DOI: 10.1126/science.1168243 , 161 (2009); 325Science

et al.Jesse W.-H. Li,or an Endless Frontier?Drug Discovery and Natural Products: End of an Era

www.sciencemag.org (this information is current as of September 26, 2009 ):The following resources related to this article are available online at

http://www.sciencemag.org/cgi/content/full/325/5937/161version of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/cgi/content/full/325/5937/161#otherarticles, 5 of which can be accessed for free: cites 41 articlesThis article

http://www.sciencemag.org/cgi/content/full/325/5937/161#otherarticles 1 articles hosted by HighWire Press; see: cited byThis article has been

http://www.sciencemag.org/cgi/collection/chemistryChemistry

: subject collectionsThis article appears in the following

http://www.sciencemag.org/about/permissions.dtl in whole or in part can be found at: this article

permission to reproduce of this article or about obtaining reprintsInformation about obtaining

registered trademark of AAAS. is aScience2009 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Sep

tem

ber

26, 2

009

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

Drug Discovery and Natural Products:End of an Era or an Endless Frontier?Jesse W.-H. Li and John C. Vederas*

Historically, the majority of new drugs have been generated from natural products (secondarymetabolites) and from compounds derived from natural products. During the past 15 years,pharmaceutical industry research into natural products has declined, in part because of an emphasis onhigh-throughput screening of synthetic libraries. Currently there is substantial decline in new drugapprovals and impending loss of patent protection for important medicines. However, untappedbiological resources, “smart screening” methods, robotic separation with structural analysis, metabolicengineering, and synthetic biology offer exciting technologies for new natural product drug discovery.Advances in rapid genetic sequencing, coupled with manipulation of biosynthetic pathways, mayprovide a vast resource for the future discovery of pharmaceutical agents.

Just over 200 years ago, a 21-year-old phar-macist’s apprentice named Friedrich Sertürnerisolated the first pharmacologically active

pure compound from a plant: morphine fromopium produced by cut seed pods of the poppy,Papaver somniferum (1). This initiated an erawherein drugs from plants could be purified,studied, and administered in precise dosagesthat did not vary with the source or age of thematerial. Pharmaceutical research expandedafter the Second World War to include massivescreening of microorganisms for new antibioticsbecause of the discovery of penicillin. By 1990,about 80% of drugs were either natural productsor analogs inspired by them. Antibiotics (e.g.,penicillin, tetracycline, erythromycin), antipara-sitics (e.g., avermectin), antimalarials (e.g., quinine,artemisinin), lipid control agents (e.g., lovastatinand analogs), immunosuppressants for organtransplants (e.g., cyclosporine, rapamycins) andanticancer drugs (e.g., taxol, doxorubicin) revo-lutionized medicine. Life expectancy in much ofthe world lengthened from about 40 years earlyin the 20th century to more than 77 years today.Although the expansion of synthetic medicinalchemistry in the 1990s caused the proportion ofnew drugs based on natural products to drop to~50% (Fig. 1), 13 natural product–derived drugswere approved in the United States between 2005and 2007, with five of them being the first mem-bers of new classes (2).

With such a successful record, it might beexpected that the identification of new metabo-lites from living organisms would be the core ofpharmaceutical discovery efforts. However, manypharmaceutical firms have eliminated their natu-ral product research in the past decade. Althoughmore than 100 natural product–based drugs arein clinical studies, this represents about a 30% drop

between 2001 and 2008. Is the era of discoveryof new drugs from natural sources ending?

What Challenges Face Drug Discoveryfrom Natural Sources?Most of the current difficulties can be dividedinto two categories: the prevailing paradigm fordrug discovery in large pharmaceutical indus-tries, and technical limitations in identifyingnew compounds with desirable activity.

The pharmaceutical environment. The double-digit yearly sales growth that drug companiestypically enjoyed until about 10 years ago hasled to unrealistically high expectations by theirshareholders and great pressure to produce“blockbuster drugs” with more than $1 billionin annual sales (3). In the blockbuster model, afew drugs make the bulk of the profit. For ex-ample, eight products accounted for 58% of

Pfizer’s annual worldwide sales of $44 billion in2007. When such drugs lose patent protection,their sales revenue can drop by 80%. About 25%of the current U.S. drug market will lose patentprotection within 4 years (4). This will removemore than $63 billion in annual income for phar-maceutical industries by 2014. Competition fromgeneric drug manufacturers, which are generallynot involved in drug discovery, accounts for 67%of all prescriptions in the United States and isencouraged by health agencies to reduce costs.

