Insect Molecular Biology (2015) 24(4), 422–431
Shifts in Buchnera aphidicola density in soybean aphids (Aphis glycines) feeding on virus-infected soybean
Bryan J. Cassone*†, Margaret G. Redinbaugh†‡, Anne E. Dorrance† and Andrew P. Michel§
among insects, plants, and plant pathogens influence endosymbiont population dynamics.
*Center for Applied Plant Sciences, The Ohio State University, OARDC, Wooster, OH 44691, USA; †Department of Plant Pathology, The Ohio State University, OARDC, Wooster, OH 44691, USA; ‡USDA, ARS Corn, Soybean and Wheat Quality Research Unit, Wooster, OH 44691, USA; and §Department of Entomology, the Ohio State University, OARDC, Wooster, OH 44691, USA
Keywords: Aphis glycines, Buchnera aphidicola, tetratrophic interactions, endosymbiont.
Abstract Vertically transmitted bacterial symbionts are common in arthropods. Aphids undergo an obligate symbiosis with Buchnera aphidicola, which provides essential amino acids to its host and contributes directly to nymph growth and reproduction. We previously found that newly adult Aphis glycines feeding on soybean infected with the beetletransmitted Bean pod mottle virus (BPMV) had significantly reduced fecundity. We hypothesized that the reduced fecundity was attributable to detrimental impacts of the virus on the aphid microbiome, namely Buchnera. To test this, mRNA sequencing and quantitative real-time PCR were used to assay Buchnera transcript abundance and titre in A. glycines feeding on Soybean mosaic virus-infected, BPMV-infected, and healthy soybean for up to 14 days. Our results indicated that Buchnera density was lower and ultimately suppressed in aphids feeding on virusinfected soybean. While the decreased Buchnera titre may be associated with reduced aphid fecundity, additional mechanisms are probably involved. The present report begins to describe how interactions
First published online 3 April 2015. Correspondence: Bryan J. Cassone, Center for Applied Plant Sciences, Department of Plant Pathology, The Ohio State University, OARDC, Wooster, OH 44691, USA. Tel.: 574-315-1179; e-mail: [email protected]
Introduction Aphids are distributed worldwide and are among the most common insect pests of cultivated plants in temperate zones (Blackman & Eastop, 2000). The key to their ecological success is largely attributed to their phloem sap-based diet, which allows them to exploit a variety of niches with few competitors (Sandstrom & Pettersson, 1994). While phloem is abundant and easily accessible, it represents an unbalanced nutrient source, rich in sugars but poor in essential amino acids (Alkhedir et al., 2013). The ability of aphids to use phloem as their exclusive food source is largely attributable to the establishment of ancient obligate associations with endosymbiont bacteria (Buchner, 1965). The symbiosis between aphids and their vertically transferred, primary endosymbiont, Buchnera aphidicola, is well characterized. The mutualism dates back over 200 million years and neither the insect nor its endosymbiont can survive independently (Baumann et al., 1995). The elimination of Buchnera with antibiotics has been shown to severely debilitate aphid development and reproduction (Prosser & Douglas, 1991; Douglas, 1996). It is thought that Buchnera benefits its aphid host by synthesizing essential amino acids lacking in phloem €ndu €z & Douglas, 2009; sap (Lai et al., 1994; Akman Gu Hansen & Moran, 2011). In turn, the aphid provides Buchnera with nitrogenous substrate that is abundant in phloem sap, including non-essential amino acids required by the symbiont to produce essential amino acids (Douglas, 1998). The interdependence of host and endosymbiont is reflected in the dramatically reduced Buchnera genome size (Charles & Ishikawa, 1999; Gil et al., 2002). Buchnera has not only lost many of the metabolic genes critical for survival in free-living C 2015 The Royal Entomological Society V
