Assembly and Analysis of the Genome Sequence of the Yeast Brettanomyces naardenensis CBS 7540

Brettanomyces naardenensis is a spoilage yeast with potential for biotechnological applications for production of innovative beverages with low alcohol content and high attenuation degree. Here, we present the first annotated genome of B. naardenensis CBS 7540. The genome of B. naardenensis CBS 7540 was assembled into 76 contigs, totaling 11,283,072 nucleotides. In total, 5168 protein-coding sequences were annotated. The study provides functional genome annotation, phylogenetic analysis, and discusses genetic determinants behind notable stress tolerance and biotechnological potential of B. naardenensis.


Introduction
Yeasts belonging to the genus Brettanomyces (teleomorph name Dekkera, which should, according to the principle "one fungus, one name", no longer be used [1]) are of general interest, both as evolutionary models for the development of the fermentative lifestyle, as spoilage or production organisms in the generation of soft drinks, alcoholic beverages, and bioethanol production [2][3][4][5]. Brettanomyces yeasts have been firstly identified as spoilage organisms in beverages, but also as part of the natural microbial population in spontaneously fermented beers [2,6].
The acetate-producing, Crabtree-negative yeast B. naardenensis was previously shown to spoil carbonated beverages and soft drinks with pH ranging 2.6-3.2 [7,8]. The B. naardenensis strain CBS 7540 was isolated in 1990 from a soft drink in South Africa (Natal). Occurrence of B. naardenensis in Cabernet Sauvignon wine was also reported [9]. On the other hand, this yeast may also have some biotechnological potential. It was shown to be able to produce ethanol from xylose [10] and, unlike other species of the genus Brettanomyces, to assimilate soluble starch. The broad substrate range of Brettanomyces can contribute to reduction of residual sugar content in beverages production.

Cultivation
Two milliliters of B. naardenensis culture grown in 100 mL YPD at 25 • C to OD 10 were harvested. Aerobic batch cultivation of the strain for RNA isolation was performed using defined minimal media supplemented with 2% glucose in 1L Multifors fermenters (Infors HT, Bottmingen, Switzerland).

Genome Sequencing
The genomic DNA of B. naardenensis CBS 7540 was prepared using a standard zymolyase and phenol-chloroform extraction. A sample quality test was made for DNA sample, and the qualified sample was used to construct libraries. Covaris or Bioruptor was used to break the genomic DNA into smaller fragments (less than or equal to 800 bps). T4 DNA Polymerase, Klenow DNA Polymerase, and T4 PNK were used for end repair of the small DNA fragments. Three prime end A addition and adapter ligation were performed using taq-polymerase. The target DNA fragments were size-selected by electrophoresis and enrichment was performed using PCR amplification and purification. The qualified library was cluster prepared and sequenced. Genome sequencing was performed using the Illumina GA II (Illumina, San Diego, CA, USA) by generating multiplexed paired-end libraries (an average insert size of 500 bp). Raw fluorescent images and call sequences was processed using a base-calling pipeline (Solexa Pipeline-version 1.0).

RNA Library Preparation and Sequencing
Isolation and processing of RNA from triplicates of batch cultures were performed as previously described [29]. The RNA was treated with the RiboMinus™ Eukaryote Kit (Thermo Fisher Scientific, Waltham, MA, USA) to remove ribosomal RNA and purified using Agencourt RNAClean XP Kit (Beckman Coulter, Indianapolis, IN, USA). The size and quantity of RNA fragments were assessed on the Agilent 2100 Bioanalyzer system (RNA 6000 Pico kit, Agilent, Santa Clara, CA, USA). Libraries were prepared using the Ion Total RNA-Seq kit for the AB Library Builder System (Life Technologies). Samples were then quantified using the Agilent 2100 Bioanalyzer system (High Sensitivity DNA kit, Agilent) and pooled followed by emulsion PCR on the Ion OneTouch™ 2 system using the Ion Proton™ Template OT2 200 v2 Kit (Life Technologies, Waltham, MA, USA) chemistry. Templated Ion Sphere particles were enriched using the Ion OneTouch™ ES (Life Technologies). Samples were loaded on a two Ion PI v2 Chips and sequenced on the Ion Proton™ System using Ion Proton™ Sequencing 200 v2 Kit (200 bp read length, Life Technologies) chemistry.

