Appl Microbiol Biotechnol (1999) 52: 146±153
MINI-REVIEW
L. Eggeling á H. Sahm
L-Glutamateand L-lysine: traditional products with
impetuous developments
Received: 8 December 1998 / Received revision: 1 March 1999 / Accepted: 5 March 1999
Abstract Amino acids have been produced with the aid
of microorganisms for nearly 40 years now. The eco-
nomic importance of these cellular building blocks is
enormous. Demand for them is rising continuously and
currently more than 106 tonnes/year are required. Con-
tinual e€orts to increase production performance are
directed towards the microorganisms themselves, as well
as towards technical improvements of the respective
processes. A special position within the amino-acid-
producing microorganisms is traditionally occupied by
Corynebacterium glutamicum. Molecular research in
conjunction with NMR studies of ¯ux has revealed
fascinating new properties of this particular organism,
including the existence of a new type of exporter and
reverse ¯uxes within the anaplerosis. The knowledge
gained will enable the further improvement of produc-
tion strains and furthermore extend fundamental in-
sights into metabolite ¯ux management within bacteria
in general.
Introduction
More than four decades ago, in 1957, Dr. S. Udeka and
Dr. S. Kinoshita in Japan discovered the bacterium
Corynebacterium glutamicum and its particular property
of excreting L-glutamate. Following this original dis-
covery, mutant strains were developed to produce
3 and price. Prices drop as a consequence of increased
L-glutamate on individual scales of more than 500 m .
C. glutamicum is a gram-positive bacterium that can be
isolated from soil. Very closely related bacteria are
L. Eggeling (&) á H. Sahm
Institut fuÈr Biotechnologie, Forschungszentrum JuÈlich GmbH,
D-52425 JuÈlich, Germany
e-mail: l.eggeling@fz-juelich.de
Tel.: +49-2461-61-5132
Fax: +49-2461-61-2710
Ó Springer-Verlag 1999
Brevibacterium lactofermentum or Brevibacterium ¯avum
(Liebl et al. 1991), Brevibacterium thiogenitalis and
Corynebacterium melassecola. Together with genera
like Streptomyces, Propionibacterium or Arthrobacter,
C. glutamicum belongs to the actinomycetes subdivision
of gram-positive bacteria. Its genome size is 3082 kb
(Bathe et al. 1996), and its entire genome sequence is
expected to be completed in 1999. The genetic tools for
C. glutamicum, enabling gene-directed mutagenesis and
transposon mutagenesis for example, are comparable to
those available for Escherichia coli (Malumbres et al.
1995; Wohlleben et al. 1993; VerteÁs et al. 1994). Cur-
rently, e€ective strains for producing selected amino
acids are available derived from C. glutamicum, E. coli,
or Bacillus subtilis. Constant research on the ¯ux prop-
erties of C. glutamicum reveal this organism to be an
outstanding example of metabolic design. In this review
we will present recent research results on the synthesis of
L-glutamate and L-lysine with C. glutamicum. We will
also include commercial aspects of the development of
amino acid production.
The amino acid market
The amounts of amino acids currently produced can be
seen in Fig. 1. The leader is still L-glutamate produced
by C. glutamicum. This is followed by L-lysine and DL-
methionine while the other amino acids trail behind.
Figure 1 also shows the relationship between volume
production volume, which leads to greater demand and
competition and reduces prices even further. The in-
crease in amino acid demand over the years is signi®-
cant. Within just 10 years the total market has
approximately doubled, with some amino acids dis-
playing a particularly large increase (Fig. 2). For in-
stance, the world market for L-lysine has increased more
than 20-fold in the past two decades. Estimates assume
that the market is currently increasing by 10%±15% per
year. There is therefore considerable movement and
 147

Ile of L-lysine made with C. glutamicum. This is also the

10 2 His
Arg
case with methionine, where chemically made methio-
nine competes with ®sh meal as the natural source of
Ser
Val Orn
Ala Cys methionine. The worldwide decrease in sardine catches,
Tyr
caused by the El NinÄo phenomenon, resulted in a de-
Price in $ / kg

Thr
Leu Phe
Asp creased availability of this natural methionine source.
10 Gly
Thus amino acid consumption is in¯uenced by livestock
and feed production throughout the world and the
prices of natural sources. In addition to the speci®c
Glu
Lys amino acids required as feed supplements, amino acids
1
Met are used for a large variety of purposes, such as chemical
1 10 102 103 104 105 106 building blocks, pharmaceutical products, or food sup-
Production capacity in tons per year plements.
Fig. 1 The amino acids with the largest market are the cheapest
(®gure kindly provided by Prof. W. Leuchtenberger, Degussa AG,
Frankfurt, Germany) A set of reactions at the anaplerotic node

As mentioned, both L-glutamate and L-lysine are made

expansion of capital by the companies producing the
amino acids. The reason for the increasing demand for
L-lysine is that this essential amino acid is required as a
feed additive for poultry and pig breeding. This in-
creased use in animal feeds re¯ects a growing meat
consumption, in particular in South America and China.
Consequently, the requirement for further essential
amino acids used as feed additives, like methionine,
L-threonine and L-tryptophan, is also increasing. This is
also true for the demand for vitamins. Part of the in-
creased amino acid demand is due to increased pollution
control since, with a balanced amino acid content in the
feed, the manure contains less nitrogen.
