ArticlePDF AvailableLiterature Review

Abstract and Figures

Natural gas is a mixture of low molecular weight hydrocarbon gases that can be generated either from fossil or anthropogenic resources. Although natural gas is used as a transportation fuel, constraints in storage, relatively low energy content (MJ/L), and delivery have limited widespread adoption. Advanced utilization of natural gas has been explored for biofuel production by microorganisms. In recent years, the aerobic bioconversion of natural gas (or primarily the methane content of natural gas) into liquid fuels by biocatalysts (methanotrophs) has gained increasing attention as a promising alternative for drop-in biofuel production. Methanotrophic bacteria are capable of converting methane into microbial lipids, which can in turn be converted into renewable diesel via a hydrotreating process. In this paper, biodiversity, catalytic properties and key enzymes and pathways of these microbes are summarized. Bioprocess technologies are discussed based upon existing literature, including cultivation conditions, fermentation modes, bioreactor design and lipid extraction and upgrading. This review also outlines the potential of Bio-GTL using methane as an alternative carbon source as well as the major challenges and future research needs of microbial lipid accumulation derived from methane, key performance index and techno-economic analysis. An analysis of raw material costs suggests that methane-derived diesel fuel has the potential to be competitive with petroleum-derived diesel.
Content may be subject to copyright.
Research review paper
Bioconversion of natural gas to liquid fuel: Opportunities and challenges
Qiang Fei, Michael T. Guarnieri, Ling Tao, Lieve M.L. Laurens, Nancy Dowe, Philip T. Pienkos
National Bioenergy Center, National Renewable Energy Laboratory, Golden, CO, USA
abstractarticle info
Article history:
Received 2 December 2013
Received in revised form 29 March 2014
Accepted 30 March 2014
Available online 12 April 2014
Keywords:
Bioconversion of natural gas into liquid fuels
(Bio-GTL)
Greenhouse gas
Renewable diesel fuel
Methanotrophic bacteria
Microbial lipids
Bioprocess optimization
Lipid extraction
Hydrotreati ng process
Techno-economic analysis
Natural gas is a mixture of low molecular weight hydrocarbon gases that can be generated from either fossil or
anthropogenic resources. Although natural gas is used as a transportation fuel, constraints in storage, relatively
low energy content (MJ/L), and delivery have limited widespread adoption. Advanced utilization of natural gas
has been explored for biofuel production by microorganisms. In recent years, the aerobic bioconversion of
natural gas (or primarily the methane content of natural gas) into liquid fuels (Bio-GTL) by biocatalysts
(methanotrophs) has gained increasing attention as a promising alternative for drop-in biofuel production.
Methanotrophic bacteria are capableof converting methaneinto microbial lipids,which can in turn be converted
into renewable diesel via a hydrotreating process. In this paper, biodiversity, catalytic properties and key en-
zymes and pathways of these microbes are summarized. Bioprocess technologies are discussed based upon
existing literature, including cultivation conditions, fermentation modes, bioreactor design, and lipid extraction
and upgrading. This review also outlines the potential of Bio-GTL using methane as an alternative carbon source
as well as the major challenges and futureresearch needs of microbiallipid accumulation derived from methane,
key performance index, and techno-economic analysis. An analysis of raw material costs suggests that methane-
derived diesel fuel has the potential to be competitive with petroleum-derived diesel.
© 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Contents
Introduction ................................................................ 597
Naturalgasstatus ............................................................ 597
Biologicalconversionofnaturalgastoliquidfuel(Bio-GTL) ......................................... 597
Applicationsofmethanotrophs ...................................................... 598
Metabolicpathwaysandenzymesofmethanotrophs.............................................. 599
Methanotrophsandspecies........................................................ 599
Carbonassimilationpathwaysandkeyenzymes .............................................. 600
Developmentofgenetictools....................................................... 603
ChallengesinBio-GTLprocess ........................................................ 604
Optimizationofculturemedium...................................................... 604
Optimizationofcultureconditions..................................................... 605
Masstransferenhancementandbioreactordesign.............................................. 605
Bioprocessdevelopment ......................................................... 606
Lipidextractionprocessdevelopment ................................................... 606
Hydrotreatingprocess .......................................................... 607
Economicconsiderations........................................................... 608
Futureprospectsandconclusions....................................................... 610
Acknowledgments ............................................................. 610
References................................................................. 610
Biotechnology Advances 32 (2014) 596614
National Bioenergy Center, National Renewable Energy Laboratory, 15013 Denver West Parkway, Golden, Colorado, 80401, USA. Tel.: +1 303 384 6269; fax: +1 303 384 6363.
E-mail address: Philip.Pienkos@nrel.gov (P.T. Pienkos).
http://dx.doi.org/10.1016/j.biotechadv.2014.03.011
0734-9750/© 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CCBY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
Contents lists available at ScienceDirect
Biotechnology Advances
journal homepage: www.elsevier.com/locate/biotechadv
Introduction
Natural gas status
With the uncertainty of available petroleum reserves and increased
demands for energy resources, renewable fuels have drawn great
attention in recent years to avoid crude oil exhaustion (Chisti, 2007).
Although multiple technologies have been studied and developed to
provide alternativefuels, the gap between supply and demandfor liquid
fuel production is estimated to signicantly increase in the next decades
(Nashawiet al., 2010; Zitteland Schindler, 2007). Naturalgas has played
a vital role of the world's supply of energy for years. As a fossil fuel,
natural gas is commonly used as an energy source for heating, cooking,
and electricity generation. In general, natural gas is colorless and
odorless in its pure form and it exists as a combustible mixture of
several hydrocarbon gases, which often contains about 8095% (v/v)
methane mixed with other heavier alkanes such as ethane, propane,
butane and pentane. Most natural gas is drawn from wells or extracted
in conjunction with crude oil production. Unwanted natural gas associ-
ated with oil extraction is often injected back to the reservoir while
awaiting a possible future market or to repressurize the well, which
can enhance oil extraction efciency. Natural gas has been considered
as one of the cleanest alternative fuels for transportation vehicles
(EPA, 2002). There are about 15 million natural gas vehicles in use
worldwide. However, natural gas-powered vehicles do not play a
dominant role globally and achieved even lower acceptance in the US
compared with vehicles powered by liquid fuels. Nowadays, only
roughly 112,000 vehicles are powered by natural gas in the US (DOE,
2014), despite the abundance of natural gas sources found recently.
This reects the chicken/egg conundrum observed with a number of
alternative fuel sources that is caused by the relative lack of infrastruc-
ture for fueling stations that discourages purchase of such vehicles and
the limited number of natural gas-compatible vehicles (EIA, 2012b),
which discourages the investment in infrastructure. A potential solution
to this stalemate can be found in the exploration of approaches to
convert natural gas into liquid transportation fuels.
One of the major obstacles for theproduction of renewable fuels will
be the supply of feedstock (EIA, 2012b). Conversely, more than
1.1 × 10
14
ft
3
(1 ft
3
gas is equal to 1000 Btu and 0.008 GGE) of natural
gas has been produced per year since 2009. As can be seen in Fig. 1,
nearly 3 × 10
13
ft
3
of natural gas was produced in the US in 2012,
representing a 30% increase since 2002. The International Energy Agen-
cy estimates that the production of natural gas will keep increasing,
with 25% of global energy derived from natural gas by 2035. Neverthe-
less, natural gas at a level of 5 quadrillion BTU of fossil fuel energy
(about 5% of the annual production) is currently ared or vented at
many places around the globe, carrying along with it the associated
greenhouse gas contributions (World_Bank, 2013). This unutilized gas
presents a $13 billion per year market value and is equivalent to 27%
US electricity production. This potentially valuable resource is wasted
because the development of the pipelines and processing facilities
needed to handle the unwanted natural gas has not kept pace with pro-
duction and there is no economic incentive to capture it at present.
However, with the anticipated depletion of liquid petroleum and the
volatility in the price of crudeoil, the utilization of natural gas as a feed-
stock for production of liquid fuels has drawn great attention in recent
years.
Due to the development of shale gas production, the cost of natural
gas has been reduced signicantly, which in turn has made natural gas
a more attractive choice for liquid fuel and chemical production when
compared with other high price raw material sources (Borgwardt,
1997; Dong and Steinberg, 1997; G. Liu et al., 2011; Li et al., 2010). As
shown in Fig. 2, there has been a tremendous increase in natural gas
production in the US since the beginning of shale gas production. Due
to the high hydrogen/carbon ratio characteristic of natural gas, it is a
very promising feedstock for liquid fuel production. The high ratio also
can help save the capital investment required to generate liquid
fuel by increasing the overall yield of carbon during the liquid fuel
production. Therefore, natural gas is an abundant and ideal alternative
source of low cost carbon source for fuel production.
The technology of the conversion of natural gas into hydrocarbon
liquid fuels has been extensively researched and developed for many
years. On a global scale, exploration of this technology has expanded
even more in recent years, since more natural gas has been found in
remote sites where gas pipelines may not be economically justied
yet. Recently, FischerTropsch (FT) technology has gathered increased
attention for the conversion of natural to liquid products (Dry, 2002;
Schulz, 1999). However, the gas to liquid (GTL) technology requires
syngas generation, syngas conversion, and hydroprocessing. Besides
the technical hurdles, the FT process also requires a very large scale
for successful commercialization, mainly due to the requirements for
large production facilities and sustainable gas supply (Vosloo, 2001).
During the syngas steps, large quantities of energy and heat are
involved, limiting the energy efciency of this process (Hall et al.,
2000, 2003). Furthermore, conventional FT technology can only achieve
carbon conversion efciency (CCE) from 25 to 50% (Steynberg and Dry,
2004). Therefore alternative GTL processes, capable of providing high
liquid production yield with higher CCE and lower energy or heat
input, bear evaluation.
Biological conversion of natural gas to liquid fuel (Bio-GTL)
Methane can be chemically oxidized to methanol by using various
catalysts (Ab Rahim et al., 2013; Benlounes et al., 2008; Hall et al.,
Fig. 1. The US natural gas gross withdrawals from 1940 to 2012.
Data collected from U.S. Energy Information Administration.
597Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
1995; Hammond et al., 2012; Periana et al., 1993). The processes of
converting natural gas to methanol have been proposed and studied
for decades, whereas all the chemical conversion processes are energy
intensive, noneconomic, and environmentally unfriendly (Adebajo
and Frost, 2012; Foster, 1985; Muehlhofer et al., 2002). In contrast, a
biological methane conversion process targeting valuable compounds,
such as next generation fuels or chemical products, is being discussed
in the scientic and industrial community (Conrado and Gonzalez,
2014; DOE, 2013). This new technology for effective conversion of
natural gas to liquid fuels or other chemical products is capable
of transforming the natural gas industry using the signicant portion
of the world's reserves of natural gas in remote regions (Conrado and
Gonzalez, 2014). Moreover, biotechnological production of fuel or
chemical products currently depends mainly on high cost sugars, such
as glucose, as an energy and carbon source, the cost of which is estimat-
ed to be about 50% of the nal products (Chang et al., 2011; Fei et al.,
2011a, 2011b). Therefore, considerable efforts to identify alternative
carbon sources, aimed at minimizing the production cost of fuels and
chemicals, are currently under way. Methane, the major component of
natural gas has been employed for the production of several chemical
products on an industrial scale as a possible alternative carbon source
for sugar-based biochemical processes (Sheldon, 2014; Thayer, 2013).
The core of the Bio-GTL technology is based upon a unique and fast-
growing class of microorganisms, namely aerobic methanotrophic
bacteria, which are able to convert methane to biomass in a well-
established natural process (Buswell and Sollo, 1948; Hamer et al.,
1967; Jiang et al., 2010).
Methanotrophic bacteria are a subset of a physiological group of
bacteria known as methylotrophs that are characterized by their ability
to utilize a variety of different C
1
substrates including methane, metha-
nol, methylated amines, halomethanes, and methylated compounds
containing sulfur (Anthony, 1982, 1986; Dijkhuizen et al., 1992;
Hanson et al., 1990) as the carbon and energy sources (Buswell and
Sollo, 1948; Hamer et al., 1967; Hanson and Hanson, 1996; Jiang et al.,
2010; Schrader et al., 2009). Methanotrophs areusually specied as aer-
obic microorganisms that can oxidize methane to serve as both energy
and carbon sources. The microbial lipid produced by methanotrophs
in aerobic cultivation could be utilized as a fuel precursor in a hydro-
treating process for the production of renewable diesel (Knothe, 2010;
Sunde et al., 2011), which is compatible with current infrastructure
and modern vehicles. A Bio-GTL process combining the technologies
of biocatalysts, synthetic biology, and advanced bioengineering process
design could operate more inexpensively and efciently than the
traditional chemical processes, providing both economic and environ-
mental advantages for liquid fuel production. Given the rapidly increas-
ing cost of fossil fuels, and food vs. fuel complications associated with
biomass-derived fuels, liquid fuels derived fromnatural gas offer a num-
ber of unique advantages. In contrast to current methods of fossil fuel
production, a methane-based liquid fuel production platform, applying
a scalable, modular, low complexity (preferably low temperature/low
pressure), low environmental impact, and low-cost GTL technology,
has the potential to compete with petroleum-based fuels (Pimentel,
2008). However successful implementation and scale-up
of commercial methane-based production processes remain to be
developed to drive the Bio-GTL technology faster and more efciently
than possible with FT processes.