Genome Assembly
The raw data (fastq format) were cleaned by FastQC and fastx_toolkit. The read size was 250 × 2 bp. We used SOAPdenovo1.05 to assemble the cleaned reads [30]. All raw reads were mapped onto the scaffolds using SOAPaligner [31] to evaluate the single base accuracy of the assembled genome sequences.

Transcriptome-Guided Genome Annotation
The annotation of the B. naardenensis CBS 7540 genome assembly was built in three steps: Prediction of candidate cDNAs using transcriptome data (assembled de-novo with Trinity and genome-guided with tophat and cufflinks) within the PASA package, followed by training of the SNAP ab-initio gene finder, and finally the actual gene build combining evidence alignments and ab-initio predictions using the Maker2 package. The annotation was deposited into European Nucleotide Archive (ENA) using EMBLmyGFF3 [32]. The annotated genome sequence of B. naardenensis CBS 7540 was deposited in the European Nucleotide Archive (ENA) with accession numbers CAACVR010000001 to CAACVR010000076 under project number PRJEB30032

Functional Genome Annotation
Functional genome annotation. Functional annotation of transcripts was performed using the Blast2Go pipeline (available online: http://www.blast2go.com/b2ghome), based on best BLAST matches of predicted cDNAs against the RefseqP fungal reference dataset. Annotation of B. naardenensis CBS 7540 was performed by searching the KOG (available online: http://www.ncbi.nlm.nih.gov/COG/) database using BLAST software.

SNP Analysis
To generate reads for analysis of variants, we resequenced the genome of B. naardenensis CBS 7540. The genomic DNA of B. naardenensis CBS 7540 was prepared using a ZYMO RESEARCH Quick DNA™Fungal/BacterialMiniprep KitCatalog No. D6005 according to manufacturer's instructions. The input DNA was quantified using Qubit dsDNA HS Assay Kit (Invitrogen, Waltham, MA, USA) and the sample purity was determined using NanoDrop. Libraries with Illumina-compatible adapters were constructed using the KAPA HyperPlus Library Prep Kit (ROCHE, Basel, Switzerland) by following the instructions of the manufacturer. The finished libraries were quantified with Qubit dsDNA HS Assay Kit (Invitrogen) and the average library size was determined using DNF-473 Standard Sensitivity NGS Fragment Analysis Kit (1 bp-6000 bp); (Agilent). The generated multiplexed paired-end libraries had an average library size of 584 bp. Whole-genome sequencing was performed using the MiSeq Reagent Kit v2, 500 Cycles (Illumina) on the MiSeq sequencing platform (Illumina, San Diego, CA, USA) with 10 pM flow cell loading. The sequencing chemistry employed 4-channel sequencing-by-synthesis (SBS) technology. Reads generated by CBS 7540 genome resequencing were mapped to the genome assembly of B. naardenensis CBS 7540 by using BWA version 0.7.4 [33]. Identification of the various variants (SNP and indels) was performed as described earlier [27].

Phylogenetic Tree Construction
The proteomes of Debaryomyces hansenii, Kluyveromyces lactis, Ogataea polymorpha, Pichia kudriavzevii, Yarrowia lipolytica, Komagataella pastoris, Candida albicans, and Saccharomyces cerevisiae were downloaded from Ensembl and NCBI databases. We first built the orthologous groups of within 12 yeast genomes. Then, 2328 gene groups, which had a strict and phylogeny-based one-to-one orthology relationship in all species included in the phylome, were respectively aligned by ClustalOmega, and trimmed using trimAl (gap-score cutoff 0.5, conservation score 0.5), and then concatenated into a single alignment. Finally, a maximum likelihood tree was constructed by MEGA. All nodes received the highest support in terms of approximate likelihood ratio tests and of a bootstrap analysis of 100 replications.

Comparative Analysis of Gene Content
Comparison of the gene content between genomes of B. naardenensis and B. bruxelensis was done using program BLAST 2.2.29+ as described earlier [27].