The actual demand for the amino acids used as feed
supplements can vary signi®cantly. L-Lysine produced
by fermentation competes with the natural L-lysine
source soybean meal, where it is present in high con-
centrations. Since in 1997 the soybean meal market
tightened globally, this led to an increased consumption
L-Glu
64%
Other
D,L-Met L-Lys
26% 8%
1982: 425.000 t
D,L-Met
24%
1991: 800.000 t
Fig. 2 The amino acid market doubles about every 10 years
with C. glutamicum. Of course, a prerequisite for their
intracellular synthesis is the constant supply of the re-
spective citric acid cycle intermediates 2-oxoglutarate
and oxaloacetate. Therefore, great attention has been
paid to the anaplerotic reactions in C. glutamicum, and
phosphoenolpyruvate carboxylase was considered to be
of major importance to replenish the cycle. Surprisingly,
use of the cloned phosphoenolpyruvate carboxylase gene
(O'Regan et al. 1989; Eikmanns et al. 1989) and the
study of an inactivation mutant revealed that this
enzyme activity is not at all essential for growth or
increased amino acid production (Peters-Wendisch et al.
1993). It was therefore initially assumed that the glyox-
ylic acid cycle could serve as an alternative anaplerotic
route. However, a de®ned deletion mutant, devoid of
isocitrate lyase, also showed unaltered amino acid pro-
duction. Only 13C-NMR analyses, which enabled 13CO2
incorporation to be traced, gave de®nite proof of the
presence of a second carboxylating reaction in C. glu-
tamicum (Peters-Wendisch et al. 1996). The investiga-
tion of this enzyme activity resulted in the detection of
pyruvate carboxylase activity (Peters-Wendisch et al.
1997) and the cloning of its gene (Peters-Wendisch et al.
1998). This carboxylase activity had been hard to detect
in the initial enzyme measurements (Tosaka et al. 1979;
Jetten et al. 1995; Cocaign-Bousquet and Lindley 1995)
2%
since it is very unstable in crude extracts and requires an
in situ enzyme assay using permeabilized cells instead of
crude axtracts for reliable determinations. Therefore,
C. glutamicum has the pyruvate dehydrogenase shu‚ing
L-Glu acetyl-CoA into the citric acid cycle, and pyruvate car-
53%
boxylase together with phosphoenolpyruvate carboxyl-
ase supplying oxaloacetate (Fig. 3). Indeed, the two
carboxylases can substitute for each other during growth
Other
on glucose, but the pyruvate carboxylase is the more
4% important reaction, since its deletion has an e€ect on
L-Lys growth. When both enzymes are deleted, growth on
19% glucose is not possible, showing that these enzymes are
the only relevant ones at the anaplerotic node. In addi-
tion to the activities of enzymes supplying the citric acid
cycle with metabolites, oxaloacetate decarboxylase,
148
Pyruvate
PEP
kinase
NADPH Pyruvate
ATP
Malic Pyruvate Oxaloacetate PEP
enzyme carboxylase decarboxylase carboxylase
ADP
Oxaloacetate
NADH Malate
dehydrogenase
Malate
Fig. 3 Enzymes at the anaplerotic node in Corynebacterium glutam-
icum that can contribute to ¯uxes between the pools of C3 units and
C4 units. PEP phosphoenolpyruvate
phosphoenolpyruvate carboxykinase and malic enzyme
activities are present in C. glutamicum catalysing the
decarboxylation of oxaloacetate (Jetten et al. 1994;
Cocaign-Bousquet et al. 1996). Such a rich diversity of
enzymes at the anaplerotic node within an organism has
not been reported previously.
This surprising diversity naturally poses the question:
what is the use of all these activities in vivo? Apparently,
genetic or enzymatic studies alone are completely inad-
equate for identifying the actual ¯uxes in such a network
(Eggeling et al. 1996). Recent extensive 13C-NMR-based
¯ux quanti®cations in a set of isogenic strains have re-
vealed that, in the living cell, in addition to the net
forward ¯ux of C4 unit generation, a reverse ¯ux of C4
unit decarboxylation is in fact operative (Marx et al.
1996). During growth of C. glutamicum the net forward
¯ux is 24 mol%, which is part of the total forward ¯ux
of 96%. Therefore, the exchange ¯ux is 72%. However,
under L-glutamate-producing conditions the net forward
¯ux is increased to 29 mol%, but the exchange ¯ux
drastically reduced to 17% (Marx et al. 1997, 1999).
Apparently the anaplerotic node serves additional
functions other than merely supplying a net oxaloacetate
synthesis. According to the cofactor demand, one can
construct an energy-consuming cycle or a transhydrog-
enase cycle (Fig. 3). To decipher the individual ¯uxes at
this central node in metabolism 13C-NMR quanti®ca-
tions using isotopomer information will be extremely
useful (Wiechert et al. 1997). This will also enable
questions of the function of the individual reactions to
be addressed.
Incorporation of ammonium
The incorporation of ammonium into L-glutamate and
the other amino acids requires the intracellular avail-
ability of this ion. It is available at high extracellular
ammonium concentrations by diffusion of the non-
protonated nitrogen form into the cell. At low ammo-
nium concentrations, however, carrier-mediated ammo-
nium uptake occurs. The corresponding amt gene of
C. glutamicum has recently been identi®ed (Siewe et al.
1996). It encodes a secondary active carrier for the up-
take of the protonated nitrogen form via a uniport
GDP
mechanism. Intracellular NH‡ 4 is incorporated into
PEP L-glutamate via glutamate dehydrogenase activity
carboxykinase
(GDH). RNA studies of the cloned gdh gene have shown
GTP that it is monocistronic and expressed in high levels to
result in a speci®c activity of 1.8 lmol min)1 mg pro-
tein)1 (BoÈrmann et al. 1992). As in other bacteria, in
addition to GDH, glutamine synthetase and glutamate-
oxoglutarate aminotransferase activities are present in
C. glutamicum to ensure L-glutamate synthesis at low
ammonium concentrations. The glutamine synthetase
gene of C. glutamicum, glnA has also been identi®ed
(Jakoby et al. 1997).
