Varroa destructor virus 1: a new picorna-like virus in Varroa mites as well as honey bees

Varroa destructor mite is an ectoparasite of the honey bee Apis mellifera. This species was recently differentiated from Varroa jacobsoni species which infests the Asian bee Apis cerana. Varroa mites feed entirely on the bee's haemolymph and have been associated with the spread of a number of viruses. Since the mites were first observed in Java, Indonesia in 1904, they have been reported in most regions of the world except Australia and the equatorial regions of Africa. V. destructor severely affects and threatens the survival of A. mellifera. The mite was spread to ?. mellifera colonies in other areas by migratory practices of bee keepers and by drifting swarms. The survival of bee colonies attacked by these mites is threatened since the mites weaken the colony if left untreated. Honey bees are important pollinators in nature and in the agricultural and horticultural industries. Bee keepers can also earn from the sale of honey, bee wax and propolis. At this moment, there is no absolute method available for controlling the mite. The methods currently employed use chemicals which contaminate honey and other bee products, and mites are developing resistance to some of the chemicals used. The economic impact of mite infestation makes it necessary to investigate alternative options for control.At the onset of the research described in this thesis, 27 nm picorna-like virus particles were observed in mite tissue apparently going through an infection cycle. The identity of this virus was unknown and it was unclear if it was infectious to the mite only or to the bee as well. The aim of this research was to isolate this virus from the mite, characterise its genome in detail and study its behaviour in mite and bee populations with the intention to determine its potential as a biological control agent against the Varroa mite.Electron microscopic examinations showed para-crystaüine aggregates of virus particles in the cytoplasm of mite tissue. The virus particles were purified from mites collected from the Wageningen University apiary and used to raise rabbit polyclonal antibodies. The antibody was applied to locate the virus in tissue sections of mites in an immunohistology examination which revealed that the virus was abundantly located in the tissues of the lower digestive tract (Chapter 2). The virus was not detected in salivary glands, indicating that this virus is not transmitted via these glands.In the next phase RNA was isolated from these virus particles and the viral genome was fully sequenced (Chapter 3). The virus has a single-stranded, positive-sense genome which is polyadenylated at the 3' terminus and can serve directly as messenger RNA. The genome has a length of 10,112 nucleotides (without the poly-A tail) and one large open reading frame (ORF) encoding a polyprotein of 2893 amino acids. The structural proteins are located in the N-terminal half and the non-structural proteins on the C-termina! half of the polyproiein, and are produced by autoproteolytic cleavage. The ORF is flanked on either side by nontranslated regions (NTRs). Phylogenetic analysis of its RNA-dependent RNA polymerase in a comparison with those of related viruses in the GenBank database revealed that this virus shows high sequence similarity to members of the genus Iflavirus. The genome organisation is also highly comparable to iflaviruses and clearly distinct from that of dicistroviruses. Since the genome sequence of this virus had not been previously reported, the virus was named Varroa destructor virus 1 (VDV-I) after the mite from which it was first isolated. VDV-I is most closely related to Deformed wing virus (DWV) which was isolated from honey bees with wing abnormalities. VDV-I and DWV have 84% genome and 95% polyprotein identity.To confirm that VDV-I is able to replicate in the mite, primers were designed to detect specifically the negative-sense RNA strand, which only occurs as replication intermediate, and hence provides evidence for viral replication. With these primers cDNA was synthesised and was further amplified by PCR. Two sets of specific primers were made to distinguish between and detect either VDV-I or DWV. In a similar way, primers, specific for the positive-sense strand, were used to detect the genomes of both VDV-I and DWV in the mites. These experiments showed that both VDV-I and DWV were replicating in the Varrao mite (Chapter 4). These findings are in good agreement with the observation of para-crystalline aggregates in electron microscopy images, which also supports the replication of both viruses in