Genome Structure
The genome of B. naardenensis CBS 7540 was assembled into 76 contigs. The details of genome assembly are summarized in Table 1. The total nucleotide content was estimated to be 11,283,072 bp, of which only 752 nucleotide residues could not be unambiguously assigned. Such ambiguous nucleotide positions were restricted to 45 N-regions, which were potentially generated at contigs links sites during scaffolding. Contigs with lengths exceeding 10 kb represented 71% of the final assembly, which is comparable to other assemblies generated by SOAPdenovo1.05 [34]. Sequence length of the shortest contig, which along with the larger contigs, compose half of genome sequence (N50), is 395,348 bp, which suggested high genome contiguity. Furthermore, 10% of the genome assembly consisted of contigs shorter than 87,012 bp. The GC-content of B. naardenensis CBS 7540 genome assembly was found to be 44.5%, which is comparable to that of B. bruxellensis [34].

Genome Annotation
An advanced annotation approach was used to increase gene prediction accuracy ( Figure S1). In total, 5168 putative protein-coding sequences were identified and annotated, which is comparable to that of B. bruxellensis CBS 2499 [25]. The annotation details of the B. naardenensis CBS 7540 genome are presented in Table 2. Statistics of the CBS 7540 genome annotation are in line with other published Brettanomyces yeasts genomes. Functional genome annotation of the B. naardenensis CBS 7540 genome based on Blast2Go and KOG database is presented is Tables S1 and S2, respectively. The functional annotation is of particular use for future studies on omics analysis in this species.

Heterozygosity
Analysis of polymorphisms of the CBS 7540 genome was performed by mapping the CBS 7540 reads to the de novo assembly of the CBS 7540 genome. Reads generated by CBS 7540 genome resequencing mapped to assembly sequence of the CBS 7540 genome with 100 x times coverage. We detected 851 variants (Table 3 and Table S3), which is much lower than that in highly dynamic genome of B. bruxellensis CBS 11270 [27]. This could indicate a haploid or highly homozygous genome. We detected 323 single-nucleotide polymorphisms (SNPs), constituting 0.002% of the genome size. Transitions (i.e., purine-purine or pyrimidine-pyrimidine exchanges) were detected more frequently than transversions (Table 4).

SNP Type Counts
Most of the heterozygous features were observed in non-coding genome regions; 736 variants, which were detected outside the open reading frames, are summarized in Table S4; 88 variants occurred in coding sequences, and in total, 49 genes with SNPs were identified (Table 5). Only four genes encoding unnamed protein products had 10 variants or more: DEKNAAT104779 (24 variants), DEKNAAT104433 (14 variants), DEKNAAT102160 (12 variants), and DEKNAAT103651 (10 variants). The corresponding best BLAST hits in B. bruxellensis AWRI1499 were a predicted subunit of the CCR4-NOT complex (AWRI1499_4082), a predicted iron homeostasis modulator (AWRI1499_1657), hypothetical protein (AWRI1499_4852), and a predicted CCR4-NOT transcription subunit 3 (AWRI1499_3655). An additional 45 genes had between one and nine variants (Table S5). Moreover, 533 indels were found in the CBS 7540 genome. Micro/mini satellites were observed among some indels (Table S3). Indels varied in size from 1 to 129 nucleotides (Figure 1), which is similar to that in B. bruxellensis CBS 11270 [27]. The length of the indels inversely correlated to the frequency.  Moreover, 533 indels were found in the CBS 7540 genome. Micro/mini satellites were observed among some indels (Table S3). Indels varied in size from 1 to 129 nucleotides (Figure 1), which is similar to that in B. bruxellensis CBS 11270 [27]. The length of the indels inversely correlated to the frequency.

Phylogenetic Analysis
Previous analyses of phylogenetic relationships within the genus Brettanomyces have been contradictory [35,36]. The availability of a whole genome sequence of B. naardenensis reported here allowed us to perform a phylogenetic analysis based on the concatenation of all gene groups ( Figure  2). Our results support previous findings on relationships between the Brettanomyces species based on the complete 26S rRNA gene sequence analysis [35]. The wine-spoiling Crabtree-positive species, B. bruxellensis and B. anomala, are closely related. The Crabtree-negative yeast B. custersianus is less related to Crabtree-positive Brettanomyces species. B. naardenensis is the most distantly related to other Brettanomyces species. The last was also suggested by the D1/D2 tree [36].

