Professional Documents
Culture Documents
www.elsevier.com/locate/biochi
Abstract
Single cell oils (SCOs) are now produced by various microorganisms as commercial sources of arachidonic acid (ARA) and docosahexaenoic acid (DHA). These oils are now used extensively as dietary supplements in infant formulas. An understanding of the underlying
biochemistry and genetics of oil accumulation in such microorganisms is therefore essential if lipid yields are to be improved. Also an
understanding of the biosynthetic pathways involved in the production of these polyunsaturated fatty acids (PUFAs) is also highly desirable as
a prerequisite to increasing their content in the oils. An account is provided of the biosynthetic machinery that is necessary to achieve oil
accumulation in an oleaginous species where it can account for lipid build up in excess of 70% of the cell biomass. Whilst PUFA production
in most microorganisms uses a conventional fatty acid synthase (FAS) system followed by a series of desaturases and elongases, in
Schizochytrium sp., and probably related thraustochytrid marine protists, PUFA synthesis now appears to be via a polyketide synthase (PKS)
route. This route is discussed. It clearly represents a major departure from conventional fatty acid biosynthesis, possibly as a means of
decreasing the amount of NADPH that is needed in the overall process.
2004 Elsevier SAS. All rights reserved.
Keywords: Fatty acid synthase (FAS); Lipid metabolon; Malic enzyme; Polyketide synthase (PKS); Polyunsaturated fatty acids
1. Introduction
The production of microbial oilsotherwise referred to
as Single Cell Oils (SCO)is now an economic reality.
Although such entities have long been suggested as viable
alternatives to plant oils and fats, the first commercial production of an SCO did not begin until 1985 and this only
lasted for 6 years before it was closed down as no longer
being cost-effective. This oil was produced using Mucor circinelloides and was designed to be rich in the polyunsaturated fatty acid (PUFA), c-linolenic acid (GLA; 18:3, n-6),
for use as an alternative to the then expensive evening primrose oil [1]. The process involved the fungus being grown in
stirred-tank fermenters of 220 m3 capacity with subsequent
harvesting and drying of the biomass, followed by oil extraction and its final refinement and purification. Although the
production did not last many years, it nevertheless established that microorganisms could be realistic commercial
alternatives to some plant oils and that the microbial oil itself
could be extracted and purified using conventional technology without the need for expensive, purpose-built units dealing only with microbial oils. All that remained for SCOs to
re-emerge as economically viable oils in their own right was
to identify an oil, or oils, that were not easy to produce agriculturally and which were intrinsically expensive: the microbial oil had to be saleable at a high price in order that it could
cover the costs of its production. These requirements are now
met by the production of very long chain PUFA which do not
occur to any extent in plant oils and which, until the development of SCO processes could only be obtained from marine
animal sources. These fatty acids are now established as important dietary nutrients for neonatal babies and their mothers and consequently a large market is available to ensure
their worldwide sales [2].
During the late 1980s and early 1990s, it became clear that
PUFAs were of increasing interest in nutrition and especially
808
in the development of newly-born babies. Clear clinical evidence began to accumulate that indicated that docosahexaenoic acid (DHA; 22:6, n-3) and arachidonic acid (ARA;
20:4, n-6) were of particular importance as these fatty acids
formed the greater majority of fatty acids found in brain tissue and were also present in mothers milk but absent from
both cows milk and infant formulas that are frequently used
in place of mothers milk. It was then established that both
these fatty acids were involved in the development of neural
and retinal functions and could thus be of benefit to babies for
them to achieve improved memory and eyesight [37]. The
problem was where one might obtain sufficient quantities of
these PUFAs to offer as nutritional supplements.
Although DHA has long been known to be a major fatty
acid of fish oils, it always occurs along with eicosapentaenoic acid (EPA; 20:5, n-3) which was contra-indicated to
be included in infant diets as it affects the uptake of DHA [8]
which is crucial for neural development. Also, there were,
and still are, severe doubts about the use of fish oils as
supplements because of the presence of environmental manmade pollutants, such as dioxins, PCBs and heavy metals including mercury compounds being taken up by fish and concentrated in the liver and other organs. Indeed some
countries, most conspicuously the USA, have an outright ban
on fish oils being given to infants and young children for
these very reasons. As the need for DHA in particular became apparent, it was appreciated by one commercial company (Martek Inc., in the USA) that DHA was a major component of some algal lipids and, if a likely organism could be
found that could be grown in large-scale industrial fermentors, then a source of DHA might be realisable [9]. Simultaneously, it was also known that ARA could be obtained from
microbial sources as work carried out in Japan in the late
1980s had shown this PUFA to be a major component of the
oil from Mortierella alpina [10].
