Fermentation of biomass
sugars to ethanol using native
industrial yeast strains.
In this paper, the feasibility of a technology for fermenting
sugar mixtures representative of cellulosic biomass
hydrolyzates with native industrial yeast strains is
This paper explores the isomerization of xylose to xylulose
using a bi-layered enzyme pellet system capable of
sustaining a micro-environmental pH gradient.
This ability allows for considerable flexibility in conducting
the isomerization and fermentation steps.
With this method, the isomerization and fermentation
could be conducted sequentially, in fed-batch, or
simultaneously to maximize utilization of both C5 and C6
sugars and ethanol yield.
Commercially-provenSaccharomyces cerevisiae strains from
the corn–ethanol industry were used and shown to be very
effective in implementation of the technology for ethanol
Saccharomyces cerevisiae is well-adapted to industrial use
due to its near theoretical ethanol yields and tolerance to a
wide spectrum of inhibitors and elevated osmotic pressure.
S. cerevisiae contains all necessary enzymes for the
conversion of xylose to ethanol, except xylose isomerase (XI).
Hence, addition of exogenous xylose isomerase for
isomerization of xylose to xylulose and fermentation of
xylulose to ethanol using native yeast strains was proposed.
Baker’s yeast (BY) (S. cerevisiae, Type II) was
purchased in dried pellet form from Sigma–
Aldrich (St. Louis, MO). The commerciallyproven yeasts from the corn–ethanol
industry, namely ethanol red (ER) (Fermentis,
Lesaffre, Marcq-en-Baroeul, France)
hydrolysate ; a substance or mixture produced by
UREASE; An enzyme that catalyzes the hydrolysis of
urea into carbon dioxide and ammonia; is present in
AZIDE; a chemical compound containing the azido
group combined with an element or radical
Azide is the anion with the formula N3−. It is
the conjugate base of hydrazoic acid. N3− is a linear
anion that is isoelectronic with CO2 and N2O
For analysis of metabolites from fermentation, a 200–
300 μl sample was collected at each time point and
centrifuged to separate yeast.
The supernatant was collected and diluted 1:3 with
deionized water and filtered through a 0.2 μm filter.
Calibration standards for glucose, xylose, xylulose, glycerol,
acetic acid, xylitol, urea and ethanol were prepared in a
All standards and samples were analyzed by HPLC using a
30 μl injection volume with a 100 μl injection loop.
For sugar, glycerol, acetic acid and ethanol detection and
analysis, Bio-Rad Aminex HPX-87 H (300 × 7.8 mm) and
Phenomenex Rezex RFQ fast acid (150 × 7.8 mm) ion
exchange columns were used in series.
The column temperature was maintained at 50 °C with a
mobile phase of 0.4 ml/min of 5 mM H2SO4.
At these operating conditions, when xylulose concentrations
are low relative to xylose, the peaks are not well-resolved
and the calculated xylulose concentrations may be subject to
For xylitol and urea analysis, a Bio-Rad Aminex HPX-87 C
(300 × 7.8 mm) column was used at 80 °C with a 0.6 ml/min
mobile phase of ultra-pure water.
The stoichiometric fermentations of glucose and xylulose to
ethanol are given by:
Xylulose fermentations with BY and ER
As noted, the additives used for isomerization of xylose had
no adverse effect on the glucose fermentation using either
BY and ER.
The experiments in this section were conducted to assess the
relative effectiveness of BY and ER in fermenting xylulose to
Xylose (60 g/l) was partially isomerized and fermented
using BY or ER.
The xylose was pre-isomerized to xylulose with 0.05 M borax
and 0.01 M urea for a 55% conversion (∼33 g xylulose/l).
By allowing a significant proportion of xylose to remain in
the media, the effectiveness of azide in suppressing xylitol
formation on the two yeast strains can be assessed.
Fermentation of the 33/27 g/l of xylulose/xylose mixtures was
initiated by adding 100 g/l yeast to the media. Ethanol
production with the two strains is shown in Figure 1.
The initial fermentation rates are higher for BY but the
overall ethanol yields are similar for both strains, as was the
case for glucose.
Both strains were able to convert the xylulose to near
theoretical yield for ethanol.
Total acetic acid and glycerol amounts of 0.06–0.08 g/g
total sugar were observed in both cases, which are similar to
literature reports for BY.
Fig. 1. Fermentation of ∼33 g/l xylulose with both ER and BY. Xylulose
was produced by the partial isomerization of 60 g/l xylose with the bilayered enzyme pellets. Both strains fermented xylulose to ethanol with
high yield, but with significantly slower kinetics than for glucose. Both
experiments contained 4.6 mM azide to block respiration.
Xylitol has been found to be a measurable by-product in
xylose/xylulose fermentations when a significant quantity of
xylose is present in the medium.
Xylitol production tends to increase as the temperature
decreases (less than 35 °C) or the pH increases (pH > 5).
For these experiments at 34 °C and pH 4.5 with azide,
xylitol formation is not likely.
