Cellular Tropism, Population Dynamics, Host Range and Taxonomic Status of an Aphid Secondary Symbiont, SMLS (Sitobion miscanthi L Type Symbiont)

SMLS (Sitobion miscanthi L type symbiont) is a newly reported aphid secondary symbiont. Phylogenetic evidence from molecular markers indicates that SMLS belongs to the Rickettsiaceae and has a sibling relationship with Orientia tsutsugamushi. A comparative analysis of coxA nucleotide sequences further supports recognition of SMLS as a new genus in the Rickettsiaceae. In situ hybridization reveals that SMLS is housed in both sheath cells and secondary bacteriocytes and it is also detected in aphid hemolymph. The population dynamics of SMLS differ from those of Buchnera aphidicola and titer levels of SMLS increase in older aphids. A survey of 13 other aphids reveals that SMLS only occurs in wheat-associated species.


Introduction
Almost all aphids (Hemiptera: Aphididae) harbor the bacterial endosymbiont Buchnera aphidicola, which supplements essential amino acids lacking in the aphids' restricted diet of phloem sap [1]. The symbiont is harbored in specialized cells called bacteriocytes (or mycetocytes) that form an organ in the aphids' abdominal cavity called the bacteriome [2]. The bacterium is transmitted from mother to offspring with perfect fidelity, and the obligate relationship between B. aphidicola and aphids has been maintained for about 150-250 million years [3,4].
In addition to B. aphidicola, aphids have about 12 vertically transmitted bacteria that are not essential for their survival. Most of these secondary or facultative symbionts are originally reported in Acyrthosiphon pisum [5,6,7,8,9,10,11,12,13,14]. The three main secondary symbionts, Serratia symbiotica (R type), Hamiltonella defensa (T type) and Regiella insecticola (U type) endow aphids with diverse abilities such as resistance to high temperatures [15], parasitoid wasps [16], and fungal pathogens [17], and R. insecticola also can broaden the spectrum of host plants [18]. In contrast, symbiotic Rickettsia and Spiroplasma negatively affect the fitness of A. pisum [10,19,20]. Similar in vivo localizations of S. symbiotica, H. defensa, R. insecticola and Rickettsia occur in embryonic A. pisum; they are housed in sheath cells and secondary bacteriocytes around the primary bacteriocytes that contain B. aphidicola, as well as in aphid hemolymph [20,21].
Recently, a new aphid secondary symbiont, SMLS (Sitobion miscanthi L type symbiont) was detected in Sitobion miscanthi and it probably represented a new genus in the family Rickettsiaceae [14]. Little taxonomic information was extracted from the 16S rRNA sequence of SMLS. In present study, we investigated in vivo localization, population dynamics and host range, and clarified the taxonomic status of SMLS using in situ hybridization along with quantitative and diagnostic PCR techniques.

Ethics Statement
No experiment involving vertebrate samples was performed in this study. An ethics statement is not required for experiments that involve insects only. The collecting of wild aphids was permitted by wheat farmers.

Materials
Aphids examined in this study were listed in Table 1. Previously, SMLS was detected in the population of S. miscanthi ZK collected from wild wheat in Zhoukou with a high frequency of infection (18/22, 81.8%). Rickettsia was detected in populations of S. miscanthi XN collected from wild wheat in Xining with a lower frequency of infection frequency (5/17, 29.4%) [14].
An ex situ SMLS-infected isofemale ZK-strain was built using one individual of S. miscanthi ZK. Aphids were reared on wheat seedlings in the laboratory at 20uC with a light:dark regime of 16:8 hr. Infections of the other three main secondary symbionts of aphids (S. symbiotica, H. defensa and R. insecticola) and the common symbiont of arthropods (Wolbachia pipientis) were tested in both the ZK-strain and Rickettsia-positive samples of the XN-population using 16S rRNA diagnostic PCR [12,22]. None of these symbionts was detected in ZK-strain and only one Rickettsia-positive sample in the XN-population was co-infected with R. insecticola (data not shown). Consequently, the gene amplifications of SMLS and Rickettsia were performed on the single-infection samples.

Molecular phylogenetic analysis
To reveal the phylogenetic position of SMLS within the Rickettsiales, nucleotide sequences of 16S rRNA, gltA and coxA representing the two main families were retrieved from GenBank for the following taxa: family  Sequences were initially aligned using CLUSTAL W as implemented in MEGA 4.0 [25] with the default parameters and then adjusted manually. Bayesian inference (BI) trees were constructed in MRBAYES 3.1.2 [26,27]. The best-fit nucleotide substitution models were selected using jMODELTEST 0.1.1 [28,29] based on Akaike Information Criterion [30]. Two independent runs including four chains were performed with initial 1,000,000 generations, and stopped when the average deviation of split frequencies fell well below 0.01. Trees were sampled every 100 generations and the initial 25% of the total trees were discarded as burn-in. Compatible groups were shown in the majority rule consensus tree. Analyses involved independent gene and the concatenated data. In the latter case, the concatenated data were partitioned as independent gene. The parameters were defined as unlinked and the prior rate was set as variable. Branch support for each node in BI trees was assessed by the frequency of nodal resolution, i.e., a Bayesian posterior probability (BPP).