To-day, there are at least three different fermentation processes each using a different microorganism for the production of DHA. For the production of ARA, various strains of
M. alpina are used in separate processes in Europe, China
and possibly Japan. These processes, together with the production organisms involved, are summarised in Table 1. For
inclusion in infant formulas, the ARASCO and
DHASCO are mixed together in the ratio of 2:1 and sold
under the trade name of Formulaid. This is incorporated into
formula preparations in over 60 countries worldwide; approximately 700 tons of this oil mixture was produced in
2003 and the expectation is that this will exceed 1000 tons in
2004 [11].
It is of importance to note that the organisms used in the
SCO processes all have very high contents of the desired
Table 1
Fatty acid profiles (given as relative % w/w) of SCOs in current production
14:0
16:0
18:0
18:1
(A) ARA-SCO processes using M. alpina strains
0.4
8
11
14
DSM
process a
0.2
6
2
4
Wuhan
Alking
process b
12:0
14:0
16:0
16:1
(B) DHA-SCO processes
4
20
18
2
Martek
process c
(DHASCO)
13
29
12
Omega-
Tech
process d
(DHASCO-S)
3
30
Nutrinova
process e
(DHActive)
18:2
18:3 (n-6)
20:3 (n-6)
20:4 (n-6)
22:0
24:0
49
70
18:0
18:1
18:2
18:3 (n-3)
20:3 (n-6)
22:5 (n-6)
22:6 (n-3)
0.4
15
0.6
39
12
25
11
46
a
The DSM process is run in Italy using fermenters of about 100 m3 capacity. The oil content of the cells is considered to be over 40% (w/w). It is sold
exclusively to Martek Biosciences Corp. and is mixed at a ratio of 2:1 (v/v) with DHASCO for incorporation into infant formulas under the trade name
Formulaid. Current production (2003) was 480 tons [11].
b
This is run by Wuhan Alking Engineering Co. Ltd., Wuhan, China. The process is run in fermenters of 50100 m3 capacity. Information about the process
and the oil is limited but refer to: http://www.alking.com.cn/english/aboutus.htm. It has, however, been recently (September 2004) announced that the US
agri-food company, Cargill, is to begin marketing this fatty acid probably outside USA and Europe.
c
Uses C. cohnii which produces over 40% of its biomass as oil. The process is run by Martek Biosciences Corp. using a number of fermenters each about
100 m3. The oil, together with the ARA oil, is used exclusively for infant formulas. Current production (2003) was 240 tons [11].
d
Uses Schizochytruim sp. This oil, formerly known as DHAGold is produced by OmegaTech Inc., Boulder, CO, now owned by Martek. The oil content of the
cells is over 40%. The oil is currently aimed at the adult nutritional supplement market although it has been previously used for improvement of poultry and
other animal feeds. Approximately 10 tons of oil were produced in 2003.
e
Uses Ulkenia sp., which is grown in a 80 m3 fermenter. The oil is produced by Nutrinova GmbH, Frankfurt, Germany and is sold under the trade name of
DHActive [23].
809
(1)
810
(2)
It is this enzyme whose activity is up-regulated at the onset of nitrogen limitation in the growth medium of the oleaginous microorganism possibly as a means of trying to scavenge additional ammonium ions from intracellular materials.
Nitrogen limitation in the cultivation of an oleaginous microorganism induces a cascade of reactions leading to the
formation of acetyl-CoA:
At the onset of nitrogen exhaustion, oleaginous cells show
an increased activity of AMP deaminase, which is up to
fivefold greater than in cells before N limitation.
The increased activity of AMP deaminase decreases the
cellular content of AMP, including its content in the mitochondrion.