The in vivo nitrogen ¯uxes have recently been in-
vestigated in C. glutamicum by sophisticated 15N-NMR
spectroscopy (Fig. 4), making it possible to quantify the
absolute ¯ux rates of the amino-carrying intermediates
important for amino acid production (Tesch et al. 1999).
In several experiments the enrichments in the a-15N-
amino group of L-glutamate and L-glutamine, as well as
that of the d-15N-amino group of L-glutamine, were
determined over time in an in vivo NMR experiment.
Unexpectedly, after the 15NH4 pulse, the labelling times
of L-glutamine in the wild type of C. glutamicum and
that of a constructed GDH deletion mutant are almost
identical. This shows that the glutamine synthetase ac-
tivity does not limit the provision of L-glutamine, which
is important since further evaluation of the labelling
kinetics revealed that a surprisingly large fraction of the
ammonium (about 30%) is assimilated via this enzyme.
The remaining part is assimilated via the GDH activity.
Glutamate
GDH 2.7 Nα
Transaminases
4.6 + direct
incorporation
NH4+
4.5
6.4 Biomass
GOGAT
Direct 1.6
incorporation
Transaminases
0.2 0
GS Nα
1.8
Nδ
Glutamine
Fig. 4 The nitrogen ¯uxes via the glutamate dehydrogenase (GDH),
glutamine synthetase (GS), and aminotransferase (GOGAT) as
quanti®ed in C. glutamicum grown in continuous culture at a dilution
rate of 0.05 1/h, growing with a surplus of ammonium. The numbers
in ovals give molar nitrogen net ¯uxes (lmol min)1 g dry weight)1),
those in hexagons the exchange ¯ux

For the latter reaction a high exchange ¯ux has been
found in vivo, approximately half as high as the net
forward ¯ux in the direction of L-glutamate. This shows
that, in C. glutamicum, the amino group present in
L-glutamate and L-glutamine is available for amino acid
synthesis at a high delivery rate and that the degree of
gdh expression due to the high exchange plays a minor
role in L-glutamate production.
The puzzle of L-glutamate ef¯ux
Although L-glutamate has been produced for almost 40
years, the mechanism of its ef¯ux by C. glutamicum is
not yet understood completely, although some details
are well known. The wild type usually does not excrete
L-glutamate. Three components are likely to be involved
in triggering export: (i) the oxoglutarate dehydrogenase
activity, (ii) the presence of a speci®c exporter and (iii)
the membrane status.
The genes encoding the oxoglutarate dehydrogenase
competing with the glutamate dehydrogenase for the
same substrate have been isolated from C. glutamicum.
In the microorganisms investigated so far the oxo-
glutarate dehydrogenase complex consists of three sep-
arable polypeptides: oxoglutarate dehydrogenase (E1o),
dihydrolipoamide succinyltransferase (E2o), and dihy-
drolipoamide dehydrogenase (E3). In C. glutamicum the
lpd gene, encoding the dihydrolipoamide dehydrogenase
domain, has a size of 1407 bp up to 42% identity at the
amino acid level with functionally characterized genes of
other bacteria (Schwinde 1995). The gene is mono-
cistronic with identical transcription and translation
initiation sites. Interestingly, the two further domains
are apparently represented in C. glutamicum by a new
type of gene structure (Usuda et al. 1996). The odhA
gene product, having approximately the size of
E1o + E2o, has 44% identity in its C-terminal part with
E1o of E. coli. In its N-terminal part there are signi®cant
identities with E2o, although the lipoyl domain and
peripheral subunit-binding domain speci®c to E2o, when
present as a single polypeptide, are absent.
This novel type of gene structure implies an unusual
quarternary structure of the oxoglutarate dehydrogenase
complex. In fact, the complex appears to be very labile,
and its puri®cation has not yet been possible. Originally
oxoglutarate dehydrogenase activity was even thought
to be absent from C. glutamicum. The unusual structure
of the oxoglutarate dehydrogenase complex might be
one reason for the superior L-glutamate-producing
properties of C. glutamicum. Exposing the cell to con-
ditions resulting in L-glutamate excretion (see below)
also reduces the 2-oxoglutarate dehydrogenase activity
to a residual activity of only 10%, whereas the activity of
the glutamate dehydrogenase is hardly affected (Kawa-
hara et al. 1997). Thus an excess conversion of 2-
oxoglutarate to succinyl-CoA is prevented, but its
conversion to L-glutamate favoured. The wild type, with
disrupted odhA, excretes L-glutamate, even in the pres-
149
ence of biotin, with productivities almost equal to that
under biotin limitation, enabling the accumulation of
250 mM L-glutamate (Kawahara et al. 1997).
Whereas originally it was assumed that L-glutamate
``leaks'' through the membrane, or that ``inverted'' up-
take systems might be involved in its ef¯ux, several
arguments and general mechanistic reasons point to the
fact that a carrier protein must be involved in the export
(KraÈmer 1994). For instance, without a carrier an ``up-
hill'' transport of L-glutamate towards the very high
concentrations obtained in fermentation broths would
not be possible (Hoischen and KraÈmer 1989). Also the
speci®city of the L-glutamate export observed requires
the presence of a carrier. However, the driving force of
the carrier is still unknown (Gutman et al. 1992), as is
any molecular detail of the carrier involved.
