Transcriptional analysis of Kluyveromyces marxianus for ethanol production from inulin using consolidated bioprocessing technology

Background Ethanol production from non-crop materials, such as Jerusalem artichokes, would make a great contribution to the energy industry. The non-conventional yeast, Kluyveromyces marxianus, is able to carry out ethanol fermentation of sugar molecules obtained from inulin-containing materials by consolidated bioprocessing. Lower inulin concentrations and micro-aeration can lead to a relatively fast and ideal fermentation process; however, it is unclear what causes the inhibition of higher concentrations of inulin and the promotion effect of aeration. Results Next-generation sequencing technology was used to study the global transcriptional response of K. marxianus Y179 under three fermentation conditions, including 120 g/L inulin without aeration (120-N), 230 g/L inulin without aeration (230-N), 230 g/L inulin with aeration by ORP controlling at −130 mV (230-130mV). A total of 35.55 million clean reads were generated from three samples, of which 4,820 predicted that open reading frames were annotated. For differential expression analysis, 950 and 1,452 differentially expressed genes were discovered under the conditions of 230-130mV and 120-N, respectively, and the sample 230-N was used as the control. These genes are mainly associated with the pathways of central carbon metabolism and ethanol formation. Increased expression of inulinase and the low activity of the autophagy-related gene, ATG8, ensured fast and ideal fermentation processes. Conclusions Despite being reported as the “crabtree-negative” species, K. marxianus Y179 could achieve favorable ethanol fermentation profiles under micro-aeration and high inulin concentrations. K. marxianus Y179 cells responded to inulin concentrations and micro-aeration that is involved in the whole ethanol metabolism network. These results will serve as an important foundation for further exploration of the regulatory mechanisms involved in ethanol fermentation from inulin by consolidated bioprocessing and also provide a valuable reference for future studies on optimization and reconstruction of the metabolism network in K. marxianus. Electronic supplementary material The online version of this article (doi:10.1186/s13068-015-0295-y) contains supplementary material, which is available to authorized users.


Background
The production of bioethanol from non-grain materials will be an important developmental trend in the fuel ethanol industry, especially in a country like China with a large population and relatively few arable lands. Recently, some inulin-containing materials, especially Jerusalem artichokes, have received increasing attention for their advantages, such as high yield, resistance to poor soil, drought, low temperatures, and pests [1][2][3]. Inulin in the roots and tubers of these plants, which is linked by fructose via a β-2,1 bond [4], is usually regarded as the main carbohydrate for ethanol fermentation.
Kluyveromyces marxianus, a sister species of Kluyveromyces lactis, is another type of "non-conventional" yeast that has been widely studied and has also been reported as a model of "crabtree-negative" yeast, in contrast to Saccharomyces cerevisiae [5]. Due to its high temperature resistance, rapid growth, and capacity to utilize various substrates, there are increasing applications of K. marxianus in biofuel industrial biotechnology. Furthermore, Open Access *Correspondence: ywj@dlut.edu.cn 1 School of Life Science and Biotechnology, Dalian University of Technology, Dalian 116024, China Full list of author information is available at the end of the article K. marxianus has been proven to be a key strain that can achieve ethanol fermentation from Jerusalem artichokes by consolidated bioprocessing technology (CBP), which combines inulinase production, inulin hydrolysis, and ethanol production [2, 6, 7].
The previous results of ethanol fermentation from the inulin of Jerusalem artichokes showed some problems under high inulin concentration conditions, such as residual inulin and a long fermentative time [7]. However, these problems did not happen under micro-aeration or lower inulin concentration conditions (our unpublished data). The molecular mechanisms of the inhibition of higher inulin concentrations and the effects of aeration are still unclear.
Although the whole-genome sequences of K. marxianus were first published in 2012 [8], publications on the transcriptional analysis of K. marxianus are very limited. Lertwattanasakul et al. [9] conducted complete genome sequencing and transcriptome analyses of K. marxianus DMKU 3-1042 to identify genes related to growth at a high temperature and the utilization of various substrates, especially xylose and arabinose. However, there have been no other reports on related information, especially ethanol fermentation by K. marxianus.
Considering the preponderance of a low background noise, less amount of RNA required, and the low cost involved in mapping the transcriptome of large genomes [10], in this article, the next-generation sequencing technology for RNA (RNA-seq) was utilized to study the global transcriptional response of K. marxianus Y179. The transcriptional response of K. marxianus Y179 was studied under various ethanol fermentative conditions [that includes low (12%) to high (23%) inulin concentration and in conditions ranging from anaerobic to micro-aeration].

