Novel approach to engineer strains for simultaneous sugar utilization
-
Upload
radhakrishnan -
Category
Documents
-
view
222 -
download
4
Transcript of Novel approach to engineer strains for simultaneous sugar utilization
![Page 1: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/1.jpg)
Author's Accepted Manuscript
Novel Approach to Engineer Strains for Si-multaneous Sugar Utilization
Pratish Gawand, Patrick Hyland, Andrew Ekins,Vincent J.J. Martin, Radhakrishnan Mahadevan
PII: S1096-7176(13)00073-6DOI: http://dx.doi.org/10.1016/j.ymben.2013.08.003Reference: YMBEN813
To appear in: Metabolic Engineering
Cite this article as: Pratish Gawand, Patrick Hyland, Andrew Ekins, Vincent J.J.Martin, Radhakrishnan Mahadevan, Novel Approach to Engineer Strains forSimultaneous Sugar Utilization, Metabolic Engineering, http://dx.doi.org/10.1016/j.ymben.2013.08.003
This is a PDF file of an unedited manuscript that has been accepted forpublication. As a service to our customers we are providing this early version ofthe manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting galley proof before it is published in its final citable form.Please note that during the production process errors may be discovered whichcould affect the content, and all legal disclaimers that apply to the journalpertain.
www.elsevier.com/locate/ymben
![Page 2: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/2.jpg)
1
Novel Approach to Engineer Strains for Simultaneous Sugar Utilization
Pratish Gawand1, Patrick Hyland1, Andrew Ekins2, Vincent J. J. Martin2, Radhakrishnan
Mahadevan1, 3*
1Department of Chemical Engineering and Applied Chemistry, University of Toronto, 200
College Street, Toronto, ON, Canada M5S 3E5
2Department of Biology, Centre for Structural and Functional Genomics, Concordia
University, 7141 Sherbrooke Street West, Montreal, QC, Canada H4B 1R6
3Institute of Biomaterials and Biomedical Engineering, University of Toronto, 164 College
Street, Toronto, ON, Canada M5S 3G9
Corresponding author
Prof. Radhakrishnan Mahadevan
Department of Chemical Engineering and Applied Chemistry, University of Toronto, 200 College Street, Toronto, ON, Canada M5S 3E5 email: [email protected] Ph: 1‐416‐946‐0996 Fax: 1‐416‐978‐8605
Abstract
Use of lignocellulosic biomass as a second generation feedstock in the biofuels industry
is a pressing challenge. Among other difficulties in using lignocellulosic biomass, one
major challenge is the optimal utilization of both 6‐carbon (glucose) and 5‐carbon
(xylose) sugars by industrial microorganisms. Most industrial microorganisms
sequentially utilize glucose over xylose owing to the regulatory phenomenon of carbon
catabolite repression (CCR). Microorganisms that can co‐utilize glucose and xylose are of
![Page 3: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/3.jpg)
2
considerable interest to the biofuels industry due to their ability to simplify the
fermentation processes. However, elimination of CCR in microorganisms is challenging
due to the multiple coordinating mechanisms involved. We report a novel algorithm,
SIMUP, which finds metabolic engineering strategies to force co‐utilization of two
sugars, without targeting the regulatory pathways of CCR. Mutants of Escherichia coli
based on SIMUP algorithm showed predicted growth phenotypes and co‐utilized
glucose and xylose; however, consumed the sugars slower than the wild‐type. Some
solutions identified by the algorithm were based on stoichiometric imbalance and were
not obvious from the metabolic network topology. Furthermore, sequencing studies on
the genes involved in CCR showed that the mechanism for co‐utilization of the sugars
could be different from previously known mechanisms.
Keywords
Strain design, metabolic modeling, glucose‐xylose co‐utilization, bilevel optimization,
carbon catabolite repression, Escherichia coli
Abbreviations
CCR – carbon catabolite repression, PTS – phosphotransferase system, PPP – pentose
phosphate pathway, G6P – glucose‐6‐phosphate, F6P – fructose‐6‐phosphate, GA3P –
glyceraldehydes‐3‐phosphate, PEP – phosphoenolpyruvate, PYR – pyruvate, R5P –
ribose‐5‐phosphate, X5P – xylulose‐5‐phosphate, E4P – erythrose‐4‐phosphate.
![Page 4: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/4.jpg)
3
1. Introduction
Lignocellulosic biomass is being increasingly sought after as an alternative feedstock for
biofuels production due to the associated benefits in sustainability (Jang et al., 2011).
Hydrolysis of lignocellulosic biomass yields two major sugars, glucose and xylose (Hahn‐
Hagerdal et al., 2001), efficient utilization of which is a major technical challenge (Kim et
al., 2010). Industrial microorganisms that can simultaneously utilize glucose and xylose
are valuable, as they can shorten and simplify the fermentation processes (Kim et al.,
2010, Kim et al., 2012, Slininger et al., 2011). However, most industrial organisms such
as E. coli (Görke and Stülke, 2008) and Saccharomyces cerevisiae (Ho et al., 1998),
(Kuyper et al., 2005) show a clear preference for glucose over xylose. Similar sequential
utilization of the sugars also exists in other industrial organisms such as Zymomonas
mobilis (Mohagheghi et al., 2002), Scheffersomyces stipitis (Slininger et al., 2011), and
Pseudomonas putida (Meijnen et al., 2008).
Sequential utilization of sugars results in a characteristic growth pattern with an
intermediate lag‐phase. Such a growth is known as diauxic growth and is controlled by
the regulatory phenomenon called carbon catabolite repression (CCR). CCR causes the
genes encoding catabolic proteins for non‐preferred sugars to be repressed during the
growth on a preferred sugar (Görke and Stülke, 2008, Deutscher, 2008, Rolland et al.,
2002, Bettenbrock et al., 2006). Mechanistic details of CCR have been rigorously studied
for many model organisms including E. coli and S. cerevisiae. In E. coli, CCR is mediated
by the interplay between the transcription activator CRP (cyclic‐AMP receptor protein),
![Page 5: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/5.jpg)
4
the signaling metabolite cAMP, adenylatecyclase, and the component EIIAGlc of the
glucose specific phoshphoenolpyruvate carbohydrate phosphotransferase system (PTS)
(Deutscher et al., 2006). In S. cerevisiae, CCR entails the main glucose repression
pathway, involving MIG1 transcriptional repressor complex, SNF1 protein kinase
complex, and protein phosphatase 1 (Rolland et al., 2002, Gancedo, 1998).
Due to the extensive information available on CCR in E. coli and S. cerevisiae,
engineering of the regulatory process has been previously attempted to generate
mutants for glucose‐xylose co‐utilization. For instance, a mutant devoid of PTS (with
deletion of ptsH, ptsI and crr, encoding HPr, EI and EIIAGlc, respectively) was constructed
to force non‐PTS mediated glucose uptake. This mutant showed diminished glucose
repression on xylose (Hernandéz‐Montalvo et al., 2001). E. coli carrying a mutation in
ptsG (encoding the glucose transporter EIIBCGlc) was found to co‐utilize glucose and
xylose at low sugar concentrations (Nichols et al., 2001). Allosteric alteration of the
global regulatory protein crp (designated as crp* or crp+) is known to confer relaxed CCR
in E. coli (Khankal et al., 2009). Recently, a crp* mutant (with an amino acid substitution
from Gly122 to Ser122) was shown to co‐utilize glucose and xylose (Yao et al., 2011).
