The molecular basis of the interactions between luteoviruses and their aphid vectors

Luteoviruses essentially replicate in the phloem tissue and are transmitted from plant to plant by aphids in a circulative, persistent manner. Virus particles are acquired when aphids feed on phloem sap. Particles are then transported from the midgut or hindgut into the haemolymph and from the haemolymph to the salivary gland, to be eventually released with the saliva to the phloem of uninfected plants. There is no evidence that luteoviruses replicate in the aphid vector. The haemolymph acts as a reservoir in which luteoviruses should persist in an infecting form during the whole lifespan of aphids.A virus overlay technique was developed for the characterization of aphid-derived proteins involved in the circulative transmission of luteoviruses by aphids (Chapter 2). Proteins from whole-body homogenates of the aphid species Myzus persicae were separated with a two-dimensional denaturing poly-acrylamide gel (SDS-PAGE) and transferred to nitrocellulose membranes. Subsequently, these membranes were incubated with purified Potato leafroll virus (PLRV; genus Polerovirus ; Family Luteoviridae ) particles. Bound virus particles were detected by incubating membranes with anti-PLRV IgG and phosphatase conjugated goat anti-rabbit IgG. Thus it was demonstrated that PLRV particles bind to five different proteins. A protein of 63 kilodalton (p63) had the highest affinity for PLRV particles and was characterized by N-terminal amino-acid sequencing and immuno-gold labeling studies. These studies revealed that this protein is a homologue of GroEL and is abundantly synthesized by the primary bacterial endosymbiont ( Buchnera sp.) of M. persicae .To show whether PLRV particles and Buchnera GroEL also interact in vivo , aphids were fed on diets containing tetracyclin (Chapter 2). This antibiotic acts as bacteriostatic by inhibiting protein synthesis. After a tetracyclin treatment, Buchnera GroEL was not detected in the haemolymph of the aphid, virus transmission was reduced by more than 70%, and the major viral capsid protein was degraded. These observations led to the suggestion that Buchnera GroEL is involved in protection of virus particles against proteolytic breakdown during circulation in the haemolymph.To study the interaction of PLRV and Buchnera GroEL of M. persicae (MpB GroEL) in more detail, the gene encoding MpB GroEL and its flanking sequences were characterized and compared to those of Escherichia coli and Buchnera spp. of other aphid species (Chapter 3). The MpB GroEL encoding gene appeared to be part of an operon with a similar organization as the groE operon of E. coli , containing another gene for a 10-kDa protein with sequence similarities to GroES of E. coli . However, a constitutive promoter sequence comparable to that of the E. coligroE operon could not be identified; only sequences comparable to the heat shock promoter of the E. coli groE operon were observed. Comparison of the deduced amino-acid sequences disclosed that MpB GroEL is approximately 98% similar to GroELs of other Buchnera spp. and 92% similar to E. coli GroEL. These results demonstrate that MpB GroEL belongs to the family 60-kDa chaperonin or heat shock protein family.Several functions of GroEL proteins have been described and the most important one is the folding of nonnative proteins inside the cytosol of prokaryotes, mitochondria and chloroplasts. MpB GroEL and other GroEL proteins have typical double-doughnut structures composed of two stacked rings of seven subunits each. Using the crystal structure of E. coli GroEL, computer-generated structural predictions of the monomer of MpB GroEL was obtained (Chapter 3). Like E. coli GroEL, each subunit of MpB GroEL consists of an apical, an intermediate and an equatorial domain. The apical domain is a continuous domain on the primary MpB GroEL protein structure, whereas the equatorial and intermediate domains are discontinuous with regions located at the N- and C-terminus of the MpB GroEL subunit. The N- and C-terminal regions of the equatorial and intermediate domains assemble in the folded structure of MpB GroEL.Functional studies of E. coli GroEL 14-mers have demonstrated that the apical domains are located at both sides of the cylindrical double-doughnut structure and contains amino acids involved in binding of nonnative proteins. The equatorial domains form the waist of the GroEL 14-mer. Intermediate domains function as hinges for moving the apical domain up and down so that amino acids in the apical domain can bind the unfolded protein. Subsequently, unfolded proteins are kept in the cavity of the GroEL 14-mer where they obtain their native structure without being disturbed by cytosolic compounds.To investigate which of the domains of MpB GroEL are involved in binding PLRV particles, deletion mutants were designed based on the primary structure of the MpB GroEL protein (Chapter 3). Full-length MpB GroEL and MpB GroEL deletion mutants were expressed in fusion with glutathione-S-transferase (GST) in E. coli and affinity-purified. The GST moiety was removed and similar amounts of recombinant protein were tested for PLRV binding in virus overlay assays. This revealed that recombinant full-length MpB GroEL proteins had a similar affinity for PLRV particles as wild type MpB GroEL proteins isolated from M. persicae . PLRV particles displayed affinity for MpB GroEL deletion mutants only if they still contained the N- or C-terminal regions of the equatorial domain. Strikingly, PLRV-binding to polypeptides containing the apical domain alone or when extended with flanking sequences did not bind PLRV. Furthermore, virus overlay assays with additional MpB GroEL deletion mutants demonstrated that determinants for PLRV binding at the C-terminal part of the equatorial domain are located between residues 408 and 475 of MpB GroEL (Chapter 4). This region comprises threeα-helices.Since the N- and C-terminal regions of the equatorial domain assemble in the folded structure of MpB GroEL, the two PLRV-binding regions may become a single PLRV-binding site. The finding that the equatorial domain was involved in binding PLRV particles and not the apical domain is surprising, since studies of E. coli GroEL showed that the apical domain is involved in binding of unfolded proteins in