Results and discussion
Physiological diversity for ethanol fermentation among various K. marxianus strains Though S. cerevisiae has been widely applied in industrial ethanol production, K. marxianus is now attracting an increasing amount of attention for its specific physiological diversity. As shown in Table 1, K. marxianus can use multiple substrates, even xylose and lactose, to produce ethanol. Moreover, although K. marxianus was known to have a poor ethanol tolerance compared to S. cerevisiae [12], the maximum ethanol concentration for K. marxianus fermentation broth is more than 100 g/L. Despite the differences of substrates [6, 11-18], the higher specific growth rates of K. marxianus guarantee superior results for various fermentative processes.

Ethanol fermentation by K. marxianus Y179 from inulin
Batch fermentation was performed to investigate the differences of gene expression between high and low inulin concentrations and from anaerobic conditions to microaeration. Higher concentrations of inulin repressed the whole process, but aeration promoted it [19,20]. Therefore, 120 and 230 g/L were selected as low and high inulin concentrations, respectively. An appropriate aeration, by controlling ORP at −130 mV [19], was set as the aeration condition for transcriptional analysis. The samples were taken at their own end of fermentation to erase the complicated chemical and biological changes that occurred during the fermentation process.
The physiology of ethanol fermentation by K. marxianus Y179 under three conditions are shown in Table 1 and Fig. 1. As reported previously [19,20], K. marxianus Y179 showed obvious catabolite repression under high inulin concentrations. Micro-aeration (ORP −130 mV) allowed K. marxianus Y179 to achieve the maximum ethanol concentration from inulin reported so far. The relatively high expression of inulinase by K. marxianus Y179 was in accordance with a previous study [20]. Consequently, the transcriptional analysis was conducted to investigate the differences of gene expression between high inulin and low inulin concentrations, and from anaerobic conditions to micro-aeration conditions. This annotation information provided a good foundation as a reference for RNA-seq ( Table 2).

Overview of transcriptional data by RNA-seq
After sequencing on the HiSeq 2000 system (Illumina), 1.44 Gb of raw reads were obtained by RNA-seq from all three samples, subsequently, and 35.55 million cleans reads were acquired after removing the adaptors and the low-quality reads, which accounted for 98% of the total reads (Additional file 1: Table S1). The clean reads were aligned to the reference genome and genes using the SOAP2 software, with a tolerance of no more than 2 bp mismatches. The number of reads that uniquely mapped the reference genome and genes accounted for ~74-89% and ~35-57%, respectively (Additional file 1: Table S1).

Quantification analysis of gene expression from K. marxianus Y179 Overall differential gene expression and validation by qPCR
The levels of gene expression, normalized as reads per kilobase per million mapped reads (RPKM), were applied to the fold changes of the differentially expressed genes (DEGs). The annotated genes were able to be covered by the clean reads from RNA-seq, ~66-93% of which achieved over a 90% coverage of the known genes ( Fig. 2a). Two comparison modules among the three samples, 230-130, mV-36 vs 230-N-72, and 120-N-24 vs 230-N-72, were performed to investigate the significantly differentially expressed genes between micro-aeration and hypoxic conditions, or between low and high inulin concentrations (Additional file 1: Figure S1). In the first module, of 1,840 DEGs, 950 genes had annotation information, including 696 genes from KEGG and 520 genes from GO, when setting the cutoff values of screening as FDR ≤0.001 and the absolute value of log2 ratio ≥1. In the other module, 1,452 of 2,658 DEGs were obtained with annotation information, including 1,095 genes from KEGG and 820 genes from GO. Thus, it could be seen that, compared with aeration and hypoxic conditions, a larger difference in gene expression between low and high inulin concentrations were observed during ethanol production by K. marxianus Y179.
A further statistical analysis of genes containing annotations in the two modules is illustrated in Fig. 2b-d, which showed that 497 mutual genes changed between two modules. Considering the up-or down-regulation of genes, it was remarkable that a great deal of genes were up-regulated (including genes related to ethanol metabolism, transcriptional factors, and so on) in the module of 120-N-24 vs 230-N-72, which might explain the inhibition of high inulin concentrations (Fig. 2c). The distribution of RPKM in 230-130mV-36 vs 230-N-72 was slightly higher than that of 120-N-24 vs 230-N-72, which might be due to the fewer DEGs in the module of 230-130mV-36 vs 230-N-72.
Six genes involved in various pathways were selected to conduct the qPCR experiments to validate the reliability of data from RNA-seq. As illustrated in Table 3, though there were a few differences between the fold changes of RNA-seq and qPCR, the trends of up-or down-regulation of all six genes chosen were the same, which consequently demonstrated the accuracy of the trends of gene expression change obtained by RNA-seq.