Mutations in various metabolic genes are also known to confer co‐utilization phenotype
to E. coli. Deletion of mgsA, for example, has been implicated in improved co‐utilization
of glucose and xylose (Yomano et al., 2009). A minimal E. coli mutant that was
engineered to produce ethanol, co‐utilized glucose and xylose using an unknown
mechanism (Trinh et al., 2008). An alternative strategy to force co‐utilization involves
cultivation of a co‐culture of substrate‐selective E. coli mutants (Eiteman et al., 2009).
![Page 6: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/6.jpg)
5
Though these mutants of E. coli show improved co‐utilization of glucose and xylose,
none of the mutants have shown complete elimination of CCR. Furthermore, similar
approaches for altering the regulatory genes in S. cerevisiae have not been particularly
successful. For instance, deletion of the regulatory gene MIG1 did not alleviate glucose
repression on xylose utilization (Roca et al., 2004). Evolutionary engineering of S.
cerevisiae also did not yield a strain capable of co‐utilizing glucose and xylose (Wouter‐
Wisselink et al., 2009). These examples suggest that the glucose‐xylose co‐utilization
problem has not been entirely solved, even in the most intensively studied organisms.
Due to the natural tendency of the organisms to use the most efficient substrate first,
which is coordinated by multiple interacting regulatory pathways, the problem of
engineering a co‐utilization phenotype is especially challenging. This problem is
compounded further by the lack of knowledge of CCR in many organisms of interest.
Thus, a method that can generate sugar co‐utilizing mutants of industrial organisms,
without any knowledge of their regulatory pathways, is of great interest.
In this study, we present a rational bilevel optimization algorithm, SIMUP, to engineer
the metabolic network of organisms in such a way that the predicted mutants are forced
to co‐utilize two sugars for growth. As a major advantage, the algorithm obviates the
knowledge of underlying regulatory mechanisms of CCR of the target organism. Many
bilevel algorithms have been previously developed to design strains with improved
production of metabolites. Notable among these are: OptKnock (Burgard et al., 2003),
OptGene (Patil et al., 2005), GDLS (Lun et al., 2009), OptORF (Kim and Reed, 2010),
optFORCE (Ranganathan et al., 2010), EMILiO (Yang et al., 2011). In addition, bilevel
![Page 7: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/7.jpg)
6
algorithms have also been used for finding enzyme capacity constraints (Yang et al.,
2008), for selection of the best objective function that explains observed experimental
fluxes (Gianchandani et al., 2008), and for the design of auxotrophy dependent
microbial sensors (Tepper and Shlomi, 2011). Though similar in framework to these
algorithms, SIMUP’s distinguishing feature is the use of multiple external media
conditions in a single optimization problem.
Solutions predicted by SIMUP were verified using E. coli as a test organism. The mutants
constructed based on the model predictions were characterized using phenotypic tests
and batch cultivation studies.
2. Materials and methods
2.1. SIMUP algorithm
SIMUP algorithm identifies strategies for deletion of metabolic reactions to force co‐
utilization of glucose and xylose. The algorithm is based on the fact that lethality of
metabolic gene deletions depends on the external nutrient conditions. For instance,
gene deletions that are lethal with glucose as a carbon source may not be lethal with
xylose as a carbon source. The algorithm uses the metabolic network of an organism to
find deletion combinations (of reactions) that are lethal when either glucose or xylose is
provided, but are not lethal only when both glucose and xylose are provided. As these
reaction deletions do not allow the organism to grow on either glucose or xylose alone,
but allow the organism to grow on a mixture of glucose and xylose, the organism is
forced to consume both the sugars simultaneously. As the algorithm uses only the
![Page 8: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/8.jpg)
7
metabolic network as an input, no knowledge of the underlying regulatory network is
required. The algorithm was formulated as a bilevel optimization problem (Fig. 1) shown
in equation (1).
2 2
,
, 1,2,3
, , 1,2,3
, , 1,2,3
, , 1,2,3 0 0,1
1
1
Where, is the growth of the mutant and is the growth of the wild‐type.
Superscripts correspond to different growth conditions: 1) with both glucose and xylose,
2) with glucose alone, and 3) with xylose alone. The outer function is subjected to the
inner constraints of the three independent flux balance analysis (FBA) (Orth et al., 2010)
problems. is the set of metabolites, and are the sets of reversible and
![Page 9: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/9.jpg)
8
irreversible reactions respectively. Variables and are the flux variables;
and are the lower and upper bounds on the reversible reaction fluxes; and
are the lower and upper bounds on the irreversible reaction fluxes; and are the
stoichiometric matrices for reversible and irreversible reactions; and and are
the coefficient vectors of the objective function. The parameter is the limit on the
number of reactions that can be deleted. Availability of substrates was altered by
changing the bounds on the exchange reactions.
The decision variables for the outer objective function were the binary variables ,
which could delete or retain a reaction by assuming a value of 0 or 1, respectively. This
was achieved by multiplying the binary variables to the corresponding upper and lower
bounds of each reaction as shown in equation (2).
. . , 2
1, 0,
The problem was solved using the strong duality theorem of LP theory by converting the
bilevel problem into a single level mixed integer linear program (MILP). A detailed, step‐
by‐step description of the solution to the problem can be found in the SI Text online.
The MILP was implemented in AMPL (ILOG) and was solved using CPLEX 11.2 (ILOG). To
solve for the bilevel optimization problem, we used a central metabolism model of E.
coli reported in a previous study with additional reactions for xylose metabolism
![Page 10: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/10.jpg)
9
(Schilling et al., 2000). To verify the solutions obtained from SIMUP on a genome‐scale
model, we used iAF1260 model of E. coli metabolism (Feist et al., 2007).
2.2. Feasibility analysis of metabolites
To understand the growth phenotypes observed for some mutants, analysis of feasibility
of formation of each biomass precursor in the metabolic network was performed. The
problem was formulated using an FBA with an objective function to maximize one
biomass precursor at a time. The stoichiometric matrix was modified by adding an
exchange flux for the biomass precursor under consideration. The maximization
problem shown in equation 3 was iteratively used to find the feasibility of formation of
each biomass precursor in the background of reaction deletions of the mutants LMSE2
and LMSE5. Similar method was used in a previous study to find network gaps in
metabolic network of E. coli (Feist et al., 2007).
. 0
3
Where, is additional exchange reaction for the biomass precursor under
consideration, is the stoichiometric matrix, is the flux vector, and are the
upper and lower bounds on the flux variable, and is the number of reactions. All the
constraints on the co‐factors were removed by allowing free exchange and supply.
![Page 11: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/11.jpg)
10
2.3. Strains, media and plasmids
All the strains and plasmids were procured from Coli Genetic Stock Center (CGSC), Yale
University. E. coli K‐12 MG1655 was the wild type control strain and was used to make
the gene knockout mutants ‐ LMSE1 (Δeda,Δpgi, Δgnd::kanR), LMSE2
(Δeda, Δpgi, Δrpe::kanR) and LMSE5 (Δeda, Δpgi, Δfbp, ΔpfkB, ΔpfkA::kanR). The
mutants were constructed using sequential generalized P1 transduction method and the
gene deletions were confirmed by PCR (sequences of primers used are provided in the
Supplementary Table 1 online). For isolation of the final knockouts for the mutants
LMSE2 and LMSE5, in addition to the antibiotic, LB plates were supplemented with
glucose (0.5%) and xylose (0.5%).
2.4. Phenotypic test and batch cultivation
The mutants LMSE1, LMSE2 and LMSE5 were characterized for growth on minimal
medium and complex medium with different carbon sources (2% glucose, 2%xylose, and
1% each of glucose and xylose).
