Energy- And cost-effective non-sterilized fermentation of 2,3-butanediol by an engineered: Klebsiella pneumoniae OU7 with an anti-microbial contamination system

Article abstract of DOI:10.1039/d0gc03044a

Microbial contamination is a serious challenge that needs to be overcome for the successful biosynthesis of 2,3-butanediol (2,3-BD). However, traditional strategies such as antibiotic administration or sterilization are costly, have high energy demands, and may increase the risk of antibiotic resistance. Here, we intend to develop a robust strategy to achieve non-sterilized fermentation of 2,3-BD. Briefly, the robust strain can metabolize unconventional chemicals as essential growth nutrients, and therefore, outcompete contaminant microbes that cannot use unconventional chemicals. To this end, Klebsiella pneumoniae OU7, a robust strain, was confirmed to rapidly exploit urea and phosphite (unconventional chemicals) as the primary sources of nitrogen (N) and phosphorus (P), and withstand deliberate contamination in the possibly contaminated systems. Secondly, metabolic engineering, pathogenicity elimination and adaptive laboratory evolution were successively performed, endowing the best strain with an excellent fermentation performance for safe 2,3-BD production. Finally, 84.53 g L-1 of 2,3-BD was synthesized with a productivity of 1.17 g L-1 h-1 and a yield of 0.38 g g-1 under the non-sterilized system. In summary, our technique reduces labor and energy costs and simplifies the fermentation process because sterilization does not need to be performed. Thus, our work will be beneficial for the sustainable synthesis of 2,3-BD. This journal is

Full text of DOI:10.1039/d0gc03044a

View Article Online
View Journal
Green
Chemistry
Cutting-edge research for a greener sustainable future
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: Z. Guo, X. Ou, P.
Xu, H. Gao, L. Zhang, M. Zong and W. Lou, Green Chem., 2020, DOI: 10.1039/D0GC03044A.
Volume 20
This is an Accepted Manuscript, which has been through the
Number
May 2018
Pages 1919-2160
9
7
Royal Society of Chemistry peer review process and has been
accepted for publication.
Green
Chemistry
Cutting-edge research for a greener sustainable future
rsc.li/greenchem
Accepted Manuscripts are published online shortly after acceptance,
before technical editing, formatting and proof reading. Using this free
service, authors can make their results available to the community, in
citable form, before we publish the edited article. We will replace this
Accepted Manuscript with the edited and formatted Advance Article as
soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes to the
text and/or graphics, which may alter content. The journal’s standard
Terms & Conditions and the Ethical guidelines still apply. In no event
shall the Royal Society of Chemistry be held responsible for any errors
or omissions in this Accepted Manuscript or any consequences arising
from the use of any information it contains.
ISSN 1463-9262
PAPER
Paul T. Anastas et al.
The Green ChemisTREE: 20 years after taking root with the 12
principles
rsc.li/greenchem

Page 1 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
Energy- and cost-effective non-sterilized fermentation of 2,3-butanediol
by an engineered Klebsiella pneumoniae OU7 with anti-microbial
contamination system
Ze-Wang Guoa, Xiao-Yang Oua, Pei Xua, Hui-Fang Gaob, Liao-Yuan Zhangb, Min-Hua
Zonga, Wen-Yong Loua,*
aSchool of Food Science and Engineering, South China University of Technology, No. 381
Wushan Road, Guangzhou 510640, Guangdong, China
bCollege of Life Sciences, Fujian Agriculture and Forestry University, No. 15
Shangxiadian Road, Fuzhou 350002, Fujian, China
*Correspondence: wylou@scut.edu.cn
†Electronic supplementary information (ESI) available.

Green Chemistry
Page 2 of 37
View Article Online
DOI: 10.1039/D0GC03044A
Abstract
Microbial contamination is a serious challenge that needs to be overcome for the
successful biosynthesis of 2,3-butanediol (2,3-BD). However, traditional strategies such as
antibiotic administration or sterilization are costly, have high energy demands, and may
increase the risk of antibiotic resistance. Here, we intend to develop a robust strategy to
achieve non-sterilized fermentation of 2,3-BD. Briefly, the robust strain can metabolize
unconventional chemicals as essential growth nutrients, and therefore, outcompete
contaminant microbes that cannot use unconventional chemicals. To this end, Klebsiella
pneumoniae OU7, a robust strain, was confirmed to rapidly exploit urea and phosphite
(unconventional chemicals) as the primary sources of nitrogen (N) and phosphorus (P), and
withstand deliberate contamination in the possibly contaminated systems. Secondly,
metabolic engineering, pathogenicity elimination and adaptive laboratory evolution were
successively performed, endowing the best strain with an excellent fermentation
performance for safe 2,3-BD production. Finally, 84.53 g/L of 2,3-BD was synthesized
with a productivity of 1.17 g/L/h and a yield of 0.38 g/g under the non-sterilized system.
In summary, our technique reduces labor and energy costs and simplifies the fermentation
process because sterilization does not need to be performed. Thus, our work will be
beneficial for the sustainable synthesis of 2,3-BD.

