Microbial Production of Primary Metabolites
Arnold L. Demain
Department of Nutrition and Food Science, Massachusetts Institute of Technology,
Cambridge, Massachusetts, 02139, U.S.A.
Microbial production of primary metabolites contrib-
utes significantly to the quality of life. Through fer-
mentation, microorganisms growing on inexpensive
carbon sources can produce valuable products such
as amino acids, nucleotides, organic acids, and vita-
mins which can be added to food to enhance its flavor
or increase its nutritive value. The contribution of
microorganisms will go well beyond the food industry
with the renewed interest in solvent fermentations.
Microorganisms have the potential to provide many
petroleum-derived products as well as the ethanol nec-
essary for liquid fuel. The role of primary metabolites
and the microbes which produce them will certainly
increase in importance.
The fantastic development of the fermentation in-
dustry has depended on three important character-
istics of microorganisms: (1) a high ratio of surface
area to volume, which facilitates the rapid uptake
of nutrients required to support high rates of metabo-
lism and biosynthesis; (2) a tremendous variety of
reactions which microorganisms are capable of carry-
ing out; and (3) a facility to adapt to a large array
of different environments, allowing a culture to be
transplanted from nature to the laboratory flask,
where it is capable of growing on inexpensive carbon
and nitrogen sources and producing valuable com-
pounds.
The power of the microbial culture in the competitive
world of synthesis can be appreciated by the fact
that even simple molecules, i.e., L-glutamic acid and
L-lysine, are still made by fermentation rather than
by chemical synthesis. Although a few minor battles
have been lost to the chemist (industrial alcohol, sol-
vents) it is obvious that the war has been won by
microbiology. Despite the efficiency of the chemical
route to riboflavin, production of this compound is
still carried out by fermentation as well as by synthe-
582
sis; almost complete chemical processes to vitamin
C and steroids still use microbial bioconversion steps,
and most natural products are so complex and con-
tain so many centers of asymmetry that they probably
will never be made commercially by chemical synthe-
sis. The key to the preservation of the viability of
the fermentation industry is our ability to modify
genetically microbial cultures to states of higher pro-
ductivity.
For many years low-molecular-weight microbial
products have been used to enhance the quality, ap-
peal, and availability of food and drink [1]. A list
of products which have been or are being used in
the food and feed industries is shown in Table 1.
The reason for this impressive use of microbes is
simple: microorganisms have an amazing ability to
utilize cheap sources of carbon and nitrogen to over-
produce valuable low- and high-molecular-weight me-
tabolites. Overproduction is the accumulation of a
metabolite, either intracellularly or extracellularly, to
a level where (based on the total volume of broth)
it exceeds by at least one order of magnitude the
normal concentration required by fastidious microor-
ganisms for optimal growth. The data in Table 2 era-
Table 1, Some fermentation products used in the food or feed
industries
Class Examples
Alcohols Ethanol
Amino acids Glutamic acid, lysine, threonine
Antioxidants Isoascorbic acid
Nucleotides 5'-Guanylic acid, 5'-inosinic acid
Organic acids Acetic acid, propionic acid, succinic acid, fumar-
ic acid, lactic acid, malic acid, tartaric acid,
citric acid, gluconic acid
Polyols Glycerol, mannitol
Protein Single-cell protein
Sugars Fructose, sorbose
Vitamins Riboflavin (B2), cyanocobalamin (B12)
Naturwissenschaften 67, 582-587 (1980) 9 by Springer-Verlag 1980
Table 2. Overproduction of some primary microbial metabolites Table 3. Production of primary metabolites
A. By partial starvation of auxotrophic mutants
Product Growth Production b Ratio
requirement a [rag/l] Produced/
Product Auxotrophic requirement
[mg/1] Required
L-Glutamic acid Glycerol or biotin
Lysine 250 50000 2.0 • 1 0 2 L-Lysine Homoserine
Glutamic acid 300 100000 3.3 • 1 0 2 L-Threonine Lysine, methionine, isoleucine
Inosinic acid 25 13000 5.2 x 102
Riboflavin 0.5 10000 2.0 x 104
Cyanocobalamin 0.001 50 5.0 • 104 B. By selecting antimetabolite-resistant mutants

These concentrations are generally used to give maximum Product Antimetabolite resistance
growth of many microorganisms. They probably are somewhat
in excess L-Threonine c~-Amino-/~-hydroxyvale ric acid
L-Methionine Ethionine
b These concentrations approximately represent the maximum L-Tryptophan 5-Fluorotryptophan
amounts reported in the literature

phasize the fantastic abilities of some overproducing Table 4. Mutation sequence for tyrosine overproduction (data from
microorganisms. [21)
Many of the molecules described above fall into the
Mutation sequence L-Tyrosine
category of primary metabolites. These are the small [g/l]
molecules of all living cells that are intermediates
or end-products of the pathways of intermediary me- Corynebacterium glutamicum (wild type) < 1.0
tabolism, or that are used as building blocks for essen-
Phenylalanine bradytrophy 3.0
tial macromolecules, or are converted into coenzymes.