Each strain was then cultivated in 5 L fermenters with 3 L working volume to find the
growth and sugar consumption profiles. The minimal medium used was described in a
previous study (Causey et al., 2003). Glucose and xylose, 5 g/L each were used as carbon
sources. The pH was maintained at 7.0 using 6 M KOH and dissolved oxygen levels were
maintained above 50 % by varying airflow rate and impeller speed. The samples were
analyzed for OD550, and glucose and xylose concentrations. The sugars were separated
using a Bio‐Rad HPX‐87H cation‐exchange column (5 mM H2SO4 mobile phase, 0.5
![Page 12: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/12.jpg)
11
mL/min flow rate, 42o C column temperature, 20 μL injection volume). The study was
carried out in duplicates. Seed cultures were prepared by inoculating a fresh colony
overnight in 3 mL minimal medium. The overnight grown culture was then transferred
to 500 mL Erlenmeyer flasks with 150 mL minimal medium containing glucose and
xylose (5 g/L each). The fermenters were inoculated with appropriate amount of
harvested biomass in exponential growth phase to get an initial OD550 of 0.1.
Additional characterization of the mutant LMSE2 was carried out for sugar utilization for
media with different ratios of glucose and xylose: 1) 3.5 g/L glucose and 7.0 g/L xylose,
and 2) 3.5 g/L glucose and 7.0 g/L xylose.
2.5. Gene sequencing
Genes that play a role in carbon catabolite repression and xylose uptake were
sequenced using Sanger sequencing. The list of genes sequenced, the regions
sequenced, and the primers used for sequencing are provided in the Supplementary
Table 2 online. The genes were PCR amplified from the wild‐type E. coli and the mutant
LMSE2. The PCR products were purified using Fermetas PCR purification kit and were
sent for sequencing at the The Sanger Sequencing Facility at The Center for Applied
Genomics (TCAG), Hospital for Sick Children, Toronto. Alignment of the sequences was
carried out using Geneious 5.5.6 (Biomatters Ltd.).
2.6. Expression of ptsG*
The mutated ptsG gene (ptsG*) was amplified from the mutant LMSE2 using primers
with restriction sites for EcoR1 (Forward: 5’‐
![Page 13: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/13.jpg)
12
AGTCATTGAATTCCGTTGAATGAGTTTTTTTAAAGC‐3’) and HindIII (Reverse:5’‐
GGATCGTAAGCTTTTAGTGGTTACGGATGTACTCA‐3’), and ligated into pBT4 (Lynch and
Gill, 2006) (Addgene plasmid number 22831). The mutation was confirmed using Sanger
sequencing before ligation into the vector. The constructed vector was transformed into
an E. coli ΔptsG mutant and the strain was designated as E. coli ptsG*. Both E. coli ΔptsG
and E. coli ptsG* were characterized in fermenters for glucose‐xylose consumption
profiles using media composition (5 g/L glucose and 5 g/L xylose) and fermentation
conditions mentioned above.
3. Results
3.1. Solution to SIMUP algorithm
Lethality in the mutants with metabolic gene deletions can be explained by the inability
of these mutants to synthesize one or more biomass precursors essential for growth.
Such auxotrophic lethal mutants can be rescued by providing the missing biomass
precursor(s) or metabolite(s) that can be converted to the missing precursor(s) (Suthers
et al., 2009). The solutions predicted by SIMUP can be considered as auxotrophies that
can be rescued only with a combination of sugars. These solutions allow only a subset of
the growth precursors to be synthesized from one sugar. The complete set of biomass
precursors can be synthesized only if both the sugars are consumed simultaneously.
Using this framework, the solutions found by the algorithm were grouped together
based on the unique set of biomass precursors that could be synthesized from each
sugar. A group of equivalent solutions was called a strategy.
![Page 14: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/14.jpg)
13
By setting the maximum gene deletion limit to three and four, a total of eleven solutions
were found, which were grouped into three strategies, A, B, and C. Three representative
mutants were constructed based on one solution from each strategy for experimental
validation. Though the strategies A and B could be explained from the connectivity of
the metabolic network, the strategy C could not be explained solely from the
connectivity. Additional analysis, involving assessment of the feasibility of formation of
each biomass precursor from the given substrates (section 2.2), was carried out to
explain the strategy C.
3.1.1. Strategy A and mutant LMSE1
Strategy A consisted of three solutions which were grouped together based on the
unique set of biomass precursors that could be synthesized from glucose and xylose. All
the solutions allowed the synthesis of only glucose‐6‐phosphate (G6P) from glucose, and
all the remaining precursors were synthesized from xylose. One of the solutions, with
three reaction deletions, pgi, gnd and eda, was used to construct the mutant LMSE1,
with three corresponding gene deletions, pgi, gnd and eda. The solutions belonging to
strategy A and the metabolic network for the mutant LMSE1 are shown in Figure 2a.
Deletion of pgi, encoding phosphoglucose isomerase, restricted the entry of glucose into
glycolysis and conversion of xylose to G6P using reductive pentose phosphate pathway
(PPP) (conversion to fructose‐6‐phosphate) and reverse activity of phosphoglucose
isomerase(Fig. 2a). Deletion of gnd, encoding 6‐phosphogluconate dehydrogenase,
prevented glucose from entering pentose phosphate pathway and no intermediate from
![Page 15: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/15.jpg)
14
PPP could be used to synthesize glucose. Finally, deletion of eda, encoding the second
enzyme of Entner‐Doudoroff pathway, ensured that glucose could not be used to
synthesize glyceraldehyde‐3‐phosphate (G3P), which could otherwise be channeled into
glycolysis. By deleting these three genes, glucose was confined to synthesis of G6P, and
could not be used in the downstream metabolic reactions. All the other biomass
precursors were provided by xylose alone, which entered PPP as xylulose‐5‐phosphate
(X5P) (Fig. 2a). The mutant was thus predicted to grow only when both glucose and
xylose were provided and not on glucose or xylose individually.
The gene deletion combination of LMSE1 was also tested on the genome scale model of
E. coli iAF1260 (Feist et al., 2007), which predicted non‐zero growth for the mutant
LMSE1 on xylose alone. This discrepancy was attributed to the fact that glycogen was
not included as a biomass precursor in the model iAF1260, thereby rendering G6P
synthesis non‐essential for growth. However, since G6P has been customarily
considered a biomass precursor and is one of the hub metabolites (Ma and Zeng, 2003),
we decided to verify the model predictions by constructing the mutant LMSE1.
The mutant LMSE1 was first phenotypically tested in the complex and minimal medium,
followed by batch cultivation to determine its sugar utilization characteristics. The sugar
consumption and growth profiles of the wild type and the mutant LMSE1 are shown in
Figure 3a and3b, respectively. The phenotypic characterization of the mutant showed
that it could grow on minimal medium with xylose as the sole carbon source (in
agreement with iAF1260 predictions). Also, the mutant did not co‐utilize glucose and
![Page 16: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/16.jpg)
15
xylose, but consumed xylose alone (Fig. 3b). These results indicated that either G6P is
not a biomass precursor or the mutant could synthesize G6P from xylose.
In the model iAF1260, G6P is not channeled into any important component except
glycogen, and since glycogen is not a biomass precursor, it was likely that the mutant
LMSE1 did not synthesize G6P. Based on these results, the central metabolism model
was updated by removing G6P as a biomass precursor. This new biomass objective
function was used for all further simulations to solve SIMUP and to identify strategy B
and C mutants.
