Skip to main content

Full text of "Bioconversion of methane to C-4 carboxylic acids using carbon flux through acetyl-CoA in engineered Methylomicrobium buryatense 5GB1C"

See other formats


Metabolic Engineering 48 (2018) 175-183 



ELSEVIER 


Contents lists available at ScienceDirect 

Metabolic Engineering 

journal homepage: www.elsevier.com/locate/meteng 



Bioconversion of methane to C-4 carboxylic acids using carbon flux through 
acetyl-CoA in engineered Methylomicrobium buryatense 5GB1C 

Shivani Garg a,b,1 5 Hao Wu 1,1 , James M. Clomburg a , George N. Bennett ’* 

a Department of Chemical and Biomolecular Engineering, Rice University, Houston 77030, United States 
b Department of BioSciences, Rice University, Houston 77030, United States 


ARTICLE INFO 


ABSTRACT 


Keywords: 

Methylomicrobium buryatense 
Methanotroph 
Crotonic acid 
Butyric acid 

Gas-to-liquid conversion 
One carbon feedstocks 


Methane, the primary component of natural gas, is the second most abundant greenhouse gas (GHG) and 
contributes significantly to climate change. The conversion of methane to industrial platform chemicals provides 
an attractive opportunity to decrease GHG emissions and utilize this inexpensive and abundantiy available gas as 
a carbon feedstock. While technologies exist for chemical conversion of methane to liquid fuels, the technical 
complexity of these processes mandate high capital expenditure, large-scale commercial facilities to leverage 
economies of scale that cannot be efficiently scaled down. Alternatively, bioconversion technologies capable of 
efficient small-scale operation with high carbon and energy efficiency can enable deployment at remote methane 
resources inaccessible to current chemical technologies. Aerobic obligate methanotrophs, specifically 
Methylomicrobium buryatense 5GB1, have recently garnered increased research interest for development of such 
bio-technologies. In this study, we demonstrate production of C-4 carboxylic acids non-native to the host, 
specifically crotonic and butyric acids, from methane in an engineered M. buryatense 5GB1C by diversion of 
carbon flux through the acetyl-CoA node of central ‘sugar’ linked metabolic pathways using reverse {/-oxidation 
pathway genes. The synthesis of short chain carboxylic acids through the acetyl-CoA node demonstrates the 
potential for engineering M. buryatense 5GB1 as a platform for bioconversion of methane to a number of value 
added industrial chemicals, and presents new opportunities for further diversifying the products obtainable from 
methane as the feedstock. 


1. Introduction 

Methane, a potent greenhouse gas (GHG) with 84 times more Global 
Warming Potential compared to carbon dioxide over a 20 year period 
(Greenhouse Gas Emissions, 2017), is also an energy rich, inexpensive 
and abundantly available carbon source. It is the primary component of 
bio-gas and natural gas, significant fractions of which are wasted due to 
flaring at production sites. For example, in 2012 3.5% of the world's 
natural gas supply was flared (Tollefson, 2016) and the amount of 
natural gas flared in the United States has more than doubled from 120 
billion cubic feet in 2005 to 278 billion cubic feet in 2015 (U.S. Natural 
Gas Vented and Flared, 2017). The flaring of natural gas is driven by the 
economics of natural gas recovery, which require a given rate of me¬ 
thane production to justify infrastructure expenditures needed to 
gather, process, and transport gaseous methane (Fei et al., 2014; 
Clomburg et al., 2017). Furthermore, the lack of gas-to-liquid (GTL) 
conversion technologies that can be efficiently deployed at the small- 
scale results in stranded gas at remote and marginal natural gas drilling 


sites (Haynes and Gonzalez, 2014; Strong et al., 2016). Existing che¬ 
mical GTL processes are technologically complex, requiring multiple- 
unit operations and resulting in high capital cost facilities that cannot 
be effectively scaled down (Haynes and Gonzalez, 2014). Thus, the 
development of bio-GTL conversion technologies that are readily de¬ 
ployable at small-scale, remote natural gas drilling sites (or other dis¬ 
tributed sources of methane emissions such as landfills or agricultural 
biogas) can enable recovery of abundantly available carbon and po¬ 
tentially allow on-site bio-conversion of otherwise wasted or flared 
methane into useful industrial platform chemicals and fuels (Fei et al., 
2014; Clomburg et al., 2017). In addition, lowering natural gas prices 
and global efforts to reduce greenhouse gases such as methane are other 
key drivers fueling development of bio-GTL technologies. 

Methanotrophs, organisms that use methane as the sole carbon and 
energy source (Hanson and Hanson, 1996; Whittenbury et al., 1970; 
Anthony, 1982), can be envisioned as promising industrial biocatalysts 
for bioconversion of methane to commercially relevant chemicals and 
fuels (Strong et al., 2016, 2015; Kalyuzhnaya et al., 2015; Conrado and 


* Corresponding author. 

E-mail address: gbennett@rice.edu (G.N. Bennett). 

1 These authors have contributed equally to this work. 

https://doi.Org/10.1016/j.ymben.2018.06.001 

Received 16 October 2017; Received in revised form 31 May 2018; Accepted 1 June 2018 
Available online 05 June 2018 

1096-7176/ © 2018 International Metabolic Engineering Society. Published by Elsevier Inc. All rights reserved. 

















S. Garg et al. 


Metabolic Engineering 48 (2018) 175-183 


Gonzalez, 2014). These organisms have been used for over 30 years for 
the production of products such as single cell protein (Norferm, Calysta: 
www.calysta.com), animal feed components (Bothe et al., 2002; Ye and 
Kelly, 2012), polyhydroxybutyrates (Criddle et al., 2013), methanol 
(Mehta et al., 1991), caretonoids, such as astaxanthin, (Ye and Kelly, 
2012) and bio-diesel (Fei et al., 2014), as well as for environmental bio¬ 
engineering applications (Hwang et al., 2007). However, these efforts 
have met minimal commercial success, in part hindered by the lack of 
genetic tools available for engineering methanotrophs, slowing progress 
and limiting the range of products. 

Recently, Methylomicrobium buryatense 5GB1 has been identified as a 
genetically tractable, aerobic and obligate methanotroph (Kaluzhnaya 
et al., 2001; Puri et al., 2015) that offers tremendous potential to be¬ 
come an industrial host for methane bio-conversion due to the avail¬ 
ability of a variety of genetic manipulations tools (Kalyuzhnaya et al., 
2015; Puri et al., 2015; Yan et al., 2016) and recent advances in un¬ 
derstanding of its metabolic pathways such as fatty acid biosynthesis 
and TCA cycle (Demidenko et al., 2016; Fu et al., 2017). As a Group I 
Gammaproteobacterium, M. buryatense assimilates carbon from me¬ 
thane at the formaldehyde level via the Ribulose Monophosphate 
(RuMP) pathway (Kaluzhnaya et al., 2001), generating fructose-6- 
phosphate that enters ‘sugar’-linked core metabolic pathways to gen¬ 
erate central metabolites such as phosphoenolpyruvate, pyruvate and 
acetyl-CoA. The presence of the Embden-Meyerhof-Parnas (EMP), En- 
tner-Doudoroff (EDD) and Pentose Phosphate (PPP) pathways 
(Kalyuzhnaya et al., 2015, 2013), in addition to the recently discovered 
Phosphoketolase (PKT) pathway (Henard et al., 2015) in this host 
means that metabolic engineering strategies established in industrial 
strains such Escherichia coli and Saccharomyces cerevisiae can be im¬ 
plemented in M. buryatense as ‘drop-ins’ to generate a variety of pro¬ 
ducts from central metabolites. Furthermore, the availability of a draft 
genome sequence (Khmelenina et al., 2013), a validated genome scale 
metabolic model (de la Torre et al., 2015), and bio-reactor performance 
parameters (Gilman et al., 2015) can enable rapid implementation and 
evaluation of metabolic engineering strategies for target product 
synthesis. These factors, combined with its tolerance to wide-ranging 
growth conditions and high resistance to contamination along with 
relatively rapid doubling time (2-3 h) compared to other methano¬ 
trophs (Kaluzhnaya et al., 2001), make it a promising host for the de¬ 
velopment of industrial processes. 

