Isolation, identification and molecular genotyping of a thermotolerant yeast strain for use in high temperature fermentation for ethanol production
To obtain personalized yeast for ethanol production in Brazil, several attributes need to be considered and evaluated, such as high yields of ethanol cell viability tolerance to stressors produced during ethanol production. Another important attribute is growth temperature. In Brazilian ethanol production, the strains currently used ferment at temperatures between 28 and 33ºC, and to maintain this temperature constant, the mills need water coolers. The use of thermotolerant yeasts could circumvent this problem improving ethanol production. The Laboratory of Biochemistry and Applied Genetics of the Federal University of São Carlos (LBGA-UFSCar) has been isolating yeasts from ethanol production since 2009, which have been deposited in the LBGA strain collection. We used this collection to screen for isolates that can grow at higher temperatures, above those used in the ethanol fermentation plants in Brazil. Within the approximately 300 strains, four isolates have been identified as thermotolerant since they are able to grow and display similar fitness at both 30 and 40ºC, unlike the industrial strain CAT-1 that is unable to growth at high temperatures [10]. These yeasts were identified as LBGA-01, LBGA-69, LBGA-157, and LBGA-175, respectively (Fig. 1)
To identify these strains, the amplification of ITS [11] and a genotyping test were carried out. The results showed that the four strains were genetically different among them and also from the industrial strain CAT-1. The ITS analysis indicated that only LBGA-01 and LBGA-69 presented the amplification of a 900 bp, expected to identify the strain as a possible Saccharomyces [11; 12]. The LBGA-157 and LBGA-175 showed a different pattern of amplification suggesting that these yeasts are non-Saccharomyces strains, but possible wild contaminating strains (see Additional File 1: Figure S1). To confirm these results, the ITS amplicons were sequenced, and we confirmed through the BLAST analyzes that LBGA-01 and LBGA-69 strains are S. cerevisiae isolates, while LBGA-157 and LBGA-175 strains are Kluyveromyces marxianus yeasts.
The thermotolerant LBGA strains present superior fermentation performance at 40ºC in comparison to the industrial strain at 30ºC
Although only strains LBGA-01 and LBGA-69 were identified as S. cerevisiae, LBGA-157 and LBGA-175 were also included in the cell growth and fermentation tests at 30 and 40ºC since it is already known that strains of Kluyveromyces also ferment at high temperatures [13]. The results of cell proliferation growth at 30ºC show that the strains LBGA-01 and LBGA-69 have the same growth profile as the industrial strain CAT-1. However, when subjected to growth at 40ºC, these strains have higher growth rates than the industrial strain. This result was already expected since it is largely reported that CAT-1 does not show good performance when subjected to high temperatures [9; 10; 14]. The strains LBGA-157 and LBGA-175 showed slower growth rates at both temperatures. In fact, the strains LBGA-01 and LBGA-69 showed a similar growth profile in both temperatures indicating a strong thermotolerant phenotype in these strains (Fig. 2a and 2b). As mentioned above, to be a yeast applicable in industrial use, strains have to present excellent fermentation characteristics and a good conversion of sugar to ethanol in addition to thermotolerant growth. It is known that in the ethanol production process, industrial yeasts perform this conversion in the first hours of fermentation, thus converting all available sugar to ethanol after approximately 4–6 hours of fermentation. To evaluate the fermentative potential of isolated yeasts, we first performed a fermentation experiment using 4% of glucose and all the isolates. The results showed that LBGA-01 and LBGA-69 had a pattern similar to the industrial strain CAT-1 in fermentation conducted at 30ºC, and had superior performance at 40ºC. Strains LBGA-157 and 175 presented a low performance in both temperatures (Fig. 2c and 2d). The yeasts used during ethanol fermentation are subjected to high concentrations of sugar and therefore are constantly exposed to osmotic stress. Since LBGA-01 showed good fermentation performance and presented a slight advantage over LBGA-69 in the glucose fermentation (Fig. 2c and 2d), we also conducted fermentative tests using 8% of sucrose concentration to simulate the conditions widely used in the standard ethanol production in Brazil, in which sucrose is the fermentable sugar used. In comparison to the CAT-1 industrial strain, LBGA-01 showed a better performance in both temperatures with a clear superiority at 40ºC (Fig. 2e and 2f). In the fermentation assays conducted at 30ºC, both strains had a similar pattern. However, at 40ºC, LBGA-01 strain converted 57.5% of the initial sugar while the industrial strain converted 45% (Fig. 2f). From an economic point of view regarding the expected yields for a sugar cane plant, these results are expressive, since the LBGA-01 strain converted 12.5% more sugar under stress conditions than the CAT-1 industrial strain. It is worth mentioning that the use of a thermotolerant yeast operating under stress conditions with good fermentation rates can represent a significant increase in ethanol production in the plant. Other advantages are a decrease in the contamination with wild yeasts and bacteria, and in water use to control the temperature in the fermentation tanks, consequently reducing energy costs, collaborating with environmental issues related to water usage and mitigating the use of natural resources.
