Received 29 October 2008/Accepted 18 March 2009

Transcription

1 JOURNAL OF VIROLOGY, June 2009, p Vol. 83, No X/09/$ doi: /jvi Copyright 2009, American Society for Microbiology. All Rights Reserved. Virion Stability Is Important for the Circulative Transmission of Tomato Yellow Leaf Curl Sardinia Virus by Bemisia tabaci, but Virion Access to Salivary Glands Does Not Guarantee Transmissibility Piero Caciagli, 1 Vicente Medina Piles, 2 Daniele Marian, 1 Manuela Vecchiati, 1 Vera Masenga, 1 Giovanna Mason, 1 Tania Falcioni, 2 and Emanuela Noris 1 * Istituto di Virologia Vegetale, Consiglio Nazionale delle Ricerche, Strada delle Cacce 73, Turin, Italy, 1 and Departament de Producció Vegetal i Ciència Forestal, Universitat de Lleida, Avda. A. Rovira Roure 191, Lleida, Spain 2 Received 29 October 2008/Accepted 18 March 2009 The capsid protein (CP) of the monopartite begomovirus Tomato yellow leaf curl Sardinia virus (TYLCSV), family Geminiviridae, is indispensable for plant infection and vector transmission. A region between amino acids 129 and 152 is critical for virion assembly and insect transmissibility. Two previously described mutants, one with a double Q129P Q134H mutation (PNHD) and another with a further D152E change (PNHE), were found nontransmissible (NT). Another NT mutant with a single N130D change (QDQD) was retrieved from a new mutational analysis. In this study, these three NT mutants and the wild-type (wt) virus were compared in their relationships with the whitefly vector Bemisia tabaci and the nonvector Trialeurodes vaporariorum. Retention kinetics of NT mutants were analyzed by quantitative dot blot hybridization in whiteflies fed on infected plants. The QDQD mutant, whose virions appeared nongeminate following purification, was hardly detectable in either whitefly species at any sampling time. The PNHD mutant was acquired and circulated in both whitefly species for up to 10 days, like the wt virus, while PNHE circulated in B. tabaci only. Using immunogold labeling, both PNHD and PNHE CPs were detected in B. tabaci salivary glands (SGs) like the wt virus, while no labeling was found in any whitefly tissue with the QDQD mutant. Significant inhibition of transmission of the wt virus was observed after prior feeding of the insects on plants infected with the PNHE mutant, but not on plants infected with the other mutants. Virion stability and ability to cross the SG barrier are necessary for TYLCSV transmission, but interactions with molecular components inside the SGs are also critical for transmissibility. Geminiviruses are transmitted in a circulative manner. Once ingested from infected tissues during feeding, virions enter the gut, cross the midgut/hemolymph barrier, and are transported through the hemolymph to all organs. After passing the hemolymph/salivary gland (SG) barrier, virions enter the SGs from which they are inoculated back into plants. Beside encapsidating the genome and transporting it in and out of the nucleus (44, 45, 49, 67, 68), the capsid protein (CP) of geminiviruses is the only viral protein mediating transmission (6, 11). The CP of the begomovirus Tomato yellow leaf curl Sardinia virus (TYLCSV), which has a monopartite genome (42), is also essential for systemic plant infection (71). As with mastre- and curtoviruses (9, 47, 51, 12), with no specific movement proteins, it is assumed that TYLCSV CP has typical movement protein properties. Its CP is also required for vector transmission and specificity (21). A region between amino acids (aa) 129 and 152, including Q129, Q134, and D152 is relevant for virion assembly, systemic infection, and transmission by the * Corresponding author. Mailing address: Istituto di Virologia Vegetale, Consiglio Nazionale delle Ricerche, Strada delle Cacce 73, Turin, Italy. Phone: Fax: e.noris@ivv.cnr.it. Supplemental material for this article may be found at Present address: Settore Fitosanitario-Regione Piemonte, Environment Park, Turin, Italy. Published ahead of print on 25 March vector, the whitefly Bemisia tabaci Gennadius (59). A single Q129P mutation prevents particle assembly and systemic infection (PQD and PQE mutants), while mutants with a further Q134H change (PHD and PHE mutants) are infectious but nontransmissible (NT) (59) (see Fig. 1A). The relevance of this region in transmission has been confirmed for Watermelon chlorotic stunt virus (41) and Abutilon mosaic virus (AbMV) (38). The interaction between B. tabaci and Tomato yellow leaf curl virus (TYLCV), another virus inducing the tomato yellow leaf curl disease, was reviewed recently (21). The pattern of association at the cellular level between B. tabaci, biotype B, and transmissible isolates of TYLCSV and between Asystasia golden mosaic virus and African cassava mosaic virus (ACMV) has been studied (27, 28, 55). Tomato mottle virus and Cabbage leaf curl virus were immunolocalized to the filter chamber and the anterior part of the midgut (39), while TYLCV CP was detected in the descending midgut lumen, close to the wall and to the rich lining of microvilli (28). TYLCSV was found inside epithelial cells of the descending midgut, while an NT isolate of ACMV was detected only in the gut lumen (55). B. tabaci has paired primary salivary glands (PSGs) and accessory salivary glands (ASGs) (35, 36), whose ultrastructure has been studied by electron microscopy (EM), identifying 13 large cells that empty into one duct lined with microvilli in each PSG and four large symmetric cells in each ASG (30). TYLCV CP was found in PSGs by immunofluorescence (14), and viral 5784

2 VOL. 83, 2009 RELATIONSHIPS OF TYLCSV MUTANTS WITH THE INSECT VECTOR 5785 DNA was detected there using in situ amplification (28). In line with this, TYLCSV CP was found by immunogold labeling in PSGs, mainly in saliva drops (55), indicating that PSGs have a central role in transmission (18, 27, 28). In this study, the biological properties of new TYLCSV CP mutants are described. Four new mutants with changes in the region from aa 129 to 134 were characterized for infectivity and transmissibility and a mutant with a single N130D substitution (mutant QDQD) was found to be nontransmissible. The NT CP mutants, including the previously described PHD and PHE mutants (59) (here renamed PNHD and PNHE mutants, respectively) and the QDQD mutant, were compared with the wild-type (wt) virus in their relationships to B. tabaci. Quantitative dot blot hybridization and immunogold labeling were used to follow their acquisition and retention in whiteflies and to localize their CP in the insect. Finally, we used competitive transmission experiments to see whether NT mutants interfered with transmission of the wt virus. MATERIALS AND METHODS Virus clones and their maintenance. The NT mutants with Q129P Q134H and Q129P Q134H D152E mutations (here renamed PNHD and PNHE, respectively) have been described