In addition, a recent study demonstrated the production of 0.8 g/L 
L-lactate from methane by M. buryatense 5GB1 through over-expressing 
a recombinant lactate dehydrogenase (Zdh) from Lactobacillus helveticus 
(Henard et al., 2016). This effort demonstrated applicability of this 
strain as a host for metabolic engineering approaches, however product 
synthesis was limited to utilizing carbon flux through the pyruvate node 
of central metabolism. To date, non-native product formation from 
acetyl-CoA, a node through which an array of promising fuel and 
chemical molecules can be synthesized, in M. buryatense or other Group 
I or Group II methanotrophs has proved challenging. In this study, we 
report the diversion of carbon from the acetyl-CoA node to product 
synthesis through a heterologous pathway in M. buryatense 5GB1C 
using methane as the sole carbon and energy source. Specifically, we 
demonstrate production of crotonic acid in M. buryatense 5GB1C by 
functional expression of a modified version of reverse P-oxidation 
pathway previously demonstrated inE. coli (Kim et al., 2016; Clomburg 
et al., 2012). This unsaturated C-4 carboxylic acid is utilized in many 
industrial applications, with its main use as a monomer for the synthesis 
of copolymers with a broad range of application such as paints, ad¬ 
hesives, coatings, ceramics, and agrochemicals (Blumenstein et al., 
2015). By over-expressing genes from various biological sources, we 
obtained up to 70mg/L of crotonic acid, in addition to the unexpected 
formation of 40mg/L butyric acid, from methane in M. buryatense 
5GB1C. In summary, we report a proof-of-concept of production of 
industrial platform chemicals from methane by utilizing the carbon flux 
through the acetyl-CoA node in an aerobic methanotroph, which opens 


avenues for expanding the product range possible via bio-conversion of 
natural gas. 

2. Materials and methods 

2.1. Strains, plasmids, and genetic manipulation 

Strain Methylomicrobium buryatense 5GB1C, a variant of strain 5GB1 
intentionally cured of its native plasmid (Puri et al., 2015), and M. 
buryatense 5GB1C Apta (Gilman et al., 2017) were used as the host 
strains. E. coli MG1655 genes atoB, fadB (Clomburg et al., 2012) and 
ydil (Kim et al., 2016), Clostridium acetobutylicum genes hbd and crt 
(Boynton et al., 1996), Methanosaeta thermophila gene acs (Berger et al., 
2012), and codon optimized NphT7 gene from Streptomyces sp. 
Okamura et al. (2010) synthesized by IDT DNA technologies, were 
cloned into the pAWP87 vector (Puri et al., 2015) with varying ribo¬ 
some binding sites (RBS1, RBS2, RBS3, RBS4 and RBS5) upstream of the 
genes, and under the control of constitutive methanol dehydrogenase 
promoter (P mxaF)> obtaining constructs pCA01-10. RBS1-5 were de¬ 
signed using RBS calculator v2.0 (Salis et al., 2009; Espah Borujeni 
et al., 2014) by selecting Methylobacterium extorquens AMI as a proxy 
host organism due to unavailability of M. buryatense 5GB1 in the RBS 
calculator program. Construction of the plasmids was performed using 
Gibson Assembly (New England Biolabs Inc., Ipswich, MA) as described 
previously (Gibson et al., 2009). The resulting plasmids and strains are 
listed in Table SI, primers used for molecular biology are listed in Table 
S2, and RBS used are listed in Table S3. 

Molecular kits for plasmid mini-preparation and gel extraction were 
purchased from QIAGEN (Hilden, Germany). DNA sequencing was 
outsourced to Lone Star Labs (Houston, TX). Transconjugation of host 
M. buryatense 5GB1C was performed with Escherichia coli S17-1 A pir 
acting as the donor strain as previously described (Puri et al., 2015). 

2.2. Medium, culture and fermentadon conditions 

M. buryatense 5GB1C cells were routinely cultured in NMS2 medium 
(Puri et al., 2015) at 30 °C with orbital shaking at 250 rpm unless 
otherwise stated. 99.5% pure CH 4 (Airgas, Houston, TX, USA) was used 
as the sole carbon source for all cultures. Strains were grown in 250 ml 
glass serum bottles (Kimble Chase, Vineland, NJ, USA) with 25% (v/v) 
CH 4 in air sealed with rubber stoppers and aluminum seals (Wheaton, 
Millville, NJ, USA). Plates were incubated in sealed jars (Oxoid Limited, 
Hampshire, United Kingdom) in an atmosphere of 25% CH 4 (v/v) in air 
at 30 °C. Antibiotics (50 pg/ml of kanamycin or 30 pg/ml rifamycin) 
were added to the growth medium as required. 

The seed culture was prepared by inoculating a single colony from a 
freshly grown plate in 30 ml NMS2 medium with 25% (v/v) CH 4 in air 
and incubated for 24 h. The primary pre-cultures were inoculated from 
this seed culture at 1% (v/v) into 250 ml serum bottle containing 40 ml 
NMS2 medium and grown with 25% (v/v) CH 4 in air. After 28 h, cells 
were harvested by centrifugation (10 min at 5875 g) and re-suspended 
in 20 ml fresh NMS2 medium to an OD of 30. During cultivation and 
fermentation, the gas phase of the serum bottle was refreshed every 
24 h by opening the bottles, re-sealing, removing 40 ml of air and in¬ 
jecting 60 ml of methane through the rubber seal septum. This resulted 
in a headspace containing 25% methane (v/v), 16% oxygen, and bal¬ 
ance nitrogen. Fermentations were monitored for 120h via culture 
samples collected every 24 h. 

For product toxicity experiments, crotonic, butyric and acetic acids 
(Sigma-Aldrich, St. Louis, MO) at concentrations 0.1, 0.2, 0.5 and 1 g/L 
were added to M. buryatense 5GB 1C cultures grown in 20 ml NMS2 
medium in glass serum bottles starting at an initial OD 0.05, incubated 
at 30 °C at 200 rpm in an orbital shaker. 


176 


S. Garg et al. 


Metabolic Engineering 48 (2018) 175-183 


Methane 

(CH 4 ) 

0T 

MMO 
•* h 2 o 
M ethanol 

(ch 3 oh) 


F-6-P 


Xu 5-P 


. )\ MDH ATP- ^ RUMP 

2e sA cycle 

\ 


Formaldehyde 

(HCHO) 


-i 

Formate 

(HCOOH) 

4 ' 




Ru-5-P 



T 




PEP 


Pyruvate 


NADH 

Carbon dioxide 

<co 2 ) 


A) Methane oxidation pathway 


B) Central metabolic pathways 

(EMP pathway, PKT pathway) 


K 


NADH + CO, Acetyl-CoA 


thiolase 




CoA 


AV 

Acetoacetyl-CoA 


3-hydroxyacyl-CoA 

dehydrogenase 


Y 


NADH 


p-Hydroxybutyryl-CoA 


enoyl-CoA 

hydratase 




Crotonic acid 


thioesterase 


_ 

Crotonyl-CoA 


CoA 


C) Engineered pathway for crotonic acid production 

using core reverse p-oxidation pathway enzymes 


Fig. 1. Product synthesis from acetyl-CoA node using engineered reverse P-oxidation pathway in M. buryatense 5GB1C. Methane oxidation pathway and central 
metabolic pathways consisting of Ribulose Monophosphate cycle pathway (RuMP; bold, black arrows),Embden-Meyerhoff-Parnas pathway (EMP; red arrows) and 
Phosphoketolase pathway (PKT; pink arrows) that are native to M. buryatense 5GB1C are shown (A, B). Engineered pathway for crotonic acid production from acetyl- 
CoA node is shown in blue (C). Abbreviations: MMO: Methane monoxygenase, MDH: Methanol Dehydrogenase, FALDH: Formaldehyde Dehydrogenase, FDH: 
Formate Dehydrogenase, Ru-5-P: Ribulose-5-Phosphate, Xu-5-P: Xylulose-5-Phosphate, F-6-P: Fructose-6-Phosphate, GAP: Glyceraldehyde-3-Phosphate, PEP: 
Phosphoenolpyruvate. 


2.3. Analytical methods 

Cell density was estimated by measurement of the optical density of 
appropriately diluted culture samples at 600 nm using a Beckman 
Coulter DU 800 UV/Vis Spectrophotometer (Beckman Coulter, Inc., 
Fullerton, CA). The OD to cell dry weight correlation was determined to 
be 0.24 gCDW/OD/L. 

