全 文 ：Biosynthesis and Genetic Engineering of Polyketides
?
ZHU Xiang-Cheng
1 , 2
, HUFFMAN Justin
1
, GERBER Ryan
1
, LOU Li-Li
1
, XIE Yun-Xuan
1
,
LIN Ting1 , 3 , JORGENSON Joel1 , MARESCH Andrew1 , VOGELER Chad1 ,
WANG Qiao-Mei
4
, SHEN Yue-Mao
3
, DU Liang-Cheng
1??
( 1 Department of Chemistry, University of Nebraska-Lincoln, Lincoln, Nebraska 68588 , USA ; 2 Department of Chemical Engineering
and Bioengineering, Zhejiang University, Hangzhou 310027 , China; 3 School of LifeScience, Xiamen University,
Xiamen 361005 , China; 4 Department of Horticulture, Zhejiang University, Hangzhou 310029 , China)
Abstract: Polyketides are one of the largest groups of natural products produced by bacteria, fungi , and plants . Many of
thesemetabolites have highly complex chemical structures and very important biological activities, including antibiotic, anti-
cancer, immunosuppressant, and anti-cholesterol activities . In the past two decades, extensive investigations have been car-
ried out to understandthemolecular mechanisms for polyketidebiosynthesis . Theseefforts have ledto thedevelopment of var-
ious rational approaches towardengineeredbiosynthesis of new polyketides .More recently, theresearchefforts have shiftedto
the elucidation of thethree-dimentional structureof thecomplex enzymemachineriesfor polyketide biosynthesisand to theex-
ploitation of new sources for polyketide production, such as filamentous fungi and marine microorganisms . This review sum-
marizes our general understanding of the biosynthetic mechanisms andthe progress in engineered biosynthesis of polyketides .
Key words: Polyketides; Polyketide Synthase; Biosynthesis; Engineered Biosynthesis
CLC number : Q 81 , Q 946 Document Code : A Article ID : 0253 - 2700 (2008) 03 - 249 - 30
Polyketides are a large family of structurally and
functionally diverse natural products produced by bac-
teria, fungi and plants .Thiswidespreadgroupof natu-
ral products possesses a broad range of important bio-
logical and medical properties, such as antibiotic, an-
tifungal , anticancer, antiparasitic, and immunosup-
pressive activities ( Fig . 1 ) (Staunton and Weissman,
2001) . Polyketides are biosynthesized fromshort chain
acyl units, such as acetate, propionate, malonate or
methylmalonate, through sequential Claisen condensati-
ons . This is similar to the biosynthesis of fatty acids .
The polyketide biosynthesis is catalyzed by a complex
enzyme machinery named polyketide synthase ( PKS) ,
which is analogous to fatty acid synthase ( FAS) . The
significanceof polyketides in agriculture, food industry
and human health has inspired extensive investigations
of the enzyme complexes in thepast two decades . Many
mechanistic insights have been revealed, which provid-
ed the foundation for rational design and manipulation
of polyketide biosynthesis to produce new products .
Polyketide biosynthesis is one of the most active
areas in current research . In the past two decades there
have been numerous papers published . For reviews
alone, more than 200 articles reviewing polyketides
have appeared in the literature, including some of the
more recent ones ( Hopwood, 1997; Khosla et al. ,
1999; Shen, 2000; Staunton and Weissman, 2001;
Baltz, 2006; Deng and Bai , 2006; Fischbach and
Walsh, 2006; Lai et al. , 2006; Donadio et al. ,
2007; Hertweck et al. , 2007; Khosla et al. , 2007;
Kopp and Marahiel , 2007; Menzella and Reeves,
2007; Van Lanen and Shen, 2008) . Readers are ad-
vised to consult the literature for amore comprehensive
understanding of thetopic . Thegoal of this reviewis to
云 南 植 物 研 究 2008 , 30 (3) : 249～278
Acta Botanica Yunnanica DOI : 10 .3724?SP. J . 1143 .2008.08053
?
?? ?Author for correspondence; E-mail : ldu@unlserve.unl . edu
Received date: 2008 - 03 - 15 , Accepted date: 2008 - 04 - 02
Dedicated to the 30th Anniversary of Acta Botanica Yunnanica
* Foun dation items: Supported in part by NSF (MCB-0614916) , Nebraska Research Initiatives ( NRI ) , Redox Biology Center ( RCB) Pilot Grant, and
NSFC OverseaYoung Scholar Award (No . 30428023) . The research was performed in facilities renovated with support fromNIH ( RR015468-
01) . JORGENSON Joel , MARESCH Andrew, and VOGELER Chad are supported by the UCARE program at University of Nebraska-Lincoln
summarize thegeneral understandingof the biosynthetic
mechanisms employedby the different typesof PKS and
to highlight themain achievementsobtained in the area
of polyketide biosynthetic engineering .