previously (59), as well as their agroinfectious clones, pbinsarphd and pbinsarphe. N130D and Q129A mutants, both with either Q or H at position 134, were obtained with the QuikChange site-directed mutagenesis kit (Stratagene, CA), according to the manufacturer s instructions. Primers SARMUT682( ) (5 -CATAAAAAAACAAGATCATACTAACC- 3 ) and SARMUT707t( ) (5 -GGTTAGTATGATCTTGTTTTTTTATG-3 ) were used for N130D mutants, while primers SARMUT677( ) (5 -GAAAA CATAAAAAAAGCAAATCATACTAACC-3 ) and SARMUT707b( ) (5 - GGTTAGTATGATTTGCTTTTTTTATGTTTTC-3 ) were used for Q129A mutants [mutant nucleotides are underlined in the ( ) primer]. Plasmids psarsst and psarqhd (59) were used as templates, and clones psarqdqd, psarqdhd, psaranqd, and psaranhd were obtained. Their agroinfectious clones were prepared by subcloning the full-length, mutated viral genomes into the pbin19 vector as described previously (59). The constructs used or cited in this work with their relevant CP sequences are shown in Fig. 1A. Tomato (Solanum lycopersicum L.; cv. Marmande) or Nicotiana benthamiana Domin. plants were inoculated with mutant agroclones and the wt TYLCSV (GenBank accession no. X61153) (42) and maintained in the greenhouse at 25 C, under a 16:8 h (light-dark) photoperiod. For insect transmission experiments, plants were used at 6 to 8 weeks postinoculation. Before acquisition, the presence of the correct mutant virus in each source plant was checked by sequencing as previously described (59). Virus purification. Virions were purified from N. benthamiana plants at 4 to 7 weeks postinoculation essentially as described previously (52). Infected tissues were powdered in a mortar with liquid N 2, resuspended in extraction buffer (0.5 M phosphate buffer [PB] [ph 6.0] containing antioxidants, 1% Triton X-100, and 0.1% Driselase; 5 ml/g of fresh tissue), and incubated overnight at 0 C. The homogenate was emulsified with 15% chloroform, and the mixture was spun for 15 min at 8,000 g (Sorvall SS-34 rotor; Du Pont). The aqueous phase was centrifuged for2hat205,000 g (Beckman 55.2 Ti rotor; Beckman, CA). Pellets were resuspended in 0.5 M PB (ph 7.0) containing 2.5 mm EDTA and centrifuged twice at low speed (with an intermediate wash between the two low-speed centrifugations). The supernatants were combined, loaded onto a preformed 20 to 50% Cs 2 SO 4 density gradient in 0.5 M PB (ph 7.0) with 2.5 mm EDTA, and spun for5hat160,000 g (Beckman SW41 rotor). Virus-containing bands were diluted in 0.1 M PB (ph 7.0), and virions were pelleted for 40 min at 390,000 g (Beckman TL100 rotor). Finally, virions were resuspended in 0.1 M PB (ph 7.0) for EM observation or in 0.05 M PB (ph 7.0) containing 15% sucrose for whitefly transmission experiments. Virus concentration of purified preparations was measured in a DU 530 spectrophotometer. Electron microscopy of virus particles. Virus particles were observed and photographed in a CM 10 electron microscope (Philips, The Netherlands), after staining with aqueous 0.5% uranyl acetate on grids coated with Formvar and carbon. Insect handling and virus transmission. B. tabaci whiteflies, biotype B, originating from Liguria, Italy, were raised on TYLCSV-immune cucumber plants (Cucumis sativus L.; cv. Marketer), in a quarantined room at 26 C 1 C, under a 16:8 h (light-dark) photoperiod (17). Trialeurodes vaporariorum Westwood insects were reared on N. tabacum L. plants (cv. White Burley) in a screen cage in a greenhouse at 23 C 2 C. Insects collected from rearing plants will be referred to as nonviruliferous. Acquisitions and inoculations, using only females, took place in a growth chamber at 26 C 1 C, with a 16:8 h (1ight-dark) photoperiod. Retention kinetics of virus DNA in whiteflies measured by quantitative dot blot hybridization. Nonviruliferous B. tabaci whiteflies were allowed a 24-h acquisition access period (AAP) on tomato source plants infected with wt or mutant TYLCSV and then transferred to healthy cucumber plants. Insect samples were taken before the AAP (day 0), at the end of the AAP (day 1), and 2, 4, 6, and 9 days after transfer to immune plants (days 3, 5, 7, and 10, respectively). Insects were killed with CO 2 and individually frozen at 20 C in microplates (conic-bottom wells) with 20 l TE (10 mm Tris-HCl, 1 mm EDTA [ph 8.0]). When needed, insects were thawed and disrupted with a sterile needle under a stereomicroscope; 80 l TE was added, and the mixture was homogenized in an ultrasonic bath (24 khz, 20 min) directly in the microplate. T. vaporariorum was treated the same way. TYLCSV DNA was detected in single whiteflies (or their dissected parts) with digoxigenin-labeled probes (2). The amount of viral DNA was calculated by measuring the total optical density (OD) of each sample on autoradiographic films (15) and analyzed with a 300A computing densitometer and ImageQuant software (release 3.2; Molecular Dynamics, Sunnyvale, CA). For membrane calibration, two series of dilutions of plasmid psarsst (59) containing one genomic unit of TYLCSV DNA in TE buffer (from 10 pg to 75 ng) or the buffer alone were used. The square root of the OD was plotted against the logarithm of known amounts of TYLCSV DNA, and the amount of DNA in each insect or insect part was calculated according to the method of Cacigli and Bosco (15). For each curve, the C, D, and upper discrimination limits were calculated, using a t value at the protection level of 0.05 for all limits (15). Analysis of variance of fitted curves and residuals was performed as described previously (15). In all cases, the F test for the variance ratio of regression gave a P value of Insects having a signal above the C limit are referred to as positive. PCR and real-time PCR. PCR was performed directly on the extracts of single whiteflies used for dot blot hybridization (5 out of 100 l) or after DNA extraction as described previously (54). Viral DNA was extracted from dissected wings and legs and from head/prothorax and abdomen samples of individual whiteflies. End-point PCR (1) and real-time PCR (54) conditions have been described previously. EM-immunogold labeling of the coat protein of wt TYLCSV and its mutants. (i) Specimen preparation. B. tabaci whiteflies exposed to the wt virus or NT mutants and nonviruliferous insects were collected, anesthetized with CO 2, and immersed in fixative (2.5% glutaraldehyde in 0.1 M PB [ph 7.2]) after the wings and legs were removed. The fixative was changed, and the insects were dehydrated in an ethanol series and cold-embedded in Lowicryl K4M resin (72). The wings and legs were analyzed for the presence of viral DNA before processing for EM. B. tabaci insects fed on tomato plants infected with wt TYLCSV, embedded in LR White resin, and used in previous experiments were also analyzed (55). For optimizing the labeling procedure, healthy or TYLCSV-infected tomato leaves (all in LR White; medium