Although traditionally the e‚ux of L-glutamate is
triggered by biotin limitation, the possibilities of
achieving excretion are surprisingly diverse: (i) growth
under biotin limitation, (ii) use of oleic acid auxotrophs,
(iii) use of glycerol auxotrophs, (vi) addition of penicil-
lin, (v) addition of detergents and (vi) addition of local
anaesthetics result in ef¯ux. To obtain massive ef¯ux the
presence of one of these compounds is not suf®cient, but
rather the cells must be grown for at least one generation
in their presence, indicating that the cell composition
must be altered. Therefore, the membrane status is ap-
parently correlated with ef¯ux. In practice, detergents
like polyoxyethylene sorbitan monopalmitate (Tween
40) or other detergents are applied. They are used to-
gether with mutants sensitive to the respective detergent
so that only a limited amount of detergent is required in
the production.
Molecular access to detergent-mediated e‚ux was
made possible by cloning dtsR1 (Kimura et al. 1996).
This gene is immediately adjacent to the orthologous
gene dtsR2. Although the catalytic function of dtsR1 is
unknown, it is likely to be related to fatty acid synthesis,
since growth of a dtsR1 mutant is promoted by addition
of oleate or oleate esters (Kimura et al. 1997). Since
overexpression of dtsR1 reduces the L-glutamate ef¯ux
when this is triggered either by biotin limitation or by
addition of detergents or penicillin, dtsR1 is an element
in the chain of events leading to L-glutamate excretion.
In the current model it is assumed that, under L-gluta-
mate-excreting conditions, enzyme complexes consisting
of DtsR1/AccBC and DtsR2/AccBC are destabilized,
thereby most likely in¯uencing fatty acid synthesis. Both
complexes have been detected experimentally (Kimura
1998, personal communication). AccBC is a biotin-
containing protein (JaÈger et al. 1996). It can be con-
cluded from the sequence identities that AccBC is the
a-chain of an acyl-CoA carboxylase, and that DtsR1
and DtsR2 are the b-chain of such an enzyme. Acyl-
CoA carboxylase activities are required to provide the
building blocks for fatty acids and mycolic acids. Al-
though the substrates of these enzymes are still unclear,
the features of the protein complexes link biotin limita-
tion with fatty acid synthesis at the molecular level.
150

Taken together, the special glutamate-ef¯ux properties dipicolinate reductase, respectively (Cremer et al. 1990).
of C. glutamicum might be attributed to the happy co- Open reading frame 2 (orf2) encodes a non-essential
incidence of several circumstances, with lipid synthesis gene (Pisabarro et al. 1993). The structure of the orf4
undoubtedly playing a major role. Although di€erent gene product, as well as the consequences of its partial
ideas and concepts are now available to explain some inactivation, suggests that orf4 is involved in transla-
speci®c features of glutamate e‚ux in C. glutamicum, tional control (Patek et al. 1997). The synthase is located
this important research target will remain a major focus at the ¯ux distribution of aspartate semialdehyde to-
in the future. wards either homoserine or L-lysine. Although the syn-
thase is not regulated in its catalytic activity, as is the
case in E. coli or plants, its total activity is pivotal for
The genes of L-lysine biosynthesis L-lysine accumulation (Eggeling et al. 1998). When a
series of graded dapA expression was used, a global re-
Almost all the genes of C. glutamicum required for sponse of the carbon metabolism to the synthase activity
L-lysine synthesis have been cloned (Fig. 5). Some of became apparent. The increased expression results in an
them are clustered, which is therefore different from increased ¯ux towards L-lysine, which is unexpectedly
E. coli where they are scattered over the chromosome. accompanied by a decreased ¯ux towards L-threonine.
The kinase-encoding gene lysC consists, in fact, of lysCa This is the reason for a decreased growth rate, with an
and lysCb in a very special gene arrangement (Kali- apparently increased intracellular pyruvate availability
nowski et al. 1991). One promoter drives lysCa and asd favourable for L-lysine synthesis. A further favourable
transcription, and a second promoter is present within consequence of increased dapA expression is that the
lysCa to result in a lysCb and asd transcript (Cremer concentrations of by-product amino acids like L-threo-
et al. 1991; Follettie et al. 1993). Since lysCb is an in- nine or L-aspartate are reduced (Kircher 1998). This
frame constituent part of lysCa, the amino acid sequence study is an example of how only slightly modi®ed ex-
of the b subunit is identical to that in the carboxy- pressions in a synthesis pathway can in¯uence the entire
terminal part of the a subunit, suggesting an interesting carbon balance of the cell.