Page 3 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
Keywords
Biocontamination management, robust, phosphorus and nitrogen metabolism, metabolic
engineering, non-sterilization fermentation

Green Chemistry
Page 4 of 37
View Article Online
DOI: 10.1039/D0GC03044A
Introduction
2,3-Butanediol (2,3-BD) is a value-added chemical because of its strong demands in
the manufacture of moisteners, explosives, fumigants and pharmaceuticals, perfumes,
plasticizers, softening agents and printing inks.1–3 Additionally, 2,3-BD is a potential
aviation fuel with a high octane number, an excellent heating value (27.2 kJ/g), and a low
freezing point.4,5 2,3-BD can be synthesized via chemical and biotechnological methods.6
Recently, the sustainable synthesis of 2,3-BD via biotechnological methods has recently
attracted growing attention because of the global environmental crisis and the shortage of
fossil fuels.7 However, unlike chemical processing, biosynthesis of 2,3-BD is generally
conducted in a sterile system to avoid microbial contamination.
Microbial contamination affects the safety and quality of the biological product.
Additionally, it results in significant economic losses to the industry. There are many
reasons for microbial contamination, such as the incomplete sterilization of equipment and
fermentation media, seeds carrying undesirable microbes, improperly sterilized air supply,
and poor sealing of fermentation equipment. Traditionally, process sterilization and the
addition of antibiotics were extensively adopted to decrease the probability of microbial
contamination.8,9 However, sterilization is a time-consuming, energy-intensive and
laborious process.10 Moreover, staff need higher-level skills to ensure successful

Page 5 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
fermentation.11 On the other hand, the overuse of antibiotics will promote the emergence
of “superbugs” and increase the risk of antibiotic resistance,12 which reduces the
competitiveness and sustainability of the 2,3-BD biosynthesis. Therefore, there is a need
to develop novel strategies to efficiently prevent biocontamination.
In previous studies, we observed a growing interest in non-sterilized fermentation
because of its advantages, including i) preventing the Maillard reaction, ii) providing
significant savings in the costs of energy, equipment, and labor, and iii) simplifying the
fermentation process.13 To this end, non-sterilization fermentation of lactic acid and acetoin
was achieved using the thermophilic B. coagulans and Bacillus sp. H15-1, respectively.14,15
Additionally, B. licheniformis was frequently reported to synthesize 2,3-BD at
thermophilic conditions, which also exhibited the potential for non-sterilized
fermentation.6,16–18 Nevertheless, these reports had some restrictions related to non-
sterilized fermentation of 2,3-BD. On the one hand, microbial contamination was still
observed under the thermophilic fermentation since the thermophilic and spore‐forming
strains are ubiquitous.9 Worse still, diacetyl and acetoin, the precursors of 2,3-BD, will
volatilize easily during thermophilic fermentation. On the other hand, fermentation was
conducted under mesophilic conditions (30-40°C) in the most of 2,3-BD producers, which

Green Chemistry
Page 6 of 37
View Article Online
DOI: 10.1039/D0GC03044A
is infeasible for thermophilic fermentation. Without medium sterilization, it is difficult to
prevent contamination completely.
Recently, essential growth nutrients (ecologically rare compounds) have been used to
provide selective pressure for the dominant growth of robust strains without antibiotics.
For this purpose, phosphite dehydrogenase (PtxD), an enzyme producing phosphate from
phosphite (an unconventional type of phosphorus), was successfully introduced into
different hosts.19–22 Additionally, the whole melamine utilization gene cluster was
assembled into E. coli or Synechococcus, endowing the modified organisms with the
capability to resist harmful contamination.19,20 Similarly, E. coli and B. subtilis were
tailored by simultaneously introducing the formamide and phosphite utilization pathways,
rendering them suitable for defending against microbial contamination.9,23
In this work, we intend to develop an effective strategy to achieve non-sterilization
fermentation of 2,3-BD (Fig. 1). First, K. pneumoniae OU7, a robust strain with an
endogenous anti-hybrid bacteria system, was confirmed to rapidly use urea and phosphite
as the sole nitrogen (N) and phosphorus (P) sources, respectively. Second, a triple mutant
of K. pneumoniae OU7 with inactivation of lactate dehydrogenase (LDH), alcohol
dehydrogenase (ADH) and phosphotransacetylase (PTA) was obtained. Moreover, the
wabG gene was disrupted to abolish the pathogenicity of K. pneumoniae OU7. Lastly, the

Page 7 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
efficiency and economic feasibility of 2,3-BD production using the engineered K.
pneumoniae OU7 were evaluated in non-sterilized fed-batch fermentation. On the whole,
the non-sterilized fermentation of 2,3-BD proposed here perfectly fulfills the principles of
environmentally benign and sustainable chemistry.
Experimental section
Strains, cultivation and medium
2,3-BD was produced using K. pneumoniae OU7, isolated in our previous work.9 The
gene‐deficient mutants of K. pneumoniae OU7 were created in this work (Table 1). All
strains were maintained on LB agar slants. The wild-type K. pneumoniae OU7 and its
mutants were cultivated in auxotrophic or conventional MOPS medium. The conventional
MOPS medium is comprised of Tricine (4 mM), MOPS (40 mM), FeSO4 (0.01 mM),
K2SO4 (0.276 mM), CaCl2 (5×10-4 mM), MgCl2 (0.523 mM), NaCl (50 mM), NH4Cl (9.5
mM), and KH2PO4 (1.32 mM). They were also supplied with 0.5 mL of trace metal solution:
0.4 mM H2BO4, 0.08 mM MnCl2, 0.01 mM CuSO4, 0.03 mM CoCl2, and 0.01 mM ZnSO4.
In auxotrophic MOPS medium, NH4Cl and KH2PO4 were replaced by urea and KH2PO3,
respectively. The seed or fermentation medium was obtained by adding 10 or 80 g/L of
glucose to MOPS medium with 2.0 mM KH2PO3, 10 g/L urea, and 0.1 g/L NiCl2.
Robustness analysis for K. pneumoniae OU7