Endosymbiont-host-plant interactions bacteria, but it has also lost most of its regulatory elements resulting in the continuous production of essential amino acids (Moran & Degnan, 2006). Virus-mediated molecular, cellular and physiological changes are well documented in infected host plants (Maule, 2007). This includes changes to the phloem sap components, such as the nutritional substances, toxins and miRNAs (Tu et al., 2013). In turn, feeding on virusinfected plants can dramatically alter the fitness of their phloem-sucking insect pests (Rubinstein and Czosnek 1997; Guo et al., 2010; Cassone et al., 2014a; Xu et al., 2014). Given the integral role some symbionts play in nutrient synthesis for their insect hosts, including Buchnera in aphids, modifications to the phloem sap components of virus-infected plants could dramatically impact the insect host symbiont profiles and in turn affect host plant utilization. Of epidemiological interest in soybean-growing regions of North America are the interactions between the invasive soybean aphid, A. glycines, and two prevalent soybean viruses, Soybean mosaic virus (SMV; Gardner & Kendrick, 1921) and Bean pod mottle virus (BPMV; Zaumeyer & Thomas, 1948). Both viruses require an insect vector for efficient transmission to new host plants. SMV is transmitted nonpersistently by A. glycines and at least 30 other aphid species in 15 different genera (Cho & Goodman, 1982; Steinlage et al., 2002; Cui et al., 2011). BPMV is beetle-transmitted, and is not transmitted by A. glycines or any other aphid species (Ghabrial & Schultz, 1983). Our previous work showed that A. glycines feeding on BPMV-infected soybean have dramatically altered cellular processes and significantly reduced early adult fecundity relative to aphids feeding on SMV-infected or healthy plants (Cassone et al., 2014a). While both SMV and BPMV detrimentally impact soybean, the mechanisms underlying the diminished fecundity of aphids feeding on BPMVinfected plants remain unknown. We hypothesize that, because Buchnera is required for normal aphid reproduction, the endosymbiont is adversely affected by aphid feeding on BPMV-infected soybean. Little information is currently available regarding endosymbiont responses to insect host feeding on virusinfected plants. In the present study, we used a combination of mRNA sequencing (mRNA-Seq) and quantitative real-time PCR to explore changes in Buchnera transcript accumulation and titre in A. glycines feeding on BPMVinfected, SMV-infected, and healthy soybean for up to 14 days. Fitness bioassays were also carried out to explore any associations between Buchnera density and fecundity of aphids feeding on virus-infected and healthy plants over a 17-day developmental period. Our results shed new light into how insect–plant–pathogen interactions alter endosymbiont population dynamics. C 2015 The Royal Entomological Society, 24, 422–431 V
Results Gross fecundity of aphids fed on virus-infected and healthy soybean Differences in cumulative fecundity were examined daily among newborn A. glycines fed on SMV-infected, BPMV-infected, and healthy soybean for 17 days using experimental procedures adapted from Cassone et al. (2014a). A. glycines develops through four nymph stages (Li et al., 2000; Wu et al., 2004), which ranged between 5 and 6 days under our controlled growth conditions. In total, 200 aphids were assayed per treatment across 20 biological replicates collected from multiple cohorts. After 8 days, fecundity was significantly lower in aphids fed on BPMV-infected soybean compared with aphids in the other two treatments (from 9 to 17 days; P < 0.001) (Fig. 1). In addition, fecundity of aphids fed on SMV-infected plants was significantly lower than the healthy fed aphids after 12 days [from 13 to 17 days; P < 0.001 (Fig. 1)]. De novo transcript assembly and functional annotation Whole-genome sequence data is available for several strains of B. aphidicola (Tamas et al., 2002; van Ham et al., 2003; Perez-Brocal et al., 2006). Although the whole genome sequence of the Buchnera A. glyines strain is not currently available, transcript data is accumulating (Bai et al., 2010; Liu et al., 2012). Approximately 343 million raw A. glycines reads were retrieved from the National Center for Biotechnology Information (NCBI) short sequence read archive under the accession number SRP031835 (Cassone et al., 2014a). After preprocessing and de novo assembly, the resultant 28,838 contigs were compared with the nr database and
Figure 1. Cumulative fecundity per aphid of newly born Aphis glycines after feeding on Bean pod mottle virus (BPMV)-infected, Soybean mosaic virus (SMV)-infected, or healthy control soybean for 17 days. Fecundity was calculated daily for each treatment/replicate/time point as the total number of offspring per trifoliate and determined per aphid based on number of surviving adults. Error bars indicate standard deviation variation between 20 replicates per treatment/fitness parameter.