Applications of methanotrophs
The products from this methane-based bioconversion process are
dependent upon the type of methanotrophic bacteria, the enzymes
employed for the oxidation of methane, and the metabolic pathways
governing the synthesis of intermediates of central metabolic routes.
To date, most of the studies of methanotrophs took place during the
development of production processes for biopolymers, vitamins, anti-
biotics, single cell protein (SCP) and carboxylic acids (Anthony, 1982;
Bothe et al., 2002; Ivanova et al., 2006; Pieja et al., 2011; Reshetnikov
et al., 2005; Tao et al., 2007; Zhang et al., 2008). A biopolymer, poly-β-
hydroxybutyrate (PHB), which can be produced by methanotrophs,
shows great promise for biomedical applications, biodegradable food
packaging, and encapsulation of agrochemicals on a large scale (Fei
et al., 2009; Gürsel and Hasirci, 1995; Shang et al., 2009, 2012;
Valappil et al., 2007). A maximum PHB content of 70% (w/w) was
obtained in a culture of Methylocystis parvus (Asenjo and Suk, 1986).
More recent efforts in methanotroph cultivation were focused upon
developing the processes for the conversion of methane into the SCP
on an industrial scale, a concept rst explored three decades ago.
UNIBIO, a company from Denmark, has demonstrated commercial
production of SCP as a nutritional food protein feed for animals
(UNIBIO, 2011).
However, until now there have been few reports discussing the
feasibility of utilizing the Bio-GTL concept for the production of liquid
fuels. This was due to limitations in strain development approaches
caused by the incomplete understanding of the metabolism of
methanotrophs and a lack of high-efciency tools and approaches to
be applied for genetic manipulation of these organisms. Although
there has been a rapid expansion of genetic and metabolic knowledge
of methanotrophs within the past decade (Jahnke et al., 1999;
Kaluzhnaya et al., 2001; Khmelenina et al., 1997), only a few studies
have been reported to use methanotrophic bacteria as biocatalysts for
lipid or fuel production through the Bio-GTL process recently
(Conrado and Gonzalez, 2014; Culpepper and Rosenzweig, 2012;
Fig. 2. The US natural gas gross withdrawals from shale gas well from 2007 to 2011.
Data collected from U.S. Energy Information Administration.
598 Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
Kalyuzhnaya et al., 2013). The reason for this is that a robust, suitable
production strain had not yet been identied. Recently, a $4 million
Advanced Research Projects Agency Energy (ARPA-E) award from
U.S. Department of Energy (DOE, Washington DC, USA) was granted
to a group led by the University of Washington (UW, Seattle, WA,
USA) to develop microbes that can convert methane into liquid diesel
fuel for transportation in 2012 (UW_Daily, 2012). The National Re-
newable Energy Laboratory (NREL, Golden, CO, USA), LanzaTech, Inc.,
a biofuel company (Chicago, IL, USA), and Johnson Matthey (London,
UK), a chemical company, have joined with UW to develop a more
efcient bioconversion process for liquid fuel production. Moreover,
ARPA-E issued a $30 million Funding Opportunity Announcement of
reducing emission using methanotrophic organisms for the transporta-
tion energy (REMOTE) in the beginning of 2013 (DOE, 2013) to fund
additional works on the development of bioconversion technologies to
turn methane into liquid fuels. The goal of this REMOTE project is to
identify transformational concepts for cost-effective, one-step conver-
sion of methane to liquid transportation fuels with low capital expendi-
ture and exible deployment to access remote, ared, or pipeline gas.
The application of methanotrophic bacteria for the production of
fatty-acid derived fuels is a novel initiative. The envisaged process
downstream of cultivation of the biomass involves a lipid extraction
process, most likely solvent-based, after which the lipid fraction will
be subjected to catalytic upgrading and conversion to renewable diesel.
For some methanotrophs, the microbial lipid fraction in the biomass can
be more than 20% on a dried cell weight basis (Kaluzhnaya et al., 2001)
and potentially even higher depending on the cultivation conditions of
the biomass and the contribution of the membrane fraction to the
whole biomass. Even though the lipid fraction relative to the biomass
can be high, the composition of the extractable lipid fraction is very
different than the lipids typically sought for biofuels from algae or
oleaginous yeasts, i.e. rich in triglycerides (Fang et al., 2000; Halim
et al., 2012). Lipid fractions typically used for biodiesel or renewable/
green diesel production start with a triglyceride-rich lipid stream,
whereas the lipid composition in methanotrophs consists mainly of
phospholipids, originating from intracytoplasmic membranes. The
membrane lipids represent a major fraction of the biomass and
consist mainly of phosphatidylglycerol (PG), phosphatidylcholine
(PC), and phosphatidylethanolamine (PE) (Fang et al., 2000; Oliver
and Colwell, 1973). This composition, in particular the presence of
high levels of heteroatoms (specically P and N) in the lipid fraction,
presents challenges to the development of suitable extraction and con-
version of lipids for downstream technologies.
Downstream processing of crude extracted lipids for renewable
diesel production follows a sequential catalytic upgrading process of
hydrotreating (hydroprocessing) followed by catalytic cracking and
isomerization of the fatty acyl chains to t the properties of renewable
diesel. The hydrotreating process, which has been used by many
petroleum reneries to remove the sulfur, nitrogen, condensed ring
aromatics or metals for the production of fossil fuel, has more recently
been employed to convert lipids into renewable diesel or jet fuel
(Knothe, 2010; Sunde et al., 2011). The resulting products are aliphatic
hydrocarbons with chain lengths similar to those found in petroleum
diesel as well as a major byproduct propane derived from the glycerol
backbone of triacylglyceride lipids (Hodge, 2006; Nylund et al., 2011).
The renewable diesel is chemically similar to traditional diesel (though
lacking in aromatic compounds) and able to blend with any petroleum
diesel in any proportion. This material differs from biodiesel in being a
hydrocarbon and lacking oxygen in the fatty acid methyl ester bond.
The added advantage of this process is that renewable diesel blends
can be produced in existing reneries by co-processing the feedstock
with petrodiesel. Renewable diesel fuels can be employed in present
pipelines, tanks, pumps, and vehicles without infrastructure changes.
Because of its aliphatic nature, renewable diesel fuel has higher energy
content than petrodiesel and biodiesel. It also has excellent oxidative
stability (because double bonds are reduced) and cold ow properties
(depending on the feedstock fatty acid prole, some isomerization
may be needed) and should contain little or no sulfur or nitrogen so
there is no concern with SOx and NOx emissions compared to
petrodiesel. A detailed comparison of different fuel properties has listed
in Table 1. It is clear that renewable diesel exhibits much better proper-
ties of cetane number, cold ow, and oxidative stability compared to the
other two diesel fuels examined. The primary advantages over
petrodiesel and biodiesel produced by conventional means are superior
cold weather properties, greater energy content, higher cetane number,
and economical capital and operating costs (Holmgren et al., 2006,
2007).
Metabolic pathways and enzymes of methanotrophs
Methanotrophs and species
The capability of microorganisms to utilize methane as the sole
carbon source for their growth was rst discovered in 1906 (Söhngen,
1906). Whittenbury et al. (1970) isolated and characterized over 100
new methane-utilizing bacteria that established the basis of the current
classication of methanotrophic bacteria. Methanotrophic bacteria are
usually considered as a unique family of gram-negative bacteria
(Anthony, 1982) and most were isolated from sewage, bogs, wetlands,
lake basins or ruminants (Hanson and Hanson, 1996). Recently, some
methanotrophic bacteria belonging to the Verrucomicrobia phylum
have been found and isolated (Duneld et al., 2007; Islam et al., 2008;
Pol et al., 2007). Although methanotrophs are now known to be
able to grow in both aerobic and anaerobic environments, most
methanotrophic bacteria have been isolated in pure aerobic cultures,
and this review will focus on this aerobic methanotrophic bacteria
(Lidstrom, 2006; McDonald et al., 2008; Murrell and Jetten, 2009).
Anaerobic methanotrophs capable of utilizing sulfate or nitrate as
electron donors have been discovered in anoxic marine water,
sediments of soda lakes, and freshwater sediments (Boetius et al.,
2000; Raghoebarsing et al., 2006). The microbially mediated anaerobic
methane oxidation cycle is the major biological sink of methane in
marine sediments, which has been estimated to consume 2 × 10
9
kg
of methane per year (Alperin and Reeburgh, 1984). Since pure cultures
capable of anaerobic methane oxidation are very difcult to character-
ize, the biochemistry of the process and products from this process re-
main to be described (Hanson and Hanson, 1996; Hinrichs and
Boetius, 2002). However, several hypotheses of various pathways
were proposed and discussed based on different lines of evidence
(Hinrichs and Boetius, 2002). Recently, the cultivation of methanotrophic
bacteria of the NC10 phylum has been described successfully, which
helps researchers elucidate and extend the understanding and
knowledge of anaerobic methane oxidizing microorganisms (Ettwig
et al., 2008, 2009, 2010; Hu et al., 2009).
Aerobic methanotrophs are now separated into two major groups
instead of using the terms of Type I, Type II, and Type Xthat were
Table 1
Comparison of different diesel fuel properties (Holmg ren et al., 2006, 2007;
UOP_Honeywell, 2013).
Specications Renewable diesel Petrodiesel Biodiesel
Cetane number 7590 4055 5065
%O atom 0 0 11
Net heating value (MJ kg
1
)44 43 38
Energy content (BTU gal
1
) 125 k 130 k 115 k
Density (kg/L) 0.78 0.84 0.88
Sulfur content b10 ppm b10 ppm b5ppm
Cloud point (°C) 30 to 555to20
NOx emission 10 to 0 Baseline +10
Cloud ow properties Excellent Baseline Poor
Oxidative stability Excellent Baseline Poor
599Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
designated based upon the physiology and phylogeny (Op den Camp
et al., 2009): Group I methanotrophs are Gammaproteobacteria (formerly
Type I and X), including the genera Methylococcus,Methylomonas,
Methylosphaera,Methylosoma,Methylomicrobium,Methylothermus,
Methylohalobius,Methylosarcina,andMethylobacter, and utilize the
ribulose monophosphate (RuMP) cycle for single carbon assimilation
(Ballows et al., 1992; Bowman et al., 1995; Bratina et al., 1992;
Bulygina et al., 1990; Hanson et al., 1993; Semrau et al., 1995;
Stanley et al., 1983; Stolyar et al., 1999; Tsuji et al., 1989); Group II
methanotrophs are Alphaproteobacteria (formerly Type II), including
the genera Methylosinus,Methylocapsa,Methylocella,andMethylocystis,
and use the serine cycle to assimilate single carbon sources (Bastien et al.,
1989; Cardy et al., 1991a,b; Fox et al., 1989; Gilbert et al., 2000; Graham
et al., 1993; Nichols et al., 1987; Whittenbury and Dalton, 1981); novel
thermoacidophilic aerobic methanotrophs Verrucomicrobia have
been found recently, which can assimilate carbon at the level of CO
2
using the Calvin Benson Bassham cycle (Duneld et al., 2007; Khadem
et al., 2011; Pol et al., 2007). One thing all methanotrophs have in
common is their traditional restricted metabolic capability; these
strains can only use methane and methanol as the carbon and energy
sources. However, two novel facultative species, Methylocella and
Methylocystis were discovered and isolated recently, which are able to
use methane and methanol as well as acetate and ethanol as their sole
energy source (Dedysh et al., 2005; Im et al., 2011). Frequently used
methanotrophic microorganisms and their optimum growth conditions
and potential fuel and chemical products are listed in Table 2.
Carbon assimilation pathways and key enzymes
By using the unique isoenzymes of methane monooxygenase
(MMO), methanotrophic bacteria can utilize C1 sources more
reduced than formic acid as sources of carbon and energy and assim-
ilate the carbon into biomass at the level of formaldehyde or formate
(Anthony, 1982, 1986; Buswell and Sollo, 1948; Dijkhuizen et al.,
1992; Hamer et al., 1967; Hanson et al., 1990; Jiang et al., 2010). As
shown in Fig. 3, the common features of methanotroph metabolism
include unique pathways allowing for the synthesis of intermediates
of central metabolic routes, and the central metabolic intermediate of
formaldehyde in both catabolism and anabolism. Methanotrophic
bacteria use MMOs to convert methane to methanol and then further
oxidize methanol to formaldehyde by using a pyrroloquinoline quinone
(PQQ)-dependent enzyme, methanol dehydrogenase. Because the
methane oxidation step requires energy in the form of NADH, the cells
break evenmetabolically with the oxidation of methanol to formalde-
hyde. Surplus energy is generated in the subsequent oxidation steps
the conversion of formaldehyde to formate and formate to carbon
dioxide through the action of formaldehyde dehydrogenase and
formate dehydrogenase, respectively.