The most important, from the industrial point of Resistance to 3-amino-L-tyrosine 5.7
view, are the amino acids, nucleotides, vitamins, sol-
vents, and organic acids. From the point of view Resistance to p-amino-DL-phenylalanine 7.5
of the microorganism, overproduction of a primary
metabolite is a wasteful process and usually occurs Resistance to p-fluoro-DL-phenylalanine 12.2

only after some breakdown in the regulatory controls
that have evolved in living systems to prevent oversyn-
thesis. Deregulated cultures are less fit to compete
in nature for survival against other organisms possess-
ing normal regulatory mechanisms. Of course, as in-
dustrial microbiologists we are interested in oversyn-
thesis, and therefore when we find a slightly deregulat-
ed culture in nature or in a culture collection we
do our utmost to further deregulate the culture. By
genetic and environmental modifications, it is possible
to alter regulatory mechanisms thus forcing the mi-
croorganisms to overproduce and excrete primary me-
tabolites of commercial importance.
Amino Acids
Feedback regulation is bypassed by limiting the cell's
ability to accumulate intracellular inhibitory and/or
repressive end-products. This is usually done by pro-
ducing an auxotrophic (i.e., nutritional) mutant and
partially starving it for its requirement. High levels
of amino acids and nucleotides are obtained by such
a procedure (Table 3). The commercial L-lysine fer-
mentation is based on this principle.
Resistance to L-tyrosine hydroxamate 13.5
A second means to bypass feedback regulation is to
produce mutants resistant to a toxic analogue of the
desired metabolite (Table 3).
Table 4 shows how successive resistance mutations
imposed on a partial auxotroph (" bradytroph ") led
to an increase in production of L-tyrosine.
Permeability alteration is very important in the pro-
duction of L-glutamic acid, the major amino acid of
commerce [3]. About 250000 t of monosodium gluta-
mate, a potent flavor enhancer, are made annually,
all by fermentation. The glutamic acid fermentation
was discovered by Kinoshita, Udaka and Shimono
[4]. Although many genera and species are included
in the group o f ' g l u t a m a t e overproducers', e.g., spe-
cies of Micrococcus, Corynebacteriurn, Brevibacterium
and Microbacterium, all are taxonomically similar and
should be included in a single genus. All glutamate
overproducers have a block in the tricarboxylic acid
cycle, i.e., they are deficient in c~-ketoglutarate dehy-
drogenase, thus the carbon flow is shunted to glutam-
ic acid. Normally, glutamic acid overproduction

Naturwissenschaften 67, 582-587 (1980) 9 by Springer-Verlag 1980 583

would not occur due to feedback regulation. How-
ever, due to a decrease in the effectiveness of the
barrier to outward passage, glutamate leaves the cell
thus allowing its biosynthesis to proceed unabated.
