Arsenophonus Interacts with Buchnera to Improve Growth Performance of Aphids under Amino Acid Stress

ABSTRACT Amino acids play a crucial role in the growth and development of insects. Aphids cannot ingest enough amino acids in plant phloem to meet their requirements, and therefore, they are mainly dependent on the obligate symbiont Buchnera aphidicola to synthesize essential amino acids. Besides Buchnera, aphids may harbor another facultative symbiont, Arsenophonus, which alters the requirement of the cotton-melon aphid Aphis gossypii for amino acid. However, it is unclear how Arsenophonus regulates the requirement. Here, we found that Arsenophonus ameliorated growth performance of A. gossypii on an amino acid-deficient diet. A deficiency in lysine (Lys) or methionine (Met) led to changes in the abundance of Arsenophonus. Arsenophonus suppressed the abundance of Buchnera when aphids were fed a normal amino acid diet, but this suppression was eliminated or reversed when aphids were on a Lys- or Met-deficient diet. The relative abundance of Arsenophonus was positively correlated with that of Buchnera, but neither of them was correlated with the body weight of aphids. The relative expression levels of Lys and Met synthase genes of Buchnera were affected by the interaction between Arsenophonus infections and Buchnera abundance, especially in aphids reared on a Lys- or Met-deficient diet. Arsenophonus coexisted with Buchnera in bacteriocytes, which strengthens the interaction. IMPORTANCE The obligate symbiont Buchnera can synthesize amino acids for aphids. In this study, we found that a facultative symbiont, Arsenophonus, can help improve aphids’ growth performance under amino acid deficiency stress by changing the relative abundance of Buchnera and the expression levels of amino acid synthase genes. This study highlights the interaction between Arsenophonus and Buchnera to ameliorate aphid growth under amino acid stress.

I nsects often harbor several species of symbionts which play important roles in fitness and behavior of host populations (1)(2)(3)(4)(5). Aphids are small sap-sucking insects which depend on symbionts to meet the challenge of low quantity and considerable imbalance of amino acids in food. The obligate symbiont Buchnera aphidicola provides aphids with amino acids and vitamin B that are usually scarce in the phloem of plants (6)(7)(8). Therefore, this symbiont has become indispensable to aphids, having established symbiosis 160 to 280 million years ago (9). Besides the obligatory symbiont, aphids may harbor other facultative endosymbionts, "Candidatus Regiella insecticola," "Candidatus Hamiltonella defensa," Serratia symbiotica, etc., which are not necessary but affect fitness of host populations with regard to characteristics such as heat tolerance (10), resistance to enemies (11), suppression of plant defense (12,13), and nutritional benefits (14). The coinfection of aphids by Buchnera with a facultative symbiont is very common (15)(16)(17)(18)(19)(20).
The interaction between an obligate and a facultative symbiont may occur due to coexistence in a host. It has been found that the facultative endosymbiont Serratia symbiotica could partially compensate for the role of Buchnera to maintain the survival and reproduction of the pea aphid Acyrthosiphon pisum when Buchnera was removed (21), and it might have performed the transition from a facultative endosymbiont to an obligate one in the aphid Cinara cedri (22). The facultative endosymbiont "Ca. Hamiltonella defensa" might improve Sitobion miscanthi fitness via stimulating the proliferation of Buchnera (23).
All cotton-melon aphid Aphis gossypii individuals harbor the obligate symbiont Buchnera, and 30 to 75% of them also harbor the facultative symbiont Arsenophonus (14,16,(24)(25)(26)(27). Arsenophonus is associated with growth performance of aphids. Arsenophonus infections overall benefit the soybean aphid Aphis glycines in population growth, but this benefit is dependent on aphid clone, plant, and time (28). This symbiont is also associated with the host range of the cowpea aphid, Aphis craccivora, and aphids infected with Arsenophonus can use black locust but cannot use alfalfa (29). Arsenophonus infections increase the survival and reproduction of the cotton-melon aphid and alter requirements of aphids for amino acids in their diet (14,16,30). On the other hand, the abundance of the obligate symbiont Buchnera may be affected by the Arsenophonus. It has been found that the Buchnera population was suppressed by the presence of the facultative endosymbiont Rickettsia or S. symbiotica in the pea aphid (21,31). The abundance of Buchnera was reduced at a rate similar to that of Arsenophonus in the soybean aphid Aphis glycines when aphids were raised on a resistant host plant and exposed to a pesticide (32). In the cotton-melon aphid, coinfection with Arsenophonus and "Candidatus Hamiltonella" did not affect the relative abundance of Buchnera (16), but the Buchnera population sizes and Arsenophonus infection rates in aphids on different host plants were significantly different (14,33). Therefore, the relationship between Buchnera and Arsenophonus still needs study.