architecture of the aspartate kinase. The regulatory
features of the kinase reside in the b subunit (Kalinowski
et al. 1990; Cremer et al. 1991; Follettie et al. 1993). The split pathway of DL-diaminopimelate synthesis
With one or more additional copies of lysC, an increased
L-lysine accumulation can be obtained (Cremer et al. In eubacteria there are three variants for the synthesis
1991; Jetten and Sinskey 1995). of DL-diaminopimelate, which is required for L-lysine
The second cluster consists of dapB-orf2-dapA-orf4 and peptidoglycan synthesis. An outstanding charac-
(Patek et al. 1997). The functionally identi®ed genes teristic of C. glutamicum is that this bacterium uses
encode the dihydrodipicolinate synthase and dihydro- both the succinylase and the dehydrogenase variant
(Fig. 5). Mutants with either inactivated dehydroge-
nase or a succinylase variant are still prototrophic
during growth with ammonia as a nitrogen source
Aspartate (Wehrmann et al. 1994), but lysine accumulation is
lysC Aspartate reduced (Schrumpf et al. 1991). As demonstrated by
13
asd
kinase
C- and 1H-NMR spectroscopy, the ¯ux partitioning
lysCα β asd
is dependent on the ammonium concentration (Son-
Aspartate semialdehyde
dapA Dihydrodi- ntag et al. 1993). Therefore, the relative use of the
dapB
picolinate dehydrogenase variant for the contribution of L-lysine
synthase
synthesis can vary from 72% to 0%. This is due to
Piperideine dicarboxylate
dapB dapA the low af®nity of the dehydrogenase towards am-
dapD monium. The luxury of having both variants together
Dehydro- gives C. glutamicum an increased ¯exibility in response
dapD aroP dapE dapE genase to changing environmental conditions and is related to
the essential need for DL-diaminopimelate as the
Diaminopimelate linking unit of peptidoglycan (Wehrmann et al. 1998).
argS lysA lysA Whereas the dehydrogenase variant is energetically less
expensive and used when ammonium is in excess, the
lysE Permease
succinylase is energetically more costly and used at
lysG lysE L-Lysine low ammonium concentrations. This system of DL-
diaminopimelate synthesis is therefore a new example
Fig. 5 Overview of the organization of lysine biosynthesis genes and of twin activities present in bacteria allowing access to
their transcripts, and the function of their gene products in the
pathway. The genes and their products not shown beside the pathway
important metabolites, such as the well-known twin
are the following: aroP aromatic amino acid uptake system; argS activities for the synthesis of glutamate, or those for
arginyl-tRNA synthetase, lysG the regulator of lysE transcription the uptake of potassium.

L-Lysine export
L-Lysine ef¯ux has been characterized as a secondary
active-transport system where lysine translocation is
thought to be coupled to H+ in¯ux or OH) e‚ux (BroÈer
and KraÈmer 1991). Recently the speci®c L-lysine export
carrier has been cloned (Vrljic et al. 1996). As expected,
the exporter LysE is a membrane protein, but quite
unexpectedly it is a comparatively small protein of
25.4 kDa. A PhoA-fusion analysis revealed a topology
so far unknown for any translocator (Fig. 6). Although
six hydrophobic domains were identi®ed on the basis of
hydropathy analyses, only ®ve membrane-spanning
helices are present, as concluded from the PhoA fusion
studies. The additional hydrophobic segment may dip
into the membrane or be localized on the surface (Vrljic
et al. 1999). Moreover, recent genome projects have re-
vealed the existence of LysE homologues enabling the
characterization of three distinct families (LysE, YahN,
and CadD). These three families, the members of which
are of similar size and have similar hydropathy pro®les
and sequence motifs, comprise the LysE superfamily, all
members of which occur in eubacteria. Most interest-
ingly, the functionally characterized members of the
LysE superfamily export L-lysine, cadmium and possibly
quarternary amines. It is therefore suggested that LysE
superfamily members will turn out to catalyse the export
of a variety of biologically important solutes.
Access to the carrier gene lysE has also solved the
puzzle of why C. glutamicum has such an exporter at all.
In a lysE deletion mutant, supplied with glucose and a
1 mM concentration of the dipeptide lysyl-alanine, an
extraordinarily high intracellular L-lysine concentration
(more than 1 M) accumulates, abolishing the growth of
that mutant (Vrljic et al. 1996). Thus the exporter serves
as a valve that excretes any excess intracellular L-lysine
present in the natural situation when growth is in the
presence of peptides, or in the producer strains where
control of L-lysine synthesis has been removed. As in
NH2
Interior
T W
N
P N N
D L D F
M
COOH Exterior
Fig. 6 Model of the topology of the lysE product encoding the
L-lysine exporter of C. glutamicum. Conserved aminoacyl residues
proposed to be of major importance for translocation are given in the
single-letter code
151
other bacteria, C. glutamicum has active peptide-uptake
systems (Erdmann et al. 1993) as well as hydrolysing
enzymes giving access to the amino acids as valuable
building blocks. However, C. glutamicum has no
L-lysine-degrading activities and therefore must prevent
the amino acid from piling-up. Of course, the exporter is
under tight control to prevent loss of L-lysine in the case
of its expensive synthesis from glucose. The auto-
genously controlled transcriptional regulator lysG serves
this purpose (Fig. 5). In conclusion, the lysine exporter
of C. glutamicum is not only an example of a new target
to improve amino acid production but also represents a
new translocator structure, which provides a new type of
intracellular amino acid control.
L-Lysine technology
The commonest carbon sources for L-lysine fermenta-
tion and also for other amino acids are molasses (cane or
sugar-beet molasses), high-test molasses (inverted cane
molasses), or sucrose and starch hydrolysates. However,
the use of molasses has severe disadvantages, since the
seasonal availability causes ageing, affecting its quality
during storage. Pro®table nitrogen sources are ammo-
nium sulfate and ammonia (gaseous or ammonia water).