Green Chemistry
Page 8 of 37
View Article Online
DOI: 10.1039/D0GC03044A
K. pneumoniae OU7 was cultivated in auxotrophic or conventional MOPS medium.
Conventional medium (with 10 g/L glucose) was prepared by adding 9.5 mM NH4Cl and
1.32 mM KH2PO4 as the sources of N and P. Auxotrophic medium (with 10 g/L glucose)
was prepared by adding 9.5 mM urea (with 0.1 g/L NiCl2) and 1.32 mM KH2PO3 as the
sources of N and P. The viable cell count was tested by plating culture fluid on LB-agar
plates and then incubating overnight at 37 °C. Subsequently, K. pneumoniae OU7 and other
microbes were classified by their unique colony morphology.23
Construction of the gene knockout strains
To construct gene knockout strains, both up- and down-stream fragments
(approximately 0.95 kB) of target genes were amplified from K. pneumoniae OU7.
Subsequently, the two fragments were combined and inserted into the suicide vector
pRE112 by in-fusion cloning, and then transferred into E. coli S17-1 to obtain the donor
strains. For the first homologous recombination, the recipient strain (300 µL) was
cocultured with the donor strain (900 µL) in a non-selective LB Petri plate (37 °C, 12 h).
This was then plated on selective LB agar plates (containing 100 µg/mL ampicillin and 30
µg/mL chloramphenicol) to obtain positive colonies. A single crossover strain was
incubated in non-selective LB broth overnight for the second homologous recombination,
then plated on 12% sucrose LB Petri plates for selection and verified by PCR method.24
All strains and plasmids are listed in Table 1. All primers are listed in Table S1.

Page 9 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
Enzyme activity assays
Cells of K. pneumoniae in different growth phases were collected by centrifugation
(10000 g, 5 min, 4°C) and completely resuspended with 50 mM precooled phosphate buffer
(pH 7.4). Subsequently, cells were disrupted by ultrasonication (10 min, 4°C), and the
lysates were centrifuged (12000 g, 20 min, 4°C) to obtain the crude cell extracts. The LDH,
ADH and PTA activity were assayed according to the method established by Guo et al. .25
The PtxD and urease activity was determined as described by Ou et al. and Pearson et al. ,
respectively.9,26
Electron microscopy analysis
Surface morphology and microstructure of KP-OU7-ME3 and KP-OU7-ME4 were
studied by field-emission scanning electron microscopy (FE-SEM, Hitachi, SU-8010) and
transmission electron microscopy (TEM, Hitachi, H‐7650), respectively. The detailed
procedures for preparing bacterial specimens for FE-SEM and TEM analysis were the same
as described by Jung et al. .27
Adaptive laboratory evolution (ALE)
ALE was conducted using a long-term serial transfer experiment. Briefly, every 12 h,
the seed culture (1% v/v) of evolved strain was rapidly transferred into a new and fresh
auxotrophic MOPS medium (containing 80 g/L glucose). The transfers were performed

Green Chemistry
Page 10 of 37
View Article Online
DOI: 10.1039/D0GC03044A
until the clones showed rapid glucose consumption, fast growth, and high production of
2,3-BD.28
Fermentation
For seed preparation, K. pneumoniae OU7 from a fresh slant tube was picked into
conical flasks (250-mL) containing 50-mL seed medium and inoculated overnight on a
rotary shaker (37°C, 200 rpm). Subsequently, seed culture (5%, v/v) was transferred into
fermentation broth.29 Batch fermentations were performed in 250-mL shake flasks with 50-
mL working volume. Fed-batch fermentation of KP-OU7-ME5 occurred in a 5-L stirred
fermenter with a 3-L initial volume. Fermentation was performed at 37℃ with the pH
maintained at 6.5 by automatic addition of 2 M NaOH, stirring at 300 rpm before 16 h and
then switched to 200 rpm with 1.0 vvm of airflow. Additionally, 2 mM KH2PO3 was fed
into the fermentation medium at 24, 40, and 56 h, to sustain long term growth for KP-OU7-
ME5 mutant.
Analytical methods
The growth rate of cells was determined using optical cell density at a 600 nm
wavelength. Measurements from a Shimadzu 00671UV‐2550 spectrophotometer
(Shimadzu Corporation, Japan) were converted to dry cell weight (DCW) according to the
equation (DCW (g/L) = 0.6348 × OD600 + 0.3257). The intracellular NADH level or NAD+

Page 11 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
level was determined using the NAD(H) Assay Kit (BC0315, Solarbio, Beijing, China).
HPLC analyses of 2,3-BD, glucose, lactate, acetoin, acetate, and ethanol were implemented
as described in a previous study.25 Briefly, RID-10A refractive index detector, Aminex
HPX-87H column (Bio-Rad) at 65°C, a mobile phase of 5 mM H2SO4, and a flow rate of
0.5 mL/min was used.
Results and discussion
K. pneumoniae OU7 grew on the auxotrophic MOPS medium
The 16S rRNA sequence of the K. pneumoniae OU7 strain was identified. Meanwhile,
further morphological and biochemical assays were also performed to identify strain (data
not shown). The alignment result (Fig. S1) showed that the 16S rRNA sequence of K.
pneumoniae OU7 (GenBank: MN999954.1) shared a strong identity (99%) with the that of
K. pneumoniae ZG26 and Klebsiella sp. F5‐1.
Previous work from our group demonstrated that the K. pneumoniae OU7 could
exploit phosphite as the sole P source for cell growth.9 Additionally, Klebsiella species
have the natural capability to utilize urea as the sole N source.30 In this study, to verify that
K. pneumoniae OU7 could simultaneously utilize phosphite and urea as the sole P and N
sources, the strain was grown in auxotrophic medium. First, we verified that K. pneumoniae
OU7 could grow normally in the phosphite-loaded medium when the concentration of
KH2PO3 was lower than 2 mM (Fig. 2A and 2B). However, there was marked inhibition