Aphis gossypii populations commonly exhibit host specialization on different host plants, and cucurbit-and cotton-specialized biotypes exist worldwide (34)(35)(36). Contents of free amino acids in cotton leaves are lower than in cucumber leaves (14,37), indicating that the cotton-and cucurbit-specialized aphids undergo different nutrient stresses. Moreover, the cotton-and cucurbit-specialized aphids are also infected with Arsenophonus at different rates, and the infection rate in aphids on cotton is significantly higher than that in aphids on cucurbits (14,27). Arsenophonus infections alter the requirements of cottonmelon aphids for some specific amino acids (14), which may contribute to the development of host specialization in aphid populations, but how Arsenophonus affects this requirement of aphids is still ambiguous. The effect of a facultative endosymbiont on aphids may be dependent on aphid genotype (18,28). Therefore, as we focused on the amino acid provider Buchnera (6-8), we hypothesized that Arsenophonus ameliorates aphid performance under amino acid stress via interaction with Buchnera and that this role might be dependent on aphid genetic background. In this study, we measured amino acid titers in three genotypes of A. gossypii infected with and cured of Arsenophonus. Then, we examined growth performance of these three genotypes infected with and cured of Arsenophonus on diets with deficient and excessive amino acids and detected the effects of Arsenophonus infections on the relative abundance and gene expression levels of Buchnera related to amino acid synthase. Finally, we detected the location of Arsenophonus and Buchnera in aphids. The results showed that Arsenophonus infections affected the titer of amino acid in aphids and growth performance of aphids under amino acid stress and that these effects were mediated by the interaction between Arsenophonus and Buchnera with regard to endosymbiont abundance and expression level of the amino acid synthase gene.
Arsenophonus affects the expression level of the amino acid synthase gene of Buchnera. The effect of Arsenophonus on the expression level of lysA, a Lys synthase gene, in Buchnera was dependent on Buchnera abundance and amino acid in diet (Fig. 5). When Buchnera was normal without antibiotic treatment (0 h), Arsenophonus infections did not affect the relative expression levels of lysA in the CA2 (t = 0.129, df = 4, P = 0.904), CA3  , and 5Â (D and E) Lys and Met diets and the correlation between relative abundance of Buchnera and Arsenophonus (F) and between relative abundance of Buchnera and body weight of aphids infected with (G) and cured of (H) Arsenophonus. The solid line is the linear regression line, and the dashed lines show the 95% confidence intervals. Differences in the relative abundance of Buchnera in aphids between Ars 1 and Ars 2 lines were analyzed by Student's t test followed by the Bonferroni correction. *, P = 0.05; **, P = 0.01; ns, no significant difference (P = 0.05).
Location of Arsenophonus and Buchnera in aphids. We dissected aphids to collect bacteriocytes. PCR based on specific primers for Buchnera and Arsenophonus detected both the endosymbionts not only in the bodies of the Arsenophonus-infected aphids but also in their bacteriocytes ( Fig. 7A; Fig. S1). Fluorescence in situ hybridization (FISH) also showed the colocalization of Arsenophonus and Buchnera (Fig. 7B to D). Both the experimental results suggest that Arsenophonus coexists in bacteriocytes with Buchnera and that interaction between them may be possible.