The financial outlook for firms doing drugdiscovery is further encumbered by extensivelitigation, costs of competitive marketing, andincreasing expectations for safety both by thepublic and by regulatory agencies such as theU.S. Food and Drug Administration (FDA). FDAapprovals of new drugs reached a 24-year lowas of 2007, and drugs approved in Europe havebeen rejected by that agency. As an example oflegal costs, after the withdrawal of the anti-inflammatory drug Vioxx because of a potentialincrease in risk of heart attack and stroke, Merckhad to set aside $970 million to pay for asso-ciated legal expenses in 2007, and another $4.85billion for upcoming U.S. legal claims (5). Oneapproach to dealing with rising costs and adwindling pipeline of new drugs is purchase ofother companies that have such resources.Recent examples include the purchase of Wyethby Pfizer for $68 billion and the acquisition ofSchering-Plough by Merck for $41 billion (6).Such large sums affect the way drug discoveryis done: Firms involved in drug discovery musthit the target not only accurately, but very quicklyand very profitably. However, for reasons out-lined below, natural product sources are currentlynot very amenable to rapid high-throughputscreening (HTS) for desirable activity as drugs

REVIEW

Department of Chemistry, University of Alberta, Edmonton,Alberta T6G 2G2, Canada.

*To whom correspondence should be addressed. E-mail:john.vederas@ualberta.ca

Natural product–derived

19810

10

20

30

40

50

60

70

80 Number of drugs

1983 1985 1987 1989 1991 1993 1995Year

Nu

mb

er

1997 1999 2001 2003 2005 2007

Natural products

Fig. 1. Number of drugs approved in the United States from 1981 to 2007.

www.sciencemag.org SCIENCE VOL 325 10 JULY 2009 161

on

Sep

tem

ber

26, 2

009

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

(7). Furthermore, in contrast to synthetic libraries,hits from natural sources are likely to have com-plex structures with numerous oxygen-containingsubstituents and an abundance of centers ofstereochemistry (8). This slows the identificationprocess and contributes to problems of supplyand manufacture.

Difficulties in discovering natural product drugcandidates. Historically, screening of natural ma-terials for biological activity hasworked well. Considering only poly-ketide metabolites, just over 7000known structures have led to morethan 20 commercial drugs with a “hitrate” of 0.3%, which is much betterthan the <0.001% hit rate for HTSof synthetic compound libraries (9).Although the hopes for useful leadsfrom unfocused combinatorial chem-istry libraries of mixtures have longsince evaporated, pharmaceutical dis-covery efforts currently favor HTSof massive libraries of pure syntheticcompounds. The output has beenquite low, but any hits are usuallyeasy to make and modify with simplechemistry. Libraries of pure com-pounds present in known amountsare also “screen friendly” and accom-modate the desire for short timelinesin examination of a large numberof molecules. The recognition thatsuch libraries are inherently lim-ited prompts increasing interest in“diversity-oriented synthesis” and li-braries of “privileged structures” (oftenbased on known drugs or naturalframeworks) to produce more com-plex molecules with a better chanceof desirable bioactivity (10).

HTS of natural sources alsopresents a variety of difficulties. Prob-lems of reliable access and supply, especially withrespect to higher plants and marine organisms,are compounded by intellectual property concernsof local governments and the Rio Convention onBiodiversity (11). Seasonal or environmental var-iations in the composition of living organismscan cause problems with initial detection of activecompounds as well as subsequent repetition ofassays or purification. Loss of source is also pos-sible: Current extinction rates for natural speciesof higher plants are estimated to be a factor of100 to 1000 over natural background (12). It hasbeen suggested that 15,000 out of 50,000 to70,000 medicinal plant species are threatenedwith extinction (12). Even if supply is easy andguaranteed, the initial extract of the natural materialusually consists of a complex mixture after frac-tionation. It may contain only very small quan-tities of a bioactive substance, often as a mixturewith structurally related molecules. The initialconcentration of an interesting compound may betoo low to be effectively detected by HTS, or theassay may be obscured by poor solubility or by

