E. minuta cells were collected in 2005 in the Mediterranean sea near Stareso, Corsica, France (Em. S1, E. minuta Stareso1), cultured in the laboratory in artificial sea water obtained from the Botanical and Zoological Garden Stuttgart (Wilhelma) and fed with Klebsiella minuta grown on nutrient agar. For the isolation of DNA, a concentrated sample of living cells was mixed with 8 M guanidiniumchloride. A 10:1 mixture with 1 M phosphate buffer pH 7.0 was made, adsorbed on a hydroxyapatite (Biorad, bio-gel HTP) column (1 cm × 0.4 cm) and washed with 4 M guanidiniumchloride, 100 mM phosphate buffer pH = 7.0, followed by washing with 4 M guanidiniumchloride, 200 mM phosphate buffer pH = 7.0. Subsequently, the bulk of DNA was eluted with 4 M guanidiniumchloride 500 mM phosphate buffer pH = 7.0. The DNA was diluted with 1 volume water and precipitated with 10 v/v% 3 M sodiumactetate pH = 5.2 and 50 v/v % propanol-2 for 10 minutes at room temperature. After precipitation and washing the pellet was air dried. Finally, the DNA pellet was dissolved in DEPC treated water (Invitrogen).
The dissolved DNA was loaded on a pulsed field agarose gel (1% agarose type II medium EEO, Sigma) and run at 170 V (145 mA) ramping from 2.5 s - 25 s for 16 hours PFGE with a LKB 2015 Pulsaphor plus control unit.
The band just below the first band of the lambda marker (Figure ) was cut out and the DNA extracted. The position of the mitochondrial band on pulsed field gel is a clear indication of a linear mitochondrial genome. Circular mitochondrial genomes of this size should run much faster in the gel. The DNA of the band was digested with Sau 3A and then size fractionated on an agarose gel. The DNA from these fractions was isolated from the gel, ligated in pUC-18 digested with BamH1 and transformed in E. coli DH101B cells. The titre of the library was 1.12 × 105. From this library, plasmid DNA from 288 different colonies was sequenced with an ABI prism 3730 online capillary sequencing machine and the mitochondrial genome was assembled as described below. The gene library was constructed by Genterprise, Mainz, Germany.
E. crassus was collected from shallow coastal waters of the sandy beach of Porto Recanati (43° 26' N, 13° 40' E) on the Italian Adriatic Coast, 50 km south of Ancona, July 1984 and cultured in the laboratory in artificial sea water (NaCl 465 mM, KCl 10 mM, MgCl2.6H2O 24.8 mM, MgSO4.7H2O 28.1 mM, CaCl2 10.4 mM, NaHCO3 2.4 mM pH 8.0).
Initially, a culture was kept in artificial seawater in an Erlenmeyer flask and fed with a small piece of raw beef. Alternatively, a set of 200 ml tissue flasks was first siliconized, filled with approximately 50 ml of artificial seawater, and inoculated with E. crassus cells. These cultures were fed with HB101 E. coli cells.
Total DNA of E. crassus was isolated by dissolving cells in 8 M guanidiniumchloride and purification by hydroxyapatite as described above for E. minuta.Four fragments of different mitochondrial genes were obtained by PCR with degenerated primers on this DNA, i.e. primers directed against the ribosomal genes rnl (5'-GTCAAGAGAGAAACAGC-3', 5'-GCATAGGGTCTTCCCGTC-3'), rns (5'-TGTGCCAGCAGCCGCGGTAA-3', 5'-TCCCMTACCRGTACCTTGTGT-3') and the complex I genes nad7 (5'-TTCGGWCCHCARCAYCCHGC-3', 5'-CTRTCRACYTCWCCRAARAC-3') and nad10 (5'-TTYGGHYTNGCHTGHTG-3', 5'-ARDGCYTCDSWDGTDGGDGGDCA-3') On these gene fragments primers for long range PCR were developed and long range PCR with LA-Taq-polymerase (5 U/μl) (Takara bio inc.) was performed. The long range PCR products were digested with different restriction enzymes, subcloned in pUC-18 (Sigma) or in pGEM-T easy (Promega) and sequenced. Sequencing was performed at the DNA diagnostics centre of the Nijmegen University Medical Center using M13 forward and reverse primers.
All sequences have been submitted to NCBI GenBank. The GenBank accession-numbers are for E. minuta GQ903130
and for E. crassus GQ903131
. The protein identifiers are displayed in additional file 4
Analysis of the sequence data
The sequences for the 18s rRNA phylogeny were aligned using the SINA Webaligner http://www.arb-silva.de/aligner
, which aligns them in accordance with the ARB/SILVA rRNA alignment [38
] which is based on a secondary structure model [39
]. Subsequently we used Gblocks [40
] to identify reliably aligned parts, using the default settings except that we did not require the coverage for every position to be 100%, but rather 80%. We then used PhyML v3.0.1 (HKY85 model, optimised equilibrium frequencies, estimated ts/tv ratio, estimated proportion of invariable sites, 4 substitution rate categories, estimated gamma distribution parameter, NNI tree topology search, 100 bootstrap iterations [41
]) to obtain the phylogeny.
The genetic code used for the translation of the Euplotes
mitochondrial DNA was derived using the standard genetic code for translation of the complete DNA sequence in 6 frames, and searching the resulting protein sequences for conserved Pfam-fs protein domains [42
] using HMMPFAM [30
]. The amino acid frequencies provided by the Pfam HMM profiles were then used to predict the translation of each codon. Averaging over all aligned occurrences of the codon, the highest scoring (i.e. most often aligned) amino acid was predicted to be the translation of the codon in vivo