Production and productivity of 1,3-propanediol from glycerol by Klebsiella pneumoniae GLC29

Article abstract of DOI:10.1016/j.cattod.2014.05.016

Interest in the development of the bioproduction of 1,3-propanediol, an important chemical intermediate with various industrial applications, has increased in recent years. Klebsiella pneumoniae is one of the most studied and efficient bacteria for 1,3-propanediol production from glycerol. A new isolate of K. pneumoniae was investigated using response surface methodology by central composite design for the production of 1,3-propanediol using glycerol. The effects of pH, temperature, stirrer speed, and glycerol concentration on the production and productivity of 1,3-propanediol were examined. Considering both production and productivity, the best conditions for glycerol conversion in 1,3-propanediol are: a pH range of 6.9-7.1, a temperature between 33 and 38.5°C, a stirrer speed of 110-180 rpm, and a glycerol concentration of 39-49 g l^-1. Batch fermentation carried out at a pH of 7.0, a temperature of 35°C, a stirrer speed of 150 rpm, and a glycerol concentration of 40 g l^-1 produced 20.4 g 1,3-propanediol l^-1, with a maximum volumetric productivity of 2.92 g l^-1 h^-1 and a yield of 0.51 g g^-1. The main byproducts were acetic acid (approximately 7.0 g l^-1) and formate (approximately 3.7 g l^-1). The newly isolated K. pneumoniae GLC29 showed potential for the conversion of glycerol into 1,3-propanediol, with high production and productivity.

Full text of DOI:10.1016/j.cattod.2014.05.016

G Model
CATTOD-9075; No. of Pages8
ARTICLE IN PRESS
Catalysis Today xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Production and productivity of 1,3-propanediol from glycerol by
Klebsiella pneumoniae GLC29
Gervásio Paulo da Silvaa,∗, Cristian J. Bolner de Limab, Jonas Contierob
a Microorganism Biotechnology Laboratory, Education Department, Bahia State University (UNEB), BR 407, Km 127, CEP 48970-000, Senhor do Bonfim, BA,
Brazil
b Industrial Microbiology Laboratory, Biochemistry and Microbiology Department, Biological Sciences Institute, São Paulo State University (UNESP), Av 24A,
1515, Bela Vista, CEP 13506-900, Rio Claro, SP, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Interest in the development of the bioproduction of 1,3-propanediol, an important chemical interme-
diate with various industrial applications, has increased in recent years. Klebsiella pneumoniae is one of
the most studied and efficient bacteria for 1,3-propanediol production from glycerol. A new isolate of K.
pneumoniae was investigated using response surface methodology by central composite design for the
production of 1,3-propanediol using glycerol. The effects of pH, temperature, stirrer speed, and glycerol
concentration on the production and productivity of 1,3-propanediol were examined. Considering both
production and productivity, the best conditions for glycerol conversion in 1,3-propanediol are: a pH
range of 6.9–7.1, a temperature between 33 and 38.5 ◦C, a stirrer speed of 110–180 rpm, and a glycerol
concentration of 39–49 g l−1. Batch fermentation carried out at a pH of 7.0, a temperature of 35 ◦C, a stir-
rer speed of 150 rpm, and a glycerol concentration of 40 g l−1 produced 20.4 g 1,3-propanediol l−1, with a
maximum volumetric productivity of 2.92 g l−1 h−1 and a yield of 0.51 g g−1. The main byproducts were
acetic acid (approximately 7.0 g l−1) and formate (approximately 3.7 g l−1). The newly isolated K. pneumo-
niae GLC29 showed potential for the conversion of glycerol into 1,3-propanediol, with high production
and productivity.
Received 27 February 2014
Received in revised form 11 May 2014
Accepted 17 May 2014
Available online xxx
Keywords:
Biodiesel
Glycerol
Fermentation
1,3-Propanediol
Klebsiella pneumoniae GLC29
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
salts such as potassium, and residual alcohol [2,3]. Glycerol must
be highly purified to be used in products such as drugs and food,
In recent years, many countries have adopted biofuels, such
as biodiesel, an alternative diesel fuel produced from renew-
able sources. Their increased use is related to environmental
pollution caused by the emission of greenhouse gases into the
atmosphere mainly through the burning of fossil fuels. Fossil oil
is a non-renewable resource, thus leading to price instability in the
international market. Biodiesel is a mono-alkyl ester of long chain
fatty acids produced from vegetable oils and animal fats. A transes-
terification reaction between the oil and an alcohol in the presence
of a catalyst produces esters (biodiesel) and about 10% glycerol
(w/w) as a byproduct [1]. Glycerol, a substance recognized as safe
for human and animal health, is a chemical with widespread uses
in several industrial sectors, like food, cosmetics and pharmaceu-
tical industries, but the increasing biodiesel production worldwide
is responsible for the generation of a glycerol surplus in the market.
Crude glycerol from biodiesel refineries consists of glycerol, water,
and the high cost of purification makes this process impractical for
the glycerol derived from biodiesel industries [4].
This glycerol surplus is an environmental problem for biore-
fineries since its disposal and/or storage generates economic
costs, raising the final biodiesel price. Therefore, the search for
new glycerol applications is one objective of researchers and
industries. Possibilities for new glycerol utilization by chemicals
[5] or industrial microbiology [1] have been reviewed. Among
possible new applications in the chemical industry is the use of
glycerol as a raw material for the production of various chemicals,
including mesoxalic acid, 1,3-dicloropropanol, glyceryl ethers,
glycerol carbonate, glyceryl esters, hydroxypyruvic acid, and
polyglycerol [5]. Biotechnologically, glycerol can be converted
by microorganisms in a number of valuable chemicals such as
1,3-propanediol, dihydroxyacetone, succinic, propionic, and citric
acids, ethanol, pigments, polyhydroxyalkanoates, biosurfactants,
2,3-butanediol, amino acids, glyceric acids, d-xylulose, hydrogen,
3-hydroxypropionaldehyde, and fatty acids [1,6–16]. Microbial
bioconversion of glycerol has a number of advantages over chem-
ical synthesis, including a greater selectivity on the final products,
∗
Corresponding author. Tel.: +55 74 3541 8929.