The permeability alteration is intentionally effected
by various manipulations including biotin starvation
(practically all glutamate excretors are biotin auxo-
trophs), glycerol starvation of glycerol auxotrophs or
addition of penicillin or fatty acid derivatives to ex-
ponentially growing cells. Apparently all of these ma-
nipulations result in a phospholipid-deficient cyto-
plasmic membrane which favors exit of glutamate
from the cell.
The bulk of the cereals consumed in the world are
deficient in the essential amino acid, L-lysine. Lysine
supplementation converts such cereals into balanced
food or feed. Thanks to the discovery by Kinoshita,
Nakayama and Kitada [5] that homoserine-requiring
mutants of the glutamate overproducer, Corynebac-
terium glutarnicum, produce large amounts of lysine
when grown under the proper conditions, an efficient
fermentation is available which produces over 50 g
lysine/1 at a weight yield of 25% of consumed glu-
cose.
The glutamate overproducers, when mutated to loss
of homoserine dehydrogenase, have proved to be the
best overproducers of lysine. However, concentra-
tions of biotin optimal for growth must be provided
to the culture. If suboptimal concentrations are pro-
vided, glutamate rather than lysine is excreted. Simi-
larly, if penicillin is added to homoserineless C. gluta-
talcum growing in medium containing optimal biotin,
glutamate is produced. The reason is that nitrogen
assimilation by the glutamate overproducers occurs
only through the reductive amination of ~-ketoglutar-
ate to glutamate, the reaction being catalyzed by the
NADP-linked glutamate dehydrogenase. Thus, the ni-
trogen of all of the natural amino acids is derived
from internal glutamate by transamination. When the
permeability mechanism is altered by limitation of
biotin or addition of penicillin, the glutamate is lost
from the cell and is no longer available as an intracel-
lular nitrogen donor for lysine synthesis.
The key to lysine overproduction is the avoidance
of feedback inhibition by: (a) limitation of a feedback
inhibitor of aspartokinase; and by (b) possession of
a feedback-insensitive dihydrodipicolinate synthetase.
Lysine belongs to the aspartic-acid family, being pro-
duced from aspartate in a branched pathway along
with threonine, methionine and isoleucine.
Flavor Nucleotides
The interest in nucleotide fermentations is due to the
ability of three purine ribonucleoside 5'-monophos-
584
phates, namely, guanylic acid (guanosine 5'-mono-
phosphate; GMP), inosinic acid (inosine 5'-mono-
phosphate; IMP); and xafithylic acid (xanthosine 5'-
monophosphate; XMP), to enhance flavor (in order
of decreasing potency). Although the corresponding
deoxyribonucleotides are also active, the following
compounds are not: adenosine 5'-monophosphate
(AMP), 2' and 3' isomers, nucleosides, free bases,
and pyrimidine derivatives.
Much of the early research on nucleotide production
dealt with excretion of nucleotide derivatives that
arise from breakdown of ribonucleic acid during stress
situations [6]. Later the emphasis turned to direct
fermentations of sugar to purine nucleosides and nu-
cleotides by bacterial mutants.
The key to effective purine accumulation is the limita-
tion of intracellular AMP and GMP. This limitation
is best effected by restricted feeding of purine auxo-
trophs. Thus, adenine-requiring mutants accumulate
hypoxanthine and inosine that results from break-
down of intracellularly accumulated IMP.
Until 1964, no successful direct fermentation leading
to production of nucleotides had been reported. An
important advance, however, came with the discovery
that the glutamate overproducers could excrete intact
nucleotides. When grown in media containing
growth-optimal concentrations of biotin, adenine
auxotrophs of C. glutamicum produce IMP. GMP
was found to be produced best by salvage synthesis
using Brevibacterium ammoniagenes. The 'salvage
synthesis' reaction :
purine + PRPP -~ purine nucleotide + pyrophosphate
had been known for years to proceed inside cells
of microbes that are provided with preformed pu-
fines ; it had been demonstrated with cell-free extracts
and with purified enzyme but had not been known
to occur extracellularly with intact cells not given
PRPP.