Green Chemistry
Page 12 of 37
View Article Online
DOI: 10.1039/D0GC03044A
on cell growth when the concentration of KH2PO3 was higher than 2 mM (Fig. 2B). To the
best of our knowledge, the reason for inhibition remains unclear, and this should be the
subject of further investigation in future work.
The growth of cells was significantly affected in the urea-loaded medium when the
medium did not have NiCl2 (Fig. 2A). Luckily, cell growth returned to normal when the
NiCl2 was added. This phenomenon can be explained since the urease produced by
Klebsiella species was activated by nickel.31 Therefore, NiCl2 is necessary when urea is
used as the N source. Moreover, it was shown that cell growth was not affected when the
concentration of urea was lower than 20.0 g/L (Fig. 2C). Finally, phosphite utilization gene
operon (ptxABCDE) and urea utilization gene cluster (ureDABCEFG) were verified from
whole genomic sequencing (Fig. 2D and 2E), both the PtxD and urease activity could be
detected (Fig. 2F). Taken together, these results strongly suggest that endogenous
phosphite and urea assimilating pathways exist in K. pneumoniae OU7.
In conclusion, K. pneumoniae OU7, a robust strain, was confirmed to rapidly exploit
phosphite and urea as P and N sources, respectively, for its growth (Fig. 1). One primary
consideration of this work’s robust strain is that phosphite is safe and inexpensive, enabling
us to perform economically viable non-sterilized fermentation on an industrial scale. On
the other hand, using urea as N source is favorable because: 1) urea has a low cost compared

Page 13 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
with ammonium salts or yeast extract; and 2) unlike ammonium salts, growth in urea-
containing medium suppresses the medium acidification rate.31
Robustness analysis of K. pneumoniae OU7
As shown in Fig. 3D, K. pneumoniae OU7 accounted for, on average, 94.9% of the
population when cultivated in a non-sterilized auxotrophic medium. However, the average
proportion of K. pneumoniae OU7 was only 88.5% when cultivated in a non-sterilized
conventional medium (Fig. 3A). Additionally, the slight growth of the undesirable
microorganisms observed is, as shown in Fig. 3D, most likely come from the starting
ingredients used to prepare the medium.
Second, K. pneumoniae OU7 was cocultured with E. coli (as a typical prokaryotic
microorganism) or S. cerevisiae (as a typical eukaryotic microorganism) to evaluate
whether the robust K. pneumoniae OU7 was able to outcompete other microorganisms.
Encouragingly, with time going on, the proportion of K. pneumoniae OU7 rapidly
increased. Eventually, the proportion of K. pneumoniae OU7 approached 100% in the
auxotrophic medium (Fig. 3E and 3F). By contrast, the proportion of K. pneumoniae OU7
was maintained at a relatively low level when cultivated in conventional medium (Fig. 3B
and 3C). The above experiments demonstrated that the simultaneous utilization of
phosphate and urea, which are less favorable to assimilate for common microbes, allowed
the K. pneumoniae OU7 a capability to overcome and resist deliberate contamination (Fig.

Green Chemistry
Page 14 of 37
View Article Online
DOI: 10.1039/D0GC03044A
1). This finding suggested that K. pneumoniae OU7 is indeed suitable for 2,3-BD
fermentation under non-sterilized conditions.
Metabolic engineering of K. pneumoniae OU7
Currently, undesirable by-products (such as lactate, ethanol, and acetate) are also
produced during the 2,3-BD fermentation process (Fig. 1), which shunts the carbon flow
and dramatically decreases the efficiency of 2,3-BD synthesis.25,28 Thus, we proposed to
block pathways that competitively consume NADH to redistribute carbon flux towards 2,3-
BD synthesis. Consequently, PCR analysis and enzymatic activity results confirmed that
ldhA (encoding lactate dehydrogenase), adhE (encoding alcohol dehydrogenase), and pta
(encoding phosphotransacetylase) were successfully disrupted (Fig. S2 and Table S2),
resulting in gene‐deficient mutants (Table 1).
To evaluate the influence of ldhA, adhE and pta inactivation on fermentation
performances, batch fermentation was performed. Compared with the KP-OU7, the
maximum DCW of KP-OU7-ME1 was elevated from 4.13 to 4.41 g/L (Fig. 4A). The
increase in DCW may result from a decrease in lactate formation, which leads to a slower
acidification rate of the medium.32 By contrast, the maximum DCW of KP-OU7-ME2
(4.27 g/L) and KP-OU7-ME3 (4.08 g/L) was lower than that of KP-OU7-ME1 (Fig. 4A).
The titer of different byproducts was different in the gene‐deficient mutants (Fig. 4A).
Firstly, the final titers of lactate in KP-OU7-ME1, KP-OU7-ME2 and KP-OU7-ME3 were

Page 15 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
only 1.38, 1.56 and 1.41 g/L, respectively, and this corresponded to 28.11%, 31.77% and
28.72% of the KP-OU7 strain (4.91 g/L), respectively. Secondly, the deletion of ADH
resulted in a nearly abolished ethanol synthesis, the ethanol concentration of KP-OU7-ME2
(0.28 g/L) decreased 88.38% compared with the parent strain (2.41 g/L). Moreover,
compared with the KP-OU7 (1.81 g/L), the deletion of the pta gene caused a slight
reduction of acetate production (1.43 g/L) in the KP-OU7-ME3 strain. Taken together, the
simultaneous inactivation of LDH, ADH and PTA could effectively alleviate lactate,
ethanol and acetate accumulation.
As shown in Fig. 4A, the final titer of acetoin in KP-OU7-ME1, KP-OU7-ME2 and
KP-OU7-ME3 were only 2.88, 2.71 and 2.28 g/L, respectively, and this corresponded to
84.96%, 79.94% and 67.26% of the KP-OU7 strain (3.39 g/L), respectively. Encouragingly,
the final 2,3-BD titer in KP-OU7-ME1, KP-OU7-ME2 and KP-OU7-ME3 reached 22.66,
22.87 and 23.41 g/L, respectively, which corresponded to a dramatic increase of 18.08%,
19.18% and 22.0%, respectively, compared with the parent strain (19.19 g/L) (Fig. 4A).
The reasons for these higher titers of 2,3-BD are: i) deletion of ldhA, adhE and pta partially
blocking the lactate, ethanol, and acetate pathway and supplying more carbon flux for the
mutants to produce 2,3-BD, and ii) the slightly increase in the NADH/NAD+ ratio that was
observed in the gene‐deficient mutants (Fig. 4D) which resulted in the decreased
accumulation of acetoin.

