The Effect of Sweet Potato Virus Disease and its Viral Components on Gene Expression Levels in Sweetpotato

Sweet potato virus disease (SPVD) is the most devastating disease of sweetpotato [Ipomoea batatas (L.) Lam.] globally. It is caused by the co-infection of plants with a potyvirus, sweet potato feathery mottle virus (SPFMV), and a crinivirus, sweet potato chlorotic stunt virus (SPCSV). In this study we report the use of cDNA microarrays, containing 2765 features from sweetpotato leaf and storage root libraries, in an effort to assess the effect of this disease and its individual viral components on the gene expression profi le of I. batatas cv. Beauregard. Expression analysis revealed that the number of differentially expressed genes (P < 0.05) in plants infected with SPFMV alone and SPCSV alone compared to virus-tested (VT) plants was only 3 and 14, respectively. However, these fi ndings are in contrast with SPVD-affected plants where more than 200 genes were found to be differentially expressed. SPVD-responsive genes are involved in a variety of cellular processes including several that were identifi ed as pathogenesisor stress-induced. Sweetpotato is the seventh most important food crop in the world, with annual world production of ≈130 million tonnes. It ranks third among root and tuber crops worldwide (Food and Agriculture Organization of the United Nations, 2005). Viral diseases, including those caused by mixed infections, are of major economic importance in most production areas around the globe. The use of vegetative cuttings as a principal propagation method provides viruses an effi cient way to perpetuate and disseminate between growing seasons as well as growing areas (Salazar and Fuentes, 2001). As many as 19 different viruses have been identifi ed in sweetpotato and 11 of these are currently recognized by the International Committee of Taxonomy of Viruses (Kreuze, 2002). The effects of these viruses on production range from minimal, to completely devastating, depending on the infecting virus, virus complexes, and sweetpotato cultivars involved. The most important and devastating viral disease affecting sweetpotatoes worldwide is sweet potato virus disease (SPVD). Yield losses of up to 90% have been reported in plants affected with SPVD (Gutiérrez et al., 2003; Hahn, 1976; Ngeve, 1990). SPVD is caused by a synergistic interaction between a potyvirus, sweet potato feathery mottle virus (SPFMV), and a crinivirus, Received for publication 7 Feb. 2006. Accepted for publication 25 May 2006. Approved for publication by the director of the Louisiana Agricultural Experiment Station as manuscript number 05-38-0752. The authors would like to thank Bryon Sosinski (Dept. of Horticultural Science, North Carolina State Univ., Raleigh) and Kornel Burg (ARC Seibersdorf research GmbH, A-2444 Seibersdorf, Austria) for making the cDNA clones available, Limei He, Regina Ali (both from CALS Genome Research Lab, North Carolina State University, Raleigh NC 27695, USA) and Joanna Jankowicz (ARC Seibersdorf research GmbH) for their hard work during development of the sweetpotato microarray, and Silvia Fluch (ARC Seibersdorf research GmbH) for printing the arrays. We would also like to thank Dr. R.A. Valverde for critically reviewing this manuscript. This research was supported in part by funds from the McKnight Foundation Collaborative Crop Research Program. 1These authors contributed equally to this research. 2Corresponding author: Don R. LaBonte; Telephone: +225-578-1024; Fax: +225578-1068; E-mail: [email protected] sweet potato chlorotic stunt virus (SPCSV). Plants co-infected with SPFMV or other sweetpotato potyviruses and SPCSV exhibit severe symptoms such as leaf strapping, vein clearing, leaf distortion, chlorosis, puckering, and stunting. The severity of symptoms, which develop fi rst in the newly emerging leaves, can be directly associated with the dramatic yield reductions observed (Salazar and Fuentes, 2001). The time from initial infection to the appearance of symptoms varies depending on age and size of the plant, with symptoms taking longer to develop on older and bigger plants (Karyeija et al., 2000). SPVD has been reported in a number of African countries, including Rwanda, Burundi, Uganda, Ghana, Nigeria, Kenya, Tanzania, Zimbabwe (reviewed by Karyeija et al., 