Fluorescence in situ hybridization
This process was generally performed as described by Koga et al. [31]. Aphid embryos were dissected from adults of the ZKstrain in cold 70% ethanol using the hooked tip of a #0 insect pin (0.3 mm diameter, 40 mm length) under a stereoscopic microscope, and then fixed in Carnoy's solution (chloroform-ethanolacetic acid [6:3:1]) for 10 hr. The fixed embryos were decolorized overnight in alcoholic 6% H 2 O 2 solution, then pre-hybridized in hybridization buffer (20 mM Tris-HCl [pH 8.0], 0.9 M NaCl, 0.01% sodium dodecyl sulfate, 30% formamide) for 3 times at 6 hr each. Embryos were then incubated overnight in hybridization buffer containing 100 pmol/ml of each fluorescent probe and 0.5 mg/ml 49,69-diamino-2-phenylindole (DAPI). Finally, the embryos were washed in a buffer (0.3 M NaCl, 0.03 M sodium citrate, 0.01% sodium dodecyl sulfate) and observed under a laser confocal microscope (LSM 510 META, Carl Zeiss). We designed two fluorescent probes that targeted B. aphidicola and SMLS 16S rRNA molecules in cells from known probes [20]: SMB-Cy5 (59-Cy5-CCTCTTTTGGGTAGATCC-39) for B. aphidicola, and SMLS-Cy3 (59-Cy3-TCCACGTCACCGTATTGC-39) for SMLS. Nuclei of aphid cells were counterstained with DAPI. No-probe and RNase digestion control experiments were employed to confirm the specificity of the detection. All manipulations were performed at room temperature.

SMLS detection in aphid hemolymph
Aphid hemolymph was collected from about 10 adult aphids of the ZK-strain following the methods described by Fukatsu et al. [7]. Bacterial DNA was extracted using the same DNA extraction kit. SMLS was detected with 16S rRNA diagnostic PCR, using the following primers: 16SA1 (59-AGAGTTTGATCMTGGCT-CAG-39) [32] and Ric16SR (59-TCCACGTCACCGTCTTGC-39) [20]. Aphid DNA of the ZK-strain served as the positive control, and sterile water used as the template in the negative control.

Quantitative PCR
DNA was extracted from a series of aphids of the ZK-strain according to days after birth. Titers of SMLS and B. aphidicola were quantified in terms of the gltA gene and molecular chaperone dnaK (dnaK) gene copies, respectively. Quantitative PCR was performed in a Mx3000 (Stratagene, USA) using the SYBR Green I method. Primers were designed by the online program Primer3 with about 98.6% amplification efficiency. Three pairs of primers had high amplification specificity as verified by unique peaks observed in respective melting curves (data not shown). Quantitative PCR reactions were carried out in a 25 ml volume containing 12.5 ml 26TransStart Green qPCR SuperMix UDG (TransGen), 0.5 ml 506Passive Reference Dye, 10 ml sterile water, 0.5 ml of each primer (10 mmol) and 1 ml DNA. Cycling conditions were 50uC for 2 min (UDG enzyme digestion), 94uC for 10 min, followed by 35 cycles at 94uC for 30 s, (55uC for gltA and ef1a, 53uC for dnaK) for 30 s, 72uC for 30 s. Finally, a melting curve was constructed. Standard curves were constructed with serial dilution plasmids, which contained 10 10 , 10 9 , 10 8 , 10 7 , 10 6 and 10 5 copies/ ml of gltA and dnaK, 10 8 , 10 7 , 10 6 , 10 5 and 10 4 copies/ml of ef1a. Sterile water was used as the template in the NTC (no template control).

Diagnostic PCR
Rickettsia and SMLS were detected within diverse species of aphids (Table 1) using 16S rRNA diagnostic PCR. Cycling conditions were the same as those used in the amplification of gltA and coxA. DNA of the aphid ZK-strain was used as the template in the positive control, and sterile water was used as the template in the negative control.

Nucleotide sequence accession numbers
The gltA and coxA sequences of SMLS, and Rickettsia from S. miscanthi, were deposited in GenBank under accession numbers HQ645970-HQ645973. The 16S rRNA sequences of SMLS isolated from Schizaphis graminum and Rhopalosiphum padi were deposited in GenBank under accession numbers JF933898-JF933900.

