Annu. Rev. Biomed. Eng. 1999. 01:535–557 Copyright q 1999 by Annual Reviews. All rights reserved 1523–9829/99/0820–0535$08.00 535 Metabolic Engineering M. Koffas, C. Roberge, K. Lee, and G. Stephanopoulos Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139; e-mail: [email protected] Key Words metabolism, flux, pathway Abstract Metabolic engineering is the science that combines systematic analysis of metabolic and other pathways with molecular biological techniques to improve cellular properties by designing and implementing rational genetic modifications. As such, metabolic engineering deals with the measurement of metabolic fluxes and elu- cidation of their control as determinants of metabolic function and cell physiology. A novel aspect of metabolic engineering is that it departs from the traditional reduc- tionist paradigm of cellular metabolism, taking instead a holistic view. In this sense, metabolic engineering is well suited as a framework for the analysis of genome-wide differential gene expression data, in combination with data on protein content and in vivo metabolic fluxes. The insights of the integrated view of metabolism generated by metabolic engineering will have profound implications in biotechnological appli- cations, as well as in devising rational strategies for target selection for screening candidate drugs or designing gene therapies. In this article we review basic concepts of metabolic engineering and provide examples of applications in the production of primary and secondary metabolites, improving cellular properties, and biomedical engineering. CONTENTS Introduction ..................................................................................... 535 Metabolic Fluxes .............................................................................. 537 Primary Metabolites .......................................................................... 539 Secondary Metabolites ....................................................................... 542 Cell and Tissue Engineering ............................................................... 545 Biomedical Applications .................................................................... 546 Inborn Errors of Metabolism ............................................................... 547 Cell and Organ Physiology ................................................................. 548 Concluding Remarks ......................................................................... 550 INTRODUCTION Altering metabolic pathways to improve cell properties and the chances of cell survival is as old as nature itself. The genomic and metabolic evolution of extre- mophiles is an example of such a natural adaptation process in bacteria (96). The Annu. Rev. Biomed. Eng. 1999.1:535-557. Downloaded from www.annualreviews.org by RENSSELAER POLYTECHNIC INSTITUTE on 06/17/14. For personal use only.
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Annu. Rev. Biomed. Eng. 1999. 01:535–557Copyright q 1999 by Annual Reviews. All rights reserved
1523–9829/99/0820–0535$08.00 535
Metabolic Engineering
M. Koffas, C. Roberge, K. Lee, and G. StephanopoulosDepartment of Chemical Engineering, Massachusetts Institute of Technology,Cambridge, Massachusetts 02139; e-mail: [email protected]
Key Words metabolism, flux, pathway
Abstract Metabolic engineering is the science that combines systematic analysisof metabolic and other pathways with molecular biological techniques to improvecellular properties by designing and implementing rational genetic modifications. Assuch, metabolic engineering deals with the measurement of metabolic fluxes and elu-cidation of their control as determinants of metabolic function and cell physiology.A novel aspect of metabolic engineering is that it departs from the traditional reduc-tionist paradigm of cellular metabolism, taking instead a holistic view. In this sense,metabolic engineering is well suited as a framework for the analysis of genome-widedifferential gene expression data, in combination with data on protein content and invivo metabolic fluxes. The insights of the integrated view of metabolism generatedby metabolic engineering will have profound implications in biotechnological appli-cations, as well as in devising rational strategies for target selection for screeningcandidate drugs or designing gene therapies. In this article we review basic conceptsof metabolic engineering and provide examples of applications in the production ofprimary and secondary metabolites, improving cellular properties, and biomedicalengineering.
CONTENTS
Introduction ..................................................................................... 535Metabolic Fluxes .............................................................................. 537Primary Metabolites .......................................................................... 539Secondary Metabolites ....................................................................... 542Cell and Tissue Engineering ............................................................... 545Biomedical Applications .................................................................... 546
Inborn Errors of Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547Cell and Organ Physiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 548
Concluding Remarks ......................................................................... 550
INTRODUCTION
Altering metabolic pathways to improve cell properties and the chances of cellsurvival is as old as nature itself. The genomic and metabolic evolution of extre-mophiles is an example of such a natural adaptation process in bacteria (96). The
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
536 KOFFAS ET AL
intentional manipulation of metabolic pathways by humans to improve the prop-erties and productivity of microorganisms is similarly an established concept.Techniques such as genetic modifications via random mutagenesis have yielded,for example, improved strains of Corynebacterium glutamicum and its relatedspecies Brevibacterium lactofermentum and Brevibacterium flavum that excretelarge amounts of amino acids into the fermentation medium (52, 80, 106). Ran-dom mutagenesis relies heavily on chemical mutagens and creative selection tech-niques to identify superior strains for achieving a certain objective. Suchtraditional genetic approaches for strain improvement have been applied exten-sively in the past also in the areas of antibiotics, solvents, and vitamin productionamong others.
Since the past decade, the development of recombinant DNA techniques hasintroduced a new dimension to pathway manipulation by offering, for the firsttime, the capability to construct specific metabolic configurations with novel,beneficial characteristics. Genetic engineering allows precise modification of spe-cific enzymatic reactions in metabolic pathways, leading to the construction ofwell-defined genetic backgrounds. The redirection of cellular metabolism to cre-ate or enhance desirable attributes has been accomplished with a variety of noveltechniques and applied towards an even greater variety of goals.
In this context, metabolic engineering has emerged as the technological andscientific discipline dealing with the introduction of specific modifications to met-abolic pathways to improve cellular properties. Metabolic engineering involvesmanipulation of enzymatic, transport, and regulatory functions of the cell by usingrecombinant DNA technology (5, 86). First, various analytical techniques are usedto identify and subsequently determine fluxes through critical metabolic pathwaysin the cell or tissue of interest. This knowledge provides the rational basis forapplying, in the second step, molecular biological techniques to enhance meta-bolic flux through a pathway of interest and minimize metabolic flow to undesiredbiosynthetically related products. Although a certain sense of direction is inherentin all strain improvement programs, the directionality of effort is a strong focalpoint of metabolic engineering, compared with random mutagenesis, because thisdirectionality plays a dominant role in enzymatic target selection, experimentaldesign, and data analysis.
Although various terms have been coined over the past two decades to rep-resent the increasing activity in pathway modification (pathway engineering, cel-lular engineering, in vitro evolution, etc) (58, 65), the term metabolic engineeringhas succeeded in capturing the ever growing interest in this area. Furthermore,although initially embodied as a collection of examples from the chemical indus-try and biomedical research, metabolic engineering is quickly becoming a distinctscientific field. Its novel contribution lies in its emphasis on complete metabolicnetworks rather than individual reactions. To elaborate, as with all traditionalfields of engineering, metabolic engineering too encompasses the two definingsteps of analysis and synthesis. Because metabolic engineering emerged withDNA recombination as the enabling technology, its initial focus was on synthesis
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 537
in the form of new pathway construction. As such, differentiation from geneticengineering was initially diffuse, and metabolic engineering could be consideredas the technological manifestation of applied molecular biology. The real contri-bution of metabolic engineering emerged as soon as a need for a more rationalapproach to identifying promising targets of metabolic manipulation was articu-lated, replacing the previous, mostly ad hoc target selection process. In this sensethe contribution of metabolic engineering emanates from pathway analysis, whichyields an enhanced perspective on metabolism and cellular function, includingconsideration of reactions in their entirety rather than in isolation. Thus, metabolicengineering seeks to analyze and then synthesize and design, using techniquesand information developed from extensive reductionist research.
Metabolic engineering has found many applications, especially in microbialfermentation. It has been applied to increase the production of chemicals that arealready produced by the host organism (e.g. 13, 15, 22, 34, 92), to produce desiredchemical substances from less expensive feedstocks (e.g. 6, 53, 107), and togenerate products that are new to the host organism (e.g. 67, 68). Other challengesassociated with metabolic engineering are the biosynthesis of secondary metab-olites, the generation of organisms with desirable growth characteristics, and themanipulation of pathways for the production of chiral compounds as intermedi-ates in the synthesis of pharmaceutical products (38). Finally, although less widelyappreciated, metabolic engineering techniques can also be applied for studyingphysiological systems and isolated whole organs in vivo to elucidate the metabolicpatterns that occur in different physiological states, such as ‘‘fed’’ or ‘‘fasted,’’as well as in disease (105).
METABOLIC FLUXES
Because metabolic pathways and fluxes are at the core of metabolic engineering,it is important to elaborate on their definition. A metabolic pathway is defined asany sequence of feasible and observable biochemical-reaction steps connecting aspecified set of input and output metabolites. The pathway flux is the rate at whichinput metabolites are processed to form output metabolites (84). The importanceof feasibility and observability should be noted in the definition, in view of thediversity and complexity of various metabolic maps. If some metabolic fluxes ina pathway or metabolic network cannot be determined independently, it is betterto lump these reactions together, because their inclusion provides no additionalinformation.
The determination of intracellular fluxes, along with analysis of factors affect-ing flux distributions, is collectively referred to as metabolic flux analysis (MFA).MFA combines data on uptake and secretion rates, biosynthetic requirements,metabolic stoichiometry, and quasi–steady-state mass balances on metabolicintermediates to determine intracellular metabolic fluxes (84). MFA has been thefocus of attention of many researchers in the past and has yielded important
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
538 KOFFAS ET AL
information on flux distribution and its control in many bacteria. The combinationof analytical methods to quantify fluxes and their control with molecular bio-logical techniques to implement suggested genetic modifications is the essenceof metabolic engineering. An iterative cycle of genetic change followed by ananalysis of its consequences and design of further modifications, analogous tothat articulated for protein engineering, can also be applied in the developmentof an optimized strain (3). The flux thus becomes a focal point of metabolicengineering and justifies further research for the development of methods for itsdetermination in vivo.