Two types of MMO enzyme systems are known (Anthony, 1986;
Dijkhuizen et al., 1992; Hanson et al., 1990; Park and Lee, 2013): a sol-
uble, cytoplasmic complex (soluble methane monooxygenase, sMMO)
and a membrane-bound, particulate system (particulate methane
monooxygenase, pMMO). The genes for pMMO are organized in
the pmoCAB operon in which pmoB,pmoA,andpmoC encode poly-
peptides corresponding to pMMO subunits, α,β,andγ, respectively
(Semrau et al., 1995). The genes for sMMO are composed of the
mmoXYBZDC operon (Stainthorpe et al., 1989 and 1990). The enzyme
of pMMO has higher afnity for methane compared to sMMO and it is
the predominant methane oxidation catalyst in nature (Leak, 1992;
Leak et al., 1985; Lipscomb, 1994).
There are two separate cycles by which formaldehyde produced as
an intermediate in methane oxidation is assimilated into microbial
biomass (Figs. 4 and 5). The ribulose monophosphate (RuMP) cycle is
usually found in Gammaproteobacterial methanotrophs that have
incompletely functioning TCA cycles and Type I internal membranes
(Hanson and Hanson, 1996; Jiang et al., 2010). Hexulose-6-phosphate
synthetase and hexulose-6-phosphate isomerase are two unique
enzymes of the RuMP cycle. The RuMP cycle, rst described by Quayle
and his colleagues (Large et al., 1961, 1962; Strøm et al., 1974)consists
of three main parts xation, cleavage and rearrangement (Fig. 4).
Fixation involves the aldol condensation of formaldehyde with D-
ribulose-5-phosphate to form 3-hexulose 6-phosphate, which is then
isomerized to fructose 6-phosphate (FMP) (Anthony, 1982). Cleavage
entails the conversion of FMP to either fructose 1,6-bisphosphate
(FBP) by phosphofructokinase, or 2-keto-3-deoxy-6-phosphogluconate
(KDPG) by the EntnerDoudoroff enzymes (White et al., 2007); these
molecules (FBP and KDPG) are then cleaved by aldolases to glyceralde-
hyde 3-phosphate along with dihydroxyacetone phosphate (DHAP)
from FBP or pyruvate from KDPG. For every three molecules of formal-
dehyde that are condensed, one molecule of FMP is cleaved, and thus
one molecule of pyruvate or DHAP is routed to biosynthesis. In the
nal step of the RuMP cycle (rearrangement), the remaining two
molecules of FMP and the glyceraldehyde phosphate undergo a series
of reactions. For example, the glyceraldehyde 3-phosphate (GAP) that
is left from the cleaved FMP can react with one of the other two FMP
molecules to form xylulose-5-phosphate (XuMP) and erythrose-
4-phosphate (EMP). EMP then reacts with the third FMP to form
septulose-7-phosphate (SMP) and GAP. These two compounds are the
substrate for another aldolase, which generates XuMP and a ribose-
5-phosphate (RiMP). The net result of these rearrangements reactions
is the production of two molecules of XuMP and one molecule of
RiMP, all of which are then converted back to RuMP by ribulose-
phosphate 3-epimerase and ribose-5-phosphate isomerase, thus
closing the cycle (Anthony, 1991; Quayle, 1980).
Cells using the serine cycle have functioning TCA cycles and Type II
internal membranes (Hanson and Hanson, 1996; Jiang et al., 2010;
Table 2
Examples of methanotrophic microorganisms, and their optimum growth conditions.
Species T
opt
pH
opt
Products Taxonomy
a
Reference
Methylomonas pelagica 25 °C 6.57.0 PHB Group I Bowman et al. (1993)
Methylosinus trichosporium OB3b 30 °C 6.87.2 PHB Group II Shah et al. (1996)
Methylococcus capsulatus (Bath) 45 °C 6.8 Lipids Group I Stanley et al. (1983)
Pseudomonas methanica 30 °C 6.9 EPS Group I Huq et al. (1978)
Methylobacter modestohalophilus 10s 25 °C 6.5 Cell mass Group I Kalyuzhnaya et al. (1998)
Methylomicrobium species 4G 30 °C 8.08.5 Cell mass Group I Kaluzhnaya et al. (2001)
Methylomicrobium alcaliphilum 20Z (Methylobacter alcaliphilus) 30 °C 9.0 Ectoine Group I Kalyuzhnaya et al. (2008),
Khmelenina et al. (1997)
Methylobacter album (Methylococcus albus) 30 °C 7.0 Cell mass Group I Bowman et al. (1993)
Methylocystis sp. 37 °C 5.7 PHB Group II Wendlandt et al. (2005)
Methylosinus trichosporium 30 °C 7.0 PHB Group II Zhang et al. (2008)
Methylosinus trichosporium OB3b 30 °C 7.0 Epoxides Group II Clark and Roberto (2003)
Methylomonas sp. 16a 30 °C 7.0 Astaxanthin Group I Rick et al. (2007)
Methylosinus trichosporium IMV 3011 30 °C 7.0 Methanol Group II Xin et al. (2004)
a
Group I methanotrophs are Gammaproteobacterial methanotrophs and Group II methanotrophs are Alphaproteobacterial.
600 Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
Matsen et al., 2013). The serine cycle operates to synthesize one
molecule 3-phosphoglycerate from 1.5 molecules of formaldehyde
plus 1.5 molecules of CO
2
(Anthony, 2011; Methanotrophy_Consortium,
2014). In this process, three molecules of formaldehyde and three of
glyoxylate give two molecules of 2-phosphoglycerate. One is assimilated
into cell material by 3-phosphoglycerate, while the other is converted to
phosphoenolpyruvate, whose carboxylation yields oxaloacetate and
subsequently malyl-CoA. In the end of the serine cycle, this malyl-CoA is
cleaved by malyl-CoA lyase into glyoxylate and acetyl-CoA. The oxidation
of acetyl-CoA to glyoxylate via the ethylmalonyl-CoA pathway completes
this glyoxylate cycle (Schneider et al., 2012). Two variants of the serine
cycle differ in how the generation of glyoxylate from acetyl-CoA is
accomplished. In some methanotrophs having the icl
+
-serine cycle, the
serine cycle for oxidation of acetyl-CoA cycles twice (Fig. 5). Isocitrate
lyase is involved rstly in the route of the oxidation of acetyl-CoA to
glyoxylate. This enzyme converts D-threo-isocitrate to form succinate
and glyoxylate. The oxaloacetate is condensed with acetyl-CoA to form
citrate by malate synthase, which eventually gets cleaved to succinate
for cell biosynthesis. Another variant, icl-serine cycle, is known in
methylotrophic bacteria such as Methylobacterium and Hyphomicrobium
strains that have no isocitrate lyase, and the route for oxidation of
acetyl-CoA to glyoxylate has not been completely elucidated (Anthony,
1982). Some Gammaproteobacterial methanotrophs (formerly Type I)
not only use RuMP cycle but also possess genes for serine cycle enzymes
(Lieberman and Rosenzweig, 2004). Recently, Conrado and Gonzalez
supposed several alternative routes for methane activation and biocon-
version pathways to improve the synthesis efciency for the production
of liquid fuels (Conrado and Gonzalez, 2014). By constructing a
dioxygenase-like enzyme, 80% energy efciency could be achieved via
an engineered methane activation pathway.
Most methanotrophs are believed to be capable of pMMO expression
when grown in the presence of copper (Dalton, 1992; Duneld et al.,
2003), but the ability to form sMMO is dominant in copper-limited
environments. The sufciency of copper leads to the synthesis of addi-
tional intracellular membranes (ICMs), appearance of the membrane
proteins associated with pMMO (Collins et al., 1991), increased CCE
and growth yield, and a loss of sMMO activity (Leak and Dalton,
1986a,b; Leak et al., 1985). Because of the formation of extensive ICM,
methanotrophs with pMMO have a relatively high lipid content
compared toother methanotrophs (and indeed to most other bacteria).
Thus, methanotrophic bacteria are among a very small subset of micro-
organisms that have a metabolic potential for the accumulation of lipids
suitable for liquid fuels, and they are the only microbial system that can
drive such production using methane as a sole carbon source.
Although agricultural oil feedstocks such as soybean, sunower,
jatropha, rape seed, and palm kernels fulll a small fraction of the liquid
fuel requirement, the enormous volumes of transportation fuels burned
(about 50% of the world's energy consumption) make it easy to see
that the use of liquid fuel from biomass can be expanded substantially
(Ezeji et al., 2007; Fortman et al., 2008). The recent focus on the food
Fig. 3. Simplied pathway for the oxidation of methane and assimilation o f formaldehyde. Major enzymes a re presented in green. Abbrevia tions: pMMO, particulate methane
monooxygenase; sMMO, soluble methane monooxygenase; PQQ, pyrroloquinoline quinone; MDH, methanol dehydrogenase; H
4
MPTP, methylene tetrahydromethanopterin pathway;
FDH, formate dehydrogenase.
Fig. 4. Simplied molecular pathway of ribulose monophosphate (RuMP) cycle for methanotrophic bacteria. Major enzymes are presented in green. Dash arrows represent possible exit
points of metabolites for biosynthetic reactions. Abbreviations: MMO, methane monooxygenase; MDH, methanol dehydrogenase; H
4
MPTP, methylene tetrahydromethanopterin path-
way; FDH, formate dehydrogenase; H 6PI, hexulose-6-phosphate isomerase; GPI, glucose phosphate isomerase; G6PDH, glucose-6-phosphate dehydrogenase; 6PGDH, 6-
phosphogluconate dehydrogenase; H6PS, hexulose-6-phosphate synthetase.
601Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
vs. fuel debate has largely caused a decline in interest in converting
edible oils and an increase in interest in non-edible oils like waste grease
and animal fats as feedstocks for biofuels. Because these sources are not
available in high volume, attention has turned to other non-edible lipid
sources from microorganisms such as oleaginous yeasts and microalgae,
where the primary class of lipids is triglycerides (like vegetable oils and
animal fats), with some contribution from polar lipids (Greenwell et al.,
2010; Lee et al., 2008). The polar lipids contain unwanted components
such as sugars, phosphorous and sulfur, which can cause process
problems like gumming or catalyst inactivation. We believe that the
lipids accumulated by methanotrophic bacteria could also be used as
the fuel precursor for the production of liquid fuel, but the high concen-
tration of phospholipids is expected to exacerbate the process problems
seen in triglyceride streams even when the phospholipid concentration
is relatively low. The lipid composition in the biomass will need to be
taken into account when developing the solvent-based extraction
Fig. 5. Simplied molecular pathwayof serine cycle for methanotrophic bacteria. Major enzymes are presented in green. Dash arrows represent exemplary possible exit points of metab-
olites for biosynthetic reactions. Abbreviations: MMO,methane monooxygenase; MDH, methanol dehydrogenase; H
4
MPTP, methylene tetrahydromethanopterinpathway; MtdA, meth-
ylene tetrahydromethanopterin dehydrogenase; FDH, formate dehydrogenase; STHM, serine hydroxymethyl transferase; HPR, hydroxypyruvate reductase; MD: malate dehydrogenase;
MTK, malate thiokinase; MCL, malyl coenzyme A lyase.
Fig. 6. Molecular structure of typical methanotroph phospholipids. (A) 1,2-di-(11Z-hexadecenoyl)-sn-glycero-3-phosphoethanolamine (PE); (B) 1-(9Z-octadecenoyl)-2-hexadecanoyl-sn-
glycero-3-phospho-N-methylethanolamine (PME); (C) 1-hexadecanoyl-2-(9Z-octadecenoyl)-sn-glycero-3-phospho-N,N-dimethylethanolamine (PDME); (D) 1,2-dioctadecanoyl-sn-
glycero-3-phospho-(1-sn-glycerol) (PG).
602 Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
system, since each organic solvent system used will have respective
selectivity for different organic lipid classes.
The fatty acids in lipids accumulated by methanotrophs are either
saturated or mono-unsaturated with different positions of the double
bond and the C14C18 chain lengths, making these fatty acids ideal
for diesel production with respect to the projected fuel properties.
Methanotrophs contained two major classes of phospholipids, i.e., PG
and PE, and two of PE-derivatives: phosphatidyl methyl ethanolamine
(PME) and phosphatidyl dimethyl ethanolamine (PDME) (Fang et al.,
2000)(Fig. 6). The exclusive presence of PG and PE-type phospholipids
is typical for gram-negative bacteria with extensive intracellular
membranes and in particular methanotrophs (Lechevalier and Moss,
1977; Weaver et al., 1975) Phosphatidylserine (PS) is lacking in
methanotrophic membranes apparently due to the presence of active
decarboxylating enzyme(s) which convert PS to PE and methylase(s)
which convert PE to PME and PDME (Goldne, 1972). The phospholipid
proles of Gammaproteobacterial methanotrophs with pMMO were
dominated by PG and PE with hexadecenoic acid (C16:1) as the major
fatty acid, whereas the Alphaproteobacterialmethanotrophshad almost
equal amounts of PG and PME, with the major fatty acids being
octadecenoate (C18:1) fatty acids (Bowman et al., 1991; Fang et al.,
2000; Guckert et al., 1991; Makula, 1978; Nichols et al., 1985;
Romanovskaia et al., 1980).