1 Classification of PKSs
The classification of PKS follows the same way
used by fatty acid synthases ( FAS) , due to the many
similarities shared by the biosynthetic process of
polyketides and fatty acids . The long carbon chains of
both polyketides and fatty acids are assembled through
repetitiveClaisen condensations using short chain acyl-
Coenzyme A ( CoA ) as building blocks . FAS normally
use acetyl-CoA as the starter and malonyl-CoA as the
extender . Each condensation reaction elongates the lin-
ear chain with a two-carbon unit derived fromdecarbox-
ylated malonyl-CoA , and thenewly formedβ-keto group
is completely reduced through ketoreduction, dehydra-
tion, and enoylreduction . PKS catalyze the same basic
chemical reactions as those catalyzed by FAS, except
that PKS are more flexible in substrate selectivity,
which leads to the difference in the reduction levels of
Fig . 1 Examples of polyketides from bacteria, fungi , and plants . The biological activity of each compound is indicated in italics under the structure
052 云 南 植 物 研 究 30 卷
theβ-keto groups, the length of the polyketide chain
and the degree of post-PKS modifications . These flexi-
bilities are the reasons for thevast diversity of chemical
structures found in polyketides .
PKS are divided into three classes (Staunton and
Weissman, 2001) . Type- I PKS (PKS-I ) are modular
enzymes usually containing multiple covalently linked
functional units (calleddomains) . PKS-I includes bac-
terial PKS-I ( BPKS-I ) and all fungal PKS ( FPKS) . A
typical BPKS-I is a large protein consisting of multiple
modules . Each module is composed of a set of doma-
ins, and each of the domains is typically used once
during the polyketide chain assembly . In contrast, a
typical FPKS only has a single module, and the doma-
ins within the module are iteratively used to catalyze
the polyketide chain formation . Type- II PKS (PKS- II )
consist of several physically discrete catalytic units that
can form a functional and repetitively used enzyme
complex . Type-III PKS ( PKS-III ) is a unique group .
Themost distinctive featureof PKS-III is theabsenceof
an acyl carrier protein (ACP) and any of theβ-keto-
processing domains normally found in other PKSs . In
essence, a PKS-III is an equivalent of theketosynthase
(KS) domain of PKS- I or PKS-II . This“domain”is
also iteratively used to synthesize products with various
chain lengths (Fig . 2) .
Fig . 2 The enzymeorganization and general chemistryof the different types of PKS . The FAS systems with similar organization and chem-
istry to the corresponding PKS typesare also indicated . Thered coloredmalonyl-moiety is themost common“building block”( bold bond)
for PKS-catalyzed reactions . Note that the SH group on thewaved line of the KS domains indicates the thiol group of active site cysteine,
whereas that of the ACP domains indicates the thiol group of the cofactor 4′-phosphopantetheine ( PPT)
1523 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
2 Bacterial Type I PKS (BPKS-I)
PKS-I are modular enzymes, including bacterial
type I PKS ( BPKS-I ) and all fungal PKS ( FPKS) .
BPKS- I are usually multi-modular enzymes, physically
residing either on the same protein or on several pro-
teins . FPKS are typically single-modular enzymes,
physically residing on the same protein . Each module
is composedof multipledomains that function atvarious
programmed steps during the carbon chain elongation .
A module minimally consists of three domains, β-keto-
synthase (KS) , acyltransferase (AT) , acyl carrier pro-
tein (ACP) . Many modules also contain the so-called
β-keto-processingdomains, β-ketoreductase (KR ) , de-
hydratase ( DH) , and enoylreductase ( ER ) ( Fig . 3)
(Hopwood, 1997) . A thioesterase (TE) domain is of-
ten located at the end of the last module . TE is also
found in some of FPKS, especially those synthesizing
fungal aromatic polyketides ( Cox, 2007 ) . TE is re-
sponsible for product release from the enzyme . In this
Section, only BPKS-I will be considered . FPKSwill be
described in Section-4 .
Throughout thebiosynthesis, the intermediates are
covalently attached to the ACP domain . In order to be
active, the ACP domain has to be modified with a 4′-
phosphopantetheinyl group ( PPT ) at its active site
serine residue . This is catalyzedby 4′-phosphopanteth-
einyl transferase ( PPTase) using Coenzyme A as the
substrate ( Lambalot et al. , 1996; Walsh et al. ,
1997) . During thebiosynthesis, the substrates (acetyl
or malonyl CoA ) as well as the growing polyketide
chains are covalently linked to PPT of ACP as acyl
thioesters . The 20?PPT serves as a long armswinging
the acyl chain fromone catalytic domain to the next to
assemble and modify the polyketide chain .
2. 1 The processive assembly line mode for BPKS-I
Among BPKS- I , 6-deoxyerythronolide B synthases
(DEBS) , the enzymes involved in antibiotic erythromy-
cin A biosynthesis in the Gram-positive bacterium Sac-
charopolyspora erythraea (Cortes et al. , 1990; Donadio
et al. , 1991 ) , are probably the best understood sys-
tem . The pioneeringwork of theLeadlay group and the
Katz group on DEBS has been themilestoneof PKS re-
search . DEBS, a 1 039 kDa giant enzyme complex, is
encoded by three eryA genes, eryAI (DEBS1 ) , eryAII
(DEBS2) and eryAIII (DEBS3) . DEBS1 and DEBS2
are separated by a 1.44 kb intergenic region, whereas
DEBS2 and DEBS3 are contiguous ( Caffrey et al. ,
1992; Donadio and Katz, 1992 ) . This huge protein
complex has a loading di-domain and six modules with
a total of 28 active domains . The co-linearity between
the domain organization in DEBS and the assembling
steps for 6-deoxyerythronolide B ( 6-DEB ) , the agly-
cone of erythromycins, is shown in Fig . 4A . The bio-
synthesis of 6-DEB is initiated by the attachment of a
propionate onto the ACP of loading didomain, which
then is transferred to the KS of module 1 ( KS1 ) . A
methylmalonate is loaded onto the ACP of module 1 ,
and the decarboxylation induced Claisen condensation
is catalyzed by the KS1 to form a diketide . Then the
diketide is transferred to thenextmodule for the second
elongation cycle . The same set of reactions is repeated
as the intermediate is passed forward onto each module
for elongation . When the 15-carbon linear chain with 7
methyl groups reaches the end, TE releases theproduct
by cyclizing it to a 14-memberedmacrolide . This mac-
rolide undergoes several modification steps, including
oxidations and glycosylations, to become the mature
product erythromycin A .