grade) were used as plant tissue controls. (ii) Antisera. The anti-tylcv (As-588) and anti-acmv (As-421) polyclonal sera (gift of S. Winter, DSMZ, Braunschweig, Germany) were cross-absorbed with healthy plant extracts or with combined healthy plant and insect tissue extracts, respectively. For immunogold labeling, ultrathin sections ( 70 nm) of plant and insect tissue that were mounted on carbon-coated gold grids as described previously (62) were incubated overnight at 4 C in the above sera, diluted 1:200 in blocking buffer (200 mm Tris-HCl [ph 7.4], 1% Tween 20, 1% bovine serum albumin, and 0.1% gelatin). Grids were washed for 1 min in distilled water and incubated for 1 h at room temperature in goat anti-rabbit immunoglobulin G serum, conjugated to 15-nm gold particles (1:20 dilution) (Janssen AuroProbe; Janssen Pharmaceuticals, NJ). (iii) EM analysis. Sections were stained with uranyl acetate (15 min) and lead citrate (2 min) and examined with an EM 910 microscope (Zeiss, Germany). A minimum of two grids with at least two ultrathin sections for each whitefly were analyzed, and at least three whiteflies per treatment were checked. All insect parts were observed for the presence of gold particles, particularly focusing on the head/prothorax and the beginning of the abdomen, where the main areas involved in transmission, i.e., SGs and filter chamber, are situated (30). Competitive transmission experiments. (i)acquisition from infected plants. To assess whether NT mutants interfered with transmission of the wt virus, nonviruliferous B. tabaci whiteflies were allowed a first AAP (16 h) on tomato plants infected with an NT mutant and a second (12 h) on tomato plants infected

3 5786 CACIAGLI ET AL. J. VIROL. with the wt virus. At the end of the second AAP, groups of five insects were transferred onto healthy tomato plants at the fourth-leaf stage for inoculation, using clip cages. Controls consisted of (i) insects allowed the first AAP on healthy tomato plants and the second on wt-infected tomato plants and (ii) insects transferred to test plants directly after the first AAP on NT mutants. Experiments were repeated three times, and data were pooled for statistical analysis after verifying that the results from the three experiments were homogeneous by the 2 test. (ii) Acquisition from purified virions. Whiteflies were allowed to feed through stretched parafilm membranes on virion preparations (wt, PNHD, and PNHE) purified from N. benthamiana plants and resuspended in 0.05 M PB (ph 7.0) plus 15% sucrose, at a virus concentration of about 160 g/ l, based on A 260 measurements (1 mg/ml 7.7 [32]). Controls consisted of (i) insects with the first AAP on PB plus 15% sucrose without virus and a second on purified wt virus or (ii) insects directly transferred to test plants after the first AAP. This experiment was done in duplicate. RESULTS Infectivity and transmissibility of new coat protein mutants. We previously showed that a single Q129P mutation in TYLCSV CP (PQD and PQE mutants) abolished systemic plant infection and that a concomitant Q134H mutation, irrespective of whether there was a D (as in the wt virus) or E at position 152 (PNHD and PNHE mutants, respectively) restored infectivity; however, PNHD and PNHE mutants could not be transmitted by B. tabaci (59). In this work, two new CP mutants, with N130D and Q or H at position 134 (QDQD and QDHD mutants) and two with Q129A, again with Q or H at position 134 (ANQD and ANHD mutants) were produced (Fig. 1A). All mutants systemically infected N. benthamiana and tomato plants following agroinoculation, with symptoms indistinguishable from those of plants infected with the wt virus. The mutants were also transmitted by B. tabaci, except for the QDQD mutant, for which none of 46 plants inoculated using more than 300 insects in all, became infected, in agreement with results described for Watermelon chlorotic stunt virus (41). This implies that alanine at position 129 (ANHD and ANHE mutants) does not prevent virion assembly and that it has a less drastic impact than proline in the noninfectious PNQD or PNQE mutant. As for Q129P mutants, a Q134H change in the QDQD context has a compensatory role, since the QDHD mutant is transmissible. Retention kinetics of virus DNA of nontransmissible CP mutants in whiteflies. The retention kinetics of the viral DNAs of the NT PNHD, PNHE, and QDQD mutants and the wt virus were compared using quantitative dot blot hybridization on single B. tabaci fed on infected plants. The number of insects that acquired or retained the viral DNAs (positive insects) and the amount of viral DNA present in positive insects over time were calculated, using the nonvector T. vaporariorum as a control. This insect is a phloem-feeding whitefly able to ingest but not transmit begomoviruses (5, 16, 63). For each virus mutant, the mean amount of viral DNA in positive insects and the number of insects tested are shown in Fig. 1B. In B. tabaci, the retention kinetics of wt virus DNA was comparable to previous results (16) and both PNHD and PNHE mutants were retained for at least 10 days after the end of the AAP, like the wt virus. Overall, no significant difference in the amount of viral DNA was found among whiteflies fed on plants infected with wt virus or by PNHD and PNHE mutants. However, it was not possible to quantify the viral DNA in individual whiteflies fed on QDQD-infected plants, since the hybridization signal, although significantly greater than the detection limit C, was lower than the quantification limit. QDQD-positive insects are therefore shown in brackets. Also, no specific trends in the number of positive insects were observed among groups fed on plants infected with wt virus or by PNHD or PNHE mutant, except again for those fed on plants infected with the QDQD mutant, for which the number of positive insects (p) decreased significantly with time (arcsin p day 0.557; P of regression 0.01 with 4 df). Thus, the QDQD mutant is acquired/retained at extremely low levels, and its DNA is degraded with time, falling below the detection limit (Fig. 1B). To further understand whether NT mutants circulated in B. tabaci, PCRs on nucleic acids extracted from whitefly wings and legs were performed. Viral DNA was amplified from the wings and legs of insects fed on plants infected with wt virus or PNHD and PNHE mutants, but not on plants infected with the QDQD mutant, from which positive signals were obtained only from whole-insect extracts (not shown). Retention kinetic experiments were also attempted on individual head/prothorax (including the SG area) and abdomen samples using real-time PCR (54). However, viral DNA could not be amplified in the single head/prothorax samples, even when the corresponding abdomens were positive (not shown). Viral DNA could be amplified only from batches of 20 to 40 heads of whiteflies exposed to