Physiological effects of overexpressed sigma factors on fermentative stress response of Zymomonas mobilis

Abstract

Zymomonas mobilis is a bacterium of industrial interest due to its high ethanol productivity and high tolerance to stresses. Although the physiological parameters of fermentation are well characterized, there are few studies on the molecular mechanisms that regulate the response to fermentative stress. Z. mobilis ZM4 presents five different sigma factors identified in the genome annotation, but the absence of sigma 38 leads to the questioning of which sigma factors are responsible for its mechanism of fermentative stress response. Thus, in this study, factors sigma 32 and sigma 24, traditionally related to heat shock, were tested for their influence on the response to osmotic and ethanol stress. The overexpression of these sigma factors in Z. mobilis ZM4 strain confirmed that both are associated with heat shock response, as described in other bacteria. Moreover, sigma 32 has also a role in the adaptation to osmotic stress, increasing both growth rate and glucose influx rate. The same strain that overexpresses sigma 32 also showed a decrease in ethanol tolerance, suggesting an antagonism between these two mechanisms. It was not possible to conclude if sigma 24 really affects ethanol tolerance in Z. mobilis, but the overexpression of this sigma factor led to a decrease in ethanol productivity.

Introduction

Zymomonas mobilis is a Gram-negative, ethanogenic bacterium with many desirable characteristics for its industrial utilization, such as low biomass production, high sugar uptake, and high ethanol tolerance. However, the mechanisms for stress response are not yet fully understood. Bioethanol fermentation may produce a number of harmful factors and directly affects productivity and commercial viability of the strains. Thus, a better understanding of the molecular regulation associated with stress response in Z. mobilis is important.

In bacteria, alternative sigma factors are responsible for promoting transcriptional changes in response to environmental conditions. In Escherichia coli, seven different sigma factors are known [1]. Sigma 70 (rpoD) is traditionally the housekeeping transcriptional factor [2]; sigma 38 (rpoS) is highly expressed in stationary phase and it is the master regulator of general stress response [3]. Sigma 32 (rpoH) and sigma 24 (rpoE) are associated with heat shock response and widely distributed in proteobacteria [4–6].

Z. mobilis ZM4 has five predicted sigma factors in the genome annotation (Table 1): sigma 70 (ZMO1623), sigma 54 (ZMO0274), sigma 32 (ZMO0749), sigma 28 (ZMO0626), and sigma 24 (ZMO1404) [7]. Sigma 70 is the only one experimentally tested so far [10, 11], being the most studied sigma protein. However, the complete mechanisms of transcription activation and even promoter recognition sequences are not entirely described in many bacteria, including Z. mobilis [12, 13]. Until now, sigma 38 has not been identified in Z. mobilis ZM4, which leads to the possibility that the heat shock regulators sigma 32 and sigma 24 might play a role in fermentative stress response.

Table 1.

Sigma factors in E. coli and their equivalent in Z. mobilis

Sigma factor	Gene	Functiona	Equivalent in Z. mobilisb	Similarity levelc
Sigma 70	rpoD	Housekeeping sigma	ZMO1623	63%
Sigma 54	rpoN	Nitrogen regulation	ZMO0274	55%
Sigma 38	rpoS	Stationary phase (stress)	–	–
Sigma 32	rpoH	Heat shock (cytoplasm)	ZMO0749	61%
Sigma 28	fliA	Flagella biosynthesis	ZMO0626	58%
Sigma 24	rpoE	Heat shock (periplasm)	ZMO1404	49%
Sigma 19	fecI	Ferric citrate transport	–	–

^aAdapted from Cook and Ussery (2013). ^bPredicted in genome annotation [7, 8]. ^cSimilarity level based on protein sequence alignment [9]

In E. coli, the transcriptional factor sigma 24 is an extra cytoplasmic stress response sigma factor (ECF), directly associated with cell envelope integrity. It controls the composition and protein folding in the cell envelope, and can be induced by heat, ethanol, and osmotic stresses [5, 14–17]. In alphaproteobacteria, a general stress response (GSR) sigma factor is frequently reported as an ECF protein [18] that can regulate the response to a number of stress conditions. Z. mobilis sigma 24 is homologous to the GSR sigma factor of other alphaproteobacteria, like Sphingomonas sp. [19], but it does not belong to any characterized ECF type. Genome analysis of Z. mobilis ZM4 has suggested that sigma 24 might be the key transcriptional factor for the response to high ethanol conditions [7], which has been endorsed by a transcriptome profile of Z. mobilis under ethanol stress [20], but no supporting experiment has been undertaken until now.