According to the co-linear model , each module
works relatively independently, with the growing acyl
chain transported fromonemoduleto thenext . Theva-
lidity of this model system has been proved by experi-
mental evidence . For example, the activesitemutation
of ER frommodule 4 ( Donadio et al. , 1993 ) and the
deletion of KR frommodule 5 (Donadio et al. , 1991)
led to the predicted 6-DEB analogs with proposed
structural changes ( Fig . 4B) . These earlier works also
opened theso-called rational biosynthetic pathway engi-
neering to produce“unnatural natural products”. In a
different approach, the TE domain was repositioned to
the end of varied modules to artificially release the na-
scent intermediates (Fig . 3A) . For example, module 1
and 2 with TE made a triketide ( Bohm et al. , 1998) ;
module 1 to 3 with TE made a tetraketide and module 1
to 5 with TE produced a hexaketide ( Kao et al. ,
1995) . The results also proved that the individual mod-
252 云 南 植 物 研 究 30 卷
ule and domain are used in the order suggested by the
genes . Similar studieson rifamycin BPKS-I that synthe-
sizes rifamycin B polyketide in Amycolatopsis mediterra-
nei further confirmed the co-linearity of BPKS- I modular
organization and product structure (Yu et al. , 1999) .
2 . 2 ?Inter-modular communications and structural
studies of BPKS-I
In order for the processive assembly line mode to
work , theremust be close intermodular communications
among the three enormous multi-modular DEBS (～350
kDa each) .Thesequence analysis has shown that there
are important linker regions between modules on differ-
ent proteins, as well as between domains within the sa-
memodule .These linkers, up to100 residues long, are
Fig . 3 The basic pathway for fatty acid and polyketide biosynthesis catalyzed by modular FAS and PKS respectively . The blue colored
moiety is derived from the starter unit; thered moieties are derived from the extension unit malonyl-CoA ; the red bold bonds indicate the
2-carbon building block that is derived frommalonyl-CoA . A to D represent the alternativeversions of the reductive cycle that lead to keto,
hydroxyl , enoyl or methylenefunctionality (modified from ( Hopwood, 1997) )
3523 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
Fig . 4 A . The assembly line mode for 6-DEB biosynthesis ( Hopwood, 1997; Khosla et al. , 2007) . The circles depict domains inDEBS and
the linker regions areomitted . The prematurepolyketides made by truncated PKS through artificial fusion of TE with different DEBS modules are
listed within the corresponding region: module 2 + TE?triketide; module 3 + TE?tetraketide; module 5 + TE?hexaketide . Thebluehydroxyl group
indicates the positionwhere the cyclization occurred; B . Thedomain mutation schemefor DEBS . The inactivation of ER in module4 leadsto the
production of a 6-DEB analogwith a residual double bond; and the deletion of KR in module 5 leads to the production of a 6-DEB analogwith
a keto group instead of the hydroxyl group . The structural change introduced by domain mutations is displayed in red color
452 云 南 植 物 研 究 30 卷
frequently rich in alanine, proline and charged resi-
dues . They are known to play a vital structural role in
the communications between modules and domains
(Gokhale et al. , 1999; Gokhale and Khosla, 2000;
Weissman et al. , 2006 ) . The mismatched linker se-
quences abolished intermodular chain transfer without
affecting the activity of individual modules, whereas
matched sequences can facilitate the channeling of in-
termediates between ordinarily nonconsecutive modules
(Okamoto et al. , 2001 ) . Some researchers have ref-
erred these linker regions as docking domains, which
arenot catalytic units but more like structural domains
involved in specific coiled-coil interactions that stabi-
lize the three-dimensional fold of proteins ( Broadhurst
et al. , 2003; Weissman, 2006) . A single amino acid
substitution inside the helix 3 region of docking domain
2 - 3 ( C-terminal docking domain of DEBS1 and N-
terminal docking domain of DEBS2) could significantly
alter the efficiency of polyketide biosynthesis ( Weiss-
man, 2006) .