B. J. Cassone et al.
genome of the Buchnera Schizaphis graminum strain (greenbug, order Hemiptera, 555 protein-coding genes). This strain was chosen for pairwise comparisons as it belongs to the same tribe (Aphidini) and is the most genetically similar to the soybean aphid Buchnera currently available in NCBI. A total of 147 de novo contigs had a best match to Buchnera in the nr database or, if no match was found, had a significant hit to the Buchnera S. graminum strain (E-value < 10220; Table S1). A total of 27,890 paired-end reads mapped uniquely to the 147 Buchnera contigs. After filtering out contigs with < 25 paired-read mapping counts across treatments, 80 remained. There were three instances in which two contigs interrogated the same Buchnera gene; for each gene, these transcripts were collapsed into a single value. The remaining 77 contigs will hereafter be referred to as ‘transcripts’ and form the dataset for subsequent analyses. The subset of 77 transcripts was functionally annotated using the Multifun classification scheme (Serres & Riley, 2000), which divides cellular functions into 10 major categories and subdivides these into hierarchical groupings. Originally developed on Escherichia coli K12, the scheme was adapted to Buchnera using BuchneraBASE (Prickett et al., 2006). This is a particularly effective annotation tool, as the gene composition of Buchnera is a subset of the K-12 genome, albeit considerably genetically differentiated (Shigenobu et al., 2000; Riley et al., 2006). The majority of transcripts function in metabolism and information transfer (77%; Fig. S1). Furthermore, most of the information transfer transcripts (81%) were protein-related and 45% of metabolism transcripts (n 5 10) were involved in building block biosynthesis (Fig. S1). The functional annotation of all 77 transcripts is shown in Table S2. Over one-quarter of the transcripts encodes proteins that function as chaperonins and biosynthetic enzymes, which was expected since they are found in high protein abundance in Buchnera (Poliakov et al., 2011; Fan et al., 2013). While the number of genes in the Buhnera A. glycines genome is currently unknown and gene composition varies dramatically among strains, the 77 identified in the present study represent 14% of the genome of the fellow Aphidini member, S. graminum (Tamas et al., 2002).
Higher Buchnera transcript abundance in aphids feeding from 4 h to 7 days on healthy soybean only DESeq (P < 0.05) was used to explore temporal differences in Buchnera transcript abundance (i.e transcript accumulation) among the aphids fed on BPMV-infected, SMV-infected, and healthy soybean for 4 h and 7 days. These time points were selected to characterize both the immediate and gradual responses of Buchnera in
aphids feeding on virus-infected plants. At 4 h, most of the transcripts (n 5 67, 90%) accumulated at similar levels across treatments. In healthy fed aphids, 40% (n 5 30) of Buchnera transcripts significantly increased in abundance from 4 h to 7 days. Moreover, nearly all of the nonsignificant transcripts also accumulated at higher levels at 7 days (n 5 44, 94%). By contrast, substantially fewer Buchnera transcripts differentially accumulated from 4 h to 7 days in the virus-fed aphids (SMV: n 5 2; BPMV: n 5 3). One transcript decreased in abundance in both the BPMV and SMV treatments (GroEL) and other three transcripts differentially accumulated in only one virus-infected treatment: nifU increased in the aphids fed on BPMV-infected, whereas amiB and prsA decreased in the SMV treatment. The differential transcript accumulation of GroEL, nifU, and prsA were validated using quantitative real-time PCR (Table S3).
