc4Ac, with ISAV. After incubation at 37°C, both Neu4,5Ac₂ and Neu5Gc4Ac were almost completely hydrolyzed, and we observed a concomitant increase of the de-O-acetylated sialic acids Neu5Ac and Neu5Gc (Fig. 3b). Incubation with heat-inactivated ISAV resulted in an unchanged HPLC elution profile (Fig. 3c) compared to mock-incubated sialic acids (Fig. 3a). This result indicated that the hydrolysis of Neu4,5Ac₂ and Neu5Gc4Ac was due to the viral esterase. We finally examined whether glycosidically bound Neu4,5Ac₂ might represent a natural substrate. As the substrate we used horse serum glycoproteins, which contained bound Neu4,5Ac₂ (22.5%), Neu5Ac (68.4%), and Neu5Gc (9.1%). Incubation at 16 or 37°C resulted in a complete de-O-acetylation of the glycosidically bound Neu4,5Ac₂ (Fig. 4b and c), while heat-inactivated ISAV was unable to hydrolyze this substrate (Fig. 4d). As expected, recombinant influenza C virus esterase was unable to hydrolyze this substrate, confirming that the horse serum preparation contained no detectable Neu5,9Ac₂ (Fig. 4a).

ISAV receptor specificity.

To determine whether the substrate for the ISAV receptor-destroying activity is similar to the determinant for the receptor binding, we performed several hemagglutination experiments. In the initial characterization work with ISAV, hemagglutination was demonstrated only with fish erythrocytes (10). However, only a limited number of erythrocytes from mammals and birds were tested. The limited shelf life of fish erythrocytes compared to mammalian erythrocytes prompted us to test more animal species. On testing of erythrocytes from cow, goat, pig, horse, rabbit, mice, and rat, both horse and rabbit erythrocytes were agglutinated by ISAV (data not shown). Tissues or sera from both of these species have been shown to contain 4-O-acetylated sialic acids (11, 15). Also, receptor-destroying activity was demonstrated, as indicated by elution of the hemagglutination. The hemagglutination titers were not significantly different from that observed with fish erythrocytes; however, elution of the hemagglutination (due to the esterase) was significantly faster with rabbit erythrocytes than with fish and horse erythrocytes. While elution took 5 to 6 h to complete with fish and horse erythrocytes, elution of rabbit erythrocytes was almost complete following 1 h of incubation at room temperature.

We then investigated whether preparations known to contain predominantly 9-O-acetylated sialic acids (BSM and rat serum) or 4-O-acetylated sialic acids (guinea pig serum, horse serum, and rabbit serum) could inhibit hemagglutination. The results are presented in Table 1 and show that guinea pig, horse, and rabbit sera as well as BSM inhibited hemagglutination. Further, inhibition was abolished following saponification of acetyl esters in guinea pig, horse, and rabbit sera, indicating that the inhibition is caused by binding to O-acetyl esters, most probably 4-O-acetylated sialic acids. In contrast, saponification did not influence the inhibition by BSM, and thus the inhibition by BSM was not mediated by O-acetyl esters.

To further confirm these results, we performed solid-phase binding assays. Here we coated serum glycoproteins from BSM, horse, and guinea pig onto polystyrene enzyme-linked immunosorbent assay plates, followed by incubation with ISAV preparations at 4°C to allow the ISAV to bind. For a control, alkali- or mock-treated glycoproteins were used. After the removal of unbound virus, incubation with 4-methyl umbelliferyl acetate was performed to detect bound ISAV. As shown in Table 2, specific binding was observed between ISAV and serum glycoproteins from horse and guinea pig but not from BSM. These binding results are in accordance with our results regarding the substrate specificity of the ISAV HE. In addition, we wanted to preincubate the serum glycoproteins with the recombinant SDAV esterase, which is specific for 4-O-acetylated sialic acids (47a). Although we did not obtain complete inhibition of ISAV binding after recombinant SDAV esterase treatment, there was a substantial reduction in binding (Table 2). A similar partial destruction of ISAV receptors was also seen when erythrocytes from horse and rabbit were treated with the recombinant SDAV esterase prior to hemagglutination assays (Table 3).