1998a), and Egypt (Ishak et al., 2003). Outside Africa, this disease has been reported in Israel (Loebenstein and Harpaz, 1960), Spain (Valverde et al., 2004), and Peru (Gutierrez et al., 2003). Since SPFMV is found wherever sweetpotatoes are grown and SPCSV has recently been reported in China (Zhang et al., 2005) and Korea (Yun et al., 2002), SPVD is thus likely to occur in these countries as well. In Argentina, a similar synergism, known as chlorotic dwarf, has been reported that also includes a third virus, sweet potato mild speckling virus (Di Feo et al., 2000). SPFMV, a member of the Potyviridae family and the Potyvirus group, is transmitted by a number of aphid species, including Aphis gossypii Glover and Myzus persicae Sulzer. Plants infected with SPFMV alone, often are symptomless or exhibit mild symptoms and the yield losses are usually minimal (Clark and Hoy, 2006; Gutiérrez et al., 2003). The titers of SPFMV in these plants are similarly low (Kokkinos and Clark, 2006). However, the titers increase dramatically when plants are co-infected with SPCSV (Karyeija et al., 2000; Kokkinos and Clark, 2004), with a corresponding increase in the severity of disease symptoms and yield loss. SPFMV is common wherever sweetpotatoes are grown (Brunt el al., 1996). In the U.S. two strains of SPFMV are recognized, the common strain (SPFMV-C) and the russet crack strain (SPFMV-RC). However, SPFMV-C does not cause septbook 657 9/27/06 11:36:51 AM 658 J. AMER. SOC. HORT. SCI. 131(5):657–666. 2006. typical SPVD symptoms in the presence of SPCSV. Symptoms are usually mild and transient or typical of single infections with SPCSV (Souto et al., 2003). Infection of sweetpotatoes with the whitefl y-transmitted (Bemisia tabaci Gennadius, Trialeurodes abutilonea Haldeman), phloem-limited crinivirus (family Closteroviridae) SPCSV alone can lead to mild to moderate symptoms, with yield losses of up to 43% (Gutiérrez et al., 2003). This virus consists of two distinct strain groups, the east African (EA) and west African (WA), both of which are able to cause synergistic disease (Ishak et al., 2003; Tairo 2005). The titers of this virus are relatively high in infected plants. Interestingly, the titers do not change signifi cantly after co-infection with SPFMV (Karyeija et al., 2000). To date SPCSV has only been found in the United States, in a tissue culture accession and not in the fi eld (Pio-Ribeiro et al., 1996). Efforts to breed for resistance to SPVD have until now focused mainly on breeding for resistance to SPFMV and many sweetpotato cultivars are reasonably resistant to SPFMV (Gibson et al., 1998). Efforts to use SPFMV resistance to breed for SPVD resistance have been unsuccessful because the SPFMV resistance is broken when plants are co-infected with SPCSV (Karyeija et al., 1998b). The mechanism underlying the synergistic interaction between SPFMV and SPCSV and its effect on the host’s response to infection are not known. It is possible that other molecular interactions in the dual infection process may provide better opportunities for resistance to SPVD than narrowly focusing on resistance to SPFMV. Understanding this phenomenon is essential if breeding for resistance to SPVD is going to be successful. An understanding of host–pathogen interactions on the molecular level can provide new insights into the effect of the synergism between SPFMV and SPCSV on the host, and can lead to new approaches in breeding for resistance to SPVD. Microarray technology (Schena et al., 1995) makes possible the assessment of relative gene expression levels of thousands of genes simultaneously. Genes from the organism under investigation (sweetpotato in this case) are spotted on a glass slide, which is then hybridized with mRNA from different treatments. The use of two different fl orescent dyes makes it possible to hybridize two treatments (or a treatment and control) on a single array. After hybridization the array is scanned using a fl uorescent scanner and computer software is used to extract intensity values from the image. Statistical analysis of the data makes it possible to determine which genes are differentially expressed between treatments. Microarrays have already been used to investigate host–pathogen interactions in plants (De Vos et al., 2005; Dowd et al., 2004; Gibly et al., 2004; Moy