the wt virus or the PNHD and PNHE mutants (not shown), but since each batch could contain different ratios of positive insects, no quantitative measurements were feasible. When tested in T. vaporariorum by dot blot hybridization, the retention kinetics of the PNHD mutant were comparable to those of the wt virus, while no DNA quantification of QDQD and PNHE mutants was feasible except, in the latter case, at the end of the AAP (Fig. 1B). In conclusion, while the DNA from PNHD and PNHE mutants circulates in the hemolymph and reaches the SG area of B. tabaci, the DNA of the QDQD mutant does not circulate and could not be quantified. The low acquisition/retention of PNHE in the nonvector indicates that the D152E change also influences virus-vector relationships. Purified QDQD virions are unstable. A prerequisite for systemic TYLCSV infection is virion assembly, and defects in CP expression and function can influence the genomic singlestranded DNA load in infected plants (12, 13). In this study, geminate QDQD virions were found in sap from infected plants (see Fig. S1 in the supplemental material), and no differences were detected in the amount of CP or viral DNA (both single-stranded and replicative supercoiled forms) in wt virus- or QDQD-infected plants (not shown). However, QDQD virions purified from infected N. benthamiana plants showed striking differences compared to plants infected with the wt virus. For the wt virus, two separate gradient bands (upper and lower bands) were obtained following sedimentation in Cs 2 SO 4 (4). Typical geminate particles were in the lower band, while in the upper band about 30% of particles appeared nongeminate and roughly circular (Fig. 2A). In contrast, for the QDQD mutant, only a single diffuse upper band was formed, containing exclusively nongeminate, roughly circular particles. Strands resembling condensed nucleic acid were seen in the background; these strands may be DNA from disrupted virions (Fig. 2A). In some experiments, a few gem-

4 VOL. 83, 2009 RELATIONSHIPS OF TYLCSV MUTANTS WITH THE INSECT VECTOR 5787 FIG. 1. (A) Alignment of the partial coat protein sequences (aa 121 to 160) of the wt TYLCSV (GenBank accession no. X61153) (42) and its CP mutants. The mutants from the QNQE mutant to the PNHE mutant were described previously (59), while ANQD, ANHD, QDQD, and QDHD mutants and their construction are described in this paper. The boxed amino acids are Q129, N130, Q134, and D152 (shown on a black background for the wt virus and on white for mutants). (B) Retention kinetics of the wt TYLCSV and its NT mutants PNHD, PNHE, and QDQD in the whiteflies B. tabaci and T. vaporariorum. Each point represents the mean amount of viral DNA detected in female insects given a 24-h AAP on infected plants, moved to nonhost plants, and then tested by quantitative dot blot hybridization at different times after the end of AAP (arrow). Bars represent standard errors. The number of positive insects/number of tested insects is shown beside each point. The numbers of QDQD-positive insects are shown in brackets.

5 5788 CACIAGLI ET AL. J. VIROL. Downloaded from FIG. 2. Comparison of wt TYLCSV and mutant QDQD virion morphology and components. (A) Micrographs of purified wt TYLCSV particles present in the upper and lower gradient bands obtained after purification (arrows point to monomers) and of the QDQD mutant particles collected from the single upper band. Bars 40 nm. (B) Sodium dodecyl sulfate-polyacrylamide gel stained with Coomassie blue and Western blot of proteins solubilized from the upper (U) and lower (L) bands of the wt virus and QDQD mutant. The Western blot (containing one-third of the amount of protein extracts of samples loaded in the Coomassie blue-stained gel) was developed with ACMV antiserum (As-421, 1:5,000). The positions of molecular markers (in kilodaltons) are indicated to the right of the gel or blot. inate QDQD particles could be obtained by reducing the relative centrifugal force during purification (not shown). Protein extracts from both the upper and lower bands of purified wt virions, analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting showed a single polypeptide of 28 kda, while those from purified QDQD mutant virions showed a cluster of bands of 24 to 28 kda, recognized by the antiserum, plus a 10-kDa polypeptide, that was not immunodetectable (Fig. 2B). Thus, the single N130D mutation does not impede twinnedvirion formation, but it causes the assembly of unstable virions that dissociate during purification, with partial degradation of the CP subunit. Immunolocalization of the CP of the wt virus and NT mutants in plants and whiteflies. The different retention kinetics of NT mutants in whiteflies, as well as QDQD virion instability, prompted us to examine the fate of virion components in whiteflies, using immunogold labeling of the CP. Sera against ACMV and TYLCV were initially tested on previously embedded plant and whitefly samples (55), overall encountering no differences between Lowicryl or LR White resins. The anti- TYLCV serum was more specific, except for a weak background in Lowicryl-embedded samples. In tomato plant tissue, weak nonspecific labeling of chloroplasts was observed with both sera, even after cross-adsorption with healthy plant extracts. With both sera, virion aggregates within the nuclei of bundle sheath cells and inside the phloem were specifically labeled, but in none of the sections analyzed could geminate virions be distinguished; sometimes, amorphous material was also labeled, probably deriving from necrotic cells with abnormal cell wall growth (see Fig. S2 in the supplemental material) as previously reported (55). The ACMV antiserum labeled chromatin in the nuclei of both viruliferous and nonviruliferous whiteflies, in different kinds of cells of the gut, cybarium, and SG areas (see Fig. S3 in the supplemental material). Further cross-absorption with nonviruliferous whitefly extracts was therefore needed to eliminate unspecific nuclear labeling. Surprisingly, this reaction was absent with the TYLCV antiserum. When background or unspecific labeling was absent on nonviruliferous insects taken as on November 2, 2018 by guest

6 VOL. 83, 2009 RELATIONSHIPS OF TYLCSV MUTANTS WITH THE INSECT VECTOR 5789 FIG. 3. Immunogold labeling of SGs of B. tabaci using ACMV antiserum (As-421) cross-absorbed with healthy plant and nonviruliferous whitefly tissue. (A) Overview of a PSG area of a nonviruliferous control whitefly. (B) Overview of a PSG area of a whitefly that fed on plants infected with wt TYLCSV. The arrow points to the basal membrane (BM). Connective tissue outside SGs is indicated by an asterisk. (C) Highermagnification view of a section of control tissue in panel A. (D) Higher-magnification view of a section of panel B, showing a mass of labeled vesicle structures, putative saliva droplets. Bars 5 m (A and B), 0.5 m (C), and 1 m (D). negative controls (Fig. 3A and C), whiteflies fed on plants infected with wt virus or NT mutants were analyzed, using only individuals whose wings and legs had been found TYLCSV positive in PCR, except for whiteflies fed on plants infected with the QDQD mutant, which were always negative (see above). B. tabaci whiteflies fed on plants infected with wt TYLCSV showed a specific general labeling of the gut lumen, filter chamber area, gut cells, and other insect parts with both

