In vitro antiviral activity against herpes simplex virus in the abalone Haliotis laevigata

INTRODUCTION

The greatest biodiversity on earth is found in oceans, with 32 of the 33 animal phyla, 14 of which are not found on land (Nybakken & Bertness, 2005). Currently there are an estimated 230 000–275 000 taxonomically described marine species, and with the new species inventory accruing 1300–1500 species a year, there is predicted to be between 1.4–1.6 million species of macrofauna and flora in the ocean (Bouchet, 2006). Viruses are also abundant in oceans, with about 10⁹ or more viruses per litre of water, exceeding numbers of bacteria and Archaea by about 15-fold (Bergh et al., 1989; Fuhrman, 1999; Suttle, 2007). Viral infections are common in marine environments, with an estimated 10²³ infections occurring every second (Suttle, 2007).

Due to natural selection, a range of antiviral agents may have evolved in marine organisms to protect them from viruses. The innate defences available to marine organisms can include secondary metabolites, bioactive peptides and proteins, thus serving as models for the development of new drugs for treating human disease. Indeed, several important antiviral compounds of marine origin have been reported, including didemnins (Caribbean tunicate, Trididemnum solidum), eudistomins (shallow-water tunicates within the genus Eudistoma), mycalamide A and B (New Zealand sponge, Mycale sp.), papuamides A (Theonella mirabilis and Theonella swinhoei sponges), avarone (Dysidea avara sponge), gymnochrome D (fossil crinoid Gymnocrinus richer), microspinosamide (Sidonops microspinosa sponge), solenolide A (gorgonian of the genus Solenopodium), hennoxazole A (Polyfibrospongia sp. sponge), thyrsiferol (red alga Laurencia venusta) and spongiadiol (deep-water Spongia sp.) (see reviews by Donia & Hamann, 2003; Dunlap et al., 2007). Other antiviral compounds such as vidarabine (adenine arabinoside), acyclovir and zidovudine (azidothymidine) have been commercially synthesized with semi-synthetic modifications from, or are structural analogues of, the arabinosyl nucleosides isolated from the sponge Cryptotethya crypta (Bergmann & Feeney, 1951; De Clercq, 2002).

Mollusca is the second largest animal phylum, with an estimated diversity of up to 200 000 extant species (Pechenik, 2000), occurring mainly in marine habitats. Like all invertebrates, molluscs only have innate immunity and do not acquire immunological protection against previous infection (Hooper et al., 2007b). Their enormous success implies compensation for the lack of adaptive immunity with effective innate defences, including humoral antiviral agents. Molluscs thus represent a great resource for discovery of antiviral compounds. Although over 900 molluscan secondary metabolites have been described (Benkendorff, 2010), few have been tested for antiviral activity. Nevertheless, antiviral activity has been reported in some molluscs, including unidentified bioactive macromolecules in the abalone Haliotis rufescens, oyster Crassostrea virginica, clams Mercenaria mercenaria and Mya arenaria, queen conch Strombus gigas, squid Loligo pealii and the sea snail Tegula gallina (Li, 1960; Li et al., 1962a, b, 1965; Prescott et al., 1964; Marderosian, 1969). Many of these molluscs are currently used to provide traditional or alternative medicines (Benkendorff, 2010).

Abalone are a major economic species in many countries (Fleming & Hone, 1996; Cook, 1998; Godoy & Jerez, 1998; Gordon & Cook, 2001, 2004). An antiviral fraction (designated paolin 2) from canned abalone was shown to inhibit replication of polyomavirus, influenza A virus and poliovirus in vitro and protected mice from infection with poliovirus and influenza virus (Li et al., 1962a). However, the location of these antiviral compounds within the abalone and their cytotoxicity, mode of action and chemical nature remain obscure. The need to understand abalone antiviral defence has become urgent, with epidemics of herpesvirus infection seriously affecting Haliotis diversicolor supertexta in Taiwan and blacklip Haliotis rubra and greenlip Haliotis laevigata in Australia, causing high rates of mortality up to 95 % within 14 days of the onset of clinical signs (Chang et al., 2005; Hooper et al., 2007a). On histopathological examination, infected abalone have severe nervous system damage, with ganglioneuritis and haemocytic infiltration (Chang et al., 2005; Hooper et al., 2007a). Electron microscopic examination of ganglia of affected abalone (Tan et al., 2008) revealed an icosahedral virus of about 150 nm in diameter, with an electron-dense core, that is, with the distinctive symmetry and appearance of a herpesvirus, and which is similar to the virus described from the oyster Crassostrea virginica (Farley et al., 1972). Initial characterization of the genome of oyster herpesvirus (named ostreid herpesvirus 1) and comparison of the gene sequence and morphological structure of herpesviruses of mammals and birds, fish and amphibians, and invertebrates including molluscs (e.g. oyster, clam) support the view that three major lineages of herpesviruses have evolved from a common ancestor (Davison, 2002; Renault, 2008). The abalone herpesvirus (AbHV) genome has been sequenced (Fegan et al., 2009; Savin et al., 2010) and a 59 682 bp portion of the sequence, deposited in GenBank (accession no. HM631982.1), shows sequence similarity to ostreid herpesvirus 1. The complete genome is estimated to be in the range of 200–210 kb in length (S. Warner, personal communication). Phylogenetic analysis of DNA polymerase proteins suggests that the abalone and oyster herpesviruses are within one family, Malacoherpesviridae, and are distantly related to other members of the family Herpesviridae (Fegan et al., 2009).