MFA also reveals the degree of pathway engagement in the overall metabolicprocess. Furthermore, elucidation of the control of flux provides a mechanisticbasis for rationalizing observed fluxes and flux distributions at key metabolicbranch points (86). As these fluxes are determined under in vivo conditions, MFAalso allows valid comparisons to be made between in vivo and in vitro enzymaticbehavior. Finally, the flux is a fundamental determinant of cell physiology and acritical parameter to use when comparing the behavior of strain variants. Even ifthe fermentation characteristics of such strains differ, such differences may berelatively unimportant if the flux distributions around key branch points have notbeen altered.
Intracellular fluxes are calculated by MFA with a stoichiometric model for theinternal reactions and mass balances around intracellular metabolites (98).Required data are obtained from measured extracellular fluxes, such as substrateuptake rates and metabolite secretion rates (85). Intracellular fluxes from reactionsthat are cyclical or that split and later converge may be determined by assayingasymmetries in the distribution of radio-labeled atoms of intermediate metabolites(77, 83). The metabolic fluxes determined by MFA provide a comprehensiveperspective of the control system at work in metabolic networks. This controlsystem can be described in terms of metabolic control coefficients (30, 41). Fluxcontrol coefficients are measures of the degree of control exercised by an enzymeon the overall network flux and can be determined by measuring flux responsesto metabolic perturbations.
As mentioned in the previous paragraphs, elucidation of the flux control struc-ture of metabolic pathways offers tremendous opportunities for the rational designof the optimal genotype of a cellular catalyst. This activity should be viewed ascomplementary to molecular biological toolboxes for implementing gene transfersand other similar modifications. In fact, recombinant technology has advanced sorapidly in recent years that rational analysis of metabolic pathways for the iden-tification of target genes and enzymes is the limiting component in the directedoptimization of cellular function. Evidence for this assertion is the observationthat currently, almost 20 years after the pioneering developments in genetic engi-neering, we have hardly begun to harness the potential of modern biotechnologyin the areas of fuels, chemicals, or materials production.
In the following sections, we review a few illustrative applications of meta-bolic-pathway manipulation. We organize the various applications of metabolic
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 539
engineering into four basic groups: (a) improving primary metabolite production,(b) improving secondary metabolite and biopharmaceutical production, (c)improving cell properties, and (d) improving the biomedical field. We have threegoals in reviewing these applications: first, to provide a sample of the truly broadrange of possibilities for biocatalyst improvement afforded by pathway manipu-lation and metabolic engineering; second, to alert the reader to the complexity ofmetabolic pathways, along with their regulation and need of coordination withoverall metabolism; and, third, to underscore the methods used for effectingdesired changes in cellular systems for industrial use or medical reasons. Thisreview concludes with some ideas and suggestions for future directions of thefield.
PRIMARY METABOLITES
A large number of mainly industrial applications can be classified as primarymetabolites. A central goal of metabolic engineering is to achieve the productionof commodity chemicals by overexpressing key metabolic pathways that alreadyexist in the host organism or by introducing new routes of metabolism. We nextreview efforts in metabolic engineering to improve the yield and productivity ofethanol, amino acids, and solvents.
A major challenge of using biotechnology in the industrial production of fuelshas long been the construction of microorganisms that are able to ferment inex-pensive and abundant carbon resources, such as lignocellulosic materials, intoethanol for use as a biofuel, among other applications. The most commonly usedethanol producer, Saccharomyces cerevisiae, cannot ferment pentoses, which mayconstitute 8%–28% of lignocellulose. On the other hand, other bacteria such asErwinia chrysanthemi, Klebsiella planticola, and Escherichia coli can grow effi-ciently on a wide range of carbon substrates that includes five-carbon sugars, butcompeting pathways divert carbon flow away from ethanol production. Initialstudies were only partially successful in redirecting fermentative metabolism toethanol production in these bacteria, by amplifying their pyruvate decarboxylaseactivity (7, 91, 92). A further improvement of the process was achieved by cloningand overexpressing the Zymomonas mobilis adhB gene, yielding recombinants ofE coli (34) and Klebsiella oxytoca (67, 68, 104) that efficiently ferment a varietyof sugars to ethanol. Ohta et al investigated the expression of the pyruvate decar-boxylase and alcohol dehydrogenase genes of Z. mobilis in a related enteric bac-terium, K. oxytoca (67, 68). Klebsiella strains have two additional fermentationpathways not present in E. coli, which are used to convert pyruvate to succinateand butanediol. As for E. coli, it was possible to divert .90% of the carbon flowfrom sugar catabolism away from the native fermentative pathways and towardethanol.
Amino acid production is also a heavily researched area. Tryptophan synthesisin E. coli is highly regulated by a complex set of feedback mechanisms. By
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
540 KOFFAS ET AL
transducing each of several mutations one at a time, researchers combined, withina single strain, a long list of alterations to these mechanisms, thus creating atryptophan overproducer (1, 82). A Corynebacterium glutamicum strain able toproduce 18 g liter11 of tryptophan has been altered to produce large amounts oftyrosine (26 g liter11) by overexpressing deregulated 3-deoxy-D-arabino-heptu-losonate-7-phosphate synthase and chorismate mutase (33). Overexpression of anadditional gene, prephenate dehydratase, in the previous construct led to the pre-dominant production of phenylalanine (28 g liter11). In a similar way, significantprogress was made recently in efforts to construct efficient threonine-producingstrains by metabolic engineering. The genes that are involved in the threonineproduction pathway of C. glutamicum were cloned, and the regulatory propertiesof the enzymes encoded by them were well characterized (e.g. 20, 22, 39). Thecloning of a deregulated (threonine-insensitive) homoserine dehydrogenase, alongwith modulation of the activity of homoserine kinase relative to that of homo-serine dehydrogenase (the first two enzymes in the threonine pathway), yieldedthreonine-secreting strains (14). Similarly, overexpression of ilvA, the first genein the isoleucine pathway, allowed significant isoleucine accumulation by B. lac-tofermentum strains (15).
Another industrially significant primary metabolite is 1,3-propanediol (1,3-PD), an intermediate in chemical and polymer synthesis, for example, in thesynthesis of polyurethanes and polyesters. 1,3-PD is currently derived from petro-leum, and it is expensive to produce relative to similar diols. Tong & Cameron(93) constructed an E. coli propanediol-producing strain carrying genes from theKlebsiella pneumoniae dihydroxyacetone (dha) regulon. These genes allow thestrain to grow anaerobically on glycerol and produce 1,3-PD. A further processimprovement via metabolic engineering in the same field is in production of 1,3-PD by using sugars as a carbon source, because sugars are significantly lessexpensive than glycerol. No known natural organism ferments sugars directly to1,3-PD. One way to replace, at least partially, the need for glycerol is by cofer-mentation of glycerol and a sugar such as glucose. Cofermentation is not possiblewith native 1,3-PD producers because glucose represses the 1,3-PD pathway. E.coli that has been transformed with the K. pneumoniae 1,3-PD oxidoreductasegene (dhaT) and the glycerol dehydratase gene (dhaB) of the same organism isable to coferment glycerol and glucose (93, 94). The glycerol is converted pri-marily to 1,3-PD, and the glucose is used for growth and regeneration of reducingpotential. Until today the production of 1,3-PD by fermenting sugars alone hasnot been possible to a commercially advantageous extent. The initial success withthe metabolic engineering of 1,3-PD production, coupled with the complexity anddifficulty of rational process improvement with Thermoanaerobacterium ther-mosaccharolyticum, the best naturally occurring organism for the fermentation ofsugars to 1,2-propanediol, led to the construction of pathways in E. coli similarto those used by T. thermosaccharolyticum. This was accomplished by overex-pressing the native glycerol dehydrogenase of E. coli or cloning the aldolasereductase gene from rat lens cDNA into E. coli. Either of these two strains led to
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 541
the production of 1,2-PD in equal amounts in batch and continuous culture. Theseresults provide the first example of recombinant organisms able to ferment sugarsto 1,2-PD and demonstrate that various strategies for the fermentation of sugarsto 1,3-PD are possible (9).
Another example of converting native metabolic intermediates to desirable endproducts is the production of the b-carotene precursor to vitamin A. By intro-ducing the three carotenogenic genes required for lycopene synthesis from farenyldiphosphate under the control of Candida utilis, a C. utilis strain producing 1.1mg of lycopene/g (dry weight) of cells has been generated. By using concepts ofmetabolic engineering with C. utilis, carbon flux in this yeast strain was redirectedaway from ergosterol formation for potential use by the carotenoid pathway. Theinfluential steps in the pathway that were manipulated were 3-hydroxy methyl-glutaryl (HMG)-coenzyme A (CoA) reductase, encoded by the HMG gene, andsqualene synthase, encoded by the ERG9 gene. A combination of ERG9 genedisruption and the overexpression of the HMG catalytic domain gave the highestlycopene yield. These findings illustrate how modifications in related biochemicalpathways can be used to enhance the production of commercially desirable com-pounds such as carotenoids (81).