Fermentation of mixed sugars
The ultimate goal of the bi-layered enzyme technology is its
implementation with biomass hydrolyzates.
Hydrolyzates may contain various fermentation inhibitors
depending on the method of biomass pretreatment
Accordingly, detoxification/conditioning may be necessary
prior to the isomerization/fermentation. Removal of
fermentation inhibitors prior to isomerization will also
eliminate the risk of inhibition of the XI and urease activities.
A recently developed ionic liquid pretreatment method was
shown to produce hydrolyzate free of the fermentation
inhibitors traditionally found with dilute acid pretreatment.
Thus, as new pretreatment methods are developed, the
need for hydrolyzate preconditioning may be eliminated.
Even in the absence of fermentation inhibitors, when S.
cerevisiae are presented with mixtures of glucose and
xylulose, the overall ethanol yields can be much different
than that achieved with either sugar separately.
Accordingly, the experiments in this section focus on
simulated glucose/xylulose mixtures.
Effect of yeast loading on fermentation of mixed sugars
Mixed sugar fermentations were conducted with 90 g/l
glucose and 30 g/l xylose both with and without the addition
WHAT IS AZIDE?
Azide is the anion with the formula N3−. It is
the conjugate base of hydrazoic acid.
N3− is a linear anion that
is isoelectronic with CO2 and N2O.
Per valence bond theory, azide can be described by
several resonance structures, an important one being
Azide is also a functional group in organic chemistry,
To ensure immediate fermentable sugar availability, the
xylose was first pre-isomerized for 24 h to obtain ∼86%
Fermentations were performed with three yeast loadings
(50, 100, and 200 g/l).
Results of these fermentations in the absence of azide are
shown in Fig. 2a–c.
Fig. 2. Mixed sugar fermentation of 90 g/l glucose and 30 g/l xylose (pre-isomerized
to xylulose) without azide (a–c) and with 4.6 mM azide (d–f). Three yeast loadings
were tested: 50 g/l (a,d); 100 g/l (b,e); and 200 g/l (c,f).
In all cases, glucose was consumed within the first 5 h. As
yeast loading increases from panels a to c, the overall
ethanol yield increases and the by-product formation
decreases by about 20%.
With 50 g/l yeast, more than 6 g/l xylulose remain after 28 h
At this low yeast loading, uptake of xylulose may limit its
At the higher yeast loadings shown in Figure2b and 2c,
xylulose utilization is much improved as is the overall ethanol
To assess the impact of azide in mixed sugar fermentations,
the fermentations shown in Fig. 2a–c were repeated with
4.6 mM azide (see Fig. 2d–f).
In pure xylulose fermentations, close to theoretical yields of
ethanol were achieved under similar conditions.
The incomplete utilization of xylulose in mixed sugar
fermentations suggests that ethanol produced from glucose
fermentation may inhibit xylulose uptake and utilization.
Effect of mode of addition and strain of yeast on mixed
The fermentation of two mixed sugar compositions: (1)
90 g/l glucose and 30 g/l xylose and (2) 130 g/l glucose and
50 g/l xylose, were conducted with fed-batch addition of
50 g/l yeast at time 0, 4, 7 and 24 h to a final concentration
of 200 g/l.
The fermentations were conducted with either BY or ER in
the presence of 4.6 mM azide. Xylose was pre-isomerized to
86% conversion prior to fermentation for all cases.
In an attempt to more-closely produce anaerobic
conditions, nitrogen was bubbled into the shake flasks for 60
s to purge oxygen after each yeast addition.
Ethanol production for 120 g/l total sugars fermented with
200 g/l yeast added either initially or in fed-batch mode are
shown in Fig. 3a.
Fig. 3. Mixed sugar fermentations with BY and ER in fed-batch yeast addition mode. Xylose
was pre-isomerized to 86% conversion for all experiments prior to fermentation; 4.6 mM
sodium azide was used in all experiments. The fed-batch yeast addition was implemented by
adding 50 g/l of yeast at 0, 4, 7 and 24 h to a total of 200 g/l. Nitrogen was purged into the
shake flask after each yeast addition and sample collection to maintain anaerobic conditions.
Ethanol production is shown for (a) for 120 g/l sugar (90 g/l glucose and 30 g/l xylose)
fermentation and (b) 180 g/l total sugar (130 g/l glucose and 50 g/l xylose) fermentation.
The overall ethanol yield increased from 0.38 to 0.42 g/g
total sugar when yeast was switched from initial to fedbatch addition.
Under these conditions, ER proved to be a marginally better
fermentor of mixed sugars (ethanol yield of 0.46 g/g total
sugar) than BY with fed-batch addition.
Ethanol yields from the fermentation of 130 g/l glucose and
50 g/l xylose with fed-batch yeast addition are shown in Fig.
Ethanol yields at the end of 72 h were 0.35 and 0.41 g
ethanol/g total sugar from BY and ER, respectively.
For BY, some reassimilation of ethanol occurred after the
first 24 h.
Fed-batch yeast addition resulted in better sugar utilization
and about 25% higher total product formation, which is
reflected as higher ethanol as well as by-product production.
