Asymmetric Interaction between Aphis spiraecola and Toxoptera citricida on Sweet Orange Induced by Pre-Infestation

Indirect interactions between herbivorous insects that share the same host have been focused on insects feeding on herbaceous plants, while few studies investigate similar interactions on woody plants. We investigated performance and feeding behavior of two citrus aphids, Aphis spiraecola Patch and Toxoptera citricida Kirkaldy, on sweet orange as affected by prior infestation of conspecifics and heterospecifics. Results showed that pre-infestation-induced interactions between A. spiraecola and T. citricida were asymmetric, with A. spiraecola gaining more fitness. In detail, pre-infestation by A. spiraecola decreased adult weight, enhanced survival rate and accelerated phloem sap acceptance of conspecifics. However, A. spiraecola pre-infestation did not affect performance or feeding behavior of T. citricida. In another infestation sequence, the pre-infestation of T. citricida did not affect conspecifics, but positively affected heterospecifics, indicated as a decreased pre-reproductive period, enhanced survival rate, adult weight, fecundity, and feeding efficiency, i.e., faster access and acceptance of phloem sap, and longer phloem sap ingestion duration. Furthermore, we found A. spiraecola pre-infestation enhanced amino acid concentration, amino acid to sugar ratio, activated salicylic acid and jasmonic acid marker gene expression, while T. citricida pre-infestation only depressed jasmonic acid marker gene expression. Changes in nutrient and phytohormone-dependent defense probably underlie the asymmetric effect.


Introduction
Infestation by herbivorous insects can indirectly affect the performance of subsequent herbivores positively, negatively, or neutrally, depending on the insect species, duration, and intensity of infestation [1,2]. In response to insect attack, plants show morphological and physiological changes [3]. These changes in plant quality may affect host selection, survival, fecundity, and population dynamics of any subsequent infested herbivores [4,5]. Therefore, these plant-mediated interactions may influence the population dynamics and community structure of herbivorous insects [6].
The performance of herbivorous insects can be affected by pre-infestation of both conspecific and heterospecific species. Many studies have measured plant-mediated inter-specific interactions between two herbivore species by transferring one insect species onto plants, and then measuring the survival, growth or insect number of the subsequently infested heterospecifics [7,8]. Similarly, the intra-specific interaction was detected by transferring insects onto plants and then measuring performance of subsequently infested conspecifics. The plant-mediated intra-specific effect may oppose or synergize that of the inter-specific effect. For example, conspecific pre-infestation increased the number of

Aphid Pre-Infestation Procedure
When shoots were approximately 2 cm long, 10 A. spiraecola or T. citricida fourth instar nymphs were transferred to a single new shoot and restricted by nylon mesh bags. After 48 h, the aphids were removed using a soft paint bush. Control shoots without aphids were similarly caged with nylon mesh bags for 48 h. Thus, there are three treatments (A. spiraecola pre-infestation, T. citricida pre-infestation, and control) in this study. The treated shoots were then infested with A. spiraecola or T. citricida to investigate plant-mediated indirect intra-and inter-specific effects. The number of treatments, aphid performance bioassay, and aphid feeding behavior analysis were described in the following sections.
Additional pre-infested and control shoots (five shoots per treatment) were excised, frozen in liquid nitrogen and stored at −80 • C for later analysis of amino acid concentration, sugar concentration, and gene expression.

Aphid Performance Bioassay
Several life history parameters, including survival rate, pre-reproductive period, fecundity, and adult weight were evaluated to determine the effects of pre-infestation on performance of subsequent aphids on the pre-infested and control shoots. To determine survival, seven newly emerged nymphs were introduced to each shoot and survival was recorded daily until the last nymph molting (six days). Survival rate (%) was calculated as the number of survived aphids divided by the number of total introduced aphids. After survival determination, the number of aphids on each shoot was adjusted to two to assess fecundity. Aphids were checked every 12 h to record the pre-reproductive period (days), and then every 24 h to record cumulative number of nymphs produced until adult aphids were 14 days old (early fecundity). Nymphs were removed daily to prevent populations increasing and facilitate counting. For the adult weight determination, six newly emerged nymphs were transferred to each pre-treated shoot. After 8 days, when aphids entered adult stage, their weights (mg) were measured individually with a microbalance (precision 0.1 mg). Aphid responses were measured from 15 shoots per treatment for each aphid species, and average data from one shoot was considered as a replicate.