Phylogenetic Analysis
Previous analyses of phylogenetic relationships within the genus Brettanomyces have been contradictory [35,36]. The availability of a whole genome sequence of B. naardenensis reported here allowed us to perform a phylogenetic analysis based on the concatenation of all gene groups (Figure 2). Our results support previous findings on relationships between the Brettanomyces species based on the complete 26S rRNA gene sequence analysis [35]. The wine-spoiling Crabtree-positive species, B. bruxellensis and B. anomala, are closely related. The Crabtree-negative yeast B. custersianus is less related to Crabtree-positive Brettanomyces species. B. naardenensis is the most distantly related to other Brettanomyces species. The last was also suggested by the D1/D2 tree [36].

Genes Associated with Food-Related Traits of B. Naardenensis CBS 7540
3.5.1. Assimilatory Pathways B. naardenensis can utilize a broad range of carbohydrates as carbon sources including galactose, maltose, xylose, trehalose, cellobiose, rhamnose, and arabinose. Sugar components of soft drinks could present a carbon source for B. naardenensis growth, resulting in turbidity or formation of microbial sediment in beverage product.
The CBS 7540 genome included a number of hypothetical genes predicted to be involved in xylose metabolism, including a putative D-xylulose reductase (DEKNAAC100143); a putative xylulokinase (DEKNAAC100346); a putative xylose/arabinose reductase (DEKNAAC102931); and a putative NADPH-dependent D-xylose reductase (DEKNAAC104786). It has previously been shown that the CBS 6042 strain of B. naardenensis could ferment xylose with 1.8 g/L of ethanol produced from 20 g/L of xylose [37]. Broad substrate utilization by non-conventional yeast species is a useful property for the reduction of residual sugar content in production of super-attenuated and lowercalorie beers [11].
Investigation of alternative nitrogen metabolism genes did not result in identification of the gene cluster (NIT genes) encoding for the nitrate assimilation pathway: Nitrate transporter (YNT1); nitrate reductase (YNR1); nitrite reductase (YNI1) in the genome of B. naardenensis CBS 7540. This is different to some B. bruxellensis strains [38] and could confer an advantage of this species over B. naardenensis in a nitrate-rich environment such as hopped beer.
B. naardenensis can survive in bottled drinks under oxygen-limited conditions. We did not identify a URA1 gene in B. naardenensis. The genome of B. bruxellensis was previously shown to contain URA9 (encoding oxygen-dependent mitochondrial dihydroorotate dehydrogenase (DHOD)) but not URA1 (oxygen-independent cytoplasmic DHOD) [39]. It remains unclear which alternative mechanism B. bruxellensis could use for uracil biosynthesis under anaerobic conditions [40]. This is different in S. cerevisiae, which has lost URA9 and employs URA1 for uracil synthesis under anaerobic conditions.

Genes Putatively Involved in Production of Volatiles
We have annotated genes of five alcohol dehydrogenases (ADH) in the B. naardenensis CBS 7540 genome: DEKNAAC101096, DEKNAAC103026, and DEKNAAC103616; ADH3 DEKNAAC105038 and a NADPH-dependent medium-chain ADH with broad substrate specificity DEKNAAC104540. Similar to B. bruxellensis, B. naardenensis has several independently duplicated ADH and ADH-like genes, which are presumably involved in metabolism of alcohols, including ethanol, and also other