et al., 2004) and other organisms (for review see Kato-Maeda et al., 2001). Virus associated host–pathogen interactions have been studied in a range of organisms, from humans (Zhu et al., 1998) to Arabidopsis thaliana (L.) Heynh., (Golem and Culver, 2003; Whitham et al., 2003). In this paper we report the use of sweetpotato cDNA microarray technology in an effort to better understand the effect of the synergistic interaction between SPFMV and SPCSV on the host’s response to infection. This study represents the fi rst effort to investigate the effect of SPVD and its viral components on gene expression of sweetpotato. Materials and Methods PLANT MATERIAL AND INOCULATIONS. Ipomoea setosa Ker-Gawl. seedlings mechanically inoculated with SPFMV-RC (isolate 952), and I. batatas cv. Beauregard plants infected with SPCSV (isolate BWFT-3) alone were grown in the greenhouse to generate the scions that were used to graft-inoculate clonally propagated plants of virus-tested [VT plants are tested for presence of viruses by grafting three times to an indicator host, I. setosa] I. batatas cv. Beauregard. Test plants were graft-inoculated 3 weeks after planting. A single wedge graft per virus was performed and individuals on which the scion(s) survived for at least three weeks were selected and used in this study. The experiment consisted of the following four treatments in a randomized complete-block design: VT (not inoculated), SPFMV-RC (VT plants graft inoculated with SPFMV-RC alone), SPCSV (VT plants graft inoculated with SPCSV alone) and SPVD (VT plants graft inoculated with SPFMV-RC and SPCSV simultaneously). Each treatment was replicated six times. Plants were grown under standard greenhouse conditions in 15-cm-diameter clay pots containing autoclaved soil mix consisting of 1 part river silt, 1 part sand, 1 part Jiffy-Mix Plus (Jiffy Products of America, Norwalk, Ohio) and 3.5 g per pot of Osmocote 14N–6.1P–11.6K (Scotts-Sierra Horticultural Products Co., Marysville, Ohio). A weekly insecticide spray program was used to control aphids and whitefl ies. At 9 weeks after inoculation the fi rst four fully opened leaves from the top of each test plant were collected, combined and immediately frozen in liquid nitrogen and stored at –80 °C until extraction. Nine weeks after inoculation was selected as the collection date to ensure better uniformity in virus titers (Kokkinos and Clark, 2006) and symptom development between biological replicates. RNA ISOLATION, LABELING, AND ARRAY HYBRIDIZATION. Total RNA was extracted from six plants of each treatment. After leaf materials were ground with a mortar and pestle in liquid nitrogen, ≈0.8 g were used to extract total RNA using the RNeasy Maxi Kit (Qiagen, Valencia, Calif.) according to the manufacturer’s instructions. The RNA was further cleaned and concentrated by using the clean-up procedure as described in the RNAeasy Mini Kit Manual (Qiagen). During both steps, DNase I digestion was carried out on the column as recommended by the manufacturer. For each sample, 10 μg of total RNA was labeled using the SuperScript Indirect cDNA Labeling System for DNA Microarrays (Invitrogen, Carlsbad, Calif.) according to the manufacturer’s protocol. Samples were labeled with Cy3 or Cy5 fl uorescent labels (Amersham Biosciences, Piscataway, N.J.) and hybridized onto arrays in a connected loop design. (Rosa et al., 2005) using the Pronto hybridization kit (Corning, Life Sciences, Corning, N.Y.). To limit dye effects, the order of the treatments in the loops, as well as the direction of labeling were varied. The order of samples in the loops and the direction of the labeling were different for different loops to ensure that a specifi c comparison in the loop is not always labeled with the same dye and hybridized together on the same array. ARRAY ARCS_SP02/2. The sweetpotato ARCS_SP02/02 array contains 3600 features, spotted in triplicate with a Genemachines Omnigrid microarray printer (GeneMachines, San Carlos, Calif.) on Corning GAPSII slides (Corning Inc.). The arrays were printed and supplied by S. Fluch at ARC Seibersdorf Research GmbH (Biogenetics/Natural Resources, Seibersdorf, Austria). The array contains 2765 features from sweetpotato leaf and sweetpotato storage root libraries as well as control features, including nonplant features, spotting buffer features and blanks. The sequence information for the sweetpotato cDNAs features spotted on the array is available online in GenBank. ARRAY SCANNING, IMAGE QUANTIFICATION, AND STATISTICAL ANALYSIS. Arrays were scanned with an AlphaArray Reader (Alpha Innotech, San Leandro, Calif.) and spots were detected and septbook 658 9/27/06 11:36:57 AM 659 J. AMER. SOC. HORT. SCI. 131(5):657–666. 