fluorescent or colored contaminants. The keycompound may be unstable in the mixture. Afurther complication can be synergistic (or an-tagonistic) activity of two constituents that maythen diminish or disappear upon separation. Forexample, a number of bacteriocins (antimicro-bial peptides from bacteria) must function astwo-component systems to display full activity(13). Finally, considerable time is often required

to complete structural characterization to deter-mine whether the molecule is already known.

A prevailing sentiment in many pharmaceu-tical organizations is that screening of naturalproduct sources is a difficult effort with a highprobability of duplication; that is, the result maybe a known compound that cannot be patented.However, fewer than 1% of microorganism spe-cies are easily cultured, and perhaps fewer than15% of higher plant species have been examinedfor bioactivity (14). Certain insects and other ani-mals have been targeted for specific bioactivities,such as toxins (15), but are not generally subjectedto HTS efforts. Clearly the biological resource isthere, but access and examination are problematic,especially if there is pressure for a short timeframe for discovery of new leads.

What Tools Are Emerging to Enhanceand Accelerate Drug Discovery fromNatural Sources?Several interlocking phases of exploration ofnatural sources can be considered: access to the

biological resource, appropriate screening of thatresource to locate a useful activity, analysis ofthe structure of the key compound, generationof analogs for optimal activity, and productionof the target drug.

Biological resources. Traditionally, soil bacte-ria (especially actinomycetes), fungi, and higherplants were main sources for drug discovery.Pharmaceutical firms increasingly abandoned

screening of microorganisms after 1990 be-cause of decreasing success rates. Common anti-biotics such as streptomycin occur in ~1% of soilactinomycetes and display activity in screens;this activity masks interesting new antimicro-bials, which may be produced at a frequency ofless than 1 in 10 million fermentations. A solu-tion pursued by Cubist Pharmaceuticals is tomassively increase the number of fermentations(to many millions per year) while miniaturizingtheir size, using calcium alginate beads as thecontainers (16). This approach is coupled to anassay with engineered Escherichia coli strainsthat are resistant to common well-known anti-biotics. However, despite the urgent need to findnew antibiotics effective against life-threateningorganisms resistant to current therapy, many phar-maceutical firms do not develop such drugs.Sales and “blockbuster potential” are limited bythe short duration of treatment with antibioticsrelative to other drugs, such as cholesterol-lowering or hypertensive agents, which are con-sumed daily for prolonged periods and relieve

H2N–CKGKGAKCSRLMYDCCTGSCRSGKC–CONH2

CH3

OCH3

OH

H3CO

H

NN

NH

S

OO

O

O

O

O

HO

HO

Conus magus

Ecteinascidia turbinata

A

B

Fig. 2. Marine sources of drugs. (A) Ziconotide (Prialt) from Conus magus. Abbreviations for amino acids: A, Ala; C,Cys; D, Asp; G, Gly; K, Lys; L, Leu; M, Met; R, Arg; S, Ser; T, Thr; Y, Tyr. (B) Trabectedin (Yondelis) from Ecteinascidiaturbinata. [Image of C. magus from (45), reprinted by permission of Macmillan Publishers Ltd.; image of E. turbinataby Florent Charpin, reprinted by permission of Florent’s Guide to the Tropical Reefs (http://reefguide.org)]

10 JULY 2009 VOL 325 SCIENCE www.sciencemag.org162

REVIEW

on

Sep

tem

ber

26, 2

009

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

symptoms rather than provide a cure. None-theless, development costs, standards for safe-ty, and requirements for limited side effectsare similar for both antibiotics and “long-term”drugs.

