The Major Antigenic Membrane Protein of “Candidatus Phytoplasma asteris” Selectively Interacts with ATP Synthase and Actin of Leafhopper Vectors

Phytoplasmas, uncultivable phloem-limited phytopathogenic wall-less bacteria, represent a major threat to agriculture worldwide. They are transmitted in a persistent, propagative manner by phloem-sucking Hemipteran insects. Phytoplasma membrane proteins are in direct contact with hosts and are presumably involved in determining vector specificity. Such a role has been proposed for phytoplasma transmembrane proteins encoded by circular extrachromosomal elements, at least one of which is a plasmid. Little is known about the interactions between major phytoplasma antigenic membrane protein (Amp) and insect vector proteins. The aims of our work were to identify vector proteins interacting with Amp and to investigate their role in transmission specificity. In controlled transmission experiments, four Hemipteran species were identified as vectors of “Candidatus Phytoplasma asteris”, the chrysanthemum yellows phytoplasmas (CYP) strain, and three others as non-vectors. Interactions between a labelled (recombinant) CYP Amp and insect proteins were analysed by far Western blots and affinity chromatography. Amp interacted specifically with a few proteins from vector species only. Among Amp-binding vector proteins, actin and both the α and β subunits of ATP synthase were identified by mass spectrometry and Western blots. Immunofluorescence confocal microscopy and Western blots of plasma membrane and mitochondrial fractions confirmed the localisation of ATP synthase, generally known as a mitochondrial protein, in plasma membranes of midgut and salivary gland cells in the vector Euscelidius variegatus. The vector-specific interaction between phytoplasma Amp and insect ATP synthase is demonstrated for the first time, and this work also supports the hypothesis that host actin is involved in the internalization and intracellular motility of phytoplasmas within their vectors. Phytoplasma Amp is hypothesized to play a crucial role in insect transmission specificity.


Introduction
Phytoplasmas, wall-less plant pathogenic bacteria of the Class Mollicutes, infect a wide variety of herbaceous and woody plants, causing significant economic losses on cultivated crops worldwide [1]. Because they have yet to be cultured axenically, phytoplasmas are described as ''Candidatus Phytoplasma spp.'' [2]. Phytoplasma chromosome [3;4;5;6] has several multi-copy genes organized in clusters of potential mobile units (PMUs), thought to be involved in host adaptation [7]. Within infected plants, phytoplasmas are restricted to phloem elements and cause growth disorders, leaf and floral alterations, and abnormal proliferation, often leading to plant death. The pathogenicity mechanisms are still unclear, but some nucleus-targeted virulence factors secreted by phytoplasma cells [8] alter plant metabolism, playing a crucial role in symptom development [9].
Phytoplasmas are transmitted by phloem-feeding insect species within the Order Hemiptera [10]. Phytoplasma transmission is persistent and propagative, involving a latent period in the vector during which ingested phytoplasmas pass from the alimentary canal through the midgut into the haemocoel, finally colonizing salivary gland cells before being transmitted to a new host plant. Vectors remain inoculative for life [11]. Phytoplasmas are usually transmitted by a narrow range of vector species [10], whereas their plant host range is usually broader [12]. Insect vector specificity plays a key role in the epidemiology of several vector-borne pathogens [13]. For this reason, identification of the molecular determinants of vector specificity is crucial to understand the epidemiology of impacting diseases worldwide. Phytoplasma colonization of the vector depends on several biological features, such as insect feeding behavior and plant host range, as well as on molecular interactions between pathogen and vector membrane proteins [11]. Since phytoplasma membrane proteins are in direct contact with host cells they are likely to be involved in specific interactions with them, as is the case for other mollicutes in which adhesion to host cells has been studied.
For example, P1 and other adhesion proteins in Mycoplasma pneumoniae [14;15], as well as other chromosomal [16] and plasmid-encoded [17;18;19] proteins in Spiroplasma citri are involved in interactions with host proteins.