Aphid Feeding Behavior
The electrical penetration graph (EPG) method can record the activity and locations of aphid stylets [28]. The feeding behaviors of A. spiraecola and T. citricida were studied according to the methods we previously used [29]. Apterous adults were starved for 8 h before experiments. Subsequently, a gold wire was attached to the dorsal side using conductive silver glue and connected to a copper extension wire inserted to an electro penetration graph (EPG) headstage amplifier. Another copper electrode was inserted into the soil. Individual aphids were placed on a shoot, and plants were placed in Faraday cages and subject to EPG monitoring over an 8 h period. The waveforms produced from EPG system included non-penetration (Np); pooled pathway phase activities (C), intracellular puncture (Pd), salivary secretion into sieve elements (E1), passive phloem ingestion (E2), and xylem absorption (G). Both con-and heterospecific pre-infestation treatments and control plants were included in the study, with 19-23 replicates per treatment.
The recorded EPGs were analyzed according to the waveform type and duration. A total of 12 parameters were analyzed. Six parameters are the total duration of the six described waveforms, another six parameters related to resistant against aphids are as follows [30]: (1) One parameter related to surface resistance is the time from onset of EPG recording to first probe (Pd); (2) Two parameters related to epidermis/mesophyll resistance and mesophyll/phloem resistance are the number of probes before first E1 and the minimum duration of C before first E1, respectively; (3) Two parameters reflecting the ease that aphids establish phloem access and acceptance are the time to first E1, and time to first sustained E2 from onset of EPG recording, respectively; (4) One parameter reflecting plant suitability is the average period of E2 (i.e., total time in E2 divided by the number of E2 waveforms per aphid).

Total Amino Acid and Soluble Sugar Concentration
Citrus leaves collected from five shoots per treatment (from 2.2) were assessed for total amino acid concentration. Leaves were used based on their similar relative amino acid concentration to that assessed from the phloem sap [31,32]. For each sample, 0.08 g leaf tissue was homogenized in a 0.72-mL phosphate buffer saline (0.01 mol/L, pH = 7) and centrifuged at 3500 rpm for 10 min. The supernatants were assessed using standard amino acid assay procedures of the total amino acid assay kit (Nanjing Jiancheng Bioengineering, Nanjing, China). Soluble sugar concentration was measured by the anthrone method using the plant soluble sugar content test kit from Nanjing Jiancheng Bioengineering. Approximately 0.07 g leaf per sample was homogenized in 1.8 mL distilled water. The mixture was boiled in water for 10 min and centrifuged at 4000 rpm for 10 min. Supernatants were diluted by distilled water and then used to detect sugar concentration according to standard procedures. Citrus leaves were assessed from five shoots per treatment.

Gene Relative Expression Detection
Total leaf RNA was isolated by TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and 1 µg of the RNA was used to synthesis cDNA using the FastQuant RT Kit with gDNase (Tiangen, Beijing, China). The transcript levels of four target genes were analyzed by fluorescent real-time quantitative PCR: Non-expressor of Pathogenesis-Related genes 1 (CtNPR1), a well-known SA receptor; Pathogenesis related protein 1 (CtPR1), a SA marker gene, which encodes the SA inducing PR protein; allene oxide synthase (CtAOS), a gene for JA biosynthesis; and cysteine proteinase inhibitor (CtPI), which works downstream of JA, and encodes the JA inducing proteinase inhibitor. PCR was performed in 20-µL reaction volumes containing 10.4-µL 2× SYBR Premix (Tiangen, Beijing, China), 7.6-µL water, 1-µL gene-specific primers and 1-µL cDNA template. Reactions were carried out on the Mx 3500P detection system (Stratagene, La Jolla, CA, USA). For each biological replicate (4 per treatment), three technical repeats were performed. The transcript changes of the target genes were compared to the reference gene glyceraldehyde-3-phosphate dehydrogenase-1 (CtGAPC1). Primers were designed from gene sequences in NCBI or from published sequences [33,34] (Table 1).

