Rat-tailed maggots were collected in liquid manure pits at cattle farms near Giessen (Hessen) and the village Siewisch (Brandenburg), Germany. Using a practicable key for determination of hoverfly larvae , we identified third instar larvae of the drone fly E. tenax. As typical characters we observed the lack of setae along the lower lateral margins and the last pair of prolegs with most of the large primary crochets facing towards the lateral margins of the body. Identification was confirmed by determination of adults which hatched at the end of October 2005.
Immune challenge of larvae and isolation of hemolymph and mRNA
Last instar E. tenax larvae, each weighing between 200–300 mg, were used for immune-challenge. 10 μl sample volume, corresponding to 100 μg LPS (purified Escherichia coli endotoxin 0111:B4, Cat. No.: L2630, Sigma, Taufkirchen, Germany) per larva were injected dorsolaterally into the hemocoel using 1 ml disposable syringes and 0.4 × 20 mm needles mounted on a microapplicator. For antibacterial activity assays, hemolyph samples were isolated by bleeding injected larvae 24 h post challenge or untreated larvae into plastic tubes. Total RNA was extracted from whole larvae 8 h post injection using the TriReagent isolation reagent (Molecular Research Centre, Cincinnati, Ohio, USA) and a further poly-(A)+ RNA preparation was performed using a mRNA Nucleotrap kit (Macherey Nagel, Germany) according to the instructions of the manufacturers. RNA integrity was confirmed by ethidium bromide gel staining and quantities were determined spectrophotometrically .
Antibacterial activity assay
Antibacterial activity of hemolymph samples was measured by an inhibition zone assay using a LPS-defective, streptomycin- and ampicillin-resistant mutant of Escherichia coli K12 strain D31 . In brief, Petri dishes (∅ 100 mm) were filled with 7 ml E. coli suspension, containing 2× YT-nutrient broth (Roth, Karlsruhe, Germany), 1% high-purity agar-agar (Roth), and 2 × 104 viable bacteria in logarithmic growth phase. Holes with a diameter of 4 mm were punched into the agar and filled with 3 μl of cell-free hemolymph. The diameters of the clear zones were measured after 24 h of incubation at 37°C and units/ml were calculated using a calibration curve obtained with dilutions of gentamycin (Sigma).
Construction of a subtracted cDNA library using the SSH method
In order to identify differentially expressed genes during immune response the SSH method was performed using mRNAs from immune challenged and control larvae and the PCR-Select cDNA Subtraction Kit from Clontech (Mountain View, CA, USA), according to the protocol of the manufacturer. Briefly, 1 μg of purified mRNA from immune-challenged and control larvae were reverse transcribed into cDNA using a cDNA synthesis primer, subsequently double stranded cDNA was generated and digested with RsaI. The double stranded cDNA from immunized larvae was ligated in separate aliquots to adaptor 1 or adaptor 2R and were denaturated at 98°C for 90 s and then hybridized at 68°C for 8 h with a 30 fold excess of double stranded cDNA from control larvae. Subsequently, both samples were mixed together again with a 10 fold excess of freshly denaturated double stranded cDNA from control larvae and hybridized in one tube at 68°C for 16 h. The sample was then subjected to two rounds of suppression PCR with PCR-primer 1 and nested primers supplied with the kit. PCR amplifications were performed in a total volume of 25 μl using a PCR cycler (Biometra, Göttingen, Germany) with a heated lid and the Advantage PCR system (Promega, Mannheim, Germany). An initial adapter extension at 72°C for 5 min was followed by a denaturation step at 95°C for 1 min and by 27 cycles of denaturation at 95°C for 15 s, annealing at 66°C for 30 s, and extension at 72°C for 90 s. A final 7-min 72°C step was added to allow complete extension of the products. The secondary PCR was performed with nested primer 1 and 2R on the diluted primary PCR products for 12 cycles under same conditions, except that 68°C was used as annealing temperature. The subtraction efficiency was confirmed by quantitative real-time PCR of the actin gene of subtracted PCR products in comparison to not subtracted PCR products. Resulting PCR products of the secondary subtractive PCR were separated on 1% (w/v) agarose gel electrophoresis with ethidium bromide staining, according to standard procedures  and five fractions were obtained by excising five gel pieces of PCR products with different lengths. Subsequently, PCR product fractions were separately purified using the NucleoSpin Extract II kit (Macherey Nagel), ligated into pGEM-T easy vector (Promega) and transformed into TOP10F' cells (Invitrogen, Carlsbad, CA, USA). The library was plated on 2×YT agar plates containing 100 μg/ml ampicillin and incubated at 37°C for 16 h. A preliminary screen of 15 colonies using the FastPlasmid Mini kit (Eppendorf, Hamburg, Germany) followed by EcoRI digestion of isolated plasmids and agarose gel electrophoresis showed that over 90% of clones contained an insert in the vector.