Assimilatory Pathways
B. naardenensis can utilize a broad range of carbohydrates as carbon sources including galactose, maltose, xylose, trehalose, cellobiose, rhamnose, and arabinose. Sugar components of soft drinks could present a carbon source for B. naardenensis growth, resulting in turbidity or formation of microbial sediment in beverage product.
The CBS 7540 genome included a number of hypothetical genes predicted to be involved in xylose metabolism, including a putative D-xylulose reductase (DEKNAAC100143); a putative xylulokinase (DEKNAAC100346); a putative xylose/arabinose reductase (DEKNAAC102931); and a putative NADPH-dependent D-xylose reductase (DEKNAAC104786). It has previously been shown that the CBS 6042 strain of B. naardenensis could ferment xylose with 1.8 g/L of ethanol produced from 20 g/L of xylose [37]. Broad substrate utilization by non-conventional yeast species is a useful property for the reduction of residual sugar content in production of super-attenuated and lower-calorie beers [11].
Investigation of alternative nitrogen metabolism genes did not result in identification of the gene cluster (NIT genes) encoding for the nitrate assimilation pathway: Nitrate transporter (YNT1); nitrate reductase (YNR1); nitrite reductase (YNI1) in the genome of B. naardenensis CBS 7540. This is different to some B. bruxellensis strains [38] and could confer an advantage of this species over B. naardenensis in a nitrate-rich environment such as hopped beer.
B. naardenensis can survive in bottled drinks under oxygen-limited conditions. We did not identify a URA1 gene in B. naardenensis. The genome of B. bruxellensis was previously shown to contain URA9 (encoding oxygen-dependent mitochondrial dihydroorotate dehydrogenase (DHOD)) but not URA1 (oxygen-independent cytoplasmic DHOD) [39]. It remains unclear which alternative mechanism B. bruxellensis could use for uracil biosynthesis under anaerobic conditions [40]. This is different in S. cerevisiae, which has lost URA9 and employs URA1 for uracil synthesis under anaerobic conditions.

Genes Putatively Involved in Production of Volatiles
We have annotated genes of five alcohol dehydrogenases (ADH) in the B. naardenensis CBS 7540 genome: DEKNAAC101096, DEKNAAC103026, and DEKNAAC103616; ADH3 DEKNAAC105038 and a NADPH-dependent medium-chain ADH with broad substrate specificity DEKNAAC104540. Similar to B. bruxellensis, B. naardenensis has several independently duplicated ADH and ADH-like genes, which are presumably involved in metabolism of alcohols, including ethanol, and also other aromatic compounds [25]. Interestingly, a previous study suggested that in B. bruxellensis, one ADH gene is involved in both ethanol production and consumption [38], which is similar to respiratory yeasts Scheffersomyces stipitis [41] and Wickerhamomyces anomalus [42].
The enzyme beta-glucosidase hydrolyzes glucose from glycosides derived from hops or fruit ingredients of beverage forming volatile aglycones. Using blast search versus beta-glucosidase AWRI1499_4190 of B. bruxellensis AWRI1499, we identified two beta-glucosidase candidate genes in the genome of B. naardenensis CBS 7540: unnamed proteins DEKNAAC101161 (e-value 0.0) and DEKNAAC101664 (e-value 2e −111 ). This indicates that B. naardenensis has genetic potential for production of flavor-active aglycones. Further metabolic conversions of monoterpenes by complex brewing cultures can contribute to the generation of fruity or floral odors [15].