Although plants remain a major source of newdrugs, with 91 compounds in clinical trials as oflate 2007, cyanobacteria (17) and marine orga-nisms (18) have been actively investigated in recentdecades, especially for neurotoxic and cytotoxiccompounds. Ziconotide (Prialt), a peptide toxinfrom cone snails (Fig. 2A), was approved in 2004for treatment of chronic pain resulting from spinalcord injury. A tropical sea squirt has yielded thecytotoxic trabectedin (Yondelis) (Fig. 2B), which isapproved in Europe since 2007 for treatment ofadvanced soft-tissue sarcoma (19). Useful orga-nisms may exist in extreme environments, suchas at great sea depths (20), in thermal vents, or insalt lakes. An appealing example is the identifica-tion of haloduracin, a two-component lantibiotic(lanthionine-containing peptide an-tibiotic) from Bacillus halodurans,which grows at an extreme pH of>9.0 (21). Awell-known lantibioticis nisin A, which is used to preservefood and is very active against Gram-positive bacteria resistant to conven-tional antibiotics. However, nisin’stherapeutic potential is blocked byits instability at neutral pH or above.Van der Donk and co-workers rea-soned that base-stable lantibioticsmay be produced by bacteria grow-ing in alkaline environments (21).Using bioinformatics, they foundhaloduracin, which can survive pHranges well above that of humanserum. Although haloduracin’s sol-ubility is limited, it provides a basisfor development of new lantibioticswith drug potential.

It may initially appear that thereare few unexplored locations to lookfor natural sources of drug can-didates, but enormous numbers ofspecies have remained unexam-ined. There are claims that morethan 99% of all bacteria cannot becultured, and that in marine envi-ronments there may be 3.7 × 1030

microorganisms, many of whichmay produce fascinating naturalproducts as drug candidates (22).To address this, in 1998 the conceptof metagenomics was proposed tolook at genes and their function insamples obtained directly from theenvironment (23). This field has ex-ploded as faster and cheaper genesequencing is becoming available(24) in combination with the abilityto rapidly sort cells from the envi-ronment and efficiently clone genesin improved vectors (25). Together

with automated screening techniques, the meta-genomic approach could afford access to the poolof 99% of unexamined microorganisms. Howev-er, at present most pharmaceutical firms do notappear to be undertaking efforts in this directionfor drug discovery.

Appropriate screening. It is likely that naturalproducts represent privileged structures for drugdiscovery. This suggestion is supported by thefact that there are a limited number of proteinfolds known, and that natural products mustbind to some of these in order to be biosynthe-sized and to fulfill their inherent function inthe producing organisms. Hence, many of themmay be structurally favored to bind to enzymesor protein receptors. However, as describedabove, the complexity of the initial natural ex-tract can make it unfavorable for HTS. The idealapproach to overcome this obstacle would beautomated separation of all constituents in theorganism into individual components, coupled

with full spectroscopic identification prior toHTS. Although this is not yet achievable, Ire-land and co-workers (26) recently automatedfractionation of crude extracts of natural mate-rials from marine sources using desalting fol-lowed by high-performance liquid chromatography(HPLC) on highly efficient monolithic columns.This was in turn coupled to mass spectrometricanalysis and collection on HTS plates. The pro-cess produces highly purified samples in 96-wellplates, which are generated as replicates for initialscreening as well as for a material archive. It isclaimed that typically only three compounds perwell are observed (26). Testing of a 15,360-memberlibrary of this type against hamster cell linesallowed identification of novel compounds withantitumor potential selective for breast cancerdespite the co-occurrence of general cytotoxinsin the same extract.

The efficiency of HTS can be greatly enhancedby using the best target to achieve “smart screen-

ing.”One approach mentioned aboveis the use of organisms resistant tocommon known antibiotics, as usedby Cubist. An innovative alternativewas developed at Merck to identifybroad-spectrum antibiotics (Fig. 3)(27). It uses a two-plate assay, inwhich one plate has Staphylococcusaureus bearing a plasmid that pro-duces antisense RNA to a key fattyacid synthase enzyme (e.g., FabF)and the other is a S. aureus controlplate without capability to producethe antisense RNA. The antisenseRNA causes degradation of themRNA at the 5′ end for the key en-zyme, thereby enhancing the orga-nism’s sensitivity for inhibitors ofthat particular protein catalyst. Thisapproach enables more sensitivedetection of activity, and also per-mits identification of individual tar-gets in complex systems throughparallel screening with different anti-sense RNAs. The hit rate for suchscreening of more than 250,000natural product extracts was high(0.3%). More important, it identi-fied a new target, bacterial fattyacid synthesis, and novel antibiot-ics, platensimycin A and platencin.