In a process based upon methanotrophic lipids, it is necessary for the
production strain to convert a signicant amount of methane into
extensive intracellular membranes and free fatty acids. Although
natural strains can have relatively high levels of ICM, it is likely that an
economically viable process will rely on engineered strains with even
higher levels of lipid. The wild type methanotrophic bacteria can
produce more than 20% lipid (w/w) in the cell body with a mole-based
CCE of 60% (Conrado and Gonzalez, 2014; Kalyuzhnaya, 2013). Accord-
ing to the stoichiometric equations taken from Leak and Dalton's
(1986a) study, the biomass yield on CH
4
is 1 g dry cell weight (DCW)/g
CH
4
. Taking into account the theoretical ethanol yield (0.51 g/g), butanol
yield (0.41 g/g), and lipid yield (0.35 g/g) on glucose (Huang and Zhang,
2011), a 35% lipid content from the engineered methanotrophic strains is
the minimum target to compete with the utilization efciency of glucose
for other biofuel production. These lipids obtained from methanotrophs
could be in forms other than phospholipids such as free fatty acids, tri-
glycerides, fatty acid esters, fatty alcohols, or even alkanes and alkenes
that could be oxidized by MMOs. All of these fuel feedstoc ks are being ex-
plored in other microbial production platforms, generally with sugars or
CO
2
as the carbon source. In addition, it will be necessary to optimize the
growth and cultivation conditions to maximize the lipid productivity and
provide optimal fuel composition.
Development of genetic tools
Metabolic engineering strategies will play a critical role in the
development of industrial methanotroph biocatalysts. Such strategies
offer means to enhance macronutrient uptake and alter metabolic ux
towards desirable bioproducts, such as lipids, biochemicals, and other
high-value co-products (Conrado and Gonzalez, 2014). For example,
upstream targeting of methane monooxygenase via overexpression
and protein engineering could enhance rates of methane oxidation to
methanol, and rates of methanol oxidation could be further enhanced
through overexpression and optimization of various methanol
dehydrogenases. To this end, numerous studies have explored the
structure and function of MMOs, lending insight into the architecture
and mechanism of its active site (Culpepper and Rosenzweig, 2012).
Nevertheless, the relatively slow growth rate and low product titers
from the cultivation of methanotrophic bacteria are still obstructing
industrial applications (Yu et al., 2003; Zhang et al., 2008). Therefore,
a functional expression system of the MMO genes in heterologous
hosts that can grow to a high cell density with additional (non-C1)
carbon sources would greatly facilitate development of Bio-GTL
technologies. So far, however, there are relatively few reports of
successful heterologous sMMO expression, in large part due to the
difculties surrounding efcient expression of the sMMO hydroxylase
component (Jahng and Wood, 1994; Jahng et al., 1996; Murrell, 2002;
Wood, 2002). Additionally, sMMO activity in heterologous hosts
(including Escherichia coli,Agrobacterium tumefaciens,Burkholderia
cepacia,Pseudomonas mendocina,Pseudomonas putida,and Sinorhizobium
meliloti) is relatively low and less robust than in methanotrophic hosts,
such as Methylosinus trichosporium,Methylomicrobium album, and
M. parvus (Murrell et al., 2000) and conserved sMMO regulatory systems
unique to methanotrophs further complicate heterologous expression
(Jahng et al., 1996; Wood, 2002). Indeed, to date, there have been no
reports of successful heterologous expression of full-length sMMO in
E. coli (Lloyd et al., 1999; Murrell, 2002; West et al., 1992; Wood,
2002). On the other hand, the successful homologous and heterologous
expression and regulation of pMMO have been reported and reviewed
before (Chistoserdova, 2011; Gou et al., 2006; Lieberman and
Rosenzweig, 2004; Murrell et al., 2000; Smith and Rosenzweig, 2013;
Smith et al., 2011), though a robust engineered methanotroph as
opposed to a natural organism has not yet been constructed. Conversely,
homologous and heterologous expression of sMMO has proven effective
in methanotrophs, including those containing only pMMO genes, in turn
facilitating a number of targeted analyses, including mutation of putative
active site residues and promoter engineering efforts (Lloyd et al., 1999;
Murrell, 2002; Smith et al., 2002). Ultimately, the ability to transfer the
phenotype of methanotrophs to a wide range of microorganisms could
have a signicant impact on industrial exploitation of this metabolic
capability (Chistoserdova, 2011; Lieberman and Rosenzweig, 2004;
Smith and Rosenzweig, 2013), but until then, bioprocesses based on
the conversion of methane to value added products depend upon natural
methanotrophs.
Additionally, enhanced incorporation of formaldehyde into organic
compounds could potentially be achieved via optimization engineering
of RuMP and serine cycles, which have been elaborated by Anthony
(1982). Lastly, an array of metabolic engineering strategies successfully
implemented in other organisms could be adopted to enhance lipid
(Fan et al., 2012; Kalscheuer et al., 2006; X. Liu et al., 2011)andco-
product (Steen et al., 2010; Weber et al., 2010) biosynthesis. Relative
to the ample, facile toolkits available for model organisms, at present,
genetic tools for methanotroph engineering are limited. However,
molecular tools for genetic engineering in Methylomicrobium sp. are
rapidly advancing, and genetic manipulation in these organisms is
now well established. An array of antibiotic resistance markers,
including tetracycline, kanamycin, and ampicillin, has been identied
as selective markers in various Methylomicrobium sp. (But et al., 2013;
Kalyuzhnaya et al., 2013; Kim et al., 2006; Rozova et al., 2010), opening
the door for stackedgenetic manipulation using multiple markers. Of
equal value is the natural resistance to chloramphenicol, observed in
many methanotrophs, which can be used a counter-selective marker
in conjugative transformation (discussed further below). Routine DNA
extraction and manipulation using standard protocols can be achieved
in methanotrophs.A recent report by Ojala et al. demonstrated effective
methanotroph gene manipulation and plasmid construction in E. coli.In
order to avoid methanotroph gene product toxicity, low copy number
vectors were employed, and utilization of strong terminators anking
the cloning sites further alleviated end-product toxicity by elimination
of insert transcription (Ojala et al., 2011).
A series of broad-host-range vectors has also been established and
implemented for methanotrophic and methylotrophic bacteria (Ali
and Murrell, 2009; Csáki et al., 2003; Marx and Lidstrom, 2001, 2002;
Nielsen et al., 1996; Ojala et al., 2011; Stafford et al., 2003). These
vectors facilitate allelic exchange, allowing for deletion and/or insertion
of genes into methanotroph genomic DNA through site-specic
recombinase systems such as CreLox recombination (Marx and
Lidstrom, 2002). Importantly, such allelic exchange, marker-recycling
vectors facilitate the generation of unmarked mutants with multiple
603Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
genetic alterations. One such example is the chromosomal integration
and heterologous expression of bacterial hemoglobin genes into
Methylomonas sp. 16a, which led to enhanced production of the high-
value co-product astaxanthin (Tao et al., 2007). At present, conjugation
offers the most effective means of methanotroph transformation,
though has been demonstrated (Methanotrophy_Consortium, 2014)
and transformation methods continue to be developed (Marx and
Lidstrom, 2001, 2002).
Challenges in Bio-GTL process
Despite the many potential advantages of methanotroph-based re-
newable diesel, lipid production with good biomass productivity from
methanotroph cultivation will face many technical hurdles, largely
due to factors associated with methanotroph cultivation and extraction
and upgrading of microbial lipids. The schematic summary of process
related research and development (R&D) efforts is shown in Fig. 7.
Generally, the R&D efforts to improve the production process would
include natural gas sourcing and delivery, natural strain screening and
construction of improved production strains, optimization of culture
medium and conditions, optimization of bioreactor design, integration
of bioprocess development of lipid purication and upgrading fuel
blendingstocks. Many process related aspectsor factors must be consid-
ered to impact economical large-scale production of the renewable
diesel fuel from natural gas. For instance, identication of optimal
culture conditions for high cell density and lipid production, and the
impact of these culture conditions upon different culture processes are
highly pertinent. The design of the bioreactor conguration and the
design of effective gas (methane) distribution for lipid production are
also crucial steps for the application on industrial scale. The quality of
the nal diesel fuel has been established by a broad array of fuel proper-
ties or standards (e.g. cloud point and cetane number), therefore the
quality of the lipid intermediates will play a major role in the overall
fuel production cost and will highly impact on the catalytic upgrading
steps. Each of the critical aspect for addressing the Bio-GTL process
challenges is discussed in the following section in details.
Optimization of culture medium
Supplying the proper nutrient components to microorganisms is
essential for any bioprocess. A fermentation medium can signicantly
inuence biomass concentration, product titer, and volumetric produc-
tivity, and the individual medium ingredients can be important contrib-
utors to process costs. To achieve high product yield, it is a prerequisite
to design an optimal production medium in an efcient fermentation
process. There are two different strategies for improving culture
medium: open and closed strategies (Kennedy and Krouse, 1999). A
closed strategy normally considers a certain number of the components
and the type of components used for the optimization work. However,
an open strategy is more complex because it considers the best combi-
nation of all possible components available. Most research starts with
the closed strategy based on the well-known recipes for selected micro-
organisms. Regardless of which medium optimization strategy is
chosen, there are three core issues which rst need to be considered:
(1) the effect of medium design on the development of the strain;
(2) the feasibility of scale up based upon shake ask medium design
data; and (3) the target variable for improvement. In addition to these
technical goals, the commercial viability of the medium (mainly cost
contributions and carbon footprint) should be also considered during
the medium optimization process.
Culture medium utilized for the fermentation with methane
is usually based upon nitrate mineral salts media developed for
methanotrophic cultivation (Park et al., 1991; Wolfe and Higgins,
1979). The media for M. trichosporium OB3b (Cornish et al., 1984),
Methylomicrobium buryatense sp. 5G (Khmelenina et al., 2000),
Methylomicrobium alcaliphilum 20Z (Khmelenina et al., 1999), and
Methylococcus capsulatus (Bath) (Park et al., 1992) are made of numer-
ous compounds that provide essential metals, minerals, and nitrogen
needed for metabolism. In addition to methane as a sole carbon source,
methanotrophic bacteria require a variety of macronutrients for
growth, such as phosphorus and potassium, and several other micro-
nutrients including copper (Cu
2+
). As noted above, copper plays
a key role in the physiology and activity of methanotrophs (Prior and
Dalton, 1985). The activity of pMMO predominates under high
Fig. 7. Research and development (R&D) map of Bio-GTL using methane as a substrate.
604 Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
concentrations of copper (Hakemian and Rosenzweig, 2007). Extensive
intracellular membranes and pMMO activity are present only when
copper concentrations exceed 0.85 to 1 μmol g
1
(dry weight) of cells
(Stanley et al., 1983). The sMMO is only expressed under low-copper
conditions (Chistoserdova et al., 2005). Copper has also been observed
to have an effect on the rate of oxidation of methane (Prior and
Dalton, 1985), the extent of intracellular membranes, and the rate
of growth in batch cultures of Methanomonas margaritae (Ohtomo
et al., 1977; Takeda and Tanaka, 1980). Most Gammaproteobacterial
methanotrophs require higher levels of copper for growth than
do Alphaproteobacterial methanotrophs (Stanley et al., 1983). Copper
effects could explain a similar switch in intracellular location observed
in M. trichosporium OB3b (Balasubramanian et al., 2011), but some
methanotrophs do not have the capacity to overcome copper stress
in this way (Stanley et al., 1983). Recently, Semrau et al. summarized
the molecular mechanism for the copper switch in terms of how
methanotrophic bacteria collect copper and utilize it in the methane
oxidation by MMOs as well as the effect of copper on proteome
(Semrau et al., 2010).
Henry and Grbic-Galic (1990) reported that the presence of a
chelator (Na-EDTA) signicantly enhanced rates of trichloroethylene
oxidation by Gammaproteobacterial and Alphaproteobacterial meth-
anotrophs. Suitable low concentrations of EDTA have also been shown
to enhance growth rates of Alphaproteobacterial methanotrophs (Zhang
et al., 2008). Wendlandt et al. (2001) reported that Mg
2+
inuenced
the PHB accumulation by Methylocystis sp. GB 25, perhaps by disturbing
the osmotic equilibrium causing a release of potassium ions into the
medium (Helm et al., 2008; Tempest and Wouters, 1981). Nitrogen
source can also affect the MMO activity. It has been indicated the NH
4
+
showed less inhibitory effect on MMO than NO
3
(Hanson and Hanson,
1996; Megraw and Knowles, 1987). King and Schnell (1994a,b) found
nitrite produced from the oxidation of ammonia by methanotrophs
could cause a more permanent inhibition of methane oxidation than the
ammonia itself.