Green Chemistry
Page 16 of 37
View Article Online
DOI: 10.1039/D0GC03044A
2,3-BD biosynthesis is highly linked with the intracellular NADH/NAD+ ratio. In Fig.
4B and 4C (all strains), both NADH and NAD+ concentrations were continuously declined
during the growth stage (≤ 32 h). However, in the non-growth stage, the levels of NADH
and NAD+ were maintained relatively constant. Furthermore, compared with the wild strain,
the inactivation of ldhA, adhE and pta resulted in the NADH/NAD+ ratios being maintained
at a higher level (Fig. 4D). These results indicated that the disruption of NADH-dependent
pathways would affect the ratio of intracellular NADH/NAD+ and facilitate the synthesis
of 2,3-BD.
Removal of the pathogenic factor by inactivating the virulence-related wabG gene
Currently, the practical application of K. pneumoniae for producing 2,3-BD is
severely hindered by its pathogenicity (generally defined as biosafety class II
microorganism), which exhibits potential harm to users and increases the cost of industrial-
scale production.27 Previous studies have shown that the virulence factors of
K. pneumoniae strains include lipopolysaccharide (LPS), capsular polysaccharide,
siderophores, and fimbriae.33 Among them, LPS is a major component of the cell wall,
which is considered an essential pathogenic determinant of K. pneumoniae. As shown in
Fig. S3C, LPS comprises three sections, O-antigenic polysaccharide (O-PS), core
oligosaccharide (OS), and lipid A. Particularly, O-antigen is responsible for the resistance
of the bacteria to complement-mediated killing. The outer core LPS is also involved in the

Page 17 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
attachment of capsules to the surface.27,34,35 Previously, extensive success stories
demonstrated that the deletion of the wabG gene (encoding glycosyltransferase) involved
in LPS biosynthetic could eliminate the pathogenicity of K. pneumoniae.27,36
Here, the core oligosaccharide (OS) biosynthetic gene cluster was verified and
annotated from whole genomic sequencing (Fig. S3A and S3B). Subsequently, based on
KP-OU7-ME3, the wabG gene was disrupted, yielding the strain KP-OU7-ME4.
Additionally, surface morphologies of KP-OU7-ME3 and KP-OU7-ME4 were examined
by SEM. Predictably, the bacterial surface of KP-OU7-ME3 was surrounded by a thick
capsular polysaccharides (CPS) layer. Moreover, the bacterial boundaries were not clear
because of mucoid CPS (Fig. 5A). By contrast, KP-OU7-ME4 showed a nonmucoid colony
morphology (Fig. 5B), consistent with the previous studies.27 On the other hand, the KP-
OU7-ME3 exhibited a clear and smooth outer membrane (Fig. 5C). However, in KP-OU7-
ME4, the outer membrane’s contour was irregular and unclear (Fig. 5D). In other words,
the integrity of the outer membrane was destroyed. Taken together, the LPS of K.
pneumoniae OU7 was successfully tailored to abolish its pathogenicity.
In batch fermentation, the maximum DCW of KP-OU7-ME3 and KP-OU7-ME4 were
4.08 and 4.15 g/L, respectively (Fig. 5E). These data indicated that the absence of outer
core LPS did not affect the cell growth of K. pneumoniae OU7. Unfortunately, the 2,3-BD
titer of KP-OU7-ME4 (20.75 g/L) was slightly lower than that of KP-OU7-ME3 (23.41

Green Chemistry
Page 18 of 37
View Article Online
DOI: 10.1039/D0GC03044A
g/L). As described by Jung et al., the main reason for the decrease of 2,3-BD titer in KP-
OU7-ME4 may result from the change of outer membrane integrity, which leads to a
slower utilization rate of glucose.27 Subsequently, adaptive laboratory evolution (ALE) was
performed to enhance 2,3-BD production. After 30 adaptation cycles, a maximum DCW
of 4.31 g/L and 2,3-BD titer of 22.96 g/L was obtained in the endpoint strain KP-OU7-
ME5 (Fig. 5E), which was 3.86% and 10.65% higher than that of KP-OU7-ME4 strain,
respectively. Therefore, the KP-OU7-ME5 strain was used in subsequent fed-batch
fermentation.
Effects of organic nitrogen on the synthesis of 2,3-BD
The yield of 2,3-BD was only 0.29 g/g in the KP-OU7-ME5 strain (Fig. 5E), and this
was lower than the theoretical yield (0.50 g/g). This data demonstrated that the nutrients
present in the MOPS medium were insufficient for 2,3-BD synthesis, even though the
MOPS medium contained all nutrients for K. pneumoniae OU7 growth. Previous studies
demonstrated a significant improvement of 2,3-BD synthesis when additional nitrogen
sources were supplemented into the fermentation medium.31
Here, both the biomass yield and 2,3-BD titer could be enhanced when organic
nitrogen was added into the MOPS medium (Fig. S4A). When yeast extract was added,
2,3-BD titer and DCW were 27.67 g/L and 7.60 g/L, respectively, corresponding to 121.41%
and 177.57% of the control (22.79 g/L and 4.28 g/L). Unfortunately, many undesirable