The concentrations of extracellular metabolites (acetate, formate, 
crotonate and butyrate) in the culture supernatant were determined via 
HPLC as previously described (Clomburg et al., 2012). Briefly, HPLC 
was conducted using a Shimadzu Prominence SIL 20 system (Shimadzu 
Scientific Instruments, Inc., Columbia, MD, USA) equipped with an 
HPX-87H organic acid column (Bio-Rad, Hercules, CA, USA) with op¬ 
erating conditions to optimize peak separation. A mobile phase of 
30 mM H 2 S0 4 at a 0.3 ml/min flow rate was used and the column was 
operated at 42 °C. 

For GC-FID and GC-MS confirmation of crotonic and butyric acid, 
transesterification and derivatization of culture samples to fatty acid 
methyl esters (FAME) were performed as described previously (Kim 
et al., 2016). GC analysis was conducted on an Agilent 7890B Series 
Custom Gas Chromatography system equipped with a 5977B Inert Plus 
Mass Selective Detector Turbo El Bundle (for identification) or a Flame 
Ionization Detector (for quantification) and an Agilent HP-5-ms capil¬ 
lary column (0.25 mm internal diameter, 0.25 pm film thickness, 30 m 
length). The following temperature profile was used with helium as the 
carrier gas at a flowrate of 1.5 ml/min: Initial 50 °C (hold 3 min); ramp 
at 20 °C/min to 270 °C (hold 6 min). The injector and detector tem¬ 
peratures were 250 °C and 350 °C, respectively. 1 pi of sample was in¬ 
jected with a 4:1 split ratio. 

2.4. Enzyme assays 

For enzyme assays, M. buryatense 5GB1C cells were grown to mid¬ 


exponential phase, pelleted and re-suspended to OD 40 in 100 mM Tris- 
HC1 buffer (pH 7.0) with 1 mM Dithiothreitol (DTT). Re-suspended cells 
were lysed by adding 0.75 g glass beads to 1 ml of cell suspension in 
1.5 ml centrifuge tube followed by homogenization of the mixture in a 
Disruptor Genie (Scientific Industries, Inc, Bohemia, NY), and cen¬ 
trifugation at 15,000 Xg at 4°C for 10 min to pellet cell debris. The 
supernatant was used as the cell extract for measurement of enzyme 
activities. Enzyme assays for thiolase, 3-hydroxybutyryl-CoA dehy¬ 
drogenase, crotonase and thioesterase activities were performed as 
described previously (Clomburg et al., 2012; Vick et al., 2015). Briefly, 
thiolase (AtoB, EC 2.3.1.9) activity was measured by monitoring the 
decrease in acetoacetyl-CoA at 303 nm. 3-hydroxybutyryl-CoA dehy¬ 
drogenase (FadB/Hbd, EC 1.1.1.35) activity was measured by mon¬ 
itoring the decrease in NADH at 340 nm using acetoacetyl-CoA as the 
substrate. Crotonase (Crt, EC 4.2.1.17) activity was measured by 
monitoring crotonyl-CoA absorbance at 263 nm (Hartmanis and 
Gatenbeck, 1984). Thioesterase (Ydil, EC 3.1.2.22) activity was mea¬ 
sured by monitoring reduction of 5,5'-dithiobis (2-nitrobenzoic acid) 
(DTNB) at 412 nm using crotonyl-CoA as the substrate. All assays were 
carried out in 200 pi reaction volume in 96 well micro-titer plates at 
30 °C using Synergy 2 Multi-Mode Microplate Reader (BioTek, Wi¬ 
nooski, VT). Protein concentration in cell extracts was determined using 
Bradford reagent (BioRad, Hercules, CA). 

3. Results 

3.1. Engineering M. buryatense 5GB1C for crotonic acid production 

M. buryatense, an obligate methanotroph, assimilates carbon from 
methane via the ribulose monophosphate pathway (RuMP) that com¬ 
bines formaldehyde with ribulose-5-phosphate resulting in the forma¬ 
tion of fructose-6-phosphate. This hexose-phosphate further enters the 
‘sugar’ linked central metabolic pathways of M. buryatense resulting in 


177 

















S. Garg et al. 


Metabolic Engineering 48 (2018) 175-183 


central metabolites such as phosphoenolpyruvate, pyruvate and acetyl- 
CoA, each of which can serve as a node for production of bio-based 
chemicals and fuels. Additionally, the recently discovered phosphoke- 
tolase pathway (Henard et al., 2015) also contributes to acetyl-CoA 
formation from sugar-phosphates of the RuMP pathway (Fig. 1). To 
demonstrate production of an industrial platform chemical from the 
acetyl-CoA node in M. buryatense 5GB1C, we sought to engineer a re¬ 
verse (5-oxidation based metabolic pathway for crotonic acid production 
(Fig. 1). This pathway has been demonstrated in the industrial strain 
Escherichia coli (Kim et al., 2016; Clomburg et ah, 2012) wherein a 
thiolase (AtoB) initiates the pathway by condensing two molecules of 
acetyl-CoA to generate acetoacetyl-CoA, followed by a 3-hydroxyacyl- 
CoA dehydrogenase (FadB) that reduces acetoacetyl-CoA to 3-hydro- 
xybutyryl-CoA, which is dehydrated by an enoyl-CoA hydratase (FadB) 
to generate crotonyl-CoA. Finally, an enoyl-CoA specific thioesterase 
(Ydil) cleaves the thiol group of crotonyl-CoA resulting in the free un¬ 
saturated fatty acid crotonic acid (Fig. 1). 

For engineering this pathway, atoB, fadB and ydil genes from E. coli 
MG1655 were initially selected to catalyze the thiolase, 3-hydroxyacyl 
dehydrogenase/enoyl-CoA hydratase and thioesterase reactions, re¬ 
spectively, as these genes have previously been used in E. coli for suc¬ 
cessful demonstration of crotonic acid production from acetyl-CoA (Kim 
et al., 2016). To optimize the expression level of each enzyme in the 
heterologous pathway, we used RBS variants of different strengths de¬ 
signed using RBS calculator 2.0 program (Salis et al., 2009; Espah 
Borujeni et al., 2014). This approach of varying RBS to optimize 
translation initiation rate and thereby increasing final product titers has 
been used previously for chemical production in E. coli (Nowroozi et al., 
2014). Additionally, we also varied the order of genes as gene clusters 
may have a bearing on transcription and may interfere in translation, 
resulting in altered levels of final product (Smanski et al., 2014). Thus, 
by using a different ribosome binding site for each gene, and varying 
the order of the three genes on the pAWP87 plasmid (Puri et al., 2015) 
under the control of a constitutive methanol dehydrogenase promoter 
( PmxaF ) (Puri et al., 2015), four strains were obtained, namely 5GB1C 
(pCAOl), 5GB1C (pCA02), 5GB1C (pCA03) and 5GB1C (pCA04) 


(Fig. 2A). 

Functional expression of AtoB, FadB and Ydil in M. buryatense 
5GB1C was confirmed via enzyme activity assays that showed increased 
thiolase (—20 pmol/mg/min) and thioesterase (—5-6 nmol/mg/min) 
activities compared to the control (Fig. 2B). FadB activity (as measured 
by (5-ketoacyl-CoA reduction) was detected at lower levels 
(0.05-0.1 pmol/mg/min) than those observed previously in E. coli (Vick 
et a] 2015) (Fig. 2B). Differences in the specific activities measured for 
each protein in response to the gene configuration and RBS utilized 
were observed, with minimal variation in thiolase and thioesterase 
activities (less than 25%) while FadB activity showed a near 2-fold 
variation depending on the configuration and RBS combination 
(Fig. 2B). Despite overall low FadB activity, crotonate production was 
detected via HPLC, and confirmed by GC-MS (Supplementary Fig. SI), 
in all four strains at titers ranging between 60 and 70 mg/L. No crotonic 
acid production was observed in the control strain containing the empty 
vector. Normalizing the crotonic acid titer relative to dry cell weight, 
90-190 pmol crotonate/gCDW was produced by these strains (Fig. 2C). 
Interestingly, the time at which crotonic acid first appeared varied with 
each strain. For example, 5GB1C (pCAOl) showed crotonic acid pro¬ 
duction only after 120 h, whereas variations in RBS and gene order 
afforded faster production, with 5GB1C (pCA03) and 5GB1C (pCA04) 
forming crotonic acid as early as 48 h (Fig. 2C). Notably, the strains 
containing the 5GB1C (pCA03) and 5GB1C (pCA04) constructs had 
nearly 2-fold higher FadB activity compared to 5GB1C (pCAOl) 
(Fig. 2B). During the course of cultivation, the OD remained stable, and 
pH decreased to 8.0-8.5 from an initial pH of 9.5. 