When a production medium is designed on a laboratory scale using
high concentrations of a substrate, little attention is paid to the limita-
tions associated with the purication processes that will result in an
unpredictable cost on the nal products. Also, the addition of metals
or salts may introduce additional contaminants or complications for
lipid purication and catalytic upgrading steps in the downstream
processing. Therefore, the medium optimization should be studied in
an integrated bioprocessing framework with consideration of improv-
ing metabolic production yield, reducing pretreatment cost, as well as
reducing unnecessary chemicals introduced in the upstream processes.
Optimization of culture conditions
To develop a bioprocess for the maximum production of desired
products, standardization of culture conditions is crucial. The most im-
portant physical variables affecting the cultures of methanotrophs are
pH, temperature, dissolved O
2
concentration, ratio of methane and O
2
,
and the time of cultivation. pH is one of the important environmental
factors for the optimal activity of microorganisms since hydrogen ion
level affects the physiological behavior of living microbes, including
methanotrophic bacteria. Depending on the species of methanotroph,
the optimum pH for methane oxidation varies between 5.0 and 10.0.
For example, M. buryatense 5G was able to grow at a pH range of 6.8
to 10.5 and optimally at pH 9.5 (Kaluzhnaya et al., 2001). Optimum
pH values for cell growth in the culture of M. alcaliphilum 20Z were
at pH 9.09.5 (Khmelenina et al., 1997). There are few reports of
methanotrophs that grew at pH values below 5.0. Pol et al. (2007)
recently isolated a new Verrucomicrobia species that could thrive at
pH value as low as 1.
Cultivation temperature is also a critical parameter because it will
affect cell growth and carbon source utilization, lipid composition, and
the solubility of methane in the culture medium. Cultivation at high
temperatures is usually favorable for commercial production, since
metabolic heat removal can be substantial for a large-scale operation.
In fact, metabolic heat generation led SCP production to methanol
from methane to avoid the rst exergonic oxidation step. The optimum
temperature for methanotrophic growth is from 25 to 30 °C in most
strains, although growth has been reported as low as 5 °C and as high
as 70 °C (Bender and Conrad, 1992; Duneld et al., 1993; King, 1992;
Op den Camp et al., 2009; Tsubota et al., 2005; Whalen et al., 1992).
The rst step in methanotrophic growth is an unusual one consider-
ing that methane combustion proceeds quite readily. MMO catalyzes
the reaction of methane and O
2
to yield methanol and water in a very
controlled fashion, utilizing a reducing equivalent from NADH, though
no biological energy is needed to make this reaction proceed (Hanson
and Hanson, 1996). Due to the mandatory requirement of a reduced
electron and the energy used for the oxidation of methane to methanol,
the energy efciency is constrained on MMO in the native MMO
pathway. A highly exothermic reaction proposed by Conrado and
Gonzalez could lead a zero net energy input through an engineered
methane activation route (Conrado and Gonzalez, 2014). Regardless of
this complication, O
2
is an essential factor that directly affects the rate
of methane oxidation. The response of methanotrophs can be consid-
ered from two aspects: the inuence of high O
2
concentration and the
capability of methanotrophs to act with lower O
2
concentration. The
detailed information on the effect of O
2
concentration on MMO-
catalyzed methane oxidation has been explored in previous reports
(Hanson and Hanson, 1996; King, 1992). A high cell growth rate
depends on sufcient O
2
supply not just for methane oxidation, but
for oxidative phosphorylation as well, but process economics requires
a balance between growth rate and the expense of maximizing O
2
supply through high powered mixing and use of pure O
2
rather than
air (Ren et al., 1997; Wilshusen et al., 2004). Amaral and Knowles
(1995) observed that Gammaproteobacterial methanotrophs were
generally associated with high ratios of O
2
to methane whereas
Alphaproteobacterial were found to favor the opposite. O
2
concentra-
tion ranging from 0.45 to 20% v/v could support maximum cell growth
rates in both Gammaproteobacterial and Alphaproteobacterial
methanotrophs (Ren et al., 1997). However, relatively low O
2
con-
centrations will favor ICM formation (Scott et al., 1981). The ratio of
methane and O
2
has shown a great inuence not only on reaction
kinetics, but also on the mass transfer coefcient (K
L
a) in methanotroph
culture system (Asenjo and Suk, 1986; Whittenbury et al., 1970).
Harwood and Pirt (1972) also found that the oxidation of methane
was markedly inuenced by the partial vapor pressures of methane
and oxygen when the culture condition changed from methane-
limited culture to oxygen-limited culture. The time of cultivation
and the age of the inoculum could also improve the mass transfer
characteristics and volumetric productivity of the desired products
(Asenjo and Suk, 1986).
Mass transfer enhancement and bioreactor design
One of the major challenges to improve production yields in Bio-GTL
process is the enhancement of mass transfer of methane from the gas
phase to the culture medium liquid phase and then into the cells. Gas
liquid mass transfer is generally recognized as a barrier in the Bio-GTL
process (Conrado and Gonzalez, 2014; Klasson et al., 1993a; Worden
et al., 1991), because limitations are inevitable at several points of the
diffusion process. In order to maximize the utilization efciency by
microorganisms, the gas molecules must transport into gasliquid in-
terface and then diffuse through the culture media (aqueous phase)
and then through the microbial cell surface, where they can participate
in metabolic reactions. The gasliquid interface mass transfer is the
major obstacle for gaseous substrate diffusion. Vega et al. (1990)
indicated that there is a strong connection between mass transfer,
kinetics, cell growth, and production of the desired products. The rela-
tionship can be concluded as follows: in the beginning of fermentation,
605Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
the generation of cell mass and products is constrained by metabolic
reaction kinetics with methane and oxygen readily available, and then
as the cell density increases, further growth depends on the mass trans-
fer rate (Slivka et al., 2011). Therefore, diffusion limitations of a gaseous
carbon source into the liquid media become apparent at high cell
densities and can cause low overall productivity.
Mass transfer rate limitations can be understood and overcome by
analysis of the mass transfer coefcient (K
L
)(m/s), which was described
by Klasson et al. (1993b) using the following equation:
dNG
S
dt¼VLKLa
HPG
SPL
S

where N
S
G
(mol) is the molar substrate transferred from the gas phase;
V
L
(L) is the volume of the reactor; P
S
G
and P
S
L
(atm) are the partial
pressures of the gaseous substrate in gasand liquid phase, respectively;
H(L × atm/mol) is the Henry's law constant; and a(m
2
/L) is the gas
liquid interface surface area for unit volume. The differential of the
partial pressures between the gaseous substrate P
S
G
and P
S
L
is considered
as the main driving force for mass transfer and thus controls the solubil-
ity of the substrate (Munasinghe and Khanal, 2010). Therefore, much
effort has gone into the design of bioreactors that can provide a higher
mass transfer coefcient by generating more gasliquid interfacial
area from smaller bubbles. Because of the various characteristics of the
microorganisms used in gas fermentation, there isno optimal bioreactor
design. High mass transfer rates, low operation and maintenance costs,
and easy scale-up are key parameters for developing an efcient bio-
reactor system (Munasinghe and Khanal, 2010). Several reactor cong-
urations have been proposed to enhance the mass transfer rate.
Continuous stirred tank reactors (CSTRs) are the most widely used
bioreactors for industrial fermentations. A general approach to improve
the mass transfer efciency in CSTR is to increase the agitation speed or
modify the impellor design (Bredwell et al., 1999), providing more
energy to generate smaller-size bubbles for increasing the gasliquid
interfacial area. However, high shear rates from excessive agitation
could damage cells and inhibit growth rate (Kadic, 2010). Moreover,
the increased power demand for this strategy greatly reduces its
economic viability in large-scale gas fermentations for low costproducts
(Munasinghe and Khanal, 2010; Ungerman and Heindel, 2007).
Consequently, air-lift reactors (Bredwell et al., 1999), micro-bubble
sparged reactors (Bredwell and Worden, 1998), bubble column reactors
(Bouaiet al., 2001; Datar et al., 2004), trickle-bed reactors (Bredwell
et al., 1999), and membrane-based reactors (Lee and Rittmann, 2002;
Nerenberg and Rittmann, 2004; Tsai et al., 2011) are some of the other
congurations that have been examined and reviewed (Munasinghe
and Khanal, 2010). Recently, a method of adding nanoparticles with
functional groups into gas fermentation system has been developed
to exploit the extensive adsorption capability of the functionalized
nanoparticles to increase the mass transfer coefciency (Zhu et al.,
2008, 2009; Zhu et al., 2010).
Bioprocess development
Despite the rapid development of bioreactors for gas fermentation
within recent years, only a few studies have focused upon the develop-
ment of bioprocess technology for fuel production by methanotrophic
bacteria. PHB production by methanotrophs was a topic of interest
20 years ago that has been reevaluated more recently (Shah et al.,
1996; Suzuki et al., 1986; Zhang et al., 2008). In those works, the re-
searchers tried to control methane concentration and O
2
level during
the cultivation of methanotroph to achieve high cell densities and
productivities. As we have noted, microbial lipids accumulated by
methanotrophic bacteria have very recently become a topic of interest,
but no process research has yet been reported.
One of the widely applied fermentation methods to achieve high cell
density, combined with high yield and productivity of the desired
products, is a fed-batch culture with controlled nutrient feeding. There
have been multiple reports of employing fed-batch culture for lipid
production in oleaginous yeasts (Fei et al., 2011a; Li et al., 2007; Lin
et al., 2011). For the synthesis of specic products, the nutrients limita-
tion strategy has been developed with the purpose of maximizing
particular product titer. A common strategy in the production of micro-
bial lipid is the separation of the process into cell growth stage and lipid
accumulation stage. A two-stage culture mode utilizing a nutrient
limitation strategy could improve lipid yield and avoid the inhibitory
effects often seen with high carbon source concentration (Chang et al.,
2012; Courchesne et al., 2009; Fei et al., 2011b), though this situation
is unlikely to be observed in methane cultures. It has been reported
that the lipid accumulation by methanotrophic bacteria could be
controlled by the concentration of O
2
, nitrogen source, and phosphate
source during the cultivation. Therefore, high lipid content could
be achieved by maintaining a limitation condition in a two-stage fed-
batch culture in the cultivation of methanotrophs.
Chang's group (Chang et al., 2011, 2014; Kang et al., 1993; Park et al.,
1985) has been developing multistage continuous high cell density cul-
ture systems (MSC-HCDC) with sugars as carbon sources for decades,
which can provide much higher productivity along with a high product
titer similar to those achieved in fed-batch cultures. This system was
proposed to be composed of n-serially connected CSTRs with either
hollow ber cell recycling or cell immobilization for a high cell density
culture.It is possible that this MSC-HCDC can be modied and employed
in the fermentation system using gases as the carbon sources for the
production of microbial lipids or fatty acids, which are intracellular
products used as the precursor for biofuel production (Amelio et al.,
2007; Lima et al., 2012). Higher cell density boosts the production
titer of intracellular products as well as the productivity of biofuel
production. The similar systems have been utilized to improve the
production titer and productivity of an intracellular product poly
(3-hydroxybutyrate) (PHB) and fatty acids in sugar-based fermentation
(Ahn et al., 2001; Fei, 2011; Wen and Chen, 2001). A 100% and 400%
improvement on titer and productivity during the gas fermentation
process can be expected by using multistage CSTRs and hollow ber
membranes proposed by Chang's group (Chang et al., 2011). Compared
with traditional fermentation monitoring sugar level during cultures,
methane and O
2
sensors are necessary in the cultivation of meth-
anotrophic bacteria. Flow and pressure controllers for supplying the
methane and O
2
can integrate with these sensors to build up an online
analysis and control system. This system could be used with
methanotrophs to supply sufcient growth substrates during the
biomass-accumulation stage and then be switched to induce a lipid-
accumulation stage by manipulating the ratio of methane and O
2
and/or by limiting certain nutrient such as nitrogen or phosphate source
during the cultivation. Besides the consideration of the optimization of
the bioprocess control, the real-time monitoring of the concentration
of methane and O
2
can also help researchers avoid the 5% (vol/vol)
lower explosive limit (LEL) and 15% (vol/vol) upper explosive limit
(UEL) of methane in air. Any methane concentration between LEL and
UEL and an ignition source can cause an explosion (Hamer et al.,
1967; Zlochower and Green, 2009).
Lipid extraction process development
Lipid extraction for a Bio-GTL process will require specictailoring
of the extraction conditions to take full advantage of all lipids in the bio-
mass. The correct mixture of solvents can give the extraction process
ideal polarity to provide high yields on fatty acids based on the recovery
of all lipid types. This is an important consideration because the quantity
and quality of the lipid stream will affect the downstream processing
options (e.g. the requirement for cleanup steps to maximize catalyst
longevity) and yields. At each step of extraction process development,
the effectiveness of extraction with respect to distribution of fuel
precursor lipids should be determined.