DISCUSSION
The facultative endosymbiont Arsenophonus exhibits a beneficial effect on aphid populations, increasing the ability of Aphis craccivora to use the host plant, locust (38), improving the reproduction of Aphis glycines (28), and enhancing the net reproductive rate and intrinsic rate of natural increase of A. gossypii populations (16,30). Here, we found that Arsenophonus infections improved growth performance of A. gossypii under amino acid deficiency stress. Among aphids growing on the Lys-or Met-deficient diet, the Arsenophonus-infected lines were generally heavier than their Arsenophonus-cured counterparts, but on the normal and excess amino acid diets, their body weights were similar, suggesting that Arsenophonus is responsible for aphid growth under amino acid stress. Amino acid in the diet is related to aphid growth. Omission of methionine in the diet led to growth arrest of Myzus persicae, and omission of lysine caused slower growth (39). In some of the aphid genotypes used in the present study, the titers of Arg, Phe, and Trp were higher in the Ars 1 line than in the Ars 2 line, but titers of Lys and Met were lower. The contents of amino acids in cotton and cucumber leaves are different, and those contents are usually lower in cotton than that in cucumber (14). In contrast, infection rates of Arsenophonus in aphids on cotton are higher than those in aphids on cucumber, and moreover, Arsenophonus infections alter the requirements of aphids for specific amino acids (12). Therefore, we suggested that the growth FIG 7 Presence of Buchnera and Arsenophonus in bacteriocytes of CA3 aphids detected using diagnostic PCR based on the Buc1 primer for Buchnera (lanes 1 to 4) and 16S rRNA (lanes 5 to 8) and fbaA (lanes 9 to 12) primers (Table S1) for Arsenophonus in the CA3 line (A) and detection using FISH (B to D). (B) Buchnera; (C) Arsenophonus; (D) merge of images in panels B and C. Lanes 1, 3, 5, 7, 9, and 11, Arsenophonus-infected aphids (Ars 1 ); lanes 2, 4, 6, 8, 10, and 12, Arsenophonus-cured aphids (Ars 2 ). For lanes 1, 2, 5, 6, 9, and 10, DNA samples were collected from an aphid body, and for lanes 3, 4, 7, 8, 11, and 12, DNA samples were collected from bacteriocytes in aphids. The primer based on fbaA is more specific to Arsenophonus than that based on 16S rRNA. regulation of nymphal aphids by Arsenophonus might be associated with the provision of amino acids.
The obligate symbiont Buchnera synthesizes amino acids in aphids (40). Those amino acids-Arg, Lys, Phe, Leu, and Met-are almost absent in cotton leaves (14), but all of them were found in the Ars 1 and Ars 2 aphid lines with a relatively high titer when aphids were fed on cotton leaves. The infection with Arsenophonus did not result in the same effect on the amino acid titer in aphids of different genotypes, suggesting that Arsenophonus does not directly supply amino acids to aphids (41,42). The amino acid stress-dependent growth amelioration of Arsenophonus-infected aphids may be associated with the obligate symbiont Buchnera.
The abundance of an endosymbiont may affect its role in host insects (43)(44)(45). The relative abundance of Buchnera is affected by host aphids, environmental factors, and interaction with other symbionts in hosts (31,33,46,47). Aphids feeding on cucumber plants harbored more Buchnera than those on cotton (33). High temperature inhibited the development of the Buchnera population in aphids (46,48). Coinfections with facultative symbionts increased or decreased the population density of Buchnera depending on the species of facultative symbiont (21,31,49). In this study, the relative abundance of Buchnera was strongly suppressed by Arsenophonus when aphids were reared on the normal amino acid diet, but on the Lys-or Met-deficient diet, the presence of Arsenophonus did not affect the relative abundance of Buchnera. Also, on the original host plant, cotton, Arsenophonus infections did not affect the relative abundance of Buchnera (14). A previous study showed that the fecundity of aphids on the normal amino acid diet used in this study was lower than that on host plant leaves (30). These results show that the normal amino acid diet used in this study is not optimal for the CA2, CA3, and CA4 aphids. The conditioned effect of facultative symbionts on obligate symbionts was also found in the pea aphid, and the facultative symbiont Regiella or "Candidatus Fukatsuia" enhanced the recovery of population density of Buchnera after heat shock (48). The effect of Arsenophonus on the abundance of the amino acid synthesis-related symbiont Buchnera may be a pathway for aphids to deal with amino acid stress in their diet, and consequently, Arsenophonus infections alter the requirements of aphids for some specific amino acids (14) and improve the growth performance of aphids.