Scientists are now beginning to elucidate the pathway of vitamin C biosyn-thesis in higher plants. Several enzymes and precursors at work in this pathwayhave recently been identified (103), and MFA may in the near future allow fortheir rational modification. To this end, Sauer et al (76) investigated fluxes in ariboflavin-producing Bacillus subtilis strain and found that generation of thedesired riboflavin was limited by the biosynthetic pathway fluxes and not by thefluxes associated with central carbon metabolism.
Another group of metabolic engineering applications is the engineering oforganisms that can use abundant byproducts of various industries as nutrients byextending the range of their substrates. Whey, with high lactose and protein con-tent, is a nutrient-rich byproduct of the dairy industry that can provide an inex-pensive carbon and nitrogen source in biotechnological processes. Although avariety of microbes can use whey, some of the most industrially prominent organ-isms are unable to do so. The E. coli lacZY operon (coding for b-galactosidaseand lactose permease) was inserted into Xanthomonas campestris, a bacteriumthat is used for xanthan gum production (24). The recombinant stain expressedhigh levels of b-galactosidase and grew well in a medium containing lactose asthe sole carbon source. In another approach, an S. cerevisiae recombinant strainwas constructed that expressed the gene for a secreted b-galactosidase (lacA) fromAspergillus niger (53). This approach offers significant advantages over earlierprocesses for the fermentation of whey by S. cerevisiae, which used either b-galactosidase–prehydrolyzed whey or yeast coimmobilized with b-galactosidase.Sucrose is another abundant and inexpensive carbon source found in, for example,cane molasses. A successful attempt to create a recombinant sucrose-metabolizingstrain involved the cloning of the scrA gene, which codes for sucrase, from E.coli B-62 onto a plasmid and then transferring the cloned DNA fragment onto
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
542 KOFFAS ET AL
the chromosome of E. coli K-12. Tryptophan producer derivatives of E. coliK-12 expressing the scrA gene grew well in sucrose medium and excretedamounts of tryptophan comparable to these from similar strains grown on glucose(95). Starch, derived from renewable sources such as corn and cereals, is a veryimportant carbon and energy source in biotechnological processes. It is a mixtureof linear and branched homopolymers of D-glucose. Because most microorgan-isms are unable to degrade this glucose biopolymer, work has focused on cloninggenes for enzymatic starch hydrolysis into various organisms (49). Along theselines, an S. cerevisiae strain was constructed that contained a glucoamylase genefrom Aspergillus sp. (35). The recombinant strain was able to grow on amylodex-trins, albeit at a lower rate than occurs when glucoamylases are added to thefermentation medium.
Cofactor engineering is a novel approach to metabolic engineering. The clon-ing of the Streptococcus mutant nox-2 gene, coding for the H2O-forming (non-toxic) NADH oxidase, under the control of the nisA promoter in Lactococcuslactis, provides a powerful tool with which to attempt the modulation of themetabolism. The main effect of overproducing the NADH oxidase was anobserved decrease in the NADH/NAD ratio under aerobic conditions. This engi-neered system could be used to provoke a shift from homolactic to mixed-acidfermentation during aerobic glucose catabolism. The magnitude of this shift wasdirectly dependent on the level of NADH oxidase overproduced. These resultsindicate that the observed shift from homolactic to mixed-acid fermentation underaerobic conditions is mainly modulated by the level of NADH oxidation resultingin low NADH/NAD ratios in the cells (57).
SECONDARY METABOLITES
The challenges associated with engineering the biosynthesis of secondary metab-olites are qualitatively different from those associated with the production ofcommodity bulk chemicals. Products of interest are often complex molecules thatare necessarily derived from biological sources. In this arena, pathway engineer-ing may increase the efficiency of existing production methods but may also leadto the development of new products. The relatively high value of these productsshifts the emphasis from economics and efficiency to innovation.
An interesting example is the production of antibiotics. Antibiotics are madeby secondary metabolic pathways that use common metabolites in less specificand, sometimes, more intricate ways than metabolites are used in primary metab-olism. Recently, it has become apparent that yields of secondary metabolites,including antibiotics, can also be improved by overcoming rate-controlling bio-synthetic steps through genetic techniques. In addition, metabolic engineeringtechniques are applied to modify known antibiotics to improve their propertiesand also to synthesize new product forms.
Among various antibiotic producers of industrial importance, Streptomycesspecies rank near the top. Actinorhodin biosynthesis genes were transferred from
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 543
S. coelicitor, the only species with well-established genetics, to S. lividans,enabling the latter strain to produce actinorhodin. Later, clustered erythromycingenes from S. erythreus were transferred to S. lividans, allowing the recombinantstrain to produce erythromycin A. Transformation of the fungi Neurospora crassaand A niger with a cosmid containing Penicillium chrysogenum penicillin bio-synthetic genes resulted in the production of penicillin V by these strains (4, 60).
The manufacture of antitumor drugs, many of which are natural products, hasalso received significant attention. The undesirable side effects of antitumor drugs,as well as the development of resistance to them, have fueled the need for thediscovery of novel therapeutic agents and the means for their synthesis. Themanipulation at a genetic level of the enzymes composing drug production path-ways within cells has emerged as a very powerful tool for achieving these goals.This method, which has been termed combinatorial biosynthesis, has yieldedderivatives and analogs of drugs such as mithramycin (21, 28), tylosin (2), eryth-romycin (19, 36, 37, 63, 99), and methymycin (108). Here, as an example, wefocus on techniques used to engineer a biological reaction pathway for the pro-duction of the cancer chemotherapy drugs epirubicin (48-epidoxorubicin) and 48-epidaunorubicin. The medical and industrial significance of epirubicin lies in itsantitumor activity, its use as a precursor of the drug candidate 4-iodoxorubicin(87), and its reduced cardiotoxicity relative to that of doxorubicin (102), a heavilyprescribed chemotherapy medication. The currently used synthetic means of pro-ducing epirubicin suffers from many complexities. The scheme involves sevendifferent synthesis steps, which are followed by still more separation and depro-tection steps. Madduri et al have described an alternative process that generates48-epidaunorubicin and epirubicin directly and relies on the fermentation of aStreptomyces peucetius strain to which has been introduced an avermectin orerythromycin biosynthetic gene (59).
This discovery was made with the use of a S. peucetius strain in which thednmV gene has been disrupted (69). This strain accumulates e-rhodomycinone, aprecursor of 48-epidaunorubicin and epirubicin, unless transformed with a plasmidcontaining the wild-type dnmV gene. In this case, daunorubicin and doxorubicin,analogs of 48-epidaunorubicin and epirubicin, respectively, are produced (97).The function of the DnmV enzyme was hypothesized to be that of a TDP-4-ketohexulose-reductase (26), much like that of the Saccharopolyspora erythraeaeryBIV (89) and Streptomyces avermitilis avrE (66) gene products. Cloning ofthese latter two genes onto plasmids used to transform the dnmV mutant yieldedbacteria that produced 48-epidaunorubicin and epirubicin. The integration of theavrE gene directly into the S. peucetius chromosome gave results similar to thoseseen with the gene acting in trans from a plasmid.
Madduri et al also found that the yield of 48-epidaunorubicin was increasedby introducing a dnrH mutation into the avrE integrant strain (59). It is believedthat the dnrH genes code for enzymes that catalyze side reactions that form gly-cosides of daunorubicin and its precursors (78). Additional productivity gainswere achieved by providing an overexpressed plasmid copy of the dnmT geneinto the avrE integrant. The enzyme DnmT is hypothesized to catalyze a limiting
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
544 KOFFAS ET AL
step in the synthesis of the daunorubin precursor, daunosamine. By incorporatingall three findings—the integration of avrE, the mutation of dnrH, and the over-expression of dnmT—into one strain, 48-epidaunorubicin titers were realized thatapproximated those of daunorubicin seen in the wild-type strain. Although theseproduct concentrations are still too low to effectively compete with the extantsynthetic process results, the gains made by metabolic engineering are significantand point the way for future improvements that can be implemented once theentire biosynthetic network has been better characterized. Extensive reviews ofcombinatorial biosynthesis advances can be found elsewhere (32, 74).
Biological processes using reactions catalyzed by enantiospecific enzymes areincreasingly being investigated as methods for the manufacture of pharmaceuticalcompounds and their intermediates. Because these compounds are often active,as a treatment, in only one particular chiral form, processes that exclusively gen-erate the desired chiral form are very advantageous. Although biologically cata-lyzed reactions, termed biotransformations, do possess this attribute, theirproductivity is often economically unfavorable relative to yields seen in organicsynthesis processes. Because of this, metabolic engineering is being called on toprovide the means of increasing the efficiencies and product concentrations ofthe pathways of interest, so that they can be more attractive for use in industrialpractice.
