The Endosymbiont Hamiltonella Increases the Growth Rate of Its Host Bemisia tabaci during Periods of Nutritional Stress

The whitefly Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) harbors several bacterial symbionts. Among the secondary (facultative) symbionts, Hamiltonella has high prevalence and high infection frequencies, suggesting that it may be important for the biology and ecology of its hosts. Previous reports indicated that Hamiltonella increases whitefly fitness and, based on the complete sequencing of its genome, may have the ability to synthesize cofactors and amino acids that are required by its host but that are not sufficiently synthesized by the host or by the primary endosymbiont, Portiera. Here, we assessed the effects of Hamiltonella infection on the growth of B. tabaci reared on low-, standard-, or high-nitrogen diets. When B. tabaci was reared on a standard-nitrogen diet, no cost or benefit was associated with Hamiltonella infection. But, if we reared whiteflies on low-nitrogen diets, Hamiltonella-infected whiteflies often grew better than uninfected whiteflies. Furthermore, nitrogen levels in field-collected whiteflies indicated that the nutritional conditions in the field were comparable to the low-nitrogen diet in our laboratory experiment. These data suggest that Hamiltonella may play a previously unrecognized role as a nutritional mutualist in B. tabaci.

By supplementing nutrients that are deficient in the diet of the insects, intracellular symbionts have contributed to the evolutionary success of their hosts [15,16]. The most developed nutritional associations are found in the obligatory symbionts (which are also referred to as primary symbionts) that provide essential amino acids and certain vitamins to insects with specialized feeding behaviors such as plant-feeding aphids [17] and blood-feeding tsetse flies [18,19]. Microbial symbionts, however, may also compensate for the nutrient deficiencies of less specialized herbivorous animals [16,20]. For example, the intracellular endosymbionts Blochmannia in carpenter ants can provide essential amino acids that are deficient in the ant diet and that thereby sustain colony fitness [21]. Similarly, cockroaches harbor intracellular bacteria (Blattabacterium sp.), which apparently recycle uric acid reserves, providing the insect with usable nitrogenous compounds during times of nitrogen famine [22][23][24]. The latter study demonstrates that a symbiont's contribution may only be apparent when the host is nutritionally compromised.
The current research concerns bacterial symbionts in the whitefly Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae). B. tabaci is a cryptic species complex of major agricultural pests that affect a wide range of crop species [25,26]. As phloem-feeders, whiteflies feed exclusively on plant sap, which is generally limited in nitrogen content [27,28]. While plant sap easily fulfills the daily energetic needs by providing ample carbohydrates [29], plant sap is unlikely to satisfy the nitrogen requirements of reproductive B. tabaci females [15,30]. Like many other sap-feeding insects, whiteflies have evolved ancient relationships with intracellular bacteria that reside within a brightly pigmented abdominal organ known as the bacteriome [15]. Such bacteria commonly provide essential nutrients by the synthesis of essential amino acids missing from the diet [15,30]. Recent sequencing has revealed that the obligate symbiont of B. tabaci, ''Candidatus Portiera aleyrodidarum'', has a very small genome that can synthesize certain essential amino acids as well as carotenoids [31,32] but not other essential amino acids or cofactors [33]. The essential amino acids and cofactors not provided by Portiera, however, might be provided by other symbionts of B. tabaci including the facultative (secondary) endosymbiont, Hamiltonella. Recent results of genome sequencing indicate that Hamiltonella may be able to synthesize amino acids and cofactors [34]. Other recent research indicates that Hamilto-nella increases whitefly fitness [35]. Together, the latter two studies are consistent with the hypothesis that Hamiltonella has nutritional roles in B. tabaci.
Although whitefly hosts are not dependent upon Hamiltonella for amino acid biosynthesis, the bacterium could supplement host stores of amino acids or play a role in nitrogen homeostasis. Here, we examine how Hamiltonella infection alters the growth of B. tabaci when the whitefly is supplied with low, standard, or high levels of nitrogen in its diet. Our goal is to test the hypothesis that Hamiltonella may function as a nutritional mutualist in whiteflies.

