All aphid lines used in this study were clonal lineages derived from a single founding female. 4106A was an insecticide-susceptible clone originating from potatoes in Scotland in 2000. Clone 5191A originated from tobacco in Greece in 2007, and exhibited 20-40 fold resistance to neonicotinoids [6, 7]. Clone FRC was derived from a sample collected from peach orchards in Southern France in 2009, following reports of reduced control with neonicotinoids. All clones were reared on Chinese cabbage under standardised conditions .
Two bioassay methods were used to establish the extent of neonicotinoid resistance in the FRC clone using insecticide alone and 5 h after pre-treatment with PBO to investigate the potential contribution to resistance of detoxifying enzymes that are inhibited by this synergist . The 'micro-application' bioassay involved dosing individual aphids on the dorsal surface with 0.25 μl of acetone plus insecticide using a micro-applicator (Burkhard) and measuring mortality 72 h later as previously described . The 'spray application' bioassay involved spraying insecticide onto aphids and foliage inside a Potter-Precision laboratory spray tower (Burkhard)  For this formulated imidacloprid (Confidor 200 g/L SL Bayer S.A. Division Agro, France) and thiamethoxam (Actara 25WG Syngenta Crop Protection AG, Basel, Switzerland) was used. For synergism assays piperonyl butoxide was formulated by Syngenta chemists as an 80% EC and applied in 0.01% non-ionic surfactant, Extravon (Syngenta Crop Protection AG, Basel, Switzerland). Test pots (45 mm diameter) were prepared with discs of Chinese cabbage on tap water agar. Twenty to thirty mixed age aphids were transferred to the dishes and allowed to settle for 24 h at 21°C with a 16:8 h light regime. Serial dilutions of insecticide were applied in 3 ml of solution at 0.6 bar with 3 second settling time (equivalent to approximately 400L/ha). Three replicates were used for each concentration tested and the responses were assessed after 72 h. Aphids that were dead or seriously affected were classed as 'dead' and LC50 values were calculated by LOGIT analysis using Syngenta software.
Imidacloprid Metabolism Assay
Imidacloprid metabolism assays were carried out to determine which imidacloprid metabolites were observed in aphids and whether increased metabolism was contributing to resistance in the FRC clone. In both experiments, individual aphids of the 4106A and FRC clones were treated by topical application of imidacloprid (0.2 μL, 1 mg/L, corresponding to the LC50 value for 4106A) using a micro-applicator as described in . Owing to the physical constraints of performing topical applications to aphids, aphids were treated in batches, which resulted in a 15 minute gap between the first and last aphid being treated from each batch.
To determine which metabolites were observed in aphids, ~500 treated aphids of each clone were collected at 0, 1, 4 and 24 h after treatment, weighed and frozen at -20°C. Due to the low treatment rate and volume used in this study, 500 aphids were pooled to provide one sample for each time point. To remove surface imidacloprid, aphid samples were washed in acetonitrile (3 × 1 mL), after which, aphid samples were placed in extraction tubes (Precellys 24, 0.5 mL, Lysing matrix VK05; Bertin Technologies, France) containing acetonitrile (0.5 mL). Samples were macerated for 5 × 20 s using a FastPrep® FP120 (QBiogene, UK), centrifuged (5590g, 15 min) and the resulting supernatant was transferred into HPLC vials for LC-MS/MS analysis. Separation was achieved with a Accela UPLC pump (Thermo Scientific, UK) using a Phenomoenex Luna (C18, 5 μm, 4.60 × 250 mm) analytical column, with a mobile phase consisting of water (+0.2 % formic acid), and a flow rate of 0.8 mL/min. The gradient elution conditions of acetonitrile:water were as follows: 0 min 10:90, 15 min 47.5:52.5, 15.1 min 80:20, 22 min 80:20, 22.1 min 10:90. The mass spectrometer used was a TSQ Vantage (Thermo Scientific) equipped with a H-ESI II source operating in positive ion mode. Analytes were detected simultaneously using selected-reaction-monitoring (SRM). SRM transitions are outlined in Additional file 1. Scan time was 0.1 s/transition and the collision gas pressure was 1.5 mTorr. As noted in Additional file 1, several metabolites were available as authentic standards. For these metabolites, extraction efficiencies were determined, by spiking untreated aphid samples with a known amount (0.05 μg) prior to performing the maceration and centrifugation steps, which were between 87-94%. For 4/5-hydroxy-imidacloprid, only 5-hydroxy-imidacloprid was available as an authentic standard and it was thought to be unlikely that chromatographic separation of the 4- and 5-hydroxy metabolites would have been achieved. Where no authentic standards were available, confirmation was made by comparison of relative retention times to previous studies .