Statistical Analysis
SPSS 20 software (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. Aphids that died or shoots that defoliated during experiment were excluded from analysis. Data for survival rate and cumulative fecundity were analyzed with repeated measures ANOVA. One way ANOVA with Tukey's HSD test was utilized for comparisons among the three treatments (conspecific pre-infestation, heterospecific pre-infestation, and control) for each aphid species. Results were deemed significantly different at p < 0.05. Proportion data associated with aphid survival rate were arcsine square root transformed to meet assumptions of normality and homogeneity of variance before analysis. The original data were used for presentation.

Discussion
Here, we found an asymmetrical interaction on the performance and feeding behavior of two citrus aphids. We observed that A. spiraecola gained enhanced survival rate from conspecific pre-infestation, and gained fitness benefits in survival rate, adult weight, and fecundity from heterospecific pre-infestation. However, T. citricida was not affected by either conspecific or heterospecific pre-infestation. Pre-infestation-induced interaction favors A. spiraecola as a superior competitor, and may influence the population and community of citrus aphid species. spiraecola pre-infested, T. citricida pre-infested, or un-infested sweet orange. Gene expression levels were evaluated relative to CtGAPC. Values represent the mean ± SE (n = 4). Bars with different letters represent significant difference (Tukey's HSD test, p < 0.05).