E-mail addresses: gpsilva@uneb.br, gpaulosilva@yahoo.com.br (G.P. da Silva).
http://dx.doi.org/10.1016/j.cattod.2014.05.016
0920-5861/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: G.P. da Silva, et al., Production and productivity of 1,3-propanediol from glycerol by Klebsiella
pneumoniae GLC29, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.05.016

G Model
CATTOD-9075; No. of Pages8
ARTICLE IN PRESS
2
G.P. da Silva et al. / Catalysis Today xxx (2014) xxx–xxx
efficiency, and a shorter reaction time. In general, chemical syn-
thesis requires several intermediate steps before reaching the final
product, specific catalysts, and extreme and well controlled con-
ditions such as high temperatures and high pressures [5,9], which
favor the biotechnological route of glycerol bioconversion [17].
One of the main products obtained through microbial bio-
conversion of glycerol is 1,3-propanediol (1,3-PDO), a polyol
with applications in the cosmetics, food, lubricant, and pharma-
ceutical industries [18–21]. The development of polypropylene
terephthalate (PPT), a thermoplastic with superior physicochem-
ical properties to those of polyethylene terephthalate (PET), used
in the production of fabrics, carpets, and engineered plastics has
created a new demand for 1,3-propanediol. Interest in the com-
mercial production of 1,3-PDO through either chemical synthesis
or microbial conversion of glycerol has increased in recent years.
Chemically, 1,3-PDO is produced through two different routes. One
of them uses acrolein (2-propenal) as a raw material, which is
hydrated to 3-hydroxypropionic acid, which is then hydrogenated
in the presence of a catalyst to 1,3-PDO. Another route involves
the hydroformylation of ethylene oxide with CO and H under high
pressure in the presence of a catalyst and a solvent [18]. This reac-
tion produces a dioxane that is hydrogenated to 1,3-PDO. These
methods are expensive, create pollutants, [22] and are dependent
on crude oil.
Bioconversion of glycerol into 1,3-propanediol has been demon-
strated only in bacteria, such as Citrobacter freundii, Klebsiella
pneumoniae, Clostridium pasteurianum, Clostridium butyricum, and
Enterobacter agglomerans [1,18,23,24]. One of the most studied and
efficient bacteria for 1,3-propanediol production is K. pneumoniae
[18,24]. A number works have focused on improving the produc-
tion and productivity of 1,3-propanediol using glycerol. Yang et al.
[25], in fed-batch fermentation at a pH of 7.0, a temperature of
37 ◦C, and microaerobiosis, using mutants of K. oxytoca deficient in
lactate formation and sucrose as a co-substrate, obtained 83.56 g
1,3-PDO l−1, with a yield of 0.62 mol mol−1, and a productivity
of 1.61 g l−1 h−1, with 60.11 g l−1 of 2,3-butanediol (2,3-BDO) pro-
duced in parallel with 1,3-PDO. Seo et al. [26] obtained mutants
deficient in the oxidative pathway, however, the production of 1,3-
PDO was not improved, probably due to a redox imbalance. Zhu
et al. [27] cloned the yqhD gene encoding 1,3-propanediol oxidore-
ductase isoenzyme (PDORI) from Escherichia coli in K. pneumoniae.
The overexpression of PDORI led to a higher 1,3-PDO production,
reaching 67.6 g l−1. In addition, the concentration of the toxic inter-
mediate 3-hydroxypropionaldehyde was reduced by 22.4% when
compared to the original strain. Huang et al. [15] examined the
effects of an over-expressed aldehyde dehydrogenase gene on the
simultaneous production of 3-hydroxypropionic acid and 1,3-PDO
by K. pneumoniae, obtaining 24.4 g l−1 and 49.3 g l−1, respectively. A
high production of 1,3-PDO using co-substrates is recorded by Oh
et al. [28] using K. pneumoniae mutant deficient in carbon catabolite
repression. Under optimized conditions, the concentration of 1,3-
PDO from glycerol was 81.2 g l−1 using molasses as a co-substrate.
Rossi et al. [20] reported concentrations up to 23.80 g 1,3-PDO l−1 in
batch fermentations under controlled pH, while in fed-batch culti-
vations the 1,3-PDO production was 36.86 g l−1 using a new strain
of K. pneumoniae. Sattayasamitsathit et al. [29] applied a statisti-
cal optimization for the simultaneous production of 1,3-PDO and
2,3-BDO using crude glycerol by a new K. pneumoniae isolate. They
reported a yield of 24.98 g 1,3-PDO l−1 and 9.54 g 2,3-BDO l−1. A
high production of 1,3-PDO by a d-lactate deficient mutant of K.
pneumoniae was reported by Xu et al. [30], obtaining 102.06 g 1,3-
PDO l−1 from aerobic fed-batch fermentation. The same approach
was tested by Durgapal et al. [31], which was used to construct a
K. pneumoniae mutant for lactate formation. In glycerol fed-batch
fermentation, the mutant strain produced 58.0 g 1,3-PDO l−1 with
a yield of 0.35 g g−1 and 2,3-BDO as the main byproduct (26.6 g l−1).
In this work, response surface methodology was used to deter-
mine the interaction effect of four independent variables (pH,
temperature, stirrer speed, and glycerol concentration) on 1,3-
propanediol production and productivity by a new K. pneumoniae
isolate.
2. Materials and methods
2.1. Isolation and identification of microorganisms
For the isolation of microorganisms that use glycerol as their
only carbon and energy source, Erlenmeyer flasks containing an
enrichment minimal medium (g l−1 of deionized water: NH4H2PO4
1.0, K2HPO4 1.0, MgSO4 0.2, NaCl 5.0, glycerol 20) were directly
inoculated with different natural samples (soil, decaying plant,
leaves, mosses, etc.) and incubated at 30 ◦C and 100 rpm on a rota-
tory shaker. Pure cultures were obtained by inoculating Petri dishes
containing the same enrichment medium. The potential of differ-
ent isolates were determined by evaluating glycerol bioconversion
in high value-added products (data not shown). The isolate GLC29
showed high potential for glycerol fermentation and was identified
as a K. pneumoniae strain using the Enterobacteriaceae identifica-
tion kit API 20E (Biomerieux, France). The new isolate GLC29 was
characterized for ethanol tolerance, growth pH, osmotolerance, and
carbon sources used. Cultures were maintained refrigerated in agar
slants containing minimal medium, with periodical transference to
new media; long time maintenance was performed in cryotubes
containing glycerol 40% (v/v) at −20 ◦C.