Sigma 32 is the traditional heat shock sigma factor, and it is very well conserved among bacteria [21]. In Bacillus subtilis, sigma 32 (Sigma-H) is essential for the initiation of sporulation at high temperatures [22], and it is found associated with cell membrane regulating the status of protein folding in E. coli [23]. In Z. mobilis, there are no published studies with sigma 32, but it was found to be overexpressed at high glucose concentration [24].

In this work, we proposed to investigate the role of sigma 32 and sigma 24 in the fermentative stress response of Z. mobilis, using an overexpression approach with physiological and metabolic analysis. Sigma 70 was also analyzed as a reference control.

Results

Construction and characterization of sigma 70–, sigma 32–, and sigma 24–expressing strains

In order to overexpress the sigma factors, an expression system was designed based on pBBR1MCS [25], a broad host plasmid that replicates in Z. mobilis. Considering that Z. mobilis ZM4 is a small bacterium with low levels of protein production, the designed expression system was tested first with a reporter gene, the alpha-amylase from Bacillus subtilis (amyE). The coding region of amyE gene was codon optimized for Z. mobilis and fused with the signal peptide of ZMO0130, a protein from Z. mobilis ZM4 [26]. This reporter gene was synthetized under the control of Z. mobilis ZM4 pyruvate decarboxylase (pdc) gene constitutive promoter and terminator and the resulting construct was cloned into pBBR1MCS, creating plasmid pB1_Amy.

The recombinant plasmid was transformed into competent Z. mobilis ZM4 and then, transformed clones were plated in selective Rich Medium with 10 g/L starch for hydrolysis observation (Fig. 1A). After transformation, 93 colonies were grown in selective media and all of them presented hydrolysis halo in starch, confirming the efficiency of the expression system.

Fig. 1.

Analysis of the expression system in Z. mobilis ZM4. (A) Colonies of Z. mobilis ZM4 transformed with pB1_Amy (1, 2, 3, and 4) and the empty plasmid pB1 (C1, C2, C3, and C4) grown in Rich Medium containing 100 mg L⁻¹ chloramphenicol and 10 g L⁻¹ starch. The plate was stained with iodine vapor. (B) Gene expression levels of Z. mobilis ZM4 strains. RNA extracted from cultures grown until OD_600nm = 1.0 were quantified by real-time RT-PCR with specific primers for the genes of sigma 70, sigma 24, and sigma 32 in the strains ZM4_p70, ZM4_p24, and ZM4_p32, respectively. mRNA fold change was calculated relatively to the same gene in the control strain (ZM4_pB1) and normalized by RNA 16S expression. Data is presented as the mean values of samples run in triplicate three times and the error bars are shown. *The difference to the control strain is statistically relevant in a t test (p < 0.001)

After this initial test, the amyE gene with Z. mobilis signal peptide was replaced by the coding region of sigma 70, sigma 32, and sigma 24, resulting in three vectors that express these sigma factors. The resulting plasmids were transformed into competent Z. mobilis ZM4 and the three resulting strains (ZM4_p70, ZM4_p32, and ZM4_p24) and an additional control strain (ZM4_pB1) containing the empty plasmid were used for further experiments (Table 2).

Table 2.