2006. quantifi ed using UCSF Spot (Jian et al., 2002). After comparing the effects of different normalization methods using MA-plots (the intensity log-ratio, M vs. the mean log intensity) (Dudoit et al., 2002), and spatial image plots, data were normalized within (printtip loess) (Smyth and Speed, 2003) and between slides (scaled). Linear models (Smyth, 2004) were fi tted for comparisons between treatments and genes were considered differentially expressed if P < 0.05 after applying the Holm (1979) multiple testing correction. All normalizations and statistical analyses were carried out using limmaGUI software (Wettenhall and Smyth, 2004). In this study, the output from limmaGUI is in the form of M-values (log2 fold change) (Wettenhall and Smyth, 2004) (Table 1). QUANTITATIVE REAL-TIME POLYMERASE CHAIN REACTION (QRT-PCR). Two-step Q-RT-PCR was carried out for seven genes using RNA from the six VT and six SPVD affected plants. Firststrand cDNA synthesis was carried out using the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen) and the resulting product was diluted by adding 40 μL water. One microliter of the dilution was used for Q-RT-PCR on the ABI PRISM 7000 Sequence Detection System using SYBR Green PCR Master Mix (Applied Biosystems, Foster City, Calif.) and 600 nM of each primer (Table 2) in a fi nal volume of 25 μL. The following PCR protocol was followed: 95 oC for 10 min, followed by 40 cycles of 95 °C for 15 s and 55 °C for 1 min. Amplifi cations from 18S ribosomal RNA specifi c primers (Applied Biosystems) were used to normalize data and dissociation curves were used to detect nonspecifi c amplifi cation. Signifi cant differences (P < 0.05) between treatments were determined using a t test (variances not assumed equal) of normalized values. FUNCTIONAL CLASSIFICATION OF GENES. Gene descriptions were obtained by comparison of sequences to GenBank and A. thaliana protein sequences (TIGR) (BLASTX E-value < 1E-5). Functional classifi cation of genes in Table 1 was based on information from the Munich Information Center for Protein Sequences (Schoof et al., 2002). Results and Discussion The number of genes differentially expressed between VT plants and the three treatments varied. Between VT and SPFMVRC, and VT and SPCSV, only 3 and 14 genes were differentially expressed, respectively, compared to 216 between VT and SPVD (Table 1). The number of differentially expressed genes was analogous to the severity of symptoms observed in the three viral treatments. At the time leaf samples were collected from SPFMV-RC-infected plants, and throughout the time period between inoculation and sample collection, no symptoms were observed, typical of single potyvirus infections (presence of the virus was confi rmed by grafting of scions from test plants to I. setosa). Symptoms of SPCSV-infected plants at the time of collection however, were distinct and characteristic of SPCSV single infections and included interveinal chlorosis and mild purpling. As expected, the most severe symptoms were observed with SPVDaffected plants, which exhibited vein clearing, leaf distortion, chlorosis, puckering, and overall stunting. When comparing VT plants and plants infected with SPCSV alone, only 3 of the 14 differentially expressed genes were suppressed by SPCSV. One of these genes, plastocyanin, was suppressed in all virus-infected treatments. Of the 216 genes differentially expressed between VT and SPVD affected plants, 93 genes were induced in SPVD and 123 suppressed. Many of the genes suppressed in SPVD affected plants are related to photosynthesis and metabolism. Of the induced genes many are involved in protein synthesis and protein fate (Table 1). Q-RT-PCR analysis was carried out for seven genes determined to be differentially expressed between VT and SPVD affected plants by microarray analysis. The results indicated that all seven genes were also signifi cantly differentially expressed (P < 0.05) using Q-RT-PCR with comparable fold changes (Table 3). This reinforces our assumptions regarding signifi cant differential expression based on limmaGUI analyses. During their infection cycles, viruses need plant proteins for accumulation and movement. Gene expression in the host is affected by virus infection. The host plant can respond to an infection by activating specifi c or general resistance pathways (Whitham et al., 2003). By determining which genes are differentially expressed in the host during infection, we hope to elucidate how the response of sweetpotato plants to dual infections of SPFMV and SPCSV differs from response to single infections. The reduction of expression levels of genes that are directly or indirectly involved in the overall photosynthetic pathway, clearly observed in the SPVD-affected plants in this study, is a phenomenon commonly observed in yellows diseases and leaves of plants showing typical chlorotic or mosaic symptoms as a result of virus infection (Hull, 2002). Our data support previous reports, which indicate that the reduction in photosynthesis, observed in virus infected plants, is correlated with the reduction of photosynthetic pigments, rubisco, and specifi c proteins associated with photosystem II (Naidu et al., 1986; van Kooten et al., 1990) and reduced activity of the crassulacean acid metabolism (CAM) (Izaguirre-Mayoral et al., 1993). As expected, the effect on expression levels of “photosynthetic” genes in plants infected with either SPFMV or SPCSV alone was minimal since these viruses, when infecting this particular sweetpotato cultivar alone, cause mild and transient symptoms. Plant resistance genes (R genes) are able to recognize pathogens carrying the corresponding avirulence genes (gene-for-gene resistance). This recognition triggers the hypersensitive response (HR), which includes programmed cell death (PCD). The HR is often preceded by the accumulation and production of reactive oxygen species (ROS), including hydrogen peroxide (H2O2) (Glazebrook, 2001). Several genes, which were differentially expressed only in plants affected by SPVD, were identifi ed as resistance-related or stress-induced genes. Interestingly, some of these genes were down-regulated whereas others were up-regulated. Two putative R genes, one belonging to the TIR-NBS-LRR class (DV036322) and the other belonging to the CC-NBS-LRR class (DV035471) were induced in SPVD affected plants. A NDR1/HIN1-like (CB330891) gene, known to be required by most CC-NBS-LRR class resistance genes in A. thaliana (Aarts et al., 1998) was also induced in SPVD. DV036322 shows homology to At5g17680.1 of A. thaliana, while DV035471 is homologous to At1g58602.1. These genes are similar to ones which encode known disease resistance proteins rpp8 and RPP1-WsB, respectively. To our knowledge, no R genes have been reported, nor is there previous evidence for gene-for-gene resistance in sweetpotato. It is probable that these two genes play some other role in sweetpotato, possibly in apoptosis or ATP-binding. One of the genes found to be down-regulated in SPVD, encodes a product belonging to the ankyrin repeat-containing protein family (DV036499). In transformed A. thaliana, an ankyrin repeat-containing protein was found to be directly associated with the oxidative metabolism of the host’s resistance to disease and stress response (Yan et al., 2002). The down-regulation of ankyrin septbook 659 9/27/06 11:37:04 AM 660 J. AMER. SOC. HORT. SCI. 131(5):657–666. 2006. S13 precursor DV034886 40S ribosomal protein S3 (RPS3C) 1E-104 -0.79 -0.69 -0.71 DV037420 40S ribosomal protein S10 (RPS10C) 1E-48 -0.50 -0.52 -0.44 DV037214 60S ribosomal protein L13A (RPL13aB) 1E-107 -1.08 -1.14 CB330735 60S ribosomal protein L26 (RPL26A) 1E-48 -0.74 -0.73 -0.68 DV036489 60S ribosomal protein L31 (RPL31A) 9E-40 0.37 0.39 0.41 CB330088 60S ribosomal protein L36a/L44 2E-45 -0.80 -0.73 -0.74 CB330146 elongation factor 1B-gamma, putative / eEF-1B gamma, putative 2E-48 0.65 0.59 0.48 CB329890 eukaryotic translation initiation factor 2B family protein / eIF-2B family protein 1E-136 0.54 0.48 CB330048 cyclophilin-type family protein 5E-42 -0.61 -0.66 -0.63 CB330102 polyubiquitin (UBQ10) (SEN3) senescence-associated protein 2E-77 -0.53 -0.77 -0.52 CB330070 subtilase family protein 1E-35 0.77 0.63 0.54