Page 19 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
microorganisms were observed when different nitrogen sources were supplemented into
the MOPS medium (Fig. S4B), which means that non-sterilized fermentation of 2,3-BD is
not suitable under these conditions. Therefore, nonsterile auxotrophic MOPS medium
(without additional nitrogen sources) was used in the subsequent fed-batch experiments.
Fed-batch fermentation of KP-OU7-ME5 in the non-sterilized system
During the fermentation process, 2,3-BD synthesis is highly dependent upon the
consumption of substrate.37 Therefore, a fed-batch experiment was conducted to
accumulate 2,3-BD using an periodic feeding strategy. The glucose was injected into the
fermenter when its concentration was less than 10 g/L. Additionally, based on the previous
studies, agitation speed was controlled using an efficient two-stage control strategy.29 As
shown in Fig. 6A, acetoin showed a sharp accumulation before 16 h with 300 rpm agitation
speed and then continuously decreased after 16 h with 200 rpm agitation speed. The main
reason is that 2,3-BD, production from acetoin, is via an NADH-linked pathway (Fig. 1).
Before 16 h, the dissolved oxygen concentration was relatively high, leading to a down-
regulation of the NADH pool (the NADH was rapidly oxidized via the electron transfer
chain), resulting in acetoin accumulation. After 16 h, the level of NADH was relatively
high, corresponding to the insufficient oxygen, thus rapid converting acetoin to 2,3-BD.
Acetate was formed as the main by-product with a final titer of 3.75 g/L. In contrast,
the titers of ethanol and lactate were lower than 2.0 g/L throughout the fermentation process.

Green Chemistry
Page 20 of 37
View Article Online
DOI: 10.1039/D0GC03044A
These phenomena suggested that other LDH, ADH and PTA manufacturing genes might
exist in K. pneumoniae OU7. Encouragingly, 2,3-BD titer continuously increased with
glucose feeding. Ultimately, 222.92 g/L of glucose was consumed, resulting in a maximum
2,3-BD titer up to 84.53 g/L with the yield of 0.38 g/g (theoretical yield of 75.84 %) and
the productivity of 1.17 g/L/h. Meanwhile, under the same conditions, sterilized
fermentation of KP-OU7-ME5 was also performed (Fig. 6C). Finally, 86.13 g/L 2,3-BD
was synthesized with the yield of 0.39 g/g and the productivity of 1.20 g/L/h under the
sterilized system, which displayed 1.89%, 2.56%, and 2.63% improvements in terms of
titer, productivity, yield, respectively, compared with non-sterilized fermentation.
Klebsiella remained the dominant genus with a relative abundance of 95.36% at the
end of non-sterilized fermentation (Fig. 6B). The slight growth of the contaminated
microbes (relative abundance of 4.64%) most likely came from the starting ingredients
used to prepare the non-sterilized medium. Moreover, the functional genes of the anti-
hybrid bacteria system were located in the chromosome, which decreased the probability
of horizontal gene transfer into environmental microorganisms. On the whole, these results
demonstrated that non-sterilized fermentation of 2,3-BD using K. pneumoniae OU7 is
viable and safe. At present, our system was unable to defend against phage infections,
which is another vexing issues in industrial fermentations. Consequently, sterilization
treatment for fermentation waste, rationally designing factories, and culture rotation are

Page 21 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
conducive to decrease the probability of phage infections. Furthermore, in future work, the
tools of abortive infection, CRISPR-Cas systems, and restriction-modification can be used
to resist phage infections.38–40
Compared with previous reports, the 2,3-BD titer from our work was relatively low
(Table 2). Among them, 178 g/L 2,3-BD was obtained by engineered S. cerevisiae with
the productivity of 1.88 g/L/h and the yield of 0.34 g/g. This is the highest titer of 2,3-BD
produced in a microbial fermentation process so far. However, the substrate costs increase
because of this engineered strain’s requirement for nutrient-rich media (which includes 20
g/L peptone and 10 g/L yeast extract) as well as antibiotics (300 mg/L hygromycin B and
0.5 mg/L aureobasidin A).41 Additionally, it would also be necessary to sterilize the
medium, which would increase the labor- and energy-costs.
Taken together, the non-sterilized fermentation of 2,3-BD presented in our work is
advantageous because: (1) it is more economically because elimination the use of antibiotic
addition or expensive nutrients; (2) labor- and energy-costs can be reduced as the need for
sterilization has been eliminated; (3) the fermentation operation can be simplified; and (4)
it is a benefit for long-lasting continuous fermentation, which often suffers from the risk of
contamination resulting from constant adjustments and requires some sophistication in
terms of instrument operation. In future work, the traditional bioreactor can be replaced by
cement, plastic or ceramic bioreactor, as Chen’s group described to reduce equipment costs

Green Chemistry
Page 22 of 37
View Article Online
DOI: 10.1039/D0GC03044A
further.11 Moreover, we believe that further improvement of 2,3-BD synthesis might be
realized by optimizing the fermentation medium or overexpressing genes of the 2,3-BD
biosynthesis pathway. Thus, the non-sterilized fermentation performed by a robust K.
pneumoniae OU7 is still an excellent alternative for the sustainable synthesis of 2,3-BD.
Conclusions
The green synthesis of 2,3-BD through biotechnological methods is attractive but
difficult to perform because of microbial contamination. In this work, by using the
systematic engineering including the evaluation of the anti-microbial contamination
system, redistribution of the metabolic flux, mutation of the pathogenicity and adaptive
evolution of the fermentation performance, we successfully redesigned the K. pneumoniae
OU7 strain for efficient 2,3-BD production. Ultimately, under the non-sterilized system,
the 2,3-BD titer and yield obtained were 84.53 g/L and 0.38 g/g, respectively. From the
perspective of industrial development, our work is competitive because it avoids antibiotic
use and reduces the energy- and labor-costs. Additionally, our work offers an insight into
the more sustainable synthesis of other chemicals using a similar strategy.
Conflicts of interest
There are no conflicts to declare
Acknowledgments