Increasingly, multiple targets arebeing investigatedwith the use of cells.An interesting example is single-cellscreening of inhibitors of phos-phorylation (kinase) signaling path-ways using flow cytometry (28).This phosphospecific flowcytometry(phosphoflow)makesmultiple quan-titativemeasurements of phosphoryl-ation levels of different signalingproteins by measuring specific fluo-rescently labeled antibodies thatrecognize them after phosphate at-

Wild type

HTS

>250,000 naturalproduct extracts

Activity-guidedfractionation of natural

product extracts

Hypersensitive mutant

Wild type

Platensimycin

O

O

O

R

OH

HO

HO2C

HN

R =

Platencin

Hypersensitive

O

R

Fig. 3. HTS using differential-sensitivity whole-cell two-plate agar diffu-sion (the Merck platensimycin assay). The strain expressing antisense RNAto FabH or FabF enzymes of fatty acid biosynthesis is hypersensitive toinhibitors of those proteins. The approach identified two new antibiotics,platensimycin and platencin.

www.sciencemag.org SCIENCE VOL 325 10 JULY 2009 163

REVIEW

on

Sep

tem

ber

26, 2

009

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

tachment. The effect of natural product librariescan be rapidly and quantitatively measured onsingle cells. Cell-based assays have also recentlybeen used to find rapamycin analogs that lack thedrug’s usual immunosuppressive activity butprotect nerve cells in models of stroke (29).

Analysis of the structure of active compounds.Structure elucidation on small quantities of ma-terial has been greatly assisted by advances inmass spectrometry and multidimensional nucle-ar magnetic resonance (NMR) spectroscopy. Itis beyond the scope of this review to discuss thedetails of these techniques. However, automatedcoupling of mass spectrometry to separation byHPLC and HTS library creation clearly accel-erates the identification of known compoundsand possible hits. Cryoprobes for NMR spec-troscopy have greatly enhanced sensitivity andhave markedly reduced the amount of material re-quired for analysis; for smaller compounds (mo-lecular weight 200), useful proton spectra require2 mg of material and carbon correlation spectracan be obtained on 0.2 mg (30). Comparison ofNMR signal positions to corresponding databasesof known compounds can hasten dereplication(i.e., recognition of known compounds). Effortsare also under way to automatically capture sub-stances from HPLC separations by solid-phaseextraction and then elute directly into an NMRcryoprobe for analysis.

Generation of analogs. Traditionally, structure-activity relationships in bioactive natural productswere examined by simple chemical transfor-mations. For example, methylation of the sidechain of the cholesterol-lowering agent lovastatinproduces simvastatin (Zocor) (Fig. 4), an im-proved drug that had sales in excess of $4.3 bil-lion in 2006 (before loss of patent protection).The developing understanding of secondary me-tabolite biogenesis allows use of biosynthetic en-zymes or genetically altered organisms to generatederivatives of drugs. Tang and co-workers usedan esterase (LovD), which normally attaches theside chain of lovastatin, to make a series of an-alogs, including simvastatin (31). The esterasenormally uses a 2-methylbutyrate attached as athioester to a large protein as its acylating agentfor the corresponding alcohol. However, a varietyof simple thioesters could be substituted to giveefficient conversion to many analogs.

Combinatorial biosynthesis, especially usingmodular polyketide synthases (PKSs) and non-ribosomal peptide synthetases (NRPSs), holdsgreat promise (32). PKSs and NRPSs are largemultidomain enzymes that sequentially condenseshort fatty acids and a–amino acids, respectively.They resemble assembly lines for making a me-tabolite by chain elongation and functional grouptransformation, and can be altered to make newcompounds. However, the results of mutationare frequently unpredictable, and the levels ofproduct formation can be very low. The growingunderstanding of linkers between individual do-mains in these proteins as well as protein-proteininteractions between domains may solve this