One potential application of biotransformations is the manufacture of Crixivan(indinavir sulfate), a protease inhibitor developed and produced by Merck & Co.and targeted for the treatment of human immunodeficiency virus (23). The chem-ical structure of Crixivan contains five chiral centers, two of which are contributedto the final compound from the intermediate 1-amino-2-(R)-indanol, a derivativeof indene. This drug precursor can be produced by transformation of indene bythe Rhodococcus sp. strain I24 (8). This soil isolate has been shown to grow withnaphthalene and toluene as its only carbon sources, and it is believed to possessat least three different oxygenases. The presence of the multiple-oxygenaseenzymes results in formation of multiple stereochemical enantiomers of thedesired 1-amino-2-indanol when indene is supplied to I24 cultures. To increaseboth the specificity with which I24 produces the 2-(R) form and the final producttiters, the pathway of indene bioconversion has been the subject of recent study.In particular, the nature of the specific oxygenases in action is being investigated,and factors controlling flux distribution to the different competing pathways ana-lyzed. These studies are aided by novel methods of flux determination that makeuse of radio-labeled tracer compounds.
The polyketide family is another rich source of bioactive molecules with anti-biotic and pharmacological properties. Reasons that polyketides are an attractivestudy model for metabolic engineering include the following: (a) their complexstructure results from simple units combined in diverse ways; (b) the modularconstruction of the enzymatic catalyst allows control of enzyme structure and,hence, polyketide type at the genetic level. Recent progress in this area has estab-lished the groundwork to generate novel polyketide structures through genetic
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 545
engineering of polyketide synthases and at the same time to derive knowledgethat elucidates the structure-function relationship in polyketide synthases (48, 64).Moreover the field provides an opportunity to bridge the fields of genetics andchemistry and, above all, promises to enable scientists to rationally design novelmolecules at the level of DNA.
CELL AND TISSUE ENGINEERING
Metabolic engineering can also be used to construct cells with desirable propertiesby altering characteristics such as growth, proliferation, tolerance to exogenousfactors, substrate utilization, etc. These characteristics are the result of complexbiological functions involving multiple gene products. This complexity mayafford greater efficiency or higher quality control or it may exist simply becausea single protein cannot provide the required function. Metabolic engineering strat-egies that coordinately modify multigene expression therefore have the potentialto achieve previously inaccessible metabolic states.
The hemoglobin (VHb) of the microorganism Vitreoscilla species has beenextensively investigated in recent years as a tool for maintaining metabolic activ-ity under hypoxic conditions. In its native host, VHb binds oxygen at a lowextracellular concentration and acts as a buffering agent at high intracellular-oxygen concentrations, allowing the bacteria to survive in the hypoxic environ-ments to which it is indigenous (100). The gene (vgb) encoding the hemoglobinwas cloned (18, 51) and used to transform a wide range of species in attempts toimprove respiration, growth, and productivity. A few of the most recent advanceshave involved increasing the total protein secretion, neutral protease, and a-amy-lase activities of B. subtilis (46), boosting antibiotic production in S. coelicolor(16) and S. lividans (60), enhancing lysine production in C. glutamicum (75), andimproving the degradation of benzoic acid by Xanthomonas maltophilia (56). Theeffectiveness of VHb in improving cellular processes was compared with that oftwo other globins, horse heart myoglobin (HMb) and yeast flavohemoglobin(YFb), and found to be the superior growth enhancer in E. coli (47). The inves-tigators theorized that this was so because, of the three globins, VHb was uniquein possessing a slow oxygen on-rate constant and a fast oxygen off-rate constant.This results in VHb being particularly effective in both scavenging sparse oxygenmolecules and donating them to the respiratory components that require oxygen.Another recent work has identified a strong effect of growth medium on theefficacy of improving processes with VHb: Wei et al (101) attempted to alter theproduction of acetoin and 2,3-butanediol in Serratia marcescens by transformingthe species with plasmids containing vgb. When the cells were grown on Luriabroth supplemented with 2% glucose, the non–vgb-bearing strains produced 15-fold as much acetoin and fourfold as much 2,3-butanediol as those cells with vgb.For growth on Luria broth supplemented with 2% casein acid hydrolysate, though,a vgb-bearing strain produced significantly more of the two products than did
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
546 KOFFAS ET AL
strains without the hemoglobin gene. This result clearly demonstrates the com-plications that can arise in using bacterial hemoglobins as metabolic engineeringtools.
In the area of cell cycle improvements, some of the most complex and impor-tant regulatory mechanisms of eukaryotic cells are those that govern cell division.Effective reprogramming of the complex regulatory apparatus to achieve biopro-cess goals, such as cessation of proliferation at high cell density to allow anextended period of high production, can require coordinated manipulation of mul-tiple genes. The overexpression of the cyclin-dependent kinase inhibitors p21 andp27 has already proven to be effective in cancer therapy. In a stable geneticconfiguration, only regulated overexpression of p27 was successful in inducinga sustained Chinese hamster ovary cell growth arrest in G1 phase, which alsoresulted in a 10-fold increase in per-cell secreted alkaline phospatase, a modelheterologous protein used during these studies. Stable overexpression of p21alone did not result in growth arrest. Recently, by tetracycline-regulated coex-pression of p21 and the differentiation factor CCAAT/enhancer-binding proteina (which both stabilizes and induces p21), Fussenegger et al achieved effectivecell cycle arrest. Production of secreted alkaline phospatase has been increased10- to 15-fold, on a per cell basis, relative to an isogenic control cell line. Becausethe activation of apoptosis response is a possible complication in a proliferation-arrested culture, the survival gene bcl-xL was coexpressed with another CDI, p27,found to enable CHO cell cycle arrest predominantly in G1 phase. CHO cellsstably transfected with a tricistronic construct containing the genes for these pro-teins and for secreted alkaline phospatase showed 30-fold higher secreted alkalinephospatase expression than control cells (25).
BIOMEDICAL APPLICATIONS
Besides manufacturing applications, metabolic engineering is having a significantimpact on the medical field. The main focus here is the design of new therapiesby identifying specific targets for drug development and by contributing to thedesign of gene therapies. Such approaches currently target a specific single enzy-matic step implicated in a particular disease. There is no assurance, however, thatthe manipulation of a single reaction will translate to systematic responses in thehuman body. Although mammalian intermediary metabolism was defined in bio-chemical terms many years ago, it is important to remember that much of thisinformation accrued from studies in vitro. We have less understanding of theorganization in vivo. Furthermore, the intersections of the central pathways ofmetabolism, such as glycolysis, gluconeogenesis, urea synthesis, tricarboxylicacid cycle, etc, cause changes in one pathway, owing to inborn error or disease,to affect pathways that may seem remote from the initial metabolic defect (73).In this regard, medical applications are no different from the ones mentionedearlier in an industrial context, and, as such, they will benefit from developments
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 547
in metabolic engineering through a better analysis of experimental results andapplications to the rational selection of targets for medical treatment. In this sec-tion, some representative examples are highlighted that illustrate the applicationof metabolic engineering tools to the study of human disease.
Inborn Errors of Metabolism
One of the earliest applications of MFA to inherited disorders of metabolismconcerned aberrations associated with glycogen storage diseases. In a series of13C-tracer experiments, Kalderon et al (42–45) examined the pathways of hepaticglucose storage and use in glycogen storage disease (GSD) types I (GSD-I) andIII (GSD-III). In children with GSD-I, the isotopomer distributions of infused [U-13C]glucose and plasma glucose were identical, indicating absence of glucoserecycling, whereas a significant change in the isotopomer distribution of plasmaglucose was observed in normal and GSD-III subjects. The absence of recycledglucose in their plasma eliminated a mechanism for glucose production in GSD-I patients involving gluconeogenesis, suggesting a deficiency of a gluconeogenicenzyme such as glucose 6-phosphatase. In contrast, gluconeogenesis was sug-gested as the major route for endogenous glucose synthesis in GSD-III patients.Moreover, these differences in glucose recycling correlated with significant dif-ferences in the glucose C-1 13C nuclear magnetic resonance splitting pattern inplasma, suggesting that such 13C nuclear magnetic resonance analysis of plasmamay be used to noninvasively diagnose defects in gluconeogenesis.
Another type of disorder to which MFA has been applied is hereditary fructoseintolerance (HFI), an inborn deficiency in the ability of aldolase B to split fruc-tose-1-phosphate. Continuous exposure of these subjects to parental fructose dur-ing infancy may result in liver cirrhosis, mental retardation, and death. In mostcases of HFI, final diagnosis of aldolase B deficiency is usually performed inliver biopsy specimens. As an alternative, Gopher et al (27) proposed a methodfor noninvasive in vivo diagnosis of HFI based on mass isotopomer analysis. Incontrol and HFI children, steps involved in fructose metabolism were quantifiedby analyzing plasma glucose isotopomer populations after nasogastric infusionof D-[U-13C]fructose. After administration of labeled fructose, the conversion offructose to glucose was significantly lower in HFI than in control children asdetermined by this method, supporting its validity as a diagnostic test. Further-more, it was found that the generally accepted pathway of fructose conversion toglucose, by fructose-1-phosphate aldolase to triose phosphate, accounts for onlyone-half of the total amount of fructose conversion in normal subjects. It is sug-gested that a direct pathway from fructose to fructose-1,6-bisphosphate by 1-phosphofructokinase exists, accounting for the remainder of fructose conversion.