Plasmids and bacterial strains used in this work

Plasmid or strain	Genotype and/or relevant featuresa	Source or reference
pBBR1MCS	Broad-host-range vector, Cmr	[25]
pB1_Amy	pBBR1MCS with pdc promoter and terminator from Z. mobilis and amyE from B. subtillis, Cmr	This work
pB1_p70	pBBR1MCS with pdc promoter and terminator and sigma 70 gene from Z. mobilis, Cmr	This work
pB1_p24	pBBR1MCS with pdc promoter and terminator and sigma 24 gene from Z. mobilis, Cmr	This work
pB1_p32	pBBR1MCS with pdc promoter and terminator and sigma 32 gene from Z. mobilis, Cmr	This work
Z. mobilis ZM4 ATCC 31821	Wild-type strain	[27]
Z. mobilis ZM4_pB1	Z. mobilis ZM4 strain containing pBBR1MCS plasmid	This work
Z. mobilis ZM4_Amy	Z. mobilis ZM4 strain containing pB1_Amy plasmid	This work
Z. mobilis ZM4_p70	Z. mobilis ZM4 strain containing pB1_p70 plasmid	This work
Z. mobilis ZM4_p24	Z. mobilis ZM4 strain containing pB1_p24 plasmid	This work
Z. mobilis ZM4_p32	Z. mobilis ZM4 strain containing pB1_p32 plasmid	This work
E. coli DH5a ATCC 47093	F−endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupGφ80dlacZΔM15 Δ(lacZYA-argF)U169 hsdR17 λ−	Invitrogen
E. coli JM110 ATCC 47013	rpsL (Strr) thr leu thi-1 lacY galK galT ara tonA tsx dam dcm supE44 ∆(lac-proAB) [F ´ traD36 proAB lacIqZ∆M15]	Stratagene

^aCm^r, chloramphenicol resistance

The first step to characterize the new strains was to quantify the expression levels of each gene. So, total RNA was extracted in the beginning of exponential phase and the mRNA levels of each sigma factor gene were quantified by RT-qPCR (Fig. 1B). The levels of expression of the sigma 70, sigma 24, and sigma 32 genes were 1.84 ± 0.20-, 161.9 ± 8.9-, and 61.5 ± 4.2-fold higher in the sigma-expressing strains than in the control strain, respectively. The lower increase in relative expression of sigma 70 is expected, considering that this protein is highly expressed during exponential growth phase. The use of the 16S RNA gene for data normalization might also have underestimated the RNA levels, since this gene might be regulated by sigma 70.

Stress response analyses

After the confirmation that all strains were expressing the specific sigma factors, an initial characterization of stress response was done based on the growth profile of the four strains in heat, osmotic, and ethanol stress conditions. The growth curves were analyzed and compared with the control strain (Fig. 2).

Fig. 2.

Growth of Z. mobilis strains ZM4_pB1, ZM4_p70, ZM4_p24, and ZM4_p32 in RM in (A) standard conditions, (B) 39 °C, (C) 50 g L⁻¹ ethanol, and (D) 160 g L⁻¹ glucose. Data is presented as the mean values of samples run in quintuplicate and the standard error bars are shown

A central composite rotatable design (CCRD) was made to analyze the relation between heat, ethanol, and osmotic stresses. This experiment can show us a better understanding about Z. mobilis growth in a wide range of combined stress conditions. A response surface was constructed by measuring biomass production rates from Z. mobilis strains under heat and ethanol stress variations in two different glucose concentrations (Figs. 3 and 4). The goal of a CCRD experiment is to give insights into the effect of multiple variables in the response profile, so to verify this tendency, three predicted points in each response surface were experimentally tested in quintuplicate. Despite a few points out of the error range, the general tendency of the proposed response model is valid for all strains (Tables 3 and 4).

Fig. 3.

Response surface from Z. mobilis ZM4 strains in a central composite design with RM medium containing 20 g L⁻¹ glucose. Biomass production (mg L⁻¹) after 10-h growth was analyzed in response to the variation of ethanol concentration and growth temperature. (A) ZM4_ pB1; (B) ZM4_p70; (C) ZM4_p24; (D) ZM4_p32. All models were reparametrized for a significance level p < 0.05

Fig. 4.