Characterization of DEGs involved in central carbon metabolism
Of 1,880 genes containing KEGG annotations in K. marxianus Y179, more than 88 KEGG metabolic pathways were categorized; 35 and 56 DEGs (fold-change ≥2) were related to the central carbon metabolic pathways (glycolysis/gluconeogenesis, TCA cycle, and PPP) or 230-130mV-36 vs 230-N-72 and 120-N-24 vs 230-N-72, respectively. According to these KEGG pathways, we reconstructed the metabolic network containing the central carbon metabolic pathway, ethanol and glycerol formation pathways, and the related intermediate enzymes (Figs. 3, 4).
Inulinase encoded by INU1 in K. marxianus Y179 might, to a large extent, affect the sugar consumption rates in ethanol fermentation. Compared with 230-N-72, micro-aeration (230-130mV-36) and low inulin concentration (120-N-24) up-regulated INU1 by 1.14and 1.94-fold, respectively, which was in accord with the inulinase activities of the supernatants (Table 1) and the previously published results [19,20]. Hexokinase and glucokinase, encoded by HXK1 and GLK1, respectively, which are crucial enzymes for controlling glucose and fructose in the central carbon metabolic pathway, were overexpressed by 2.2-and 3.0-fold under micro-aerobic and low inulin conditions in our systems, respectively. The overexpression of these three genes might explain the phenomenon of fast inulin consumption rates.
Phosphoglycerate mutase (encoded by GPM1) catalyzed the transition between 2-phosphoglycerate and 3-phosphoglycerate. Three kinds of isozymes were found in S. cerevisiae, encoded by GPM1, GPM2, and GPM3 [21,22]. Gpm1p and Gpm3p were detected in K. marxianus Y179, the identities of which were 87 and 57%, respectively, in contrast to that in S. cerevisiae. The expression of GPM1 under aerobic and low inulin conditions was down-regulated by 1.89-and 2.83-fold, respectively. Papini et al. [23] also made it clear that the GPM1 of S. cerevisiae were substantially up-regulated when using ethanol as the only carbon source. Consequently, the differential expression of GPM1 in K. marxianus Y179 was inferred to be associated with respiratory growth on non-fermentable substrates, like ethanol, to maintain the energy balance in the central carbon metabolic pathway. To enhance a high ethanol yield, silencing GPM1 in K. marxianus Y179 is an apparent alternative option to decrease the loss of ethanol as the substrate.
Pyruvate is the pivotal intermediate metabolite of the entire central carbon metabolism and determines the carbon fluxes to the TCA cycle or to ethanol formation. Pyruvate kinase, which catalyzes phosphoenolpyruvate into pyruvate, is regarded as the rate-limiting enzyme. Two kinds of pyruvate kinases, Pyk1p and Pyk2p (encoded by CDC19 and PYK2, respectively), have been observed in S. cerevisiae so far [24,25]. It was indicated previously that the expression of Pyk1p had nothing to do with the accumulation of ethanol [26], but it might be concerned with glucose utilization under the conditions of high temperature or resistance to environmental stress [24]. Only one kind of pyruvate kinase was detected in K. marxianus Y179, KmPyk1p, the identity of which was 86% homologous with Pyk1p in S. cerevisiae. However, the expressions of KmPYK1 under three conditions demonstrated few differences (Additional file 1: Table S2).
To our surprise, the expression of malate dehydrogenase (Mdh1p) in the sample of 230-130mV-36 increased by up to 4.6-fold. The enhanced transcription of mdh1p, a key enzyme in the TCA cycle, might be speculated to have other biological functions, because there was no obvious enhancement of the TCA cycle in 230-130mV-36. Previous results showed that MDH could be regarded as a transhydrogenase-like shunt, which regulated the redox state in S. cerevisiae cells [27]. What is more important, research from Lushchak et al. [28] indicated that the expression level of Mdh1p in SOD-defective S. cerevisiae was up-regulated to work as a protective mechanism under oxidative stress. Consequently, the up-regulation of KmMDH1 in 230-130mV-36 might have something to do with this. On the other hand, PPP provided and balanced NADP + /NADPH or NAD + /NADH ratio in cells, so the expression of genes involved in PPP were promoted under aerobic conditions (230-130mV-36) for the regulation and balance of the redox state in the cell (Figs. 3, 4).