7 5790 CACIAGLI ET AL. J. VIROL. sera (not shown), confirming the ability of ACMV antiserum to recognize TYLCSV (55). In spite of some variability among specimens, a strong signal was clearly distinguishable in a vacuolated area, concentrated in a type of vesicle or droplet structure, putative saliva drops, present inside PSGs (Fig. 3B and D). B. tabaci fed on PNHD-infected plants also showed labeling in some PSG cells close to the basal membrane (Fig. 4A to C), with strong signals localized in similar droplet structures (Fig. 4A and B). For the PNHE mutant, in one of five viruliferous insects checked, labeling was found in the cytoplasm of some PSG cells (Fig. 4D and E), particularly those close to the basal membrane of cells numbered 1 and 2 (30). Finally, for the QDQD mutant, only diffuse and weak labeling, indistinguishable from background with no specific signals around SGs was found (not shown), suggesting that partially disassembled or free virion components do not accumulate in specific tissues. These data are in agreement with the poor recovery of QDQD DNA from whiteflies. Notably, with both antisera, light labeling of electron-dense drops in eggs was detected in nonviruliferous insects (Fig. 5A) or in those fed on wt virus-infected plants (Fig. 5B). This reaction was especially strong for whiteflies fed on PHNE- and PNHD-infected plants (Fig. 5C and D), but the reaction was again weak for QDQD-infected plants (Fig. 5E) and for plants infected with an NT isolate of ACMV (not shown). Although the nature of this reaction is unknown, the strong but reproducible differences indicate that this labeling was indeed specific. In conclusion, no differences were found in CP localization in the gut lumen, gut cells, and SGs among insects fed on plants infected with wt virus or PNHD and PNHE mutants; notably, both mutants, or at least their CPs, reached the SGs, as did the wt virus, while the QDQD-CP was not detected in any of the whitefly tissues. The PNHE mutant inhibits transmission of the wt virus. Competitive transmission experiments were designed to analyze whether NT mutants interfered with transmission of the wt virus by B. tabaci. When whiteflies acquired virus from PNHEinfected tomato plants and were then fed on plants carrying wt TYLCSV, transmission of the wt was reduced by about 35% Virus (P 0.001) (Fig. 6A). In the control, transmission efficiency of the wt virus fell within the expected values for a 12-h AAP (17). No inhibition was observed with the PNHD mutant. To correctly interpret this result, virus titers should be similar in the source plants. Although detailed quantitative analysis was not carried out, no substantial differences in the virus titers were noticed in plants infected with NT mutants or the wt virus, using either Southern or Western blots (see results above) (59). However, to validate this interference, competitive transmissions were also performed using purified virions. In this case, the first acquisition on QDQD virions was omitted, since QDQD did not interfere using infected plant sources and its particles were unstable following purification. In this case, prefeeding on PNHE virions completely abolished transmission of the wt virus, while no effect was observed with PNHD (Fig. 6B). Therefore, the PNHE mutant interferes with transmission of the wt virus when acquired either from infected plants or from purified preparations. DISCUSSION Mutants that can systemically infect plants but are not transmitted by whiteflies may reveal domains involved in the interaction between viruses and insects and may help identify molecular components relevant for transmission. In this study, we analyzed the fate of three nontransmissible TYLCSV CP mutants in whiteflies and tested whether they interfered with the transmission of the wt virus. The overall results are summarized in Table 1. The main parameters governing acquisition and transmission of TYLCSV and the related species TYLCV are already known (17, 19, 20). The kinetic experiments reported here showed that the infectious and virion-forming NT mutants PNHD and PNHE, or at least their DNA, persisted in whiteflies as the wt virus did, indicating that they circulated in B. tabaci hemocoel and reached the head/prothorax organs. In contrast, the QDQD mutant was strikingly impaired in retention kinetics, and even immediately after the whitefly fed, QDQD mutant DNA could not be quantified. QDQD genome fragments could be detected from whole-insect extracts (but TABLE 1. Summary of data on virion formation, transmission, retention kinetics, and localization of TYLCSV CP mutants a Systemic infection Particle formation Biological data Retention kinetics in whiteflies b EM localization in B. tabaci c Transmission by B. tabaci Interference in transmission of wt virus B. tabaci vector T. vaporariorum nonvector wt // Up to 10 days Up to 10 days / Mutants ANQD NA NA NA NA NA NA NA ANHD NA NA NA NA NA NA NA QDHD NA NA NA NA NA NA NA PNHD Up to 10 days Up to 10 days PNHE Up to 10 days NQ QDQD d NQ NQ / PNQD // // // // // // // PNQE // // // // // // // Gut Salivary glands Eggs a, positive;, negative; //, not testable; NA, not assayed. b Dot blot hybridization with digoxigenin-labeled probes on single B. tabaci extracts. NQ, not quantifiable. c Detection with ACMV (As-421) and TYLCV (As-588) antisera., localized; /, signal similar to that of the nonviruliferous whitefly control;, labeling mainly on electron-dense drops in eggs;, not localized. d Nongeminate particles following purification.

8 VOL. 83, 2009 RELATIONSHIPS OF TYLCSV MUTANTS WITH THE INSECT VECTOR 5791 FIG. 4. Immunogold labeling of SGs of B. tabaci fed on plants infected with the PNHD mutant (A to C) or PNHE mutant (D and E), using the ACMV antiserum (As-421) cross-adsorbed with healthy plant and nonviruliferous whitefly tissues. (A to C) Peripheral part of a SG, with cytoplasmic labeling of different cells shown in the boxed areas (B and C) or in the oval. (D) Peripheral part of a SG. (E) Higher-magnification view of a section of panel D, with strong cytoplasmic labeling close to the basal membrane (black arrows). Connective tissue outside SGs is indicated by an asterisk. Abbreviations: BM, basal membrane; N, nucleus; EDG, electron-dense granules; RER, rough endoplasmic reticulum. Bars 3.63 m (A), 1.45 m (B and C), 4.6 m (D), and 2.3 m (E).