LBGA-01 is resistant to stressors produced during 1G and 2G ethanol production process.
Yeasts undergo constant stressing conditions during the Brazilian fermentation process to produce first (1G) and second (2G)-generation ethanol, directly resulting in a decrease of final yield in the process. To evaluate the tolerance of LBGA-01 strain, we analyzed its growth and survival under different concentrations of stressors such as ethanol, sugar, lactic acid, acetic acid, HMF and furfural (the latter three inhibitors of the 2G ethanol production process) and compared them to the industrial strain CAT-1 and the laboratory haploid strain Sc9721. The concentration of each stressor was established according to the literature, as described in the methodology section. The results obtained through the dropout analysis showed that the LBGA-01 strain is more resistant in all tested stressors (Fig. 3).
The overall positive results exceeded the initial expectations, especially in the presence of 2% acetic acid, where we observed that the LBGA-01 strain was more resistant than both the CAT-1 strain and the laboratory strain. When subjected to 4% acetic acid, all of the strains suffered stress and had their growth inhibited. For the lactic acid test, the LBGA-01 strain was also more resistant in the two tested concentrations (2% and 4%), reveling that under conditions of contamination by Bacillus spp or other lactic acid-producing bacteria [15], this yeast would be resistant, and not affected during alcoholic fermentation. However, a wider range of concentration needs to be tested to determine in which concentration of lactic acid the LBGA-01 strain can survive. The high ethanol stress test showed that LBGA-01 has a similar resistance profile to that of the industrial strain for all the used ethanol concentrations: 12, 14 and 16%. We observed higher resistance of LBGA-01 in comparison to CAT-1 strain at 16% of ethanol. In this trial, the control laboratory strain was drastically affected by ethanol stress. When subjected to different concentrations of sucrose, the LBGA-01 strain was also more resistant than CAT-1 and SC9721 strains in the three tested sugar concentrations (20, 25 and 30%), corroborating once more the results obtained during the fermentative tests. For the HMF test, the LBGA-01 had a slight increase in resistance at the lowest evaluated concentration (40 mM). Growth tests with furfural showed that the LBGA-01 strain is more resistant than the industrial and the laboratory strain at both tested concentrations of 0.3 mM and 0.9 mM. Furfural are potent inhibitors of Saccharomyces during lignocellulosic fermentation [16]. For this reason, the thermotolerance conjugated to furfural resistance, as described for LBGA-01, could be an important feature for the production and improvement of 2G ethanol. Taken together, these results demonstrate the robustness of the LBGA-01 strain in the presence of several stressors during the fermentation process, thus becoming a potential strain for the production of 1G ethanol production. Industrial strains have shown to be more robust towards the main lignocellulosic inhibitors produced during biomass pre-treatment processes [17]. The present results are in accordance to these findings, in which the laboratory strain (Sc-9721) was more affected by the presence of HMF, furfural and acetic acid than the industrial LBGA-01 and CAT-1 strains. Regarding the stressors of 2G ethanol production, further tests need to be carried out to confirm the resistance to the same stressors. Nonetheless, our initial results point out that the LBGA-01 strain is more resistant than the CAT-1 industrial strain when submitted to HMF, furfural and acetic acid concentrations, thus calling attention to its potential use in the 2G ethanol production.
Transcriptional responses of LBGA-01 under high temperature fermentation conditions.
To better understand the metabolic changes in LBGA-01 strain during high temperature fermentation, the expression of genes involved in efficiency of fermentation and membrane biosynthesis and sucrose assimilation were evaluated using qPCR, and compared with the industrial yeast CAT-1 at both fermentation temperatures (30ºC and 40ºC). The mRNA levels for genes involved in the efficiency of fermentation are summarized in Fig. 4.