Earlier studies using limited proteolysis and ac-
tive-site labeling have suggested that the individual
DEBS modules can forma homodimer and thereby gen-
erate two equivalent and independent clusters of active
sites for polyketidebiosynthesis (Aparicio et al. , 1996;
Kao et al. , 1996) . More recently, the 3-D structural
studies of BPKS-I havestarted to emerge . In spiteof the
complexity and huge size of BPKS-I , there have been
somesuccesses . In 2001, the first crystal structure of a
BPKS-I domain was reported ( Tsai et al. , 2001 ) . It
was the 66 kDa TE fromDEBS with a 2 .8?resolution .
Later, the samegroup reported the 1 .7? crystal struc-
tureof the 51 kDa KR withbound NADPH frommodule
1 of DEBS ( Keatinge-Clay and Stroud, 2006 ) . In
2003 , Weissman and colleagues reported the NMR so-
lution structureof the 120-residue docking domain 2 -
3 , which functions in the interaction between DEBS2
and DEBS3 (Broadhurst et al. , 2003 ) . In 2006 , the
Ban group reported the 4 .5? structure for the entire
270 kDamammalian FAS (Maier et al. , 2006) , which
was a remarkable breakthrough . The Khosla group
solved the 2 .7? crystal structure of the 194 kDa ho-
modimeric fragment from module 5 of DEBS (Tang et
al. , 2006 ) . Based on the available structures and dou-
ble helical homodimer model , Sherman proposed a
three-dimensional model for BPKS-I ( Sherman and
Smith, 2006 ) . The postulated structure contains two
successivemodules with a complete set of domains, a
TE domain at the end, and docking domains between
themodular interfaces . Theoverall modular homodimer
structure including KS, AT, DH, ER and KR is de-
rived fromthe FAS-I structure; the special regions like
KR, KS-AT and TE are modified based on the corre-
sponding structures from DEBS; the ACP conformation
is consulted fromtheACP solution structureof the acti-
norhodin synthase; the conformationof the docking do-
mains are derived fromthedock 2 - 3 mode structureof
DEBS . Although an intact modular structureof BPKS-I
is still not available, theseworks have provided a rela-
tively clear picturefor thestructureof themegasynthas-
es as well as mechanistic insights into the biosynthetic
mechanismof BPKS- I .
2 . 3 Engineered biosynthesis using BPKS-I
So far, more than 200 new polyketides have been
synthesized through engineereingof BPKS-I ( Weissman
and Leadlay, 2005 ) . There are three main prerequi-
sites for engineeredpolyketidebiosynthesis (Hershberg-
er, 1996) . The first is the availability of cloned genes
for a secondary metabolic pathway fromdifferent organ-
isms; the second is a genetically defined host strain
able to support the production of polyketides; and the
third is the ability to modify and recombine specific
genes from different pathways using recombinant DNA
technology . DEBS is the most extensively enginnered
BPKS- I for biosynthesis; it has been tackled through
domain manipulation, module swapping, precursor di-
rected biosynthesis, and heterologous expression .
2 . 3 .1 Heterologous expression systems for BPKS-I
a . The E. coli expression system
E. coli normally doesn′t produce polyketides,
which is advantageous as this provides a clean back-
ground for heterologous products . In order to produce
active BPKS-I in E. coli , the first requirement is to be
able to convert the apo-ACP into the active holo-ACP
by adding the 4′-phosphopantetheine ( PPT ) group
(Lambalot et al. , 1996) . Although E . coli has its own
5523 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
PPTases (ACPS and EntD) , they are specific for the
activation of E . coli FAS- II and enterobactin syn-
thetase . The heterologous ACPs can′t beproperly modi-
fied and aregenerally inactive .Therefore, a“promiscu-
ous”PPTase, such as Sfp from Bacillus subtilis, which
is able to convert almost any apo-ACP to its holo form,
has to be co-expressed with BPKS-I in E. coli (Quadri
et al. , 1998 ) . In addition, there are several other
problems with the production of complex polyketides in
E. coli, such as the availability of the precusors and
building blocks for polyketide biosynthesis .
A successful example was carried out by the Kho-
slagroup in 2001 ( Pfeifer et al. , 2001 ) . They inte-
grated sfpgene from B. subtilis into the prp operon in
E. coli and deleted the prpRBCD genes that are puta-
tively responsible for propionate catabolism . Only the
prpE gene which encodes the putative propionyl-CoA
synthase was conserved . Both sfp and prpE wereunder
control of a T7 RNA polymerase promoter inducible by
isopropyl-β-D-1-thiogalactopyranoside ( IPTG ) . Two
plasmids, pTR132 and pYC216 , were also introduced
to producemethylmanoyl-CoA . Thesegenes were coex-
pressed with DEBS genes in E. coli BL21 ( DE3 )
named BAP1 . With the feeding of propionate, BAP1
produced 6-DEB at a level even higher than the wild
type S. erythraea and comparable with the industrial
strain that overproduces erythromycin . This example
demonstrated the power of the E . coli systemfor engi-
neered PKS biosynthesis .
b . The Streptomyces coelicolor CH999 expression system
The strain CH999 was a genetically engineered
host derived from Streptomyces coelicolor (McDaniel et
al. , 1993 ) . It is ideal for production of polyketides
that areoriginally produced by a Streptomyces species .