Buchnera transcript accumulation is inhibited in aphids feeding on virus-infected soybean The relative accumulation of four Buchnera transcripts (tktB, gpmA, thrA and htpG) was assessed by quantitative real-time PCR in aphids feeding on BPMV-infected, SMV-infected, and healthy soybean at six time points over a 10-day period. Accumulation was also assessed in aphids that fed for 10 days on soybean of their respective treatment and then transferred to new V2-stage healthy plants for 2 days (10 1 2 days) or 4 days (10 1 4 days). The transcripts were randomly selected from the subset identified by mRNA-Seq as similarly expressed at 4 h and >2-fold higher in the 7 days of healthy treatment relative to virus-infected treatments. Transcripts were chosen based on accumulation patterns across time points and treatments rather than their functional roles because the vast majority of transcripts exhibited a similar pattern. The primer pairs of selected transcripts had PCR efficiencies (E-values) between 1.9 and 2.15. The abundance of all four Buchnera transcripts increased in aphids fed on healthy plants from 1 to 3 days and then decreased slightly at 5 days (Fig. 2). One-tailed independent t-tests (P < 0.05) indicated that this pattern of increase/decrease was significant for all four transcripts. The Buchnera transcript profiles of aphids fed on virus-infected soybeans were relatively uniform, with transcripts exhibiting a small decrease in abundance from 1 to 3 days (P < 0.05, for all but thrA and gpmA of the BPMV treatment). When aphids were transferred from virus-infected plants to fresh, healthy soybean at 10 days, expression levels of all four transcripts increased significantly (P < 0.05), and approached the levels observed in aphids moved from healthy to healthy plants. In contrast, aphids maintained on the same virus-infected plant (or a new plant of the C 2015 The Royal Entomological Society, 24, 422–431 V
Figure 2. Accumulation of Buchnera tktB, gpmA, thrA, and htpG transcripts in Aphis glycines. Aphids were fed on Soybean mosaic virus (SMV)-infected, Bean pod mottle mosaic virus (BPMV)infected, or healthy soybean for 4 h to 10 days (solid lines), and then moved to fresh, healthy soybean for 2 or 4 days (dotted lines). Data presented are for relative log2-fold transcript levels. Error bars indicate standard deviation variation between three replicates per treatment/time.
same virus treatment) did not show increased transcript accumulation (data not shown). Differences in transcript accumulation largely reflect changes in Buchnera titre To determine if the observed changes in transcript abundance during aphid development (at 4 h, 7 days, and 10 1 4 days) reflect mRNA levels or endosymbiont titre, gDNA levels of the Buchnera single-copy gene dnaK were measured relative to the single-copy aphid gene RpL7. For the aphids feeding on healthy plants, ANCOVA analyses of relative threshold cycle (CT) values indicated that dnaK copy number increased from 4 h to 7 days (F 5 206.6, P < 0.0001, R2 5 0.76), but did not differ between 7 days and 10 1 4 days (F 5 0.61, P 5 0.439, R2 5 0.78). For both virus-infected treatments, dnaK copy number did not differ between 4 h and 7 days (SMV, F 5 2.59, P 5 0.121, R2 5 0.77; BPMV, F 5 0.25, P 5 0.620, R2 5 0.73) but increased from 7 days to 10 1 4 days (SMV, F 5 458.1, P < 0.0001, R2 5 0.93; BPMV, F 5 178.5, P < 0.0001, R2 5 0.82). Relative abundance of dnaK in the healthy-fed aphids increased roughly threefold from 4 h to 7 days and twofold from 7 days to 10 1 4 days in the virus-infected aphids (Fig. 3). These temporal changes in dnaK copy number closely mirrored the patterns and magnitudes of transcript accumulation, indicating that the differences among treatments predominately reflect Buchnera titre rather than mRNA levels. Discussion Virus-induced changes to infected plants have been well characterized, and information is increasing for the C 2015 The Royal Entomological Society, 24, 422–431 V
responses of insects feeding on virus-infected plants (Maule, 2007; Brault et al., 2010; Luan et al., 2011; Gotz et al., 2012; Xu et al., 2012; Cassone et al., 2014a, 2014b); however, little is known regarding how these tritrophic interactions affect the insect microbiome. Endosymbiosis with microorganisms is common in insects, with > 10% of species requiring the metabolic capabilities of intracellular bacteria in their nutrition (Baumann et al., 2006). Aphids harbour an obligate symbiosis with
Figure 3. Relative abundance of the Buchnera aphidicola dnaK gene compared to a single-copy aphid host gene for A. glycines. Aphids were fed on Soybean mosaic virus (SMV)-infected, Bean pod mottle mosaic virus (BPMV)-infected, and healthy soybean for 4 h and 7 days, then moved to fresh, healthy soybean for 4 days. At total of 24 aphids were assayed per treatment/time point. P values represent difference between time points for the healthy and virus-infected treatments. *P < 0.05; **P < 0.01; ***P < 0.001. Error bars indicate standard error variation between 24 replicates per treatment/time.