Diabetes has also been the subject of many studies using stable isotopomermethods. A series of extensive studies by Cohen (10–12) using streptozotocin-diabetic rats has shown that the increase in relative flux through pyruvate car-boxylase and the inhibition of flux through pyruvate kinase that prevents
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
548 KOFFAS ET AL
reconversion of phosphoenolpyruvate into pyruvate may be concerted actionsleading to the enhanced gluconeogenesis found in this model of diabetes. Tayek& Katz (90) compared the relative contributions of gluconeogenesis and glyco-genolysis with postabsorptive glucose production in normal and non–insulin-dependent diabetes mellitus (NIDDM) humans. They found that total glucoseproduction was elevated in NIDDM patients compared with control subjects andthat fractional contribution of gluconeogenesis was comparable, raising doubtsabout the widely held notion that synthesis of hepatic glycogen is seriouslyimpaired in NIDDM. On the other hand, Landau et al (54) estimated the contri-bution of gluconeogenesis to glucose production in insulin-dependent diabetesmellitus (IDDM) patients to be significantly less than in normal subjects, alsobased on analyses of isotopomer distribution. More recently, Peroni et al (70)measured gluconeogenic fluxes in postabsorptive and starved normal and strep-tozotocin-diabetic rats, in which it was found that the increased gluconeogeniccontribution to glucose production in diabetic rats relative to control rats in thepostabsorptive state was abolished in the starved state.
Once inherited disorders of metabolism are diagnosed and the affected bio-chemical pathways have been identified, these diseases may be treated by usinggenetic-engineering tools. The therapy for many inherited liver enzyme deficien-cies requires the removal of toxic intermediate metabolites from the blood ofaffected individuals. Recent research tries to focus on the removal of circulatingtoxins by expressing the missing enzymes in tissues other than the liver. This willhopefully positively influence the disease phenotype. Harding et al (29) success-fully expressed the phenylalanine hydroxylase (PAH) activity in skeletal and car-diac muscle of mice under the control of the mouse muscle creatine kinasepromoter. When they bred the muscle PAH-expressing mice with liver PAH-deficient mice, a progeny was created that lacked PAH activity in liver butexpressed PAH in muscle. These mice exhibited hyperphenylalaninemia at base-line, but serum phenylalanine levels decreased significantly when the mice weresupplemented with tetrahydrobiopterin, a required cofactor for PAH. This resultsuggests that gene therapy targeted to heterologous tissue, such as muscle, willbe an effective treatment of selected inborn errors of metabolism.
Cell and Organ Physiology
A powerful feature of metabolic engineering is that it allows systematic investi-gation of metabolic control and regulation in intact tissues. Given that manymetabolic disorders, such as liver cirrhosis, post-traumatic hypermetabolism andmuscle wasting, cancer cachexia, etc, have no clearly identifiable genetic origins,it is clear that to develop therapeutics or treatment strategies, quantitative char-acterization has to be achieved at a biochemical level. In this regard, the resultsof cell or organ physiological studies within the framework of MFA could bevery valuable. For example, an important focus of tumor biology has been onunderstanding the ability of tumors to adapt to adverse growth environments such
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 549
as hypoxic or hypoglycemic conditions. It has been suggested that the energymetabolism of tumors is altered in such a way as to accommodate high levels ofanaerobic metabolism and alternate substrate utilization. Using AS-30D hepatomacells, Holleran et al (31) investigated the quantitative importance of acetoacetateand glucose as substrates of energy metabolism in tumors. It was found thatacetoacetate diverted pyruvate from pyruvate dehydrogenase (PDH) to pyruvatecarboxylation. In contrast, dichloroacetic acid, a metabolic analog of acetoacetatethat is an activator of PDH, increased the oxidation of glucose largely throughPDH, indicating that PDH is not maximally active in the presence of dichloroac-etic acid. Isotopomer spectral analysis of lipid synthesis demonstrated that, in theabsence of acetoacetate, glucose supplied 65% of the acetyl-CoA used for denovo lipogenesis. Isotopomer spectral analysis refers to a variant of 13C-isoto-pomer analysis particularly suited to the study of lipid metabolism (50). In thepresence of high levels of acetoacetate, glucose was replaced as a lipogenic pre-cursor, and acetoacetate supplied 85% of the acetyl-CoA for lipogenesis vs only2% for glucose. Thus AS-30D cells may have a large capacity for acetoacetateuse for de novo lipogenesis, leading to greater capacity for fat storage, whichmay help survival in cachexic conditions.
Portais et al (71) used a mathematical model of the TCA cycle in combinationwith 1H-13C-nuclear magnetic resonance to calculate the flux distribution in ratbrain tumor cells. In their study, it was found that the pyruvate carboxylase activ-ity and the efflux from the TCA cycle in C6 glioma cells are estimated to be verylow, suggesting a lack of glutamine production in these cells. In a subsequentpublication (72), they reported that glutamine and glucose are metabolized com-plementarily in C6 cells in that glutamine is mainly used for anaplerosis but nota substrate for energy metabolism, whereas glucose is poorly anaplerotic and isessentially used as energetic fuel. Using a similar method, Bouzier et al (5) foundthat, unlike normal astrocytes, C6 cells preferentially use lactate as a substratefor oxidative metabolism. Such characterizations of the differences in interme-diary metabolism between tumors and their normal counterparts could lead tobetter understanding of tumor proliferation and may be exploited to control tumorgrowth in vivo.
MFA has also been used to characterize the effects of acute metabolic stresses,such as hypoxia or reperfusion injury. Malloy et al (62) developed a model basedon isotopomer distribution of glutamate in heart tissue extracts to show that, inperfused hearts exposed to a combination of substrates (lactate, acetate, glucose),ischemia-reperfusion leads to an increase in the contribution of acetate and adecrease in the contribution of lactate as sources of acetyl CoA. Ischemia-reper-fusion injury also causes an increase in anaplerotic flux. However, exogenouslyadded aspartate or glutamate are not significantly metabolized. This is a findingthat establishes a protective role for aspartate and glutamate on myocardial ische-mia that does not result from a direct mechanism involving the TCA cycle in theheart tissue. Also in a perfused heart model, Laplante et al (55) established thatthe cardioprotective effect of fumarate during ischemia or hypoxia occurs through
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
550 KOFFAS ET AL
its reduction to succinate. Another technique developed by Sherry et al (79, 88),showed that lipoamide—an agent being considered to enhance recovery afterinfarctus—prevented, in large part, the switch from lactate to acetate use inducedby ischemia.
Finally, over the past several years, protocols have been developed that allowthe extension of 13C-labeling–based flux analysis to noninvasive monitoring ofspecific organs without biopsies. One such approach involves the use of xeno-biotics such as phenylacetate, which in primates is excreted in urine as a glutamineconjugate. Conjugation occurs specifically in the liver. Glutamine is synthesizedfrom a-ketoglutarate via glutamate without rearrangement of carbons. Thus, glu-tamine carbons reflect the carbons of a-ketoglutarate. Consequently, by analyzingthe labeling pattern in the glutamine conjugate in urine, liver metabolic fluxesmay be estimated. This method was first validated by Magnusson et al (61), whoestimated the fluxes in and around the TCA cycle in human subjects by analyzingthe 14C distribution in excreted phenylacetate after infusion with [3-14C]lactateand oral administeration of phenylacetate. Jones et al (40) improved the methodby identifying a stable isotope tracer, [U-13C]propionate, which is quantitativelyextracted into the liver from portal circulation. Coined as ‘‘chemical biopsy’’ byDiDonato et al (17), these noninvasive organ-monitoring protocols used in con-junction with 13C-labeling experiments could become an important research anddiagnostic tool, providing a more detailed assessment of the metabolic state ofliver in human metabolic disorders.
CONCLUDING REMARKS
The success of biotechnology and biomedical engineering applications of the typeaddressed in this review—the production of primary and secondary metabolites,the alteration and improvement of cellular properties, and the investigations ofthe causes and potential treatments of diseases and injuries—are all subject to theunderstanding of complex networks of metabolic pathways. Because its focus isnot on isolated reactions but on the interrelationships of reactions in networks,the emerging field of metabolic engineering should be seen as particularly relevantto these applications. The examples described above clearly demonstrate thebreadth of the range of advances realized through metabolic engineering.
Although metabolism and cell physiology provide the main context for ana-lyzing reaction pathways, it should be pointed out that results of flux determi-nation and control have still broader applicability. Thus, besides the analysis ofmaterial and energy fluxes through metabolic pathways, the concepts of metabolicengineering are equally useful in the analysis of information fluxes and of thoseencountered in signal transduction pathways. Because the latter have not yet beenwell defined, the main focus of this article has been on applications to metabolicpathways. However, once the concepts of information pathways have crystallized,we expect that many of the ideas and tools presented herein will find good use
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 551
in the study of the interactions of signal transduction pathways and the elucidationof the complex mechanisms by which external stimuli control gene expression.
Metabolic engineering is the science aiming at a holistic understanding ofmetabolic functions and cellular physiology. As such it provides a much neededframework for analyzing measurements of differential gene expression obtainedthrough the application of emerging technologies such as DNA microarrayhybridization. These data, in combination with detailed flux measurementsobtained by available methods and others under development, offer the best prom-ise for a systematic study and elucidation of metabolic networks. As accuratemeasurements of gene expression, protein content, and in vivo fluxes becomereadily available, intricate regulatory structures at the genetic and metabolic levelswill be gradually better understood using the principles of metabolic engineering.This will have profound implications for the rational modification of metabolicand signaling pathways both in a biotechnological context and in refining currentmethods for target selection in drug development and gene therapy.
Visit the Annual Reviews home page at http://www.AnnualReviews.org.
