Complete Genome Sequence of Amycolatopsis mediterranei S699 Based on De Novo Assembly via a Combinatorial Sequencing Strategy

Abstract

The genome of Amycolatopsis mediterranei S699 was resequenced and assembled de novo. By comparing the sequences of S699 previously released and that of A. mediterranei U32, about 10 kb of major indels was found to differ between the two S699 genomes, and the differences are likely attributable to their different assembly strategies.

GENOME ANNOUNCEMENT

Isolated in 1957 at St. Raphael, France (8), Amycolatopsis mediterranei original strain ATCC 13685 synthesizes a collection of rifamycin complexes, while a single fermentation product of rifamycin B, whose derivatives are effectively used to treat mycobacterial infections (11), could be obtained only if sodium diethyl barbiturate was added to the medium (9). In 1975, a mutant strain designated ATCC 21789 that is able to produce a sole product of rifamycin B without the addition of barbiturate was isolated (11), and a related strain, S699 (5), has been widely used in laboratory research ever since (1, 2, 4, 6).

In 2011, the whole-genome sequence of A. mediterranei S699 (CP002896) presenting a 10,236,779-bp chromosome, assembled by mapping the shotgun sequences to the 10,236,715-bp genome of a rifamycin SV-producing strain, A. mediterranei U32 (CP002000) (14), was released (10). In this project, an S699 strain obtained from Giangarlo Lancini at Lepetit Laboratories, Geranzano, Italy, was resequenced via a combinatorial strategy and de novo assembly.

A total of 589,827 reads were generated using a Roche 454 GS FLX Titanium system and were assembled using the Newbler Program (version 2.3), which resulted in 67 contigs with an average size of 152,345 bp. Sanger-based sequencing was employed to facilitate gap closing, to amend the low-quality regions (score < 40), and to verify the variations between the draft sequence and the corresponding genome regions of other A. mediterranei strains. Finally, a consensus sequence containing 10,246,920 bp with an estimated error rate of <0.5 per 100,000 bases providing 31-fold coverage was acquired.

Beside a few different sequences predicted by Glimmer (3) and Genemark (7), most of the S699 genome was annotated by BLASTN based on the annotation of U32 followed by BLASTP functional assignment using KEGG and NR databases. A total of 9,227 protein-coding gene loci (AMES_CDSs) with an average length of 989 bp were identified.

Compared to the S699 (CP002896) genome, 218 single nucleotide polymorphisms (SNPs) and 51 indels were found in our S699 (CP003729) sequence. All 12 large indels (>40 bp and mostly repeat sequences) are insertions compared to not only the S699 (CP002896) genome but also that of U32. Except for the three insertions that failed to be specifically amplified by PCR, we confirmed that all of the other nine insertions, including a 7.5-kb insertion, were present in the genomes of ATCC 13685 and ATCC 21789. Therefore, we propose that the major indel variations between these two S699 genomic sequences could be caused by their different assembly strategies (10).

In comparison to the U32 genome, 234 SNPs and 48 indels were identified in our S699 (CP003729) genome. In particular, in the rif cluster, unlike that of U32, the Rif-Orf16 (cytochrome P450) encoded by gene locus AMES_0651 essential for the conversion of SV to B is a wild-type protein, as is the case in other rifamycin B-producing strains (13, 14). In addition, a frameshift of an aminotransferase (Rif-Orf9) was found at the 3′ region of its encoding AMES_0639 gene locus, but it may not affect the production of rifamycin (1, 12).

Nucleotide sequence accession number.

The genome sequence of A. mediterranei S699 in this project has been deposited in the GenBank database under accession number CP003729.