3.1.2. Strategy B and mutant LMSE2
The strategy B was a group of five solutions which allowed synthesis of only ribose‐5‐
phosphate (R5P) from glucose and only xylulose‐5‐phosphate (X5P) from xylose. The
mutant LMSE2, with three gene deletions, pgi, rpe and eda, each corresponding to a
single reaction, was constructed to verify the strategy B. The solutions belonging to the
strategy B and the metabolic network for the mutant LMSE2 are shown in Figure 2b.
The genes pgi and eda were common to the mutant LMSE1 and served the same
purpose of restricting the entry of glucose into glycolysis and xylose to G6P (Fig. 2b).
Glucose could not be used to synthesize downstream metabolites beyond R5P in PPP.
Deletion of rpe prevented xylose from entering PPP, and the only metabolite that could
be synthesized from xylose was X5P. In presence of both glucose and xylose, R5P (from
glucose) and X5P (from xylose) could be used to synthesize all the remaining PPP
![Page 17: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/17.jpg)
16
metabolites, which then could be used to synthesize the remaining biomass precursors.
The solution was verified on the model iAF1260. The equivalent strategy needed
deletion of six reactions (The strategies based on genome scale models are provided in
Supplementary Table 3 online).
For the mutant LMSE2, deletion of only three genes was required to confer the
expected phenotype. Additional deletions, based on iAF1260, were not needed,
indicating that the additional reactions could not support growth, most likely due to a
limited flux–carrying capacity. The characterization results of the mutant LMSE2 are
shown in Figure3c and Table 1. The mutant showed model predicted phenotype and did
not grow on either glucose or xylose minimal medium. However, it grew well on
minimal medium containing both glucose and xylose. When cultivated in a batch, it
showed simultaneous utilization of glucose and xylose, thereby confirming SIMUP
prediction.
The trend of the sugar uptake rates in the mutant LMSE2 showed that xylose was
consumed at a higher rate than glucose (Table 1). Even though the stoichiometry
suggests that glucose and xylose are required in almost equal amounts (for synthesis of
R5P and X5P, respectively) in the mutant LMSE2, the excess xylose utilization rate can be
explained by the fact that additional xylose could be utilized using the reaction tkt2 (Fig
2b). Using the reaction tkt2, X5P and erythrose‐4‐phosphate (E4P) could react to give
fructose‐6‐phosphate (F6P) and glyceraldehydes‐3‐phosphate (G3P) which were
channeled into glycolysis, thereby, resulting in a higher xylose to glucose ratio.
![Page 18: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/18.jpg)
17
It is also clear from the simulation results that the deletions of the genes in the mutants
LMSE2 slightly decreased the biomass yields per mmol C when compared to the wild
type (Table 1). Accordingly, experimentally observed values for biomass yields were low
for the mutant LMSE2. Thus, the co‐utilization phenotype in the mutant LMSE2, which
required rerouting of the fluxes in the metabolic network, came at the cost of decreased
efficiency of the metabolic network.
3.1.3. Strategy C and mutant LMSE5
The strategy C consisted of three solutions, two with three reaction deletions and one
with four reaction deletions. The solutions in the strategy C could not be explained from
the network connectivity like the strategies A and B. The mutant LMSE5 was constructed
based on the solution with four reaction deletions, pgi, gnd, fbp, and pfk. Due to the
presence of two isozymes, pfkA and pfkB, the mutant LMSE5 had five gene deletions
(pgi, gnd, fbp, pfkA, and pfkB). The solutions in the strategy C and the related metabolic
network for the mutant LMSE5 are shown in Figure 2c.
With the deletion of the four reactions, glucose could not be used to synthesize
fructose‐6‐phosphate (F6P). Furthermore, gluconeogenic flux and entry of glucose into
PPP were stopped. The gene deletions hence were lethal in the glucose minimal
medium. Xylose, however, could enter through PPP as X5P, followed by conversion to
R5P (using the reactions rpi and rpe) (Fig. 2c). With X5P and R5P synthesized from
xylose, all the remaining metabolites of PPP could possibly be synthesized and
channeled into glycolysis. Thus, considering the connectivity of the metabolic network,
![Page 19: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/19.jpg)
18
it appears that the strain would grow on xylose minimal medium. However, SIMUP
predicted that the deletion of the four reactions would be lethal on xylose. An
equivalent strategy on the genome scale model iAF1260 needed deletion of seven
reactions (Supplementary Table 3 online). As the metabolic network was fully
connected with respect to xylose, lethality on xylose in the strategy C was most likely
due to stoichiometric constraints.
On construction of the mutant LMSE5, it was found that the mutant showed no growth
on glucose, limited growth on xylose (Supplementary Fig. 1), and significant growth only
when both glucose and xylose were provided (Fig. 3d). The mutant co‐utilized glucose
and xylose in batch cultivation, in agreement with the SIMUP predictions.
Similar to the mutant LMSE2, the predicted growth yield of the mutant LMSE5 was
lower than the wild type, and reduced growth yield was observed even in experimental
results (Table 1). Additionally, the sugar uptake rates of the mutant LMSE5 were
different from the mutant LMSE2 due to different stoichiometric requirements. As
glucose supported the synthesis of most of the biomass precursors in the mutant
LMSE5, and xylose was required mainly for synthesis of F6P (Fig 2c), glucose was
consumed at much a higher rate than xylose.
To investigate the stoichiometric constraints of the mutant LMSE5, further analysis of
the metabolic network was carried out.
![Page 20: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/20.jpg)
19
3.2. Feasibility analysis of biomass precursors
Lethality caused due to metabolic gene deletions can be explained by the mutants’
inability to synthesize the complete set of biomass precursors (Suthers et al., 2009). To
identify the cause of lethality of the mutant LMSE5 (on xylose), the feasibility of
formation of its biomass precursors was analyzed under three different substrate
conditions: glucose alone, xylose alone, and glucose and xylose. Similar analysis was also
carried out for the mutant LMSE2.
Figure 4 shows the results for the feasibility analysis for the mutants LMSE2 and LMSE5.
As expected, the mutant LMSE2 could not synthesize any precursor except R5P from
glucose and X5P (not a biomass precursor) from xylose (Fig. 4a). However, in presence
of both glucose and xylose, all the precursors could be synthesized and hence growth of
the mutant was possible. For the mutant LMSE5, growth on glucose was not possible
because R5P and erythrose‐4‐phosphate (E4P) could not be synthesized from glucose
(as deletion of gnd prevented glucose from entering PPP) (Fig. 4b). The only metabolites
that were feasible from xylose were R5P and F6P. No other metabolites, such as G3P or
E4P, could be synthesized, despite the existing reactions that connected these
metabolites to xylose. To identify whether any stoichiometric imbalance was causing the
lethality on xylose, we iteratively added secretion reaction for each biomass precursor
to the metabolic network of the mutant LMSE5 and observed its effect on the growth on
xylose.
![Page 21: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/21.jpg)
20
Figure 4c shows the feasibilities of biomass precursors in the mutant LMSE5 after
addition of a secretion reaction for F6P. Allowing secretion of F6P enabled the mutant
LMSE5 to grow on xylose alone, suggesting that the cause of reduced growth of the
mutant LMSE5 on xylose was intracellular accumulation of F6P. The mutant was most
likely forced to synthesize excess F6P to meet the imposed stoichiometric constraints
due to gene deletions. As sugar phosphates cannot be secreted out of the cell, a likely
accumulation of F6P resulted in the limited growth phenotype of the mutant LMSE5 on
xylose.
Thus, the co‐utilization phenotype in the mutant LMSE5 can be attributed to
stoichiometric imbalance in addition to the topological constraints of the metabolic
network observed in the mutant LMSE2.