2.2. Glycerol fermentation
For the preparation of the inoculum, the bacterium K. pneumo-
niae GLC29 from agar slants was grown in 500 ml Erlenmeyer flasks
containing 200 ml of autoclaved (121 ◦C/15 min) minimal medium
(g l−1 of deionized water: NH4H2PO4 5.0, K2HPO4 1.0, MgSO4·7H2O
0.2; NaCl 1.0, yeast extract 1.0, glycerol 20) and 200 ␮l of trace ele-
ment solution (g l−1 of deionized water: EDTA 0.5, CaCl2·2H2O 0.5,
CoCl2·6H2O 0.16, MoNH4·4H2O 0.1, CuSO4·5H2O 0.16, FeSO4·7H2O
0.5, MnSO4·H2O 0.5, ZnSO4·7H2O 0.22, NiCl2·6H2O 0.03, H3BO3
0.12). Seed culture flasks were incubated overnight on a rotatory
shaker at 30 ◦C and 100 rpm.
The fermentation medium was the same as for seed inocula.
Glycerol was added in different concentrations, and the pH was
adjusted with NaOH, in agreement with each experiment (Table 1).
Batch fermentations were carried out on a 2 l reactor containing
500 ml final working volume, with a 5% (v/v) inoculum. The pH was
maintained through the automatic addition of 5 M NaOH, and the
temperature was controlled by a water bath with microprocessor-
based temperature control. The fermentation medium was stirred
magnetically with cylindrical stir bars (12 mm × 55 mm). Each fer-
mentation experiment was run in duplicate or triplicate.
2.3. Analytical methods
Fermentation samples were withdrawn periodically for
growth monitoring and to determine residual substrates and
metabolites. Samples for HPLC were centrifuged (approximately
12,000 × g/8 min), and the supernatant was frozen for posterior
analysis. Cell growth was monitored at 600 nm (OD600) on a Bel
Photonics SP-220 UV/vis spectrophotometer. A standard curve
was constructed relating OD600 to cell dry weight (CDW). The pH
was recorded for fermentation monitoring (pH rises after carbon
source depletion), and 5 M NaOH consumption was recorded for
the correction of dilution due to base addition in the reactors.
Residual substrate and metabolites (succinate, lactate, formate,
Please cite this article in press as: G.P. da Silva, et al., Production and productivity of 1,3-propanediol from glycerol by Klebsiella
pneumoniae GLC29, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.05.016

G Model
CATTOD-9075; No. of Pages8
ARTICLE IN PRESS
G.P. da Silva et al. / Catalysis Today xxx (2014) xxx–xxx
Table 1
3
acetate, ethanol, 1,3-propanediol, and 2,3-butanediol) were quan-
tified through high-performance liquid chromatography (HPLC
Thermo Separation Products) equipped with an in-line UV/vis
and RI detectors and an ion-exchange column (300 mm × 7.8 mm
Rezex Roa organic acid – Phenomenex, Torrance, CA, USA). The
samples were run at 60 ◦C and the mobile phase was 0.005 M
Experimental range and levels of the independent variables used in central compos-
ite design.
Independent variables
Range and levels
−˛ −1
5.5 6.0
0
+1
+˛
7.5
40.5
270
55
pH
X1
X2
X3
X4
6.5
33.5
7.0
37
220
45
H2SO4 at 0.4 ml min−1
.
Temperature (◦C)
Agitation (rpm)
26.5
70
30
120
25
170
35
Glycerol (g l−1
)
15
2.4. Experimental design
The central composite design (CCD) was performed to test the
effect of the independent variables pH, temperature, stirrer speed,
and glycerol concentration on the production and productivity of
1,3-PDO by K. pneumoniae GLC29 using glycerol as the only car-
bon and energy source. Each independent variable was arranged at
five levels with eight star points and three replicates at the center
point. A second order polynomial model was used to determine the
optimal interval of each variable for 1,3-PDO production and pro-
ductivity. The variables of the experiments were coded according
to the following equation:
using the CCD method, with five levels each, namely pH (X1), tem-
perature (X2), stirrer speed (X3), and glycerol concentration (X4)
(Tables 1 and 2).
2.4.1. Statistical analysis
The Statistica 7.0 software package (StatSoft, Tulsa, USA) was
used for the experimental design and regression analysis of the
experimental data. Response surface and contour plots were gen-
erated to understand the interaction between different variables.
Statistical analysis of the model was performed to evaluate the
analysis of variance (ANOVA). The quality of the polynomial model
equation was statistically tested by the coefficient of determina-
tion R2 and the statistical significance determined by an F-test. The
significance of the regression coefficients was tested by a t-test.
Xi − Xcp
xi =
where i = 1, 2, . . ., K
(1)
ꢀXi
where xi is the coded value of an independent variable; Xi is the
real value of an independent variable; Xcp is the real value of an
independent variable at the center point; and ꢀXi is the step change
value.
The behavior of the system was explained by the following
quadratic equation:
3. Results
3.1. K. pneumoniae GLC29
ꢀ
ꢀ
ꢀ
Y = B0
+
BiXi +
BiiXi2
+
BijXiXj
(2)
The isolated GLC29 showed a high potential for glycerol fer-
mentation. GLC29 was isolated from bryophytes collected near
the soil, growing at the base of Terminalia catappa and was iden-
tified as a newly isolated K. pneumoniae strain. This bacterium
grows quickly in minimal medium containing only glycerol as its
where Y is the predicted response, i.e., 1,3-PDO concentration
and/or productivity; B0 is the offset term; Bi is the linear effect;
Bii is the squared effect; and Bij is the interaction effect. Xi is ith
independent variable. A total of 27 experiments were carried out
Table 2
CCD values for the optimization of four mathematically predicted and experimentally observed variables for production and productivity of 1,3-propanediol by K. pneumoniae
GLC29 using glycerol.