The most abundant (immunodominant) membrane proteins of several phytoplasmas have been sequenced and classified into three types, Imp, IdpA and Amp, based on chromosomal gene organization and membrane anchor structure [20], but the specific functions of these proteins are still unknown. A strong positive selection on the Amp and Imp genes [21;22] is compatible with a role in host pathogen recognition. Amp and Imp are both present in the ''Ca. P. asteris'' genome, but because Amp is expressed at a higher level than Imp it is considered the predominant antigenic protein in this species [21]. A role in species-specific recognition between vector and phytoplasma has been proposed for Amp of ''Ca. P. asteris'', onion yellows strain (OY), based on the in vitro interaction of OY Amp with actin of the vector and co-localization of the phytoplasma with actin filaments in the insect gut [23]. Other components of the complex phytoplasma membrane proteome also may play a role in phytoplasma-vector interactions. In 'Ca. P. asteris', plasmid-encoded transmembrane proteins are expressed to a greater degree in insects than in infected plants [24;7] and probably have a role in adaptation to life within the insect niche and/or in interaction of the phytoplasma membrane with the vector. The latter hypothesis is supported by the lack of an entire plasmid in a non-vector transmissible line of OY, probably due to reductive evolution [25].
Chrysanthemum yellows phytoplasma (CYP), 16SrI-B (''Ca. P. asteris''), is associated with a disease of ornamental plants in northwestern Italy, where Macrosteles quadripunctulatus Kirschbaum and Euscelidius variegatus Kirschbaum are the most important and efficient vectors of this pathogen [26]. We have used CYP as an experimental model due to its high transmission efficiency, the rapid and clearly recognizable symptoms produced in plants and the ease of continuous insect maintenance under controlled conditions. To confirm the interaction of Amp with vector proteins, identify the insect partner molecules in the interaction and explore their role in phytoplasma transmission specificity, we used a recombinant CYP protein expressed in Escherichia coli [27]. Far Western blotting and affinity chromatography showed interactions of CY Amp with several proteins of the vector, among which the a and b subunits of ATP synthase were identified for the first time. No interactions with proteins of nonvector insects were observed. The presence of ATP synthase on the plasma membranes of the midgut and salivary gland cells of the insect vector was demonstrated indicating that this protein may act as a receptor for Amp. This is the first report of such a role in bacteria. Amp also interacted with actin of vector, but not nonvector, insect species, consistent with previous findings for the closely related OY phytoplasma and several of its vector species [23]. These results suggest the presence of a complex network of vector proteins that interact with Amp in the determination of vector specificity.

Results and Discussion
All insect species used in this study were characterized for their ability to acquire and transmit CYP, for the interactions of their proteins with CYP recombinant membrane proteins (CYP antigenic membrane protein, Amp, and arginine transporter, Art) in dot far Western blots and for the capability of their proteins of binding the CYP Amp in affinity chromatography, evaluated by SDS polyacrylamide gel electrophoresis (PAGE) and Western blots of affinity column elution phases (Table 1).

Vector status
M. quadripunctulatus, E. variegatus, Empoasca decipiens Paoli and Scaphoideus titanus Ball are leafhopper vectors of CYP belonging to the family Cicadellidae; the first two species are very efficient [26] while the latter two are poor vectors under experimental conditions [28;29]. The leafhopper Zyginidia pullula (Boheman), and the aphids Aphis craccivora Koch and Myzus persicae (Sulzer), putative non-vectors, were assessed for CYP transmission competence in this work; M. quadripunctulatus always acquired and transmitted CYP with a 100% efficiency in our experimental conditions (Table 1). Z. pullula (Cicadellidae) was used as a phylogenetically closely related species control; aphids were used as phloem feeder controls. All Z. pullula individuals (n = 11) fed on CYP infected plants were PCR-negative for the presence of the phytoplasma and, consistently, no transmission to test plants was obtained with this species. Failure of CYP transmission is explained by lack of phytoplasma acquisition by this parenchyma-feeder leafhopper [30].