3.3. Characterization of the mutant LMSE2 on different ratios of glucose and xylose
The mutant LMSE2 showed co‐utilization of glucose and xylose when the two sugars
were provided in equal concentrations. However, it was important to characterize the
mutant on different concentrations of glucose and xylose as lignocellulosic hydrolysate
contains unequal concentrations of glucose and xylose (Hahn‐Hagerdal et al., 2001). We
characterized the mutant LMSE2 on two different sugar ratios, 1:2 glucose:xylose and
2:1 glucose:xylose. The growth characteristics obtained from the characterization
studies are summarized in Figure 5 and Table 1. The mutant LMSE2 grew slower in both
1:2 glucose:xylose and 2:1 glucose:xylose media as compared to the medium containing
equal concentrations of glucose and xylose. In addition, glucose and xylose uptake rates
![Page 22: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/22.jpg)
21
were also found to be lower in media containing unequal concentrations of glucose and
xylose. From the values of glucose and xylose utilization rates, it is apparent that growth
of the mutant LMSE2 was controlled by xylose metabolism and excess glucose did not
improve the growth rate of the mutant (Table 1). The growth yield, however, was found
to be better when the mutant LMSE2 grew on different sugar mixtures, most likely due
to slower metabolism and no production of acetate. In addition, excess glucose and
xylose were found to accumulate in both batches with unequal glucose and xylose
concentrations. These results indicate that further engineering of the mutant LMSE2 is
required to adjust to different sugar ratios. Interestingly, with different concentrations
of sugars, the mutant LMSE2 changed the ratios of sugar uptake rates suggesting that
the mutant does not need to utilize sugars in strictly equal proportion. This result
indicates a possibility of further engineering to channel the excess sugar consumed into
useful products.
4. Investigation of the mechanism of CCR‐elimination in the mutant LMSE2
To identify whether the molecular mechanism of CCR‐elimination in our co‐utilizing
strain LMSE2 was similar to the previously reported mutants (Hernandéz‐Montalvo et
al., 2001, Nichols et al., 2001, Khankal et al., 2009, Yao et al., 2011, Yomano et al., 2009,
Trinh et al., 2008), where the key genes implicated in CCR were mutated, we sequenced
eleven genes (and their relevant upstream and downstream regions) involved in CCR in
E. coli (Supplementary Table 2). The only gene that was found to carry a mutation was
ptsG, which encodes the glucose transporter EIIBCGlc. A base‐pair substitution from T to
![Page 23: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/23.jpg)
22
G at position 89 within the gene was identified, which resulted in the amino acid
substitution from Lys30 to Arg30. A sequence comparison of the wild‐type and LMSE2
ptsG genes is provided in the Supplementary Data online. The protein EIIBCGlc consists
of two domains: a hydrophobic transmembrane EIIC domain and a hydrophilic EIIB
domain, which are connected by a conserved linker (Jeckelmann et al., 2011). The
identified mutation was present in the transmembrane EIIC domain of the protein.
Mutations in this region have been associated with non‐specific utilization of ribose,
mannitol and fructose (Siebold et al., 2001). Mutations in ptsG are also known to induce
co‐utilization of glucose and xylose (Nichols et al., 2001).
To characterize the identified mutation in ptsG, the mutated gene (ptsG*) was expressed
in the wild‐type E. coli devoid of ptsG (E. coli ΔptsG). Characterization of the two
mutants (E. coli ΔptsG and E. coli ptsG*) indicated that the mutation in ptsG* was not
responsible for the elimination of CCR in the mutant LMSE2. This conclusion could be
drawn as glucose still repressed xylose‐utilization in the E. coli ptsG* mutant (Fig. 6).
Thus, no CCR gene was found to carry any significant mutation that would eliminate CCR
in the mutant LMSE2. This result is in contrast to the earlier attempts to eliminate CCR,
which involved gene deletions of critical CCR genes or employed mutated versions of
CCR genes. These results suggest involvement of novel genes or mechanisms in the
observed co‐utilization phenotype of the mutant LMSE2.
![Page 24: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/24.jpg)
23
5. Discussion
Essentiality of metabolic genes is known to be dependent on the external nutrient
conditions (Harrison et al., 2007). SIMUP algorithm extends this notion to find the gene
deletion combinations that force co‐utilization of glucose and xylose. Two of the three
mutants constructed in this study showed the desired co‐utilization phenotype.
Whereas the mutant LMSE2 showed strict co‐utilization of the two sugars, the mutant
LMSE5 exemplified a metabolic engineering strategy that was not obvious from
metabolic network topology. Under different concentrations of sugars the mutant
LMSE2 consumed sugar at different rates suggesting a possibility of channeling the
excess sugar flux towards products. Supplementary Table 4 online compares the
important properties of the mutants constructed in this study with the previously
reported co‐utilizing mutants.
One of the limitations of the mutants constructed in the study was that the mutants
consumed glucose and xylose at a lower rate than the wild‐type, thereby increasing the
batch times of the fermentation process. Thus, the co‐utilization phenotype observed in
the mutants came at the cost of decreased productivity of any potential products. This
limitation of the mutants constructed can be attributed to the steady‐state metabolic
modeling approach that was used to design the mutants. The steady‐state metabolic
models cannot simulate the dynamic nature of the metabolism, and thus cannot be
used to predict or optimize the dynamic growth of the metabolic mutants (Song and
![Page 25: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/25.jpg)
24
Ramkrishna, 2012). To improve the fermentation productivities along with sugar co‐
utilization phenotypes, more detailed dynamic models of metabolism need to be
implemented (Song and Ramkrishna, 2012, Anesiadis et al., 2008).
The most important feature of the mutants constructed in this study was that the co‐
utilization phenotype in these mutants was achieved without any prior knowledge of
the regulatory mechanisms involved in CCR. As the gene deletions identified in the study
force the mutants to co‐utilize glucose and xylose, the mutants are anticipated to
accommodate the change by accumulating mutations in their regulatory genes. Thus,
this study can be useful to find novel gene targets by which mutants can co‐utilize
glucose and xylose. This approach is especially important when the targeted mutations
in the regulatory pathways are unsuccessful in forcing the co‐utilization phenotype.
To identify the molecular mechanism used by the mutant LMSE2, the genes involved in
CCR in E. coli were sequenced. Two major mechanisms are implicated in CCR in E. coli: a
global regulatory mechanism (involving cAMP‐CRP complex), and operon‐specific
inducer exclusion mechanisms (Görke and Stülke, 2008). A key participant in CCR of E.
coli is the protein EIIAGlc encoded by crr. For an unrepressed uptake of secondary sugars,
it is important that the protein EIIAGlc remains in its phosphorylated state
(Supplementary Fig. 2), which can be sustained by deactivating the subsequent
phosphate acceptor: EIIBCGlc (encoded by ptsG). Mutations in ptsG have been previously
reported to induce co‐utilization of glucose and other secondary sugars such as ribose,
fructose and xylose (Siebold et al., 2001). Characterization of the mutated ptsG in the
![Page 26: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/26.jpg)
25
mutant LMSE2 showed that the identified mutation was silent and was not responsible
for the co‐utilization phenotype. In addition, no other mutation in the known CCR genes
was identified. The mutant LMSE2 thus co‐utilized glucose and xylose most likely by
accumulating mutations in non‐CCR genes exerting a pleotropic effect on sugar
utilization, or by varying its gene expression levels without any permanent mutations.