Fig. 1. A scheme to show how the citrate/malate cycle and the cytosolic transhydrogenase cycle could provide sufficient precursors of acetyl-CoA and
NADPH for lipogenesis in oleaginous microorganisms (from Refs. [14,15]). Enzymes: 1, pyruvate decarboxylase; 2, malate dehydrogenase; 3, malic enzyme;
4, pyruvate dehydrogenase; 5, citrate synthase; 6, ATP:citrate lyase; 7, citrate/malate translocase. Net carbon balance: pyruvate acetyl-CoA + CO2. Net
reaction for NADPH production: NADH + NADP+ + ATP NAD+ + NADPH + Pi. The transhydrogenase cycle can operate independently of the carbon flux
from citrate in the mitochondrion to acetyl-CoA in the cytosol and consequently can provide all the NADPH needed for both fatty acid biosynthesis and fatty
acid desaturation and elongation reactions.
811
Fig. 2. A diagram to show the organisation of a hypothesised lipogenic metabolon. The flux of carbon from the mitochondrion, via citrate efflux and acetyl-CoA
formation in the cytosol, thence into fatty acids and finally into long chain PUFAs (LCPUFA) occurring in the membranes of the endoplasmic reticulum, is
shown by the continuous lines. The system uses pyruvate (from glycolysis) as the provider of intramitochondrial acetyl-CoA and for citric acid production as is
shown in Fig. 1 but which is not repeated here for clarity. OAA: oxaloacetate; AC-CoA: acetyl-CoA; Mal-CoA: malonyl-CoA; FAS: fatty acid synthase; ACL:
ATP:citrate lyase; ACC: acetyl-CoA carboxylase; CMT: citrate/malate translocase; ME: malic enzyme; PC: pyruvate carboxylase; MDH: malate dehydrogenase. The yellow enzymes form the citrate/malate cycle while the purple enzymes represent the cytosolic transhydrogenase cycle (see also Fig. 1). (From
Ref. [14].)
(3)
812
Fig. 3. Pathways for the formation of PUFA in microorganisms using the conventional FAS route. Fatty acids are synthesised from acetyl-CoA and malonylCoA (which is itself formed from acetyl-CoA via acetyl-CoA carboxylasesee Fig. 2) using the FAS complex of enzymes. The saturated fatty acid, stearic acid
is then successively desaturated and elongated through a series of reactions leading to the formation of various PUFAs. PUFAs fall into two categories, the
n-3 and n-6 series, depending on the position of the final double bond nearest the terminal methyl group. In M. alpina, which is used for the production of ARA,
there is a D17 (n-3) desaturase that can then form EPA under certain conditions. It is possible that a similar n-3 desaturase may occur in some organisms that
produce DPA(n-6), which would then be converted by this enzyme into DHA (see text for further discussion).
which are formed from the fatty acids via phosphatic acid
and its subsequent dephosphorylation to a diacylglycerol followed by a final acylation step ([14], and see also [17] for a
detailed review of this subject). In some photosynthesising
microalgae, not under current consideration for commercial
production of PUFA-rich oils, triacylglycerols are not the
major lipid component and, instead, there are numerous lipid
types present that are involved in the make-up of the photosynthesising membrane systems. This complexity of lipids in
these organisms is disadvantageous to them being used commercially. Also these organisms, being obligate phototrophs,
pose severe difficulties for their large-scale cultivation which
have not yet been overcome.
As is shown in Table 1, a key feature of the range of fatty
acids found in Schizochytrium sp. and Ulkenia sp., both of
which are heterotrophic eukaryotic organisms belonging to
the order of Thraustochytriales within the phylum of Labyrinthulomycota, is the occurrence of a second PUFA, docosapentaenoic acid (DPA, 22:5) along with DHA. This fatty acid
occurs, not as a member of the expected n-3 series (i.e. like
DHA), but as the n-6 isomersee Fig. 3. Although DPA is
unusual fatty acid, it also occurs in small amounts in animal
lipids and extensive animal trials in the safety of this oil
[18,19] have established that DPA is not detrimental to the
dietary usefulness of DHA. Although it does not act as a
813
Fig. 4. A suggested scheme to account for the formation of DHA by the PKS
route of synthesis in Schizochytrium sp. and probably other thraustochyrid
organisms. (See text for further information concerning the genetic organisation of the PKS in Schizochytrium sp.) The sequence would begin with the
transfer of both acetyl-CoA and malonyl-CoA into their corresponding esters with ACP. These are then condensed via KSalso sometimes known as
the condensing enzymeto give the 3-ketobutyryl intermediate, which is
then converted into butyryl-ACP using KR, followed by a DH and finally by
an ER. This is then the equivalent of what occurs in a conventional FAS
mediated system. However, in successive additions of malonyl-CoA to the