606 Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
Analytical extractions reported in the literature for microbial lipids
are reported to be highly effective and primarily utilize a traditional
method based on chloroform: methanol extraction of harvested wet
biomass (Bligh and Dyer, 1959). Although effective, an extraction
system based on chlorinated solvents may not be considered scalable
due to toxicity and cost. Hexane isa common solvent used forextraction
because it is easy to recover in a solvent recycling step, thanks to its low
boiling point. However, limited sol ubility of polar lipids in hexane leaves
behind potentially valuable by-products. Supercritical uid extraction
(SFE) can be considered as a potential green technology to replace
solvent extraction, which has been reviewed and compared recently
(Halim et al., 2012). The advantage of SFE as an extraction method is
that CO
2
acts as a solvent for lipid extraction at high pressures
(72.9 atm) and subsequent depressurization makes this solvent
removal relatively easy. Nevertheless, the non-polar properties of
supercritical CO
2
make this extraction type selective for neutral lipids
in the absence of entrainers, and thus in its pure form SFE with only
CO
2
may be less applicable for methanotroph polar lipid fraction.
There have been studies on the utilization of co-solvents, such as etha-
nol to aid with the extraction of polar lipids (Montanari et al., 1999),
whereas the SFE system has not been optimized for extraction of
polar lipids. A study of the extractability of different lipid types in a
SFE extraction system indicated that it could only extract at most
about 60% of what a conventional extraction (e.g. Soxhlet extraction
using polar/non-polar solvent mixtures with or without cell-disruption
prior to extraction) yields (Soh and Zimmerman, 2011). In addition,
concerns about cost and scalability are considerable for this process to
be deployed. Dried cell mass is often chosen for extracting the lipids
for analytical-scale extraction, whereas due to the high cost and energy
requirement for drying the biomass, emphasis for technology develop-
ment ought to be placed on the development of a liquid or wet extraction
system.
Experimentation with cell disruption as a means to increase the
accessibility of lipids to the extraction solvent has been covered in the
literature for microalgae (Greenwell et al., 2010; Halim et al., 2012),
and some parallels can be drawn between algal biomass extraction
and methanotroph biomass extraction. The majority of these technolo-
gies were developed for biomass containinghigh levels of neutral lipids,
which are easier to extract from the surrounding water environment
with a nonpolar solvent like hexane. Yoo et al. (2012) indicated that a
1.5 fold increase of lipid recovery can be achieved by using an osmotic
shock pretreatment prior to solvent extraction of microalgal biomass
from cell-wall-less mutant strains. Additional cell disruption steps can
be found in the literature that can increase the efciency of solvent-
based lipid extraction (Soh and Zimmerman, 2011). Although, there
are parallels between lipid extraction technologies developed for
microalgae and oleaginous yeasts and those that are needed to render
a methanotroph-biomass process economical, the differences between
prokaryotic (bacterial) and eukaryotic cells could have implications on
the effectiveness of extractions.
One of the major differences between bacterial and eukaryotic cells
can be found in the cell wall structures. With the absence of structural
polysaccharide structures in the cell wall of bacteria, the cells will not
need the same level of energy (and cost) intensive cell disruption
steps to rupture or hydrolyze the cell wall. However, in addition to a
cell wall that is easy to breach, there is the additional challenge of
lipid composition and solvent separation steps. It is possible that the
application of mechanical disruption will not work for the isolation of
the polar lipids as in methanotroph biomass, due to their amphiphilic
nature and close interaction with hydrophobic membrane proteins.
Another potential problem may occur if polar lipids act as surfactants
and cause tight water/solvent emulsions requiring unacceptable energy
inputs to resolve. Alternative pathways for extracting lipids may be
found in a chemical treatment of the biomass prior to an extraction
process. It has been proposed that acid and/or base hydrolysis ofthe bio-
mass can enhance lipid extractability by release of free fatty acids
derived from the microbial lipids (Harun and Danquah, 2011; Harun
et al., 2011; Sathish and Sims, 2012). An acid followed by base hydroly-
sis process successfully recovered 79% lipids, of which 74% was found in
the solvent and 26% could be recovered from different steps in the pro-
cess, either in the solid precipitate or in the aqueous phase. This process
was demonstrated on wet microalgal biomass with 84% moisture
(Sathish and Sims, 2012). Saponication of the methanotroph lipids
may reduce the amount of potential catalyst poisons (especially
phosphorous and sulfur) in the extract and thus reduce downstream
clean up steps and increase catalyst longevity. However, unless
this process can be combined with the lipid extraction process, this
additional chemical conversion step prior to catalytic hydroprocessing
might be cost-prohibitive.
In summary, developing and tailoring an extraction process should
be carried out in concert with a upstream and downstream process in
mind, where for example the harvesting process can be linked with a
wet extraction process and the extracted lipid fraction can be directly
fed to a catalytic upgrading process that is tailored to the lipid composi-
tion generated in the production strain. A thorough characterization of
the lipid fraction composition is needed atall steps to tailor the biomass
processing and lipid extraction technology and to achieve the optimum
efciency of fatty acid extraction.
Hydrotreating process
A hydrotreating process follows the extraction of the methanotroph
biomass for upgrading the fuel lipids into hydrocarbon fuels. A simpli-
ed hydrotreating process is shown in Fig. 8. The hydrotreating process,
also known as hydroprocessing or hydrodeoxygenation accompanied
by hydrocracking (Soltes and Lin, 1987), is an established renery pro-
cess to reduce contaminants such as sulfur and nitrogen (by reduction
of C\N and C\S bonds), condensed ring aromatics, or metals, while
enhancing cetane number, density, and smoke point. Hydrotreating is
also employed to convert feedstocks containing double bonds and
oxygen moieties into parafnic hydrocarbons by saturating the double
bonds and removing the oxygen with release of either CO
2
or H
2
O. A
typical reaction of catalytic hydrotreating process of a lipid-based
feedstock-triglyceride is listed in Eq. (1), where R represents unsaturated
alkyl groups and Rrepresents saturated alkyl groups. The triglyceride
(CH
2
(COOR)CH(COOR)CH
2
(COOR)) consists of the fatty hydrocarbon
tail (R) that varies in the length of carbon chains and in the number
of unsaturated regions. As shown in Eq. (1), by catalytically removing
oxygen and saturating double bonds of fatty acid chains with hydrogen
Fig. 8. Hydrotreating process for the production of renewable diesel fuel.
607Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
(H
2
), straight-chain renewable diesel (RH
3
) and propane (C
3
H
8
) are
formed along with H
2
O. The reactions involved in the hydrotreating
process are mainly dependant on the feedstocks and catalysts, which
have been in some detail (Choudhary and Phillips, 2011; Furimsky,
2000; Popov and Kumar, 2013).
CH2COORðÞCH COORðÞCH2COORðÞþ12H2
Catalyst C3H8þ3R0H3þ6H2O
ð1Þ
The chemical composition of lipid feedstocks can be changed in
various ways using renery processes for the production of renewable
diesel or jet fuel under both elevated temperature and elevated pres-
sure. Several catalyst systems can be applied to obtain different liquid
fuels. This process is usually divided into a deoxygenation stage and
isomerization stage. In the case of renewable diesel, feedstocks such as
lipids, fatty acids or triglyceride molecules are rst hydrotreated. An
isomerization process often catalyzed by sulded catalysts (Şenol
et al., 2005a, 2005b) may be needed following hydrotreating to convert
the n-alkanes to branched chain alkanes for improved cold weather
characteristics. In some cases, a third processing step, hydrocracking,
is also needed. This step breaks C\C bonds and reduces the chain
length, to meet product specications. Sometimes, a separate cracking
step may not be necessary because the feedstock fatty acids may meet
the product specications. In addition, some cracking can be expected
to occur in the hydrotreating and isomerization steps. But if the micro-
bial lipid feedstock contains very long chain fatty acids (outside the
typical range of C14C18 which matches reasonably well with diesel
and jet fuel) or if the target fuel is gasoline, hydrocracking would be
necessary.
Signicant challenges exist in the effectiveness of hydrotreating for
oils that are high in polar lipids, as most catalysts are rapidly poisoned
by the presence of heteroatoms such as P, S, and N in the lipid stream
(Qian et al., 2001). In particular the phosphorus-containing molecules,
found in the lipid fractions extracted from methanotrophs, are known
to be potent catalyst poisons (Angele and Kirchner, 1980; Bouwens
et al., 1988; Koritala, 1975). A potential hurdle for a Bio-GTL process is
likely to be faced in the presence of large amounts of phospholipids,
which can cause rapid catalyst deactivation in hydroprocessing
(Kubička and Horáček, 2011). It remains to be seen whether novel
catalysts can be developed, which are robust for the co-processing of
the complex lipid streams. Developing a technology to convert these
lipids to fuel range hydrocarbons would be critical for the successful
development of the Bio-GTL process. A small number of publications
have described homogeneous catalyst systems for the hydroprocessing
of unsaturated carbon bonds in phospholipids (Nádasdi and Joó, 1999;
Vigh et al., 1987), but the selective conversion of phospholipids to
hydrocarbons has not been addressed in the literature and thus the
development of robust and resistant catalysts is absolutely necessary.
An alternative route to improvements in lipid conversion is tailoring
the extraction system to the downstream hydrotreating process. As
discussed above, the composition of the lipid fractionwill highly depend
on the extraction and conversion process introduced and thus the
hydrotreating process will need to be modied or developed to be
robust to the range of lipid compositions encountered. An overall
process conceptual design would be a good tool to help synergize
process development among fermentation, extraction, and catalytic
upgrading processes.
Many industrial companies are developing this hydrotreating
process as the basis for their renewable diesel projects. For example,
Tyson Foods is working with Dynamic Fuels and ConocoPhillips to
turn waste animal fat into renewable diesel (Tyson, 2007). UOP, a
Honeywell company, has been focusing on converting natural oils, fats
or grease to an advanced drop-in renewable diesel product by using
their patented green diesel processes (UOP_Honeywell, 2013). Other
companies utilizing this system include Shell in the US (Eilers et al.,
1990; Sie et al., 1991), Petrobas in Brazil (NREL, 2006), Eni in Italy
(Lane, 2012), Aemetis Inc., agreement with Chevron Lummus Global
(Aemetis, 2012), Dynamic Fuels, a 50/50 joint venture between
Syntroleum and Tyson Food (Syntroleum Corporation, 2011), and
Neste Oil Corporation in Finland (Neste_Oil, 2007).
Economic considerations
The future of alternative diesel fuels hinges on various factors such
as feedstock availability, processing costs, and supportive political
framework. Economic analysis for a process in early stage development
requires a conceptual level design to develop a detailed process ow
diagram, rigorous materials and energy balance calculations, capital
and project cost estimations, and economic analysis using appropriate
nancial assumptions. A techno-economic analysis (TEA) is commonly
used as a process economic analysis tool to guide investors and policy
makers in order to down select the most effective technology
(Aden et al., 2002; Bowonder and Sharif, 1988; Davis et al., 2011;
Dutta et al., 2011; Peeters et al., 2007; Tao et al., 2012). The economic
consideration of a methane-based biotechnology for liquid fuel pro-
duction could be compared with the gas fermentation process, as well
as with production processes in petro-diesel industry. Price and
availability are two major concerns for a feedstock that can be used in
a large-scale production of commodities. In the US, 5.4 × 10
13
BTU of
methane is ared every year (World_Bank, 2013). This ared methane
might be considered as a zero or low-cost value feedstock. The cost of
liquid fuels derived from the wasted natural gas could be signicantly
reduced and competitive compared to fossil fuels or biofuels (Tao and
Aden, 2009). The Bio-GTL can also avoid a global impact on climate
change caused by the 4 × 10
8
tons of CO
2
and CO
2
equivalents released
annually from aring and venting methane (World_Bank, 2013). The
price of most chemical feedstocks has risen dramatically, trending
with the price of petroleum, whereas the wellhead price of natural gas
as shown in Fig. 9 droppedfrom $7.9 to $2.6 per 1000 ft
3
(corresponding
to $2.5/MMBtu) (EIA, 2012b). Because of the development of shale gas
technology, the production capacity of natural gas has been improved
signicantly, leading to reduced costs, which are expected to remain
stable for some time. Adding to this economic situation, the price of
natural gas has shown a unique trend during the past ve years
(EIA, 2012b). It has been suggested that the methane-based diesel
produced by methanotrophs could be produced for less than half the
cost of other biofuels from microalgae and other microorganisms
grown on sugars, and could even compete directly with petro-based
fuels (Calysta_Energy, 2013). Recently, Calysta Energy, LLC. (Menlo
Park, CA, US) announced its plans to invest in the research of transpor-
tation fuels and chemicals production from natural gas. Calysta
representatives have stated that liquid hydrocarbons can be produced
from natural gas on a large scale via their proprietary Bio-GTL platform,
which is claimed to be faster and more efcient than conventional
routes, such as algae-or sugar-based methods (Calysta_Energy, 2013).