the cytosol of E. coli cells. PLRV particles may have different binding characteristics because of the size limitation of the central cavity of the GroEL molecule and the fact that binding occurs extracellularly in the haemolymph.The interaction between PLRV particles and MpB GroEL was investigated in more detail (Chapter 4). Virus overlay studies with additional MpB GroEL deletion mutants revealed that regions between amino acid residues 1 and 57, and 427 and 457 of the N- and C-terminal regions of the equatorial domain, respectively, contain the determinants for PLRV binding. To determine which amino acids are involved in PLRV binding, overlapping decameric peptides of PLRV-binding regions were synthesized and incubated with virus particles in a virus overlay based experiment (Chapter 4). Alanine replacement studies of binding peptides showed that amino acids R13, K15, L17 and R18 of the N-terminal region of the equatorial domain, and R441 and R445 of the C-terminal region of the equatorial domain are responsible for PLRV binding. Alanine replacement of R13, K15, L17 and R18 eliminated PLRV binding of MpB GroEL(1-408) completely, whereas replacement of R441 and R445 reduced, but not eliminated, virus binding of MpB GroEL(122-548). This suggests that besides R441 and R445 other residues in the C-terminus are part of the PLRV-binding site.These still unknown residues are likely to be located in the region between amino acids 427 till 474, which comprises oneα-helix located to the outside of GroEL 14-mers. Residues R13, K15, L17 and R18 are located in a longα-helix that is present more internally of GroEL 14-mers. The N- and C-terminal amino acids are positioned behind each other in a cavity, which might be accessible for the readthrough domain (RTD) which protrudes from the surface of a luteovirus particle.The luteovirus protein capsid is composed of a major 23-kDa coat protein (CP), and lesser amounts of a ~54-kDa readthrough protein, expressed by translational readthrough of the CP into the adjacent open reading frame encoding the RTD. The RTD is exposed on the surface of the virus particle and contains the determinants necessary for virus transmission by aphids. To study whether the highly conserved major CP or the RTD of the minor 54-kDa protein are involved in GroEL binding, BWYV mutants devoid of the RTD were synthesized and tested for GroEL affinity in a GroEL-ligand assay (Chapter 5). It was found that the BWYV RTD mutants did not bind GroEL, indicating that the RTD contains the GroEL-binding determinants. BWYV mutants lacking the RTD domain were also injected into the haemolymph of aphids and the persistence of these mutants was compared with those of wild-type virus particles (Chapter 5). These studies clearly showed that BWYV mutants devoid of the RTD were more rapidly degraded than wild-type viruses, indicating that the RTD, containing the GroEL-binding sites, is crucial for the persistency in the aphid.To reveal whether conserved domains of the RTD are involved in GroEL binding, five luteoviruses belonging to the genus Polerovirus and Pea enation mosaic virus (PEMV; Enamovirus ) were tested for binding to Buchnera GroEL proteins isolated from several aphid species using GroEL-ligand assays (Chapter 5). All luteoviruses displayed a specific but differential affinity for the GroEL homologues isolated from the endosymbiotic bacteria of both vector and non-vector aphid species, and for E. coli GroEL. This indicates that GroEL is not involved in vector specificity. Sequence alignment of the RTDs of different luteoviruses and PEMV revealed that only the N-terminal half of the RTDs is conserved, whereas the C-terminal halves have no global sequence identity. This C-terminal region is also lacking from the PEMV RTD. The highest overall level of sequence similarity in the RTD extends from position 184 to 223 where about 23% of the residues are identical.To assess whether the viral determinants required for the interaction of luteoviruses with Buchnera GroEL reside in the conserved region of the RTD, GST-fusions of the RTD and mutants thereof were expressed in E. coli (Chapter 6). After affinity purification, the GST moiety was cleaved and the resulting RTD protein tested for MpB GroEL affinity using a GroEL-ligand assay. This showed that the conserved region of the RTD plays a crucial role in binding GroEL.The knowledge derived from the binding studies of GroEL and luteoviruses is valuable for the development of specific control methods. The fact that Buchnera GroEL and luteoviruses directly interact in vitro suggests that this occurs in the haemolymph of aphids as well. Consequently, peptides or antibodies that interfere in this interaction by binding to the equatorial domain of Buchnera GroEL or the RTD of luteoviruses reduce specifically the transmission efficiency of luteoviruses by aphids. It is possible to produce these interfering compounds by plants so that aphids acquire them while feeding. Further studies should reveal whether there are possibilities for transporting peptides or antibodies from the gut to the haemolymph.Chapter 7 of this thesis describes an investigation that may lead to an alternative control strategy. In this chapter the effects of neem ( Azadirachta indica A. Juss) seed kernel extracts (NSKE) and its major active compound, azadirachtin, on the ability of M. persicae to transmit PLRV is studied. This secondary plant metabolite has major effects on bacterial symbionts of leafhoppers. Since endosymbiotic bacteria play a major role in the performance of aphids and luteovirus transmission by aphids, it was investigated whether treatments with these compounds would exert an effect on aphid larval growth and mortality, and on the aphid intracellular symbionts. The neem metabolites displayed a 100% mortality at doses higher than 2560 ppm., and morphological aberrations on the bacterial endosymbionts were observed. At doses lower than 160 ppm of NSKE or azadirachtin, the endosymbiont population of M. persicae , and mortality, growth and feeding behavior was similar to that of the untreated groups of aphids. However, PLRV transmission was inhibited by 40-70%. These observations raise the possibility that interfering with the relationship between endosymbionts and aphids may contribute to the control of luteovirus transmission by aphids.