problem. One approach to overcoming difficul-ties resulting from introduction of a modified do-main is directed evolution to increase productionby the chimera. Only three rounds of mutagen-esis and screening of modest libraries (103 to 104

clones) of an NRPS domain were sufficient togive substantial improvement of production of anisoleucine-containing analog of andrimid (Fig. 4)having better antibiotic properties (33). In a dif-ferent example, direct mutation of a PKS domaincombined with inactivation of a gene responsiblefor post-assembly oxidation gave nystatin ana-

logs (Fig. 4) with improved antifungal activityand lower toxicity (34). Inactivation of a post-PKS P450 that oxidizes a methyl to the C16 car-boxyl in nystatin, coupled with mutation of theenoyl reductase activity in the NysC PKS mod-ule, affords methyl derivatives with a conjugatedheptaene moiety (as opposed to an interruptedhexaene). The best compounds are effective againstdisseminated candidosis in mouse models andare considerably less toxic than amphotericin

B. However, such generation of new analogsusing combinatorial biosynthesis can require ex-tensive time and effort, and is currently notmatched to the requirements of HTS.

Production of target compounds. The tradi-tional approach to optimal production of drugsby microorganisms is to “mutate and screen” forstrain improvement, as was done in the 1940sfor penicillin. This methodology can readily yieldenhancement of drug production by two to threeorders of magnitude, and in some cases by fourto five orders of magnitude (35). At present, this

tactic is difficult to beat, and molecular biolog-ical manipulations have been more successfulonly in isolated cases. An interesting example isthe improvement of production of doramectin(Dectomax) (Fig. 4), an antiparasitic agent in theavermectin family (36). The initial productionstrain was a mutant of Streptomyces avermitilislacking a dehydrogenase and an O-methyl trans-ferase. This strain can be fed cyclohexane car-boxylic acid as the initiator to give doramectin

Lovastatin R = HSimvastatin R = CH3

Nystatin R = COOHNystatin analog R = CH3, 28,29-didehydro

Doramectin

Andrimid R = HAndrimid - isoleucine R = CH3

Fig. 4. Drug analogs produced by modification of biosynthetic genes. Shaded areas indicate sitesof altered structure; Me, methyl.

10 JULY 2009 VOL 325 SCIENCE www.sciencemag.org164

REVIEW

on

Sep

tem

ber

26, 2

009

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

together with a side product having a hydroxylat C-23 instead of a double bond. DNA shuf-fling of the aveC gene and screening gave a newgene that encoded 10 amino acid mutations andreduced the unwanted contaminant for an over-all improvement of production by a factor of 23.Natural products, including those from bacteria,can of course also be made or improved syn-thetically. For example, a large amount of efforthas gone into synthetic analogs of epothilone B,a promising anticancer drug (37).

Drugs with complex structures from higherplants present a challenge if they are found insmall concentrations. An example is the anti-tumor agent taxol (paclitaxel), which occurs inthe bark of the pacific yew tree Taxus brevifolia(38). Initially it was predicted that treatment ofovarian cancer and melanoma cases in the UnitedStates would require the destruction of more than360,000 yew trees annually. However, a semi-synthetic route to taxol from 10-deacetylbaccatin,which was isolated from the needles of the Euro-pean yew T. baccata, was developed that avertedthe devastation of trees.

In the past decade, taxol has been commer-cially produced by Bristol-Myers Squibb by plantcell fermentation (PCF) technology. However,for a number of other drugs there are substantialdifficulties in getting sufficient production byPCF (39). An alternative approach is heterolo-gous expression of the biosynthetic pathway inyeast or bacteria, frequently with modificationof the genes to optimize production (syntheticbiology). A well-known example is the effortto produce the antimalarial drug artemisinin inE. coli and yeast (40). Artemisinin from Artemesiaannua is a sesquiterpenoid that is effectiveagainst multi–drug-resistant Plasmodium speciesbut is expensive for Third World patients. TheKeasling group (40) has engineered E. coli toproduce its precursor, artemisinic acid, inconcentrations of up to 300 mg per liter. Thisrequired extensive work, including engineering amevalonate pathway to produce ample amountsof precursor for isoprenoid synthesis, optimizationof expression of amorphadiene synthase (the keyterpene cyclase), and incorporation of a modifiedversion of amorphadiene oxidase (the P450 thatconverts amorphadiene to artemisinic acid). Thework demonstrates that many complex naturalproducts from plants can be engineered into het-erologous hosts for fermentative production.