Strategies to use methanotrophs to produce SCP, chemicals or PHB
from methane have been under development since the middle of the
20th century (Anthony, 1982; Gürsel and Hasirci, 1995; Oliver and
Colwell, 1973). However, no economic studies have been published
for Bio-GTL processes. Several advances in upstream processes such
as the constructionof production strains, optimization of culture condi-
tions, and reactor design may be needed to achieve a cost-competitive
process. Proper sourcing of low cost of raw materials (especially at
sites where natural gas is ared or vented) can also reduce the produc-
tion cost. Optimizing the culture medium and conditions can improve
overall production cost and energy utilization. Constructing a robust
strain for lipid production by methanotrophs can improve the CCE and
production yield. Maximum product titer and productivity can be
accomplished through synergistic development of an appropriate
production reactor with optimal gas transfer capabilities, and strain
development. Besides the factors mentioned above, engineering factors
608 Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
to reduce the risk of explosion, cooling water usage, the source of O
2
(air, versus O
2
-enriched air or pure O
2
), the type of nitrogen sources,
and fatty acid prole would also contribute to the total production
cost of lipids. Downstream processes, such as lipid extraction and
upgrading will also require signicant work as the precedents from
other lipid-production strains such as microalgae and oleaginous
yeast, are not perfectly instructive for a polar lipid based process. Lipids
produced by methanotrophs can be pretreated for the hydrocarbon
fuels. High efcient and cost competitive catalysts used for lipid extrac-
tion and hydrotreating process as well as conversion efciency of these
processes can reduce the price of the nal liquid fuel and increase the
prots of the investment.
Since detailed economic analysis is not available for this specic
pathway technology, a preliminary cost analysis summarizes the diesel
production cost from various CH
4
prices in cost year of 2012 for the
Bio-GTL process (Table 3) based solely on raw material costs and yields
(excluding all other cost of production). When CH
4
is used as a carbon
source for diesel production by methanotrophs, raw material cost
contributions range from $0.7 to $10.8/gal, shown in Table 3.While
petroleum-derived diesel is currently sold at $4/gal in the US
(EIA, 2014), the methanotrophic process could have potential to be
economically compatible if all other cost of production would be held
at or below $2.3/gal, which is the price of crude oil purchased by reners
for the production of petroleum-derived diesel (EIA, 2014). Except the
potential advantage of the low cost, methane-derived diesel can also
help reduce the demandsof the petroleum-derived diesel. For instance,
3.86 kg natural gas can produce 1 kg diesel theoretically, which means
600 ft
3
natural gas can be converted into 1 gal of diesel fuel. More
than 1.1 × 10
14
ft
3
global natural gas has been produced in 2011. If all
converted to diesel, roughly 1.8 × 10
11
gal of diesel fuel could be
made. Therefore, taking account of the 2.9 × 10
11
gal of diesel fuel
produced in 2012 (EIA, 2012b), an estimated 60% petroleum-derived
diesel can be potentially replaced annually by the methane-derived
diesel via the Bio-GTL process. In consideration of 2.1 × 10
11
ft
3
natural
gas ared and vented in 2012 in the US (EIA, 2012a), 3.5 × 10
8
gal of
diesel fuel could be contributed from the wasted natural gas. However,
in order to achieve the cost target of being compatible with petroleum-
derived diesel and the long-term sustainable supply, research and
development should be focused in the area of increasing DCW/CH
4
yield, lipid content, extraction yield as well as hydrotreating yield,
shown in Table 3 as majorcost driven factors. Sincethe prices and yields
related with this preliminary assessment are simply assumed based
upon literature studies and none of the capital costs and operation
costs have been considered carefully here, signicant uncertainty of
the results should be expected. Also note that the cell components
after lipid extraction, such as protein, pigments, and waste cells, could
be potentially collected and converted to value added coproducts.
The preliminary analysis shown in Table 3 only suggests that
methane-derived diesel fuel has the potential to be competitive with
petroleum-derived diesel, especially when natural gas prices stay low.
A more comprehensive TEA should be performed for an integrated
Bio-GTL process to identify key technical barriers that will have implica-
tions on overall process economics.
Other approaches that can support TEA and life cycle assessment
(LCA) efforts involve the recovery of byproduct credits, integration of
wastewater treatment, consideration of CO
2
and methane credits,
residual biomass after lipid extraction, and propane from hydrotreating
process, etc. For instance, residual biomass can be employed in an
anaerobic digestion process for the production of volatile fatty acids
(Chang et al., 2010; Lim et al., 2008a, 2008b) or recycling of carbon as
methane in biogas (Holm-Nielsen et al., 2009), sugars (Li et al., 2009),
or SCP (Bothe et al., 2002). Although a biological catalyst is utilized for
the production of lipids and fatty acids, natural gas is still considered
as a fossil fuel feedstock, and the product of a natural gas Bio-GTL
process would likely remain a fossil fuel. If biogas from anaerobic
digestion process were used as the feedstock, the product could be
considered a biofuel or renewable fuel. In addition, there is a growing
awareness of important environmental issues, including determining
the greenhouse gas (GHG) emissions from the conversion processes.
Syngas to liquid fuel using FT technology has been studied and reported
Fig. 9. The US natural gas wellheadprice from 1940 to 2012.
Data collected from U.S. Energy Information Administration.
Table 3
Preliminary cost assessment of the renewable diesel production cost from the Bio-GTL process for the best and worst scenarios in cost year of 2012.
Scenario CH
4
price Biomass yield
(DCW/CH
4
)
Lipid content
(lipid/DCW)
Extraction yield
(FAME/lipid)
Hydrotreating yield
(diesel/FAME)
Estimated raw material
cost/gal of diesel
Best case $100/ton
a
1g/g
b
50%
c
0.95 g/g
d
0.95 g/g
e
$0.7/gal
f
Worst case $200/ton
a
0.6 g/g
b
20%
g
0.7 g/g
d
0.7 g/g
e
$10.8/gal
f
a
U.S. Natural Gas Wellhead Price from Jan to Dec of 2012 (EIA, 2012b). The content of CH
4
in natural gas could be as high as 80% (IEA, 2013).
b
Estimated based on the 60% CCE reported in pervious reports (Conrado and Gonzalez, 2014; Kalyuzhnaya et al., 2013).
c
Data assumed in this study.
d
Data taken from Im et al. (2014).
e
Data taken from Choudhary and Phillips (2011).
f
Results calculated by using the following equation: Cost = CH
4
price / biomass yield / Lipid content / Extraction yield / Hydrotreating yield. Diesel density is 0.84 kg/L.
g
Data taken from 2013 ARPA-e workshop (Kalyuzhnaya, 2013).
609Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
with high life cycle GHG emission potentials (Marano and Ciferno,
2001; Xie et al., 2011), but limited studies have been reported with
environmental judgments for Bio-GTL related pathways using natural
gas or biogas as the feedstock.
Low cost and high availability of natural gas argue for the potential of
methanotroph pathways to liquid fuels (potentially diesel blend
stocks). Such a process can be economically viable as well as sustainable
in the near future if biological productivity is competitive with algae-
based or sugar-based pathways. Synergistic efforts that consider all
the process steps in a detailed process framework could help the design
of the biorenery process in a more cost effective way and provide
proper guidance in the research and development to drive down the
production cost. Variation of process alternatives with economic
consideration could help research to align better not only with near
term cost goals but also with long-term environmental goals.
Future prospects and conclusions
Considering the recent volatility of crude oil prices and the potential
for future shortages, the utilization of natural gas/methane as a sub-
strate for liquid fuel production has tremendous potential. Hydrocarbon
liquid fuel production from natural gas/methane could replace a signif-
icant amount of petroleum-based liquid fuel usage in the US, while at
the same time capturing value from a wasted resource and mitigating
climatechangeissuesexacerbatedbyventedandared natural gas.
Nevertheless, the challenges in moving from proof of concept to scale-
up and commercialization still remain to besolved. In contrast to the ef-
forts to develop biofuels in last century, the pathway to fuels from
methanotrophs remains in its infancy. With the advent of economic
and efcient tools of systems biology especially genomics, transcripto-
mics, and metabolomics, the potential to construct recombinant
methanotrophic bacteria for fuel production is becoming much more
accessible. The most important prerequisite for the production of micro-
bial lipids and hydrocarbon fuels from methanotrophs will be a robust
strain with a stable phenotype for rapid growth and lipid production.
It is likely that stable overexpression of pMMO will play a large role in
this phenotype both for increased methane oxidation rates and for in-
creased lipid from ICM serving as a support for the higher levels of
pMMO (Chistoserdova et al., 2009), but additional manipulations to im-
prove carbon ux to lipids and fatty acids will likely also be needed to
increase the yields. By constructing de novo fatty acid synthesis path-
ways with gene knocking-in or/and knocking-out techniques, the pro-
duction of liquid fuels is becoming viable (Fei et al., 2013; Lu et al.,
2012). However, typical approaches for metabolic engineering result
in the interruption of the natural metabolic networks and imbalance
of some important metabolites, such as ATP/ADP, NAD +/NADH,
NADP+/NADPH, and Acyl-CoA (Lee et al., 2008). Therefore, a compre-
hensive understanding of the central metabolism of methanotrophs is
needed, which could minimize the inuences and maximize the pro-
duction of desired products. Another area that could help exploit
methanotrophs as fuel producers lies in observations that some
methanotrophs are also capable of growing chemolithoautotrophically
with CO
2
as a carbon source. Khadem et al. (2011) observed that
Methylacidiphilum fumariolicum strain has the ability to use the Cal-
vinBensonBassham (CBB) cycle for CO
2
xation with CH
4
as the ener-
gy source. This study indicated that cultivation of this strain using a
mixture of methane and CO
2
could improve the cell growth and product
yield, which could reduce the production cost of liquid fuels. This could
be especially valuable if biogas is the source of methane because CO
2
is a
signicant component. Meanwhile, efforts to optimize culture medium
and conditions as well as exploring bioprocess technology are being
pursued to enhance lipid productivity and reduce production cost. Var-
ious reactors designed for gas-phase fermentation are offering great
contributions to improve mass transfer rate during the cultivation.
Because high cell density culture of methanotrophic bacteria, such as
M. capsulatus for the production of SCP has been achieved by applying
UniBio's U-Loop technology (UNIBIO, 2005), high volumetric titer and
productivities of microbial lipids and liquid fuels can be expected in
industrial applications. Moreover, development of catalysts for lipid
extraction and upgrading will continue to make methane-based liquid
fuel more competitive with other biofuels on TEA consideration.
Acknowledgments
This work was supported by a funding (project number: 0670-5169)
from Advanced Research Projects Agency Energy (ARPA-E) and U.S.
Department of Energy (DOE).
References
Ab Rahim MH, Forde MM, Jenkins RL, Hammond C, He Q, Dimitratos N, etal. Oxidation of
methane to methanol with hydrogen peroxide using supportedgoldpalladium alloy
nanoparticles. Angew Chem Int Ed 2013;52:12804.
Adebajo MO, Frost RL. Recent advances in catalytic/biocatalytic conversion of greenhouse
methane and carbon dioxide to methanol and other oxygenates. Greenhouse gases:
capturing, utilization and reduction; 2012. p. 3156.
Aden A, Ruth M, Ibsen K, Jechura J, Neeves K, Sheehan J, et al. Lignocellulosic biomass to
ethanol process design and economics utilizing co-current dilute acid prehydrolysis
and enzymatic hydrolysis for corn stover. DTIC document; 2002.
Aemetis. Aemetis expands biojet license agreement with Chevron LG (http://www.
biodieselmagazine.com/articles/8752/aemetis-expands-biojet-license-agreement-with-
chevron-lg) Biodiesel Magazine; 2012.
Ahn W, Park S, Lee S. Production of poly(3-hydroxybutyrate) from whey by cell recycle
fed-batch culture of recombinant Escherichia coli. Biotechnol Lett 2001;23:23540.
Ali H, Murrell JC. Development and validation of promoter-probe vectors for the study of
methane monooxygenase gene expression in Methylococcus capsulatus Bath.
Microbiology 2009;155:76171.
AlperinM,ReeburghW.Geochemical observations supporting anaerobic methan e
oxidation. Microbial growth on C-1 compounds. 1984;282:289.
Amaral JA, Knowles R. Growth of methan otrophs in methane and oxygen counter
gradients. FEMS Microbiol Lett 1995;126:21520.
Amelio M, Morrone P, GallucciF, Basile A. Integrated gasication gascombined cycle plant
with membrane reactors: technological and economical analysis. Energy Convers
Manag 200 7;48:268093.
Angele B, Kirchner K. The poisoning of noble metalcatalysts by phosphoruscompounds
I chemical processes, mechanisms and changes in the catalyst. Chem Eng Sci 1980;
35:208991.
Anthony C. The biochemistry of methylotrophs. London: Academic Press; 1982.