the cytoplasm of mite cells (Chapter 2).In Chapter 5, the structural proteins of the isolated virus were examined by SDS-PAGE and it was observed that the two largest proteins (VPl and VP2) were present in relatively equal amounts In the virus. These proteins were N-terminally sequenced to reveal the amino acids that determined the proteolytic cleavage sites. VP2 was mapped directly N-terminal to the non-structural helicase, while VPl was located immediately upstream of VP2. Through Western blot analysis it was demonstrated that only VPl reacted very strongly to the antiserum prepared against the virus. In a next experiment the structural proteins were expressed individually fused to glutathione S-transferase. The resulting fusion proteins were tested in Western blot analysis using the antiserum against purified virus. This study revealed that VP1 was the only structural protein which was recognised, implying that this protein probably covered the entire surface of the virus particle, hiding the other structural proteins beneath, or that the other structural proteins were not immunogenic. The viral 3C-like protease was also expressed as a fusion protein and used to raise antibodies, which efficiently detected the protease polypeptide in a control experiment. This antiserum might be used to detect viral replication using an antibody-based method such as ELISA, or to localise replication in the mite body using histochemical methods in addition to molecular techniques.So far there is no report on the structure and function of the 5' nontranslated region (5' NTR) of the RNA of iflaviruses. One of the aims of this investigation was to predict whether conserved secondary structures occurred in the 5' NTR of four iflaviruses (Chapter 6). The predictions revealed two types of structures. VDV-I and DWV have long 5' NTRs with complex structures that resemble those of enteroviruses (Picornavtridae), particularly that of Poliovirus. Perina nuda picorna-Iike virus (PnPV) and Ectropis obliqua picorna-iike virus (EoPV) have shorter 5' NTRs with simpler structures that are unique and do not resemble any 5' NTR structures among picornaviruses. The translation of the ORF of genome organisation is also highly comparable to iflaviruses and clearly distinct from that of dicistroviruses. Since the genome sequence of this virus had not been previously reported, the virus was named Varroa destructor virus 1 (VDV-I) after the mite from which it was first isolated. VDV-I is most closely related to Deformed wing virus (DWV) which was isolated from honey bees with wing abnormalities. VDV-I and DWV have 84% genome and 95% polyprotein identity.To confirm that VDV-I is able to replicate in the mite, primers were designed to detect specifically the negative-sense RNA strand, which only occurs as replication intermediate, and hence provides evidence for viral replication. With these primers cDNA was synthesised and was further amplified by PCR. Two sets of specific primers were made to distinguish between and detect either VDV-I or DWV. In a similar way, primers, specific for the positive-sense strand, were used to detect the genomes of both VDV-I and DWV in the mites. These experiments showed that both VDV-I and DWV were replicating in the Varrao mite (Chapter 4). These findings are in good agreement with the observation of para-crystalline aggregates in electron microscopy images, which also supports the replication of both viruses in the cytoplasm of mite cells (Chapter 2).In Chapter 5, the structural proteins of the isolated virus were examined by SDS-PAGE and it was observed that the two largest proteins (VPl and VP2) were present in relatively equal amounts In the virus. These proteins were N-terminally sequenced to reveal the amino acids that determined the proteolytic cleavage sites. VP2 was mapped directly N-terminal to the non-structural helicase, while VPl was located immediately upstream of VP2. Through Western blot analysis it was demonstrated that only VPl reacted very strongly to the antiserum prepared against the virus. In a next experiment the structural proteins were expressed individually fused to glutathione S-transferase. The resulting fusion proteins were tested in Western blot analysis using the antiserum against purified virus. This study revealed that VP1 was the only structural protein which was recognised, implying that this protein probably covered the entire surface of the virus particle, hiding the other structural proteins beneath, or that the other structural proteins were not immunogenic. The viral 3C-like protease was also expressed as a fusion protein and used to raise antibodies, which