Characterization of DEGs related to ethanol, glycerol, and by-products formation
Pyruvate was first converted to acetaldehyde by pyruvate decarboxylase (pdc), which was then reduced to ethanol by alcohol dehydrogenase (adh) in the pathway of ethanol formation. The pyruvate decarboxylase, PDC1, played was enhanced by roughly threefold, which agreed with previous results [26]. In addition, an appropriate aeration resulted in a slightly higher RPKM value of PDC1 (25,820 vs 22,334). As mentioned by Suleau et al.
[5], the up-regulation of PDC1 ensured carbon fluxing to the fermentative pathways rather than the TCA cycle, which indicated the relatively ideal results of micro-aeration by controlling ORP at −130 mV. Acetaldehyde is reversibly reduced to ethanol by alcohol dehydrogenase, which is a vital enzyme for the metabolism of ethanol in yeasts. At least seven kinds of alcohol dehydrogenase were found in S. cerevisiae, Adh1p-Adh7p [29]. Seven alcohol dehydrogenases detected in K. marxianus Y179 were analyzed using a phylogenetic tree (Fig. 5). The levels of expression (RPKM values) demonstrated the remarkable functions of four ADH genes (Adh1p-Adh4p), which could be classified into two types according to the phylogenetic tree. Adh1p and Adh2p were found to be located in the cytosol, were the main functional alcohol dehydrogenases under anaerobic conditions, and kept a high transcriptional level under very high gravity conditions [26,30], which was in accordance with our results in this article. Additionally, Adh3p and Adh4p were located in the mitochondria. Adh4p showed no structural or functional homologies with the other three enzymes [29], and its low RPKM value obtained in this study demonstrated a small role in ethanol metabolism in the presence of nonrespiratory carbon sources, as described previously [31]. As reported previously, ADH3-disrupted K. marxianus was more sensitive to the reactive oxygen species [31]. The expression of ADH3 in K. marxianus Y179 was dramatically up-regulated (over 10-fold) under aerobic conditions (230-130mV-36 against 230-N-72 and 120-N-24) (Fig. 5; Additional file 1: Table S2), which hence participated in the enhancement of respiration and played a significant role in the reduction reactions.
Glycerol, as a protective mechanism responding to environmental stress, played a key role in keeping high cell viabilities during ethanol fermentation. Two homologies of glycerol-3-phosphate dehydrogenase, key enzymes in glycerol metabolism, were discovered in S. cerevisiae (Gpd1p and Gpd2p), which exhibited an enhanced expression under high osmolarity and anaerobic conditions, respectively [32][33][34]. Our results revealed that there were also two kinds of glycerol-3-phosphate dehydrogenase in K. marxianus Y179, KmGpd1p and KmGpd2p. GPD1 from S. cerevisiae was reported to be significantly down-regulated, due to a lack of dissolved oxygen [26]; however, in our study, the KmGPD1 gene was apparently up-regulated under high inulin concentrations and anaerobic conditions (Additional file 1: Table  S2), which agreed with the concentration obtained by HPLC in the supernatant (data not shown).

