4% NaCl, 2.1% MgSO4·7H2O, 1.8% MgCl2·6H2O, 0.42% KCl, 0.056% CaCl2, and 12 mM Tris-HCl, pH 7.5). Solid media were prepared by the addition of 1.5% agar (Difco). If required, novobiocin was added at 0.3 μg mL−1. Escherichia coli was routinely grown in Luria–Bertani medium (0.5% yeast

extract, 1% peptone, 1% NaCl); if required, 100 μg mL−1 ampicillin was added. For the construction of plasmids, E. coli JM109 (F′traD36 proA+B+ lacIqΔ(lacZ)M15/Δ(lac-proAB) glnV44 KU-57788 chemical structure e14− gyrA96 recA1 relA1 endA1 thi hsdR17) was used. To prepare unmethylated DNA for efficient transformation of H. volcanii, E. coli ER2925 (New England Biolabs, Hitchin, UK) was used. Transformation of E. coli (Sambrook & Russel, 2001) and H. volcanii (Cline et al., 1989) was performed as described. General DNA techniques were performed as described (Sambrook & Russel, 2001). AmyH was produced CX-5461 solubility dmso in H. volcanii by transforming this strain with the plasmid pSY-AmyH, which has been described before (Kwan et al., 2008). All mutations in the signal-peptide encoding region of the amyH gene were carried out using the Quickchange mutagenesis system (Stratagene, La Jolla, CA). To visualize AmyH secretion on plates, 0.5% starch was added to YPC-agar. After 2 days of growth, starch-YPC plates were stained for 30 s with iodine solution (2% KI, 0.2% I2). Proteins were separated by sodium dodecyl sulphate polyacrylamide gel

electrophoresis (SDS-PAGE) and immunoblotted onto polyvinylidene difluoride membranes (Millipore, Watford, UK) using a semi-dry system. Amylase was visualized with specific antibodies and horseradish peroxidase anti-rabbit IgG conjugates (Promega, Southampton, UK), using the Pico West detection system (Perbio Science, Cramlington, UK). Proteomes from E. coli K-12 MG1655, Haloarcula marismortui ATCC 43049, Natromonas pharaonis DSM2160, and Halobacterium salinarum NRC1 were obtained through the European Bioinformatics Institute (http://www.ebi.ac.uk/genomes). Proteomes were analysed firstly with tatfind 1.4 at http://signalfind.org/tatfind.html (Dilks et al., 2003). To avoid false-positives, two additional steps were adopted. Firstly, very few (if any) Tat substrates

are polytopic integral membrane proteins, and proteins showing one or more additional membrane-spanning domains (using TMHMM at http://www.cbs.dtu.dk/services/TMHMM/) were therefore removed from Fludarabine chemical structure the dataset. Secondly, proteins in the dataset were analysed for signal peptides using the Hidden Markov model of signalp 3.0 (Bendtsen et al., 2004; http://www.cbs.dtu.dk/services/SignalP/). Any proteins below the threshold score of 0.5 were also removed. For archaea, it is not clear whether the Gram-negative or the Gram-positive model is better; for this reason, both were tested and proteins scoring below the threshold in either model were removed. The final datasets contained 24 Tat substrates for E. coli, 94 for H. marismortui, 41 for H. salinarum, and 74 for N. pharaonis.