Anthony C. Bacterial oxidation of methane and methanol. Adv Microb Physiol 1986;27:
113210.
Anthony C. Assimilation of carbon by methylotrophs. Biotechnology 1991;18:79.
Anthony C. How half a century of research was required to understand bacterial growth
on C 1 and C 2 compounds; the story of the serine cycle and the ethylmalonyl-CoA
pathway. Sci Prog 2011;94.
Asenjo JA, Suk JS. Microbial conversion of methane into poly-β-hydroxybutyrate (PHB):
growth and intracellular product accumulation in a type II methanotroph. J Ferment
Technol 1986;64:2718.
Balasubramanian R, Kenney GE, Rosenzweig AC. Dual pathways for copper uptake by
methanotrophic bacteria. J Biol Chem 2011;286:373139.
Ballows A, Truper H, Dworkin M, Harder W, Schleifer K. The prokaryotes, a handbook on
the biology of bacteria: ecophysiology, isolation, identication, applications. Berlin
Heidelberg New York: Springer; 1992.
Bastien C, Machlin S, Zhang Y, Donaldson K, Hanson R. Organization of genes required
for the oxidation of methanol to formaldehyde in three type II methylotrophs. Appl
Environ Microbiol 1989;55:312430.
Bender M, Conrad R. Kinetics of CH
4
oxidationin oxic soils exposed to ambient air or high
CH
4
mixing ratios. FEMS Microbiol Lett 1992;101:2619.
Benlounes O, Mansouri S, Rabia C, Hocine S. Direct oxidation of methane to oxygenates
over heteropolyanions. J Nat Gas Chem 2008;17:30912.
Bligh E, Dyer WJ. A rapid method of total lipid extraction and purication. Can J Biochem
Physiol 1959;37:9117.
Boetius A, Ravenschlag K, Schubert CJ, Rickert D, Widdel F, Gieseke A, et al. A marine
microbial consortium apparently mediating anaerobic oxidation of methane. Nature
2000;407:6236.
Borgwardt RH. Biomass and natural gas as co-feedstocks for production of fuel for fuel-
cell vehicles. Biomass Bioenergy 1997;12:33345.
Bothe H, Jensen KM, Mergel A, Larsen J, Jørgensen C, Bothe H,et al. Heterotrophic b acteria
growing in association with Methylococcus capsulatus (Bath) in a single cell protein
production process. Appl Microbiol Biotechnol 2002;59:339.
BouaiM, Hebrard G, Bastoul D, Roustan M. A comparative study of gas hold-up, bubble
size, interfacial area and mass transfer coefcients in stirred gasliquid reactors and
bubble columns. Chem Eng Process Process Intensif 2001;40:97111.
Bouwens SM, Vissers JP, de Beer VH, Prins R. Phosphorus poisoning of molybdenum
sulde hydrodesulfurization c atalysts supported on carbon and alumina. J Cata l
1988;112:40110.
Bowman JP, Skerratt JH, Nichols PD,Sly LI. Phospholipid fatty acid and lipopolysaccharide
fatty acid signature lipids in methane-utilizing bacteria. FEMS Microbiol Lett 1991;
85:1521.
610 Q. Fei et al. / Biotechnology Advances 32 (2014) 596614
Bowman JP, Sly LI, NicholsI PD, Hayward A. Revised taxonomy of the methanotrophs:
description of Methylobacter gen. nov., emendation of Methylococcus, validation of
Methylosinus and Methylocystis species, and a proposal that the family Methylococcaceae
includes only the group I methanotrophs. Int J Syst Bacteriol 1993;43:73553.
BowmanJP,SlyLI,StackebrandtE.The phylogenetic position of the family
Methylococcace ae. Int J S yst Bacte riol 1995;45:18 25.
Bowonder B, Sharif MN. Basis for techno-economic policy analysis. Sci Public Policy 1988;
15:2179.
Bratina BJ, Brusseau GA, Hanson RS. Use of 16S rRNA analysis to investigate phylogeny of
methylotrophic bacteria. Int J Syst Bacteriol 1992;42:6458.
Bredwell MD, Worden RM. Masstransfer properties of microbubbles. 1. Experimental
studies. Biotechnol Prog 1998;14:318.
Bredwell MD, Srivastava P, Worden RM. Reactor design issues for synthesisgas fermen-
tations. Biotechnol Prog 1999;15:83444.
Bulygina ES, Galchenko VF, Govorukhina NI, Netrusov AI, Nikitin DI, Trotsenko YA, et al.
Taxonomic studies on methylotrophic bacteria by 5S ribosomal RNA sequencing. J
Gen Microbiol 1990;136:4416.
Buswell A, Sollo Jr F. The mechanismof the methane fermentation. J Am ChemSoc 1948;
70:177880.
But SY, Khmelenina V, Mustakhimov I, TrotsenkoYA. Construction and characterization of
Methylomicrob ium alcaliphil um 20Z knockout mu tants defecti ve in sucrose and
ectoine biosynthesis genes. Microbiology 2013;82:2535.
Calysta_Energy. Biological Gas-to-Liquids® (http://www.calystaenergy.com/technology-
fuels.html) ; 2013.
Cardy D, Laidler V, Salmond G, Murrell J. Molecula r analysis of the me thane mono-
oxygenase (MMO) gene cluster of Methylosinus trichosporium OB3b. Mol Microbiol
1991a;5:33542.
Cardy D, Laidler V, Salmond G, Murrell J. The methane monooxygenase gene cluster of
Methylosinus tri chosporium: cloning and sequencing of the mmoC gene. Arch
Microbiol 1991b;156:47783.
Chang HN, Kim NJ, Kang J, Jeong CM.Biomass-derived volatile fatty acid platform for fuels
and chemicals. Biotechnol Bioprocess Eng 2010;15:110.
Chang HN, Kim NJ, Kang J, Jeong CM, Fei Q, Kim BJ, et al. Multi-stage high cell continuous
fermentation for highproductivity and titer. Bioprocess BiosystEng 2011;34:41931.
Chang HN, Fei Q, Choi JDR, Kim NJ, Kim W. Method for producing microbial intracellular
products from volatile fatty acids. US Patent 20,120,219,993; 2012.
Chang HN, Jung K, Lee JC, Woo H-C. Multi-stage continuous high cell density culture sys-
tems: a review. Biotechnol Adv 2014;32:51425.
Chisti Y. Biodiesel from microalgae. Biotechnol Adv 2007;25:294306.
Chistoserdova L. Modularity of methylotrophy, revisited. Environ Microbiol 2011;13:
260322.
Chistoserdova L, Vorholt JA, Lidstrom ME. A genomic view of methane oxida tion by
aerobic bacteria and anaerobic archaea. Genome Biol 2005;6.
Chistoserdova L, Kalyuzhnaya MG, Lidstrom ME. The expanding world of methylotrophic
metabolism. Annu Rev Microbiol 2009;63:47799.
Choudhary TV, Phillips CB. Renewable fuels via catalytic hydrodeoxygenation. Appl Catal
Gen 2011; 397:112.
Clark TR, Roberto FF. Microbial production of epoxides. US Patent 2003.
Collins MLP, Buchholz LA, Remsen CC. Effect of copper on Methylomonas albus BG8. Appl
Environ Microbiol 1991;57:12614.
Conrado RJ, Gonzalez R. Envisioning the bioconversion of methane to liquid fuels. Science
2014;343:6213.
Cornish A, Nicholls KM, Scott D, Hunter BK, Aston WJ, Higgins IJ, et al. In vivo 13C NMR
investigations of methanol oxidation by the obligate methanotroph Methylosinus
trichosporium OB3B. J Gen Microbiol 1984;130:256575.
Courchesne NMD, Parisien A, Wang B, Lan CQ. Enhancement of lipid production using
biochemical, genetic and transcription factor engineering approaches. J Biotechnol
2009;141:3141.
Csáki R,Bodrossy L, Klem J, MurrellJC, Kovács KL. Genesinvolved in the copper-dependent
regulation of soluble methane monooxygenase of Methylococcus capsulatus (Bath):
cloning, sequencing and mutational analysis. Microbiology 2003;149:178595.
Culpepper MA, Rosenzweig AC. Architecture and active site of particula te methane
monooxygenase. Crit Rev Biochem Mol Biol 2012;47:48392.
Dalton H. Methane oxidation by methanotrophs. Physiological and mechanistic implications.
Biotechnology Hand Book; 1992.
Datar RP, Shenkman RM, Cateni BG, Huhnke RL, Lewi s RS. Fermentation of biomass
generated producer gas to ethanol. Biotechnol Bioeng 2004;86:58794.
Davis R, Aden A, Pienkos PT. Techno-economic analysis of autotrophic microalgae for fuel
production. Appl Energy 2011;88:352431.
Dedysh SN, Knief C, Duneld PF. Methylocella species are facultatively methanotrophic.
J Bacteriol 2005;187:466570.
Dijkhuizen L, Le vering P, De Vries G. The physiol ogy and biochemistry of aerobic
methanol-utilizing gram-negative and gram-positive bacteria. Biotechonl Hand
Book; 1992.
DOE. Reducing Emissions using Methanotrophic Organisms for Transportation Energy
(REMOTE). DE-FOA-0000881 (https://arpa-e-foa.energy.gov/Default.aspx?Search=
DE-FOA-0000881&SearchType=) US. Department of Energy; 2013.
DOE. Natural gas vehicles (http://www.afdc.energy.gov/vehicles/natural_gas.html) US.
Department of Energy; 2014.
Dong Y, Steinberg M. Hynolan economical process fo r methanol production from
biomass and natural gas with reduced CO
2
emission. Int J Hydrogen Energy 1997;22:
9717.
Dry ME. The FischerTropsch process: 19502000. Cata l Today 2002;71:22741.
Duneld P, Dumont R, Moore TR. Methane production and consumption in temperate
and subarctic peat soils: response to temperature and pH. Soil Biol Biochem 1993;
25:3216.
Duneld PF, Khmelenina VN, Suzina NE, Trotsenko YA, Dedysh SN. Methylocella silvestris
sp. nov., a novel methanotroph isolated from an acidic forest cambisol. Int J Syst
Evol Microbiol 2003;53:12319.
Duneld PF, Yuryev A, Senin P, Smirnova AV, Stott MB, Hou S, et al. Methane oxidation by an
extremely acidophilic bacterium of the phylum Verrucomicrobia. Nature 2007;450:
87982.
Dutta A, Talmadge M, Hensley J,Worley M, Dudgeon D, BartonD, et al. Pr ocess design and
economics for conversion of lignocellulosic bi omass to ethanol . NREL/TP-5100 -
51400; 2011 [Contract No. DE-AC36-08GO28308].
EIA. Natural gas gross withdrawals and production. U.S. Energy Information Administra-
tion; 2012a (http://www.eia.gov/dnav/ng/ng_prod_sum_a_epg0_vgv_mmcf_a.htm).
EIA. U.S. Energy Information Administration_Annual Energy Outlook 2013 with Projec-
tions to 2035; 2012b [Document No DOE/EIA-0383 (2012)].
EIA. Gasoline and diesel fuel update (http://www.eia.gov/petroleum/gasdiesel/)U.S.En-
ergy Information Administration; 2014.
Eilers J, Posthuma S, Sie S. The shell middle distillate synthesis process (SMDS). Catal Lett
1990;7:25369.
EPA. Clean alternative fuels: compressed natural gas (http://www.afdc.energy.gov/pdfs/
epa_cng.pdf) U.S Environmental Protection Agency; 2002.
Ettwig KF, Shima S, De Pas-Schoon en V, Katinka T, Kahnt J, Medema MH, et al.
Denitrifying bacteria anaerobically oxidize methane in the absence of Archaea. Envi-
ron Microbiol 2008;10:316473.
Ettwig KF, van Alen T, van dePas-Schoonen KT, Jetten MS, Strous M. Enrichment and mo-
lecular detection of denitrifying methanotrophic bacteria of the NC10 phylum. Appl
Environ Microbiol 2009;75:365662.
Ettwig KF, Butler MK, Le Paslier D, Pelletier E, Mangenot S, Kuypers MM, et al. Nitrite-
driven anaerobic methane oxidation by oxygenic bacteria. Nature 2010;464:5438.
Ezeji TC, Qureshi N, Blaschek HP. Bioproduction of butanol from biomass: from genes to
bioreactors. Curr Opin Biotechnol 2007;18:2207.
Fan J, Yan C, Andre C, Shanklin J, Schwender J, Xu C. Oil accumulation is controlled by car-
bon precursor supply for fatty acid synthesis in Chlamydomonas reinhardtii. Plant Cell
Physiol 2012;53:138090.
Fang J, Barcelona MJ, Semrau JD. Characterization of methanotrophic bacteria on the basis
of intact phospholipid proles. FEMS Microbiol Lett 2000;189:6772.
Fei Q. Biodiesel production from microbial lipids accumulated by oleaginous microorgan-
isms with various carbon sources. KAIST Library: KAIST; 2011.