CH999 wasgeneratedby deleting thewhole actinorhod-
inbiosynthetic gene cluster ( act) through homologous
recombination . A shuttle vector called pRM5 was de-
signed to express thePKS genes in CH999 . This vector
contains all necessary genetic elements for the expres-
sion of the PKS genes . CH999 was initially used to ex-
plore the biosynthesis of actinorhodin and the roles of
PKS-II system ( McDaniel et al. , 1993 , 1993; Mc-
Daniel et al. , 1994; McDaniel et al. , 1994 , 1995 ) .
Subsequently, becauseof its versatility and compatibili-
ty, CH999 was exploited as ageneral expression system
and has been successfully applied in the studies of bio-
synthesis of erythromycin (Kao et al. , 1994; Pieper et
al. , 1995) , epithilone (Tang et al. , 2000) and 6-meth-
ylsalicylic acid (6-MSA) (Bedford et al. , 1995) .
c . Other expression systems
A S. erythraea strain called JC2 was genetically
engineered to removeall of theBPKS-I genes except for
the chain-terminating TE region ( Rowe et al. , 1998) .
Since S. erythraea cannot transcribegenes that are in-
troduced on autonomous plasmids, any new gene to be
expressed must be integrated into thegenome . An inte-
grativevector pCJR24 was developed for use with JC2 .
The Leadlay group reintroduced the entire DEBS gene
set into JC2 and found that the strain produced a large
amount of erythromycins andprecursor macrolides simi-
lar to the wild type ( Rowe et al. , 1998 ) . This pro-
vides another way to increase theyieldof erythromycins
or its engineered analogs . Another eukaryotic expres-
sion system is Saccharomyces cerevisiae . 6-MSA syn-
thase ( 6-MSAS ) has been expressed in this host
(Kealey et al. , 1998; Wattanachaisaereekul et al. ,
2007) . The gene was cloned into a yeast expression
vector under the control of alcohol dehydrogenase 2
(ADH2 ) promoter and terminator, and the sfp gene
was also cloned and coexpressed with 6-MSAS in yeast
to produce 6-MSA .
It should benoted that although theseheterologous
expression systems are discussed under the BPKS-I
section, most of these systems have also been used for
other PKS systems, which will be described in follow-
ing sections . The only exception is the JC2 system,
which is limited to DEBS-like BPKS-I and can not be
applied for other PKSs .
2 . 3 .2 Genetic engineering strategies for BPKS-I
The co-linear mode of action of BPKS-I offers an
obvious advantage for manipulating the domains and
modules . A number of genetic engineering strategies
have been exploited for the production of new
polyketides . The strategies include deletion, substitu-
tion, addition and reposition of domains, modules, or
even entire subunits .
652 云 南 植 物 研 究 30 卷
a . Domain manipulation
It has already been shown in the DEBS systemthat
the inactivationor deletionof theβ-keto processingdoma-
ins (KR, DH and ER) in DEBS can arrest polyketides
at different reduction levels ( Donadio et al. , 1991;
Donadio et al. , 1993) . The substitution of similar do-
mains with different properties is a common strategy to
expand the functionof BPKS- I . For example, the load-
ing domain of spiramycin synthase from S. ambo-
faciens, which uses acetate as the starter unit, was
substituted by the counterpart from tylosin synthase
from S. fradiae, which uses propionate as the starter
unit (Kuhstoss et al. , 1996; Gandecha et al. , 1997) .
The resulting hybrid PKS produced the spiramycin analog
usingpropionateinsteadof natural acetatetostart thebio-
synthesis . Similar work has been done in the DEBS sys-
tem (Oliynyk et al. , 1996; Ruan et al., 1997; Stassi et
al., 1998) . Furthermore, the strategy of addition and re-
positionof domains, notably theTE domain, has also led
to theproductionof newproducts andbetter understanding
of the biosynthetic mechanism (Kao et al. , 1995; Bohm
et al., 1998) . However, it shouldbepointedout that not
every domain can be manipulated . It is clear that many
such manipulations lead to incompatibility between the
domains of different origins and conflictof substratespec-
ificities due to the unnatural substrates?intermediates . In
addition, the boundaries between domains are difficult
to define due to the lack of high resolution structures
for a complete BPKS- I .
b . Module and subunit manipulation
The strategyof manipulating intactmodules (Mars-
den et al. , 1998 ) and even entire subunits would help
alleviate some of the problems associated with domain
manipulation . Bimodular and trimodular BPKS-Is were
constructed by mixing the appropriate modules from
DEBS and rapamycin synthase from Streptomyces hygro-
scopicus, and thehybrid enzymewhich preserved an in-
tact ACP-KS didomain that spans the junction between
successivemodules was able to synthesize the predicted
lactones ( Ranganathan et al. , 1999) .The control muta-
nts without thedidomain failed to produce a polyketide .