9 5792 CACIAGLI ET AL. J. VIROL. FIG. 5. Immunogold labeling of eggs using ACMV antiserum (As- 421) cross-adsorbed with healthy plant and nonviruliferous whitefly tissues. Nonviruliferous B. tabaci (A) and B. tabaci fed on plants infected with the wt virus (B), PNHE mutant (C), PNHD mutant (D), and QDQD mutant (E). All gold particles are circled. Bars 1.45 m (A, C, and D), 2.9 m (B), and 2.3 m (E). FIG. 6. Effects of the NT mutants on the transmission of the wt virus. Percentages of infected plants/inoculated plants obtained after serial acquisition of viruses from infected plants (A) or from purified virus preparations (B). Whiteflies were allowed an initial 16-h AAP on plants infected with NT viruses (I AAP) or on healthy plants (H) and then a second 12-h AAP (II AAP) on plants infected with the wt virus. Controls for purified virions consisted of I AAP on 15% sucrose. Groups of insects (five insects per group) were transferred to healthy tomato plants for inoculation, and the plants were tested for TYLCSV infection 3 weeks later. Data were pooled from three independent experiments. The infection rates shown above each bar with different superscripts (a, b, and c) differ significantly by the 2 test (the ratio of number of infected plants and the number of inoculated plants was significantly different). not from wings and legs) only by PCR, which is about 1,000 times more sensitive than dot blot hybridization (54), meaning that QDQD movement within whiteflies was severely limited. For TYLCSV and other monopartite geminiviruses (9, 12, 47, 51, 64, 65), assembled virions are needed for systemic infection of plants and may also be required for circulation within the vector (28). Although no differences in wt or QDQD viral loads were detected in infected plants and QDQD geminate particles were found in plant sap by EM (see Fig. S1 in the supplemental material), QDQD virions appeared nongeminate following purification, with partially degraded CP. In contrast, only a few nongeminate particles were seen in the upper band of wt virus preparations, with no signs of CP degradation. Nongeminate particles have already been reported for other geminiviruses (37) and found to contain nucleic acid, though nontransmissible (46). Furthermore, nongeminate particles smaller than monomeric ones and with a T1 symmetry have been found to encapsidate defective interfering DNA molecules, half the genome size (23, 40, 53), but in our case, defective interfering DNAs were never found and nongeminate QDQD particles were approximately half-virion size. The geminivirus packaging/unpackaging process from CP oligomerization to capsomer assembly and virion building and stability is far from clear, and nucleic acids and host or vector factors may be required. Since TYLCV CP cooperatively binds single-stranded DNA (61), this interaction could be fundamental for virion assembly. Furthermore, to our knowledge, there have been no reports of empty geminivirus particles. Using TYLCV CP deletion mutants in a two-hybrid system, domains involved in CP self-assembly have been identified at the N and

10 VOL. 83, 2009 RELATIONSHIPS OF TYLCSV MUTANTS WITH THE INSECT VECTOR 5793 C termini, and full-length capsid proteins from the QNQE, QNHD, and QNHE mutants were found capable of self-interaction, although the interaction was loose (34). However, the corresponding TYLCSV mutants formed infectious and transmissible virions (59). Whether the CPs of the PNQD and PNQE mutants, for which virions were indeed not detected (59), or of the QDQD mutant, whose particles are unstable (this work), are defective in self-interaction, remains to be determined. The architecture of certain plant viruses can change with ph or metal ion concentration (50, 60, 66). Recently, ACMV virions, stable between ph 4.0 and 8.5, were found to disassemble in 12-nm pentameric capsomers, without forming intermediate nongeminate particles similar to those described here (43). The structure of Maize streak virus has been determined, and its CP is the only structural model available for geminiviruses, with an eight-stranded, antiparallel -barrel shape, similar to the CP of the RNA virus Satellite tobacco necrosis virus (74). This model, further improved by fitting it to the CP of ACMV (8), shows a 6-aa sequence (aa 128 to 133) protruding as a loop from the E strand of Satellite tobacco necrosis virus, where amino acids 129 and 130 could be the most relevant residues for virion architecture and possibly for interaction with whitefly components. Taking advantage of the anatomical and ultrastructural descriptions of B. tabaci and its SG system (35, 36) and of previous immunolocalization studies (30, 55), we have analyzed the localization of the CP NT mutants by immunogold labeling in whitefly tissues or organs known to be involved in transmission. Careful control of experimental conditions aimed at reducing the nonspecific reactions of the antisera, particularly with insect cell nuclei, was necessary to avoid misleading conclusions concerning, for instance, virus multiplication in the vector (22, 27). In the gut area, including midgut and microvilli, the localization patterns of the wt TYLCSV and its NT mutants (PNHD and PNHE mutants) were similar to that of TYLCV (27). No labeling was found over the ASGs, confirming that they are not involved in transmission (18, 27), in contrast to the data on luteoviruses in aphids (31). Also, we confirmed the presence of the wt virus in PSGs and in the cytoplasm of PSG cells (27, 55). More importantly, the PNHD and PNHE mutants or at least their CPs were found inside PSGs. Combining the DNA retention kinetics and the CP localization results, it is highly probable that the PNHD and PNHE mutants move in the hemolymph and cross the hemocoel/sg wall as virions, as the wt virus does. It is noteworthy that the presence of NT mutants in PSGs as virions implies that crossing of the SG wall may not guarantee transmission as previously postulated (28). Interaction with SG molecular factors may be fundamental to maintain infectivity. Direct contact with saliva components could compromise the stability of some virion structures. However, while the saliva of certain aphid species has been at least partially characterized (56, 70), there is little biochemical data available for whiteflies, with only one report of alkaline phosphatase activity in the basal regions of B. tabaci SGs and connecting ducts and in saliva secreted during feeding (25). The discrepancy in the retention kinetics of the PNHE mutant in