Page 23 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
This work was supported by the National Key Research and Development Program of
China (number: 2018YFC1603400, 2018YFC1602100); National Natural Science
Foundation of China (number: 21676104, 21878105); Science and Technology Program
of Guangzhou (number: 201904010360); Fundamental Research Funds for the Central
Universities (number: 2019PY15); and the Key Research and Development Program of
Guangdong Province (number: 2019B020213001).
References
1.
J. W. K. Oliver, I. M. P. Machado, H. Yoneda and S. Atsumi, Proc. Natl. Acad.
Sci., 2013, 110, 1249–1254.
2.
J. Feng, Y. Gu, P.-F. Yan, C. Song and Y. Wang, ACS Sustain. Chem. Eng., 2017,
5, 11221–11225.
3.
4.
X.-J. Ji, H. Huang and P.-K. Ouyang, Biotechnol. Adv., 2011, 29, 351–364.
L. Zhang, R. Singh, D. Sivakumar, Z. Guo, J. Li, F. Chen, Y. He, X. Guan, Y. C.
Kang and J.-K. Lee, Green Chem., 2018, 20, 230–242.
J. Haider, G. R. Harvianto, M. A. Qyyum and M. Lee, ACS Sustain. Chem. Eng.,
2018, 6, 14901–14910.
5.
6.
Y. Ge, K. Li, L. Li, C. Gao, L. Zhang, C. Ma and P. Xu, Green Chem., 2016, 18,
4693–4703.

Green Chemistry
Page 24 of 37
View Article Online
DOI: 10.1039/D0GC03044A
7.
8.
9.
P.-F. Yan, J. Feng, S. Dong, M. Wang, I. A. Khan and Y. Wang, ACS Sustain.
Chem. Eng., 2017, 5, 11016–11023.
S. Gopinath, M. V Kim, T. Rakib, P. W. Wong, M. van Zandt, N. A. Barry, T.
Kaisho, A. L. Goodman and A. Iwasaki, Nat. Microbiol., 2018, 3, 611–621.
X. Ou, X. Wu, F. Peng, Y. Zeng, H. Li, P. Xu, G. Chen, Z. Guo, J. Yang, M. Zong
and W. Lou, Biotechnol. Bioeng., 2019, 116, 3333–3348.
10. W. Zeng, W. Li, L. Shu, J. Yi, G. Chen and Z. Liang, Bioresour. Technol., 2013,
142, 697–700.
11. G.-Q. Chen and X.-R. Jiang, Curr. Opin. Biotechnol., 2017, 50, 94–100.
12. H. C. Neu, Science, 1992, 257, 1064–1073.
13. B. Zhao, L. Wang, C. Ma, C. Yang, P. Xu and Y. Ma, Bioresour. Technol., 2010,
101, 6494–6498.
14. Y. Zhang, X. Chen, B. Qi, J. Luo, F. Shen, Y. Su, R. Khan and Y. Wan, Bioresour.
Technol., 2014, 163, 160–166.
15. Z. Xiao, R. Gu, X. Hou, J.-Y. Zhao, H. Zhu and J. Lu, J. Chem. Technol.
Biotechnol., 2017, 92, 2845–2852.
16. L. Li, L. Zhang, K. Li, Y. Wang, C. Gao, B. Han, C. Ma and P. Xu, Biotechnol.
Biofuels, 2013, 6, 123.

Page 25 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
17. L. Li, C. Chen, K. Li, Y. Wang, C. Gao, C. Ma and P. Xu, Appl. Environ.
Microbiol., 2014, 80, 6458–6464.
18. J. O’Hair, Q. Jin, D. Yu, N. Poe, H. Li, S. Thapa, S. Zhou and H. Huang, ACS
Sustain. Chem. Eng., 2020, 8, 11244–11252.
19. A. Shaw, F. Lam, M. Hamilton, A. Consiglio, K. MacEwen, E. Brevnova, E.
Greenhagen, W. LaTouf, C. South, H. Dijken and G. Stephanopoulos, Science,
2016, 353, 583–586.
20. T. T. Selão, A. Wlodarczyk, P. J. Nixon and B. Norling, Metab. Eng., 2019, 54,
255–263.
21. K. Motomura, K. Sano, S. Watanabe, A. Kanbara, A.-H. G. Nasser, T. Ikeda, T.
Ishida, H. Funabashi, A. Kuroda and R. Hirota, ACS Synth. Biol., 2018, 7, 2189–
2198.
22. S. I. González-Morales, N. B. Pacheco-Gutiérrez, C. A. Ramírez-Rodríguez, A. A.
Brito-Bello, P. Estrella-Hernández, L. Herrera-Estrella and D. L. López-
Arredondo, Biotechnol. Biofuels, 2020, 13, 119.
23. Z.-W. Guo, X.-Y. Ou, S. Liang, H.-F. Gao, L.-Y. Zhang, M.-H. Zong and W.-Y.
Lou, ACS Synth. Biol., 2020, 9, 2537–2545.
24. X. Feng, L. Jiang, X. Han, X. Liu, Z. Zhao, H. Liu, M. Xian and G. Zhao, Microb.
Cell Fact., 2017, 16, 209.