In addition to crotonic acid, expected by-products such as acetate 
and formate and the unexpected by-product butyrate were detected in 
culture medium of these strains. All four strains primarily formed 
acetate (0.05-0.2 g/L) and formate (0.01-0.02 g/L) prior to producing 
crotonic acid. Butyric acid at titers of approximately 0.04 g/L was de¬ 
tected either at the time of crotonic acid production (in 5GB1C 
(pCA04)) or after formation of crotonic acid (in 5GB1C (pCAOl), 
5GB1C (pCA02) and 5GB1C (pCA03)). The identity of butyric acid was 
confirmed via GC-MS (Supplementary Fig. S2). The empty vector 


A PmwF 

p 

pCAOl -L R1 I fadB | - R2 | atoB ‘ - R3 f Ydil 




pCA02 - 1 - R1 I atoB / — | R2 | fa dB - | R3 | y dil 

pCA03 -J-Rl ydil — R2 fadB - R3 atoB 

P 

' mxof 

pCA04 Xr 2 atoB - R3 fadB — Rl | ydil 


B 


Strain 

Thiolase 

3-hydroxybutyryl- 
CoA dehydrogenase 

Thioesterase 

(Crotonyl-CoA) 

WT 

0.004 ±0.001 

0.003 ± 0.000 

Not detected 

5GB1C (pCAOl) 

21.26 ±0.02 

0.039 ± 0.001 

0.0059 ± 0.0001 

5GB1C (pCA02) 

17.3 ±0.2 

0.054 ± 0.002 

0.0050 ± 0.0001 

5GB1C (pCA03) 

20.0 ± 0.4 

0.075 ± 0.003 

0.0057 ± 0.0002 

5GB1C (pCA04) 

19.1 ± 0.2 

0 073 ± 0 001 

0.0062 ± 0.0001 


All activities expressed in p mol/mg protein/mi n 


c 


D 



5GB1C (pCAOl) 
5GB1C (pCA02) 
5GB1C (pCA03) 
5GB1C (pCA04) 
Wild-type 



Time (hr) 


♦ Acetate 
■ Formate 

• Butyrate 


Fig. 2. Production of crotonic acid in engineered M. buryatense 5GB1C strains. Four strains expressing atoB, fadB and ydil genes with varying RBS and gene-orders 
under the control of constitutive mxaF promoter (P^of) were designed (A). Expression of thiolase (AtoB), 3-hydroxybutyryl-CoA dehydrogenase (FadB), thioesterase 
(Ydil) in M. buryatense 5GB 1C was confirmed via enzyme activity assays (B). Crotonate production profile is shown for all four strains grown in 20 ml culture medium 
in 250 ml glass serum bottles at high cell density (OD 30), 30 °C, 200 rpm with 25% methane-in-air with headspace refreshed every 24 h (C) Additional by-products 
profiles of the best performing strain, 5GB1C (pCA04), (dashed lines) is compared with that of the empty vector control strain (solid lines) (D). Each data point is 
average of replicate experiments, and error bars represent standard deviation. 


178 









































S. Garg et al. 


Metabolic Engineering 48 (2018) 175-183 


AtoB / 

NphT7 


Acetyl-CoA 


thiolase 


tc 


Acetyl-CoA/ 

Malonyl-CoA 

CoA 


XV 

Acetoacetyl-CoA 

/ 3-hydroxyacyl-CoA NADH 
Hbd dehydrogenase ? 

oh o 

3-Hydroxybutyryl-CoA 


Crt 


enoyl-CoA 

hydratase 


thioesterase 


h 2 o 

Crotonyl-CoA 

— T.nA 


Crotonic acid 


B 



pCA04 pCA05 pCA06 pCA07 


AtoB 

NphT7 


Hbd, Crt 
Ydil 


Formate 
Acetate 
Butyrate 
■ Crotonate 


Thiolase 

- Acetoacetyl-CoA synthase 

3-Hydroxyacyl-CoA 
~ dehydrogenase/ 
Enoyl-CoA reductase 

Thioesterase 


Fig. 3. Expression of alternative genes for 
crotonic acid production in M. buryatense 
5GB1C via acetyl-CoA node. Genetic manip¬ 
ulations are shown that target replacement of 
E. coli MG1655 atoB with Streptomyces sp. 
NphT7 for acetoacetyl-CoA synthesis, E. coli 
MG1655 fadB with hbd, crt from Clostridium 
acetobutylicum for 3-hydroxyacyl-CoA dehy¬ 
drogenase and enoyl-CoA hydratase reactions 
in engineered M. buryatense 5GB1C host (A). 
Evaluation of metabolite production profiles 
(at 72 h) in resultant strains harboring plas¬ 
mids pCA05, pCA06 and pCA07, as compared 
to strain with parent plasmid, pCA04 (B). 


control strain formed acetate and formate but did not accumulate bu¬ 
tyrate or crotonate (Fig. 2D). This suggests that the engineered strains 
accumulated butyric acid most likely by action of an endogenous enoyl- 
CoA reductase that reduced crotonyl-CoA to butyryl-CoA that subse¬ 
quently underwent thiol cleavage reaction catalyzed either by a native 
thioesterase, or the recombinant thioesterase, Ydil which has been 
shown to be highly specific for four carbon-chain acyl-CoAs (Kim et al., 
2016). 

As strain 5GB1C (pCA04) formed both crotonic acid and butyric 
acid at 48 h, the earliest timepoint at which crotonic acid was observed, 
this strain was selected for further genetic manipulations aimed at in¬ 
creasing crotonic acid titers. 

3.2. Alternative genes to enhance crotonic acid production 

To improve crotonic acid titers, several gene replacements were 
attempted targeting replacement of E. coli atoB and fadB genes with 
alternative versions from other organisms (Fig. 3A). Specifically, as low 
FadB activity was detected in the initial strains 5GB1C (pCA01-04), we 
first sought to replace E. coli MG1655 fadB with NADH-dependent Hbd 
and Crt from Clostridium acetobutylicum that are involved in butyryl- 
CoA synthesis from acetyl-CoA (Boynton et ah, 1996; Hartmanis and 
Gatenbeck, 1984), generating strain 5GB1C (pCA05). These enzymes 
have been utilized in numerous metabolic engineering studies aimed at 
C4 product synthesis (e.g. butanol) due to their high specificity for four- 
carbon intermediates (Shen et al., 2011). Additionally, given the ther¬ 
modynamic constraints of the condensation reaction (Lan and Liao, 
2012) and the potential for the acetyl-CoA pool to be a limiting factor, 
we replaced E. coli AtoB that catalyzes reversible condensation of two 
molecules of acetyl-CoA, with a KASHI homolog, the acetoacetyl-CoA 
synthase (EC 2.3.1.194) NphT7 from soil bacterium Streptomyces sp. 
CL190, that carries out the ATP-dependent irreversible condensation of 
malonyl-CoA with acetyl-CoA to form acetoacetyl-CoA (Okamura et al., 
2010). As M. buryatense 5GB1 forms relatively high amounts of lipids 
(8% FAME content per g CDW) compared to many other bacteria 
(Demidenko et al., 2016), we reasoned that this organism may have a 
large pool of malonyl-CoA available for NphT7 driven condensation 
with acetyl-CoA. Strain 5GB1C (pCA06) was constructed that expressed 
NphT7 in place of AtoB, and NphT7 was further combined with Hbd, 
Crt to generate strain 5GB1C (pCA07). 