The lack of a suitable molluscan cell line presents an obstacle to culturing AbHV. Therefore, in view of the innate immunity in abalone, we chose to investigate their activity against herpes simplex virus type 1 (HSV-1), which is readily measured using a plaque assay (Russell, 1962). This technique has been used to measure the antiviral effect of various natural products including peptides secreted by the African clawed frog (Xenopus laevis) (Egal et al., 1999) and recently for detecting anti-HSV activity in oyster (Crassostrea gigas) haemolymph (Olicard et al., 2005a). We report here that abalone haemolymph and lipophilic extract of the abalone digestive gland contain significant activity against HSV-1.

RESULTS

Cytotoxicity assays were carried out to determine the concentration range of abalone haemolymph and of peptide, lipophilic and non-lipophilic extracts of different tissues from adult greenlip abalone (H. laevigata) for antiviral screening against HSV-1 in the non-toxic range for Vero cells. We considered an acceptable level of cytotoxicity to be when the abalone extract caused less than 10 % of cell death (CD₁₀), as determined by a trypan blue exclusion assay. DMSO (0.5 % in distilled water) caused no cell death and no change in cell morphology, and thus was used for dissolving the lipophilic extract before the antiviral assay. The maximal concentrations of crude haemolymph and non-lipophilic and lipophilic extracts from the digestive glands of abalone that were not cytotoxic were 20 % (v/v) and 2000 and 3000 μg ml⁻¹, respectively (Table 1⇓). Other abalone extracts appeared to be non-cytotoxic within the examined concentration range (CD₁₀ >1000 μg ml⁻¹ for all peptide extracts, >4000 μg ml⁻¹ for lipophilic and non-lipophilic extracts; Table 1⇓).

Within the non-cytotoxic range, antiviral activity was detected in haemolymph and lipophilic extract of the digestive glands of abalone using a plaque reduction assay (Table 1⇑). A series of twofold dilutions of haemolymph and lipophilic extract of the digestive gland was examined further for antiviral activity. Eagle's minimum essential medium (EMEM) and 0.5 % DMSO (in EMEM) were used as negative controls for the antiviral assay of haemolymph and lipophilic extract of the digestive gland, respectively. As there was no significant difference in the number of plaques in the negative-control plates [t=0.52, degrees of freedom (d.f.)=6, P=0.624], DMSO (0.5 %, v/v) did not interfere with HSV-1 plaque formation. Crude haemolymph (20 %, v/v) and lipophilic extract of the digestive gland (3000 μg ml⁻¹) induced up to approximately 96 and 100 % protection of Vero cells from HSV-1 infection, respectively (Fig. 1⇓). The positive control, acyclovir, induced 50 % protection of Vero cells from HSV-1 infection at 0.48 μg ml⁻¹ (Table 1⇑), and at 2 μg ml⁻¹ totally blocked HSV-1 replication (Fig. 1⇓).

Acyclovir (2 μg ml⁻¹), abalone haemolymph (20 %, v/v) and lipophilic extract of the digestive gland (3000 μg ml⁻¹) were tested for antiviral activity. There was a reduction in the number of plaques (>82 %) and plaque size (>32 %) in all test samples in comparison with relevant controls (EMEM for acyclovir and haemolymph; 0.5 % DMSO in EMEM for the lipophilic extract of the digestive gland; Table 2⇓). Significant decreases in plaque numbers were confirmed for the positive control, acyclovir (t=15.99, d.f.=6, P<0.0001), abalone haemolymph (t=19.97, d.f.=3.78, P<0.0001) and lipophilic extract of the digestive gland (t=15.91, d.f.=6, P<0.0001). Similarly, a reduction in plaque size was detected for acyclovir (t=5.75, d.f.=14.57, P<0.0001), abalone haemolymph (t=2.17, d.f.=15, P=0.047) and lipophilic extract of the digestive gland (t=14.38, d.f.=10, P<0.0001).