The similar modular manipulations were carried out be-
tween DEBS and rifamycin synthase from Amycolatopsis
mediterranei ( August et al. , 1998 ) . As mentioned
above, the linkers and docking domains play a key role
in inter-modular and subunit communications ( Gokhale
et al. , 1999; Weissman, 2006) . Therearetwo kinds of
factors to connect modules, intermodular linkers for the
modules in the same subunit and docking domains for
thesubunit interfaces . Once the communication between
the fusedmodules is established, thehybridBPKS-I can
be harnessed to make expected polyketides . However,
the principlebehind these linkers and docking domains
is not totally clear and their compatibilitywith incoming
modules is not always predictable .
c . Precursor directed mutational biosynthesis
This strategy combines the power of chemical syn-
thesis with engineered biosynthesis and has shown the
ability to markedly increase the diversity of new
polyketides . By supplementingsynthetic precursors that
are analogous to natural substrates to PKS mutants in-
activated in an important step of polyketide pathway,
new polyketides with desired structural features could
be produced . This technique was first applied in the
soil bacterium Streptomyces avermitilis, the producer of
avermectin, which is a commercially important antipar-
asitic macrolide used as human andveterinary medicine
(MacNeil et al. , 1992) . To generate novel avermectin
analogs, carboxylic acids were supplemented to an
avermectin null-mutant with inactivated branched-chain
α-keto acid dehydrogenase ( BKD) . Without BKD, the
2-oxo acid cannot beconverted to isobutyric or methyl-
butyric acids that are the starter units for avermectin .
The usage of precursor directed mutational biosynthesis
in this null mutant produced 36 novel avermectin ana-
logs from 800 substrates (Dutton et al. , 1991 ) . This
strategy was further extended to improve the production
of the antiparasitic agent doramectin, an avermectin
analogmade by a S. avermitilis (Δbkd) mutant in the
presence of the cyclohexanecarboxylic acid ( CHC )
( Cropp et al. , 2000 ) . In the DEBS model system,
DEBS1 exhibits the flexibility to use various starter
units such as acetyl and butyryl CoA , or corresponding
N-acetylcysteamine (SNAC) thioesters . This flexibility
has been exploited in precursor directed biosynthesis
(Kinoshita et al. , 2001) . A mutant with inactivated KS
7523 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
Fig . 5 Combinatorial biosynthesis of polyketides . Theproduced polyketidebackboneportionsare indicatedwith thesame color corresponding to the enzymes
( Hopwood, 1997; Staunton and Weissman, 2001; Fischbach andWalsh, 2006 ) . A . Modular engineering of DEBS (black) , rapamycinPKS ( red) , and ri-
famycin PKS ( blue) . Solid ellipse is the natural linker between DEBS module 1 and 2 . ( 1) bimodular PKS of DEBS module 1 and 3 withmodule1 linker .
(2 ) bimodular PKS of DEBS module 1 and rapamycin PKS module 12 with preserved ACP1-KS2 intermodular region . (3) bimodular PKS of DEBS module
1 and rifamycin PKSmodule 5 withmodule 1 linker; B . Hybrid BPKS-Is with combined subunits from pikromycin PKS ( green) andDEBS (black) . The al-
tered version of DEBS3 withDH and ER domains fromrapamycinPKS (red) was also used; C . Precursor directedmutational biosynthesisof 6-DEB analogs
using DEBS-KS1°mutant . The in vivo incorporation of synthetic substrates ( orange) resulted in the production of novel 6-DEB like macrolides
852 云 南 植 物 研 究 30 卷
in module 1 (DEBS-KS1°) was constructed to disable
the first chain extension and stop the6-DEB biosynthe-
sis . Chemically synthesized non-natural diketide or
triketide SNAC-thioesters were supplied and recognized
by downstream modules of DEBS to restore the
polyketide chain assembly . Further tailoring of these
aglycons by erythromycin post-DEBS enzymes resulted
in thegeneration of novel erythromycin derivatives with
biological activities comparableto their natural counter-
parts . In spite of the success, this strategy also has
limitations . For example, the synthetic substrates must
beefficiently transported across thecells andshould not
be toxic to the host .
3 Type II PKS
While BPKS- I are mega-enzymes with a multiple
modular composition similar to animal FAS-I , BPKS- II
are relatively small and comprised of several individual
enzymes similar to bacterial and plant FAS-II . BPKS-I
typically synthesize macrolides such as erythromycins,
whereas BPKS- II usually synthesize aromatic polyketides .
Each of the physically separate enzymes (subunits) that
makeup a BPKS-II complex works iteratively during the
polyketide chain assembly .The S. coelicolor CH999 ex-
pression system that has been described above was ini-
tially used tostudy BPKS- II (McDaniel et al. , 1993) .
So far, BPKS- II areonly found in bacteria, such
as the soil-borne and marinegram-positive actinomyce-
tes (Staunton and Weissman, 2001; Hertweck et al. ,
2007) . Polyketides produced by BPKS-II are classified
based on the polyphenolic ring cyclization patterns and
biosynthetic pathways . Table 1 lists the common
groups, including anthracyclines, such as the impor-
tant antitumor drug daunorubicin from S. peucetius,
angucyclines, such as the antitumor agent landomycin
from S. cyanogenus S136 , aureolic acids, such as the
anticancer drug chromomycin from S. griseus, tetracy-
clines (such as the broad-spectrum antibiotics tetracy-
cline from S. aureofaciens, polyphenols ( such as the
broad-spectrumantifungal agent pradimicin from Actino-
madura hibisca, and benzoisochromanequinones ( such
as the antibiotic actinorhodin from S. coelicolor) .