What Are the Future Prospects forNatural Product Drugs?With the current framework of HTS in majorpharmaceutical industries and increasing govern-ment restrictions on drug approvals, it is pos-sible that the number of new natural product–derived drugs could go to zero. However, this islikely to be temporary, as the potential for newdiscoveries in the longer term is enormous. Ac-cess to rapid and inexpensive genome sequenc-

ing via 454 sequencing (41) or single-moleculereal-time (SMRT) (42) methods will fully en-able metagenomics for unculturable organisms. Itwill also uncover “silent pathways” in plants (43)to afford access to a large collection of newproducts and biocatalysts. It will allow preserva-tion of any threatened species, through catalogingof its genetic blueprint, and may permit recoveryof extinct organisms. Estimates for the total num-ber of living species range from 2 million to 100million, with one claim of 30 million species justfor insects. Hence, the number of biosyntheticproducts and enzymes remaining to be examinedis huge. Systems biology could eventually mapthe likely metabolism for most species. Such alibrary of biochemical transformations could be amagnificent tool for the design and generation ofnew products. Just as synthetic chemists currentlyplan total syntheses of a target compound usingestablished reagents and well-precedented trans-formations, synthetic biologists will be able tocall on vast arrays of enzymes to rationally makecomplex molecules. Directed evolution and site-specific mutation can optimize the desired activ-ity of such proteins. Rapid gene sequencing ofindividual humans will enhance the developmentof personalized medicine—the use of an individ-ual’s DNA sequence as a basis for drug selection(44). This could satisfy the current expectationsfor high levels of safety by predicting side effectsand assisting the correct choice of therapeuticdrugs. New gene-mapping techniques will facil-itate speedy diagnostic tests to determine causesof illness, including infections. This could lessenthe indiscriminate use of antibiotics and therebyreduce the development of bacterial resistance tosuch drugs.

To achieve the potential of facile genome se-quencing, the development of robust platformsis essential for heterologous expression of genesof novel biosynthetic pathways. Expression ofbiosynthetic enzymes often requires considerableeffort. Problems of codon usage and optimization(requiring gene synthesis), of protein localiza-tion and modification, and of metabolite toxicityto the producing organism are only a few of thedifficulties. The solutions will involve modifi-cations of organisms that are already widely used(e.g., E. coli, Saccharomyces cerevisiae). For ex-ample, enhancement of primary precursor pro-duction can be coupled with programmed controlof promoters and modifications to enhance thestability of foreign proteins. Active export of toxictarget metabolites could potentially be engineeredusing drug resistance transporters. It is likely thatsuch expression platforms will initially be dem-onstrated in academic laboratories prior to usein pharmaceutical industry. Although the cur-rent industry model for drug discovery does notfavor natural products, the resource is so vast asto seem unlimited, and these emerging tools willprovide exhilarating discoveries leading to newmedicines.

References and Notes1. G. R. Hamilton, T. F. Baskett, Can. J. Anaesth. 47, 367

(2000).2. A. L. Harvey, Drug Discov. Today 13, 894 (2008).3. N. N. Malik, Drug Discov. Today 13, 909 (2008).4. S. Houlton, Chem. World (January 2009), p. 12.5. J. M. Wolsing, Defense Counsel J. (July 2008), p. 209.6. L. M. Jarvis, Chem. Eng. News 87, 5 (16 March 2009).7. F. E. Koehn, G. T. Carter, Nat. Rev. Drug Discov. 4, 206

(2005).8. M. S. Butler, J. Nat. Prod. 67, 2141 (2004).9. K. J. Weissman, P. F. Leadlay, Nat. Rev. Microbiol. 3, 925

(2005).10. D. J. Newman, J. Med. Chem. 51, 2589 (2008).11. B. E. Kirsop, J. Ind. Microbiol. Biotechnol. 17, 505