Response surface from Z. mobilis ZM4 strains in a central composite design with RM containing 160 g L⁻¹ glucose. Biomass production (mg L⁻¹) after 10-h growth was analyzed in response to the variation of ethanol concentration and growth temperature. (A) ZM4_ pB1; (B) ZM4_p70; (C) ZM4_p24; (D) ZM4_p32. All models were reparametrized for a significance level p < 0.01

Table 3.

Experimental validation of the CCRD models with 20 g L⁻¹ glucose

Strains	Experimental conditions
	39 °C and 10 g L−1a		34 °C and 30 g L−1a		32 °C and 55 g L−1a
	Predictedb	Obtainedb	Predictedb	Obtainedb	Predictedb	Obtainedb
ZM4_pB1	106.70 ± 7.24	111.41 ± 0.87	160.85 ± 3.40	155.13 ± 1.80	104.94 ± 4.94	99.43 ± 1.20
ZM4_p70	112.63 ± 9.23	111.39 ± 1.03	101.07 ± 5.21	95.95 ± 0.34	87.91 ± 8.30	95.22 ± 0.80
ZM4_p24	151.10 ± 8.11	162.02 ± 1.70	135.28 ± 3.21	127.63 ± 0.79	93.34 ± 5.15	94.03 ± 1.77
ZM4_p32	147.67 ± 9.39	150.88 ± 0.51	126.02 ± 3.72	125.89 ± 0.56	70.19 ± 5.97	58.19 ± 0.26

^aEthanol concentration at the beginning of the experiment

^bBiomass production (mg L⁻¹) after 10-h growth. All experiments were run in biological quintuplicate and the standard error is shown

Table 4.

Experimental validation of the CCRD models with 160 g L⁻¹ of glucose

Strains	Experimental conditions
	39 °C and 10 g L−1a		34 °C and 30 g L−1a		32 °C and 55 g L−1a
	Predictedb	Obtainedb	Predictedb	Obtainedb	Predictedb	Obtainedb
ZM4_pB1	53.54 ± 6.60	57.09 ± 0.15	87.06 ± 3.10	91.02 ± 0.48	54.68 ± 4.50	51.30 ± 0.33
ZM4_p70	20.48 ± 8.70	1.67 ± 0.29	0.00 ± 3.84	0.00 ± 0.01	0.00 ± 8.17	0.01 ± 0.01
ZM4_p24	77.01 ± 5.00	73.95 ± 0.21	90.99 ± 1.98	89.73 ± 0.12	63.44 ± 3.18	60.87 ± 0.69
ZM4_p32	82.10 ± 9.75	88.64 ± 0.34	93.50 ± 4.58	91.98 ± 0.36	41.89 ± 6.65	33.16 ± 0.08

^aEthanol concentration at the beginning of the experiment

^bBiomass production (mg L⁻¹) after 10-h growth. All experiments were run in biological quintuplicate and the standard error is shown

At last, the effects of episomal expression of sigma factors were evaluated in a flask fermentation batch for analysis of growth rate, glucose consumption, ethanol production, and metabolite accumulation (Tables 5 and 6). Considering that the experiments for each sigma factor are independent, the analysis was conducted individually.

Table 5.

Fermentation parameters of Z. mobilis strains expressing sigma factors

Strains	μa (h−1)	Biomass productivityb (mg L−1 h−1)	Glucose consumptionb (g L−1 h−1)	Ethanol productivityb (g L−1 h−1)
ZM4_pB1	0.3327 ± 0.0134	265.5 ± 2.8	6598 ± 0.058	3193 ± 0.028
ZM4_p70	0.3598 ± 0.0086	255.4 ± 2.5	6435 ± 0.060	3137 ± 0.022
ZM4_p24	0.3464 ± 0.0108	238.0 ± 4.9**	6439 ± 0.094	2937 ± 0.036**
ZM4_p32	0.3396 ± 0.0151	255.9 ± 2.8	6900 ± 0.057***	3177 ± 0.060

^aAverage growth rate in exponential phase for 24 h of flask fermentation in RM with 100 g L⁻¹ glucose. ^bAverage productivity and consumption from samples after 14 h of fermentation. Statistically different from the control (ZM4_pB1) in a t test (**p < 0.01 and ***p < 0.001). All samples were run in biological triplicate

Table 6.