Ethics Statement
The research complies with all laws of the country (China) in which it was performed and was approved by the 'Department of Scientific Management of Chinese Academy of Agricultural Sciences, China' (permit number: 20090112). The authority responsible for a national park or other protected area of land or sea, the relevant regulatory body concerned with protection of wildlife. The four field populations in our study were farmlands and were all in the permitted range. The field studies did not involve endangered or protected species.

Whitefly Colony
The laboratory population of B. tabaci Q (recently termed the Mediterranean species) used in this study was originally collected on poinsettia in Beijing, China, in 2009 and has been reared on poinsettia (Euphorbia pulcherrima Wild. ex Klotz.) at 2662uC with a 12-h light/12-h dark photoperiod in individual insect-proof cages. Periodic diagnostic screening revealed that this population harbored only Portiera and Hamiltonella [12]. In our previous article, rifampicin treatments were performed to generate a genetically identical whitefly strain that dramatically reduced the Hamiltonella numbers, while the primary symbiont Portiera had an obligatory relationship with B. tabaci and could not be removed by antibiotic treatment [12]. With the possibility of the Hamiltonella recover, prior the experiment, the adults had received rifampicininfused sucrose with 50 mg/ml of rifampicin for 48 h in three successive generations to persistently suppress the Hamiltonella number. In this way, Portiera was maintained in the strains of the whitefly B. tabaci, while Hamiltonella can be completely removed, and qPCR assays demonstrated that most of the adults in the F4 generation were Hamiltonella-free (Fig. S1).
Four B. tabaci field populations from Haidian, Changping, Langfang, and Nankou, respectively were collected near Beijing, China during the 2013 crop season. At each site, three subsamples of whiteflies were collected, with an approximately 500-1000 m distance between each subsample. The whiteflies from three subsamples were combined into one collection per site. At least 100 whiteflies per collection site were preserved in 95% ethanol and stored at 220uC.

Biotype and Symbiont Determination for B. tabaci Field Collection
Genomic DNA extraction from each of 30 individual whiteflies per collection site was performed on individual adult whiteflies as described by White et al. (2009) [37]. The whitefly biotype was determined by the CAPS (cleavage amplified polymorphic sequence) of mtCOI (mitochondrial cytochrome oxidase I) with the restriction endonucleases VspI [38]. The presence of Hamiltonella was determined using a diagnostic PCR protocol according to Pan et al. (2012) [39]. Reactions were performed in 25 mL volume containing 2.5 mL 106PCR Buffer (Mg 2+ Plus), 2 mL dNTP mix (2.5 mM of each nucleotide), 0.5 mL of each primer (10 mM each), and 0.125 mL of TaKaRa Taq (5 U/mL) (TaKaRa Biotechnology (Dalian) Co., Ltd). The PCR cycling conditions for detection of these symbionts were summarized in Pan et al. (2012) [39]. All PCRs included negative and positive controls. The resultant PCR products were electrophoresed on a 2.0% agarose gel in a 0.56TBE buffer and visualized by Gelview staining. Because the obligate symbiont Portiera should be present in all extractions, any samples that failed to amplify Portiera were considered to be of poor quality and discarded.

Whitefly Growth Rate Assays
The effect of Hamiltonella infection and nitrogen supply on the growth rate of B. tabaci Q was assessed using whiteflies that were three generations removed from the last rifampicin treatment. Newly-emerged adult females of B. tabaci were reared on four types of artificial diets. The standard diet contained 30% (w/v) sucrose and 5% yeast extract (YE) solution (Oxoid, Hampshire, England) in distilled water; this diet supports maximum survival of whiteflies [36]. A low-nitrogen diet contained 2% YE and 30% sucrose, and a high-nitrogen diet contained 10% YE and 30% sucrose. Finally, a no-nitrogen diet contained only 30% sucrose. Stock solutions of the diets were prepared under aseptic conditions using doubledistilled water (DDW) and were sterilized by autoclaving.
Female whiteflies (genetically identical and differing only in the presence or absence of Hamiltonella infection) were reared individually on the four diets described in the previous section for 5 days after adult eclosion. The diets were provided in Parafilm-membrane sachets in a feeding chamber [12]. An MT5 microbalance (Mettler) was used to weigh each whitefly to the nearest microgram at the start and at the end of the 5-day feeding period. Whitefly relative growth rate (RGR) was assessed by the formula ln (weight on day 5/weight on day 0)/5. A separate experiment was performed for the high-, low-, and no-nitrogen diet, and in each case the standard diet was included as a control. These experiments were performed three times (trials 1, 2, and 3) with the high-nitrogen, low-nitrogen and no-nitrogen diets, respectively.