The absence of ATF1 and ATF2 genes involved in production of acetate esters disables Brettanomyces from producing the banana-flavored isoamyl acetate and honey-flavored 2-phenylacetate. However, the presence of esterase could potentially enable the formation of ethyl esters with tropic fruit-like odors: Ethyl acetate, hexanoate, and octanoate. Esterase DEKNAAC103380 was annotated in the genome of B. naardenensis CBS 7540. Esterification of cheesy-flavored long-chain fatty acids (C9, C10, C12, C14, C16) by Brettanomyces yeasts was suggested to contribute in formation of grape-like beer flavor [15]. Two other unnamed genes, DEKNAAC104471 and DEKNAAC100147, have high sequence similarity to DEKNAAC103380 and could potentially encode additional esterases.
Both S. cerevisiae and B. bruxellensis can convert cinnamic acids derived from grape skin into hydroxycinnamic acids using esterase following conversion into hydroxysterene using cinnamate decarboxylase. However, unlike S. cerevisiae, B. bruxellensis displays vinyl phenol reductase (VPR) activity, which catalyzes further conversion of hydroxystyrene into a volatile ethyl derivate [6]. Ethyl phenols are characteristic "Brett" off-flavor markers, which are described as "mousy" and represent molecular agents of wine or beer spoilage. VPR reaction was recently shown to be catalyzed by an enzyme with dual superoxide dismutase and NADH-dependent reductase activity [43]. A BLAST search of the B. naardenensis CBS 7540 genome with the B. bruxellensis ST05.12/26 VPR sequence identified a likely orthologous gene (DEKNAAC101290, e-value 5 × 10 −100 ), which suggested that this enzyme may play a role in off-odors formation in B. naardenensis as well.
Sensation of acetate is a matter for discussion-its presence is desirable for certain brewing styles and in other beverages is perceived as a sign of spoilage. In contrast to B. bruxellensis, which was reported as a potent acetate producer, B. naardenensis was described to generate moderate amounts of acetate under aerobic conditions [44]. The reason for the acetate overproduction phenotype is thought to be the insufficient activity of the acetyl-CoA synthetase responsible for the conversion of acetate to acetyl-CoA [45,46]. Several aldehyde dehydrogenases, DEKNAAC103751, DEKNAAC103492, and DEKNAAC101110; aldehyde dehydrogenase (NADP+) Ald4 DEKNAAC100618; and mitochondrial aldehyde dehydrogenase DEKNAAC103573 were annotated in the genome of B. naardenensis CBS 7540. We identified several B. naardenensis CBS 7540 gene candidates for acetyl-CoA synthetase, DEKNAAC103441, DEKNAAC102074, DEKNAAC102090.
The inhibition of alcoholic fermentation under anaerobic conditions-commonly known as the Custer effect-characteristic of yeasts belonging to the Brettanomyces/Dekkera genus [47] is associated with overproduction of NADH during acetate production. Low expression of NADH oxidizing enzymes of glycerol production pathway in B. bruxellensis [48] is not sufficient to provide a sink for NADH.
Alternative respiration involving salicylhydroxamic acid (SHAM)-sensitive alternative oxidase (AOX) was suggested to present an additional mechanism behind the Custer effect [39]. The hypothetical gene DEKNAAC101381 likely encodes AOX in B. naardenensis CBS 7540 (best BLAST hit is alternative oxidase AWRI1499_1979 with an e value of 10 -162 ).