Another explanation for the co‐utilization phenotype could be changes in the metabolite
concentrations that could allosterically regulate the proteins involved in CCR. For
example, phosphorylation state of the protein EIIAGlc depends on the concentration
ratio of pyruvate and phopshpoenolpyruvate (PEP) (Görke and Stülke, 2008). The
possibility of co‐utilization of glucose and xylose, either by varying gene expression
levels or by changing concentrations of metabolites, is especially likely since the mutant
was not adaptively evolved. Both these phenomena provide interesting possibilities to
investigate the mechanism involved in co‐utilization of glucose and xylose. In the future,
we expect to perform genome re‐sequencing, transcriptomic and metabolomic studies,
that can potentially distinguish the mechanism underlying the co‐utilization phenotype.
Finally, as the gene deletion strategies identified by SIMUP strongly couple sugar co‐
utilization to growth, by not allowing growth on individual sugars, it is ensured that the
mutants do not revert back to their original CCR phenotype. To the best of our
knowledge the mutants LMSE2 and LMSE5 are the only mutants that have growth‐
coupled co‐utilization of glucose and xylose, thereby making them the best candidates
for adaptive evolution.
![Page 27: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/27.jpg)
26
6. Conclusions
We developed a rational algorithm that uses conditional synthetic lethality to force the
co‐utilization of two sugars. We demonstrated a successful use of the algorithm by
constructing E. coli mutants capable of co‐utilizing glucose and xylose; however, the
mutants consumed sugars at lower rate than the wild‐type. No previous knowledge of
the regulatory network was needed to construct these mutants. We believe that this
novel approach is not limited to E. coli and can be conveniently extended to other
industrially relevant organisms with genome‐scale metabolic models and other sugar
combinations.
![Page 28: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/28.jpg)
27
Table 1. Growth characteristics of the mutants LMSE2 and LMSE5.
Mutant Sugar ratio
μexp
(h‐1)
qglc(mmol/
g dcw/h)
qxyl (mmol/ g dcw/h)
Yx/s (g/mmol C)
μFBA (h‐1)
YX/S (predicted)(g/mmol C)
Wild type 1G:1X 0.60 (glc)0.28 (xyl)
9.3 6.53 0.011 (glc)0.008 (xyl)
0.86 0.49
0.0150.015
LMSE2 1G:1X 0.38 3.89 4.91 0.008 0.67 0.014
LMSE2 2G:1X 0.26 2.17 2.48 0.010 0.32 0.012
LMSE2 1G:2X 0.34 2.76 3.78 0.010 0.51 0.014
LMSE5 1G:1X 0.20 3.86 1.12 0.007 0.36 0.012
Growth rates and biomass yields were calculated using genome‐scale metabolic model iAF12610 of E. coli. Experimentally obtained substrate uptake rates were used for growth rate calculations.
![Page 29: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/29.jpg)
28
Figure legends
Figure 1. The schematic shows the logic behinds the SIMUP algorithm. The algorithm
finds an optimum strategy that allows growth only when both glucose and xylose are
provided together and does not allow growth on either glucose or xylose alone. This
strategy causes the organism to co‐utilize glucose and xylose. The problem was
formulated as a bilevel optimization problem. The outer objective function maximized
the growth on both the substrates and penalized the growth on either one substrate, by
deleting the reactions in the metabolic network. The inner optimization problems find
the maximum possible growth under each substrate condition while conforming to the
reaction deletions by the outer optimization problem.
Figure 2. The solutions predicted by the SIMUP algorithm. Based on how each solution
splits the metabolic network, they have been divided into three strategies: a) Strategy A
in which glucose is used to synthesize glucose‐6‐phosphate and all other metabolites are
synthesized from xylose. Among the three gene combinations shown, the highlighted
(red) solution was used to construct mutant LMSE1. The metabolic map of the mutant
LMSE1 is shown with equivalent solutions. b) Strategy B has five solutions. The
highlighted solution was used to construct the mutant LMSE2. Glucose was used to
synthesize ribose‐5‐phosphate and xylose was used to synthesize xylulose‐5‐phosphate,
and using these two metabolites all the remaining biomass precursors were synthesized.
c) Strategy C contains three solutions. The four gene deletion strategy highlighted was
![Page 30: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/30.jpg)
29
used to construct mutant LMSE5. The strategy has non‐intuitive solutions as xylose was
connected to all the metabolites. However, the strain was capable of growth only when
glucose and xylose both were provided.
Figure 3. Results for batch cultivation studies and phenotypic results for the wild type
and the mutants. The phenotypic test results are shown as insets in the plots ( ‐ no‐
growth/ lethal phenotype; ‐ growth/non‐lethal phenotype; ‐ limited
growth/synthetic‐sick phenotype). The batch cultivation studies show time profiles of
dry cell weight (•), glucose (♦), and xylose (▲). The abbreviations used are LB – Luria‐
Bertani broth; MM – minimal medium; G – glucose; X – xylose; GX – glucose + xylose; NS
– no sugars. a) Batch cultivation results for wild type E. coli. b) Characterization results
for LMSE1. c) Characterization results for LMSE2. d) Characterization results for LMSE5.
Figure 4. Stoichiometric feasibilities for the mutant strains LMSE2 and LMSE5. G –
glucose as carbon source; X – xylose as carbon source; GX – glucose and xylose as
carbon source. a) Feasibilities of biomass precursors for the mutant LMSE2. b)
Feasibilities of biomass precursors for the mutant LMSE5. c) Feasibilities for all the
biomass precursors with additional secretion reaction for F6P in the metabolic network
of the mutant LMSE5. F6P – fructose‐6‐phosphate; GA3P – glyceraldehydes‐3‐
phosphate; 3PG – 3‐phosphoglycerate; PEP – phosphoenolpyruvate; PYR – pyruvate;
R5P – ribose‐5‐phosphate; E4P – erythrose‐4‐phosphate; OA – oxaloacetate; AKG – α‐
ketoglutarate.
![Page 31: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/31.jpg)
30
Figure 5. Results for batch cultivation studies for the LMSE2 mutant on unequal
concentration of glucose and xylose. The batch cultivation plots show time profiles of
dry cell weight (•), glucose (♦), and xylose (▲). a) batch cultivation results on 7 g/L
xylose and 3.5 g/L glucose, b) batch cultivation results on 3.5 g/L xylose and 7 g/L
glucose.
Figure 6. Results for batch cultivation studies for E. coli ΔptsG and E. coli ptsG* mutants.
The batch cultivation studies show time profiles of dry cell weight (•), glucose (♦), and
xylose (▲). a) characterization results for E. coli ΔptsG, b) characterization results for E.
coli ptsG*.
Acknowledgements
We thank Laurence Yang (University of Toronto) for the insightful discussions on the
computational work, and Echo Da Zhang and Saba Khan (University of Toronto) for their
help in experiments. This research was funded by ABIP Canada and NSERC
Bioconversion Network.
![Page 32: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/32.jpg)
31
References
Anesiadis, N., Cluett, W.R., Mahadevan, R., 2008. Dynamic metabolic engineering for increasing bioprocess productivity. Metab. Eng. 10, 255‐266.
Bettenbrock, K., Fischer, S., Kremling, A., Jahreis, K., Sauter, T., Gilles, E.D., 2006. A quantitative approach to catabolite repression in Escherichia coli. J. Biol. Chem. 281, 2578‐2584.
Burgard, A.P., Pharkya, P., Maranas, C.D., 2003. Optknock: a bilevel programming framework for identifying gene knockout strategies for microbial strain optimization. Biotechnol. Bioeng. 84, 647‐657.
Causey, T.B., Zhou, S., Shanmugam, K.T., Ingram, L.O., 2003. Engineering the metabolism of Escherichia coli W3110 for the conversion of sugar to redox‐neutral and oxidized products: Homoacetate production. Proc. Natl. Acad. Sci. U. S. A. 100, 825‐832.