ER clearly shows less ethanol inhibition at these high sugar
loadings than BY.
This result is consistent with the hypothesis that an ethanoltolerant yeast strain may be less susceptible to the synergistic
inhibition effects of azide and ethanol.
This inhibitory effect may be more clearly seen in Fig.
4 where xylulose uptake rates are compared for BY with and
without azide on pure pentose and mixed sugars.
Fig. 4. Xylulose uptake in pure and mixed sugar fermentations. All
experiments were conducted with an initial yeast loading of 100 g/l
and with 30 g/l xylose pre-isomerized to yield approximately 25.8 g/l
xylulose; Experiments A and B contained 4.6 mM azide.
Xylulose uptake in mixed sugar fermentations for BY
Transient xylulose concentrations for three different
fermentations with 100 g/l BY loading and the same initial
xylulose concentration are compared in Fig. 4.
Curve C shows that the most rapid uptake of xylulose
occurred in the mixed sugar fermentation without azide (see
also Fig. 2b).
In contrast, when azide is added, xylulose uptake rate is
reduced (Fig. 4, curve B, see also Fig. 2e).
Indeed, the xylulose uptake in curve B is even slower than
observed with pure xylulose alone (curve A) where ethanol
inhibition resulting from glucose fermentation is not possible.
It may be that xylulose utilization kinetics are hampered by
the combined effects of azide and accumulated intracellular
Based on the experiments summarized in Fig. 2, Fig. 3
and Fig. 4, it was concluded that the most efficient
implementation of SIF for mixed sugar fermentations with
BY is high yeast loadings without addition of azide to the
Under such conditions, xylulose uptake is benefited by the
availability of sufficient transporters, and as well as reduced
Simultaneous-isomerization-and-fermentation of mixed
Isomerizations and fermentations were conducted
simultaneously on mixed sugars containing 60 g/l glucose
and 30 g/l xylose in media without azide.
Three separate experiments were conducted to assess the
effect of the extent of xylose isomerization to xylulose on
In these experiments, the bi-layered pellets were (a) not
added to the fermentation medium (control run); (b) added
with borax and yeast; or (c) added first, followed by yeast
addition after partial isomerization of the sugars with borax.
Fig. 5. Simultaneous isomerization and fermentation (SIF) of 60 g/l
glucose and 30 g/l xylose. Fermentation with (a) no isomerization of
xylose; (b) SIF; and (c) pre-isomerization of xylose followed by SIF. Data
shown are the average values of duplicate experiments; glucose values
have standard deviation within ±8%; all other values have standard
deviation within ±3%. For all experiments shown in Fig. 3, carbon
balances closed to within 3% (data not shown). Ethanol produced in (c)
is slightly lower than (b) due to lower initial sugar concentrations. Yield
of ethanol per total sugar for the experiments are (a) 67%, (b) 74%,
and (c) 81%.
As shown in Fig. 5, glucose was completely consumed within
1–2 h for all three cases.
In the absence of added bi-layer pellets (Fig. 5a), the
ethanol yield based on total sugar (glucose and xylose) was
67%; about 20% of the xylose was consumed.
Byproducts formed included acetic acid, glycerol and xylitol.
When the bi-layer pellets were added with 0.05 M borax
(Fig. 5b), the overall ethanol yield increased from 67% to
74% (±1%), with about one third of the initial xylose
remaining at 10 h.
The facilitated transport shuttle set up by borax allows
appreciable conversion of xylose to xylulose, presenting yeast
with xylulose to ferment.
However, as the SIF proceeds, the rate of xylose
isomerization decreases due to xylitol accumulation within
the fermentation broth.
Xylitol formation began after glucose was depleted and
rose gradually to ∼2.5 g/l.
These experiments, while proving the viability of SIF, point
to the need to address issues associated with the slow rate of
isomerization in this configuration.
In Fig 5c, the xylose was pre-isomerized to 70% xylulose and
then yeast were added.
These high initial xylulose concentrations led to a further
improvement over Fig. 5b with ethanol yield approaching
Indeed, the partial isomerization to xylulose prior to
fermentation leads to lower xylitol production contributing
to the increased ethanol productivity.
In proportions found in biomass hydrolyzates (concentration
of glucose 3–6 times higher than xylose), isomerization of
xylose to xylulose will dominate.
Moreover, since glucose is rapidly fermented by yeast,
fructose formation is likely to be very low (fructose was not
detected in the experiments described inFig. 5b and c).
Even if small quantities of fructose are formed, S.
cerevisiae ferments fructose as well as glucose.
Recently, a method for isomerizing xylose to xylulose at high
yield was demonstrated wherein the xylulose complexing
agent, pH, and temperature are compatible with
Thus, isomerization and fermentation could be conducted
sequentially, in fed-batch, or simultaneously to maximize
utilization of both C5 and C6 sugars and ethanol yield.
For the first time, it was demonstrated that industrially
proven yeast strains such as ethanol red can convert high
concentrations of mixed sugars to ethanol in times of about
30 h with yields comparable to engineered strains.