(1996).12. V. Brower, J. Natl. Cancer Inst. 100, 838 (2008).13. S. Garneau, N. I. Martin, J. C. Vederas, Biochimie 84, 577

(2002).14. A. Saklani, S. K. Kutty, Drug Discov. Today 13, 161

(2008).15. S. Mouhat, N. Andreotti, B. Jouirou, J. M. Sabatier,

Curr. Pharm. Des. 14, 2503 (2008).16. R. H. Baltz, Curr. Opin. Pharmacol. 8, 557 (2008).17. N. V. Wase, P. C. Wright, Expert Opin. Drug Discov. 3,

903 (2008).18. J. W. Blunt et al., Nat. Prod. Rep. 26, 170 (2009).19. D. S. Dalisay, S. L. Lievens, J. P. Saludes, T. F. Molinski,

Nat. Rev. Drug Discov. 8, 69 (2008).20. D. Skropeta, Nat. Prod. Rep. 25, 1131 (2008).21. L. E. Cooper, A. L. McClerren, A. Chary, W. A. van der

Donk, Chem. Biol. 15, 1035 (2008).22. J. Kennedy, J. Marchesi, A. D. W. Dobson, Microb. Cell

Fact. 7, 27 (2008).23. J. Handelsman, M. R. Rondon, S. F. Brady, J. Clardy,

R. M. Goodman, Chem. Biol. 5, R245 (1998).24. N. Blow, Nature 453, 687 (2008).25. P. Hugenholtz, G. W. Tyson, Nature 455, 481 (2008).26. T. S. Bugni et al., J. Nat. Prod. 71, 1095 (2008).27. H. Jayasuriya et al., Angew. Chem. Int. Ed. 46, 4684

(2007).28. P. O. Krutzik, J. M. Crane, M. R. Clutter, G. P. Nolan,

Nat. Chem. Biol. 4, 132 (2008).29. B. Ruan et al., Proc. Natl. Acad. Sci. U.S.A. 105, 33

(2008).30. G. Lang et al., J. Nat. Prod. 71, 1595 (2008).31. H. Zhou, X. Xie, Y. Tang, Curr. Opin. Biotechnol. 19, 590

(2008).32. K. J. Weissman, R. Müller, ChemBioChem 9, 826 (2008).33. M. A. Fischbach, J. R. Lai, E. D. Roche, C. T. Walsh,

D. R. Liu, Proc. Natl. Acad. Sci. U.S.A. 104, 11951(2007).

34. T. Brautaset et al., Chem. Biol. 15, 1198 (2008).35. C. Olano, F. Lombo, C. Mendez, J. A. Salas, Metab. Eng.

10, 281 (2008).36. K. Stutzman-Engwall et al., Metab. Eng. 7, 27 (2005).37. J. Mulzer, K. H. Altmann, G. Höfle, R. Müller, K. Prantz,

C. R. Chim. 11, 1336 (2008).38. 2004 Greener Synthetic Pathways Award: Bristol-Myers

Squibb Company, Development of a Green Synthesis forTaxol Manufacture via Plant Cell Fermentation andExtraction (www.epa.gov/greenchemistry/pubs/pgcc/winners/gspa04.html), accessed 6 April 2009.

39. S. C. Roberts, Nat. Chem. Biol. 3, 387 (2007).40. J. D. Keasling, ACS Chem. Biol. 3, 64 (2008).41. J. M. Rothberg, J. H. Leamon, Nat. Biotechnol. 26, 1117

(2008).42. J. Eid et al., Science 323, 133 (2009); published online

20 November 2008 (10.1126/science.1162986).43. E. Lewinsohn, M. Gijzen, Plant Sci. 176, 161 (2009).44. S. S.-J. Lee, A. Mudaliar, Science 323, 342 (2009).45. K. Garber, Nat. Biotechnol. 23, 399 (2005).46. Supported by the Natural Sciences and Engineering

Research Council of Canada, the Canada Research Chairsprogram, and the University of Alberta.

10.1126/science.1168243

www.sciencemag.org SCIENCE VOL 325 10 JULY 2009 165

REVIEW

on

Sep

tem

ber

26, 2

009

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om