Discussion
Here, we found an asymmetrical interaction on the performance and feeding behavior of two citrus aphids. We observed that A. spiraecola gained enhanced survival rate from conspecific pre-infestation, and gained fitness benefits in survival rate, adult weight, and fecundity from heterospecific pre-infestation. However, T. citricida was not affected by either conspecific or heterospecific pre-infestation. Pre-infestation-induced interaction favors A. spiraecola as a superior competitor, and may influence the population and community of citrus aphid species.
The reciprocal effects of different herbivore species that share the same host plant have gained much attention. We found a positive effect on A. spiraecola and neutral effect on T. citricida induced by heterospecific pre-infestation. This is different to the most reported detrimental effect on other species, such as whitefly pre-infestation negatively affected performance of M. persicae [35], leaf miners [36,37], and Pieris rapae [38], which may help whitefly become a strong competitor. Here, A. spiraecola gains fitness benefit in the interspecific interaction through promoting performance of itself rather than inhibiting other species. In addition to the performance change, feeding behavior detected by EPG technique allows quantification of aphid response, and may help explain how aphid are affected by pre-infestation [39]. Aphis spiraecola on heterospecific pre-infested plant spent less time on penetration or pathway phase, spent more time on phloem sap ingestion, and gained quicker access and acceptance of phloem sieve elements. These feeding behavior parameters indicate heterospecific pre-infestation enhanced plant susceptibility to A. spiraecola [40,41], which was consistent with enhanced performance. On the other hand, in line with the performance, feeding behavior of T. citricida was also not affected by A. spiraecola pre-infestation. Toxoptera citricida pre-infestation showed a positive effect on A. spiraecola; thus, the management of T. citricida may reduce the fitness of A. spiraecola, indicating additional secondary benefits of controlling T. citricida.
Sap-sucking insects like aphids often aggregate in the feeding site, thus conspecific pre-infestation usually occurs to affect insect performance [42]. Con-specific pre-infestation enhanced the survival rate, decreased adult weight, and decreased the time spent before phloem sap ingestion for A. spiraecola compared with control. It has been shown that A. spiraecola feeding caused leaf curling, which would reduce the space provided for aphids. In addition, the enhanced survival rate caused by conspecific pre-infestation increased the number of aphids. These factors would enhance the density of A. spiraecola, a factor important in producing the winged form of aphids [43,44]. The winged form would help it disperse to more new shoots or plants. Therefore, conspecific pre-infestation is thought to favor A. spiraecola fitness. In contrast, T. citricida pre-infestation did not affect performance or feeding behavior of con-specifics. When considering both the intra-and inter-specific effects, for A. spiraecola, the conspecific pre-infestation effect is less strong than that of heterospecific (only the enhanced survival rate vs enhanced performance of all the tested life history parameters), while for T. citricida, the conspecific and heterospecific effect did not differ (both are unchanged). Thus, the pre-infestation may intensify intermore than intra-specific competition between citrus aphids. Besides, the conspecific and heterospecific pre-infestation induced a positive synergistic effect on A. spiraecola, which may help explain why it has become a dominant species in the citrus groves in America [23].
Herbivore-induced changes in host plant morphology, nutrition, defense, or some combination of these changes mediated an alteration of the performance of subsequent infested insects [45][46][47]. In our study, T. citricida pre-infestation did not affect SA signaling marker genes expression, but depressed those involved in JA signaling. As the JA-dependent defense is considered effective in conferring resistance against phloem-sucking insects [48], the depressed defense may benefit subsequently infested aphids. Concurrently, we observed that A. spiraecola feeding on heterospecific infested plants had better performance and enhanced feeding efficiency. Particularly, the less time before reaching phloem sap and passive phloem ingestion also indicate the decreased mesophyll/phloem resistance it encountered [30]. Interestingly, plants pre-infested with A. spiraecola showed increased SA and JA signaling gene expression, but did not affect the performance of subsequent infested T. citricida, indicating that other factors such as nutrition change may involve. Amino acid concentration and amino acid to sugar ratio are considered as an index of host plant nutrient quality for aphids [18]. Positive correlations between growth, reproduction and plant amino acid concentration were established in Rhopalosiphum insertum and M. persicae [9,49]. The infestation of A. spiraecola enhanced the amino acid concentration and amino acid to sugar ratio. We hypothesis that the beneficial effect of enhanced nutrition counteract the detrimental effect of induced defense, which results in unchanged performance of T. citricida by heterospecific pre-infestation. For the conspecific interaction, even feeding on plant with enhanced phytohormone-dependent defense, A. spiraecola feeding behavior reflects the fact that it encountered a decreased mesophyll/phloem resistance. It is possible that A. spiraecola can overcome the induced defense, and the decreased adult weight reflects energy cost of detoxification [50]. Therefore, both pre-infestation-induced changes in defenses and nutrition were involved in the indirect interaction between citrus aphids. However, although T. citricida repressed JA defense, the performance and feeding behavior of subsequent infested conspecific is not affected, may be T. citricida is less likely to be affected by plant defense through the long-term co-evolution with host plant. The quality of the plant after insect feeding on a pre-infested plant will help fully explain pre-infestation effect on conspecifics and heterospecifics.

Conclusions
To our knowledge, this is the first study to investigate plant-mediated effects on performance and feeding behavior of different citrus aphids. A. spiraecola and T. citricida show asymmetric interaction induced by pre-infestation. Particularly, T. citricida pre-infestation caused A. spiraecola to gain more fitness. Furthermore, pre-infestation-induced changes in phytohormone-dependent defense and nutritional quality probably underlie the asymmetric interaction. Moreover, the two citrus aphids manipulate host plant physiology in distinct ways, which may relate to the different adaptive strategy between oligophagous and polyphagous insects through the long term co-evolution with host plant [51].

Conflicts of Interest:
The authors declare no conflict of interest.