Colony PCR and blot hybridization
Colony PCR was performed on 288 randomly picked colonies with vector specific primers T7-promotor: 5'-TAATACGACTCACTATAGGG-3' and SP6: 5'-ATTTAGGTGACACTATAG-3' (purchased from Thermo electron, Waltham, MA, USA) using a Biometra PCR cycler and the Red Taq PCR system (Sigma, Taufkirchen, Germany). Used PCR conditions were: denaturation at 95°C for 3 min followed by 30 cycles of denaturation at 95°C for 15 s, annealing at 43°C for 15 s, and extension at 72°C for 60 s. A final 7-min 72°C step was added to allow complete extension of the products. 1 μl of resulting PCR products were identically spotted onto two sheets of positively charged nylon membranes (Roche, Lewes, United Kingdom). Membranes were dried and UV cross-linked using a BioRad UV cross-linker (BioRad, München, Germany), according to the instructions of the manufacturer. Digoxigenin labeled probes for hybridization were generated using secondary PCR products of subtracted and non-subtracted cDNAs and the Dig-High Prime Labelling kit (Roche, Lewes, United Kingdom). Hybridization, washing, and detection of digoxigenin labeled DNA was performed in accordance to the user guide instructions of the Dig Easy Hyb Granules, Dig-Wash and Block Buffer Set, Anti-Digoxigenin-AP and NBT/BCIP ready-to-use tablets (Roche).
Sequencing and computer analysis of cDNA sequence data
Plasmid isolation of 117 positively screened colonies was performed with the FastPlasmid Mini kit (Eppendorf) and purified plasmids were custom sequenced by Macrogen Inc. (Seoul, South-Korea). Sequences were used to identify similar sequences of the National Center for Biotechnology Information databases using BLASTX program (BLASTX 2.2.13; http://www.ncbi.nlm.nih.gov/BLAST/). InterProScan http://www.ebi.ac.uk/InterProScan/ was used for an integrated search in PROSITE, Pfam, and PRINTS databases at EMBL-European Bioinformatics Institute and to predict signal sequences and transmembrane regions.
Sequence alignments and phylogenic analyses
Sequence alignments were computed using the blosum62 algorithm  at http://bioinfo.genopole-toulouse.prd.fr. For phylogenetic reconstruction, we used the software package MrBayes 3.1.2 , which combines Bayesian inference and Markov chain Monte Carlo convergence acceleration techniques known as Metropolis coupling. The best fixed-rate model of amino acid evolution was determined by model jumping among nine possible models. The model with the overall highest posterior probability was WAG  for PGRPs and Blosum62 model  for defensins. Generations were sampled with the current tree saved at intervals of 100 generations. We used convergence diagnostic (i.e., the standard deviation of split frequencies) to determine whether the run length was sufficient. The average standard deviation of split frequencies at 2 × 106 generations was 0.0076 for PGRPs and at 107 generations 0.0023 for defensins, respectively. This therefore indicated that the two chains that were run converged on similar results. The 50% majority rule tree presented here was constructed from all sampled trees with the first 25% of all trees. Trees were visualized with TREEVIEW 1.6.6 . Posterior probabilities plotted at the nodes can be interpreted as the probability that the tree or clade is correct .
Quantitative real time RT-PCR
Quantitative PCR was performed with the real-time PCR system Mx3000P (Stratagene, La Jolla, California, USA) using the FullVelocity SYBR® Green QRT-PCR Master Mix (Stratagene, La Jolla, California, USA), according to the protocols of the manufacturer. In order to confirm the subtraction efficiency of constructed cDNA library 1 ng of unsubtracted and subtracted cDNA, respectively, was used to amplify α-tubulin, Et-AMP10, and eristalin. Used primers were: the universal primers α-tubulin-forward: 5'-GCCAACCAGATGGTCAA-3' and α-tubulin-reverse: 5'-GCTTGGTCTTGATGGTG-3', eristalin-forward: 5-ATGGCTACATGTGATCTGCT-3', eristalin-reverse: 5'-ACGGCAATTGCAGACT-3', Et-AMP10-forward: 5'-ATGGACCCTCTTCTGTGG-3', Et-AMP10-reverse: 5'-TGGGCATCTGACAATA-3'. For gene expression analyses, we used 100 ng total RNA per well and primers described above. In addition, insect universal primers actin-forward: 5'-ATCCTCACCCTGAAGTACCC-3 and actin-reverse: 5-CCACACGCAGCTCATTGTA-3' were used to amplify Eristalis actin. Primers were selected using the primer3 software  and were purchased from Thermo electron (Waltham, MA, USA).