However, these gene replacements did not improve crotonic or 
butyric acid titers compared to the parent strain, 5GB1C (pCA04) 
(Fig. 3B). As FadB activity was low in the parent strain, we determined 


3-hydroxybutyryl-CoA dehydrogenase activity levels in all the strains 
expressing alternative versions of FadB, i.e. Hbd and Crt. Enzyme ac¬ 
tivity assays of 3-hydroxybutyryl-CoA dehydrogenase in 5GB1C 
(pCA05) and 5GB1C (pCA07) expressing Hbd and Crt revealed slightly 
higher activity, 0.150 ± 0.048 pmol/min/mg and 0.133 ± 0.029 
pmol/min/mg in the two strains, respectively, compared to 
0.073 ± 0.001 pmol/min/mg in the parent strain, 5GB1C (pCA04). 
However, crotonase activity in both 5GB1C (pCA05) and 5GB1C 
(pCA07) was low, 0.007 ± 0.004 pmol/ming/mg and 0.027 ± 0.002 
pmol/min/mg respectively, which could be responsible for low flux 
through the pathway. Even though crotonic and butyric acid levels 
were unaffected, differences were noted in the number of by-products 
produced by these strains (Fig. 3B). Strain 5GB1C (pCA05) expressing 
atoB, hbd, crt and ydil formed the highest amount of acetate compared 
to other strains that expressed nphT7 with fadB/hbd,crt and ydil. Spe¬ 
cifically, 5GB1C (pCA05) formed 0.3 g/L acetate (equivalent to —500 
umol/gCDW) which was 3-fold and 1.5-fold higher than that produced 
by 5GB1C (pCA06) and 5GB1C (pCA07) respectively. The amount of 
formate generated by 5GB1C (pCA05) (0.01 g/L) was also lower than 
that formed by 5GB1C (pCA06) (0.02 g/L) and 5GB1C (pCA07) (0.04 g/ 
L). Given that 5GB1C (pCA05) channeled more carbon to acetate for¬ 
mation and with lower formate production compared to other strains, 
this strain was chosen for further genetic manipulations targeted at 
blocking acetate formation and up-regulating acetate conversion to 
acetyl-CoA, thus potentially improving carbon flux through acetyl-CoA 
node. 

3.3. Genetic manipulations to alter acetate formation and acetate uptake 

To increase crotonate production, the excess carbon being lost to 
acetate could be recovered and channeled towards crotonate either by 
i) blocking the pathway for acetate formation by deleting the phos- 
photransacetylase (pta ) gene in the host strain, a strategy commonly 
used to reduce the flux to acetate (Dittrich et al., 2005; Shams Yazdani 
and Gonzalez, 2008; Singh et al., 2011), or ii) by recycling acetate back 
to acetyl-CoA by overexpressing acetyl-CoA synthetase (acs). The latter 
strategy has been reported in literature as a tool for lowering acetate 
accumulation in E. coli and in Saccharomyces cerevisiae (Lin et al., 2006; 
Shiba et al., 2007). Both of these strategies were evaluated to improve 
crotonic acid production in M. buryatense 5GB1C (Fig. 4A). 

The plasmid expressing atoB, hbd, crt and ydil genes (pCA05) was 
introduced into an engineered M. buryatense 5GB1C host strain con¬ 
taining a pta deletion and product formation was evaluated. While 


179 


















S. Garg et ai 


Metabolic Engineering 48 (2018) 175-183 



ADP 


atoB 


hbd 


crt 


1 

J 

1 



u 


Crotonyl-CoA 

NADH 


enoyl-CoA 

reductase 


yd/7 


( 

o 


Butyric acid 


Butyryl-CoA 


B 



0.5 
__ 0.4 


§ 0.2 


0.1 


0 •' 
0 


0.5 

___ 0.4 

n 

S 0.3 
o> 

5 02 

o 
u 
< 


0.1 

0 •' 
0 


48 72 

Time (hr) 


- 4 


I 


24 


48 72 

Time (hr) 


96 


Time (hr) 


0.1 ? 

O) 


0.08 


0.06 


o 
** 

rz 

2 

0.04 | 

0.02 § 
O 

0 O 



■*Crotonate mm Butyrate -♦--Acetate 

Fig. 4. Genetic manipulation of acetate formation and uptake system to increase crotonic acid production in engineered M. bwyatense 5GB1C. Pathway showing 
genetic manipulations incorporated in M. buryatense 5GB1C host strain to i) block acetate formation by deleting pta gene, or ii) re-utilize acetate by over-expressing 
acs from E. coli or Methanosaeta thermophila (A). Metabolite profile of 5GB1C (pCA05) is shown (B), impact of Apta gene deletion on acetate formation and 4-C 
carboxylic acids formation in 5GB1C strain expressing pCA05 plasmid (C), and impact of M. thermophila acs over-expression in combination with atoB, hbd, crt and 
ydil in 5GB1C (pCA09) is shown (D). Each data point is average of replicate experiments, and error bars represent standard deviation. 


acetate production was reduced approximately 50% in the A pta host 
strain harboring pCA05 compared to the wild-type host strain har¬ 
boring the same plasmid, there was no improvement in crotonic acid 
titer (Figs. 4B, 4C). Furthermore, crotonic and butyric acid production 
was not observed until 96 h and 120 h, respectively. 

In an alternative strategy aimed at assimilating acetate, an AMP- 
forming acetyl-CoA synthetase gene sourced either from E. coli (Kumari 
et al., 1995) or Methanosaeta thermophila, a thermophilic methanogen, 
(Berger et al., 2012) was overexpressed in combination with atoB, hbd, 
crt and ydil genes resulting in strains 5GB1C (pCA08) and 5GB1C 
(pCA09), respectively. While strain 5GB1C (pCA08) that over-expressed 
E. coli acs did not show any crotonic or butyric acid production, strain 
5GB1C (pCA09) over-expressing M. thermophila acs formed crotonic and 
butyric acids as early as 24 h. Even though crotonic acid titer in 5GB1C 
(pCA09) was similar to that produced by the parent strain, 5GB1C 
(pCA05), butyric acid titer increased to 0.08 g/L by 96 h in 5GB1C 
(pCA09), a 2-fold increase compared to 5GB1C (pCA05) (Figs. 4B, 4D). 
Acetate production in 5GB1C (pCA09) was reduced by 60-90% com¬ 
pared to 5GB1C (pCA05) (Figs. 4B, 4D). A cumulative titer of 0.15 g/L 
of 4-C carboxylic acids was eventually obtained in 5GB1C (pCA09) 
strain after 96 h through overexpression of an acetyl-CoA synthetase 
aimed at shuttling carbon from acetate back to acetyl-CoA. 


3.4. Process optimization to improve crotonic acid production 

Using the 5GB1C (pCA09) strain, various process modifications 
were evaluated for their impact on product formation. Changing the 
methane-to-air ratio in the headspace, increasing nitrate concentration 
from 1 x to 8 x in the medium (to overcome nitrogen limitation if 
any), and variation in culture volume (10, 15, 30 ml) did not improve 
crotonic acid production (data not shown). 

Product toxicity was also evaluated by growing wild-type M. bur¬ 
yatense 5GB 1C in varying concentrations of acetic, butyric and crotonic 
acids. While acetic and butyric acids up to 0.5 g/L did not impact M. 
buryatense 5GBlC's growth, 1 g/L of both these acids resulted in slight 
growth impairment compared to the control to which no acids were 
added (Fig. 5A, B). In contrast, crotonic acid above 0.1 g/L proved to be 
highly inhibitory to M. buryatense's growth (Fig. 5C). While the max¬ 
imum amount of crotonic acid formed in this study is less than 0.1 g/L, 
suggesting that bottlenecks in the pathway, rather than product toxi¬ 
city, prevent accumulation/formation of crotonic acid at titers higher 
than 0.07 g/L, this toxicity is an important consideration in further 
improving crotonic acid production. In addition to improving pathway 
function (see Section 4), strain engineering efforts can possibly 
target alteration of membrane fluidity to block toxic product entry into 


180 








S. Garg et al. 


Metabolic Engineering 48 (2018) 175-183 


E 

c 

o 

o 

S£- 

Q 

O 



-»■■ Og/L 
0.1 g/L 
-a- 0.2 g/L 
0.5 g/L 
1 g/L 


B 


E 

c 

o 

o 

S2. 

Q 

O 



-i Og/L 
0.1 g/L 
0.2 g/L 
0.5 g/L 
1 g/L 



Og/L 
0.1 g/L 
0.2 g/L 
0.5 g/L 
1 g/L 


Fig. 5. Crotonic, butyric and acetic acids tolerance of M. buryatense 5GB1C. Impact of varying concentrations (0, 0.1, 0.2, 0.5 and 1 g/L) of acetic acid (A), butyric 
acid (B) and crotonic acid (C) on growth of M. buryatense 5GB1 is shown. Each data point is average of two replicates, and error bars represent standard error. 


cells, and heterologous expression of effluent pumps to remove excess 
product out of cells (Dunlop, 2011). Furthermore, a rational approach 
aimed at identifying toxicity targets of crotonic acid in combination 
with random mutagenesis to identify crotonic acid tolerant mutants can 
also be used (Connor et al., 2010; Stephanopoulos, 2002). Adaptive lab 
evolution can also be used to improve strain tolerance to increased 
crotonic acid concentrations (Hu et al., 2016). Understanding the me¬ 
chanism^) of product toxicity and these types of approaches to alle¬ 
viate toxic effects will be critical for both crotonic acid production as 
well as other non-native products that can potentially be produced via 
metabolic engineering of this methanotroph. 