LITERATURE CITED
1. Aiba S, Imanaka T, Tsunekawa H. 1980.Enhancement of tryptophan productionin Escherichia coli as an application ofgenetic engineering. Biotechnol. Lett.2:525–30
2. Arisawa A, Kawamura N, Takeda K,Tsunekawa H, Okamura K, Okamoto R.1994. Cloning of the macrolide antibioticbiosynthesis gene acya, which encodes3-O-acyltransferase, from Streptomycesthermotolerans, and its use for directfermentative production of a hybrid mac-rolide antibiotic. Appl. Environ. Micro-biol. 60:2657–60
3. Bailey JE. 1991. Towards a science ofmetabolic engineering. Science 252:1668–74
4. Bartel PL, Zhu CB, Lampel JS, DoschDC, Connors N, et al. 1990. Biosynthesisof anthraquinones by interspecies clon-ing of actinorhodin biosynthesis genes inStreptomyces: clarification of actinor-hodin gene functions. J. Bacteriol.172:4816–26
5. Bouzier AK, Voisin P, Goodwin R, Can-ioni P, Merle M. 1998. Glucose and lac-
tate metabolism in C6 glioma cells:evidence for the preferential utilizationof lactate for cell oxidative metabolism.Dev. Neurosci. 20:331–38
6. Brabetz W, Liebl W, Schleifer KH. 1991.Studies on the utilization of lactose byCorynebacterium glutamicum, bearingthe lactose operon of Escherichia coli.Arch. Microbiol. 155:607–12
7. Brau B, Sahm H. 1986. Cloning andexpression of the structural gene forpyruvate decarboxylase of Zymomonasmobilis in Escherichia coli. Arch. Micro-biol. 144:296–301
8. Buckland BC, Drew SW, Connors NC,Chartrain MM, Lee C, et al. 1999. Micro-bial conversion of indene to indandiol: akey intermediate in the synthesis of Crix-ivan. Metab. Eng. 1:63–74
9. Cameron DC, Altaras NE, Hoffman ML,Shaw AJ. 1998. Metabolic engineeringof propanediol pathways. Biotechnol.Prog. 14:116–25
10. Cohen SM. 1987. 13C and 31P NMRstudy of gluconeogenesis: utilization of13C-labeled substrates by perfused liver
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
552 KOFFAS ET AL
from streptozotocin-diabetic and un-treated rats. Biochemistry 26:563–72
11. Cohen SM. 1987. 13C NMR study ofeffects of fasting and diabetes on themetabolism of pyruvate in the tricarbox-ylic acid cycle and the utilization of pyru-vate and ethanol in lipogenesis inperfused rat liver. Biochemistry 26:581–89
12. Cohen SM. 1987. Effects of insulin onperfused liver from streptozotocin-dia-betic and untreated rats: 13C NMR assayof pyruvate kinase flux. Biochemistry26:573–80
13. Colon GE, Follettie MT, Jetten MSM,Stephanopoulos G, Sinskey AJ. 1993.Redirections of carbon flux at a Cory-nebacterium glutamicum threonine met-abolic branch point by controlled enzymeoverexpression. In Annu. Meet. Am. Soc.Microbiol., 93, Atlanta, GA, p. 320.Washington, DC: Am. Soc. Microbiol.
14. Colon GE, Jetten MSM, Nguyen TT,Gubler ME, Sinskey AJ, StephanopoulosG. 1995. Effect of inducible thrB expres-sion on amino acid production in Cory-nebacterium lactofermentum 21799.Appl. Environ. Microbiol. 61:74–78
15. Colon GE, Nguyen TT, Jetten MSM,Sinskey AJ, Stephanopoulos G. 1995.Production of isoleucine by overexpres-sion of ilvA in a Corynebacterium lac-tofermentum threonine producer. Appl.Microbiol. Biotechnol. 43:482–88
16. DeModena JA, Gutierrez S, Velasco J,Fernandez FJ, Fachini RA, et al. 1993.The production of cephalosporin C byAcremonium chrysogenum is improvedby the intracellular expression of abacterial hemoglobin. Bio-Technology11:926–29
17. DiDonato L, Des Rosiers C, Montgom-ery JA, David F, Garneau M, Brunengra-ber H. 1993. Rates of gluconeogenesisand citric acid cycle in perfused livers,assessed from the mass spectrometricassay of the 13C labeling pattern of glu-tamate. J. Biol. Chem. 268:4170–80
18. Dikshit KL, Webster DA. 1988. Cloning,characterization, and expression of thebacterial globin gene from Vitreoscilla inEschericia coli. Gene 70:377–86
19. Donadio S, McAlpine JB, Sheldon PJ,Jackson M, Katz L. 1993. An erythro-mycin analog produced by reprogram-ming of polyketide synthesis. Proc. Natl.Acad. Sci. USA 90:7119–23
20. Eikmanns B, Metz M, Reinscheid D,Kricher M, Sahm H. 1991. Amplificationof three threonine biosynthetic genes inCorynebacterium glutamicum and itsinfluence on carbon flux in differentstrains. Appl. Microbiol. Biotechnol.43:617–22
21. Epp J, Huber MLB, Turner JR, GoodsonT, Schoner BE. 1989. Production of ahybrid macrolide antibiotic in Strepto-myces ambofaciens and Streptomyceslidivans by introduction of a cloned car-bomycin biosynthetic gene from Strep-tomyces thermotolerans. Gene 85:293–301
22. Follettie MT, Shin HK, Sinskey AJ.1988. Organization and regulation of theCorynebacterium glutamicum hom-thrBand thrB loci. Mol. Microbiol. 2:53–62
23. Fox JL. 1996. Two more HIV proteaseinhibitors approved. Nat. Biotechnol.14:427
24. Fu JF, Tseng YH. 1990. Construction oflactose-utilizing Xanthomonas campes-tris and production of xanthan gum fromwhey. Appl. Environ. Microbiol. 56:919–23
25. Fussenegger M, Schlatter S, Datwyler D,Mazur X, Bailey J. 1998. Controlled pro-liferation by multigene metabolic engi-neering enhances the productivity ofChinese hamster ovary cells. Nat. Bio-technol. 16:468–72
26. Gallo MA, Ward J, Hutchinson CR.1996. The dnrM gene in Streptomycespeucetius contains a naturally occurringframeshift mutation that is suppressed byanother locus outside of the daunorubi-
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 553
cin-production gene cluster. Microbiol-ogy 142:269–75
27. Gopher A, Vaisman N, Mandel H, Lap-idot A. 1990. Determination of fructosemetabolic pathways in normal and fruc-tose-intolerant children: a 13C NMRstudy using [U-13C]fructose. Proc. Natl.Acad. Sci. USA 87:5449–53
28. Hara O, Hutchinson CR. 1992. A mac-rolide 3-O-acyltransferase gene from themidecamycin-producing species Strep-tomyces mycarofaciens. J. Bacteriol.174:5141–44
29. Harding CO, Wild K, Chang D, MessingA, Wolff JA. 1998. Metabolic engineer-ing as therapy for inborn errors of me-tabolism—development of mice withphenylalanine hydroxylase expression inmuscle. Gene Ther. 5:677–83
30. Heinrich R, Rapoport TA. 1974. A linearsteady-state treatment of enzymaticchains. Eur. J. Biochem. 42:105–40
31. Holleran AL, Fiskum G, Kelleher JK.1997. Quantitative analysis of acetoace-tate metabolism in AS-30D hepatomacells with 13C and 14C isotopic tech-niques. Am. J. Physiol. 272:E945–51
32. Hutchinson CR. 1998. Combinatorialbiosynthesis for new drug discovery.Curr. Opin. Microbiol. 1:319–29
33. Ikeda M, Katsumata R. 1992. Metabolicengineering to produce tyrosine or phe-nylalanine in a tryptophan-producingCorynebacterium glutamicum strain.Appl. Environ. Microbiol. 58:781–85
34. Ingram LO, Conway T, Clark DP, SewellGW, Preston JF. 1987. Genetic engi-neering of ethanol production in Esche-richia coli. Appl. Environ. Microbiol.53:2420–25
35. Innis MA, Holland MJ, McCabe PC,Cole GE, Wittman VP, et al. 1985.Expression, glycosylation and secretionof an Aspergillus glucoamylase by Sac-charomyces cerevisiae. Science 228:21–26
36. Jacobsen JR, Hutchinson CR, Cane DE,Koshia C. 1997. Precursor-directed bio-
synthesis of erythromycin analogs by anengineered polyketide synthase. Science277:367–69
37. Jacobsen JR, Keatinge-Clay AT, CaneDE, Khosla C. 1998. Precursor-directedbiosynthesis of 12-ethyl erythromycin.Bioorg. Med. Chem. 6:1171–77
38. Jacobsen JR, Khosla C. 1998. New direc-tions in metabolic engineering. Curr.Opin. Chem. Biol. 2:133–37
39. Jetten MSM, Gubler ME, McCormickMM, Colon GE, Follettie MT, SinskeyAJ. 1993. Molecular organization andregulation of the biosynthetic pathwayfor aspartate derived amino acids in Cor-ynebacterium glutamicum. In IndustrialMicroorganisms: Basic and AppliedMolecular Genetics, ed. RH Batltz, GHegeman, PL Skatrud, p. 97. Washing-ton, DC: Am. Soc. Microbiol.