Essay
pH
Temperature (◦C)
Stirrer speed (rpm)
Glycerol (g l−1
)
1,3-Propanediol
Production (g l−1
Observed
X1
X2
X3
X4
)
Productivity (g l−1 h−1
)
Expected
Observed
Expected
1
2
3
4
5
6
7
8
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
5.5
7.5
6.5
6.5
6.5
6.5
6.5
6.5
6.5
6.5
6.5
30.0
30.0
30.0
30.0
37.0
37.0
37.0
37.0
30.0
30.0
30.0
30.0
37.0
37.0
37.0
37.0
33.5
33.5
26.5
40.5
33.5
33.5
33.5
33.5
33.5
33.5
33.5
120
120
220
220
120
120
220
220
120
120
220
220
120
120
220
220
170
170
170
170
70
25
45
25
45
25
45
25
45
25
45
25
45
25
45
25
45
35
35
35
35
35
35
15
55
35
35
35
13.22
21.88
14.30
22.91
14.19
20.98
12.53
20.50
13.10
23.55
11.98
22.19
14.09
25.05
11.99
21.73
10.29
18.36
17.01
14.45
18.25
17.65
7.73
12.33
20.92
13.5
1.47
1.62
1.94
1.62
1.95
2.22
2.04
1.84
2.46
2.82
1.93
2.82
3
1.66
1.29
2.17
1.67
1.89
2.25
1.76
1.99
2.17
3.13
1.92
2.74
2.97
4.66
2.08
3.63
1.28
3.43
1.76
2.87
3.19
2.68
1.31
2.49
3.53
3.53
3.53
22
12.17
20.14
11.54
19.43
12.99
23.91
12.19
23.02
14.37
24.67
11.77
21.99
13.62
16.84
17.24
16.04
19.62
18.1
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25C
26C
27C
5.03
1.89
3.8
1.33
3.27
1.86
2.66
2.98
2.78
1.34
2.35
3.36
3.96
3.28
270
170
170
170
170
170
9.1
27.46
18.96
19.89
18.78
27.91
19.21
19.21
19.21
Please cite this article in press as: G.P. da Silva, et al., Production and productivity of 1,3-propanediol from glycerol by Klebsiella
pneumoniae GLC29, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.05.016

G Model
CATTOD-9075; No. of Pages8
ARTICLE IN PRESS
4
G.P. da Silva et al. / Catalysis Today xxx (2014) xxx–xxx
Table 3
coefficients for 1,3-PDO production.
Model coefficient estimated by multiple linear regression for 1,3-propanediol pro-
duction by Klebsiella pneumoniae GLC29.
Y(Production) = 19.210 + 0.804X1 − 0.299X2 − 0.380X3 + 4.702X4
+ 0.384X1X2 − 0.491X1X3 + 0.583X1X4 − 0.449X2X3
− 0.154X2X4 − 0.020X3X4 − 0.994X1X1 − 0.643X2X2
Factor
Regression
coefficient
Standard
error
t value
P value
Intercept
X1
X12
X2
X22
X3
X32
X4
X42
X1X2
X1X3
X1X4
X2X3
X2X4
X3X4
19.2100
0.8046
0.9452
0.3342
0.3544
0.3342
0.3544
0.3342
0.3544
0.3342
0.3544
0.4093
0.4093
0.4093
0.4093
0.4093
0.4093
20.3237
2.4076
0.0000
0.0330
0.0158
0.3876
0.0947
0.2772
0.8080
0.0000
0.6269
0.3661
0.2526
0.1797
0.2937
0.7126
0.9606
−0.9943
−0.2996
−0.6430
−0.3804
−0.0880
4.7020
−2.8051
−0.8964
−1.8141
−1.1383
−0.2483
14.0705
−0.4987
0.9391
− 0.088X3X3 − 0.176X4X4
(3)
The quadratic model in Eq. (3) with fourteen terms contains four
linear terms, six factorial interactions and four quadratic terms,
where Y(Production) is the predicted response, i.e., the 1,3-PDO con-
centration in the fermentation medium, and X1, X2, X3, and X4 are
respectively the coded values of the tested variables pH, tempera-
ture, stirrer speed, and glycerol concentration.
Table 3 shows the Student’s t distribution and the correspond-
ing values, along with the estimated parameters. The probability (P)
values were used as a tool to check the significance of each coeffi-
cient. A larger magnitude of the t-test and a smaller P-value denote
greater significance of the corresponding coefficient. As seen in
Table 3, temperature of fermentation (X2) and stirrer speed (X3)
had no effect on the final 1,3-PDO concentration, while the inde-
pendent variables pH (X1) and glycerol concentration (X4) had a
strong positive linear effect on the response (P < 0.05). The same is
observed with the squared variable (X12), where the negative sign
revealed a reduction in 1,3-propanediol production when pH was
increased in the system. The non-significant terms (on the basis
of P-values above 0.05) were neglected, and the model Eq. (3) was
modified to the reduced fitted model (YAdjusted production) of Eq. (4).
−0.1767
0.3844
−0.4918
0.5831
−0.4494
−0.1543
−0.0206
−1.2018
1.4247
−1.0979
−0.3772
−0.0504
X1 = pH; X2 = temperature; X3 = agitation; X4 = glycerol concentration. Meaningful
values at 95% confidence are in italics.
carbon and energy source. GLC29 grows between pH 4 and 10,
with optimal growth between pH 5 and 8. NaCl osmotolerance
varies between 0 and 4.0% (p/v) after 24 h of incubation, but still
grows in higher NaCl concentrations when incubated for a longer
time, that is, 5.5% in 44 h, up to a maximum of 6.5% in 72 h, and
not growing in higher NaCl concentrations. K. pneumoniae GLC29
grows in up to 8% ethanol in mineral medium containing a glyc-
erol concentration of 20 g l−1. In a phenol red carbohydrate broth
containing Durham tubes, this bacterium produces acid and gas
from the following tested carbon sources: glucose, xylose, sucrose,
maltose, lactose, mannitol, mannose, sorbitol, and glycerol. The
products from glycerol fermentation are the 1,3-PDO, 2,3-BDO,
ethanol, and the organic acids acetate, formate, lactate, and suc-
cinate.