4. Discussion 

In this study, we have demonstrated product formation from the 
acetyl-CoA node through heterologous pathway expression in an ob¬ 
ligate methanotroph using methane as the carbon source. For this, a 
modified version of the reverse P-oxidation (r-BOX) pathway 
(Dellomonaco et al., 2011) was used where one turn of the r-BOX cycle 
was combined with an enoyl-CoA specific cycle termination thioes- 
terase, Ydil, (Kim et al., 2016) to generate the four carbon a,P-un- 
saturated carboxylic acid crotonic acid from crotonyl-CoA. In addition 
to crotonic acid, the four-carbon saturated carboxylic acid butyric acid 
was also formed as the by-product of the engineered pathway, with the 
cumulative titer of these two products reaching 0.15 g/L in the best 
performing strain. The formation of butyrate from crotonic acid was 
unexpected as no trans-enoyl-CoA reductase (TER) was over-expressed. 
This suggests that reduction of crotonyl-CoA to butyryl-CoA is being 
driven by an endogenous trans-enoyl-CoA reductase. Previously, it has 
been shown that E. coli enoyl-ACP reductase (FabI) possesses inherent 
TER activity (Vick et al., 2015) to support reduction of crotonyl-CoA to 
butyryl-CoA. It is possible that M. buryatense FabI (METBU_RS0100175, 
NCBI Reference Sequence: NZ_KB455575.1) also possesses TER activity 
similar to that possessed by E. coli FabI, thus catalyzing reduction of 
crotonyl-CoA to butyryl-CoA, thereby leading to butyrate production, 
most likely by the action of over-expressed Ydil on butyryl-CoA. 

Despite the successful demonstration of product synthesis from the 
acetyl-CoA node, the overall titers achieved remain low. Based on the 
measurement of specific activities of each enzyme involved in the 
pathway from acetyl-CoA to crotonic acid, our results indicate that 
functional expression levels of heterologous genes may limit pathway 
performance in this organism. Specifically, the measured 3-hydro- 
xybutyryl-CoA dehydrogenase activities, regardless of the enzyme uti¬ 
lized (i.e. FadB or Hbd), were significantly lower (by more than an 
order of magnitude) than those observed upon overexpression of these 
enzymes in E. coli (Vick et al., 2015; Shen et al., 2011). In contrast, 
thiolase activity observed upon AtoB overexpression in M. buryatense is 
remarkably similar to that obtained in E. coli (Vick et al., 2015; Shen 
et al., 2011), indicating that functional expression of enzymes in this 
methanotroph may be highly enzyme specific. Thus, future approaches 
to improve strain performance can target evaluation of additional gene 
variants for each step of the pathway including evaluation of codon 


usage in these genes to determine the best gene candidates. Our results 
also showed that varying RBS and gene order impacted the rate of 
product formation. This means while a complete understanding of M. 
buryatense physiology is still lacking, a combinatorial gene assembly 
approach as used in E. coli and S. cerevisiae, by varying promoters, RBS, 
gene order and gene variants, for heterologous expression of multiple 
genes (Nowroozi et al., 2014; Zhang et al., 2015; Keasling, 2012) would 
be useful for future metabolic engineering efforts in M. buryatense 
5GB1. These efforts can be aided by a more rigorous examination of 
transcriptional and translational responsiveness of gene and protein 
expression to various synthetic designs (e.g. promoters, RBS, gene 
configuration, etc.) to provide an improved understanding of key ele¬ 
ments for optimal heterologous gene expression. 

In addition to functional expression of the desired pathway en¬ 
zymes, the availability of acetyl-CoA also appears to be a major factor 
dictating product synthesis. Our experiments showed that while de¬ 
leting phosphotransacetylase reduced acetate formation, crotonic and 
butyric acid formation were also reduced as a result. These results in¬ 
dicate that decreasing acetate formation through this approach did not 
enable increased acetyl-CoA availability. This supports the role and 
importance of phosphotransacetylase within the phosphoketolase 
pathway for acetyl-CoA formation in which phosphotransacetylase 
catalyzes the conversion of acetyl-phosphate to acetyl-CoA (Henard 
et al., 2015). Given the recent report demonstrating a 2-fold increase to 
the intracellular acetyl-CoA concentration and enhanced biomass and 
lipid accumulation via overexpression of a phosphoketolase isoform in 
M. buryatense (Henard et al., 2017), this pathway appears to play a large 
role in acetyl-CoA formation. As such, an alternative approach for re¬ 
ducing acetate formation while still retaining phosphotransacetylase for 
converting acetyl-phosphate to acetyl-CoA can be explored through the 
deletion of acetate kinase. 

In contrast to a pta deletion, over-expression of M. thermophila acs 
(Berger et al., 2012) for the conversion of acetate to acetyl-CoA enabled 
increased C4-carboxylic acid formation, demonstrating the importance 
of increasing acetyl-CoA availability and viability of this approach for 
this purpose. To further increase flux through acetyl-CoA node, addi¬ 
tional pathways that divert carbon away from acetyl-CoA can be 
blocked (Kalyuzhnaya et al., 2015). For example, by deleting lactate 
dehydrogenase, more pyruvate can be channeled to acetyl-CoA for¬ 
mation via EMP pathway. Additionally, elimination of the glycogen 
synthesis pathway can also increase availability of carbon for central 
metabolites formation. M. buryatense also has a highly active fatty acid 
synthesis pathway that draws from the acetyl-CoA pool (Demidenko 
et al., 2016), thus down-regulation of the FAS pathway, in combination 
with deletion of native beta-oxidation pathway genes (Demidenko 
et al., 2016) could further improve acetyl-CoA available for target 
product synthesis. However, the alteration of FAS pathway flux may 
have the unintended consequence of limiting intracellular membrane 
formation thereby inhibiting particulate methane monooxygenase's 
functioning. In such a case, strategies such as switching to copper-free 
cultivation of M. buryatense to activate sMMO expression, or fine-tuning 
of FAS down-regulation to enable sufficient expression of pMMO can be 


181 









S. Garg et al. 


Metabolic Engineering 48 (2018) 175-183 


explored. 

Increased knowledge on the physiology and metabolic network of 
M. buryatense can also aid in optimizing flux to maximize acetyl-CoA 
and target product synthesis. Based on experimental data, two routes 
are known to generate acetyl-CoA in M. buryatense'. i) EMP pathway 
where pyruvate dehydrogenase converts pyruvate to acetyl-CoA 
(Kalyuzhnaya et al., 2013), and ii) PKT pathway where FPkt or XPkt 
catalyze conversion of fructose-6-phosphate or xylulose-5-phosphate to 
acetyl-CoA (Henard et al., 2015). Each of these pathways have varying 
requirements for reducing equivalents, and the current lack of knowl¬ 
edge about the exact requirements of reducing equivalents for ATP 
generation in M. buryatense 5 GB1 limits metabolic engineering efforts 
for optimizing flux through these three pathways. Theoretically, by 
modulating the carbon flux through these pathways, different carbon 
and energy efficiencies can be achieved. For example, if acetyl-CoA is 
primarily formed through the EMP pathway, then production of 1 mol 
of crotonyl-CoA results in the net generation of 3 NADH molecules, 2 
ATP and 2 CO 2 per 6 mol of formaldehyde entering the RuMP pathway, 
achieving 67% carbon efficiency. At this maximum theoretical carbon 
efficiency, for every mole of crotonic acid produced, 6 mol of methane 
are consumed and 2 mol of CO 2 released. In contrast, the formation of 
crotonyl-CoA from the PKT pathway consumes 1 NADH per 4 mol of 
formaldehyde, with 100% carbon conversion efficiency (4HCHO + 
NADH -» Crotonyl-CoA). By distributing carbon flux through these two 
pathways (equations (i), (ii) and (iii)), a redox balanced and ATP 
generating conversion of formaldehyde to crotonyl-CoA can be 
achieved with maximum theoretical carbon efficiency of 89%. 

