Selective suppression of bacterial contaminants by process conditions during lignocellulose based yeast fermentations

Viability of bacterial contaminants with or without cocultivation with yeast

Even though bacteria such as Lactobacilli frequently contaminate industrial yeast fermentations, none of the tested Lactobacilli strains included in this study could grow or even maintain its viability in the hydrolysate without the presence of yeast (Figure 1). In fact, it is worth mentioning that none of the more than 15 Lactobacilli isolates obtained managed to multiply in the absence of yeast (data not shown). Acetobacter, showed an entirely different behavior and the viability was well preserved, or even slightly enhanced, during the incubation time of 3 days (Figure 1).

In cocultures of bacteria and yeast there was a tendency to a slight improvement in preservation of viability of L. fermentum although on average the increase is less than 1 log unit (Figures 1 and 2). However, the most striking effect of these coculture experiments was the severe effect on the viability of the yeast S. cerevisiae from the presence of Acetobacter (Figure 2). Without bacteria, S. cerevisiae managed to grow and multiply close to five generations in the hydrolysate used, but in the presence of Acetobacter there was a drastic reduction in the viability of the yeast after a growth period of 1 to 2 days (Figure 2).

The effect of nutrient supplementation (yeast extract) on viability of bacteria

One reason for the somewhat increasing viability of Lactobacilli in the presence of yeast cells could be that yeast will offer some additions of essential nutrients to the bacteria [10]. To test this hypothesis we added yeast extract to the hydrolysate, but no effect whatsoever on viability of Lactobacilli could be detected (Figure 3). Most probably, the effect of yeast additions on viability of bacteria originate from the fact that the activity of S. cerevisiae will detoxify the hydrolysate [11, 12] perhaps with some added value from nutrient availability.

Similarly, no effect on yeast viability and multiplication could be detected as a result of yeast extract additions to the hydrolysate (Figure 3).

Growth and product formation during cocultivation of yeast and Lactobacillus or Acetobacter

Coculture experiments revealed large effects on growth and viability of yeast and bacteria when coexisting in the lignocellulosic medium. How did this affect metabolism and major catabolic products such as ethanol, acetate and lactic acid?

The ethanol concentrations reached about 12 g/l and there was no negative effect due to the presence of Lactobacilli (Figure 4A, 4B). Lactic acid concentrations were very low, below 0.1 g/l, throughout the duration of the experiment and no production or consumption was detected. Acetate was found at initial concentrations of about 3.0 g/l and a slow consumption resulted in final concentrations between 2.0 to 2.3 g/l (Figure 4A).

A completely different picture emerged when S. cerevisiae was mixed with Acetobacter in cocultures. In this case there was production of ethanol during the initial 24 h but this was followed by declining ethanol concentrations concomitant with an increase in acetate concentrations (Figure 4B), similar to what is normally observed during growth of Acetobacter [13]. Obviously, the Acetobacter not only has a very negative effect on the viability of the yeast (Figure 2) but also consumes a substantial part of the previously formed ethanol.

The selective effect on yeast and bacteria from addition of NaCl, sugar, ethanol, and low pH

Due to the potentially detrimental effect of bacterial contaminations on yeast fermentations it was decided to assess whether it was possible to choose process conditions that would selectively inhibit bacteria with minimum effects on the yeast. A systematic study on the effect of low pH and additions of NaCl, sugar, ethanol on the growth of yeast and bacteria isolated from industrial ethanol production plants was performed. The results regarding stress tolerance in various yeast strains can be found in [14]. Concerning bacteria, more than 15 isolates of Lactobacilli as well as 3 strains reported to be relevant in similar environments, together with several isolates of Acetobacter were included in this investigation (Albers and Larsson, unpublished results). Of these, a subset of eight representative Lactobacilli, six yeast, and two Acetobacter strains were chosen for a wider analysis at several treatment levels in rich yeast extract/peptone/dextrose (YPD) or MRS media. These screening experiments were performed using pure cultures. The results were evaluated using multiple regression analysis. No single parameter could be expected to selectively favor yeast at the expense of bacteria. However, the fitted response surfaces (Figure 5) indicated that Lactobacilli may be more sensitive than yeast to combinations of elevated NaCl and ethanol concentrations. The Acetobacter strains grew very slowly in this experimental set up in comparison to the Lactobacilli and yeasts in all tested conditions.