By adding haemolymph or lipophilic extract of the digestive gland at differing times (before, during and after infection), and varying the length of time the abalone substances were in contact with the virus/cell culture, we tried to address the question of when during infection these abalone substances exert their antiviral activity (Fig. 2⇓). The five different protocols with differing times of contact of the abalone substances with virus/cells are listed in Methods. Haemolymph acted at around the time of viral entry, whilst lipophilic extract of the digestive gland blocked viral activity subsequent to entry (Fig. 3⇓). The antiviral activity of haemolymph and lipophilic extract at different concentrations depended on the time during the assay at which these substances were present [two-way analysis of variance (ANOVA) interaction between time and concentration of haemolymph, F=21.69, d.f.=16, P<0.0001; and of lipophilic extract of the digestive gland, F=36.56, d.f.=16, P<0.0001]. Pre-incubation of haemolymph with cells for 2 h prior to infection, with removal of haemolymph immediately before the addition of virus to cells (protocol i) did not reduce the plaque number or size. Likewise, addition of haemolymph to the assay 1 h after virus was added to the cells (protocol v) did not reduce the plaque size or number (Tukey's HSD test, P>0.5). However, the presence of haemolymph (at concentrations of 20 %, v/v) during the time of virus entry into cells, whether pre-incubated with virus for 1 h before viral infection or added simultaneously with virus (protocols ii and iii), in each case left in the culture for the duration of the assay, or present during the first hour of virus/Vero cell incubation (protocol iv), gave rise to significant antiviral activity (Fig. 3a⇓, P<0.05).

Additional assays were carried out to determine whether haemolymph affected virus binding or entry, and whether it had a direct virucidal effect. Because lipophilic extract was not active during the first hour of viral infection, it was tested for virucidal effect but not in the binding/entry assay. The assay for virus attachment showed that abalone haemolymph at the maximum active concentration of 20 % (v/v) reduced the attachment of HSV-1 to Vero cells (73.9 % reduction in plaque formation). The assay for virus entry showed that haemolymph did not significantly interfere with viral entry (14.6 % reduction in plaque numbers). Haemolymph did not have a virucidal effect (<1 % reduction in plaque numbers).

In contrast to haemolymph, lipophilic extract of the digestive gland at 3000 μg ml⁻¹ had significant antiviral activity (Dunnett's T3 test, P<0.05) when used in infection protocols (ii), (iii) and (v), in which it was in contact with virus and cells from 1 h after infection (Fig. 3b⇑). However, as seen in Fig. 3(b)⇑, this abalone lipophilic extract had insignificant antiviral activity when present during the first hour of virus/Vero cell incubation (P=0.67). The virucidal assay showed that the lipophilic extract at the maximum active concentration of 3000 μg ml⁻¹ had a virucidal effect, reducing plaque numbers by 37.4 % in comparison with the 0.5 % DMSO control.

Experiments were carried out to characterize the antiviral substances within the extracts. Three haemolymph plasma fractions were obtained by centrifuging crude haemolymph at 3000, 28 000 and 60 000 r.p.m. Crude haemolymph was treated with heat (120 °C, 30 min), or with proteinase K and/or trypsin. There was no difference in antiviral activity against HSV-1 between crude haemolymph and each of the treated haemolymphs (one-way ANOVA, F=0.377, d.f.=7, P=0.91). Treatment of haemolymph with proteinase K or trypsin remarkably reduced protein content (OD₅₉₅ reduced from 2.49 to 0.85 and 1.64, respectively) (Table 3⇓). Gel electrophoresis confirmed that, after overnight incubation of abalone haemolymph or BSA with proteinase K and/or trypsin, degradation of protein occurred, producing protein fragments smaller than 5 kDa (data not shown). Nevertheless, the antiviral activity of the haemolymph was not affected by digestion with proteinase K, trypsin or combined protein digestion. Inactivated proteinase K and trypsin did not interfere with HSV-1 plaque formation in controls (Table 3⇓). The protein concentration in haemolymph samples was not significantly correlated with antiviral activity (Pearson's correlation coefficient, r²=0.44, P=0.07).