3HCHO -» 0.5Crotonyl-CoA + C0 2 + ATP + 1.5NADH (i) EMP 
pathway 

6HCHO + 1.5NADH -* 1.5Crotonyl-CoA (ii) PKT pathway 
9HCHO -» 2Crotonyl-CoA + C0 2 + ATP (iii) = (i) + (ii) 

Determination of flux distribution between these pathways and 
understanding of reducing equivalent requirements by M. buryatense 
5GB1 can enable future efforts targeting improved carbon efficiency 
and restoration of redox balance through these pathways, in combina¬ 
tion with regulating the expression of formate dehydrogenase and 
formaldehyde dehydrogenase genes. 

Furthermore, based on genome scale metabolic reconstruction of M. 
buryatense a partial serine cycle is predicted to contribute to the acetyl- 
CoA pool (de la Torre et al., 2015), indicating overexpression of serine 
cycle enzymes may also improve titers of products formed through 
acetyl-CoA. Per mole of formaldehyde, the serine cycle consumes a 
mole of CO 2 to produce a mole of acetyl-CoA while consuming NADH 
and ATP (Kalyuzhnaya et al., 2015) (equation iv). Thus, while activa¬ 
tion of the serine pathway for crotonic acid production may allow 
capturing of C0 2 along with CH 4 , this pathway will be very demanding 
in terms of reducing equivalents and ATP requirements (equation vi) 
making it unfavorable for target product synthesis. 

HCHO + C0 2 + 2NADPH + 3ATP -» Acetyl-CoA (iv) Serine cycle 
2Acetyl-CoA + NADH -* Crotonyl-CoA (v) Crotonic acid production 
pathway 

2HCHO + 2C0 2 + 4NAD(P)H + NADH + 6ATP -» Crotonyl-CoA 
(vi) = (iv) + (v) 

Building on the results of this study, which has demonstrated pro¬ 
duct synthesis from acetyl-CoA node in an obligate methanotroph, these 
types of approaches can be exploited to both improve product synthesis 
and open up avenues for further diversifying target products from 
acetyl-CoA node (Kalyuzhnaya et al., 2015). For example, by utilizing 
multiple turns of r-BOX cycle, medium chain saturated and unsaturated 
carboxylic acids can be produced (Kim et al., 2015), and by varying 
starter substrates for the r-BOX pathway or for fatty acid synthesis 
pathway, functionalized carboxylic acids and alcohols can be generated 


(Cheong et al., 2016; Garg et al., 2016). Implementation of these ap¬ 
proaches for improving product synthesis and diversification will also 
rely on further advancement in the understanding of and ability to 
engineer M. buryatense to unlock the potential of this organism for the 
biological conversion of methane to fuels and chemicals. 

Acknowledgements 

This work was funded by an award from the U.S. Department of 
Energy ARPA-E REMOTE program (Award No. DE-AR000439). The 
authors would like to thank Dr. Ka-Yiu San for providing the M. ther- 
mophila acs gene (US Patent Application #62/635417). Authors would 
also like to thank Mr. William Duong for assistance with fermentations. 
Authors have no competing interests in this study. 

Appendix A. Supporting information 

Supplementary data associated with this article can be found in the 
online version at http://dx.doi.Org/10.1016/j.ymben.2018.06.001. 

References 

Anthony, C., 1982. The Biochemistry of Methylotrophs. Academic Press, London, New 
York, pp. 431. 

Berger, S., Welte, C., Deppenmeier, U., 2012. Acetate activation in Methanosaeta ther- 
mophila: characterization of the key enzymes pyrophosphatase and acetyl-coa syn¬ 
thetase. Archaea 2012, 315153. 

Blumenstein, J., Albert, J., Shulz, R.P., Kohlpaintner, C., 2015. Crotonaldehyde and 
Crotonic Acid. Ullmann’s Encyclopedia of Industrial Chemistry. 

Bothe, H., et al., 2002. Heterotrophic bacteria growing in association with Methylococcus 
capsulatus (Bath) in a single cell protein production process. Appl. Microbiol. 
Biotechnol. 59, 33-39. 

Boynton, Z.L., Bennett, G.N., Rudolph, F.B., 1996. Cloning, sequencing, and expression of 
clustered genes encoding beta-hydroxybutyryl-coenzyme A (CoA) dehydrogenase, 
crotonase, and butyryl-CoA dehydrogenase from Clostridium acetobutylicum ATCC 
824. J. Bacteriol. 178, 3015-3024. 

Cheong, S., Clomburg, J.M., Gonzalez, R., 2016. Energy- and carbon-efficient synthesis of 
functionalized small molecules in bacteria using non-decarboxylative Claisen con¬ 
densation reactions. Nat. Biotechnol. 34, 556-561. 

Clomburg, J.M., Vick, J.E., Blankschien, M.D., Rodriguez-Moya, M., Gonzalez, R., 2012. A 
synthetic biology approach to engineer a functional reversal of the {3-oxidation cycle. 
ACS Synth. Biol. 1, 541-554. 

Clomburg, J.M., Crumbley, A.M., Gonzalez, R., 2017. Industrial biomanufacturing: the 
future of chemical production. Science 355. 

Connor, M.R., Cann, A.F., Liao, J.C., 2010. 3-Methyl-1-butanol production in Escherichia 
coli: random mutagenesis and two-phase fermentation. Appl. Microbiol. Biotechnol. 
86, 1155-1164. 

Conrado, R.J., Gonzalez, R., 2014. Chemistry. Envisioning the bioconversion of methane 
to liquid fuels. Science 343, 621-623. 

Criddle, C.S., 2013. et al. (USA, vol. US20130071890 Al. 

Dellomonaco, C., Clomburg, J.M., Miller, E.N., Gonzalez, R., 2011. Engineered reversal of 
the {3-oxidation cycle for the synthesis of fuels and chemicals. Nature 476, 355-359. 
Demidenko, A., Akberdin, I.R., Allemann, M., Allen, E.E., Kalyuzhnaya, M.G., 2016. Fatty 
acid biosynthesis pathways in Methylomicrobium buryatense 5G(B1). Front. 
Microbiol. 7, 2167. 

Dittrich, C.R., Bennett, G.N., San, K.Y., 2005. Characterization of the acetate-producing 
pathways in Escherichia coli. Biotechnol. Prog. 21, 1062-1067. 

Dunlop, M.J., 2011. Engineering microbes for tolerance to next-generation biofuels. 
Biotechnol. Biofuels 4, 32. 

Espah Borujeni, A., Channarasappa, A.S., Salis, H.M., 2014. Translation rate is controlled 
by coupled trade-offs between site accessibility, selective RNA unfolding and sliding 
at upstream standby sites. Nucleic Acids Res. 42, 2646-2659. 

Fei, Q., et al., 2014. Bioconversion of natural gas to liquid fuel: opportunities and chal¬ 
lenges. Biotechnol. Adv. 32, 596-614. 

Fu, Y., Li, Y., Lidstrom, M., 2017. The oxidative TCA cycle operates during methano- 
trophic growth of the Type I methanotroph Methylomicrobium buryatense 5GB1. 
Metab. Eng. 42, 43-51. 

Garg, S., et al., 2016. Microbial production of bi-functional molecules by diversification of 
the fatty acid pathway. Metab. Eng. 35, 9-20. 

Gibson, D.G., et al., 2009. Enzymatic assembly of DNA molecules up to several hundred 
kilobases. Nat. Methods 6, 343-345. 

Gilman, A., et al., 2015. Bioreactor performance parameters for an industrially-promising 
methanotroph Methylomicrobium buryatense 5GB1. Microb. Cell Fact. 14, 182. 
Gilman, A., et al., 2017. Oxygen-limited metabolism in the methanotroph 
Methylomicrobium buryatense 5GB1C. PeerJ 5, e3945. 

Greenhouse Gas Emissions, 2017. Understanding Global Warming Potentials. 
Environmental Protection Agency. 

Hanson, R.S., Hanson, T.E., 1996. Methanotrophic bacteria. Microbiol. Rev. 60, 439-471. 
Hartmanis, M.G., Gatenbeck, S., 1984. Intermediary metabolism in Clostridium 


182 


S. Garg et al. 


Metabolic Engineering 48 (2018) 175-183 


acetobutylicum: levels of enzymes involved in the formation of acetate and butyrate. 
Appl. Environ. Microbiol. 47, 1277-1283. 