Metabolites production of Z. mobilis strains expressing sigma factors

Strains	Lactate concentrationa (mg L−1)	Glycerol concentrationa (mg L−1)	Acetate concentrationa (mg L−1)
ZM4_pB1	125.4 ± 1.9	211.4 ± 3.1	1061.8 ± 12.0
ZM4_p70	3.2 ± 2.6***	226.7 ± 14.1	640.2 ± 2.3***
ZM4_p24	11.1 ± 5.4***	141.3 ± 4.5***	1001.6 ± 12.2
ZM4_p32	95.1 ± 1.2***	213.3 ± 2.2	977.0 ± 15.5

^aAverage of the maximum concentration after 24 h of fermentation with the standard error. Statistically different from the control (ZM4_pB1) in a t test (***p < 0.001). All samples were run in biological triplicate

Sigma 70

The additional expression of sigma 70 in strain ZM4_p70 resulted in a faster growth compared with the control strain in standard condition (Fig. 2A). There is an increase of growth rate, probably due to the major role of this sigma factor in the exponential growth phase.

Strain ZM4_p70 also showed growth inhibition by high glucose concentration in all experiments (Figs. 2D and 4B). Intriguingly, glucose consumption rates were not affected during flask fermentation (Table 5), suggesting that glucose transporters expression is not related to the growth inhibition.

The additional sigma 70 might be competing with other transcriptional factors for the RNA polymerase core, disturbing the synthesis and transport of osmoprotectants, leading to a slow adaptive response to osmotic shock. Further investigation is necessary to test this hypothesis.

Another observed effect in the strain ZM4_p70 was the drastic reduction in lactate accumulation; there was only about 4% of the lactate amount observed in the control strain (Table 6). The expression of sigma 70 might inhibit lactate dehydrogenate or even directly change the pyruvate flux to the Krebs cycle or to another not identified metabolite.

Sigma 24

Strain ZM4_p24 showed significative improvement in tolerance to high temperature conditions (Figs. 2B, 3C, and 4C). However, there is no clear evidence that ethanol tolerance is also controlled by this sigma factor. In fact, the response surface indicates no significative variation in ethanol tolerance. Despite the hypothetical relationship between heat and ethanol stress responses in Z. mobilis, our results do not associate it with sigma 24, and considering the reduction of ethanol productivity observed in ZM4_p24 fermentation (Table 5), it might be possible that the excess of sigma 24 disturbs ethanol adaptation mechanisms.

Sigma 32

Sigma 32 overexpression also improved temperature tolerance, confirming the predicted role in the heat shock response system (HSRS), but the most remarkable result is the improvement in high sugar adaptation and the significant reduction in ethanol tolerance of ZM4_p32 (Figs. 2, 3D, and 4D).

These results suggest that HSRS, modulated by sigma 24 and sigma 32, has no direct relation to high ethanol concentration response mechanisms in Z. mobilis. Furthermore, a more detailed analysis of the response surfaces of ZM4_p32 shows an antagonistic tendency between the HSRS and high ethanol concentration response mechanisms (Fig. 3D).

Analyzing the osmotic shock response, the strain ZM4_p32 showed more tolerance to high glucose concentration and also a significant increase in glucose consumption rates (Table 5), but no major alteration was observed in any other fermentation parameters or metabolite accumulation (Table 6). Sigma 32 might induce a change in the plasma-membrane composition, as observed in other bacteria [23, 28], which could lead to a rise in glucose influx and a reduction in ethanol tolerance without modifications in the primary metabolism in the ZM4_p32 strain.

Discussion

In this work, we have demonstrated that sigma 32 and sigma 24 might be related to heat shock response in Z. mobilis. This evidence corroborates the current ideas that Z. mobilis presents a canonical transcriptional control for the response to temperature stress and that this response overlaps, at least in part, with other mechanisms of stress response, such as osmotic pressure and high ethanol concentration.