YAdjusted
= 18.241 + 0.804X1 + 4.702X4 − 0.812X1X1 (4)
production
The statistical significance of Eq. (4) was checked using an F-
test and the analysis of variance (ANOVA) for the response surface
quadratic model, which is summarized in Table 4. Fisher’s F-test
(F(14,12) = Sm2 /SS2 = 15.828 > Ft(14,12) = 2.62 with a very low prob-
ability value [(Pmodel > F) < 0.0000] indicated that the model was
significant. The goodness-of-fit of the model was checked by the
determination coefficient (R2) and the multiple correlation coeffi-
cient (R). In this case, the R2 value (0.943) indicates that the sample
variation of 94.3% for 1,3-propanediol was attributed to the inde-
pendent variables, and only 5.7% of the total variation cannot be
explained by the model. The adjusted determination coefficient
(Adj. R2 = 0.888) was also satisfactory for confirming the signifi-
cance of the model.
The surface and contour plots were constructed from the
ANOVA results (Table 4), showing the best pH interval for 1,3-PDO
production and the interaction effect between pH and glycerol con-
centration (Fig. 1). Analysis of Fig. 1 shows that the best pH interval
for glycerol conversion in 1,3-PDO was 6.4–7.1, and a higher pro-
duction of 1,3-PDO occurred in higher concentrations of glycerol,
in spite of the fact that productivity is negatively affected by high
glycerol concentrations, as can be seen in the following section.
3.2. Production of 1,3-propanediol by K. pneumoniae GLC29
Preliminary fermentations in a rotatory shaker showed that
GLC29 consumed glycerol quickly, converting it to 1,3-propanediol
as the main fermentation product (data not shown). The central
composite design was used to determine values of pH, temperature,
agitation, and glycerol concentration for optimal bioconversion
of this substrate to 1,3-PDO. Linear and quadratic effects of each
variable and the effect of interaction between them on 1,3-PDO
production and productivity were analyzed. Table 2 presents the
matrix of experiments, observed and expected production (g l−1),
and productivity (g l−1 h−1) of 1,3-PDO. Results of the dependent
variable were used to determine regression coefficients for 1,3-PDO
production (Table 3), to calculate ANOVA (Table 4), and to construct
response surface and contour plots (Fig. 1) for significant results. Eq.
(3) is presented the complete equation as a function ofregression
Table 4
Analysis of variance of the quadratic model for 1,3-propanediol yield by Klebsiella pneumoniae GLC29 using glycerol.
Source
Sum of squares
Degrees of freedom
Mean square
F value
P > F
Model
593.9063
32.1624
21.5609
0.7098
14
12
9
42.4218
2.6802
2.3957
0.3549
15.8278
0.0000
Residual
Lack of Fit
Pure error
6.750
0.1356
2
Total
626.0688
26
R
2 = 0.948; Adj. R2 = 0.888; R = 0.973.
Please cite this article in press as: G.P. da Silva, et al., Production and productivity of 1,3-propanediol from glycerol by Klebsiella
pneumoniae GLC29, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.05.016

G Model
CATTOD-9075; No. of Pages8
ARTICLE IN PRESS
G.P. da Silva et al. / Catalysis Today xxx (2014) xxx–xxx
5
Fig. 1. Contour and surface plots of glycerol fermentation by K. pneumoniae GLC29, showing the best pH intervals (A) and the effect of glycerol concentration on 1,3-propanediol
production (B). Temperature and stirrer speed were not significant for fermentation.
3.3. Productivity of 1,3-propanediol by K. pneumoniae GLC29
terms, six factorial interactions, and four quadratic terms.
Y(Productivity) = 3.533 + 0.539X1 + 0.279X2 − 0.129X3 + 0.296X4
+ 0.143X1X2 − 0.191X1X3 + 0.331X1X4
− 0.161X2X3 + 0.183X2X4 − 0.033X3X4
− 0.295X1X1 − 0.305X2X2 − 0.150X3X3
− 0.408X4X4
Glycerol bioconversion in 1,3-PDO was also analyzed with
respect to maximum productivity, determined in the interval of
time where the largest slope of the straight line plotting 1,3-PDO
productivity versus fermentation time was observed. The results
for productivity, observed and predicted by the statistical model,
are shown in Table 2. At 95% probability, there were statistically
significant effects on 1,3-PDO productivity from pH, temperature,
glycerol concentration, and the quadratic coefficients of all inde-
pendent variables, and interactions between pH and stirrer speed,
pH and glycerol concentration, and temperature and glycerol con-
centration (Table 5). Table 6 presents regression coefficients for the
productivity of 1,3-PDO by K. pneumoniae GLC29 using glycerol.
In order to determine the maximum 1,3-propanediol produc-
tivity corresponding to the optimum interval of pH, temperature,
stirrer speed, and glycerol concentration, a second order polyno-
mial model was proposed to calculate the optimum levels of these
variables. By applying a multiple regression analysis on experimen-
tal data, a second order polynomial model explains the role of each
variable and their second order interactions in producing 1,3-PDO.
The quadratic model in Eq. (5) with 14 terms contains four linear
(5)
Non-significant terms (on the basis of P-values above 0.05 as
shown in Table 5) were neglected and the model Eq. (5) was mod-
ified to the reduced fitted model Eq. (6):
Y(Adjusted
= 3.533 + 0.539X1 + 0.279X2 + 0.296X4
− 0.191X1X3 + 0.331X1X4 − 0.161X2X3
+ 0.183X2X4 − 0.295X1X1 − 0.305X2X2
− 0.150X3X3 − 0.408X4X4
productivity)
(6)
Surface and contour plots for interaction effects between inde-
pendent variables were constructed from ANOVA results (Table 6).
