Strains LMG 1590^T and LMG 1591^T, type strains of A. pasteurianus subsp. ascendens and A. pasteurianus subsp. paradoxus, respectively, were cultured for 3 days in YPGD media (0.5% yeast extract, 0.5% peptone, 0.5% glycerol, and 0.5% d-glucose) containing 4% (v/v) ethanol. Genomic DNA was extracted using phenol-chloroform extraction and ethanol precipitation. The quality of purified genomic DNA was tested by using NanoDrop 2000 UV–Vis spectrophotometer (Thermo Scientific, MA, USA) and Qubit 2.0 fluorometer (Life Technologies, MA, USA).

The genomes of the two strains were sequenced at Macrogen using two different technologies: Illumina HiSeq and the PacBio single-molecule real-time technique with a 10-kb library (South Korea). De novo assembly of the read sequences was carried out using the hierarchical genome assembly process workflow. The annotation of the sequences was carried out using a modified version of the Prokka annotation pipeline, which incorporated Prodigal 2.60, Aragorn, and RNAmmer 1.2 for the prediction of open reading frames, tRNAs, and rRNAs, respectively (Seemann, 2014). Genome comparison among the two strains and other fully sequenced A. pasteurianus genomes was carried out by using Mauve software (Darling et al., 2010).

The complete genome sequences have been deposited in GenBank under the accession numbers CP015164-CP015167 (A. pasteurianus subsp. ascendens LMG 1590^T) and CP015168-CP015171 (A. pasteurianus subsp. paradoxus LMG 1591^T) in October 2016. The BioProject designations for strains LMG 1590^T and LMG 1591^T are PRJNA322127¹ and PRJNA317328,² respectively. A. pasteurianus subsp. ascendens LMG 1590^T and A. pasteurianus subsp. paradoxus LMG 1591^T are available from the BCCM/LMG Bacteria Collection under accession numbers LMG 1590 and LMG 1591, respectively.

We obtained 218,360 raw reads covering a total of 1,387,777,653 bp with 228× genome coverage for strain LMG 1590^T and 146,922 raw reads covering a total of 897,929,341 bp with 124× genome coverage for strain LMG 1591^T. The complete genome sequence of strain LMG 1590^T contained a circular chromosome of 2,859,878 bp with 53.1% G + C content and three circular plasmids with 55.4% G + C content. The genome sequence of LMG 1591^T was also assembled into a circular chromosome of 2,810,721 bp with 53.2% G + C content and three circular plasmids with 54.3% G + C content. The general features of the genomes are summarized in Table 1. Briefly, the analyses of the strain LMG 1590^T genome identified 2,931 genes. Among them 2,856 genes were annotated as coding DNA sequences (CDSs). A total of 3,163 genes were predicted from the genome of strain 1591^T of which 3,088 genes were identified as CDSs. The genome sequences data are available in FASTA, annotated GenBank flat file, graphical, and ASN.1 formats.

To investigate the overall genomic differences between the sequenced A. pasteurianus strains, including LMG 1590^T and LMG 1591^T, and the previously sequenced A. pasteurianus species, a global alignment of genome sequences from 13 strains was performed using Mauve software (Darling et al., 2010). The results showed that A. pasteurianus NBRC 101655, A. pasteurianus 386B, and eight A. pasteurianus IFO strains were quite similar with respect to genome structure in the chromosome (Figure 1), which was inconsistent with previous reports (Illeghems et al., 2013; Wang et al., 2015). Moreover, rearrangements, deletions, amplifications, and insertions occurred frequently in A. pasteurianus Ab3, A. pasteurianus subsp. ascendens LMG 1590^T, and A. pasteurianus subsp. paradoxus LMG 1591^T (Figure 1). Similar phenomena were also observed in the plasmids of these strains (Figure S1 in Supplementary Material).

The pyrroloquinoline quinone (PQQ)-alcohol dehydrogenase (ADH) and aldehyde dehydrogenase (ALDH) are responsible for the oxidative metabolism of ethanol to produce acetic acid in AAB. These proteins generally consist of three subunits: a quinohemoprotein catalytic subunit, a triheme cytochrome c subunit, and a third subunit with unknown function. Sequence analysis of the catalytic subunits of the ADHs from strains LMG 1590^T and LMG 1591^T by a combined transmembrane topology and signal peptide predictor indicated that both of the proteins had a signal peptide and were located in the periplasmic space (Käll et al., 2004). The ADHs were aligned with the well-studied ADH from Pseudomonas putida (50% identity to ADHs from LMG 1590^T and LMG 1591^T) (Xia et al., 1996), and the final output was processed using the program ESPript 3.0 (Robert and Gouet, 2014). The results indicated that the PQQ bound to the N-terminal portion, whereas the C-terminal end bound the heme c (Figure S2 in Supplementary Material). The ADHs and ALDHs from the 13 strains were further aligned using Clustal Omega (Sievers and Higgins, 2014). Both ADHs and ALDHs from the strains showed high identity (>98%; Tables S1 and S2 in Supplementary Material), although Acetobacter species are known to exhibit genetic instability (Azuma et al., 2009). Because ethanol and acetic acid tolerance could be partly attributed to the intrinsic properties of the amino acid sequences of the two proteins and high concentrations of ethanol would not cause mutations in their sequences (Trcek et al., 2006; Zheng et al., 2015), we proposed that the high conservation of the proteins may contribute to the stable industrial performance of A. pasteurianus.

In conclusion, the complete genomes of two A. pasteurianus subspecies were sequenced and assembled into one chromosome and three plasmids. Comparative genome and sequence analyses showed that rearrangements occurred in the A. pasteurianus strains and that the ADHs and ALDHs responsible for acetic acid production were highly conserved in these strains.