A minimal BPKS-II consists of two ketosynthase
units KSα , KSβ ( also known as chain-length factor,
CLF ) and ACP . KSα and KSβ normally form a het-
erodimer (KS-heterodimer) mimic to theKS homodimer
in bacterial FAS-II ( Pan et al. , 2002) . Other catalyt-
ic subunits, such as KR , CYC, and ARO ( aro-
matase) , as well as the post-PKS tailoring enzymes,
such as oxygenases, glycosyl and methyl transferases,
may bepresentedvariably in theBPKS-II (Hertweck et
al. , 2007) . Theminimal PKS is responsible for the it-
Table 1 Polyketides derived from PKS-I I catalyzed biosynthesis ( Hert-
weck et al. , 2007) . Structures, names, activities, and originated species
of typical compounds of each group are presented . Only the aglycone
structures of these compounds are elucidated, and the R group indicates
the sugar side chain in the metaboliteoverall structure
9523 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
erative decarboxylative condensation of malonyl-CoA
and extends the polyketide chain . When theintermedi-
ate polyketide chain reaches a certain length, KR ,
CYC, and ARO are believed to define the folding pat-
tern andoxidation level of the nascent polyketide chain
and release the final products through cyclization
(Meurer et al. , 1997) .
3 . 1 Biosynthetic mechanism of BPKS-II
3 .1 .1 Initiation of biosynthesis
The actinorhodin PKS is probably the best studied
system in PKS- II (Fig . 6) . The biosynthesis is initat-
ed by priming theminimal PKS (Fig . 7A) . It is gen-
erally accepted that most BPKS-II areprimedwith ace-
tate (Moore and Hertweck, 2002 ) . KSα and KSβ are
highly similar except that KSβ has aglutamine residue,
instead of a cysteine, at the highly conserved active
siteof KS . This subtle change leads to the inability of
KSβ to catalyze Claisen type C-C bond formation ( Bi-
sang et al. , 1999) . The presenceof glutamine instead
of cysteine in KSβ might render it amalonyl-CoA decar-
boxylase to initiatetheBPKS-II priming . More interest-
ingly, the condensation deficient KSQ as a KSβ analog
has also been found in the loading module of some
BPKS- I , such as pikromycin PKS-I , and provides an
acetate starter unit by decarboxylation of malonyl-CoA
(Xue et al. , 1998) .
Although acetate is themost common starter unit,
there are numerous BPKS-II that employ alternative
starter units . Two alternative pathways have been prop-
osed for non-acetate PKS- II priming . The first is KSIII
priming, which is derived fromsimilar initiationof fatty
acid biosyntheses catalyzedby dissociated FAS- II found
in E. coli and Streptomycetes ( Han et al. , 1998 ) .
Several BPKS-II systems are known to utilize the alter-
nativepriming (Bao et al. , 1999; Tang et al. , 2003) .
The other pathway is through direct loadingof substrate
by acyl-CoA ligase . This pathway is used in enterocin
from S. maritimus ( Piel et al. , 2000 ) , in which an
acyl-CoA ligase catalyzes the formation of benzoyl-CoA
starter unit and loading onto the enterocin BPKS- II
(Xiang and Moore, 2003 ) . These alternative pathways
lead to a functional cross-talk between fatty acid and
polyketidemetabolisms .
3 . 1 .2 Substrate loading in BPKS-II
Malonyl-CoA is usually the substrate for chain
elongation catalyzedby BPKS-II (Fig . 7A) . The load-
ing of substrate to ACP requires malonyl-CoA∶ACP
transferase ( MAT) . Interestingly, the MAT gene is
absent in most BPKS-II gene clusters, except in the
cluster for daunorubicin ( DpsD) (Ye et al. , 1994 ) ,
R1128 complex ( ZhuC ) ( Meadows and Khosla,
2001) , and enterocin ( EncL ) ( Xiang and Moore,
2003) . There had been a long-going debate regarding
the involvement of MAT as a potential fourth component
of minimal PKS in the BPKS-II catalyzed biosynthesis .
One opinion is that an endogenous MAT is recruited
from a fatty acid biosynthetic pathway to act with min-
imal PKS, such as in tetracenomycin ( TCM ) PKS- II
from S. glaucescens ( Bao et al. , 1998 ) . Some in vitro
studies also supported this view ( Carreras and Khosla,
1998; Dreier et al. , 1999 ) . On theother hand, there
was evidence tosuggest that theACP of BPKS- II is ca-
pable of self-malonylation and a MAT from the FAS
system is not necessary (Hitchman et al. , 1998; Zhou
et al. , 1999) . Both the TCM-ACP and MAT seem to
contain the same acyl-transferase activity . The differ-
ence is that the catalytic efficiency of MAT is much
higher than the one of TCM-ACP, which explains the
observation that MAT-independent polyketide biosyn-
thesis only proceeds at high concentrations of a PKS- II-
ACP (Zhou et al. , 1999) .Thereseems to beno abso-
lute requirement of MAT in the in vitro polyketide bio-
synthesis ( Matharu et al. , 1998; Arthur et al. ,
2006 ) . Furthermore, chemically synthesized acti-
norhodin (ACT) ACP which excludes any contamina-
tion of trace MAT also exhibited the self-malonylation
activity (Arthur et al. , 2005) . This result convincing-
ly proved that self-malonylation is an inherent activity
of PKS-II-ACP . However, the relevant roles of MAT
and PKS- II-ACP self-malonylation in polyketide bio-
synthesis still remain unknown in vivo .