Ten of 13 batches of A. craccivora and all 10 batches of M. persicae were PCR-positive after feeding on CYP infected daisy. However, none of the 10 aphid-exposed test plants for each species developed symptoms. One month after inoculation, all aphidexposed plants tested negative in PCR for the presence of CYP. To estimate the phytoplasma titer in the aphids, PCR positive samples (10 batches each of A. craccivora and M. persicae) were analysed by quantitative real time PCR. CYP ranged between 36 and 232 cells/ng of insect DNA in A. craccivora, and from 256 to 1246 in M. persicae. Mean quantities of CYP cells per ng of insect DNA were 93 (6 SE 19) and 643 (6 SE 106) for A. craccivora and M. persicae, respectively. As a comparison, CYP titre in infective vector leafhoppers ranges between 2.0610 4 and 1.0610 5 phytoplasma cells/ng vector DNA, among the different species [26]. Moreover, the non transmitter individuals of the leafhopper E.  variegatus hosted a significantly lower number of CYP cells (averaging 1800 phytoplasma cells/ng insect DNA) compared to transmitter ones [31]. Overall, pathogen titer in aphid samples was significantly lower than in infective M. quadripunctulatus and E. variegatus [26] (One way ANOVA, F = 8.927, P,0,001) and even below the amount found in non transmitter E. variegatus [31]. Phytoplasma acquisition by aphids is not surprising, since they feed on phloem sap, and since three species of apple aphids acquired, but did not transmit, Apple proliferation phytoplasma [32].

Identification of interacting proteins
Far Western blots. To highlight any interaction between CYP fusion Amp (CYPfAmp) and insect proteins, and to observe possible differences in this interaction between vector and nonvector species, leafhopper and aphid native total proteins (serving as the prey) were spotted on a membrane and assayed in dot far Western blot (FWB) experiments, with CYPfAmp as bait ( Figure 1A). Following incubation with CYPfAmp, interaction signals were detected from proteins of all vector leafhopper species (M. quadripunctulatus, Mq; E. variegatus, Ev; E. decipiens, Ed; S. titanus, St) and from the non-vector M. persicae (Mp), but not from the other non-vectors, Z. pullula (Zp) and A. craccivora (Ac). Incubation controls with i) the other CYP membrane-protein fusion (Art), ii) His-tag fusion antigen (His) and iii) buffer only, devoid of bait (K-), did not produce any signal ( Figure 1A; Table 1). Dot blotting is a simple technique to demonstrate protein-protein interactions, but it does not provide information on the number and size of binding partners [33]. To gain clues for further identification of interacting partners, insect total proteins from vector species (Mq, Ev, Ed, St) and M. persicae were then separated onto one dimensional SDS-PAGE (1D FWB) and incubated with CYPfAmp ( Figure 1B). Several proteins of all insect species, with molecular weights ranging between 85 and 20 kDa, interacted with CYPfAmp. In the absence of bait (K-) no interaction signal was detected. When the membrane fraction of insect proteins was assayed in 1D FWB with CYPfAmp ( Figure 1C), only membrane proteins from leafhopper vector species showed interaction signals. In the absence of bait (K-) no specific interaction signal was recorded. A similar pattern of interacting membrane proteins was shared by vector species, showing intense bands of about 50 and 30 kDa.
CYP Amp always interacted with proteins of all vectors, irrespective of the insects' different transmission efficiencies [28;26;29]. No interaction with CYPfAmp was recorded with proteins of the aphid A. craccivora in dot FWB, whereas the interaction signal, always evident with total proteins of the aphid M. persicae, disappeared when membrane fractions were analysed. Therefore, we can speculate that the interacting M. persicae proteins may be cytosolic.
Affinity chromatography. To identify vector proteins interacting with CYP Amp, an affinity chromatography assay was performed. Several CYPfAmp-interacting proteins from the two highly efficient vectors M. quadripunctulatus and E. variegatus, with apparent molecular weights between 25 and 90 kDa, were evident in the elution phases analysed by SDS PAGE (Figure 2A, lanes Mq and Ev). P90, p55, p50 and p42, the most abundant interacting proteins of E. variegatus, were excised from the gel and identified by mass spectrometry. Weak to very weak bands were present in the elution phases of the two poorly efficient vectors E. decipiens and S. titanus (Figure 2A, lanes Ed and St). In the elution phases obtained after loading the column with total extracts of non-vector species, no proteins were retained and observed in SDS-PAGE ( Figure 2A, lanes Zp, Ac, Mp; Table 1). Bands were absent in SDS PAGE of elution phases following loading of vector proteins on control columns covalently linked to His-tag and CYPfArt (not shown).