During the fermantation assays, in both glucose-limited chemostats, and in fed-batch fermentation, increase in glycerol production rate and glycerol titer were observed at 40ºC, respectively, when compared to control condition (30ºC),. Since GPD1 and GPD2 are key enzymes in glycerol synthesis, we hypothesized that the expression of GPD1 and GPD2 would increase during the fermentations. However, our results indicate that the expression of these genes did not changed (Fig. 4). GPD2 and GPD1 are paralog genes encoding the isoenzyme NAD-dependent glycerol-3-phosphate dehydrogenase, which has an important role in osmoadaptation (GPD1) and anoxic growth conditions (GPD2) [18; 19; 20] Mutants lacking both GPD1 and GPD2 do not produce detectable glycerol, leading to the accumulation of dihydroxyacetone phosphate (DHAP). This DHAP can be converted to methylglyoxal, a cytotoxic compound that can inhibit the yeast growth [21]. Differently, the growth of the LBGA-01 was not affected in either temperature. Interestingly, our results show that under high temperature (40 ºC), expression of SNF1 was highly increased in fermentations using the LGBA-01 strain (Fig. 4). The kinase SNF1 was described as an repressor of GPD2 via phosphorylation to halt glycerol production when nutrients are limited [22]. Therefore, we hypnotize that the unchanged GPD2 abundance is because its repression is exacerbated in the LBGA-01 strain due to the increase of SNF-1 expression (Fig. 5d).
As expected, a decrease in SUC2 (invertase) expression was observed in all the strains after 4 hours of fermentation (Fig. 4), what was due to the inhibition of SUC2 expression caused by the accumulation of glucose and fructose during the first hours of fermentation, as a result of the invertase activity [23]. However, a different pattern was found between CAT-1 and LBGA-01 after 8 hours of fermentation. During CAT1 fermentation, the SUC2 gene is reactivated as glucose concentration decreases and the invertase resumes the metabolization of the residual sucrose. The metabolic shift caused by glucose concentration and by inactivation and reactivation of SUC2 is accompanied by the expression of SNF1. The activation of this kinase is glucose-dependent and directly related to the inactivation of glucose transporters and activation of genes involved in the utilization of alternative carbon sources [23; 24; 25]. As previously described, the SNF1 expression in LBGA-01 strain is maintained at high levels during the fermentation. Meanwhile, the SUC2 expression decreases, as mentioned above, and the sucrose consumption remains unchanged (Fig. 2e and 2f). In fact, the ethanol production rate is higher than at 40ºC than at 30ºC in this strain, accompanying the SNF1 expression that is increased in this temperature. We suggest that this process happens because there is an augment in the internalization of sucrose by MAL31 and AGT1 transporters since both proteins are able to actively transport sucrose, maltose, and maltotriose, although this process naturally occurs in absence of glucose [26; 27]. Our hypothesis is also supported by the expression of the SNF1 gene, that is highly expressed in LBGA-01 strain during the whole fermentation process, even in the presence of glucose, thus possibly activating the receptors [28; 29; 30]. Interestingly the expression of SNF1 was higher in the middle and in the end of the fermentation conducted with LBGA-01 in both temperatures, when sucrose was used as carbon. Therefore, we argue that LBGA-01 can be used in higher levels of this sugar since sucrose would be inverted by SUC2 and transported by AGT1 at the same time, and later inverted by intracellular SUC2.
When the expression of genes involved in the formation of secondary products of fermentation such as glycerol (GPD2), acetate (ALD6 and ALD4), and acetil-CoA (ACS2) was evaluated, we found repression of all of these genes at 40 ºC (Fig. 5a-c). These results suggest that the alternative pathways for glucose utilization are inhibited at high temperatures in the LBGA-01 strain, which preferentially uses the available carbon source for the ethanol production pathways.
Quantitative physiological parameters of LBGA-01 during anaerobic glucose-limited chemostat at high temperature
Chemostat cultivations have been broadly applied on quantitative study of physiological parameters in S. cerevisiae. We deemed to investigate the impact of a high temperature (40 ºC) on the anaerobic physiology of LGBA-01 in comparison to a control temperature (30ºC) and to the experiment conducted by Della-Bianca et al. [31] that used the industrial strain PE-2, largely used for Brazilian ethanol production, using glucose-limited chemostat cultures (Table 1). An advantage of studying microbial cells under continuous culture instead of batch culture is that in the former, the specific growth rate can be held constant under different stressful conditions [32].
Table 1
– Physiology of S. cerevisiae strains in glucose-limited.