A typical lysine fermentation is shown in Fig. 7. After
consumption of the initially supplied sugar, the sub-
strates are added continuously. L-Lysine accumulates in
concentrations up to 170 g/l. Together with sugar, am-
monium has to be fed and is supplied in the form of
ammonium sulfate to provide the counterion to neu-
tralize the accumulating basic amino acid. Therefore,
L-lysine is present in the fermentation broth as its sul-
fate. Since amino acids are fairly cheap products, a high
sugar conversion yield is a very important criterion for
the entire production process. Appropriate restrained
growth strategies can be used to maximize yield or
productivity (Kiss and Stephanopoulos 1991), and yields
200
Growth
150
L-Lysine (g/l)
20
100
Sugar (g/l)
10
50
10 20 30 40
Time (h)
Fig. 7 Time course of L-lysine accumulation in a production plant
152
of 54% L-lysine from glucose have been reported (Ka-
wahara et al. 1990).
For the work-up procedure of the L-lysine formed,
i.e. the downstreaming, several basically different pro-
cesses have been developed that lead to different forms
of the product. Three processes are currently in use to
supply L-lysine in a form suitable for feed purposes. One
is the isolation of lysine by ion-exchange chromatogra-
phy, evaporation and crystallization to yield a prepara-
tion containing 98.5% L-lysine hydrochloride. The
second process involves biomass separation, evapora-
tion and ®ltration to give an alkaline solution of con-
centrated L-lysine containing (50%). The third, very
ef®cient procedure uses the direct spray-drying of the
entire fermentation broth followed by granulation to
yield a lysine sulfate preparation consisting of 47%
L-lysine. This is the most ef®cient procedure since there
are no losses in downstreaming and the waste volume is
minimized.
Outlook
The constant e€orts to understand and improve amino
acid production have proved worthwhile. Integrated
research activities ± the only possible way to study the
whole cell ± have revealed the eciency of the bacteria
very much more clearly. Molecular studies on amino
acid synthesis have provided a few surprises. Exciting
questions are, for example, whether there are other
amino acid exporters or what part is played by the LysE
homologues within the eubacteria. In general, the prin-
ciples of how carbon metabolism is engineered in central
reactions and biosynthesis pathways are largely under-
stood (Eggeling and Sahm 1999). Further attractive
targets of research are, for instance, the linkage of car-
bon and nitrogen metabolism, using information about
the genome, the role of the energy budget and global
regulatory interactions. In conclusion, research on the
amino-acid-producing microorganisms is a ¯ourishing
and rapidly developing ®eld, as is the production of the
amino acids themselves.
Acknowledgements Our own work appearing in this review was
supported by grants from Degussa AG, Frankfurt. We thank B.J.
Eikmanns, and W. Pfe€erle for discussions, and E. Kimura and
A.A. de Graaf for contributing data prior to publication.
References
Bathe B, Kalinowski J, PuÈhler A (1996) A physical and genetic map
of the Corynebacterium glutamicum ATCC 13032 chromosome.
Mol Gen Genet 252: 255±265
BoÈrmann E, Eikmanns BJ, Sahm H (1992) Molecular analysis of
the Corynebacterium glutamicum gdh gene encoding glutamate
dehydrogenase. Mol Microbiol 6: 317±326
BroÈer S, KraÈmer R (1991) Lysine excretion by Corynebacterium
glutamicum. 2. Energetics and mechanism of the transport
system. Eur J Biochem 202: 137±143
Cocaign-Bousquet M, Lindley ND (1995) Pyruvate over¯ow and
carbon ¯ux within the central metabolic pathways of Coryne-
bacterium glutamicum during growth on lactate. Enzyme
Microbiol Technol 17: 260±267
Cocaign-Bousquet M, Guyonvarch A, Lindley ND (1996) Growth-
rate dependent modulation of carbon ¯ux through central me-
tabolism and the kinetic consequences for glucose-limited che-
mostat cultures of Corynebacterium glutamicum. Appl Environ
Microbiol 62: 429±436
Cremer J, Eggeling L, Sahm H (1990) Cloning the dapA dapB
cluster of the lysine-secreting bacterium Corynebacterium glu-
tamicum. Mol Gen Genet 220: 478±480
Cremer J, Eggeling L, Sahm H (1991) Control of the lysine bio-
synthetic sequence in Corynebacterium glutamicum as analyzed
by overexpression of the individual corresponding genes. Appl
Environ Microbiol 57: 1746±1752
Eggeling L, Oberle S, Sahm H (1998) Improved L-lysine yield with
Corynebacterium glutamicum: Use of dapA resulting in in-
creased ¯ux combined with growth limitation. Appl Microbiol
Biotechnol 49: 24±30
Eggeling L, Sahm H (1999) Amino acid production: principles of