Complex patterns of vector proteins interacting with CYPfAmp were obtained with FWB and affinity chromatography approaches. Similarly, several proteins interacting with Amp of the OY strain of the ''Ca. P. asteris'' have been reported for Macrosteles striifrons and Hishimonus spp. [23]. Under native conditions, CYP Amp did not interact with proteins from non-vectors ( Figure 2A). Similar results were reported for OYP Amp and two non-vector leafhopper species [23].
Peptide mass fingerprinting and Western blots. Peptide mass fingerprinting analysis identified E. variegatus p55 and p50 as ATP synthase a and b subunits, respectively, and p42 as actin. Three peptides from p55 shared homology with a putative mitochondrial ATP synthase a subunit precursor of Toxoptera citricida, four peptides of p50 matched the ATP synthase b subunit from Drosophila melanogaster, and four peptides of p42 had homology with actin from Culex pipiens pipiens (Tables 2 and 3). Peptides from p90 did not match any known protein, and therefore p90 could not be identified.
BLAST based analysis of the DNA sequences of the partial E. variegatus actin (HQ451984) and ATP synthase b (HQ451985) genes confirmed the MS/MS identification, and all the peptides from mass spectrometry analysis were identified in the corresponding deduced amino acid sequences. Conserved domains were found in the former sequence, confirming that the deduced protein was actin of E. variegatus. ClustalW alignment showed 98% similarity with the actin of the aphid Acyrthosiphon pisum (ACYPI006035), with one conserved and two semi-conserved substitutions. Conserved domains were found in the deduced ATP synthase b subunit protein of E. variegatus. ClustalW alignment showed 95% similarity with the ATP synthase b subunit of the aphid A. pisum (NP_001119645), with five conserved, three semiconserved and four not conserved substitutions ( Figure 3). PredictProtein analysis of the two ATP synthase b subunit sequences identified a phosphorylation site unique in the E. variegatus sequence, a disulphide bond site (aa 122) present only in A. pisum, and different profiles of protein-protein binding sites between the two insect species (Figure 3).
To further confirm protein identification, Western blot (WB) assays were carried out with actin and ATP synthase b subunit specific antibodies. The ability of these sera to detect their antigens from all the insects included in this study was preliminarly assessed: no evident difference was observed in the molecular masses of the expected proteins among the different species (data not shown).
WB analysis was performed on eluates from different phases of CYPfAmp affinity columns ( Figure 2B). In the elution phases of all vectors (Mq, Ev, Ed, St), the protein p42 was recognized by the antibody against actin, and p50 by the antibody against the ATP synthase b subunit. Neither actin nor the ATP synthase b subunit were detected in the eluate of the non-vector species Z. pullula.
Lack of leafhopper sequence information hampered the identification of CYPfAmp-interacting p90.
Actin, and the frequently associated molecule myosin, are constitutive cytoskeleton proteins involved in several physiological processes of the cell including movement, cell shape, intracellular motility [34], phagocytosis and endocytosis [35]. Actin and myosin of leafhopper vectors interacted with Amp of the ''Ca. P. asteris'' OY strain, a phytoplasma closely related to CYP, and their role as determinants of vector specificity has been suggested [23].
Actin is a conserved gene and its interaction with Amp, a highly variable phytoplasma protein, may be necessary for conserved function. For example, actin of the vector Circulifer haematoceps plays a role in the internalization of Spiroplasma citri [16]. Involvement of host cell actin in bacterial invasion is known also for other intracellular pathogens such as Shigella and Listeria [35], Salmonella [36;37], and Streptococcus [38;39], among others. Actin is also involved in the movement of a multiple nucleopolyhedrovirus of the alfalfa looper, Autographa californica, towards the host nucleus during the initial phases of the infection, and towards the cell membrane tips of actin-rich surface spikes later in infection [40]. Moreover, actin and ATP synthase are both binding proteins of the CrylAc toxin found in the proteome of the midgut membrane fraction of the tobacco budworm, Heliothis virescens [41].