Separation of abalone haemolymph by hydrophobic interaction chromatography revealed that antiviral activity was mostly in the fraction that passed straight through the dianion column (Table 4⇓). No antiviral activity was detected in the water and subsequent lipophilic fractions extracted from the column using water, methanol and/or acetone (Table 4⇓).

DISCUSSION

This study demonstrated that the abalone H. laevigata has activity against HSV-1 in its crude haemolymph (EC₅₀=6.23 %, v/v) and lipophilic extract of the digestive gland (EC₅₀=667 μg ml⁻¹), but not in other tissues. This expands on previous studies that found activity in H. rufescens against polyomavirus, influenza A virus and poliovirus (Li, 1960; Li et al., 1962a). Consequently, abalone provide a source of antiviral compounds. Further experiments will be required to determine whether activity against HSV-1 provides a suitable model for screening abalone for potential resistance to abalone viral ganglioneuritis. Studies are under way to test abalone extracts against marine herpesviruses such as koi herpesvirus and salmonid herpesvirus types 1 and 2.

A trypan blue assay revealed that incubation of haemolymph or the lipophilic or non-lipophilic extract of the digestive gland of abalone at high concentrations for 2 days could cause Vero cell death. As this cytotoxicity could interfere with HSV-1 replication, we used a maximum test concentration of 20 % (v/v) for haemolymph, and 3000 and 2000 μg ml⁻¹ for the lipophilic and non-lipophilic extract of the digestive gland, respectively, for the antiviral assays. Within the non-cytotoxic range, antiviral activity was detected in the haemolymph and lipophilic extract of the digestive gland of greenlip abalone in a concentration-dependent manner. Indeed, increasing the concentration of crude haemolymph to 20 % (v/v) and the lipophilic extract of the digestive gland to 3000 μg ml⁻¹ in the antiviral assay resulted in a reduction in the number of HSV-1 plaques by about 98 and 95 %, respectively. Similar activity against HSV-1 is also found in Crassostrea gigas haemolymph (protein concentration 600 μg ml⁻¹; Olicard et al., 2005a) and a protein extract, the L₄₀ fraction, of periwinkle Littorina littorea haemolymph (protein concentration of 530 μg ml⁻¹; Defer et al., 2009). However, at these concentrations, Crassostrea gigas haemolymph and the L₄₀ fraction of Littorina littorea haemolymph cause much higher cytotoxicity (∼20 and 34 %, respectively) than H. laevigata haemolymph and lipophilic extract of the digestive gland (∼10 %).

To enter a cell, HSV-1 binds to cellular glycosaminoglycan chains (WuDunn & Spear, 1989) such as heparin sulfate (Lindahl et al., 1994) or chondroitin sulfate (Banfield et al., 1995; Stringer & Gallagher, 1997; Mårdberg et al., 2002) via the viral glycoproteins gC and gB (Tal-Singer et al., 1995; Qie et al., 1999). HSV-1 entry is mediated by fusion of the viral envelope with the cell membrane, involving viral glycoproteins gB, gD, gH and gL and specific cellular receptors (Montgomery et al., 1996; Spear, 2004; Browne, 2009). Fusion of HSV-1 with Vero cells is a relatively fast process, mostly taking place within the first 10–15 min of virus/cell incubation at 37 °C (Koyama & Uchida, 1987; Andersen et al., 2004). Thus, different times of addition of abalone extracts, relative to the time of addition of virus to cells, were used here to identify the stage of infection at which their antiviral activity was exerted. The antiviral activity of haemolymph was significant if it was present during the first hour of virus/cell incubation (Fig. 3a⇑), suggesting that its action may be due to inhibition of an early stage of infection. The attachment and entry assay results are consistent with haemolymph reducing HSV-1 binding to Vero cells, although at this point we cannot absolutely preclude the possibility that a haemolymph component is internalized simultaneously with the virus to exert its effect on a post-entry event in infection. Antiviral action resembling that described here has been found in the milk protein lactoferrin, its pepsin cleavage product lactoferricin and a number of other highly cationic α-helical peptides, which mediate their activity by binding heparin sulfate on the cellular surface and blocking attachment or entry of HSV-1 (Marchetti et al., 1996; Andersen et al., 2002, 2004; Jenssen et al., 2004a, b; Jenssen, 2005). Prevention of viral entry could also be due to sugar or other compounds with a high affinity for viral attachment/entry glycoproteins competing with cellular receptors, for example chondroitin sulfate type E derived from squid cartilage (Bergefall et al., 2005).