Identification of optimum combinations of NaCl + ethanol selectively inhibiting bacteria

In order to test the hypothesis that enhanced levels of NaCl together with ethanol could selectively inhibit bacteria, a number of cocultures with yeast, L. buchneri, L. plantarum Acetobacter syzygii and Acetobacter tropicalis (all isolated from Domsjö Fabriker AB) were performed using a variety of combinations between NaCl and ethanol (Figure 6). To evaluate the effectiveness of this treatment a calculation was performed after 24 h of incubation concerning the increase in viable yeast cells and decrease in viable bacteria, respectively. The conditions finally chosen for more rigorous testing were 25 g/l NaCl + 40 g/l ethanol and 50 g/l NaCl + 20 g/l ethanol, since these combinations resulted in the largest decrease in bacterial viability while yeast could still maintain its viability. The former condition resulted in an increase in viable yeast cells almost similar to the increase obtained by the control without additions while the number of bacteria showed a more than tenfold decrease (Figure 6). Raising the NaCl concentration to 50 g/l and reducing ethanol to 20 g/l provoked a very drastic reduction in viable bacteria, close to 1,000-fold, while there was still an increase in viability of yeast during the 24 h period of testing (Figure 6).

Enhancing yeast viability and productivity by a combination of NaCl and ethanol additions

The successful treatment in terms of increased yeast viability and reduction of bacteria by these additions were verified by cultivations of cocultures in bioreactors. In the absence of NaCl and ethanol the bacteria multiplied in the lignocellulosic medium while yeast viability showed a negative trend (Figure 7). A totally opposite effect was seen on addition of NaCl + ethanol, that is, there was a substantial decrease in bacteria and improved growth of yeast (Figure 7). Also in terms of ethanol production, NaCl + ethanol additions turned out to be successful as there was indeed an increase in the amount of ethanol produced compared with results obtained using lignocellulosic medium without these additions (Figure 8). To some extent these results were somewhat surprising since both NaCl as well as ethanol additions in pure cultures also showed a negative effect on the yeast S. cerevisiae. Apparently, reducing the viability of bacteria is crucial to having a net positive effect on the yeast performance in these mixed cultures.

Verification of laboratory results in an industrial ethanol production plant

In order to test whether additions of NaCl together with ethanol could be a strategy to reduce the amount of bacteria also at a real industrial ethanol production plant, experiments were carried out at Domsjö Fabriker AB in Örnsköldsvik, Sweden. Samples were obtained from the production plant at this site and various combinations of NaCl/ethanol additions were tested. It should be kept in mind that this kind of test will include the response of the entire microbial community at the selected site and not only some selected isolates. Preferably, these tests should be repeated at different time points as the conditions and composition of the microbial flora change over time. As it turned out it was indeed possible to restrict bacterial contaminants by additions of NaCl and ethanol in a real industrial environment (Figure 9). However, it did require extensive testing and fine tuning of the concentrations used, and in our case a combination of 25 g/l NaCl and 12.5 g/l ethanol added was the most effective combination (Figure 9).

Strains

An industrial strain of S. cerevisiae was used (CCUG53310, Culture Collection University of Göteborg, Göteborg, Sweden) [15]. This strain was originally isolated from an industrial ethanol production plant, Domsjö Fabriker AB, Örnsköldsvik, Sweden. Lactic acid bacteria were obtained from a culture collection, Lactobacillus fermentum (ATCC14931) or isolated from Domsjö Fabriker AB, L. buchneri, L. plantarum. Acetic acid bacteria, Acetobacter tropicalis, A. syzygii, were isolated from the same industrial plant. The isolated bacteria were species determined by a combination of API test (bioMerieux, France) and 16S RNA gene sequencing.