Amplification and identification of the gltA and coxA sequences
Putative gltA and coxA sequences of SMLS and Rickettsia from S. miscanthi were amplified, cloned and sequenced. A 475 bp fragment (excluding primer sequences) was obtained with primers gltAF3 and gltAR11 for the ZK-strain and it was most similar (70%) to that of R. typhi when searched by Blastn in NCBI (http:// www. ncbi.nih.gov/BLAST/). A 1049 bp fragment was obtained with primers coxAF1 and coxAR1 for the ZK-strain. It was 80% similar to the sequence of O. tsutsugamushi. The primers gltAF11 and gltAR11 amplified a 1023 bp fragment from the DNA of Rickettsia-positive S. miscanthi. The fragment was 97% similar with that of R. bellii. Finally, a 404 bp fragment was obtained from Rickettsia-positive S. miscanthi with primers coxAF4 and coxAR1 and it was most similar (96%) to R. bellii. All sequences were converted into amino acids to confirm translation.

Phylogeny
Using jModelTest, the GTR substitution model with rate variation among sites (+G) was selected for 16S rRNA and coxA, and TIM3+G was selected for gltA. The parameters of these models were estimated in MRBAYES. In addition to BI trees, we also constructed neighbor joining (NJ) trees with Kimura 2-parameter substitution model in MEGA 4.0, and searched maximum parsimony (MP) trees with heuristic method, TBR algorithm in PAUP 4.0b10* [34]. The three methods obtained almost identical topologies for 16S rRNA and coxA. Moreover, BI obtained better resolution than NJ and MP when using gltA and the concatenated data. Herein, we only provided the BI trees ( Figure 1). Monophyly of the Rickettsiales was supported in all analyses, although not highly supported using 16S rRNA alone (BPP,0.5). The monophyly of Rickettsiaceae was supported in all analyses, while the monophyly of Anaplasmataceae was not supported in the coxA gene tree. SMLS usually clustered with O. tsutsugamushi in the Rickettsiaceae. However, SMLS also clustered with the Anaplasmataceae in the gltA tree, in which O. tsutsugamushi was not included. The close affinity between Rickettsia from S. miscanthi and R. bellii was highly supported in all analyses.

SMLS detection in aphid hemolymph
SMLS was detected in hemolymph. No product was amplified in negative controls, removing the possibility of contamination during amplification.

Population dynamics of SMLS and B. aphidicola
The quantitative PCR results (Figure 3) revealed that the population of B. aphidicola ( Figure 3A) increased during nymphal growth, peaked at the 9 day-stage when aphids matured, declined in the active reproduction day-stages (from 9 to11 day-stages), resurged at the 13 day-stage, and declined again in the remaining stages (from 13 to 29 day-stages). When normalized by titers of the host gene (ef1a), the density of B. aphidicola ( Figure 3C) exhibited similar dynamics but declined from 5 to 9 day-stages. The population of SMLS ( Figure 3B) increased from 1 to 13 day-stages, declined from 13 to17 day-stages, then increased again to attain its highest density in the 29 day-stage. When normalized by the titers of host gene ( Figure 3D), the density of SMLS exhibited the same dynamics.

Diagnostic PCR of 16S rRNA
To estimate the incidence of Rickettsia and SMLS infections across species of aphids, 141 samples of 13 species of aphids were subjected to diagnostic PCR for 16S rRNA. Taken together with the sequencing results, none of these aphids appeared to be infected with Rickettsia but SMLS was detected in S. graminum and R. padi with infection rates of 10% and 51.2%, respectively.