In contrast, the antiviral activity of lipophilic extract of the digestive gland was most notable when added 1 h after viral infection. This suggested that the extract may be acting at a stage of infection subsequent to cellular entry (e.g. blocking the uncoating of viral genomes or the trafficking of virion components). Recently, bovine lactoferrin, in addition to inhibiting virus attachment and/or entry, has been shown to inhibit cell-to-cell virus spread, possibly by interacting with the structural viral proteins ICP5 (major capsid protein) and VP16 (viral tegument protein) (Jenssen et al., 2008). Lactoferrin and its cleavage product lactoferricin also interfere with trafficking of HSV-1 capsids along microtubules towards the nucleus (Marr et al., 2009). As components of innate immunity, it is possible that the mechanism of antiviral activity in the lipophilic extract of the abalone digestive gland resembles that of lactoferrin. In contrast to lactoferrin, the lipophilic extract showed no effect on viral attachment or entry but had a virucidal effect when incubated with free virus. The antiviral compound in the lipophilic extract could be internalized simultaneously with the virus to exert its effect at an early stage of post-entry events such as transport, capsid uncoating or transcription. Similar to abalone haemolymph, the lipophilic extract did not give rise to an antiviral state in Vero cells. We were unable to exclude the possibility of the lipophilic extract having an antiviral effect later in infection.

The digestive gland is the largest organ in gastropods and is a relatively complex organ, linked by ducts to the stomach. It is a site for intracellular and extracellular digestion, absorption of digestive products, excretion and storage of reserves (Kay et al., 1998). A range of enzymes is excreted by the digestive gland, including lipases, proteases and carbohydrases. However, these enzymes are not lipophilic and thus are unlikely to be responsible for the antiviral activity observed in the digestive gland. In contrast, there is potential for dietary-derived secondary metabolites to contribute activity in the digestive gland. Grazing gastropods are well known to acquire bioactive compounds from their algal diets for use in their own defence (e.g. Faulkner, 1984, 1992). Antiviral activity (or antiviral compounds) has also been reported previously from a range of red and green algae (Ivanova et al., 1994; Hayashi et al., 1996a, b; Ohta et al., 1998; Smit, 2004; Park et al., 2005), which are consumed by abalone. However, it should be noted that the abalone used in this study were sourced from an aquaculture farm and fed an artificial pellet diet, with no known antiviral properties. Thus, it seems likely that the antiviral factors in the digestive gland are intrinsic to abalone.

A range of antiviral proteins and peptides are found in many marine species, such as littorein produced by the common periwinkle Littorina littorea (Defer et al., 2009), defensins produced by the Mediterranean mussel Mytilus galloprovincialis (Roch et al., 2004) and haemocycanin produced by the shrimp Penaeus monodon (Zhang et al., 2004). However, we have shown that anti-HSV-1 activity was retained after most protein in the abalone haemolymph was destroyed by heat and proteinase treatment. In addition, no correlation existed between antiviral activity and the protein concentration of crude haemolymph, three different plasma fractions (I, II and III), and heat- or proteinase-treated haemolymph samples. A possible interpretation is that small heat-resistant peptides, remaining after proteinase K and trypsin digestion, rather than large peptides, are responsible for the abalone anti-HSV-1 activity. Indeed, small peptides including lactoferricin (25–49 aa), defensins (29–33 aa), indolicidin (13 aa), brevenin-1 (24 aa), protegrins (18 aa), tachyplesin-1 (17 aa), melittin (26 aa), clavalin (23 aa) and megainin (23 aa) have anti-HSV activity (Yasin et al., 2000; Andersen et al., 2003; Albiol Matanic & Castilla, 2004; Marr et al., 2009). Separation of haemolymph on a hydrophobic dianion resin column further revealed that antiviral activity was not due to lipophilic active compounds (e.g. fatty acids and aliphatic or aromatic alkaloids), but was more probably due to sugars, acids or small polar peptides/proteins, which would pass through the column or elute in the water fraction (Einbond et al., 2004). Further investigation is needed to elucidate the chemical structure of the antiviral compounds and the details of their mechanism of action.

In summary, two types of antiviral activity were detected in abalone (H. laevigata), with different modes of action against HSV-1. Many antiviral compounds produced by marine invertebrates are part of their innate defence against viruses (Pan et al., 2000; Zhang et al., 2004); however, it remains to be determined whether the compounds with activity against HSV-1 in this study could also be effective against AbHV and other viruses. As phlebotomy of abalone is quick and non-lethal, their haemolymph can also be used in future studies aimed at understanding molluscan defence systems.

METHODS