efficiently detected the protease polypeptide in a control experiment. This antiserum might be used to detect viral replication using an antibody-based method such as ELISA, or to localise replication in the mite body using histochemical methods in addition to molecular techniques.So far there is no report on the structure and function of the 5' nontranslated region (5' NTR) of the RNA of iflaviruses. One of the aims of this investigation was to predict whether conserved secondary structures occurred in the 5' NTR of four iflaviruses (Chapter 6). The predictions revealed two types of structures. VDV-I and DWV have long 5' NTRs with complex structures that resemble those of enteroviruses (Picornavtridae), particularly that of Poliovirus. Perina nuda picorna-Iike virus (PnPV) and Ectropis obliqua picorna-iike virus (EoPV) have shorter 5' NTRs with simpler structures that are unique and do not resemble any 5' NTR structures among picornaviruses. The translation of the ORF of picorna(-like) viruses in general is initiated by the recognition of one or two internal ribosome entry sites (IRES). The cap-dependent mechanism of translation initiation employed by most cellular mRNAs is not used by these viruses. The IRES activity in the 5' NTRs of iflaviruses had not yet been determined experimentally. Therefore, IRES activity in the 5' NTR of VDV-I was examined. This research was also aimed at identifying a permissive cell line with the ability to support VDV-I IRES function, which is a prerequisite for the translation of the polyprotein ORF, and which might potentially support viral replication. The activity of the IRES of VDV-I was investigated by cloning the 5' NTR between two reporter genes: enhanced green fluorescent protein (EGFP) and firefly luciferase (Fluc). EGFP was cloned directly downstream of the OpIE2 promoter, which is activated by cellular factors, and is translated via a cap-dependent mechanism. The translation of Fluc was dependent on IRES activity in the 5'NTR of VDV-I in the respective cell lines. The presence of the 5' NTR of VDV-I greatly improved the expression levels of the second reporter gene (Fluc) in Lymantria dispar Ld652Y cells, showing that the 5' NTR of VDV-I contains a functional IRES element. This IRES element was active in a host specific manner since it showed much lower activity in Spodoptera frugiperda Sf21 cells and no activity in Drosophila melanogaster S2 cells.The transmission ofVDV-1 between the mite V. destructor and the honey bee A. mellifera was surveyed in comparison with DWV (Chapter 7). To determine the spread of these viruses in mites (adults, nymphs and eggs) and bees (eggs, larvae, pupae and adults), the antiserum raised against purified virus was used in ELISA analyses. Infections by VDV-1 or DWV could not be distinguished using this immunology technique due to the high similarity (97%) in the immunodominant protein VPl. The proportion of Varroa mites (71%) infected with VDV-I and/or DWV at the Wageningen University apiary was slightly higher than that of bees (65%). Vertical transmission to the next generation was established in the mite population since virus was detected in mite eggs using a dot-blot immunoassay. Virus could not be detected in bee eggs, implying that no vertical transmission occurred in this species. Subsequently, nested RT-PCR was used to distinguish VDV-I from DWV and with this technique both viruses could be detected even in the same individual. Eighty eight percent of the mites had VDV-I, 19% of the mites were co-infected with VDV-I and DWV, and no mites with only DWV infection were detected. Seventy nine percent of the adult bees had VDV-I, 26% of the adult bees were co-infected with VDV-I and DWV and there was no DWV only infection in the adult bees either, in the specimens tested. The conclusion from this experiment is that VDV-I and DWV are able to co-exist in an individual mite or bee and can replicate in both organisms. A limited survey of these viruses among mites from different regions of mainland Europe indicated that the two viruses exist together in hives across this part of the world.The research described in this thesis compared two closely related viruses able to infect both the mite and bee. One of these viruses (VDV-I) is new. In this study the pathogenicity ofVDV-1 in mites and bees was not analysed in detail, but so far clear symptoms have not been found in the mite or the bee that could be attributed to VDV-I infection. In these studies, DWV also did not result in clear symptoms in mites and bees. Due to the results obtained, VDV-1 is not a prime candidate for biological control of the V. destructor mite.