The model can be considered suitable because the percent-
age of variation explained is 94.5% for the complete equation
Table 5
Model coefficient estimated by multiple linear regression for 1,3-propanediol pro-
ductivity by Klebsiella pneumoniae GLC29.
and 88.2% for the adjusted model. The Fisher’s F-test (F(14,12)
=
Sm2 /SS2 = 14.892 > Ft(14,12) = 2.62 with a very low probability
value [(Pmodel > F) < 0.0000] indicated that the model was sig-
nificant. These results allowed defining the best intervals for
independent variables to obtain the highest 1,3-PDO productivity
Factor
Regression
coefficient
Standard
error
Computed t value
P value
Intercept
X1
3.5333
0.5387
0.1771
0.0626
0.0664
0.0626
0.06647
0.0626
0.0664
0.0626
0.0664
0.0767
0.0767
0.0767
0.0767
0.0767
0.0767
19.9496
8.6036
0.0000
0.0000
0.0008
0.0007
0.0006
0.0621
0.0437
0.0004
0.0000
0.0866
0.0286
0.0010
0.0581
0.0342
0.6734
X12
−0.2946
−4.4369
X2
X22
X3
X32
X4
X42
X1X2
X1X3
X1X4
X2X3
X2X4
X3X4
0.2787
4.4515
−0.3046
−0.1287
−0.1496
0.2962
−4.5875
−2.0561
−2.2537
4.7310
Table 6
Analysis of variance for 1,3-propanediol productivity by Klebsiella pneumoniae
GLC29 from glycerol.
Source
Sum of
squares
Degrees of
freedom
Mean
square
F value
P > F
−0.4084
−6.1495
0.1431
1.8662
−0.1906
0.3306
−2.4856
4.3111
Model
19.6194
1.1293
15.7774
0.2762
14
12
19
2
1.4013
0.0941
0.8304
0.1381
14.8915
6.0115
0.0000
0.1521
Residual
Lack of fit
Pure error
−0.1606
−2.0944
0.1831
2.3878
−0.0331
−0.4319
Total
20.7486
26
X1 = pH, X2 = temperature, X3 = stirrer speed, X4 = glycerol concentration. Meaningful
values at 95% confidence are in italic.
R
2 = 0.945, Adj. R2 = 0.882.
Please cite this article in press as: G.P. da Silva, et al., Production and productivity of 1,3-propanediol from glycerol by Klebsiella
pneumoniae GLC29, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.05.016

G Model
CATTOD-9075; No. of Pages8
ARTICLE IN PRESS
6
G.P. da Silva et al. / Catalysis Today xxx (2014) xxx–xxx
Fig. 2. Contour and surface plots of 1,3-propanediol productivity by K. pneumoniae GLC29 using glycerol. Effects of the interaction between pH and temperature (A), pH and
agitation (B), glycerol and pH (C), temperature and agitation (D) and temperature and glycerol (E).
d), and a glycerol concentration between 39 and 49 g l−1 (Fig. 2c
and e). Considering these results, batch fermentation was carried
out using the best intervals for 1,3-PDO production and productiv-
ity.
(Fig. 2). Analyzing contour and surface plots, the best intervals for
higher 1,3-PDO productivity were as follows: a pH between 6.7 and
7.5 (Fig. 2a–c), a temperature ranging from 33 to 38.5 ◦C (Fig. 2a, d
and e), a stirrer speed ranging from 110 to 180 rpm (Fig. 2b and
Please cite this article in press as: G.P. da Silva, et al., Production and productivity of 1,3-propanediol from glycerol by Klebsiella
pneumoniae GLC29, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.05.016

G Model
CATTOD-9075; No. of Pages8
ARTICLE IN PRESS
G.P. da Silva et al. / Catalysis Today xxx (2014) xxx–xxx
7
of temperature had shown no effect on 1,3-PDO production by K.
pneumoniae Blh-1. The results for temperature and pH for 1,3-PDO
production found in this work is different from that of Zhang
et al. [32], who observed that optimal cultivation parameters for
temperature and pH using K. pneumoniae XJ-Li were 40 ◦C and pH
8.0, respectively. Zhang et al. [32] obtained 12.2 g of 1,3-PDO l−1
from 20 g glycerol l−1 consumed in about 8 h, with a productivity
of 1.53 g l−1 h−1 and a yield of 0.75 mol mol−1 (0.61 g g−1). In batch
fermentation at pH 7.0 and 35 ◦C, K. pneumoniae GLC29 produced
20.4 g of 1,3-PDO l−1 from 40 g glycerol l−1 consumed in about 9 h,
with 2.92 g l−1 h−1 of maximal productivity and a yield of 0.51 g
1,3-PDO g−1 glycerol (0.61 mol mol−1). The difference in pH found
between the XJ-Li and GLC29 strains could be explained by the ori-
gin of the isolates, since XJ-Li was isolated from micro-alkaline soil.
A negative effect on 1,3-PDO production when pH was increased
was also observed by Rossi et al. [20].
The pH is one of the parameters that determines the profile
products in fermentation. For example, 2,3-BDO formation starts
at pH 6.6 and reaches a maximum yield at pH 5.5, thereby produc-
ing fewer toxic products such as alcohols or glycols as a strategy
to escape progressive acidity of the medium [33]. However, when
acetate (and formate) is the main byproduct of fermentation, the
1,3-PDO yield is higher, because conversion of pyruvate to acetate
produces ATP but does not consume the reducing equivalents
(NADH) required for 1,3-PDO production [24,34]. In order to mini-
mize the formation of 2,3-BDO (produced between pH 5.0 and 6.5)
and lactic acid (produced between pH 7.1 and 8.0), Ji et al. [35] eval-
uated the production of 1,3-PDO, periodically alternating the pH of
the medium to between 6.3 and 7.3 during fermentation. The best
Fig. 3. Batch glycerol fermentation by K. pneumoniae GLC29 at pH 7.0, 35 ◦C, 150 rpm
and 40 g of glycerol l−1
.
3.4. Batch fermentation
Considering the results obtained for production (g l−1) and pro-
ductivity (g l−1 h−1) of 1,3-PDO by K. pneumoniae GLC29 using
glycerol, batch fermentation was performed at a pH 7.0, a tempera-
ture of 35 ◦C, a stirrer speed of 150 rpm and a glycerol concentration
of 40 g l−1, which are within the optimal intervals for both param-
eters as seen in Figs. 1 and 2. In these conditions, GLC29 grew well
producing about 2.5 g CDW l−1. Glycerol was quickly consumed,
taking just over 9 h of fermentation (Fig. 3). The final concen-
tration of 1,3-PDO was about 20.4 g l−1, with a yiel−d1 of 0.51 g
results obtained were 70 g of 1,3-PDO l−1, a yield of 0.70 mol mol−1
,
and a productivity of 0.97 g l−1 h−1, with low concentrations of 2,3-
BDO and lactate. Another parameter that increases lactate as well as
2,3-BDO production is an excess of glycerol in the medium [24,33].