Recent studies have been conducted to elucidate the mechanisms of response to heat shock in Z. mobilis, using genomic and transcriptomic approaches to identify several genes and proteins that are essential for heat tolerance, especially in the thermotolerant strain Z. mobilis TISTR 548 [29–31]. To date, the evidence indicates that Z. mobilis has a heat shock system similar to the one found in E. coli, with mechanisms that contribute to the stabilization of the membrane and the maintenance of macromolecules [30].

It is tempting to assume that membrane changes and the maintenance of macromolecules are also related to cellular response to high ethanol and glucose concentrations, but so far, the experimental data are based on few works with Z. mobilis in the presence of high concentrations of glucose [24, 32] or ethanol [10, 20, 33, 34], and the overlap of these data is not yet conclusive.

With the data provided here, it is possible to suggest a correlation between these different forms of stress, but contrary to what is proposed in the literature, the correlation between ethanol and other stresses (osmotic shock and heat shock) seems to be linked in an antagonistic way to the control of sigma 32. This duality between ethanol tolerance and osmotic shock induced by glucose makes sense physiologically, since cell cultures generally start with high glucose concentration and, over time, glucose concentration decreases, and ethanol concentration increases.

The main mechanism of sigma 32 control described for E. coli is based on the degradation of this sigma factor under normal temperature conditions by the FtsH protease, which maintains the very low intracellular levels of sigma 32. After the heat shock, sigma 32 levels increase significantly (induction phase), which leads to the production of heat shock proteins. Then, there is a rapid deactivation of the induction phase by the action of chaperones dnaK and dnaJ, which inhibit sigma 32 activity through a reversible multimeric association, controlling the amount of this sigma factor [28, 35, 36].

In a study with transposon insertions in Z. mobilis, some heat shock proteins were identified as essential for growth at 39 °C [30]. Another study showed an increase in the expression of sigma 32 factor and some heat shock proteins (including some with DnaJ domain) when Z. mobilis is under osmotic stress induced by glucose [24]. However, the transcriptome of Z. mobilis did not present any differential expression of these proteins when exposed to 5% ethanol [20]. As there is not enough information about the control of sigma 32 in Z. mobilis, it is only possible to conclude from our results that the presence of this protein activates both the heat shock and the response to osmotic stress mechanisms, increasing the rate of glucose consumption, and also reducing tolerance to ethanol.

Sigma 24, sigma 70, sigma 54, and sigma 28 are differentially expressed when Z. mobilis is grown in the presence of ethanol [20]. In our work, the overexpression of only sigma 24 did not result in a significative increase in the ethanol tolerance of Z. mobilis ZM4, making a conclusion about the real role of this transcriptional factor in the ethanol response impossible. But considering that ethanol tolerance is a complex feature involving a set of responses at different levels, it is likely that the combination of various sigma factors at specific concentrations is required to have a significant increase of this phenotype. Thus, more comprehensive analysis techniques may have greater success for the study of ethanol response, such as global transcriptional machinery engineering (gTME) [37].

The only ethanol tolerance gTME experiment performed in Z. mobilis so far was successful in constructing a strain more resistant to ethanol stress using a mutant version of sigma 70 factor under pdc promoter control, but no transcriptomic analysis was performed with this strain [10]. When using this technique with sigma 70 from Z. mobilis, one must be careful with the glucose concentration used. As it has been demonstrated here, the strain that overexpresses sigma 70 loses adaptability to osmotic shock induced by high concentrations of glucose. Thus, the mutant versions of sigma 70 could have increased growth compared with a control by the simple inactivation of this factor, and not by a real increase in the mechanisms of tolerance to ethanol. The use of gTME with other sigma factors could provide relevant information and help to elucidate these mechanisms of stress response.