However, the effect of Arsenophonus on aphid growth performance was not determined by the relative abundance of endosymbionts. Although Buchnera was positively correlated with Arsenophonus in terms of relative abundance, their abundance was not correlated with the body weight of aphids. A previous study showed that the removal of Arsenophonus delayed the development of immature stages and reduced nymphal survival rates of the date palm hopper, Ommatissus lybicus (50). Elimination of Arsenophonus and "Ca. Hamiltonella" shortened the total life span of A. gossypii (16). The presence of Arsenophonus is beneficial to aphids' development and survival. Buchnera is distributed in specific bacteriocytes in aphids (51). Arsenophonus was also restricted to the bacteriocytes, which the obligatory symbiont "Candidatus Portiera" occupies in whiteflies (2,52,53). Here, we found the coexistence of Arsenophonus and Buchnera in bacteriocytes in A. gossypii. The coexistence and correlation in population dynamics imply that there may be interaction between Arsenophonus and Buchnera.
The interaction between Arsenophonus and Buchnera affected the expression level of the amino acid synthase genes of Buchnera. Arsenophonus infection may affect amino acid synthesis in Buchnera to enhance the adaptability of aphids to amino acid stress in the diet and thus improve growth performance of aphids. The Buchnera genome contains most genes for biosynthesis of essential amino acids, such as Arg, Lys, Met, and Trp (6,40,54,55). Arsenophonus infections affected the relative expression levels of the Lys and Met synthase genes lysA and metE of Buchnera, although this effect was still dependent on aphid genotype. Moreover, these Arsenophonus-mediated changes in gene expression levels were different in aphids reared on the normal and deficient amino acid diets, suggesting that Arsenophonus is likely to play a role for aphids in dealing with amino acid stress. In CA2 aphids, the lysA expression level was significantly higher in the Arsenophonus-infected line than that in its Arsenophonus-cured line when aphids were reared on the Lys-deficient diet. The metE expression level was significantly lower in the Arsenophonus-infected CA3 line than in the Arsenophonus-cured line on the normal amino acid diet, but it became equal on the Met-deficient diet. Moreover, when Buchnera was suppressed by an antibiotic, the effect of Arsenophonus on the gene expression level of amino acid synthases of Buchnera also changed remarkably. These results suggest that Arsenophonus may prompt the change of the amino acid synthesis level of Buchnera when aphids undergo amino acid stress.
The Buchnera genome completely lost various regulatory systems, such as regulatory genes of biosynthesis of essential amino acids, indicating that Buchnera does not respond to environmental changes (55,56). Therefore, the expression regulation of amino acid synthase genes of Buchnera may likely be controlled by hosts or other coinfecting symbionts. The aphid genotype-dependent effects of Arsenophonus on abundance and gene expression level of Buchnera imply that the regulation of amino acid synthase genes may be mediated by host aphids. The Arsenophonus genome does not carry a whole set of amino acid synthase genes (41,42), meaning that this symbiont is unable to supply amino acids for aphids, but it affected the expression level of the amino acid synthase genes of Buchnera and altered aphid performance under amino acid deficiency stress. Therefore, Arsenophonus affected the regulation of Buchnera genes. It is still worth studying whether this effect is a direct one, from Arsenophonus to Buchnera, or an indirect one, from Arsenophonus to the host aphid and then Buchnera.
It has been confirmed that Arsenophonus-mediated growth performance of aphids under amino acid stress concerned the abundance and gene expression level of Buchnera. When aphids undergo amino acid stress, Arsenophonus abundance changes, and this is linked to changes in Buchnera. On the other hand, Arsenophonus infections lead to the change in expression levels of amino acid synthase genes of Buchnera which directly increase or decrease the supply of amino acids for aphids. Arsenophonus improves aphid growth performance under amino acid stress, possibly via interaction with the obligate symbiont Buchnera (Fig. 8). The higher abundance of Buchnera and high expression levels of amino acid synthase genes may result in the higher synthesis of amino acids which aphids require. The finding that the effect of Arsenophonus on expression levels of amino acid synthase genes in the aphids with reduced levels of Buchnera was not the same as that in the aphids with normal levels of Buchnera indicates that the role of Arsenophonus is indeed dependent on the abundance of Buchnera.