The best intervals for a higher 1,3-PDO productivity were
between pH 6.7 and 7.5, a temperature range of 33 and 38.5 ◦C,
a stirrer speed of 110–180 rpm, and a glycerol concentration of
between 39 and 49 g l−1. When considering medium agitation, a
higher stirrer speed is harmful to productivity, probably due to
a higher aeration of the medium, which impairs fermentation.
According to Biebl et al. [33], at a neutral pH the 1,3-PDO for-
mation is reduced in the presence of oxygen, which is probably
caused by the oxygen sensitivity of the two enzymes of the 1,3-
PDO pathway. The authors speculate that under microaerobiosis
and low pH, glycerol is converted to 2,3-BDO only, i.e., the NADH is
recycled by oxygen and no longer by glycerol. According to Seifert
et al. [36], inactivation of diol dehydratase by O2 in the absence
of a substrate occurs by a reaction of the activated Co-C bond of
the enzyme-bound coenzyme B12 with O2. Hongwen et al. [37]
recorded maximal activity of key enzymes for 1,3-PDO production
at pH 7.0 and that adequate oxygen favors cell growth, whereas
excess oxygen is not suitable for accumulation of 1,3-PDO. Activity
of the three key enzymes and biomass increased with shaker speed,
with maximum biomass and activities of GDH and PDOR attained
with a shaker speed of 200 rpm. Maximal activities of key enzymes
were obtained at 37 ◦C; in temperatures above 37 ◦C, biomass and
activities greatly decreased [37].
1,3-PDO/g glycerol. Maximal productivity was 2.92 g l
h
−1, and
the maximum rate of consumption of glycerol was 5.26 g l−1 h−1
.
The main byproducts obtained were acetic acid (approximately
7 g l−1) and formate (approximately 3.7 g l−1). Small amounts of
succinate (approximately 1.59 g l−1) and lactate (1.51 g l−1) were
produced. The production of 2,3-BDO and ethanol was less than
1.0 g l−1
.
4. Discussion
The bacteria K. pneumoniae and C. butyricum are the most effi-
cient microbial producers of 1,3-PDO using glycerol. Recently, these
bacteria have been extensively studied for industrial applications
in glycerol fermentation for their high production and productiv-
ity. In this work, preliminary batch fermentations in Erlenmeyer
flasks and a rotatory shaker showed a new strain to be efficient in
fermenting glycerol. This microorganism was identified as K. pneu-
moniae strain GLC29. The fermentation products were 1,3-PDO,
2,3-BDO, ethanol, acetate, lactate, succinate, and formate. In order
to optimize the conditions for 1,3-PDO production, response sur-
face methodology was performed to determine the effect of the
independent variables pH, temperature, stirrer speed, and glycerol
concentration on 1,3-PDO production and productivity.
The fermentation temperature and stirrer speed had no signifi-
cant effect on 1,3-PDO production from glycerol, while the best pH
was between 6.4 and 7.1. A higher production of 1,3-PDO occurred
in higher concentrations of glycerol (Fig. 1a and b and Table 3).
The fact that temperature did not interfere significantly with the
1,3-PDO concentration is interesting for industrial applications,
because it allows working within a wide range of temperatures
without losing yield in the process, reducing the cost of fer-
mentation with respect to processes for which strict control of
this parameter is required. However, productivity is negatively
affected by both high and low temperatures. These results are in
accordance with Rossi et al. [20], who observed that the variation
Productivity decreases with lower and higher glycerol concen-
trations. At high concentrations, glycerol could pass the cell mem-
brane by passive diffusion, and at low concentrations, by glycerol
facilitators [38]. Low substrate concentration has an obvious effect
on transport across the cell membrane and, consequently, there
is little substrate remaining for enzymatic reactions. At high con-
centrations, the enzyme glycerol dehydratase (GDHt, E.C. 4.2.1.30,
convert glycerol to 3-hydroxypropionaldehyde) is the limiting step
for the consumption of glycerol and consequently to produce 1,3-
PDO in K. pneumoniae [39]. However, GDHt is subject to suicide
Please cite this article in press as: G.P. da Silva, et al., Production and productivity of 1,3-propanediol from glycerol by Klebsiella
pneumoniae GLC29, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.05.016

G Model
CATTOD-9075; No. of Pages8
ARTICLE IN PRESS
8
G.P. da Silva et al. / Catalysis Today xxx (2014) xxx–xxx
inactivation by glycerol during catalysis, which involves irre-
versible cleavage of the Co-C C bond of the coenzyme B12, forming
5ꢀ-deoxyadenosine and an alkylcobalamin-like species [36].
The newly isolated K. pneumoniae GLC29 strain contains several
characteristics of a good industrial strain. For example, glycerol was
fermented without strict anaerobic conditions, and no rich nutri-
ent was required for fermentation. These characteristics make the
fermentation process less expensive, which is pursued in industrial
strains. Nevertheless, there are possibilities for better fermentation
results, evaluating new medium compositions, supplementations,
use of co-substrates, inocula, and by genetic engineering. Currently,
the excess of glycerol produced as a byproduct of the biodiesel
industry is of global concern, and several alternatives are being
researched. One of the potential applications is to use glycerol as
a carbon source in industrial microbiology to obtain products of
high added value, and 1,3-propanediol appears to be one of the
most promising alternatives. In this study, K. pneumoniae GLC29
presented a high potential for industrial application.
[11] M.J.A. Stevens, S. Vollenweider, P. Mertes, C. Lacroix, Biochem. Eng. J. 79 (2013)
239–245.
[12] S. Abad, X. Turon, Biotechnol. Adv. 30 (2012) 733–741.
[13] B.T. Maru, A.A.M. Bielen, M. Constantí, F. Medina, S.W.M. Kengen, Int. J. Hydro-
gen Energy 38 (2013) 5563–5572.
[14] H. Habe, T. Fukuoka, D. Kitamoto, K. Sakaki, J. Oleo Sci. 58 (2009) 595–600.
[15] Y. Huang, Z. Li, K. Shimizu, Q. Ye, Bioresour. Technol. 103 (2012) 351–359.