Investigators soon found that, as long as certain fer-
mentation requirements were met, several purine
bases could be added to wild-type cultures of B. am-
moniagenes for conversion to their respective nucleo-
tides [7]. Thus GMP, the most potent nucleotide fla-
voring agent can be produced by adding guanine to
cells of B. ammoniagenes.
Vitamins
Riboflavin (vitamin B2) is produced commercially by
both fermentation and chemical synthesis. Whereas
the chemically synthesized material is generally used
for pharmaceuticals and food, riboflavin produced
by fermentation is used in animal feed.
Naturwissenschaften 67, 582-587(1980) 9 by Springer-Verlag 1980
The vitamin is generally present in microbial cells
as the coenzymes flavin mononucleotide (FMN) and
flavin adenine nucleotide (FAD), which are bound
to protein. Riboflavin overproducers contain as much
F M N and FAD as normal microorganisms but they
also have a large intracellular pool of free riboflavin.
Normal microorganisms excrete less than 10 mg ri-
boflavin/1. The overproducers can be divided into
three groups. The low overproducers include the clos-
tridia, which can produce, at best, about 100 mg/1.
Yeasts, especially Candida species, are moderate over-
producers that can be made to yield about 600 mg
riboflavin/1. The high overproducers are two yeast-
like molds, Eremothecium ashbyii and Ashbya gossy-
pii, which synthesize riboflavin in concentrations over
10000 mg/1 [8].
The biochemical key to riboflavin overproduction ap-
pears to involve iron. Ferrous ion severely inhibits
riboflavin production by low overproducers; it inhib-
its formation by the moderate overproducers to a
lesser extent, and has no inhibitory action against
E. ashbyii and A. gossypii. I favor a hypothesis that
an iron-flavoprotein is a repressor of riboflavin syn-
thesis. If such an enzyme or a non-enzyme iron-flavo-
protein were the repressor of riboflavin synthesis,
growth in iron-deficient media would produce cells
with little or no repressor, and flavin synthesis would
be derepressed. Another possibility is that iron re-
presses the riboflavin biosynthetic enzymes whereas
riboflavin or a derivative inhibits the first enzyme
of the pathway.
V i t a m i n B12 is the other vitamin made by fermenta-
tion. Two very different types of bacteria are used
by the industry, i.e., Propionibacterium shermanii and
Pseudomonas denitrificans. The key to the P. shermanii
fermentation is apparently avoidance of feedback re-
pression by vitamin B12. Thus the early stage is con-
ducted under anaerobic conditions in the absence of
the precursor 5,6-dimethylbenzimidazole. These con-
ditions prevent vitamin B12 synthesis and allow for
the accumulation of the intermediate, cobinamide.
Then the air is turned on and dimethylbenzimidazole
is added, converting the cobinamide to the vitamin.
In the P. denitr~'cans fermentation, the entire process
is carried out under conditions of low oxygen transfer
but the key to the fermentation is the methyl donor,
betaine (or choline). Vitamin B12 formation by P.
denitr(ficans is totally dependent upon betaine or cho-
line but the mechanism of control is completely un-
known [9].
Organic Acids
Filamentous fungi have been widely used for the com-
mercial production of organic acids. With respect to
Naturwissenschaflen 67, 582 587 (1980) 9 by Springer-Verlag 1980
amounts, citric acid appears to be the major product
made by mycelial fungi. It has been estimated that in
excess of 100 000 t of citric acid are produced per year.
The commerical process!' employs Aspergilhts niger
or one of its mutants. Various processes have been
developed for citric acid production. These are the
Koji fermentation process, the liquid culture shallow
pan process, and the submerged fermentation process.
Shu and Johnson [10] demonstrated that in sub-
merged culture, the production of citric acid took
place in media deficient i0 iron and manganese. The
main features of the fermentation are a high initial
concentration of sugar (about 15%), low levels of
magnesium and iron and :2 pH value below 3.5. Phos-
phate and nitrogen are usually supplied at low levels.