40. Jones JG, Solomon MA, Sherry AD,Jeffrey FM, Malloy CR. 1998. 13CNMR measurements of human gluco-neogenic fluxes after ingestion of[U-13C]propionate, phenylacetate, andacetaminophen. Am. J. Physiol. 275:E843–52
41. Kacser H, Burns JA. 1973. The controlof flux. Symp. Soc. Exp. Biol. 27:65–104
42. Kalderon B, Gopher A, Lapidot A. 1987.A quantitative analysis of the metabolicpathways of hepatic glucose synthesis invivo with 13C-labeled substrates. FEBSLett. 213:209–14
43. Kalderon B, Korman SH, Gutman A,Lapidot A. 1989. Estimation of glucosecarbon recycling in children with glyco-gen storage disease: a 13C NMR studyusing [U-13C]glucose. Proc. Natl. Acad.Sci. USA 86:4690–94
44. Kalderon B, Korman SH, Gutman A,Lapidot A. 1989. Glucose recycling andproduction in glycogenosis type I and III:stable isotope technique study. Am. J.Physiol. 257:E346–53
45. Kalderon B, Lapidot A, Korman SH,Gutman A. 1988. Glucose recyclingand production in children with gly-
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
554 KOFFAS ET AL
cogen storage disease type I, studiedby gas chromatography/mass spec-trometry and (U-13C)glucose. Biomed.Environ. Mass. Spectrom. 16:305–8
46. Kallio PT, Bailey JE. 1996. Intracellularexpression of Vitreoscilla hemoglobin(VHb) enhances total protein secretionand improves the production of a-amy-lase and neutral protease in Bacillus sub-tilis. Biotechnol. Prog. 12:31–9
47. Kallio PT, Tsai PS, Bailey JE. 1996.Expression of Vitreoscilla hemoglobin issuperior to horse heart myoglobin oryeast flavohemoglobin expression forenhancing Escherichia coli growth ina microaerobic bioreactor. Biotechnol.Bioeng. 12:751–57
48. Kao CM-F. 1997. Structure, function andengineering of modular polyketide syn-thases. PhD thesis. Stanford Univ., PaloAlto, Calif.
49. Kennedy JF, Cabalda VM, White CA.1988. Enzymatic starch utilization andgenetic engineering. Trends Biotechnol.6:184–89
50. Kharroubi AT, Masterson TM, Aldagh-las TA, Kennedy KA, Kelleher JK. 1992.Isotopomer spectral analysis of triglyc-eride fatty acid synthesis in 3T3-L1 cells.Am. J. Physiol. 263:E667–75
51. Khosla C, Bailey JE. 1988. The Vitreos-cilla hemoglobin gene: molecular clon-ing, nucleotide sequence and geneticexpression in Escherichia coli. Mol. Gen.Genet. 214:158–61
52. Kinoshita S, Nakayama K, Kitada S.1958. L-Lysine production using micro-bial auxotroph. J. Gen. Appl. Microbiol.4:128–29
53. Kumar V, Ramakrishnan S, Teeri TT,Knowles JKC, Hartey BS. 1992. Saccha-romyces cerevisiae cells secreting anAspergillus niger galactosidase grown onwhey permeate. Bio-Technology 10:82–85
54. Landau BR, Chandramouli V, SchumannWC, Ekberg K, Kumaran K, et al. 1995.Estimates of Krebs cycle activity and
contributions of gluconeogenesis tohepatic glucose production in fastinghealthy subjects and IDDM patients.Diabetologia 38:831–38
55. Laplante A, Vincent G, Poirier M, DesRosiers C. 1997. Effects and metabolismof fumarate in the perfused rat heart. A13C mass isotopomer study. Am. J. Phys-iol. 272:E74–82
56. Liu S-C, Webster DA, Wei M-L, StarkBC. 1996. Genetic engineering to con-tain the Vitreoscilla hemoglobin geneenhances degradation of benzoic acid byXanthomonas maltophilia. Biotechnol.Bioeng. 49:101–5
57. Lopez F, Kleerebezem M, De Vos W,Hugenholtz J. 1998. Cofactor engi-neering: a novel approach to metabolicengineering in Lactococcus lactis bycontrolled expresion of NADH oxidase.J. Bacteriol. 180:3804–8
58. MacQuitty JJ. 1988. Impact of biotech-nology on the chemical industry. ACSSymp. Ser. 362:221–33
59. Madduri K, Kennedy J, Rivola G,Inventi-Solari A, Filippini S, et al. 1998.Production of the antitumor drug epirub-icin (48-epidoxorubicin) and its precursorby a genetically engineered strain ofStreptomyces peucetius. Nat. Biotechnol.16:69–74
60. Magnolo SK, Leenutaphong DL, De-Modena JA, Curtis JE, Bailey JE, Gal-azzo JL. 1991. Actinorhodin productionby Streptomyces coelicitor and growth ofStreptomyces lividans are improved byexpression of a bacterial hemoglobin.Bio-Technology 9:473–76
61. Magnusson I, Schumann WC, BartschGE, Chandramouli V, Kumaran K, et al.1991. Noninvasive tracing of Krebscycle metabolism in liver. J. Biol. Chem.266:6975–84
62. Malloy CR, Sherry AD, Jeffrey FM.1990. Analysis of tricarboxylic acidcycle of the heart using 13C isotope iso-mers. Am. J. Physiol. 259:H987–95
63. McAlpine JB, Tuan JS, Brown DP, Greb-
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 555
ner KD, Whittern DN, et al. 1987. Newantibiotics from genetically engineeredactinomycetes. 1. 2-Norerythromycins,isolation and structural determinations. J.Antibiot. 40:1115–22
64. McDaniel R, Ebert-Khosla S, HopwoodDA, Khosla C. 1993. Engineering bio-synthesis of novel polyketides. Science262:1546–50
65. Nerem RM. 1991. Cellular engineering.Ann. Biomed. Eng. 19:529–45
66. Occi JL, Gerwain KM, Ruby CL,MacNeill DJ. 1991. Abstr. Annu. Meet.Soc. Ind. Microbiol., 42, Philadelphia,Abstr. P49, p. 93. New York: Soc.Indust. Microbiol.
67. Ohta K, Beall DS, Mejia JP, ShanmugamKT, Ingram LO. 1991. Genetic improve-ment of Escherichia coli for ethanol pro-duction: chromosomal integration ofZymomonas mobilis genes encodingpyruvate decarboxylase and alcoholdehydrogenase II. Appl. Environ. Micro-biol. 57:893–900
68. Ohta K, Beall DS, Mejia JP, ShanmugamKT, Ingram LO. 1991. Metabolic engi-neering of Klebsiella oxytoca M5A1 forethanol production from xylose andglucose. Appl. Environ. Microbiol. 57:2810–15
69. Otten SL, Gallo MA, Madduri K, Liu X,Hutchinson CR. 1997. Cloning and char-acterization of the Streptomyces peuce-tius dnmZUV genes encoding threeenzymes required for biosynthesis of thedaunorubicin precursor thymidine di-phospho-L-daunosamine. J. Bacteriol.179:7316–21
70. Peroni O, Large V, Diraison F, Beylot M.1997. Glucose production and gluconeo-genesis in postabsorptive and starvednormal and streptozotocin-diabetic rats.Metabolism 46:1358–63
71. Portais JC, Schuster R, Merle M, CanioniP. 1993. Metabolic flux determination inC6 glioma cells using carbon-13 distri-bution upon [1–13C]glucose incubation.Eur. J. Biochem. 217:457–68
72. Portais JC, Voisin P, Merle M, CanioniP. 1996. Glucose and glutamine metab-olism in C6 glioma cells studied by car-bon 13 NMR. Biochimie 78:155–64
73. Reeds PJ, Berthold HK, Boza JJ, BurrinDG, Jahoor F, et al. 1997. Integration ofamino acid and carbon intermediarymetabolism: studies with uniformlylabeled tracers and mass isotopomeranalysis. Eur. J. Pediatr. 156(Suppl. 1)S50–58
74. Salas JA, Mendez C. 1998. Geneticmanipulation of antitumor-agent biosyn-thesis to produce novel drugs. TrendsBiotechnol. 16:475–82
75. Sander FC, Fachini RA, Hughes DE,Galazzo JL, Bailey JE. Expression of theVitreoscilla hemoglobin in Corynebac-terium glutamicum increases the finalconcentration and yield of L-lysine.1993, Florence, Italy. ECB6: Proc. Eur.Congr. Biotechnol., 6th, pp. 607–10Elsevier, New York 1994
76. Sauer U, Hatzimanikatis V, Bailey JE,Hochuli M, Szyperski T, Wuthrich K.1997. Metabolic fluxes in riboflavin-producing Bacillus subtilis. Nat. Biotech-nol. 15:448–52
77. Schmidt K, Carlsen M, Nielsen J, Villad-sen J. 1997. Modeling isotopomer distri-butions in biochemical networks usingisotopomer mapping matrices. Biotech-nol. Bioeng. 55:831–40
78. Scotti C, Hutchinson CR. 1996. En-hanced antibiotic production by manip-ulation of the Streptomyces peucetiusdnrH and dnmT genes involved in doxo-rubicin (Adriamycin) biosynthesis. J.Bacteriol. 178:7316–21
79. Sherry AD, Malloy CR, Zhao P, Thomp-son JR. 1992. Alterations in substrateutilization in the reperfused myocar-dium: a direct analysis by 13C NMR.Biochemistry 31:4833–37