Another possibility to study ethanol tolerance in Z. mobilis would be with a random gene disruption technique, using a transposon, in the ZM4 strain and the ZM4_p32 strain (which is sensitive to ethanol) and verify which mutations would inhibit growth in high ethanol concentrations in the wild type strain and which would recover the tolerance to ethanol in the ZM4_p32 line, thus identifying the genes essential for ethanol tolerance and the genes whose overexpression disturbs the activation of this mechanism.

Conclusion

We confirmed sigma 24 and sigma 32 as stress response sigma factors in Z. mobilis. Sigma 32 is important for heat shock–induced and glucose-induced osmotic shock response but is inhibitory for ethanol stress response. Sigma 24 triggered only the heat shock response, but did not affect ethanol tolerance, suggesting that only parts of these stress response mechanisms overlap and that probably a complex arrangement of sigma factors is necessary to control ethanol tolerance in Z. mobilis.

Experimental procedures

Strain construction

The coding region of alpha-amylase gene from B. subtillis [38] was codon optimized for Z. mobilis ZM4 using the program CodonCode Aligner. This sequence was fused to the signal peptide of the gene ZMO0130 of Z. mobilis ZM4 [26] and to the promoter (500 bp) and terminator (200 bp) regions of pyruvate decarboxylase gene of Z. mobilis ZM4. The cassette was synthetized by Genone (Rio de Janeiro, Brazil) and cloned into plasmid pBBR1MCS [25].

The coding regions of sigma 70 (ZMO1623), sigma 24 (ZMO1404), and sigma 32 (ZMO0749) genes of Z. mobilis ZM4 were amplified by PCR and cloned replacing the alpha-amylase gene in the plasmid pB1_Amy using the In-Fusion® Cloning System (Clontech, USA) and transformed into E. coli DH5a by RbCl₂ chemical transformation. The clones were selected in Luria-Bertani (LB) agar medium with chloramphenicol (25 mg L⁻¹). All primers are shown in Table 7.

Table 7.

Primers used in this work

Primer name	Sequence (5′ → 3′)
rpoD_F	ATATGGAGTAAGCATATGGCAGAGACGACTACGGCT
rpoD_R	TATTTAAAAAGTCGACCTAGTGGTCGAGGAAGCTCCG
rpoE_F	TATGGAGTAAGCATATGATCGAAAATCATGAAAAAGAGACAAATTCTCC
rpoE_R	TATTTAAAAAGTCGACCTAGCGACCGTTACTAAGGTCG
rpoH_F	ATATGGAGTAAGCATATGGCCACCAGTAGCACCCT
rpoH_R	TATTTAAAAAGTCGACCTATGCCATAGCGGGTAACAGC
Ppdc_F	GCGGTGGCGGCCGCTCTAGACTAGATGCGGTATA
Ppdc_R	GTGTTTTGAATATATGGAGTAAGCA
Tpdc_F	GTCGACTTTTTAAATAAACTTAGAGC
Tpdc_R	TACCGGGCCCCCCCTCGAGTTCTTTTTATCGGTC
R16S_F	AGAACATAGAAGAGGTAAGT
R16S_R	TCAACTATAGACCAGTAAGT
RTrpoD_F	GATCATACTTTGGAAGAAGT
RTrpoD_R	GAGAAGGATGTTTCAGTTT
RTrpoE_F	CCTTAGAAAAATTGGTGGA
RTrpoE_R	CGCTTACTTCATCCATAATC
RTrpoH_F	GAAGCGATGGATAAACTG
RTrpoH_R	TAGACCTGACTTAATTCTTC

All plasmids (Table 2) were transformed in E. coli JM110 to demethylate DNA and enhance Z. mobilis transformation efficiency [39]. These plasmids were transformed into Z. mobilis ZM4 (ATCC 31821) by electroporation (200 Ω, 25 μF, and 2.5 kV in a 0.2-cm cuvette) and the colonies selected in solid Rich Medium (RM) (20 g L⁻¹ glucose; 10 g L⁻¹ yeast extract; 2 g L⁻¹ KH₂PO₄; 1 g L⁻¹ MgSO₄·7H₂O; 1 g L⁻¹ (NH₄)₂SO₄) with agar (15 g L⁻¹) and chloramphenicol (100 mg L⁻¹) incubated for 96 h at 30 °C. The strain containing the plasmid pB1_AMY had starch (10 g L⁻¹) added to the medium and after 96-h incubation at 30 °C, the plate was stained with iodine vapor.