Effects of Arsenophonus on growth performance of aphids and the relative abundance and expression levels of amino acid synthase genes of Buchnera are dependent on aphid genotype. The aphid genotype-dependent roles of Arsenophonus may be attributable to the strain of symbiont or to the aphids' requirements for amino acids. It has been found that the S-type Arsenophonus strain significantly decreased the insecticide resistance of the brown planthopper Nilaparvata lugens, but the N-type strain did not (57). In this study, the Arsenophonus strains in the three genotypes of aphids had slightly different fbaA and 16S rRNA gene sequences (30). This small difference may change the role of Arsenophonus in hosts, and it is worth exploring. Notably, we found that the titers of amino acid in three genotypes on cotton leaves were not the same, suggesting that requirements of aphids for amino acids are different. A symbiont which can meet the requirement of aphids in a specific environment may be reserved under natural selection. Due to various requirements of aphids, the significance of Arsenophonus-mediated improvement of growth performance was not wholly similar among different genotypes under a specific amino acid stress. Therefore, Arsenophonus-mediated changes in the population abundance and amino acid synthase gene expression of Buchnera were also dependent on aphid genotype.

MATERIALS AND METHODS
Cotton-melon aphids and host plants. Three genotypes of the cotton-melon aphid, Aphis gossypii (CA2, CA3, and CA4, based on the polymorphisms of six microsatellite loci), which had been collected from cotton and reared in the laboratory 4 years ago were used in this study (14). Aphids of the three genotypes were infected with Arsenophonus, one of nine known facultative symbionts; this was tested frequently using the diagnostic PCR method (14), which confirmed that Arsenophonus was still present. Therefore, these three genotypes are called the Arsenophonus-infected (Ars 1 ) lines. The fbaA and 16S rRNA gene sequences of Arsenophonus in these three genotypes had a high degree of similarity, and only one and four mutations of bases, respectively, were found (30). We added antibiotics (400 mg/mL ampicillin, 200 mg/mL cephalosporin, and 200 mg/mL gentamicin) to the artificial diet to cure the Arsenophonus infection of the three Ars 1 lines 3 years ago and tested them frequently to confirm that the Arsenophonus was still absent recently; thus, the three corresponding Arsenophonus-cured (Ars 2 ) lines were established. The antibiotics used to cure the Arsenophonus infections in the CA2, CA3, and CA4 genotypes did not affect the abundance of Buchnera when aphids were reared on cotton leaves (14). All Ars 1 and Ars 2 aphid lines were reared on cotton leaves at 25°C with a photoperiod of 14 h of light and 10 h of darkness.
Examination of amino acids in aphids. One hundred 6-day-old aphids each of the CA2, CA3, and CA4 genotypes in the Ars 1 and Ars 2 lines reared on cotton leaves were collected, weighed, and put into an Eppendorf micro test (EP) tube. One milliliter of 0.02 N HCl was added, and the aphids were ground. The aphid mixture was kept at 4°C for 6 h and then centrifuged at 14,000 rpm at 4°C for 15 min. A 500-mL portion of supernatant fluid was transferred into a tube and centrifuged at 14,000 rpm at 4°C for 15 min. The supernatant fluid was filtered using a 0.22-mm membrane and used to examine the titers of amino acids. Seventeen amino acids (Fig. 1) were measured using a Hitachi automatic amino acid analyzer (L-8900; Hitachi, Japan) by the method of Tian et al. (14). Three replicates were performed for each genotype in each line.
Measurement of body weight of aphids. To control amino acid intake, an artificial diet (14) was used to rear aphids. Here, two amino acids, lysine (Lys) and methionine (Met), in the diet were controlled, and each amino acid was used at three concentrations, 0, 1, and 5 times the amount in the referenced diet (14). The Ars 1 and Ars 2 aphid lines of the CA2, CA3, and CA4 genotypes were reared using these amino acid-controlled diets to explore the effect of Arsenophonus infections on aphid growth under amino acid stress. Ten newborn aphids were reared on each kind of amino acid-controlled artificial diet using two layers of thin Parafilm (58), and 40 replicates were performed for each aphid line on each kind of artificial diet. One hundred 3-day-old aphids reared on one kind of amino acid diet were collected as a sample to measure their total body weight using a Sartorius BSA 124S-CW (0.1 mg) balance to exhibit the growth performance of aphids. Three or four samples were collected and measured for each aphid line on each amino acid-controlled diet. These 3-day-old aphids were also collected and stored at 280°C for the examination of the relative abundance of Buchnera and relative expression levels of genes using real-time qPCR.