fatty acyl chain, the double bond introduced by the DH can be retained. If
the sequence follows the accepted route of fatty acid biosynthesis in bacteria, and to which this pathway is related, then the double bond would initially be in the trans configuration and would moreover be at the 2,3position. An isomerase is therefore conjectured which will convert the bond
into the cis configuration and simultaneously move it to the 3,4-position as
shown, for example, in the conversion of 6:1(n-4) to 6:1(n-3). Successive
additions of malonyl-CoA will lead to a lengthening chain in which the double bonds are retained if they are in the right position. It is suggested that in
the diagram after the formation of 8:1(cis-5), the sequence will proceed to
10:2(cis-4,cis-7), then to 12:3(c3,c6,c9) followed by 14:3(5,8,11),
16:4(4,7,10,13), 18:5(3,6,9,12,15), 20:5(5,8,11,14,17) and finally DHA at
22:6(4,7,10,13,16,19). To accommodate the presence of the D4-desaturase
identified by Qui et al. [25] one would have to speculate that 20:5(n-3), the
penultimate intermediate would be converted first to 22:5(7,10,13,16,19)
and that this would be the substrate for this final desaturase. However, to
create 22:5(n-6), which is found in the thraustochytrids, an entirely different
sequence to the one given in the figure and indicated above would have to be
followed in which a n-6 series of fatty acids [beginning with 10:1(cis-4)]
were synthesised. Such a sequence would not involve the participation of a
D4-desaturase. (A similar PKS route to DHA biosynthesis also occurs in
Ulkenia sp.see Table 1 and also text.)
814
possibly synthesised in this particular species by a conventional series of elongases and desaturases similar to that
shown above in Fig. 3. This, however, ignores the close taxonomic similarity between Schizochytrium and Thraustochytrium and also ignores that the profiles of the fatty acids
from each organism are somewhat similar [26] with both organisms having DPA(n-6) rather than DPA(n-3). It would
therefore seem reasonable to suggest that pathways of fatty
acid biosynthesis in these two organisms should be the same.
No genetic analysis, though, was carried out on the Thraustochytrium sp. examined by Qiu et al. [25] and thus some
caution should be exercised before assuming that this species
does not have the PKS system as identified in the
Schizochytrium sp. by Metz et al. [20]. However, the presence of the D4 desaturase would appear incontrovertible but
no evidence has yet been provided for any other fatty acid
desaturase being detected, either at the genetic or protein
level in any of the DHA-synthesising microorganisms mentioned in Table 1.
With regard to the synthesis of the C14 to C18 fatty acids
which occur in all the DHA-synthesising microorganisms, it
is not clear if there is a separate conventional FAS system
that operates alongside the PKS system or whether these
fatty acids are also produced by the latter pathway. Analysis
of appropriate genetic mutants would clearly provide an unambiguous answer to this and other outstanding questions
concerning PUFA biosynthesis.
The operation of a PKS system helps to explain why many
of the DHA-producing organisms do not have intermediate
chain-length fatty acyl groups in their lipids; the PKS system
would seem to be like a drain-pipe; the building units (acetyl
and malonyl-CoA) go in at one end and the final product
emerges at the other. Intermediates are apparently too tightly
bound to be released as free CoA esters and to be available
for triacylglycerol biosynthesis. Although no work has yet
been reported on the mechanism of DHA biosynthesis in C.
cohnii that would satisfactorily account for DHA being the
sole PUFA being produced by this organism, it will be of
clear interest to determine whether it uses a similar PKS system to the Schizochytrium one; if this is found to be the case
then it would also help to explain why it has always been
difficult to identify conventional, eukaryotic-like desaturases
in C. cohnii (C. Ishiwaka, J.P.Wynn and C. Ratledge, unpublished work).
A PKS system would also serve to decrease the overall
requirement for NADPH being needed both for the reduction
of the growing fatty acid acyl chain (mainly at the level of the
reduction of the 2,3-enoyl intermediate to leave an unsaturated fatty acyl chain for the next condensation reactionsee
Fig. 4) and for the operation of desaturases used in the formation of unsaturated fatty acids in all other systems. As
these desaturases are also linked to O2, the PKS system of
PUFA production is therefore, like the bacterial FAS system
an anaerobic (non-oxygen requiring) system. From an evolutionary point of view, it would seem sensible to avoid having to use NADPH in two separate sequences: first in produc-
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
815