Putative Stress Tolerance Genes
B. naardenensis was shown to survive in pH ranging from 2.6-3.2 [7,8]. The mechanism behind this notable stress tolerance is unclear. At low extracellular pH, undissociated weak acids become able to diffuse into the cytosol. The higher intracellular cytoplasm pH leads to dissociation of weak organic acids, which diminishes growth by the drop in intracellular pH, ATP depletion, intracellular accumulation of anions [49], and inhibition of DNA replication [50]. Plasma membrane ATPase pumps out protons, consuming ATP and thus declining biosynthetic processes. Weak organic acids can cause additional destructive effects. Acetic acid was shown to induce generation of reactive oxygen species (ROS) in the cell [51].
Studies on S. cerevisiae determined several genes conferring high acetic acid tolerance: A transcriptional activator involved in the response to weak acid stress, HAA1; CUP2, paralog of HAA1; glyoxylase gene GLO1 (an enzyme responsible for the detoxification of methylglyoxal, a side-product of the triose phosphate isomerase reaction in glycolysis); oxidative stress-resistance gene DOT5 encoding nuclear thiol peroxidase, and vacuolar membrane Atpase VMA7 [52]. The corresponding orthologous genes in the genome of B. naardenensis CBS 7540 were identified as DEKNAAC104025, DEKNAAC102923, DEKNAAC104208, and DEKNAAC101910, respectively.
On other hand, B. naardenensis has remarkably low thermotolerance, which could limit its dissemination in human-related habitats. B. naardenensis was able to grow at 25 • C but not 30 • C. This is different to B. bruxellensis, which can grow at 37 • C [53]. Since data on ethanol tolerance of B. naardenensis are scarce, it is unclear if ethanol levels can act as additional limiting factor.
The genome size 11.3 Mbp of B. naardenensis CBS 7540 is in the range of that determined for other Brettanomyces species [19][20][21][22][23][24][25][26][27][28]. Several strains of B. naardenensis, CBS 6040, CBS 6041, CBS 6042, CBS 6107, and CBS 6117, have been isolated from carbonated soft drinks in USA, Belgium, Netherlands, Norway, and Denmark, respectively [44]. CBS 7540 investigated in this study is a strain of African origin. A recent study showed that geographical localization and industrial fermentation environment of origin had similar impacts on the total genetic variance of the B. bruxellensis population, suggesting an anthropic influence on the spatial biodiversity of this microorganism [56]. Similar to B. bruxellensis, almost no isolates from "natural" non-human related habitats were reported for B. naardenensis. All above-mentioned B. naardenensis strains were described as soft drink spoilage microorganisms, suggesting notable species tolerance to low pH and osmotic stress. The low number of reports on the occurrence of B. naardenensis in wine [9] and beer [6] indicates that the species is less competitive in these environments than B. bruxellensis, B. anomala, and B. custersianus [57]. A genome hybridization event resulting in highly heterozygous allotriploid genotype of some Australian B. bruxellensis wine strains could confer a selective advantage of this species in the presence of the high levels of sulfur dioxide, a preservative used in winemaking [22,23,58]. In the present study, we identified genes involved in tolerance to low pH, which might provide deeper understanding of molecular basis behind occurrence of B. naardenensis in soft drinks.
The recent rise in craft and specialty beer production has awoken an interest in non-conventional yeast species and resulted in more careful consideration of organoleptic properties of fermented beverages flavor [15]. Organoleptic properties of beverages such as turbidity, presence of sediments, acidic taste, and special aroma profiles depends on brewing style and is the matter of personal preferences. We have determined the B. naardenensis genes putatively involved in metabolism of alternative carbon sources, which could contribute to production of innovative beverages with high attenuation degree and low alcohol content. A number of identified candidate genes responsible for generation of volatile compounds suggests that further substrate optimizations might allow for suppressing undesirable and enhancing pleasant aromas generated by B. naardenensis.
Interestingly, the fermentative lifestyle has evolved in one species of this genus, B. bruxellensis, while the other species such as B. naardenensis are still respiratory, i.e., they are only forming ethanol under oxygen limitation. In contrast to the Saccharomyces line, development of fermentative lifestyle in B. bruxellensis was not preceded by a whole genome duplication. There are indications that promoter re-wiring and amplification of fermentative genes such as alcohol dehydrogenase genes played a role in developing the fermentative live style [5]. Those re-organizations of metabolic regulation might have been promoted by a highly dynamic genome of B. bruxellensis [27]. In contrast to the highly dynamic genome structure of B. bruxellensis, a low number of identified variants points towards high genome stability of B. naardenensis CBS 7540. Low genetic polymorphism could limit the physiological plasticity of the species abolishing evolution towards fermentative lifestyle-this is consistent with preservation of respiratory phenotype of B. naardenensis.
The teleomorphic state of B. naardenensis [59] was shown to be an artefact caused by the staining method [60]. The high genome stability is rather uncommon for asexual Brettanomyces species. A high level of homozygosity between Saccharomyces strains was shown to be associated with sporulation and selfing phenomena [61,62]. Phylogenetic analysis showed that B. naardenensis diverged from other Brettanomyces species earliest, and therefore molecular physiology of this species might be distinct from other species of this genus.
The genome sequence of B. naardenensis provides a basis for further research efforts to understand notable stress tolerance and uncover the full biotechnological potential of this yeast in the production of new-generation beverages.

Conclusions
Despite frequent isolation of B. naardenensis from food-related environments, this species remains poorly investigated. Genome survey of B. naardenensis provides the first insights into the genetic landscape of this yeast. We present highly refined RNA-guided genome annotation. Additionally, functional annotations provide tools for further system biology studies on this species. Genome size and gene content of B. naardenensis is comparable to other annotated Brettanomyces species. However, high genome homozygosity is different from dynamic heterozygous genome of B. bruxellensis. Apart from highly diverged DNA and protein sequences between these two Brettanomyces species, promotor rewiring boosted differentiation of B. naardenensis and B. bruxellensis lifestyles into respiratory and fermentative, respectively. Our analysis supports distant phylogenetic relations between these species. Identification of genes involved in tolerance to low pH explains ability of B. naardenensis to spoil soft drinks. We have determined B. naardenensis genes that could contribute to production of innovative beverages with high attenuation degree and low alcohol content. Candidate genes responsible for generation of volatile compounds were identified, which provides the basis for future exploration the full potential of B. naardenensis.