Examination of the relative abundances of Arsenophonus and Buchnera in aphids. The relative abundance of Arsenophonus was examined following the method of Chang et al. (30). Copy numbers of the Arsenophonus ftsK gene and the aphid ef1a gene were detected using qPCR. The ratio of ftsK copy

Arsenophonus Affects Growth of Aphids
Microbiology Spectrum number to that of ef1a was considered the relative abundance of Arsenophonus in aphids. Similarly, the absolute copy numbers of the groEL gene of Buchnera and the ef1a gene of aphids were also examined, and the relative abundance of Buchnera in aphids was estimated using the ratio of the groEL copy number to that of ef1a (14). Five Arsenophonus-infected and Arsenophonus-cured aphids of the CA2, CA3, and CA4 genotypes reared using controlled amino acid diets with 0Â, 1Â, and 5Â amino acids for 3 days were collected as samples during the body weight experiment described above, and their DNA was extracted to examine the relative abundance of Arsenophonus and Buchnera. Three separate biological repeats were performed for the Ars 1 and Ars 2 lines of the genotypes CA2, CA3, and CA4 reared on two kinds of amino acid-controlled diet (Lys and Met). qPCR followed the method of Tian et al. (14) using SYBR premix Ex Taq (TaKaRa) in an ABI 7500 real-time PCR system (Thermo Fisher Scientific) using the ftsK forward primer ftsK-F (59-TCAAGGTGGCGCTGAATCTT-39) and the reverse primer ftsK-R (59-CGGG CTTACCTCTAGCTTTCC-39), with 103% amplification efficiency; the groEL primers GroEL-F (59-GCCATCCA AAGCCGTATTAGT CA-39) and GroEL-R (59-AGTACCGCAACACCACCAGATA-39), with 107% amplification efficiency; and the ef1a primers ef1a-F (59-TCACCATCATTGACG CACCTG-39) and ef1a-R (59-CCAGTACCA GCAGCAACGAT AAG-39), with 108% amplification efficiency (14). The qPCR procedures were 95°C for 0.5 min, followed by 40 cycles of 95°C for 5 s and 60°C for 34 s. Each DNA sample was run in three separate qPCRs as technical repeats. Standard curves (log concentration of DNA on the x axis and PCR cycle number on the y axis) for the ftsK, groEL, and ef1a genes were set up based on the method of Zhang et al. (33). The absolute copy numbers of groEL and ef1a in each sample were computed by the method of Whelan et al. (59). Detection of the expression levels of amino acid synthase genes of Buchnera. Five aphids reared on one amino acid-controlled diet for 3 days were pooled in one RNase-free EP tube, and there were 3 biological repeats of each. Three hundred microliters of TRIzol and zirconia beads was put into the tube and used to grind the aphids. Total RNA of the five aphids was extracted using the TRIzol reagent (TaKaRa, Japan). The integrity of extracted RNA was tested by agarose gel electrophoresis, and its concentration met the standard for qPCR. Then, cDNA synthesis was carried out according to TaKaRa PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Japan). The synthesized cDNA was stored at 280°C for qPCR to detect the relative expression levels of amino acid synthase genes of Buchnera.
Two genes of Buchnera were chosen, lysA and metE, which encode the amino acid synthesis of Lys and Met, respectively (40). Their relative expression levels in the Ars 1 and Ars 2 aphids on the amino-acid-controlled diet were examined using a qPCR method based on two reference genes, the actin gene and ef1a, of aphids (48,60,61). The qPCR was performed in a thermocycler (no. 7500; Applied Biosystems, Carlsbad, CA) according to the instructions for the SYBR premix Ex Taq (TaKaRa). The PCR volume was 20 mL, containing 10 mL SYBR green PCR master mix (TaKaRa), 2 mL cDNA templates, 0.4 mL forward and reverse primers (10 mM), 0.4 mL ROX reference dye II, and 6.8 mL double-distilled H 2 O (Table S1). The qPCR programs were 95°C for 30 s, followed by 40 cycles of 95°C for 5 s and 60°C for 34 s. Each qPCR was performed in three technical replicates. The relative expression level of a gene was calculated using the 2 2DDCT method (62).