S. cerevisiae strain | LBGA-01 (This work) | Saccharomyces cerevisiae PE-2 [31] |
Temperature | 30 °C | 40 °C | 30 °C |
q glucose | -5.28 ± 0.50 | -7.22 ± 0.93 | -5.06 ± 0.15 |
q CO2 | 7.98 ± 0.69 | 12.02 ± 1.04 | 8.51 ± 0.28 |
q Ethanol | 8.79 ± 1.03 | 11.50 ± 1.72 | 7.70 ± 0.26 |
q Glycerol | 0.89 ± 0.22 | 1.38 ± 0.32 | 0.89 ± 0.04 |
q Lactate | 0.05 ± 0.03 | 0.11 ± 0.02 | 0.05 ± 0.00 |
q Pyruvate | 0.01 ± 0.00 | 0.08 ± 0.03 | Not reported |
q Acetate | 0.00 ± 0.00 | 0.01 ± 0.02 | 0.00 ± 0.00 |
X (g DW L− 1) | 2.64 ± 0.30 | 2.09 ± 0.26 | 2.63 ± 0.01 |
YX/S (g DW g glucose− 1) | 0.11 ± 0.01 | 0.08 ± 0.01 | 0.11 ± 0.00 |
YETH/S (g ethanol g glucose− 1) | 0.43 ± 0.12 | 0.41 ± 0.01 | Not reported |
YG/S (g glycerol g glucose− 1) | 0.09 ± 0.01 | 0.10 ± 0.01 | Not reported |
Residual glucose (mM) | 0.17 ± 0.23 | 2.4 ± 0.58 | Not reported |
C recovery (%) | 101.97 ± 1.73 | 101.28 ± 1.35 | 100.9 ± 0.7 |
The assays were conducted using anaerobic chemostats with synthetic medium at a dilution rate of 0.1 h− 1. Specific q rates are given in mmol g− 1 h− 1. Data are the average value from duplicate or triplicate experiments ± deviation of the mean.
In anaerobic glucose-limited chemostat cultures of the LGBA-01 strains, carbon was mainly diverted to ethanol and CO2, and minor amounts of glycerol and lactic acid with a concomitant formation of yeast biomass were produced. When comparing the data obtained from LBGA-01 strain cultivated at 40ºC and 30ºC (control), we observed an increase in consumption of glucose (38%) as well as in the production rates of CO2 (51%), glycerol (54%) and ethanol (36%). On the contrary, we observed a substantial decrease in biomass yield (25%), and no effect on glycerol yield or on the maximum specific growth rate during the batch phase (Table 1). Interestingly, we did not observe difference between 40 ºC and the control condition in ethanol yield during steady-state in cultures of the strain LGBA-01. In contrast, we observed that glucose concentration was higher at 40ºC the during steady-state, suggesting a possible inhibition of glucose uptake.
Stressing conditions such as high temperature can generate perturbations in the redox balance inside the cells [33]. The increase in the rate of synthesis of by-products (such as acetate and lactate), which are involved in the reoxidation of NADH, are indicative of how cells are responding to this stressful condition, as well as the differential gene expression of ACS2 gene reported above.
A similar experimental setup was performed by Bianca et al. (2014) [31] using the industrial S. cerevisiae strain PE-2, known as highly stress-tolerant [8]. The results obtained in the present study showed that LGBA-01 presents higher ethanol and glycerol production rates than S. cerevisiae PE-2 under similar conditions, i.e., at 30ºC. These data suggest an advantage on the industrial process. Moreover, as mentioned by Bianca et al. (2014), the absence of acetic acid in all cultivations is a remarkable phenotypic characteristic found in a strain that grows in acidic environments, such as those found in the industrial ethanol process.
A previous report analyzed the thermotolerance of industrial S. cerevisiae strains isolated from Brazilian ethanol plants, such as CAT-1, PE-2, BG-1, and JP-1 in synthetic media with glucose as the sole carbon and energy source [14]. Although the authors have used different conditions from those reported in this study during the batch phase, i.e., oxygen-limited shake-flask cultures as opposed to anaerobic bioreactors, the growth rates of some strains (JP-1 and CAT-1) were higher at 37 ºC than at 30ºC (0.39 and 0.38 h− 1, respectively). Furthermore, they were lower than the growth rates obtained for the LBGA-01 strain at both 30 ºC and 40ºC. In respect to ethanol yield, JP-1 and BG-1 presented an increase in cultivations at 37ºC when compared to 30ºC, differently from our results. Instead, PE-2 presented a small increase in ethanol yield at 37 ºC than at 30 ºC.