metabolic engineering. In: Papoutsakis ET, Lee HF (eds)
Metabolic engineering. Dekker, New York Basel (in press)
Eggeling L, Sahm H, Graaf AA de (1996) Quantifying and di-
recting metabolite ¯ux: application to amino acid overproduc-
tion. Adv Biochem Eng Biotechnol 54: 1±30
Eikmanns BJ, Follettie MT, Griot MU, Sinskey AJ (1989) The
phosphoenolpyruvate carboxylase gene of Corynebacterium
glutamicum: molecular cloning nucleotide sequence and ex-
pression. Mol Gen Genet 218: 330±339
Erdmann A, Weil B, KraÈmer R (1993) Lysine secretion by wild-
type Corynebacterium glutamicum triggered by dipeptide up-
take. J Gen Microbiol 139: 3115±3122
Follettie MT, Peoples OP, Agoropoulou C, Sinskey AJ (1993)
Gene structure and expression of the Corynebacterium ¯avum
N13 ask-asd operon. J Bacteriol 175: 4096±4103
Gutmann M, Hoischen C, KraÈmer R (1992) Carrier-mediated
glutamate secretion by Corynebacterium glutamicum under
biotin limitation. Biochim Biophys Acta 1112: 115±123
Hoischen C, KraÈmer R (1989) Evidence for an e‚ux carrier system
involved in the secretion of glutamate by Corynebacterium
glutamicum. Arch Microbiol 151: 342±347
JaÈger W, Peters-Wendisch PG, Kalinowski J, PuÈhler A (1996) A
Corynebacterium glutamicum gene encoding a two-domain
protein similar to biotin carboxylases and biotin-carboxyl-car-
rier proteins Arch Microbiol 166: 76±82
Jakoby M, Tesch M, Sahm H, KraÈmer R, Burkowski A (1997) Iso-
lation of the Corynebacterium glutamicum glnA gene encoding
glutamine synthetase I. FEMS Microbiol Lett 154: 81±88
Jetten MSM, Sinskey AJ (1995) Recent advances in the physiology
and genetics of amino acid-producing bacteria. Crit Rev Bio-
technol 15: 73±103
Jetten MSM, Pitoc GA, Follettie MT, Sinskey AJ (1994) Regula-
tion of phospho(enol)-pyruvate- and oxaloacetate-converting
enzymes in Corynebacterium glutamicum. Appl Microbiol Bio-
technol 41: 47±52
Jetten MSM, Follettie MT, Sinskey AJ (1995) E€ect of di€erent
levels of aspartokinase on the lysine production by Corynebac-
terium lactofermentum. Appl Microbiol Biotechnol 41: 76±82
Kalinowski J, Bachmann B, Thierbach G, PuÈhler A (1990) As-
partokinase genes lysCa and lysCb overlap and are adjacent to
the aspartate semialdehyde dehydrogenase gene asd in Co-
rynebacterium glutamicum. Mol Gen Genet 224: 317±324
Kalinowski J, Cremer J, Bachmann B, Eggeling L, Sahm H, PuÈhler
A (1991) Genetic and biochemical analysis of the aspartokinase
from Corynebacterium glutamicum. Mol Microbiol 5: 1197±
1204
Kawahara Y, Yoshihara Y, Ikeda, Yoshii H, Hirose Y (1990)
Stimulatory e€ect of glycine betaine on L-lysine fermentation.
Appl Microbiol Biotechnol 34: 87±90
Kawahara Y, Takahasi-Fuke K, Shimizu E, Nakamatsu T, Nak-
amori S (1997) Relationship between the glutamate production
and activity of 2-oxoglutarate dehydrogenase in Brevibacterium
lactofermentum. Biosci Biotechnol Biochem 61: 1109±1112

Kimura E, Abe C, Kawahara Y, Nakamatsu T (1996) Molecular
cloning of a novel gene, dtsR, which rescues the detergent
sensitivity of a mutant derived from Brevibacterium lac-
tofermentum. Biosci Biotechnol Biochem 60: 565±1570
Kimura E, Abe C, Kawahara Y, Nakamatsu T, Tokuda H (1997)
A dtsR gene-disrupted mutant of Brevibacterium lac-
tofermentum requires fatty acids for growth and eciently
produces L-glutamate in the presence of an excess of biotin.
Biochem Biophys Res Commun 234: 157±161
Kircher M (1998) Amino acids as feed additives. Mag Soc Ind
Microbiol 48: 4±11
Kiss RD, Stephanopoulos G (1991) Metabolic activity control of
the L-lysine fermentation by restrained growth fed-batch strat-
egies. Biotechnol Prog 7: 501±509
KraÈmer R (1994) Secretion of amino acids by bacteria: physiology
and mechanism. FEMS Microbiol Rev 13: 75±79
Liebl W, Ehrmann M, Ludwig W, Schleifer KH (1991) Transfer of
Brevibacterium divaricatum DSM 20297 Brevibacterium ¯avum
DSM 20412 and DSM 1412 and Corynebacterium lilium DSM
20137 to Corynebacterium glutamicum and their distinction by
rRNA gene restriction patterns. Int J Syst Bacteriol 41: 225±260
Malumbres M, Mateos L, Martin JF (1995) Microorganisms for
amino acid production: Escherichia coli and corynebacteria. In:
Hui YH, Khachtourians GG (eds) Food biotechnology. Verlag
Chemie, Weinheim New York, pp 423±469
Marx A, Graaf AA de, Wiechert W, Eggeling L, Sahm H (1996)
Determination of the ¯uxes in the central metabolism of Co-
rynebacterium glutamicum by NMR spectroscopy combined
with metabolite balancing. Biotechnol Bioeng 49: 111±129
Marx A, Striegel K, Graaf AA de, Sahm H, Eggeling L (1997)
Response of the central metabolism of Corynebacterium glu-
tamicum to di€erent ¯ux burdens. Biotechnol Bioeng 6: 168±180