In order to confirm the integrity of vectors within Z. mobilis, plasmids were extracted from selected clones and sequenced by Sanger’s method. All plasmids and strains are shown in Table 2.

Gene expression quantification

Gene expression was measured using two-step qRT-PCR with a QuantiFast SYBR Green Kit (Quiagen) in a Rotor-Gene Q thermocycler (Quiagen) according to the manufacturer’s standard protocol. Total RNA was harvested during exponential growth (OD_600nm = 1.0) of a flask culture of Z. mobilis ZM4 in Rich Medium using a RNeasy Kit (Quiagen). All primer pairs were designed to amplify ~ 100 bp of the specific target gene and are shown in Table 7. Gene ZMOr003, encoding 16S ribosomal RNA, was used as an endogenous control to normalize for differences in total RNA quantity. All samples were run in biological triplicate and technical triplicate.

Stress response analysis

All stress response experiments were performed in 1.0 mL Rich Medium, in a cell repellent surface 24-well plates and a microplate spectrophotometer (EPOCH™—BioTek) with incubation (30 °C or 39 °C), agitation (237 cpm), and absorbance (600 nm) reads. The initial inoculum was of 10 μL (1:100) from a culture grown for 24 h and normalized to 3.0 of absorbance (600 nm). Ethanol (50 g L⁻¹) and glucose (160 g L⁻¹) were supplemented to RM right before the experiment began. All experiments were done in biological quintuplicate.

Central composite rotatable design analysis

Central composite rotatable design (CCRD) was constructed based on two variables (ethanol and temperature) and one response (biomass production after 10-h culture growth) to generate a quadratic response surface [40] in two different glucose concentrations: 20 g L⁻¹ and 160 g L⁻¹. Ethanol and temperature range matrices are shown in Table 8.

Table 8.

Central composite rotatable design variable matrix

Variables	Matrix value
	− 1.41	− 1.00	0.00	+ 1.00	+ 1.41
Ethanol (g L−1)	1.72	10.00	30.00	50.00	58.28
Temperature (°C)	28.34	30.00	34.00	38.00	39.66

All data points were done in biological quintuplicate with four repetitions in the central point. In order to design the experiment and to analyze data, the software Protimiza Experimental Design [41] was used.

Cell growth conditions were the same as those used in the experiment described in the “Stress response analysis” section, with temperature and ethanol concentration ranging according to the design matrix. Biomass was calculated based on the predicted OD_600nm conversion [42].

Flask fermentations

All flask fermentations were performed in 250-mL flasks, with 100 mL Rich Medium supplemented with 100 g L⁻¹ glucose and 100 mg L⁻¹ chloramphenicol. Flasks were incubated for 24 h under agitation of 100 rpm and 30 °C. The inoculum was adjusted to 0.10 of initial absorbance (660 nm), and every 2 h, 1 mL of sample was collected to be analyzed.

Glucose, ethanol, glycerol, lactate, and acetate were quantified in the supernatant of the samples using high-performance liquid chromatography (HPLC) with an ion-exchange column (Rezex organic acids (ROA) 15 cm), H₂SO₄ (5 mM) as mobile phase, 0.6 mL min⁻¹ of flow rate, and a refractive index detector (RID-20A, Shimadzu). All flask fermentations were done in biological triplicate.

Funding information

This study was funded by Conselho Nacional de Desenvolvimento Científico e Tecnológico (400711/2014-1), by Fundação de Apoio à Pesquisa do Distrito Federal (193000830/2015) and by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). The plasmid pBBR1MCS was kindly donated from Professor Henrique Ferreira (UNESP-Rio Claro, Brazil).

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflicts of interest.