B. J. Cassone et al.
B. aphidicola, which provides its host essential amino acids that are deficient in a phloem-based diet. A major invasive pest of soybean, the soybean aphid, A. glycines, is now established in most North American soybean-growing regions (Ragsdale et al., 2011). In the present study, we explored changes in Buchnera transcript abundance and titre in newborn A. glycines feeding on BPMV-infected, SMV-infected, and healthy soybean for up to 2 weeks. Aphid fecundity on healthy and virus-infected plants was also assayed over a 17day developmental period. As expected, Buchnera titre predominately explained the developmental stage-specific variation in transcript abundance. Consistent with most reduced microbial genomes, Buchnera has lost nearly all mechanisms for regulating the transcription of genes and consequently the expression levels of most genes is stable (Moran & Mira, 2001; Moran et al., 2005; Moran & Degnan, 2006); however, differential expression was detected for a handful of Buchnera transcripts. Transcripts of two genes (prsA and nifU) increased or decreased in abundance in only the BPMV or SMV treatment, suggesting their expression may be altered in response to virusspecific changes to infected plants. Notable was the GroEL gene, which was the only gene showing decreased transcript accumulation with development in aphids feeding on both SMV- and BPMV-infected plants. GroEL (an orthologue of HSP60) is a universally distributed chaperonin that functions in the folding of new and destabilized proteins (Ohtaka et al., 1992). Since the amount of GroEL increased with aphid development on healthy plants in this and a previous study using S. graminum (Baumann et al., 1996), its transcription appears impeded with increased feeding time on virus-infected plants. There is some evidence that GroEL binds virus particles in the haemolymph of their insect vector host, and that this interaction prevents proteolytic degradation of the virions which, in turn, may promote transmission (van den heuvel et al., 1994; Filichkin et al., 1997; Morin et al., 1999; Hogenhout et al., 2000; Gottlieb et al., 2010); however, neither SMV nor BPMV are thought to breach the midgut barrier and circulate in the haemolymph of the aphid. Future studies are needed to determine the mechanisms underlying reductions in GroEL transcription in virus-exposed aphids. Buchnera density was higher in newly adult A. glycines feeding on healthy soybean than in younger aphids. Although Buchnera titre varied across host development among Acyrthosiphon pisum clones (Vogel & Moran, 2011), a similar pattern was reported in Sitobion miscanthi (Li et al., 2011). Increased Buchnera titre was also found between 5- and 7-day-old S. graminum, although distinct stages of development were not defined (Baumann & Baumann, 1994). The higher
density may be attributed to increased genomic copy number of Buchnera during post-embryonic development of insects to adulthood, as detected in A. pisum (Komaki & Ishikawa, 2000); and/or the important role of Buchnera in parthenogenesis (Douglas, 1996). Buchnera are vertically transmitted and embryos represent a large contribution to the biomass of aphids. Aphids initiate embryogenesis as an embryo or larva, and consequently new adults contain multiple embryos of different developmental ages (Bermingham & Wilkinson, 2009). It is conceivable that the increased Buchnera density in new adults reflect increases in embryo titre as the aphids gear up for reproduction. Buchnera density in A. glycine feeding on virusinfected (SMV and BPMV) soybean was relatively stable across the 10-day development period. Accordingly, the spike in Buchnera density between the fourth-instar nymph and adult stages on healthy plants was not found. Only after aphids were transferred to healthy plants did endosymbiont density significantly increase, suggesting aphid feeding on virus-infected soybean suppresses Buchnera population growth and/or aphid development. We initially hypothesized that aphid host feeding on BPMV-infected soybean detrimentally impacted Buchnera, which in turn altered aphid fecundity. Our previous study indicated Buchnera density was lower in aphids fed on both BPMV-infected and SMVinfected soybean but decreased fecundity was only detected in the BPMV treatment (Cassone et al., 2014a); however, by extending the fitness assays to include an additional 10-day period, reductions in fecundity