Deutscher, J., 2008. The mechanisms of carbon catabolite repression in bacteria. Curr. Opin. Microbiol. 11, 87‐93.
Deutscher, J., Francke, C., Postma, P.W., 2006. How phosphotransferase system‐related protein phosphorylation regulates carbohydrate metabolism in bacteria. Microbiol. Mol. Biol. Rev. 70, 555‐1031.
Eiteman, M.A., Lee, S.A., Altman, R., Altman, E., 2009. A substrate‐selective co‐fermentation strategy with Escherichia coli produces lactate by simultaneously consuming xylose and glucose. Biotechnol. Bioeng. 102, 822‐827.
Feist, A.M., Henry, C.S., Reed, J.L., Krummenacker, M., Joyce, A.R., Karp, P.D., Broadbelt, L.J., Hatzimanikatis, V., Palsson, B.O., 2007. A genome‐scale metabolic reconstruction for Escherichia coli K‐12 MG1655 that accounts for 1260 ORFs and thermodynamic information. Mol. Syst. Biol. 3, 121.
Gancedo, J.M., 1998. Yeast carbon catabolite repression. Microbiol. Mol. Biol. Rev. 62, 334‐361.
Gianchandani, E.P., Oberhardt, M.A., Burgard, A.P., Maranas, C.D., Papin, J.A., 2008. Predicting biological system objectives de novo from internal state measurements. BMC Bioinformatics 9.
Görke, B., Stülke, J., 2008. Carbon catabolite repression in bacteria: Many ways to make the most out of nutrients. Nat. Rev. Microbiol. 6, 613‐624.
![Page 33: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/33.jpg)
32
Hahn‐Hagerdal, B., Wahlbom, C.F., Gardonyi, M., van Zyl, W.H., Cordero Otero, R.R., Jonsson, L.J., 2001. Metabolic engineering of Saccharomyces cerevisiae for xylose utilization. Adv. Biochem. Eng. Biotechnol. 73, 53‐84.
Harrison, R., Papp, B., Pál, C., Oliver, S.G., Delneri, D., 2007. Plasticity of genetic interactions in metabolic networks of yeast. Proc. Natl. Acad. Sci. U. S. A. 104, 2307‐2312.
Hernandéz‐Montalvo, V., Valle, F., Bolivar, F., Gosset, G., 2001. Characterization of sugar mixtures utilization by an Escherichia coli mutant devoid of the phosphotransferase system. Appl. Microbiol. Biotechnol. 57, 186‐191.
Ho, N., Chen, Z., Brainard, A., 1998. Genetically engineered Saccharomyces yeast capable of effective cofermentation of glucose and xylose. Appl. Environ. Microbiol. 64, 1852‐1859.
Jang, Y.S., Park, J.M., Choi, S., Choi, Y.J., Seung, D.Y., Cho, J.H., Lee, S.Y., 2011. Engineering of microorganisms for the production of biofuels and perspectives based on systems metabolic engineering approaches. Biotechnol. Adv. .
Jeckelmann, J.M., Harder, D., Mari, S.A., Meury, M., Ucurum, Z., Müller, D.J., Erni, B., Fotiadis, D., 2011. Structure and function of the glucose PTS transporter from Escherichia coli. J. Struct. Biol. 176, 395‐403.
Khankal, R., Chin, J.W., Ghosh, D., Cirino, P.C., 2009. Transcriptional effects of CRP* expression in Escherichia coli. Journal of Biological Engineering 3, 13.
Kim, J.H., Block, D.E., Mills, D.A., 2010. Simultaneous consumption of pentose and hexose sugars: An optimal microbial phenotype for efficient fermentation of lignocellulosic biomass. Appl. Microbiol. Biotechnol. 88, 1077‐1085.
Kim, J., Reed, J.L., 2010. OptORF: Optimal metabolic and regulatory perturbations for metabolic engineering of microbial strains. BMC Syst Biol 4, 53‐53.
Kim, S.R., Ha, S.J., Wei, N., Oh, E.J., Jin, Y.‐., 2012. Simultaneous co‐fermentation of mixed sugars: a promising strategy for producing cellulosic ethanol. Trends Biotechnol. .
Kuyper, M., Toirkens, M.J., Diderich, J.A., Winkler, A.A., Van Dijken, J.P., Pronk, J.T., 2005. Evolutionary engineering of mixed‐sugar utilization by a xylose‐fermenting Saccharomyces cerevisiae strain. FEMS Yeast Res. 5, 925‐934.
Lun, D.S., Rockwell, G., Guido, N.J., Baym, M., Kelner, J.A., Berger, B., Galagan, J.E., Church, G.M., 2009. Large‐scale identification of genetic design strategies using local search. Mol. Syst. Biol. 5, 296.
![Page 34: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/34.jpg)
33
Lynch, M.D., Gill, R.T., 2006. Broad host range vectors for stable genomic library construction. Biotechnol. Bioeng. 94, 151‐158.
Ma, H., Zeng, A.P., 2003. Reconstruction of metabolic networks from genome data and analysis of their global structure for various organisms. Bioinformatics 19, 270‐277.
Meijnen, J.P., De Winde, J.H., Ruijssenaars, H.J., 2008. Engineering Pseudomonas putida S12 for efficient utilization of D‐xylose and L‐arabinose. Appl. Environ. Microbiol. 74, 5031‐5037.
Mohagheghi, A., Evans, K., Chou, Y.C., Zhang, M., 2002. Cofermentation of glucose, xylose, and arabinose by genomic dna‐integrated xylose/arabinose fermenting strain of Zymomonas mobilis AX101. Applied Biochemistry and Biotechnology ‐ Part A Enzyme Engineering and Biotechnology 98‐100, 885‐898.
Nichols, N.N., Dien, B.S., Bothast, R.J., 2001. Use of catabolite repression mutants for fermentation of sugar mixtures to ethanol. Appl. Microbiol. Biotechnol. 56, 120‐125.
Orth, J.D., Thiele, I., Palsson, B.O., 2010. What is flux balance analysis? Nat. Biotechnol. 28, 245‐248.
Patil, K.R., Rocha, I., Forster, J., Nielsen, J., 2005. Evolutionary programming as a platform for in silico metabolic engineering. BMC Bioinformatics 6, 308.
Ranganathan, S., Suthers, P.F., Maranas, C.D., 2010. OptForce: An optimization procedure for identifying all genetic manipulations leading to targeted overproductions. PLoS Computational Biology 6.
Roca, C., Haack, M.B., Olsson, L., 2004. Engineering of carbon catabolite repression in recombinant xylose fermenting Saccharomyces cerevisiae. Appl. Microbiol. Biotechnol. 63, 578‐583.
Rolland, F., Winderickx, J., Thevelein, J.M., 2002. Glucose‐sensing and signalling mechanisms in yeast. FEMS Yeast Res. 2, 183‐201.
Schilling, C.H., Edwards, J.S., Letscher, D., Palsson, B.O., 2000. Combining pathway analysis with flux balance analysis for the comprehensive study of metabolic systems. Biotechnol. Bioeng. 71, 286‐306.
Siebold, C., Flükiger, K., Beutler, R., Erni, B., 2001. Carbohydrate transporters of the bacterial phosphoenolpyruvate: Sugar phosphotransferase system (PTS). FEBS Lett. 504, 104‐111.