Characterization of DEGs response to oxidative stress
Our previous fermentation indicated the delay of the process under relatively anoxic conditions and in turn an appropriate aeration volume promoted cell growth and ethanol production. From yet another angle, aeration resulted in the production of toxic compounds, called reactive oxygen species (ROS), including superoxide, peroxide, hydroxyl radical, and so on [35,36]. As a consequence, a thorough protective mechanism exists in K. marxianus Y179 cells to protect it from damage by ROS.
The Trx-TrxR system is widely present in most yeast mitochondria and acts as a defense against ROS damages. Under aerobic conditions, the TRXR gene increased by 1.0-fold (230-N-72) and 3.4-fold (120-N-24), respectively. Glutathione reductase achieved the transition between oxidized and reduced glutathione, which was up-regulated by 1.5-to 2.4-fold under aerobic conditions. Peroxidase consumed ROS like O 2 2− via the reduction of H 2 O 2 , which mainly includes catalase and glutathione peroxidase, and aeration enhanced them by 2.2-to 4.3-fold (230-N-72) and 2.3-to 2.6-fold (120-N-24), respectively. Cu-Zn superoxide dismutase was also up-regulated by roughly 1.0-fold under aerobic conditions to remove the superoxide anion in cells. Blanco et al. [35] reported the responses of gene transcriptional levels in K. lactis to hypoxic and oxidative stress, and they found that the TRXR and CTT1 genes in wild-type K. lactis were obviously up-regulated during the aerobiosis-hypoxia shift. In spite of ROS being generated  Table 5.
by aeration (230-130mV-36), the obvious up-regulation of various genes related to oxidative stresses removed most ROS damages and guaranteed relatively high cell viabilities to achieve an ideal result. A type of alkyl hydroperoxide reductase (encoded by KmallA2475) was worthy of attention because of its high transcriptional level (RPKM up to 20,000), which increased by 1.3-to 3.3-fold compared to that under anoxic conditions. On the basis of sequence analysis, it belonged to one of the five types of peroxiredoxin (Tas1p) [37], which occupied the highest concentration in cells [38]. The homologous identity of KmallA2475 with that in S. cerevisiae was up to 90%, and it could be predicted that KmallA2475 in K. marxianus Y179 might be one of the most important enzymes for defense against ROS damages.
In addition, the high transcriptional level of a polyamine transporter (encoded by TPO1) piqued our interests (Additional file 1: Table S2). Endogenous polyamine protects cells from damage by some fermentation inhibitors and ethanol by binding to plasma membranes, ribosomes, and DNA [39]. Aranda et al. [40] found that the addition of acetaldehyde upregulated TPO1 in S. cerevisiae. Moreover, a TPO1defective strain showed an enhanced tolerance of fermentation inhibitors, such as furfural and acetic acid [41]. However, the aerobic conditions increased the TPO1 gene expression by 2.6-to 3.7-fold accordingly in our study. Though the specific underlying reasons and mechanisms are still unclear, we predict that this may be related to the acetaldehyde content in ethanol metabolism and that it may belong to a kind of protective mechanism for cells, especially for defending against damages by ROS from aeration by transporting a special type of polyamine into cells.

Characterization of DEGs associated with sugar transporters
Considering the homology between K. marxianus and K. lactis, we aligned the predicted hexose transporter genes in K. marxianus Y179 to the known transporters in K. lactis and obtained a similar transport system (Table 4). K. lactis embraced a relatively simple transport system and related genes [42]. The transport of hexoses by K. lactis mainly relied on two transporters: a low-affinity glucose transporter, rag1p [43], and a highaffinity glucose transporter, hgt1p [44]. In most K. lactis cells, rag1p was usually replaced by two high homologous proteins, Kht1p and Kht2p [42]. In addition, K. lactis also expressed the FRT1 gene, which encoded a hexose transporter with a high affinity with fructose, which contributed to parts of hexose transport [45].
Rag1p/kht1p in K. marxianus Y179 was induced by high concentrations of glucose or fructose, and we detected two predicted open reading frames (ORFs) that encoded the protein (KmallA1640 and KmallA1641). Kht2p, encoded by three ORFs (KmallA1637, KmallA1638, and KmallA1639), was up-regulated under low glucose concentrations. The up-regulation of two high-affinity transporters, Hgt1p and Frt1p, in the sample of 120-N-24 might account for the low residual sugar concentration. Therefore, we speculated that if the expression of these transporters were increased at the first stage of fermentation, the process of ethanol production might be accelerated to some extent.