Green Chemistry
Page 26 of 37
View Article Online
DOI: 10.1039/D0GC03044A
25. X. Guo, C. Cao, Y. Wang, C. Li, M. Wu, Y. Chen, C. Zhang, H. Pei and D. Xiao,
Biotechnol. Biofuels, 2014, 7, 44.
26. M. A. Pearson, I.-S. Park, R. A. Schaller, L. O. Michel, P. A. Karplus and R. P.
Hausinger, Biochemistry, 2000, 39, 8575–8584.
27. S.-G. Jung, J.-H. Jang, A.-Y. Kim, M.-C. Lim, B. Kim, J. Lee and Y.-R. Kim,
Appl. Microbiol. Biotechnol., 2013, 97, 1997–2007.
28. K. Jantama, P. Polyiam, P. Khunnonkwao, S. Chan, M. Sangproo, K. Khor, S. S.
Jantama and S. Kanchanatawee, Metab. Eng., 2015, 30, 16–26.
29. X.-J. Ji, H. Huang, J. Du, J.-G. Zhu, L.-J. Ren, N. Hu and S. Li, Bioresour.
Technol., 2009, 100, 3410–3414.
30. S. Brisse, F. Grimont and P. A. D. Grimont, eds. M. Dworkin, S. Falkow, E.
Rosenberg, K.-H. Schleifer and E. Stackebrandt, Springer New York, New York,
NY, 2006, pp. 159–196.
31. X.-J. Ji, H. Huang, J. Du, J.-G. Zhu, L.-J. Ren, S. Li and Z.-K. Nie, Bioresour.
Technol., 2009, 100, 5214–5218.
32. T. Yang, Z. Rao, G. Hu, X. Zhang, M. Liu, Y. Dai, M. Xu, Z. Xu and S.-T. Yang,
Biotechnol. Biofuels, 2015, 8, 129.
33. K. Khaertynov, V. Anokhin, A. Rizvanov, Y. Davidyuk, D. Semyenova, S. Lubin
and N. Skvortsova, Front. Med., , DOI:10.3389/fmed.2018.00225.

Page 27 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
34. M. Regué, L. Izquierdo, S. Fresno, N. Piqué, M. M. Corsaro, T. Naldi, C. De
Castro, D. Waidelich, S. Merino and J. M. Tomás, J. Bacteriol., 2005, 187, 4198–
4206.
35. E. Frirdich, E. Vinogradov and C. Whitfield, J. Biol. Chem., 2004, 279, 27928–
27940.
36. L. Izquierdo, N. Coderch, N. Piqué, E. Bedini, M. M. Corsaro, S. Merino, S.
Fresno, J. M. Tomás and M. Regué, J. Bacteriol., 2003, 185, 7213–7221.
37. J. Yuan, Y.-Z. He, Z.-W. Guo, H.-F. Gao, F.-B. Chen, L.-Z. Li, Y.-Y. Li and L.-Y.
Zhang, BioResources, 2017, 12, 4926–4942.
38. R. Sorek, C. M. Lawrence and B. Wiedenheft, Annu. Rev. Biochem., 2013, 82,
237–266.
39. H. Harvey, J. Bondy-Denomy, H. Marquis, K. M. Sztanko, A. R. Davidson and L.
L. Burrows, Nat. Microbiol., 2018, 3, 47–52.
40. S. J. Labrie, J. E. Samson and S. Moineau, Nat. Rev. Microbiol., 2010, 8, 317–327.
41. Y.-G. Lee and J.-H. Seo, Biotechnol. Biofuels, 2019, 12, 204.
42. J. Fu, G. Huo, L. Feng, Y. Mao, Z. Wang, H. Ma, T. Chen and X. Zhao,
Biotechnol. Biofuels, 2016, 9, 90.
43. J.-J. Dai, J.-S. Cheng, Y.-Q. Liang, T. Jiang and Y.-J. Yuan, Bioresour. Technol.,
2014, 167, 433–440.

Green Chemistry
Page 28 of 37
View Article Online
DOI: 10.1039/D0GC03044A
44. L. Zhang, J. Sun, Y. Hao, J. Zhu, J. Chu, D. Wei and Y. Shen, J. Ind. Microbiol.
Biotechnol., 2010, 37, 857–862.
45. M.-Y. Jung, H.-M. Jung, J. Lee and M.-K. Oh, Biotechnol. Biofuels, 2015, 8, 106.
46. X.-J. Ji, L.-G. Liu, M.-Q. Shen, Z.-K. Nie, Y.-J. Tong and H. Huang, Biotechnol.
Bioeng., 2015, 112, 1056–1059.

Page 29 of 37
Green Chemistry
View Article Online
DOI: 10.1039/D0GC03044A
Tables
Table 1. Strains and plasmids used in this study.
Strains, plasmids Description
Strains
Source
KP-OU7
Wild strain of K. pneumoniae OU7, deposited in China Center for Lab collection
Type Culture Collection (CCTCC NO. M2017449)
KP-OU7-ME1
KP-OU7-ME2
KP-OU7-ME3
KP-OU7-ME4
KP-OU7-ME5
E. coli S17-1 λpir
K. pneumoniae OU7 (ΔldhA)
This study
This study
This study
This study
This study
K. pneumoniae OU7 (ΔldhAΔadhE)
K. pneumoniae OU7 (ΔldhAΔadhEΔpta)
K. pneumoniae OU7 (ΔldhAΔadhEΔptaΔwabG)
KP-OU7-ME4 at the 30th transfer during metabolic evolution
Host strain for pRE112, SmR TpR thi-1 pro recA hsdR-M+RP4: 2- Lab collection
Tc:Mu: Km Tn7 λpir
Plasmids
pRE112
Suicide vector, R6k origin, SacB CmR oriV oriT
Suicide vector for ldhA knockout
Suicide vector for adhE knockout
Suicide vector for pta knockout
Lab collection
This study
This study
This study
This study
pRE112-ΔldhA
pRE112-ΔadhE
pRE112-Δpta
pRE112-ΔwabG
Suicide vector for wabG knockout