Marx A, Eikmanns BJ, Sahm H, Graaf AA de, Eggeling L (1999)
Response of the central metabolism of Corynebacterium glu-
tamicum to the use of an NADH-dependent glutamate dehy-
drogenase. Metabol Eng 1: 35±48
O'Regan M, Thierbach G, Bachmann B, Villeval D, Lepage P,
Viret F, Lemoine Y (1989) Cloning and nucleotide sequence of
the phosphoenolpyruvate carboxylase-coding gene of Coryne-
bacterium glutamicum ATCC13032. Gene 77: 237±251
Patek M, Bilic M, Krumbach K, Eikmanns B, Sahm H, Eggeling L
(1997) Isolation and transcriptional analysis of the dapB-orf2-
dapA-orf4 operon of Corynebacterium glutamicum encoding
two enzymes involved in L-lysine synthesis. Biotechnol Lett 19:
1113±1117
Peters-Wendisch P, Eikmanns BJ, Thierbach G, Bachmann B,
Sahm H (1993) Phosphoenolpyruvate carboxylase in Coryne-
bacterium glutamicum is dispensible for growth and lysine
production. FEMS Microbiol Lett 112: 269±274
Peters-Wendisch PG, Wendisch VF, Graaf AA de, Eikmanns BJ,
Sahm H (1996) C3-carboxylation as an anaplerotic reaction in
phosphoenolpyruvate carboxylase-de®cient Corynebacterium
glutamicum. Arch Microbiol 165: 387±396
Peters-Wendisch PG, Wendisch VF, Paul S, Eikmanns BJ, Sahm H
(1997) Pyruvate carboxylase as an anaplerotic enzyme in Co-
rynebacterium glutamicum. Microbiology 143: 1095±1103
Peters-Wendisch P, Kreutzer C, Kalinowski J, Patek M, Sahm H,
Eikmanns BJ (1998) Pyruvate carboxylase from Corynebacte-
rium glutamicum: characterization, expression and inactivation
of the pyc gene. Microbiology 134: 1095±1103
153
Pisabarro A, Malumbres M, Mateos LM, Oguiza JA, Martin JF
(1993) A cluster of three genes (dapA, orf2, and dapB) of
Brevibacterium lactofermentum encodes dihydrodipicolinate
synthase, dihydrodipicolinate reductase, and a third polypep-
tide of unknown function. J Bacteriol 175: 2743±2749
Schrumpf B, Schwarzer A, Kalinowski J, PuÈhler A, Eggeling L,
Sahm H (1991) A functionally split pathway for lysine synthesis
in Corynebacterium glutamicum. J Bacteriol 173: 4510±4516
Schwinde J (1995) PhD thesis, University of DuÈsseldorf, Germany
Shiio I, Ujigawa-Takeda K (1980) Presence and regulation of
a-ketoglutarate dehydrogenase complex in a glutamate-pro-
ducing bacterium Brevibacterium ¯avum. Agric Biol Chem 44:
1897±1904
Siewe RM, Weil B, Burkovski A, Eikmanns BJ, Eikmanns M,
KraÈmer R (1996) Functional and genetic characterization of the
(methyl)ammonium uptake carrier system of Corynebacterium
glutamicum. J Biol Chem 271: 5398±5403
Sonntag K, Eggeling L, De Graaf AA, Sahm H (1993) Flux par-
titioning in the split pathway of lysine synthesis in Corynebac-
terium glutamicum: quanti®cation by 13C- and 1H-NMR
spectroscopy. Eur J Biochem 213: 1325±1331
Tesch M, de Graaf AA Sahm H (1999) In vivo activities of the
ammonia assimilatory pathways in Corynebacterium glutamic-
um studied with 15N.NMR. Appl Environ Microbiol 65: 1099±
1109
Tosaka O, Morioka H, Takinami K (1979) The role of biotin-
dependent pyruvate carboxylase in L-lysine production. Agric
Biol Chem 43: 1513±1519
Usuda Y, Tujimoto N, Abe C, Asakura Y, Kimura E, Kawahara
Y, Kurahashi O, Matsui H (1996) Molecular cloning of the
Corynebacterium glutamicum (Brevibacterium lactofermentum
AJ12036) odhA gene encoding a novel type of 2-oxoglutarate
dehydrogenase. Microbiology 142: 3347±3354
VerteÁs AA, Asai Y, Inui M, Kobayashi M, Kurusu Y, Yukawa H
(1994) Transposon mutagenesis of coryneform bacteria. Mol
Gen Genet 245: 397±405
Vrljic M, Sahm H, Eggeling L (1996) A new type of transporter
with a new type of cellular function: L-lysine export in Co-
rynebacterium glutamicum Mol Microbiol 22: 815±826
Vrljic M, Bellmann A, Garg J, Wachi S, Freudl R, Malecki MJ,
Sahm H, Kozina VJ, Eggeling L, Saier MH Jr (1999) The LysE
superfamily: topology of the lysine exporter LysE of Coryne-
bacterium glutamicum, a paradyme for a novel superfamily of
transmembrane solute translocators. J Mol Microbiol Bio-
technol (in press)
Wehrmann A, Eggeling L, Sahm H (1994) Analysis of di€erent
DNA-fragments of Corynebacterium glutamicum complement-
ing dapE of Escherichia coli. Microbiology 140: 3349±3356
Wehrmann A, Phillip B, Sahm H, Eggeling L (1998) Di€erent
modes of diaminopimelate synthesis and their role in cell wall
integrity: a study with Corynebacterium glutamicum. J Bacteriol
177: 3159±3165
Wiechert W, Graaf AA de (1997) Bidirectional reaction steps in
metabolic engineering: modeling and simulation of carbon
isotope labeling experiments. Biotechnol Bioeng 55: 101±117
Wohlleben W, Muth G, Kalinowski J (1993) Genetic engineering of
gram positive bacteria. In: PuÈhler (ed) A genetic engineering of
microorganisms. VCH, Weinheim New York Basel