B. tabaci and T. vaporariorum (Fig. 1) and especially the ability of only this mutant to inhibit transmission of the wt virus (Fig. 6) are indicative of the complexity of the relationships between begomoviruses and whiteflies. Although even the existence of whitefly components interacting with begomoviruses is still speculative, it is possible that PNHE virions or virion components irreversibly bind and block putative cell receptors inside PSG cells of B. tabaci. Ongoing whitefly functional genome projects are expected to help to identify genes involved in Begomovirus transmission (26, 48). Viral DNA fragments have been amplified from ovary tissue of whiteflies that acquired TYLCV (29) and TYLCSV (7). However, we found no specific labeling of the CP in the ovary area for the wt virus. The strong reaction observed with the PNHD and PNHE CPs suggests a direct association with egg components, but since neither the QDQD mutant nor an NT strain of ACMV reacted similarly, binding to egg constituents cannot be considered a hallmark of nontransmissibility. Several symbiotic relationships exist among bacteria and arthropods. For B. tabaci, symbionts were found in bacteriocytes, specialized cells transmitted through eggs (33). Insect symbionts synthesize GroEL-type chaperonins that are thought to bind to virions and protect them in the hemolymph. For luteoviruses, GroEL homologues of Buchnera sp. play a role in transmission (69). In whiteflies, prior feeding with anti-groel antibodies strongly reduced TYLCV transmission (58), and binding of geminivirus CP to GroEL confirmed the specificity of this interaction (3, 57). However, since the CP of NT isolates of AbMV still bound to GroEL (57), its role in transmission appears controversial. Since the QDQD mutant is not detected in the hemolymph, its putative interaction with GroEL homologues can be disregarded, and QDQD virion instability is the most probable cause of its low persistence in whiteflies. Other cases of NT geminiviruses have been reported. AbMV probably lost its ability to be transmitted because of extended vegetative propagation (73). An Israeli NT isolate of AbMV persisted in the gut of B. tabaci for more than a week but did not circulate in its hemolymph (57). In a West Indian AbMV isolate (24), residues 124, 149, and 174 were found responsible for transmissibility (38). By mutating residues expected to be surface exposed and relevant for protein-protein interaction in the CP of the curtovirus Beet mild curly top virus, an NT mutant (CP 25 28) was identified that was infectious, formed virions, and was taken up by its natural vectors (65), like the PNHD and PNHE TYLCSV mutants, but unfortunately, details on its stability in the vector or on the localization of Beet mild curly top virus and its mutants in the insect were not reported. Transcytosis across the gut and ASG membranes seems to be the main mechanism governing luteovirus movement in aphids (10). However, the difficulties of directly visualizing geminivirus particles in whiteflies by transmission electron microscopy and the paucity of information on molecular factors involved do not allow, at the moment, demonstration that begomoviruses rely on a similar mechanism. In conclusion, our findings strengthen the idea that virion formation is necessary but not sufficient for begomovirus transmissibility (59) and demonstrate that virion stability is important and that crossing the SG barrier may not be sufficient for transmission. Other factors, such as the ability to cross gut epithelia and the interaction with chaperones or molecular components within SGs and helping virions to preserve infectivity, may also influence geminivirus transmissibility.

11 5794 CACIAGLI ET AL. J. VIROL. ACKNOWLEDGMENTS We thank Bob Milne, Gian Paolo Accotto, and Jozsef Burgyan for critical discussions. B. Milne also revised the English. This work was supported in part by the Spanish government (CICYT project; AGF ) and by Consiglio Nazionale delle Ricerche for a Short Term Mobility Grant 2006 to V.M. REFERENCES 1. Accotto, G. P., J. Navas-Castillo, E. Noris, E. Moriones, and D. Louro Typing of tomato yellow leaf curl viruses in Europe. Eur. J. Plant Pathol. 106: Accotto, G. P., A. M. Vaira, E. Noris, and M. Vecchiati Using nonradioactive probes on plants: a few examples. J. Biol. Chem. 13: Akad, F., N. Dotan, and H. Czosnek Trapping of Tomato yellow leaf curl virus (TYLCV) and other plant viruses with a GroEL homologue from the whitefly Bemisia tabaci. Arch. Virol. 149: Al-Bitar, L., and E. Luisoni Tomato yellow leaf curl geminivirus: serological evaluation of an improved purification method. EPPO Bull. 25: Antignus, Y., M. Perlsman, R. Ben-Yoseph, and S. Cohen Interaction of Tomato yellow leaf curl virus with its whitefly vector, Bemisa tabaci. Phytoparasitica 21: Azzam, O., J. Frazer, D. De la Rosa, J. S. Beaver, P. Ahlquist, and D. P. Maxwell Whitefly transmission and efficient ssdna accumulation of bean golden mosaic geminivirus require functional coat protein. Virology 204: Bosco, D., G. Mason, and G. P. Accotto TYLCSV DNA, but not infectivity, can be transovarially inherited by the progeny of the whitefly vector Bemisia tabaci (Gennadius). Virology 323: Böttcher, B., S. Unseld, H. Ceulemans, R. B. Russell, and H. Jeske Geminate structures of African cassava mosaic virus. J. Virol. 78: Boulton, M. I., H. Steinkellner, J. Donson, P. G. Markham, D. I. King, and J. W. Davies Mutational analysis of the virion-sense genes of Maize streak virus. J. Gen. Virol. 70: Brault, V., E. Herrbach, and C. Reinbold Electron microscopy studies on luteovirid transmission by aphids. Micron 38: Briddon, R. W., M. S. Pinner, J. Stanley, and P. G. Markham Geminivirus coat protein replacement alters insect specificity. Virology 177: Briddon, R. W., J. Watts, P. G. Markham, and J. Stanley The coat protein of Beet curly top virus is essential for infectivity. Virology 172: Brough, C. L., R. J. Hayes, A. J. Morgan, R. H. A. Coutts, and K. W. Buck Effects of mutagenesis in vitro on the ability of cloned Tomato golden mosaic virus DNA to infect Nicotiana benthamiana plants. J. Gen. Virol. 69: Brown, J. K., and H. Czosnek Whitefly transmission of plant viruses. Adv. Bot. Res. 36: Caciagli, P., and D. Bosco Quantitative determination of tomato yellow leaf curl geminivirus DNA by chemiluminescent assay using digoxigenin-labeled probes. J. Virol. Methods 57: Caciagli, P., and D. Bosco Quantitation over time of tomato yellow leaf curl geminivirus DNA in its whitefly vector. Phytopathology 87: Caciagli, P., D. Bosco, and L. Al-Bitar Relationships of the Sardinian isolate of tomato yellow leaf curl geminivirus with its whitefly vector Bemisia tabaci Gen. Eur. J. Plant Pathol. 