Discussion
In the gltA tree, O. tsutsugamushi was not included because the species lost its functional gltA gene [35]. SMLS clustered with the Anaplasmataceae, perhaps due to long branch-attraction or repulsion. Considering the robust supports in phylogenetic analyses of 16S rRNA, coxA and the concatenated data, SMLS most likely belonged to the Rickettsiaceae and had a sibling relationship with O. tsutsugamushi. The high level of sequence divergence (6%) between 16S rRNA from O. tsutsugamushi and SMLS previously indicated that SMLS might best be classified as a new genus [14]. Due to the absence of a coxA gene standard in bacterial classification, divergences of coxA sequences in Rickettsia were used to evaluate those between SMLS and O. tsutsugamushi. All 29 rickettsial coxA sequences were downloaded from GenBank on 25 Jan 2011 (Table S1). The uncorrected p-distance between the coxA sequences of SMLS and O. tsutsugamushi was 0.207, and this was much larger than the largest p-distance within Rickettsia (0.171 for R. bellii vs. R. prowazekii; Table S1). Assuming the divergence in coxA sequences of Rickettsia reflected intrageneric variation in the family Rickettsiaceae, then both coxA and 16S rRNA divergences between SMLS and O. tsutsugamushi reached an intergeneric level.
In situ hybridization revealed that SMLS was housed in two types of embryonic cells-sheath cell and secondary bacterio-   cytes-both of which were located near primary bacteriocytes that contained B. aphidicola. This discovery implied a probable interaction between SMLS and B. aphidicola. Further, SMLS was also detected in hemolymph. The in vivo localizations were very similar to those of other, thoroughly investigated aphid secondary symbionts including S. symbiotica, H. defensa, R. insecticola and Rickettsia [7,12,20,36]. Although speculative, the same mechanisms of infection, proliferation and vertical transmission may be shared by SMLS and the other secondary symbionts. In vivo localizations indicate that aphid secondary symbionts may have identical traits. Herein, ZK-strain aphids are discovered to be infected with SMLS only; no infection of Rickettsia, the other three main secondary symbionts of aphids and W. pipientis is detected. The two controls confirmed the hybridization's specificity. However, the probe target SMLS used in present study was designed referring to the one target Rickettsia, its use in distinguishing Rickettsia and SMLS must be taken with caution.
In general, the population of SMLS and B. aphidicola exhibit different developmental dynamics in their hosts. Buchnera aphidicola provides nutrition essential for aphid survival, particularly for the rapid production of embryos [1]. The population dynamics of B. aphidicola appear to be typical of aphids, as evidenced by patterns in pea aphids [20,37]; the symbiont's density increases during nymphal growth, peaks during the active reproduction of young adults and declines in older stages. The resurgence of B. aphidicola at the 13 day-stage is probably due to the mismatch of rates of proliferation and consumption. When normalized with host gene titers, the density of B. aphidicola declines during the 5 to 9 daystages. Apparently, B. aphidicola's proliferation cannot match the rapid growth of young aphids in those stages. In comparison, the population of SMLS exhibits an increase-decline-increase curve, and the highest density occurs at the 29 day-stage, the last daystage examined herein. Moreover, the same density dynamic is obtained after normalization with host gene titers. Two other aphid secondary symbionts (S. symbiotica and Rickettsia) have population dynamics that differ from that of B. aphidicola [20,37]. Whereas, the infection level of Rickettsia maintains in older aphids, the population of S. symbiotica increases in older aphids and this is coincident with that of SMLS. Thus, whereas B. aphidicola is an obligate symbiont of aphids, the secondary symbiotic relationship of SMLS differs. This difference may drive the divergent population dynamics.
In addition to S. miscanthi, only the two wheat-feeding species (S. graminum and R. padi) among 13 tested species of aphids appear to be infected by SMLS, and no infection is obtained for Rickettsia. All three strains of SMLS have identical 16S rRNA sequences suggesting a recent horizontal transmission among the three wheat-feeding aphids. Secondary symbionts can be transferred between species of aphids [11,12], yet the mechanisms of these interspecific transmissions remains undiscovered [38]. Wolbachia pipientis may be transferred via feeding on plants [39,40]. Because all of the three SMLS-infected aphids feed on wheat, it is possible that either feeding habits or wheat seedlings are responsible for SMLS transmission.
We collected fresh wheat seedlings and those that had been fed to aphids of the ZK-strain and then froze them in liquid nitrogen. Extracted bacterial total DNA was subjected to 16S rRNA diagnostic PCR. SMLS was not detected on either fresh wheat seedlings or those that had been fed to ZK-strain aphids. Thus, wheat seedlings could not be associated with the transmission of SMLS. Another route must have been responsible for the horizontal transmission of SMLS among wheat-feeding aphids.
We could not test whether SMLS specifically infected wheatfeeding aphids only or not. A large-scale survey of SMLS in aphids was not possible, and infection rates of SMLS within host species vary with geography, as documented in R. padi (Table 1) and S. miscanthi [14]. These tests would have required wide-scale sampling, both taxonomically and geographically. Regardless of why, SMLS widely infected wheat-feeding aphids.
In insects, vertically transmitted bacteria promote their transmission either by manipulating their host's reproduction (e.g. W. pipientis) [41], or by increasing the fitness of infected hosts (e.g. S. symbiotica, H. defensa, and R. insecticola) [15,16,17,18]. SMLS is vertically transmitted from mother to offspring with high fidelity, at least under laboratory rearing conditions. Sitobion miscanthi is largely parthenogenetic making it is unlikely that SMLS spreads by manipulating the reproductive systems of S. miscanthi, as W. pipientis does in arthropods. Further studies are required to investigate whether or not SMLS infections increase the fitness of S. miscanthi.

Supporting Information
Table S1 Matrix of uncorrected p-distance of coxA sequences in genus Rickettsia. (DOC)