ATP synthase is an enzymatic complex responsible for ATP synthesis. Although it is generally found in the inner membrane of mitochondria [42], several recent reports describe the localization of ATP synthase components on the outer face of the plasma membrane of several human, mouse and rat cell types, where they function as receptors for multiple ligands [43;44] and participate in diverse processes such as regulation of lipid metabolism and cholesterol homeostasis [45;46], control of proliferation and differentiation of endothelial cells [47], immune recognition of tumors [48], or human innate immunity [49], as reviewed in [42] and [50]. Among arthropods, in the mosquito (Aedes aegypti) midgut brush border, ATP synthase b is part of a complex network of cell surface proteins, including actin, binding Dengue-2 Flavivirus [51]. In fat body cells of the greater wax moth, Galleria mellonella, ATP synthase a and b subunits have been identified in mitochondria, in the cytosol and in the plasma membrane, where they interact with juvenile hormone binding protein and have a probable role in translocation of this molecule [52]. In the crustacean Pacifastacus leniusculus, ATP synthase b is expressed on the membrane surface of a subpopulation of hematopoietic cells, where it acts as a receptor of a cytokine [53]. It is noteworthy that, in the Pacific white shrimp, Litopenaeus vannamei, ATP synthase b binds to white spot syndrome whispovirus and has a role in the host invasion process [54].

Subcellular localization of ATP synthase
To investigate whether the b subunit of ATP synthase is exposed on the outer cell membrane of CYP insect vectors, WB assays and confocal microscopy were performed on mitochondrial protein fractions and plasma membrane.
SDS PAGE and Western blots. Figure 4A shows SDS PAGE profiles of plasma membrane (P) and mitochondrial (M) fractions isolated from whole bodies, salivary glands and midguts of E. variegatus. In WB assays, flotillin 1 and cytochrome C, markers of plasma membranes and mitochondria [55], respectively, were detected only in their corresponding fraction ( Figure 4B), confirming the quality of the organelle isolation protocol. ATP synthase b was detected in mitochondrial and plasma membrane fractions from whole bodies as well as from salivary glands and midgut ( Figure 4B), thus confirming the expression of this protein in the plasma membranes of the two leafhopper organs most involved in phytoplasma transmission.
Immunofluorescence. Whole permeabilised salivary glands ( Figure 5B) and midguts ( Figure 6B) of E. variegatus were strongly labelled by the anti-ATP synthase b antibody in confocal immunofluorescence microscopy. In contrast, using a rabbit preimmune serum on similar whole organ preparations, no signal was detected ( Figures 5C, 6C). Strong signals were also observed on sections of permeabilised salivary glands ( Figure 5D, E) and midguts ( Figure 6D, E), where mitochondria were visible as concentrated spots of signal. The anti-ATP synthase b antibody recognised the antigen also from a section of not permeabilised salivary glands ( Figure 5F) and midguts ( Figure 6F): in these cases the signal was most intense on cell edges, especially in the midgut, where the internal lumen brush border was strongly labelled. Finally, the same section of not permeabilised salivary glands was doubly labelled with antibodies against ATP synthase b (green, Figure 5G) and anti-flotillin 1 (red, Figure 5H), a cell membrane marker [55]: the merged image ( Figure 5I) shows, in yellow, the  co-localization of the two proteins. Similar observations were carried out on a section of not permeabilised midgut ( Figure 6G, H, I), though the flotillin 1 signal was weaker in this organ, even in single labelling (data not shown). The presence of ATP synthase b on the outer surface of the cell membranes of E. variegatus is the first demonstration of ectopic expression of this protein in the Order Hemiptera, although localization of ATP synthase was reported for plasma membrane fractions of D. melanogaster embryos [56].
We showed the presence of ATP synthase b in both midgut and salivary glands of a phytoplasma insect vector. These organs are essential in phytoplasma invasion of insects, the midgut being the site of hemocoel entry following feeding from the phloem of infected plants, and the salivary glands being sites of obligate colonization prior to transmission with saliva [11]. In other biological systems [57;58;43;59], including the freshwater shrimp [53], cell membrane ATP synthase b contributes to the production of extracellular ATP. In S. citri, an ATP binding cassette transporter is involved in salivary gland colonization of the vector C. haematoceps [60]. However, phytoplasmas lack the ATP synthetic pathway [61;3], and depend partly on their host for their energy metabolism [1;3]. Host extracellular ATP in the gut lumen and hemocoel may be required for extracellular survival of phytoplasma cells before their entrance into the midgut and salivary glands, respectively.