The incubation temperature is around 30 ~ and high-
ly aerobic conditions are necessary. In approximately
8 to 10 days the major portion (80-90%) of the sugar
is converted to citric a.c'id, the titers reaching about
1 O0 g/1. ii
Citric acid is easily assimilated, it is palatable and
exhibits low toxicity. Consequently, it is widely used
in the food and pharmaceutical industry. Citric acid
is employed as an acidifying and flavor enhancing
agent. It also serves as an antioxidant for inhibiting
rancidity in fats and oils. Citric acid and its salts
are used as buffers in such food products as jams
and jellies as well as stabilizers in a variety of foods.
The pharmaceutical industry uses approximately 16%
of the available supply of citric acid.
In recent years, new processes have been developed
for the production of citric acid by Candida yeasts,
especially from hydrocarbons. Such yeasts are able
to convert n-paraffins to citric and isocitric acids in
extremely high yields (150-170% on a weight basis)
[11]. Production of citric acid over isocitric acid is
favored by selecting mutants which are deficient in
the enzyme aconitase [12].
Bioconversions
In addition to the multireaction sequences of fermen-
tations, microorganisms are useful in carrying out
processes in which a compound is converted into a
structurally related product by the use of one or a
small number of enzymes contained in cells. Such
processes are called bioconversions or microbial
transformations and may be carried out with growing
cells, resting cells, spores or dried cells. One of the
earliest bioconversions known was the quantitative
conversion of ethanol to vinegar by the acetic acid
bacteria. This group of bacteria is especially useful
in carrying out incomplete oxidations of organic com-
pounds and is used commercially in the oxidation
585
Table 5. Yields of some bioconversions
Substrate Product
Sorbitol Sorbose
Mannitol Fructose
Glycerol Dihydroxyacetone
Glucose 5-Ketogluconic acid
Glucose 2-Ketogluconic acid
Glucose Gluconic acid
L-Tyrosine L-DOPA
Progesterone 1 l~-Hydroxyprogesterone
of sorbitol to sorbose, the single biological step in
the otherwise chemical production of vitamin C (as-
corbic acid).
Bioconverting organisms are known for practically all
types of chemical reactions. The reactions are stereo-
specific, the ultimate in specificity being exemplified
by the steroid bioconversions. This specificity is ex-
ploited in the resolution of racemic mixtures of amino
acids and intermediates of prostaglandin synthesis.
In many cases, the bioconversion reaction is preferred
over a chemical step when a specific isomer rather
than a racemic mixture is desired. Bioconversions are
characterized by high yields as shown in Table 5.
Other attributes include mild reaction conditions and
the coupling of reactions using a microorganism con-
taining several enzymes working in series.
In developing bioconversions, it is important to exam-
ine the regulation of enzyme synthesis during growth
since the " q u a l i t y " of a bioconverting cell population
depends on the concentration of enzyme in those cells.
Often inducers are useful and it is imperative to avoid
catabolite repression. Mutation can be used to elimi-
nate further catabolism of the desired product. Perme-
ability is often a problem with respect to contact
of the substrate with the enzyme in the cell. In certain
processes, Mn 2 § deficiency or addition of sm-face-ac-
rive agents has been used to decrease the effect of
the permeability barriers. It is sometimes desirable
to grow cells on one substrate and convert a different
substrate; this is known as "cometabolism". Prob-
lems of product inhibition of bioconversions have
been solved by addition of ion-exchange resins or
by dialysis culture. Mixed cultures or sequential addi-
tion of cells have been used to carry out bioconver-
sions involving several steps in series catalyzed by
different cultures. The problem of insoluble sub-
strates, especially prevalent in the steroid field, can
be resolved by using finely divided suspensions of
substrates, suspensions in surface-active agents such
as Tweens, or soluble complexes or esters of sub-
strates. In recent years, there has developed a tremen-
dous interest in immobilized cells to carry out such
586
Microorganism Wt yield [%]
Gluconobacter suboxydans 98
Gluconobacter suboxydans 95
Gluconobaeter suboxydans 95
Gluconobacter suboxydans 90
Pseudomonas mildenbergii 100
Aspergillus niger 97
Aspergillus oryzae 100
Rhizopus nigricans 90
processes. These are usually much more stable than
either free cells or enzymes and are more economical
than immobilized enzymes.