To explore the interaction between Arsenophonus and Buchnera, the Ars 1 and Ars 2 lines were reared on an artificial diet with 2 mg/mL rifampicin for 24 h and 48 h to reduce the abundance of Buchnera, and then these aphids were reared on 0Â and 1Â Lys and Met diets for 3 days. The aphid samples were collected to examine the relative expression levels of lysA and metE using qPCR as described above. Three biological repeats were performed for rifampicin and amino acid treatments for CA2, CA3, and CA4.
Location of Arsenophonus and Buchnera. We used diagnostic PCR and FISH to examine the location of Arsenophonus and Buchnera in aphids. DNA samples from an aphid body and 1,000 bacteriocytes collected from aphids according to the method of Douglas and Dixon (63) were extracted. Bacteriocytes were washed three times using absolute alcohol before DNA extraction. DNA from the Arsenophonus-infected and cured CA2, CA3, and CA4 aphids was extracted. The PCR detection of Buchnera used the primers Buc1F and Buc1R, and the detection of Arsenophonus used two pairs of primers based on 16S rRNA and fbaA (Table S1). PCR products were examined using agarose gel electrophoresis to confirm the presence or absence of Buchnera and Arsenophonus in the DNA samples. If both Buchnera and Arsenophonus were detected in bacteriocytes, it indicated that these two endosymbionts coexist in the same bacteriocytes. Buchnera and Arsenophonus were visualized in CA2, CA3, and CA4 aphids using FISH as described previously (64). The fluorochrome-labeled oligonucleotide probes for Buchnera and Arsenophonus were 59-Cy3-CCTCTTTTGGGTAGATCC-39 and 59-Cy5-CCTTAACACCTTCCTCACGAC-39, respectively (21,65,66).
Data analyses. Effects of aphid genotype and Arsenophonus infection on amino acid titers in aphid bodies were analyzed using multivariate analysis of variance (MANOVA) in the general linear model (GLM), and the titers of 17 amino acids were dependent variables. The difference between amino acid titers in the Ars 1 and Ars 2 lines was analyzed using Student's t test. The effect of Arsenophonus infection on body weights of aphids of a given genotype was analyzed using Student's t test followed by the Bonferroni correction between the Ars 1 and Ars 2 lines reared on the diets containing 0Â, 1Â, and 5Â concentrations of one amino acid. The effects of amino acid deficiency or excess and aphid genotype on the relative abundance of Arsenophonus in aphids were analyzed using two-way ANOVA in the GLM, and the differences in the relative abundance of Arsenophonus in an aphid genotype reared on different amino acid-controlled diets were analyzed using ANOVA followed by the post hoc Tukey's test. The effect of aphid genotype and Arsenophonus infection on the relative abundance of Buchnera in aphids on a 1Â amino acid diet was analyzed using the GLM. The difference in the relative abundance of Buchnera between Ars 1 and Ars 2 lines was analyzed using Student's t test followed by the Bonferroni correction. The correlation analyses between the relative abundance of Arsenophonus and Buchnera in all three genotypes (CA2, CA3, and CA4) on diets with deficient (0Â), normal (1Â), and excess (5Â) Lys and Met and between the relative abundance of Arsenophonus or Buchnera and aphid body weight were performed using the Pearson method, and the Ars 1 and Ars 2 lines were analyzed separately when data collected from aphids reared on 0Â, 1Â, and 5Â amino acid diets were pooled. Effects of Arsenophonus infection and Buchnera treatment time by antibiotic on the relative expression levels of amino acid synthase genes of Buchnera in aphids were analyzed using the GLM, and the difference between Ars 1 and Ars 2 lines was checked by Student's t test. All data analyses were performed using IBM SPSS Statistics V25.
ACKNOWLEDGMENT This work was supported by the National Natural Science Foundation of China (grant no. 31672034). The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
We declare no conflict of interest.