Conclusion
Several lines of evidence support a role for ATP synthase as a receptor of the phytoplasma antigenic membrane protein in the transmission of CYP: i) in vitro interaction between Amp and insect vector proteins; ii) absence of Amp interaction with proteins of non-vector insect species; iii) localization of ATP synthase b subunit on the plasma membrane of midgut and salivary glands of the leafhopper vector E. variegatus; and iv) different features in the amino acid sequence of E. variegatus ATP synthase b compared to that of a non-vector aphid, despite high sequence similarity. Although ATP synthase activity on the surface of vector gut and salivary glands has not been investigated, extracellular ATP production in arthropods has been reported [53]. During their extracellular phase, in both the gut lumen and the hemocoel, phytoplasmas might profit from this exogenous energy source for more efficient colonization of the intestine and salivary glands, respectively. Consistent with this hypothesis, the strong interaction between CYP Amp and ATP synthase of efficient vectors correlates with high phytoplasma multiplication and transmission rates [26].
The complex profile of binding proteins retained in the affinity column in this study suggests that phytoplasma Amp interacts with a network of insect vector proteins. Among these, ATP synthase a and b were identified for the first time, and actin was confirmed in CYP as already shown in OY [23]. P90, found in this study, could not be identified, most probably due to the lack of leafhopper peptide sequence information in the databases.
We cannot exclude the existence of a highly variable host receptor among the unidentified and non conserved interacting proteins. In this case, the interaction of CYP Amp with the highly conserved proteins ATP synthase, actin, and myosin could be mediated by such a receptor. Nevertheless, the freeze-thawing disruption of putative insect protein complexes before affinity chromatography did not alter the profile of Amp-interacting proteins (data not shown), suggesting that Amp interaction with actin and ATP synthase a and b could be direct. Moreover, the phytoplasma membrane proteome is complex, including several transport systems [3] and at least two antigenic membrane proteins [21], and it may also be dynamic, as suggested by the recent report of membrane-targeted proteins expressed differentially in the vector and encoded by extrachromosomal potential mobile units [7]. Recently, a role in phytoplasma transmission has been suggested also for a transmembrane plasmid-encoded protein of OY [24]. The role of other phytoplasma membrane proteins in interactions with insect proteins must be further explored to fully understand the mechanism of phytoplasma transmission by vector insects.

Phytoplasma strain, insect species and transmission experiments
Chrysanthemum yellows phytoplasma (CYP) [26] was maintained in daisy, Chrysanthemum carinatum Schousboe, by periodic insect transmission using M. quadripunctulatus. CYP vectors M. quadripunctulatus and E. variegatus were reared on oat (Avena sativa  Visual symptom assessment and PCR assays of total DNA from leaf veins were used to assess phytoplasma infection of the exposed plants. DNA extraction, diagnostic nested PCR and quantitative real time PCR Total DNA was extracted from leaf veins of exposed plants, from individual Z. pullula adults and from batches of 5/10 aphids, as described [62]. The presence of CYP was assayed by nested PCR driven with universal and specific primers [63;64]. The concentration of total DNA in PCR positive insects was measured using quantitative real time PCR, as described in [62]. CYP DNA was expressed as number of CYP cells per nanogram of insect DNA. To compare phytoplasma titre in aphids and in infective M. quadripunctulatus and E. variegatus [26], one-way analysis of variance (ANOVA) followed by Tukey test for multiple comparison was performed with SIGMAPLOT 11 (Systat Software).

Protein extraction
Total proteins were extracted from batches of all insect species (20 M. quadripunctulatus, 10 E. variegatus, 25 E. decipiens, 10 S. titanus, 30 Z. pullula and about 100 individuals of each aphid species) with the same procedure to perform dot far Western blots (FWB), one dimensional (1D) FWB, Western blots (WB) and affinity chromatography experiments. Insects were homogenized with a mortar and pestle in 200 ml of Rx buffer [23], and centrifuged for 1 min at 13,000 g; proteins in the supernatant were quantified with Bradford reagents (Bio-Rad).