Hartmanis, M.G.N., Gatenbeck, S., 1984. Intermediary metabolism in Clostridium acet¬ 
obutylicum - levels of enzymes involved in the formation of acetate and butyrate. 
Appl. Environ. Microbiol. 47, 1277-1283. 

Haynes, C.A., Gonzalez, R., 2014. Rethinking biological activation of methane and con¬ 
version to liquid fuels. Nat. Chem. Biol. 10, 331-339. 

Henard, C.A., et al., 2016. Bioconversion of methane to lactate by an obligate metha- 
notrophic bacterium. Sci. Rep. 6, 21585. 

Henard, C.A., Freed, E.F., Guamieri, M.T., 2015. Phosphoketolase pathway engineering 
for carbon-efficient biocatalysis. Curr. Opin. Biotechnol. 36, 183-188. 

Henard, C.A., Smith, H.K., Guamieri, M.T., 2017. Phosphoketolase overexpression in¬ 
creases biomass and lipid yield from methane in an obligate methanotrophic bioca¬ 
talyst. Metab. Eng. 41, 152-158. 

Hu, B., et al., 2016. Comprehensive molecular characterization of Methylobacterium 
extorquens AMI adapted for 1-butanol tolerance. Biotechnol. Biofuels 9, 84. 

Hwang, J.W., Choi, Y.B., Park, S., Choi, C.Y., Lee, E.Y., 2007. Development and mathe¬ 
matical modeling of a two-stage reactor system for trichloroethylene degradation 
using Methylosinus trichosporium OB3b. Biodegradation 18, 91-101. 

Kaluzhnaya, M., et al., 2001. Taxonomic characterization of new alkaliphilic and alka- 
litolerant methanotrophs from soda lakes of the Southeastern Transbaikal region and 
description of Methylomicrobium buryatense sp.nov. Syst. Appl. Microbiol. 24, 
166-176. 

Kalyuzhnaya, M.G., et al., 2013. Highly efficient methane biocatalysis revealed in a 
methanotrophic bacterium. Nat. Commun. 4, 2785. 

Kalyuzhnaya, M.G., Puri, A.W., Lidstrom, M.E., 2015. Metabolic engineering in metha¬ 
notrophic bacteria. Metab. Eng. 29, 142-152. 

Keasling, J.D., 2012. Synthetic biology and the development of tools for metabolic en¬ 
gineering. Metab. Eng. 14, 189-195. 

Khmelenina, V.N., et al., 2013. Draft genome sequence of Methylomicrobium buryatense 
strain 5G, a haloalkaline-tolerant methanotrophic bacterium. Genome Announc. 1. 

Kim, S., Clomburg, J.M., Gonzalez, R., 2015. Synthesis of medium-chain length (C6-C10) 
fuels and chemicals via {5-oxidation reversal in Escherichia coli. J. Ind. Microbiol. 
Biotechnol. 42, 465-475. 

Kim, S., Cheong, S., Gonzalez, R., 2016. Engineering Escherichia coli for the synthesis of 
short- and medium-chain a,{5-unsaturated carboxylic acids. Metab. Eng. 36, 90-98. 

Kumari, S., Tishel, R., Eisenbach, M., Wolfe, A.J., 1995. Cloning, characterization, and 
functional expression of acs, the gene which encodes acetyl coenzyme A synthetase in 
Escherichia coli. J. Bacteriol. 177, 2878-2886. 

Lan, E.I., Liao, J.C., 2012. ATP drives direct photosynthetic production of 1-butanol in 
cyanobacteria. Proc. Natl. Acad. Sci. USA 109, 6018-6023. 

Lin, H., Castro, N.M., Bennett, G.N., San, K.Y., 2006. Acetyl-CoA synthetase over¬ 
expression in Escherichia coli demonstrates more efficient acetate assimilation and 
lower acetate accumulation: a potential tool in metabolic engineering. Appl. 
Microbiol. Biotechnol. 71, 870-874. 

Mehta, P.K., Ghose, T.K., Mishra, S., 1991. Methanol biosynthesis by covalently im¬ 
mobilized cells of Methylosinus trichosporium: batch and continuous studies. 


Biotechnol. Bioeng. 37, 551-556. 

Nowroozi, F.F., et al., 2014. Metabolic pathway optimization using ribosome binding site 
variants and combinatorial gene assembly. Appl. Microbiol. Biotechnol. 98, 
1567-1581. 

Okamura, E., Tomita, T., Sawa, R., Nishiyama, M., Kuzuyama, T., 2010. Unprecedented 
acetoacetyl-coenzyme A synthesizing enzyme of the thiolase superfamily involved in 
the mevalonate pathway. Proc. Natl. Acad. Sci. USA 107, 11265-11270. 

Puri, A.W., et al., 2015. Genetic tools for the industrially promising methanotroph 
Methylomicrobium buryatense. Appl. Environ. Microbiol. 81, 1775-1781. 

Salis, H.M., Mirsky, E.A., Voigt, C.A., 2009. Automated design of synthetic ribosome 
binding sites to control protein expression. Nat. Biotechnol. 27, 946-950. 

Shams Yazdani, S., Gonzalez, R., 2008. Engineering Escherichia coli for the efficient 
conversion of glycerol to ethanol and co-products. Metab. Eng. 10, 340-351. 

Shen, C.R., et al., 2011. Driving forces enable high-titer anaerobic 1-butanol synthesis in 
Escherichia coli. Appl. Environ. Microbiol. 77, 2905-2915. 

Shiba, Y., Paradise, E.M., Kirby, J., Ro, D.K., Keasling, J.D., 2007. Engineering of the 
pyruvate dehydrogenase bypass in Saccharomyces cerevisiae for high-level produc¬ 
tion of isoprenoids. Metab. Eng. 9, 160-168. 

Singh, A., Cher Soh, K., Hatzimanikatis, V., Gill, R.T., 2011. Manipulating redox and ATP 
balancing for improved production of succinate in E. coli. Metab. Eng. 13, 76-81. 

Smanski, M.J., et al., 2014. Functional optimization of gene clusters by combinatorial 
design and assembly. Nat. Biotechnol. 32, 1241-1249. 

Stephanopoulos, G., 2002. Metabolic engineering by genome shuffling. Nat. Biotechnol. 

20 , 666 - 668 . 

Strong, P.J., Xie, S., Clarke, W.P., 2015. Methane as a resource: can the methanotrophs 
add value? Environ. Sci. Technol. 49, 4001-4018. 

Strong, P.J., Kalyuzhnaya, M., Silverman, J., Clarke, W.P., 2016. A methanotroph-based 
biorefinery: potential scenarios for generating multiple products from a single fer¬ 
mentation. Bioresour. Technol. 215, 314-323. 

Tollefson, J., 2016. 'Flaring' wastes 3.5% of world's natural gas. Nature. 

de la Torre, A., et al., 2015. Genome-scale metabolic reconstructions and theoretical in¬ 
vestigation of methane conversion in Methylomicrobium buryatense strain 5G(B1). 
Microb. Cell Fact. 14, 188. 

U.S. Natural Gas Vented and Flared, 2017. U.S. Department of Energy, Washington, DC. 

Vick, J.E., et al., 2015. Escherichia coli enoyl-acyl carrier protein reductase (FabI) sup¬ 
ports efficient operation of a functional reversal of {5-oxidation cycle. Appl. Environ. 
Microbiol 81, 1406-1416. 

Whittenbury, R., Phillips, K.C., Wilkinson, J.F., 1970. Enrichment, isolation and some 
properties of methane-utilizing bacteria. J. Gen. Microbiol. 61, 205-218. 

Yan, X., Chu, F., Puri, A.W., Fu, Y., Lidstrom, M.E., 2016. Electroporation-based genetic 
manipulation in type I methanotrophs. Appl. Environ. Microbiol. 82, 2062-2069. 

Ye, R.W., Kelly, K., 2012. Construction of carotenoid biosynthetic pathways through 
chromosomal integration in methane-utilizing bacterium Methylomonas sp. strain 
16a. Methods Mol. Biol. 892, 185-195. 

Zhang, B., et al., 2015. Ribosome binding site libraries and pathway modules for shikimic 
acid synthesis with Corynebacterium glutamicum. Microb. Cell Fact. 14, 71. 


183