In terms of the specific ethanol production rate, our results revealed that LBGA-01 strain has a higher rate at 40ºC than at 30ºC, although reaching similar values of ethanol yield under both conditions. Similarly, increased specific rates of glycerol production were also observed under such conditions, although glycerol yield was not affected. The deviation of carbon away from biomass formation at 40ºC seems to be due to pyruvate and lactate production. This result can be explained by the reduced expression of genes encoding the enzymes responsible for the production of secondary products (Fig. 5).
Fermentative performance of LBGA-01 in conditions mimicking the Brazilian industrial ethanol process under high temperature
Different aspects of S. cerevisiae strains as specific growth, yields in ethanol, glycerol, and cell productivity are commonly investigated under laboratory conditions, using a batch mode operation and synthetic defined culture media offering conditions containing all nutrients in adequate amount, to enable maximum growth rate. In the synthetic medium, the carbon source is usually a limiting nutrient [34].
Industrial conditions are not reproducible and vary from batch to batch. There is insufficient data reported for conditions that reproduce the different characteristics found in industrial environments. Specifically in Brazilian 1G ethanol production, sugarcane juice and molasses are often used as a lower cost carbon source for fermentation [35]. Its composition and quality also vary among different batches and harvesting periods; therefore, synthetic laboratory medium replicates of conditions are poor and may lead to misinterpreted conclusions [36]. Also, other stresses are associated with industrial production, such as toxicity of products, non-aseptic conditions, substrate inhibition, cell recycle, acid treatment, bacterial contamination, and temperature stress [34]. Thus, high tolerance for such a great variety of stressful conditions is a desirable feature for a yeast strain in the fuel ethanol industry [34; 37; 38].
To asses and study physiologic aspects and performance of LBGA-01 under highly stressful conditions, we scaled down the Brazilian 1G ethanol production with sugarcane molasses as carbon source using protocol described by Raghavendran et al. [34]. Thermotolerance was investigated by submitting cells to 34 °C and 40 °C that are unusual laboratory temperatures, commonly found in Brazilian sugarcane mills and inside reactors due to the exothermic reactions of ethanol production [36].
Fermentation capacity was monitored by plotting the produced CO2 per gram of wet biomass as a function of fermentation time. As shown in other studies, there could be an increase or decrease in biomass weight, and the plot of the total amount of CO2 against time could not represent well the fermentation capacity [34]. In this case, the normalization of the specific mass is necessary. The fermentation at 34 °C showed a slightly lower fermentation capacity (Fig. 6a) compared to fermentation at 40 °C in the first cycle. At both temperatures, the yeast started with virtually the same viability of approximately 80%. The fermentation capacity of the first cycle at 34 °C was kept in the other sequential cycles, so with a viability of about 100% (Fig. 6b). The fermentation capacity decreased cycle to cycle at 40 °C, as it can be seen from the reduction of the experimental data slope, probably associated with a reduction of viability. At 40 °C, LBGA-01 viability slightly decreased after the first cycle (54.9%), reaching 28.7% of viability after the fourth cycle.
During the experiments, both fermentations started with similar weights. (4 ± 0.06 g). Corroborating constant viability, the biomass at 34 °C had a negligible decrease at first, followed by a small increase (less than 5%) in all subsequent cycles (Fig. 6c), thus resulting in a total weight increase of 7%. At the high temperature, biomass decreased, similarly to viability, with a mean decrease of 6.2% for each cycle and a total reduction of 24% (Fig. 6c).
Ethanol yield and glycerol production were also assessed during fermentation. As mentioned above, glycerol production is associated with a part of the physiological response of cells to osmotic shock. Thus, its formation occurs in many kinds of stressful situations. As a cell response, intracellular glycerol is thought to decrease water activity in the cytosol, leading to higher water uptake [39]. Glycerol levels were monitored and showed an increase with cycle in both temperatures, and were higher at 40 °C than at 34 °C. As a protection mechanism, increased glycerol levels are caused by the high-temperature stressful condition that cells are submitted to (Fig. 6d).
Ethanol yield for each cycle was calculated as described elsewhere [35]. A correction factor for high cell density was applied as previously reported, and a specific volume of 0.7 mL g− 1 (wet basis) was considered for yeast cells. Thus, the ethanol yield accounts for ethanol from centrifuged wine and pelleted yeast biomass. A mass balance for ethanol is applied as a difference between ethanol content at the end of the cycle and ethanol in the beginning (returned wine plus pelleted yeast biomass from the previous cycle). The ethanol yield is expressed as a percentage of the maximum theoretical ethanol that could be produced by the total sugar content (Eq. 1):