[16] T. Chookaew, S. O-Thong, P. Prasertsan, Int. J. Hydrogen Energy 37 (2012)
13314–13322.
[17] V.E.T. Maervoet, M. De Mey, J. Beauprez, S. De Maeseneire, W.K. Soetaert, Org.
Process Res. Dev. 15 (2011) 189–202.
[18] R.K. Saxena, P. Anand, S. Saran, J. Isar, Biotechnol. Adv. 27 (2009) 895–913.
[19] D.M. Rossi, J.B. da Costa, E.A. de Souza, M.C.R. Peralba, M.A.Z. Ayub, Renew.
Energy 39 (2012) 223–227.
[20] D.M. Rossi, E.A. de Souza, S.H. Flôres, M.A.Z. Ayub, Renew. Energy 55 (2013)
404–409.
[21] G. Kaur, A.K. Srivastava, S. Chand, Biochem. Eng. J. 64 (2012) 106–118.
[22] G. Chotani, T. Dodge, A. Hsu, M. Kumar, R. LaDuca, D. Trimbur, W. Weyler, K.
Sanford, Biochim. Biophys. Acta 1543 (2000) 434–455.
[23] T. Homann, C. Tag, H. Biebl, W.-D. Deckwer, B. Schink, Appl. Microbiol. Biotech-
nol. 33 (1990) 121–126.
[24] H. Biebl, K. Menzel, A.-P. Zeng, W.-D. Deckwer, Appl. Microbiol. Biotechnol. 52
(1999) 289–297.
[25] G. Yang, J. Tian, J. Li, Appl. Microbiol. Biotechnol. 73 (2007) 1017–1024.
[26] M.-Y. Seo, J.-W. Seo, S.-Y. Heo, J.-O. Baek, D. Rairakhwada, B.-R. Oh, P-S. Seo,
M.H. Choi, C.H. Kim, Appl. Microbiol. Biotechnol. 84 (2009) 527–534.
[27] J.-G. Zhu, S. Li, X.-J. Ji, H. Huang, N. Hu, World J. Microbiol. Biotechnol. 25 (2009)
1217–1223.
[28] B.-R. Oh, W.-K. Hong, S.-Y. Heo, L.H. Luo, A. Kondo, J.-W. Seo, C.H. Kim, Bioresour.
Technol. 130 (2013) 719–724.
[29] S. Sattayasamitsathit, P. Prasertsan, P. Methacanon, Process Biochem. 46 (2011)
608–614.
Acknowledgments
This work was financially supported by the Banco do Nordeste
do Brasil (BNB) (grant no. BNB/Uneb 059/2007) and the Fundac¸ ão
de Amparo a Pesquisa do Estado da Bahia (Fapesb) (grant no. PES
0072/2008).
[30] Y.-Z. Xu, N.-N. Guo, Z.-M. Zheng, X.-J. Ou, H.-J. Liu, D.-H. Liu, Biotechnol. Bioeng.
104 (2009) 965–972.
[31] M. Durgapal, V. Kumar, T.H. Yang, H.J. Lee, D. Seung, S. Park, Bioresour. Technol.
159 (2014) 223–231.
References
[32] G. Zhang, B. Ma, X. Xu, C. Li, L. Wang, Biochem. Eng. J. 37 (2007) 256–260.
[33] H. Biebl, A.-P. Zeng, K. Menzel, W.-D. Deckwer, Appl. Microbiol. Biotechnol. 50
(1998) 24–29.
[34] M. González-Pajuelo, I. Meynial-Salles, F. Mendes, P. Soucaille, I. Vasconcelos,
Appl. Environ. Microbiol. 72 (2006) 96–101.
[35] X.-J. Ji, H. Huang, J.-G. Zhu, N. Hu, S. Li, Appl. Biochem. Biotechnol. 159 (2009)
605–613.
[36] C. Seifert, S. Bowien, G. Gottschalk, R. Daniel, Eur. J. Biochem. 268 (2001)
2369–2378.
[37] C. Hongwen, F. Baishan, H. Zongding, Biochem. Eng. J. 25 (2005) 47–53.
[38] Y.Q. Sun, W.T. Qi, H. Teng, Z.L. Xiu, A.P. Zeng, Biochem. Eng. J. 38 (2008)
22–32.
[39] K. Ahrens, K. Menzel, A.-P. Zeng, W.-D. Deckwer, Biotechnol. Bioeng. 59 (1998)
544–552.
[1] G.P. da Silva, M. Mack, J. Contiero, Biotechnol. Adv. 27 (2009) 30–39.
[2] M. Metsoviti, K. Paraskevaidi, A. Koutinas, A.-P. Zeng, S. Papanikolaou, Process
Biochem. 47 (2012) 1872–1882.
[3] S. Papanikolaou, S. Fakas, M. Fick, I. Chevalot, M. Galiotou-Panayotou, M.
Komaitis, I. Marc, G. Aggelis, Biomass and Bioenergy 32 (2008) 60–71.
[4] Z. Chi, D. Pyle, Z. Wen, C. Frear, S. Chen, Process Biochem. 42 (2007) 1537–1545.
[5] Y. Zheng, X. Chen, Y. Shen, Chem. Rev. 108 (2008) 5253–5277.
[6] K. Petrov, P. Petrova, Appl. Microbiol. Biotechnol. 84 (2009) 659–665.
[7] D. Rittmann, S.N. Lindner, V.F. Wendisch, Appl. Environ. Microbiol. 74 (2008)
6216–6222.
[8] M. Khamduang, K. Packdibamrung, J. Chutmanop, Y. Chisti, P. Srinophakun, J.
Ind. Microbiol. Biotechnol. 36 (2009) 1267–1274.
[9] H. Habe, T. Fukuoka, D. Kitamoto, K. Sakaki, Appl. Microbiol. Biotechnol. 84
(2009) 445–452.
[10] S.J. Sarma, S.K. Brar, E.B. Sydney, Y. Le Bihan, G. Buelna, C.R. Soccol, Int. J. Hydro-
gen Energy 37 (2012) 6473–6490.
Please cite this article in press as: G.P. da Silva, et al., Production and productivity of 1,3-propanediol from glycerol by Klebsiella
pneumoniae GLC29, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.05.016