Gluconic acid is produced via bioconversion of glu-
cose by a large variety of fungi including the Aspergil-
lus niger group as well as many species of Penicilliurn.
The medium often contains glucose concentrations
approaching 30%. Gluconic acid in the form of its
calcium salt is used in cases of calcium deficiency.
Sodium gluconate is employed as a sequestering agent
to prevent the deposition of soap scums on cleaned
surfaces. The free acid finds use as a mild acidulant
in a variety of industrial processes (metal processing,
leather tanning, and foods).
Solvents
Ethyl alcohol represents a primary metabolite that
can be produced by fermentation of any carbohydrate
material containing a fermentable sugar or a polysac-
charide that can be depolymerized to a fermentable
sugar. Yeasts are preferred for these fermentations
but the species to be utilized is determined by the
substrate added to the medium. Saccharomyces cere-
visiae is generally employed when the alcohol is derived
from the fermentation of hexoses, whereas Kluyvero-
myces fragilis may be utilized if lactose is the sub-
strate.
Under optimum conditions, within a realistic period
time, approximately 10 to 12% alcohol by volume
is readily obtained. This concentration of alcohol
slows down growth and the fermentation ceases. With
special yeasts, the fermentation can be continued t o
alcohol concentrations in the range of 20% by vol-
ume. However, these concentrations are attained only
after months or years of fermentation, such as in
the case of wines. In general the commercial produc-
tion of alcohol by fermentation is completed within
a five-day period, alcohol concentrations approximat-
Naturwissenschaften 67, 582587 (1980) 9 by Springer-Verlag 1980
ing 12% by volume. At present alcohol is mainly
manufactured by the petrochemical industry from
ethylene. However, with the prices of petroleum in-
creasing, and an expected abundance of grains, it
appears likely that the fermentation process for ethyl
alcohol will be re-instituted. The prospects are that
the fermentation process will be competitive with
those of the chemical process industries. In this re-
gard, organisms such as clostridia are being reex-
amined after years of neglect. Clostridium thermocel-
lure, an anaerobic thermophile, can convert waste cel-
lulose directly to ethanol. Other clostridia produce
acetate, lactate, acetone and butanol and will be uti-
lized more and more as the liquid-fuel crisis deep-
ens.
Naturwissenschaften 67, 58~587 (1980) ~ by Springer-Verlag 1980
i. Demain, A.L. : Syrup. Soc. Gem Microbiol. 21, 77 (1971)
2. Hagino, H., Nakayama, K. : Agr. Biol. Chem. 37, 2013 (1973)
3. Demain, A.L., Birnbaum, J.: Can. Top. Microbiol. Immunol.
46, 1 (1968)
4. Kinoshita, S., Udaka, S., Shimono, M. : J. Gen. Appl. Micro-
biol. 3, 193 (1957)
5. Kinoshita, S., Nakayama, K., Kitida, S. : ibid. 4, 128 (1958)
6. Demain, A.L. : Progr. Ind. Microbiol. 8, 35 (1968)
7. Nara, T., Misawa, M., Kinoshita, S.: Agr. Biol. Chem. 32,
561 (1968)
8. Demain, A.L. : Ann. Rev. Microbiol. 26, 369 (1972)
9. Demain, A.L, White, R.F. : J. Bacteriol. 107, 456 (1971)
10. Shu, P., Johnson, M.J.: ibid. 56, 577 (I948)
I 1. Ikeno, Y., et al. : J. Ferm. Technol. 53, 752 (1975)
12. Akiyama, S., et al. : Agr. Biol. Chem. 37, 885 (1973)
Received June 10, I980
587