For 1D FWB membrane proteins were extracted, according to a protocol adapted from [65]. Insects were crushed in 200 ml of Buffer 1 (1% Triton X-114, 150 mM NaCl, 10 mM Tris, pH 7.4), sonicated for 1 min and centrifuged for 3 min at 13,000 g at 0uC. The supernatant was layered onto 300 ml of a sucrose cushion (6% sucrose, 0.06% Triton X-114, 150 mM NaCl, 10 mM Tris, pH 7.4), incubated 3 min at 30uC and centrifuged at room temperature (RT) at 300 g for 3 min. The top (aqueous) phase, about 200 ml, was collected, mixed with 300 ml of fresh Buffer 1 and incubated on ice for 3 min; the bottom (detergent) phase kept at RT. The aqueous phase was loaded, for a second separation, on the same detergent phase. Following an incubation of 3 min at 30uC and centrifugation at RT (300 g for 3 min), the detergent phase containing membrane proteins was collected (,30 ml) and the protein concentration was estimated using the Bradford assay, as before.

Subcellular fractionation
To determine the cellular localization of ATP synthase, mitochondria and plasma membranes were separated from whole bodies or excised salivary glands and midguts of E. variegatus, as described [57]. Whole insects (50 specimens) were crushed in 600 ml of 9% TES buffer (20 mM Tris HCl, 1 mM EDTA, 9% sucrose, pH 7.4), and centrifuged at 500 g for 5 min at 4uC. The supernatant was homogenized in 9% TES buffer in a Potter-Elvehjem homogenizer and centrifuged at 14,0000 g for 30 min at 4uC. The pellet was resuspended in 1 ml of 9% TES buffer layered on a cushion of 1 ml of TES buffer containing 50% sucrose and 1 ml of TES containing 38.5% sucrose, and centrifuged at 100,000 g in a TLA-100.3 rotor (Beckman-Coulter) for 60 min at 4uC. Plasma membranes were collected from the top of the sucrose cushion, suspended in 9% TES buffer and pelletted by centrifugation at 31,000 g for 60 min at 4uC. Mitochondria were collected from the bottom of the sucrose cushion, suspended in 9% TES buffer and pelletted by centrifugation at 5,000 g for 60 min at 4uC. Excised salivary glands and midguts from 30 and 100 E. variegatus, respectively, were homogenized in 9% TES buffer in a Potter-Elvehjem homogenizer and treated as described above. Far Western blots Partial fusion proteins of CYP, Amp (Antigenic membrane protein) and Art (Arginine transporter), produced previously [27], were used as bait in FWB experiments, following a protocol slightly modified from [66]. For dot FWB, 40 mg total proteins of each insect species was spotted on polyvinyl difluoride (PVDF, Bio-Rad) membrane; 1 ng of CYPfAmp was also spotted, as positive control. For 1D FWB, insect total or membrane protein preparations were subjected to SDS-PAGE (40 mg/well) and blotted onto PVDF membrane. Dot FWB and 1D FWB membranes were blocked for 1 h at 4uC with 3% BSA in TBS containing 0.1% Tween (TBST) and then incubated overnight at 4uC with bait recombinant proteins: i) CYPfAmp, ii) CYPfArt, iii) His-tag purified from pRSetC expression vector (Invitrogen), or iv) buffer only (negative control). Four mg of all baits were suspended in BSA-TBST mixed with EDTA-free antiprotease cocktail Complete I (Roche). The PVDF membranes were washed for 1 h with BSA-TBST and for 10 min with TBST, then incubated

Western blots
Western blots (WB) were carried out to confirm the identification of Amp interacting insect proteins, using antisera specific for actin and for ATP synthase b. This technique also was used to analyse the subcellular localization of ATP synthase; in addition to the anti-ATP synthase b antibody, an antiserum specific for flotillin 1, a protein expressed only in the plasmalemma and used as a cell membrane marker, and another antiserum specific for cytochrome C, a mitochondrial marker [55], also were used. Following SDS-PAGE on 12-15% polyacrylamide gels, proteins were blotted on PVDF membrane. Membranes were blocked, incubated with primary and secondary antibodies diluted in BSA-TBST, washed and developed as detailed above for the FWB assays. Polyclonal primary antibodies, specific for ATP synthase b (ab43177, Abcam plc, diluted 1:5000), actin (A5060, Sigma-Aldrich, diluted 1:1000) and flotillin 1 (ab41927, Abcam plc, diluted 1:1000), were detected with HRP conjugated antibody A0545 (Sigma-Aldrich, diluted 1:10,000). The monoclonal anticytochrome C (ab13575, Abcam plc, diluted 1:1000) was recognized by HRP conjugated A4416 (Sigma-Aldrich, diluted 1:5000).