![Page 35: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/35.jpg)
34
Slininger, P.J., Thompson, S.R., Weber, S., Liu, Z.L., Moon, J., 2011. Repression of xylose‐specific enzymes by ethanol in Scheffersomyces (Pichia) stipitis and utility of repitching xylose‐grown populations to eliminate diauxic lag. Biotechnol. Bioeng. 108, 1801‐1815.
Song, H.S., Ramkrishna, D., 2012. Prediction of dynamic behavior of mutant strains from limited wild‐type data. Metab. Eng. 14, 69‐80.
Suthers, P.F., Zomorrodi, A., Maranas, C.D., 2009. Genome‐scale gene/reaction essentiality and synthetic lethality analysis. Molecular Systems Biology 5.
Tepper, N., Shlomi, T., 2011. Computational design of auxotrophy‐dependent microbial biosensors for combinatorial metabolic engineering experiments. PLoS ONE 6, e16274.
Trinh, C.T., Unrean, P., Srienc, F., 2008. Minimal Eschenchia coli cell for the most efficient production of ethanol from hexoses and pentoses. Appl. Environ. Microbiol. 74, 3634‐3643.
Wouter Wisselink, H., Toirkens, M.J., Wu, Q., Pronk, J.T., Van Maris, A.J.A., 2009. Novel evolutionary engineering approach for accelerated utilization of glucose, xylose, and arabinose mixtures by engineered Saccharomyces cerevisiae strains. Appl. Environ. Microbiol. 75, 907‐914.
Yang, L., Cluett, W.R., Mahadevan, R., 2011. EMILiO: A fast algorithm for genome‐scale strain design. Metab. Eng. 13, 272‐281.
Yang, L., Mahadevan, R., Cluett, W.R., 2008. A bilevel optimization algorithm to identify enzymatic capacity constraints in metabolic networks. Computers and Chemical Engineering 32, 2072‐2085.
Yao, R., Hirose, Y., Sarkar, D., Nakahigashi, K., Ye, Q., Shimizu, K., 2011. Catabolic regulation analysis of Escherichia coli and its crp, mlc, mgsA, pgi and ptsG mutants. Microb Cell Fact. 10, 78.
Yomano, L., York, S., Shanmugam, K., Ingram, L., 2009. Deletion of methylglyoxal synthase gene (mgsA) increased sugar co‐metabolism in ethanol‐producing Escherichia coli. Biotechnol. Lett. 31, 1389‐1398.
![Page 36: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/36.jpg)
35
Highlights
• We developed an algorithm, SIMUP, to engineer organisms for sugar co‐utilization.
• The algorithm predicted multiple strategies including some non‐obvious strategies.
• Two model‐based Escherichia coli mutants showed co‐utilization of glucose and
xylose.
• No CCR gene had significant mutations, suggesting novel mechanisms of co‐
utilization.
![Page 37: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/37.jpg)
Maximize Growth on glucose and xylose
Subject to Network stoichiometry
Reaction bounds Gene knockouts identified by outer problem
Maximize Growth on glucoseSubject to
Network stoichiometryReaction bounds
Gene knockouts identified by outer problem
Maximize Growth on xylose
Subject to Network stoichiometry
Reaction bounds
Gene knockouts identified by outer problem
Maximize (Growth on glucose and xylose) –
(Growth on glucose) – (Growth on xylose) by deleting genes
Maximum gene knockouts ≤ Limit
Subject to
Figure
![Page 38: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/38.jpg)
PGI
RPE
FBP EDD EDA TPI FBA
eda
pts
zwf pgl
gnd
pgi
GLC
G6P
F6P
FDP
GA3PDHAP
6PGA 6PG Ru5P
GA3P
rpe
XuL
2DDG6Pedd
rpi
xkspfkA
fbp
fba
tpi
To other biomass precursors
tktA2
X5PR5P
S7P
F6P E4P
tktA1
talA
XYL
xylA
eda
pts
zwf pgl
gnd
pgi
GLC
G6P
F6P
FDP
GA3PDHAP
6PGA 6PG Ru5P
GA3P
rpe
XuL
2DDG6Pedd
rpi
xkspfkA
fbp
fba
tpi
To other biomass precursors
tktA2
X5PR5P
S7P
F6P E4P
tktA1
talA
XYL
xylA
Deleted reactions
Glucose flux
Glucose and xylose flux
Xylose flux
F6P Biomass precursor
rpe Deleted gene
PGI
GND
EDD EDA FBP
eda
pts
zwf pgl
gnd
pgi
GLC
G6P
F6P
FDP
GA3PDHAP
6PGA 6PG Ru5P
GA3P
rpe
XuL
2DDG6Pedd
rpi
xkspfkA
fbp
fba
tpi
To other biomass precursors
tktA2
X5PR5P
S7P
F6P E4P
tktA1
talA
XYL
xylA
PGI
GND
FBA FBP TPI
PFK
a b
c
Figure
![Page 39: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/39.jpg)
0.00
0.50
1.00
1.50
2.00
2.50
0.00
1.00
2.00
3.00
4.00
5.00
6.00
0 4 8 12 16 20 24
bio
mass (g/L
)
glu
cose ,
xylo
se (g/L
)
Time h
0.00
0.50
1.00
1.50
2.00
2.50
3.00
0.00
1.00
2.00
3.00
4.00
5.00
6.00
0 5 10 15 20
Dry
cell w
eig
ht
(g/L
)
glu
cose ,
xylo
se (g/L
)
Time h
a
0.00
0.50
1.00
1.50
2.00
2.50
3.00
0.00
1.00
2.00
3.00
4.00
5.00
6.00
0 5 10 15 20
bio
mass (g/L
)
glu
cose ,
xylo
se (g/L
)
Time h
c
b
dMM
LB
G X GX
NS
MM
LB
G X GX
NS
MM
LB
G X GX
NS
0.00
0.50
1.00
1.50
2.00
2.50
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
0 5 10 15 20
bio
mass (g/L
)
glu
cose ,
xylo
se (g/L
)
Time h
Figure
![Page 40: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/40.jpg)
G X G+X
F6P - - +
GA3P - - +
3PG - - +
PEP - - +
PYR - - +
R5P + - +
E4P - - +
OA - - +
AKG - - +
G X G+X
F6P - + +
GA3P + - +
3PG + - +
PEP + - +
PYR + - +
R5P - + +
E4P - - +
OA + - +
AKG + - +
G X G+X
F6P - + +
GA3P + + +
3PG + + +
PEP + + +
PYR + + +
R5P - + +
E4P - + +
OA + + +
AKG + + +
a b c
Not feasibleFeasible+ -
Figure
![Page 41: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/41.jpg)
0.00
0.50
1.00
1.50
2.00
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
0 5 10 15 20
bio
mass (g/L
)
glu
cose, xylo
se (g/L
)
Time h
0.00
0.50
1.00
1.50
2.00
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
0 5 10 15 20
bio
mass (g/L
)
glu
cose, xylo
se (g/L
)
Time h
a b
Figure
![Page 42: Novel approach to engineer strains for simultaneous sugar utilization](https://reader030.fdocuments.us/reader030/viewer/2022020119/575094c31a28abbf6bbbe4fe/html5/thumbnails/42.jpg)
0.00
0.50
1.00
1.50
2.00
2.50
3.00
0.00
1.00
2.00
3.00
4.00
5.00
6.00
0 5 10
bio
mass (g/L
)
glu
cose, x
ylo
se (g/L
)
Time h
0.00
0.50
1.00
1.50
2.00
2.50
0.00
1.00
2.00
3.00
4.00
5.00
6.00
0 5 10
bio
mass (g/L
)
glu
cose, x
ylo
se (g/L
)
Time h
a b
Figure