Characterization of DEGs referring to transcriptional regulations
Transcriptional regulation may be associated with transcription factors (TFs). The phenomenon of glucose repression exists widely in yeast cells, where mig1p, encoded by MIG1, may be a crucial TF. The MIG1 gene had been detected in various yeasts, including S. cerevisiae, K. lactis, and K. marxianus [46,47]. Cassart et al. [47] demonstrated that the expression of MIG1 was regulated by glucose in the media, and the promoters of invertase (encoded by SUC2) and inulinase (encoded by INU1) were repressed by Mig1p in the presence of glucose. In our experiments, the enhanced expression of KmMIG1 was observed under high inulin and anoxic conditions (RPKM = 1,308, Additional file 1: Table S2). After the employment of low inulin and aerobic conditions, the expression of KmMIG1 was down-regulated by 1.4-to 2.5-fold. Consequently, an appropriate manipulation of the MIG1 gene (for example, lowering or silencing its expression at the initial stage of fermentation) may improve some key enzymes' activities, like inulinase in the process of ethanol production from inulin, thus fundamentally solving the complex problems.
In contrast to the behavior of MIG1, the CCR4/POP2 regulation system contributes to glucose derepression. Ccr4p is a complex that regulates the expression of various genes in yeast cells. Known to be part of the Ccr4p complex [48,49], similar to that of Pop2p, Ccr4p can resist glucose repression with the assistance of glucoserepressible alcohol dehydrogenase (encoded by AHD2) [49]. The relatively low transcriptional levels of the two TFs in K. marxianus Y179 (KmPOP2 and KmCCR4), which were homologous to those in S. cerevisiae, indicated that the CCR4/POP2 system did little to regulate gene expression in K. marxianus Y179. The inconspicuous differences in the expression of ADH2 among three samples had also proven the above conclusion ( Fig. 5; Additional file 1: Table S2).
On the other hand, the expression of some functional proteins will also affect the gene expression and the maintenance of their activities. The heat shock protein (HSP) family is a family of chaperones that assist proteins to fold correctly and maintain activities under some strict environmental stresses. Hsp26 had a very high transcriptional level in K. marxianus Y179 during ethanol fermentation, up to 53% compared to that in S. cerevisiae, which was one of the ten highest expressed genes in all three samples (Additional file 1: Table S2). Amorós et al. [50] discovered that HSP26 in S. cerevisiae was up-regulated under carbon source starvation and oxidative and osmotic stresses. As a consequence, it can be inferred that the enhanced expression of KmHSP26 in K. marxianus Y179 may be related to the low inulin and high ethanol conditions at the final stage of fermentation.
Hsp42 was also previously reported in S. cerevisiae under ethanol fermentation conditions [26]. The expression profile of Hsp42 was also found to be correlated with an increase in ethanol concentrations, and its expression was induced by high ethanol concentrations, which was consistent with our results in this article. A 1.0-fold up-regulation of KmHSP42 expression was observed under high ethanol and anoxic conditions (Additional file 1: Table S2), which might be conducive to conserving the enzyme activities of cells under very high ethanol concentrations. Hsp31 was induced by increasing oxidative stress in S. cerevisiae [51]. In accordance with that, we found that KmHSP31 was up-regulated by 2.7-to 3.6-fold under aerobic conditions (230-130mV-36) in contrast to those in anoxic conditions, which might reduce the damages of enzymes in cells under high oxidative stress.

Characterization of DEGs related to autophagy
Autophagy is supposed to be a conserved catabolic process in eukaryotes to maintain a steady state by the degradation of damaged cellular organelles and other dysfunctional macromolecules, like proteins [52]. Autophagy has been proven to make great differences and to guarantee the survival of yeasts under various stressful conditions [53]. There were great differences observed between the viability of cells under three conditions (Fig. 6a), which might result in different fermentative profiles. It was inferred that cell viabilities were greatly related to the fermentation time and the residual sugar concentration at the end. Consequently, having a better understanding of DEGs related to autophagy may provide references for optimizing the fermentative process. It was detected that non-selective autophagy played a vital role for maintaining the cell survival in the systems of ethanol production from inulin by K. marxianus Y179 (Fig. 6). Non-selective autophagy, also called macroautophagy, engulfs the cytoplasm to form autophagosomes, which comprise the main degradation process in eukaryotes [54]. Three DEGs related to autophagy, ATG3, ATG7, and ATG8 (Fig. 6b; Table 5), were identified to contribute to the formation of autophagosomes [55]. Atg8, a ubiquitin-like protein, was up-regulated by almost 3.0-fold under the condition of 230-N-72, which was in line with the results of cell viabilities (Fig. 6a). Moreover, it was reported that when macroautophagy occurred in cells, the expression of ATG8 would greatly increase to demonstrate a positive correlation with the size of autophagosomes [55,56]. Hence, the expression of ATG genes