80. Shiio I, Yoshino H, Sugimoto S. 1990.Isolation and properties of lysine-pro-ducing mutants with feedback-resistantaspartokinase derived from Brevibacter-
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
556 KOFFAS ET AL
ium flavum strain with citrate synthase-and pyruvate kinase-defects and feed-back resistant phosphoenol pyruvate car-boxylase. Agric. Biol. Chem. 54:3275–82
81. Shimada H, Kondo K, Fraser PD, MiuraY, Saito T, Misawa N. 1998. Increasedcarotenoid production by the food yeastCandida utilis through metabolic engi-neering of the isoprenoid pathway. Appl.Environ. Microbiol. 64:2676–80
82. Shio I. 1986. Production of individualamino acids: tryptophan, phenylalanineand tyrosine. In Biotechnology of AminoAcid Production, ed. K Aida, I Cjibata,K Nakayama, K Takinami, H Yamada,pp. 188–206. Tokyo: Elsevier
83. Sonntag K, Eggeling L, deGraaf AA,Sahm H. 1993. Flux partitioning in thesplit pathway of lysine synthesis inCorynebacterium glutamicum–quantifi-cation by 13C NMR and 1H-NMR spec-troscopy. Eur. J. Biochem. 213:1325–31
84. Stephanopoulos G. 1999. Metabolicfluxes and metabolic engineering. Metab.Eng. 1:1–11
85. Stephanopoulos G, Simpson TW. 1997.Flux amplification in complex metabolicnetworks. Chem. Eng. Sci. 52:2607–27
86. Stephanopoulos G, Valino JJ. 1991. Net-work rigidity and metabolic engineeringin metabolite overproduction. Science252:1675–81
87. Suarato A, Penco S, Arcamone F. 1982.U.S. Patent No. 4,345,068
88. Sumegi B, Butwell NB, Malloy CR,Sherry AD. 1994. Lipoamide influencessubstrate selection in post-ischaemic per-fused rat hearts. Biochem. J. 297:109–13
89. Summers RG, Donadio S, Staver M,Wendt-Pienkowski E, Hutchinson CR,Katz L. 1997. Sequencing and muta-genesis of genes from the erythromycinbiosynthetic gene cluster of Saccharo-polyspora erythraea that are involved inL-mycarose and D-desosamine produc-tion. Microbiology 143:3251–62
90. Tayek JA, Katz J. 1996. Glucose produc-tion, recycling, and gluconeogenesis in
normals and diabetics: a mass isoto-pomer [U-13C]glucose study. Am. J.Physiol. 270:E709–17
91. Tolan JS, Finn RK. 1987. Fermentationof D-xylose and L-arabinose to ethanol byErwinia chrysanthemi. Appl. Environ.Microbiol. 53:2033–38
92. Tolan JS, Finn RK. 1987. Fermentationof D-xylose to ethanol by geneticallymodified Klebsiella planticola. Appl.Environ. Microbiol. 53:2039–44
93. Tong IT, Cameron DC. 1992. Enhance-ment of 1,3- propanediol production bycofermentation in Escherichia coli ex-pressing Klebsiella pneumoniae dharegulon genes. Appl. Biochem. Biotech-nol. 34/35:149–59
94. Tong IT, Liao HH, Cameron DC. 1991.1,3-propanediol production by Esche-richia coli expression genes from theKlebsiella pneumonia dha regulon. Appl.Environ. Microbiol. 57:3541–46
95. Tsunekawa H, Azuma S, Okabe M, Oka-moto R, Aiba S. 1992. Acquisition ofsucrose utilization system in Escherichiacoli K-12 derivatives and its applicationto industry. Appl. Environ. Microbiol.56:2081–88
96. Van der Oost J, Ciaramella M, MoracciM, Pisani FM, Rossi M, de Vos WM.1998. Molecular biology of hyperther-mophilic Archaea. Adv. Biochem. Eng.Biotechnol. 61:87–115
97. Vara J, Lewandowska-Skarbek M,Wang Y-G, Donadio S, HutchinsonCR. 1989. Cloning of genes governingthe deoxysugar portion of the eryth-romycin biosynthesis pathway in Sac-charopolyspora erythraea (Streptomy-ces erythreus). J. Bacteriol. 171:5872–81
98. Varma A, Boesch BW, Palsson BO.1993. Stoichiometric interpretation ofEscherichia coli glucose catabolism un-der various oxygenation rates. Appl. En-viron. Microbiol. 59:2465–73
99. Weber JM, Leung JO, Swanson SJ, IdlerKB, McAlpine JB. 1991. An erythro-
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
METABOLIC ENGINEERING 557
mycin derivative produced by targetedgene disruption in Saccharopolysporaerythraea. Science 252:114–17
100. Webster DA. 1987. Structure and func-tion of bacterial hemoglobin and relatedproteins. In Advances in InorganicChemistry, ed. GC Eichhorn, LG Mar-zilli, pp. 245–65. New York: Elsevier
101. Wei M-L, Webster DA, Stark BC. 1998.Metabolic engineering of Serratia mar-cescens with the bacterial hemoglobingene: alterations in fermentation path-ways. Biotechnol. Bioeng. 59:640–46
102. Weiss RB. 1992. The anthracyclines—will we ever find a better doxorubicin.Semin. Oncol. 19:670–86
103. Wheeler GL, Jones MA, Smirnoff N.1998. The biosynthetic pathway of vita-min C in higher plants. Nature 393:365–69
104. Wood BE, Ingram LO. 1992. Ethanolproduction from cellobiose, amorphouscellulose, and crystalline cellulose by re-combinant Klebsiella oxytoca containingchromosomally integrated Zymomonas
mobilis genes for ethanol production andplasmids expressing thermostable cellu-lase genes from Clostridium thermocel-lum. Appl. Environ. Microbiol. 58:2103–10
105. Yarmush ML, Berthiaume F. 1997. Met-abolic engineering and human disease.Nat. Biotechnol. 15:525–28
106. Yokota A, Shiio I. 1988. Effects of re-duced citrate synthase activity and feed-back-resistant phosphoenol pyruvate car-boxylase on lysine productivities ofBrevibacterium flavum mutants. Agric.Biol. Chem. 52:455–63
107. Zhang M, Eddy C, Deandra K, Finkel-stein M, Picataggio S. 1995. Metabolicengineering of a pentose metabolismpathway in ethanologenic Zymomonasmobilis. Science 267:240–43
108. Zhao LS, Sherman DH, Liu HW. 1998.Biosynthesis of desosamine: construc-tion of a new methymycin/neomethy-mycin analogue by deletion of a deso-samine biosynthetic gene. J. Am. Chem.Soc. 120:10265–57
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
Annual Review of Bimedical Engineering Volume 1, 1999
CONTENTSA Dedication in Memoriam of Dr. Richard Skalak, Thomas C. Skalak 1
Tissue Engineering: Orthopaedic Applications, C. T. Laurencin, A. M. A. Ambrosio, M. D. Borden, J. A. Cooper Jr. 19
Airway Wall Mechanics, Roger D. Kamm 47
Biomechanics of Microcirculatory Blood Perfusion, Geert W. Schmid-Schönbein 73
Engineering and Material Considerations in Cell Transplantation, Elliot L. Chaikof 103
Bioreactors for Haematopoietic Cell Culture, Lars Keld Nielsen 129
Implanted Electrochemical Glucose Sensors for the Management of Diabetes, Adam Heller 153
Injectable Electronic Identification, Monitoring, and Stimulation Systems, Philip R. Troyk 177
Robotics for Surgical Applications, Robert D. Howe, Yoky Matsuoka 211
Transport of Molecules, Particles, and Cells in Solid Tumors, Rakesh K. Jain 241
Nucleic Acid Biotechnology, Charles M. Roth, Martin L. Yarmush 265
Fluid Mechanics of Vascular Systems, Diseases, and Thrombosis, David M. Wootton, David N. Ku 299
Automatic Implantable Cardioverter-Defibrillators, William M. Smith, Raymond E. Ideker 331
Engineering Aspects of Hyperthermia, Robert B. Roemer 347
3-D Visualization and Biomedical Applications, Richard A. Robb 377
Microfabrication in Biology and Medicine, Joel Voldman, Martha L. Gray, Martin A. Schmidt 401
Engineering Design of Optimal Strategies for Blood Clot Dissolution, Scott L. Diamond 427
Cellular Microtransport Processes: Intercellular, Intracellular and Aggregate Behavior, Johannes M. Nitsche 463
New Strategies for Protein Crystal Growth, J. M. Wiencek 505
Metabolic Engineering, M. Koffas, C. Roberge, K. Lee, G. Stephanopoulos 535
Ultrasound Processing and Computing: Review and Future Directions, George York, Yongmin Kim 559
Telemedicine, Seong K. Mun, Jeanine W. Turner 589
Imaging Transgenic Animals, T. F. Budinger, D. A. Benaron, A. P. Koretsky 611
Instrumentation for the Genome Project, J. M. Jaklevic, H. R. Garner, G. A. Miller 649
Ann
u. R
ev. B
iom
ed. E
ng. 1
999.
1:53
5-55
7. D
ownl
oade
d fr
om w
ww
.ann
ualr
evie
ws.
org
by R
EN
SSE
LA
ER
PO
LY
TE
CH
NIC
IN
STIT
UT
E o
n 06
/17/
14. F
or p
erso
nal u
se o
nly.
Related Documents