Amino Acid Analysis
Laboratory Hamiltonella-infected female whiteflies that were reared on the three types of diets (high, standard, and low nitrogen) or that were collected from the four field sites were subjected to amino acid analysis in batches (50 adult females per batch, six replicate pools per combination of Hamiltonella infection status and diet or per field site). Each batch was homogenized in 0.1 ml of ice-cold 80% ethanol in a glass, hand-held tissue grinder. After centrifugation at 12,000 g for 15 min to remove debris and precipitated protein, the supernatant was retained for subsequent amino acid analysis. The 800 mL aliquots of extract were dried and hydrolysed in 6 mol l 21 HCl at 110uC for 24 h in a sealed ampoule. The hydrolysate was neutralised with NaOH, dried in a Speed-Vac and dissolved in 80% ethanol, and then filtrated

Statistical Analysis
In the three laboratory experiments concerning the effect of Hamiltonella and diet on growth rate, growth rates between Hamiltonella-infected and noninfected B. tabaci Q were compared with independent sample t-tests. This was done separately for each trial and was also done to compare growth rates of infected and noninfected whiteflies on the standard diet (the control) that was included in each trial. Data for amino acid content were subjected to a one-way ANOVA; if the ANOVA was significant, means were compared with a Tukey's HSD test. All data were analyzed using SPSS software package (ver.17, SPSS Inc, USA). Statistical significance was determined at P,0.05. Means and standard errors are reported.

Survey of Biotype and Symbionts for B. tabaci Field Collection
All the 4 field populations were of pure Q biotype and all individuals of the 4 field populations had Portiera and 87% individuals had Hamiltonella (Data not shown).

Whitefly Growth Rate as Affected by Hamiltonella Infection and Nitrogen Level in an Artificial Diet
When B. tabaci females were reared on the standard diet (control), the RGR showed no significant difference between Hamiltonella-infected and noninfected whiteflies in any trials of the three experiments ( Fig. 1-3). When B. tabaci females were reared on diets that contained high levels of nitrogen, Hamiltonella-infected whiteflies showed 45% higher RGR than did noninfected whiteflies in trial 1, but it showed no differences in trial 2 and 3 (Fig. 1). On a low-nitrogen diet, the RGR of Hamiltonella-infected whiteflies was 42% and 65% higher than that of noninfected ones in trial 1 and 3, respectively, but showed no difference in trial 2 (Fig. 2). On a no-nitrogen diet, the RGR of Hamiltonella-infected whiteflies were significant higher than that of noninfected ones in all three trials (Fig. 3). Regardless of diet, RGR was never lower for Hamiltonella-infected than noninfected B. tabaci females (Fig. 1-3).

Total Amino Acid Content in Adult B. tabaci Females Reared in the Laboratory or Collected from the Field
The free amino acid content in the Hamiltonella-infected whiteflies reared in laboratory was positively correlated with the total nitrogen concentration in the diet (Fig. 4). In addition, Hamiltonella did not significantly influence the free amino acid content of the whiteflies reared on standard diets (t-test, p.0.05); the free amino acid content was 21.1461.30 nmol mg 21 whitefly mass with Hamiltonella infection and 25.4761.60 nmol mg 21 whitefly mass without Hamiltonella infection.
The free amino acid content of adult female whiteflies collected from three of the four field sites (Haidian, Changping, and Langfang) was less than or similar to that for whiteflies reared on the low-nitrogen diet in the laboratory (Fig. 4). The free amino acid content was higher in whiteflies at Nankou than at the other three field sites and was similar to that observed for whiteflies reared on the standard diet.