In the second experiment, aphids were treated as described above, with the following modifications: 250 treated aphids of each clone were collected at 0, 4 and 24 h after treatment, weighed and frozen at -20°C. Treatments were replicated three times. Aphid samples were washed in acetonitrile (3 × 0.5 mL), placed in extraction tubes (Precellys 24, 0.5 mL, Lysing matrix VK05) containing acetonitrile (0.25 mL) and prepared as described above. Recoveries of imidacloprid and 4/5-hydroxy metabolite were determined using LC-MS/MS. Separation was achieved using Ultra Performance LC® (ACQUITY UPLC-System; Waters, UK) using an ACQUITY UPLC column (HSS T3, 1.8 μm, 100 × 2.1 mm), with a mobile phase consisting of water (+0.2% formic acid), with a flow rate of 0.6 mL/min. The gradient elution conditions of acetonitrile:water were as follows: 0 min 0:100, 0.5 min 0:100, 3.5 min 95:5, 4.5 min 95:5, 4.6 min 0:100, 5.0 min 0:100. The mass spectrometer used was a Finnigan TSQ Quantum Discovery (Thermo Scientific) equipped with an Ion Max source operating in positive ion mode. Imidacloprid and 4/5-hydroxy metabolite were detected simultaneously using SRM.
The microarray (Agilent Technologies) used in this study was designed by the Georg Jander Lab and is based on a previously described array containing probes for > 10, 000 M. persicae unigenes produced by Sanger sequencing  augmented with an additional 30, 517 probe set designed on EST unigene sequences identified in a 454 sequencing project . Microarray experiments consisted of four biological replicates and for each of these, two hybridisations were done in which the Cy3 and Cy5 labels were swapped between samples for a total of eight hybridisations between resistant and susceptible clones. Arrays were processed, scanned and the raw data normalised as described previously . The TM4 suite of software from the Institute of Genomic Research was used for statistical analysis using the SAM (Significance Analysis of Microarrays) module and applying a False Discovery Rate (FDR) of zero to detect significantly differentially expressed genes . Genes identified by SAM and showing a transcription ratio > 2 fold in either direction were considered to be differentially transcribed between the two clones. Microarray data were submitted to the Gene Expression Omnibus (GEO) database (series no. GSE24629).
Quantitative RT-PCR was used to validate microarray data by examining the expression profile of selected genes. Primers were designed to amplify a fragment 90-150bp in size and are listed in Additional file 4. PCR reactions were carried out as described previously . Data were analysed using the geometric mean of three selected housekeeping genes (actin, para which encodes the voltage gated sodium channel, and ace, which encodes acetylcholinesterase) for normalisation according to the strategy described previously . Quantitative PCR was used to determine CYP6CY3 gene copy number using genomic DNA as the template as described previously .
Membrane preparation and [3H] imidacloprid binding
Membranes from resistant (FRC) and susceptible (4106A) aphid clones were prepared and radioligand binding experiments performed based on previously described methods  with the following modifications. Aphids, (~5 g) were homogenized on ice in 16 ml of pre-chilled homogenisation buffer (0.05 M Tris base, pH 7.4 with HCl, 200 μM PMSF) with a motor-driven Ultra Turrax (5 × 20 sec bursts, level 4.5). The homogenate was centrifuged (3,000 g, 30 mins at 4°C) and resulting supernatant filtered through two layers of pre-wetted mira cloth before a final centrifugation (83,000 g, 60 mins at 4°C). The pellet (crude membrane) was re-suspended in chilled freshly prepared binding buffer (0.05 M Tris, 0.12 M NaCl, 100 μM EDTA, pH7.4 HCl), beaded into liquid N2 and stored at -80°C. The protein concentration was determined by Bradford assay using BSA as a standard. Comparative saturation isotherms were performed in side-by-side fashion with appropriate time delay. [3H]-imidacloprid (1.1 TBq/mmol, purity > 95%, in-house synthesis) and methyllycaconitine (MLA, Sigma) (non-specific ligand) were prepared in binding buffer. Crude aphid membranes were defrosted on ice and diluted to 5 mg/ml in binding buffer containing 0.2% BSA. Incubations were carried out in a final volume of 1 ml to avoid ligand depletion at very low concentrations of added [3H]-imidacloprid (returned bound counts were in all cases less than 10% of added). Order of addition was binding buffer, 100 nM MLA or control, 100 μg membranes followed by [3H]-imidacloprid at indicated concentrations. Reactions equilibrated over 90 mins at room temperature with constant shaking and were terminated by filtration using a Brandel harvester onto Whatman GF/B filters pre-soaked in cold binding buffer with 0.25% polyethylenimine (no BSA). Filters were rapidly rinsed in a two wash cycle (3 ml) with ice cold binding buffer (no BSA), removed, allowed to air-dry for 12 hrs and soaked (6 hrs) in 10 ml scintillant (Scintsafe Gel, Fisher) prior to counting.
Sequence analysis of M. persicae nAChR subunits
Total RNA was isolated from three pools of five adult aphids from all clones listed above and first strand cDNA synthesised as described previously . PCR amplification was performed using gene-specific oligonucleotide primers designed to the published nAChRα1 (X81887), nAChRα2 (X81888), nAChRα3 (AJ236786), nAChRα4 (AJ236787), nAChRα5 (AJ880084) and nAChRβ1 (AJ251838) as described previously . Amplicons of the expected size were ethanol precipitated and sequenced directly in both directions using the BigDye mix (Applied Biosystems).