The cDNA was then synthesized, cloned, and packed using the ZAP-cDNA synthesis kit and the ZAP-cDNA

Gigapack III Gold Packaging Extract (Stratagene, La Jolla, CA, USA) following the manufacture’s instructions. After packaging, the obtained P. nordestina skin cDNA library was plated and the isolated phage clones were randomly collected in SM buffer (10 mM NaCl; 8 mM MgSO4; 50 mM Tris-HCl pH 7.5; 0.01% gelatin) containing 0.3% chloroform, before the recovery of the phagemid Doramapimod cost containing the recombinant cDNA by in vivo excision. Alternatively, some of the clones were isolated after mass excision of the library. After in vivo excision the plasmid cDNA clones were then amplified and purified by alkaline lysis using Wizard Minipreps DNA Purification kit (Promega, Madison, WI, USA). The nucleotide

sequence was determined by the dideoxy chain-termination method using the BigDye™ Terminator Cycle Sequence Kit and the ABI 310 automatic system (Applied Biosystems, Foster City, CA, USA). The analysis of sequences was conducted using a set of web based analysis programs. Sequence quality was first analyzed MG-132 in vivo with the Phred and Crossmatch software packages to remove low quality ends (Green, 1996). After this preliminary analysis, only good quality sequences (phred > 20) with a length longer than 150 bp were considered for definitive annotation. The collection of good quality sequences was organized into clusters by using CAP3 software. We took into account overlaps of 50 bp that had at least 98% identity (Huang

and Madan, 1999). The obtained sequences were compared to protein GenBank NR (http://www.ncbi.nih.gov) and Swissprot release 44 (ftp.ebi.ac.uk/pub/databases/swissprot/release/) Dynein databases using the BLASTx program (Altschul et al., 1990). Gene descriptions and EC numbers from Swissprot best hits and their associated product names were automatically assigned using 10−10 as the e-value cutoff. Thereafter, the ESTs were manually inspected by comparing the BLAST results with the automatically annotated EC numbers for functional classification. After this, an additional annotation allowing the alignment was conducted comparing the predicted protein sequences of clusters with Uniprot database and Swiss-Prot, SP-TrEMBL and stable Ensembl proteomes databases using the SMART software (Schultz et al., 1998). The average readable sequence length was of 390 bp, and only those considered having good quality were used to proceed with annotation. In this work, a total of 212ESTs or clusters were analyzed.