Discussion
Metabolic provisioning of hosts by endosymbionts is common in obligate associations [17,40]. Analysis of the genome sequence of the obligate symbiont of B. tabaci revealed that it lacks metabolic pathways for cofactors and some essential amino acids [33]. These results suggested that the genome of the insect or of the secondary endosymbionts might encode the enzymes needed to synthesize the missing cofactors and amino acids. As shown in another report, these missing metabolic pathways might exist in Hamiltonella [34]. Given the predictions of nitrogen as potential interaction points for Hamiltonella and their hosts, we experimentally determined whether Hamiltonella could influence the nitrogen homeostasis and growth of B. tabaci females. Our results demonstrate that Hamiltonella which naturally infects B. tabaci can act as a nutritional mutualist, i.e., Hamiltonella significantly increased the growth of B. tabaci when the whiteflies were subjected to low nitrogen environments. The results obtained with the low-nitrogen diet seem most relevant because fieldcollected whiteflies contained low amounts of amino acids. Although decreases in dietary nitrogen severely reduced the growth of both infected and noninfected females in our laboratory experiments, the reduction was less for infected females than for noninfected females, suggesting that Hamiltonella might provide protection against nitrogen deficiency. This is the first report of Hamiltonella having a compensatory effect on an insect host during periods of nutritional deficiency.
Another recent study from our laboratory demonstrated that Hamiltonella infection could substantially enhance whitefly performance; when the bacterium was removed by antibiotic treatment, adult development was impaired and reproductive output was reduced [35]. Considering the latter results and those provided in the current report, it seems clear that Hamiltonella infection increases host fitness. Increases in B. tabaci fitness also occur with infection by Rickettsia, which is another vertically transmitted endosymbiotic bacterium [13]. While insects may often benefit from harboring symbionts, they also may experience a cost [41][42][43][44]. The cost can be greater in older animals [45], and in contrast to the results reported here, the costs can sometimes be greater in nutritionally stressed animals [43]. Whereas most of the time they may gain the main benefits from their bacterial companions (e.g. nutrient provisioning and nitrogen recycling) [46][47][48][49].
The supplement or complement the essential nutrients provided by a primary symbiont was also reported for secondary symbiont in sharpshooters [50], cedar aphid [51], and pea aphid [52]. In the case of two endosymbiotic bacteria in the sharp shooter Homalodisca vitripennis, McCutcheon and Moran (2007) [53] showed that coevolution resulted in complementarity of function in that one bacterium encodes for metabolic pathways that are missing in the other and in the insect. Given the results of the current study and given that Hamiltonella inhabits bacteriocytes alongside Portiera [35,54,55], we suspect that the primary endosymbiont Portiera and the second endosymbiont Hamiltonella may perform complementary functions in B. tabaci hosts.
Taken together, our findings suggest that Hamiltonella may produce essential nutrients that are not produced or are insufficiently produced by B. tabaci or Portiera under nutrient stress. If this inference is correct, it could explain the high frequency of Hamiltonella in whitefly populations in China [39,56] and would also suggest that Hamiltonella might be considered an obligate or nearly obligate endosymbiont from the perspective of the whitefly. Hamiltonella spp. is the main bacterium associated with whiteflies [57-60] but does not obligatorily depend on the whitefly hosts for survival [30]. Within the whitefly host, however, this bacterium enjoys a secure and stable environment with an abundance of nutrients and is also provided with a ready means of dispersal. That this bacterium is present in all stages of the B. tabaci life cycle (QS, unpublished data), that at least some elements are transmitted effectively from parents to offspring, and that this bacterium increases host fitness in general and increases the host growth rate under nutrient stress suggests that the association benefits both partners. Figure S1 Portiera and Hamiltonella densities across treatments with antibiotics. To quantify Portiera and Hamiltonella, total DNA was extracted and used for quantitative PCR. The mean number of genome of Portiera and Hamiltonella was given per actin copies.