101: Cicero, J. M., and J. K. Brown Squash leaf curl virus localizes in primary salivary gland compartments, and at midgut and filter chamber brush border cells in viruliferous Bemisia tabaci. J. Insect Sci. 8: Cohen, S., and I. Harpaz Periodic, rather than continual acquisition of a new tomato virus by its vector, the tobacco whitefly (Bemisia tabaci Gennadius). Entomol. Exp. Appl. 7: Cohen, S., and F. E. Nitzani Transmission and host range of the tomato yellow leaf curl virus. Phytopathology 56: Czosnek, H Interactions of Tomato yellow leaf curl virus with its whitefly vector, p In H. Czosnek (ed.), Tomato yellow leaf curl virus disease: management, molecular biology, breeding for resistance. Springer, Dordrecht, The Netherlands. 22. Czosnek, H., M. Ghanim, S. Morin, G. Rubinstein, V. Fridman, and M. Zeidan Whiteflies: vectors, and victims (?), of geminiviruses. Adv. Virus Res. 57: Frischmuth, T., M. Ringel, and C. Kocher The size of encapsidated single-stranded DNA determines the multiplicity of African cassava mosaic virus particles. J. Gen. Virol. 82: Frischmuth, T., G. Zimmat, and H. Jeske The nucleotide sequence of abutilon mosaic virus reveals prokaryotic as well as eukaryotic features. Virology 178: Funck, C. J Alkaline phosphatase activity in whitefly salivary glands and saliva. Arch. Insect Biochem. 46: Ghanim, M., and S. Kontsedalov Gene expression in pyriproxyfenresistant Bemisia tabaci Q biotype. Pest Manag. Sci. 63: Ghanim, M., and V. Medina Localization of Tomato yellow leaf curl virus in its whitefly vector Bemisia tabaci, p In H. Czosnek (ed.), Tomato yellow leaf curl virus disease: management, molecular biology, breeding for resistance. Springer, Dordrecht, The Netherlands. 28. Ghanim, M., S. Morin, and H. Czosnek Rate of Tomato yellow leaf curl virus translocation in the circulative transmission pathway of its vector, the whitefly Bemisia tabaci. Phytopathology 91: Ghanim, M., S. Morin, M. Zeidan, and H. Czosnek Evidence for transovarial transmission of tomato yellow leaf curl virus by its vector, the whitefly Bemisia tabaci. Virology 240: Ghanim, M., R. C. Rosell, L. R. Campbell, H. Czosnek, J. K. Brown, and D. E. Ullman Digestive, salivary, and reproductive organs of Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) B type. J. Morphol. 248: Gildow, F. E., and S. M. Gray The aphid salivary gland basal lamina as a selective barrier associated with vector-specific transmission of barley yellow dwarf luteoviruses. Phytopathology 83: Goodman, R. M., and J. Bird Bean golden mosaic virus. CMI/AAB Descriptions of Plant Viruses, no Association of Applied Biologists, Wellesbourne, Warwick, United Kingdom. 33. Gottlieb, Y., M. Ghanim, G. Gueguen, S. Kontsedalov, F. Vavre, F. Fleury, and E. Zchori-Fein Inherited intracellular ecosystem: symbiotic bacteria share bacteriocytes in whiteflies. FASEB J. 22: Hallan, V., and Y. Gafni Tomato yellow leaf curl virus (TYLCV) capsid protein (CP) subunit interactions: implications for viral assembly. Arch. Virol. 146: Harris, K. F., Z. Pesic-Van Esbroeck, and J. E. Duffus Anatomy of a virus vector, p In D. Gerling and R. T. Mayer (ed.), Bemisia 1995: taxonomy, biology, damage control and management. Intercept Ltd., Andover, Hants, United Kingdom. 36. Harris, K. F., Z. Pesic-Van Esbroeck, and J. E. Duffus Morphology of the sweet potato whitefly, Bemisia tabaci (Homoptera, Aleyrodidae) relative to virus transmission. Zoomorphology 116: Hatta, T., and R. I. B. Francki The fine structure of chloris striate mosaic virus. Virology 92: Höhnle, M., P. Höfer, I. D. Bedford, R. W. Briddon, P. G. Markham, and T. Frischmuth Exchange of three amino acids in the coat protein results in efficient whitefly transmission of a nontransmissible Abutilon mosaic virus isolate. Virology 290: Hunter, W. B., E. Hiebert, S. E. Webb, J. H. Tsai, and J. E. Polston Location of geminiviruses in the whitefly Bemisia tabaci (Homoptera: Aleyrodidae). Plant Dis. 82: Jovel, J., W. Prieß, and H. Jeske Characterization of DNA intermediates of an arising geminivirus. Virus Res. 130: Kheyr-Pour, A., K. Bananej, G. A. Dafalla, P. Caciagli, E. Noris, A. Ahoonmanesh, H. Lecoq, and B. Gronenborn Watermelon chlorotic stunt virus from the Sudan and Iran: sequence comparisons and identification of a whitefly-transmission determinant. Phytopathology 90: Kheyr-Pour, A., M. Bendahmane, V. Matzeit, G. P. Accotto, S. Crespi, and B. Gronenborn Tomato yellow leaf curl virus from Sardinia is a whitefly-transmitted monopartite geminivirus. Nucleic Acids Res. 19: Kittelmann, K., and H. Jeske Disassembly of African cassava mosaic virus. J. Gen. Virol. 89: Kotlizky, G., M. I. Boulton, C. Pitaksutheepong, J. W. Davies, and B. L. Epel Intracellular and intercellular movement of maize streak geminivirus V1 and V2 proteins transiently expressed as green fluorescent protein fusions. Virology 274: Kunik, T., K. Palanichelvam, H. Czosnek, V. Citovsky, and Y. Gafni Nuclear import of the capsid protein of tomato yellow leaf curl virus (TYLCV) in plant and insect cells. Plant J. 13: Larsen, R. C., and J. E. Duffus A simplified procedure for the purification of curly top virus and the isolation of its monomer and dimer particles. Phytopathology 74: Lazarowitz, S. G., A. J. Pinder, V. D. Damsteegt, and S. G. Rogers Maize streak virus genes essential for systemic spread and symptom development. EMBO J. 8: Leshkowitz, D., S. Gazit, E. Reuveni, M. Ghanim, H. Czosnek, C. McKenzie, R. L. J. Shatters, and J. K. Brown Whitefly (Bemisia tabaci) genome project: analysis of sequenced clones from egg, instar, and adult (viruliferous and non-viruliferous) cdna libraries. BMC Genomics 7: Liu, H., M. I. Boulton, C. L. Thomas, D. A. M. Prior, K. J. Oparka, and J. W. Davies Maize streak virus coat protein is karyophyllic and facilitates nuclear transport of viral DNA. Mol. Plant-Microbe Interact. 12: Liu, H., C. Qu, J. E. Johnson, and D. A. Case Pseudo-atomic models of swollen CCMV from cryo-electron microscopy data. J. Struct. Biol. 142: Liu, L., J. W. Davies, and J. Stanley Mutational analysis of bean yellow dwarf virus, a geminivirus of the genus Mastrevirus that is adapted to dicotyledonous plants. J. Gen. Virol. 79: Luisoni, E., R. G. Milne, and M. Vecchiati Purification of tomato yellow leaf curl geminivirus. Microbiologica 18: MacDowell, S. W., R. H. Coutts, and K. W. Buck Molecular charac-