were also evident in the aphids feeding on SMVinfected plants. While the factors governing the temporal differences in the onset of diminished fecundity between the virus-infected treatments are unknown, a potential molecular signature could be observed in the aphid transcriptional profiles of Cassone et al. (2014a). In that study, the downregulation of the transcriptome and associated significant reduction in fecundity were detected considerably earlier in the BPMV treatment. Taken collectively, it is feasible that the lower Buchnera density plays a role in the reduced fecundity of A. glycines feeding on virus-infected plants; however, we suggest additional mechanisms are at play given Buchnera density is similarly suppressed in aphids feeding on BPMV- and SMV-infected soybean for 17 days but fecundity differs, with aphids feeding on BPMV-infected plants requiring longer aphid developmental cues. Antagonistic effects of BPMV and SMV on A. glycines performance have been previously reported (Donaldson & Gratton, 2007). In the field experiments, aphid naturally colonized the virus-infected plants significantly less frequently than healthy plants. Moreover, the laboratory assays indicated aphid population growth on C 2015 The Royal Entomological Society, 24, 422–431 V
Endosymbiont-host-plant interactions BPMV-infected soybean was significantly lower compared to aphids feeding on healthy soybean (data not generated for SMV). The detrimental impacts to aphid performance manifested across several different soybean cultivars and experimental assays, and were also detected for aphids feeding on soybean infected with a different virus (Alfalfa mosaic virus). This suggests that, for this insect–host plant pathosystem, the negative consequences to A. glycines fitness may be a general response to changes in the virus-infected plants. This is one of the first studies to examine how insect host feeding on virus-infected plants impacts endosymbiont population dynamics. Buchnera population growth was impeded in aphids feeding on BPMV- and SMVinfected plants, which may be partially associated with the fecundity of adult aphids. Future molecular, behavioural, and physiological studies are needed to determine how the interactions between insect and virus alter Buchnera population growth, and the biological consequences of such changes. Notably, BPMV is not transmitted by A. glycines and SMV is nonpersistently transmitted; thus, neither virus is thought to cross membrane barriers and reach the aphid bacteriocyte. Consequently, the impacts to Buchnera population growth may not be direct; rather, they might reflect the indirect effects of virus infection on host soybean. Targeted proteomic and metabolomic studies of soybean phloem components (e.g. nutritional substances, toxins, miRNAs) may provide evidence of a causal link between Buchnera density, aphid fitness and the virus-induced changes to host plants. Experimental procedures Aphid colony, soybean and virus maintenance Laboratory assays used a colony of A. glycines biotype 3 established from field collections in Indiana and maintained in a growth chamber at the Ohio Agricultural Research and Development Center, Wooster, OH (Hill et al., 2010). Aphids were maintained on ‘Sloan’ seedlings grown in 15 3 7.5 3 15 cm cages under controlled conditions of 25 C and 75% relative humidity with a 16 h light/8 h dark cycle. The BPMV isolate was maintained in Sloan soybean by serial leaf-rub inoculation with inoculum made from leaves of infected plants (Louie et al., 2000). Inoculum was made by grinding infected leaf tissue into 10 mM KHPO4, pH 7 (1:4 w/v) and then mechanically inoculated onto 1.5-week-old soybean plants. SMV was maintained in soybean through serial transmission by A. glycines. Experimental plants of the virus-infected treatments were generated by serial leaf-rub inoculation 10 days after planting. At 10 days post-inoculation, plants were visually inspected for symptom development on their youngest trifoliate, and virus presence was validated using enzyme-linked immunosorbent assays (Todd et al., 2010). The following day (11 days post-inoculation) experimental plants were used in for subsequent studies. C 2015 The Royal Entomological Society, 24, 422–431 V
Fitness studies The fitness studies were adapted from Cassone et al. (2014a) with the experimental design extended to include an additional 8 days. Ten apterous newborn adult A. glycines (