Affinity chromatography and protein identification
One mg of total proteins from different insect species was loaded on an affinity column (AminoLink Kit, Pierce) covalently linked to 2 mg of CYPfAmp, according to the manufacturer's instructions. Six fractions of about 1.5 ml were collected from each elution. Proteins in each fraction were precipitated with acetone, subjected to SDS-PAGE and either colloidal Coomassie blue stained or blotted on PVDF membrane for WB. Affinity chromatography assays were repeated at least three times for each insect species, always with freshly extracted proteins.
Four visible bands in the elution phases of E. variegatus were excised from the gel and shipped to the Protein ID Service (Department of Biology Technology Facility, University of York, UK) for trypsin digestion, and MALDI-MS and MS/MS analysis with a Bruker Autoflex III MALDI-TOF/TOF. Protein identification was accomplished using the MASCOT database search engine (www.matrixscience.com).

Primer design, RT-PCR, and sequence analysis
To further confirm the MS/MS protein identification we obtained the corresponding DNA coding sequences. Actin and ATP synthase b genes from several insect species, available in Genbank, were aligned with ClustalW2 [67] and degenerated primers were designed on consensus DNA fragment sequences encoding peptides identified by mass spectrometry of E. variegatus proteins. Selected primers were ActFw1 (59-ATGTGTGACGAW-GAKRTWGCMGC-39) ActRv1100 (59-CCDGGGCCGGAY-TCGTCGT-39) for actin, and BetaFw220 (59-CAGCAY-TTGGGWGAAAAYAC-39) BetaRv1110 (59-GGATCBACAG-CWGGRTARATACC-39) for ATP synthase b. Total RNA was extracted from five E. variegatus adults with TRIzolH reagent (Invitrogen) and reverse transcribed to cDNA with MuLVH Reverse Transcriptase and Random Hexamers (Applied Biosystems) following manufacturer's recommendations. E. variegatus cDNA was amplified with degenerated primers and PCR products, corresponding to actin and ATP synthase b partial genes, were sequenced with a capillary 3730 DNA Analyzer (ABI) by Bio-Fab Research (Pomezia, RM, Italy). Programs ClustalW2 [67], Conserved Domain Database [68] and PredictProtein [69] were used to analyse sequence features.

Immunofluorescence
Salivary glands and midguts were excised from E. variegatus adults and fixed quickly in 4% paraformaldehyde, 0.1 M phosphate buffer, pH 7.4, 0.1% Triton X-100 overnight at 4uC. Organs were washed three times in phosphate-buffered saline, pH 7.4 (PBS) and some were permeabilised overnight at 4uC with PBS and 1% Triton X-100. Confocal microscopy was performed on whole organs and on 100 mm thick sections, made with a Balzer Vibratome series 1.000 from 8% agarose embedded organs. Whole organs or sections were blocked in PBS containing 1% BSA for 30 min, incubated overnight with anti-ATP synthase b antibody (ab43177, Abcam plc, diluted 1:200), washed three times with PBS, blocked, incubated for 2 h with a 1:80 dilution of the FITC conjugated antibody (F1262, Sigma-Aldrich), and finally washed five times in PBS. Samples were mounted on microscope slides in Citifluor (Molecular Probes, Invitrogen) and observed. Control sections were treated in a similar manner but incubated in a rabbit pre-immune serum (diluted 1:200) instead of primary antibody. A sequential labelling protocol was employed to double label ATP synthase and flotillin 1 and confirm the ectopic expression of the former. We first labelled ATP synthase with FITC as described above. Sections were then incubated overnight with anti-flotillin 1 antibody (ab41927, Abcam plc) diluted 1:150, and 3 h with Alexa 633 conjugated antibody (A-21071, Molecular Probes, Invitrogen) diluted 1:50. The control sections were treated in a similar manner, except that either the second or both antibodies were omitted from the incubation mixture. All samples were observed with a Leica TCS SP2 confocal microscope, using a 406 water-immersion objective (HCX Apo 0.80). Laser bands of 488 nm Ar and a 633 nm He/Ne were used to excite FITC and Alexa 633, respectively.