Conclusions
A global transcriptome analysis of the non-conventional yeast, K. marxianus Y179, was conducted for ethanol production under various fermentative conditions. Hundreds of differentially expressed genes associated with the whole ethanol metabolism network were identified. Despite being reported as the "crabtree-negative" species, K. marxianus Y179 could achieve favorable fermentative profiles with an appropriate aeration volume, which had been verified by its transcriptional response. Genes related to oxidative stress and autophagy were apparently up-regulated or down-regulated to guarantee high cell viabilities, which were supposed to be the primary reasons of the ideal result. Moreover, the release of glucose repression and increased inulinase activity at the initial stage of fermentation, by controlling the transcription factor Mig1, may help to improve productivity. Consequently, information on the genome annotation and transcriptome data provided a valuable reference for future studies on the optimization and reconstruction of the metabolism network in K. marxianus.

Strains, media, and culture conditions
Kluyveromyces marxianus Y179 was deposited at the China Center for Type Culture Collection (CCTCC) with a reference number of M202031. The strain was maintained on a YPD agar slant composed of 20 g/L glucose, 20 g/L peptone, and 10 g/L yeast extract at 4°C for routine use. For long-term preservation, cultures were stored at −80°C in 20% glycerol.

Genome annotation
The genome information of K. marxianus was downloaded from the NCBI database (http://www.ncbi.nlm. nih.gov/Taxonomy/Browser/wwwtax.cgi?id=1162310) [8]. The whole-genome sequences above were annotated according to the Gene Ontology (GO) database and Kyoto Encyclopedia of Genes and Genomes (KEGG) database to obtain the reference genome and annotated gene information.

Transcriptome analysis of K. marxianus Y179 Cell growth conditions
Yeast cells that were cultivated in 100 mL of YPD medium in 250-mL flasks for 16-18 h were inoculated into a 3-L fermenter, with a 1.0-L working volume, at 30°C and 150 rpm. To study the transcriptional differences at high concentrations of substrate and aeration, three fermentation conditions were carried out: 230 g/L inulin in YP medium (20 g/L peptone, 10 g/L yeast extract) without aeration (230-N-72), 230 g/L inulin in YP medium controlling the redox potential at −130 mV (230-130mV-36), and 120 g/L inulin in YP medium without aeration (120-N-24). The ORP throughout the process was controlled by a sterile ORP electrode and a PID (Proportional, Integral, and Derivative) control algorithm, as described previously [26].

RNA isolation
Samples were taken after fermentation, under three conditions, at 72, 36, and 24 h, respectively. Cell pellets were collected by centrifugation at 5,000g at 4°C for 5 min, and were then frozen by liquid nitrogen. The total RNA of every sample was extracted by the RNeasy ® Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The 2130 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) was used to determine the RNA quality and integrity, and the RNA integrity number (RIN) of each sample used for RNA-seq was no less than 6.5.

Preparation of cDNA library and sequencing
The cDNA libraries were constructed and sequenced at the Beijing Genomics Institute (BGI, Shenzhen, China). The final cDNA libraries were qualified and quantified using the 2130 Bioanaylzer (Agilent Technologies) and the StepOnePlus Real-Time PCR System (Applied Biosystems). Afterward, the library products were ready for sequencing via the HiSeq 2000 system (Illumina). To detect the differences of gene expressions among various samples, single-end technology was used to obtain about 40-50 bp reads in a single run.