Cultivations

Inoculum cultures were grown in YPD medium (10 g/l yeast extract (Sigma-Aldrich St. Louis, USA), 20 g/l peptone (Nordic Biolabs, Taby, Sweden), 20 g/l glucose (Merck, Darmstadt, Germany)) for yeast and MRS medium (Oxoid, Hampshire, England) for bacteria for 1 to 1.5 days at 30°C in falcon tubes or shake flasks depending on culture volume.

The cultivations were performed in a lignocellulosic hydrolysate of spruce chips pretreated with dilute acid, a composition determined previously [14]. The pH was adjusted to 5.0 with ammonia (Merck, Darmstadt, Germany) and the hydrolysate medium was filter sterilized before usage.

For cocultivations on a small scale in 50 ml falcon tubes, the cells in the inocula were harvested after 1 day by centrifugation, resuspended in sterile water and the optical density at 610 nm (OD₆₁₀) was measured. The hydrolysate (6 or 10 ml) was inoculated with cells (yeast and/or bacteria) and water in a total of 45 μl/ml hydrolysate to give an initial OD₆₁₀ of 0.05 for yeast and at 0.09 for bacteria. The lid was closed and the tubes were incubated at 30°C in a rotary shaker and monitored for up to 4 days. Samples for medium analyses were centrifuged (2 min, at minimum 14,000 g) and stored at -20°C before analysis.

Cocultivations on a large bench scale were performed with 1 l of medium using 3 l bioreactors (Belach Bioteknik AB, Stockholm, Sweden) operated at 30°C with a stirring rate of 300 rpm and no gas inlet. The initial pH was set to 5.0 with ammonia (Merck, Darmstadt, Germany) during the preparation of medium and the decrease during cultivations was always less than 0.5 pH units.

Determination of bacteria and yeast viability at an industrial production plant

The start inoculum was a mixture of microorganisms harvested from the Domsjö Fabriker industrial ethanol production plant located in Örnsköldsvik, Sweden. This mixture (sludge) contained the complete microbiological community existing in an industrial ethanol fermentation plant: mainly yeast, lactic acid bacteria and acetic acid bacteria.

This microbiological community was cultivated for 32 h at 30°C in spent sulfite liquor supplemented with 10.2 ml 25% ammonium (Merck, Darmstadt, Germany) and 171 mg/l KH₂PO₄ (Merck, Darmstadt, Germany) with and without addition of NaCl (Merck, Darmstadt, Germany) (25 g/l) + ethanol (VWR, Leuven, Belgium) (12.5 g/l). The pH was adjusted to 5.0 by addition of 5 M NaOH (Merck, Darmstadt, Germany) prior to fermentation. The cultivations were performed in 300 ml Erlenmeyer flasks with a total volume of 200 ml.

Measurements of the cell viability were performed by colony forming unit (CFU) count.

Analyses

Metabolites in the medium (ethanol, acetic acid, lactic acid) were analyzed using commercial enzymatic kits assays (R-Biopharm GmbH, Darmstadt, Germany) with adapted volumes in microtiter plates. Absorbance was measured with a Fluostar Galaxy plate reader (BMG Labtechnologies, Offenburg, Germany).

CFU determinations were performed on agar plates with YPD for yeast (when bacteria was present in large numbers, 20 μl of 50 g/l ampicillin (AppliChem, Darmstadt, Germany) was added to each plate) and MRS for bacteria with 0.1 g/l cycloheximide (Merck, Darmstadt, Germany)to suppress growth of yeast. Each dilution was spread on two or three plates. The plates were incubated at 30°C for 2 days for yeast and for 3 days for bacteria to establish distinct colonies before counting.

Multiple regression analysis of the specific growth rate as a function of pH, temperature, and concentrations of NaCl, glucose, ethanol and lactic acid was performed using the software Modde 9.0 (Umetrics AB, Umea, Sweden).