Quantification analysis of gene expression by RNA-seq
To annotate and quantify the genes expressed in K. marxianus Y179, we first calculated the percentage of the gene covered by reads (gene coverage), which was determined as the ratio of the base number in a gene covered by unique mapping reads to the total base numbers of that gene. The levels of the differentially expressed genes (DEGs) were normalized by RPKM, defined as the number of reads per kilobase of the exon region per million mapped reads. Furthermore, in this article, the log2 ratio of RPKM between different samples was used to calculate the fold-change values. To screen the DEGs among various samples, we used a false discovery rate (FDR) of ≤0.001 and the absolute value of the log2 ratio ≥1 as the threshold to judge the significance of differential gene expression. More stringent criteria with a smaller FDR and a larger fold-change value can be used to identify DEGs.

Analysis of the DEGs of K. marxianus Y179 grown in different culture conditions
Expression pattern analysis of DEGs was clustered using Cluster software [57] and Java Treeview software [58]. The hierarchical clustering of the chosen experimental conditions and genes was carried out using the Euclidean distance as the formula of the distance matrix. Gene Ontology (GO) enrichment analysis and functional classification helped to simplify huge amounts of data from RNA-seq and to identify the most significant DEGs from thousands of genes. To understand the distribution of gene functions of the species on the macro level, GO functional classification for DEGs was applied using the WEGO software [59] after obtaining the GO annotation of DEGs, as described above.
Kyoto Encyclopedia of Genes and Genomes was used as the major public pathway-related database to conduct the pathway enrichment analysis to further understand the biological functions of the genes. KEGG enrichment analysis identifies the significantly enriched metabolic pathways or the signal transduction pathways in DEGs, compared with the whole-genome background, of which the Q value was defined as no greater than 0.05.
A brief introduction and procedure of the transcriptome analysis of K. marxianus Y179 is illustrated in Additional file 1: Figure S1.

RT-PCR and real-time quantitative PCR
RT-PCR and real-time quantitative PCR (qPCR) were performed to verify the accuracy of the gene expression levels obtained by RNA-seq. The same total RNA samples for RNA-seq were utilized to conduct the PCR reactions. The total RNA of each sample (1 µg) was reverse transcribed to cDNA using the Prime-Script ® RT reagent Kit (Takara Bio Inc.) according to the manufacturer's protocol. Afterward, six differentially expressed genes involved in EMP, TCA, glycerol and ethanol formation, and stress response pathways were selected to perform qPCR, and the information of the primers is shown in Additional file 1: Table S3. A 10 times dilution of cDNA solution was used as the DNA template for the qPCR reaction and ddH 2 O was used as the negative control. A two-step PCR reaction was employed, and the system and conditions were as described in the manufacturer's protocol of SYBR ® Premix Ex Taq ™ II (Takara Bio Inc.). Finally, the actin gene was selected as the endogenous reference gene, and the data analysis of the fold change of a specific gene was determined by the method of 2 −∆∆C T as described by Livak and Schmittgen [60]. Triplicate experiments were performed to guarantee the reproducibility of all the results.

Biomass, sugar, and ethanol determination
The cell concentration was measured using the optical density at 620 nm. Concentrations of the reducing sugar in inulin were determined by the dinitrosalicylic acid method [61], and the total sugar was firstly hydrolyzed by 0.2 M H 2 SO 4 at 100°C for 1 h, then measured by the methods of Miller [61] after adding an equivalent amount of 0.4 M NaOH.
The ethanol was analyzed by gas chromatography (Agilent 6890A, Agilent Technologies) as previously described [19]. Briefly, a hydrogen flame ionization detector and isothermal capillary column (solid phase: cross-linked polyethylene glycol, carrier gas: nitrogen, and injector temperature: 250°C) were operated at 250 and 120°C, respectively.

Measurement of inulinase activity
Inulinase activity was measured according to the previous method [19,20]. Briefly, 0.5 mL of culture supernatant was incubated with 2% (w/v) inulin prepared in 0.02 M sodium acetate buffer (pH 4.6) at 55°C for 10 min, and the reducing sugar was analyzed by the dinitrosalicylic acid method [61]. One enzyme unit was defined as the amount of fructose (μM) hydrolyzed per minute under the above conditions. Fructose was used as the standard substance to plot a standard curve.

Cell viability detection
The viabilities of the K. marxianus Y179 cells were roughly detected by the methylene blue stain technique [62].
Analysis of the biomass, sugars, ethanol, inulinase activity, and cell viabilities were done in duplication, and the mean values are shown in the results section.