3 . 1 .3 Chain elongation in BPKS-II
Following starter unit priming of KSα and initial
acylation of ACP, the BPKS-II catalyzed biosynthesis
proceeds to the iterative chain elongation (Fig . 7B) .
062 云 南 植 物 研 究 30 卷
Fig . 6 Proposed pathway for the biosynthesis of actinorhodin and isolated polyketides from defectivepathways lacking other BPKS-II components .
The different cyclization pattern in related pathway has also been indicated; the bold bonds represent the carbon unit derived from manoyl-CoA .
The organization of act PKS-II gene cluster is listed, minimal PKS is shown in red with other enzymes in blue ( adapted from Hopwood, 1997)
1623 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
The mechanismfor chain elongation is not totally clear
in spite of the extensive studies . The studies of ACT
minimal PKS in priming, elongation, and termination
of theoctaketide biosynthesis have proposed that C169
and H346 form a catalytic dyad for acyl chain attach-
ment (Dreier and Khosla, 2000 ) . In addition, H309
locates themalonyl-ACP in the active site and supports
carbanion formation by interacting with the thioester
carbonyl , whereas K341 enhances the rate of malonyl-
ACP decarboxylation via electrostatic interaction . The
catalytic cycles for chain elongation appear to involve
thedynamic KSα?KSβ?ACP ternary complex with repeat-
ed assemblyanddisassembly . TheACP and theKS-het-
erodimer dissociate after each C-C bond formation, and
thenewly extended acyl chain is transferred back from
the ACP pantetheine armto the KSα cysteine active site
prior to the dissociation of the ternary complex .
3 . 1 .4 Chain length determination in BPKS- II
The chain length determination mechanism is
probably themost extensively studied aspect inBPKS- II
(Fig . 7C and D) . Many studies were carried out in
the S. coelicolor CH999 system . Initially, ACP was
the target in many studies . When actinorhodin (ACT,
C16) ACP was substituted with tetracenomycin (TCM ,
C20) ACP or frenolicin ( FRE , C18 ) ACP, the enz-
yme complex remained active and generated aromatic
polyketides with the same chain length as actinorhodin
or shunt products produced by mutants defective in the
actinorhodin pathway ( Khosla et al. , 1993 ) . These
results indicated that ACP is probably not a key factor
controlling chain length . Furthermore, the solution
structure of ACT-apo-ACP also excluded the possibility
that ACP can directly control the size of final product
( Crump et al. , 1997) . Instead, ACP may help stabi-
lize the growing polyketide chain to avoid self-cycliza-
tion becauseof thehighly active unreduced keto group .
In the subsequent studies, the KS subunits were
the targets . ACT-KS mutants were complemented with
other KS, and the results suggested that KSα may be
relatively generalized in function, whereas the KSβ may
provide specificity in polyketide chain construction
(Kim et al. , 1994 ) . Further engineered biosyntheses
of novel polyketides in CH999 system with different
combination of BPKS-II components fromACT, GRA ,
and TCM PKS-II systems revealed that the product
chain length is at least partially dictated by KSβ (Mc-
Daniel et al. , 1993 ) . KSβ was thus termed as chain
length factor ( CLF ) . Later experiments showed that
the size of the final polyketides was mainly determined
by CLF , but also required the contributionof KSα . The
minimal PKS sets with the sole change in CLF abol-
ished the biosynthesis of polyketides . TheKS-CLF het-
erodimer together determine the chain length . Re-
cently , the determination of the crystal structure of
ACT KS-CLF heterodimer has provided mechanistic in-
sights into the polyketide chain control ( Keatinge-Clay
et al. , 2004 ) . The structure revealed highly comple-
mentary contacts of the two subunits, which probably
explains why most heterologous KS and CLF are unable
to compose a functional dimer and make an active hy-
brid BPKS- II system . A trapped intermediate in the
heterodimer also revealed a polyketide binding pocket .
Overall , the catalytic site Cys169 in the KS monomer
and the gating residue Phe116 in the CLF monomer
mark the beginning and end of the 17-? polyketide
tunnel . It is apparent that the keto groups of growing
polyketide chain probably contact with the KS-CLF
tunnel as enols, so the activeketo groups are separated
to prevent the spontaneous cyclization due to the in-
tramolecular reactions . The sequence comparisonof the
known KS-CLF enzymes around helix-turn-loop region
identified several residues in the tunnel as gates to con-
trol the length of the extending chain . Two phenylala-
nines (F109?F116) and one threonine (T112 ) were as-
sumed to be the main factors to decide the cavity of the
tunnel (Pan et al. , 2002; Tang et al. , 2003; Keatinge-
Clay et al. , 2004) . For KS-CLF heterodimer associated
with longer chain lengths (from16 to24 carbons) , these
residues are replaced with less bulky amino acids .The-
oretically, the sizes of the gating residues control the
cavity of the tunnel , thus the chain length of the
polyketide is also regulated ( Keatinge-Clay et al. ,
2004) . This structural hypothesis is supported by site-
directed mutagenesis of these key residues in different
PKS-IIs . The octaketide synthase ACT PKS-II ( C16)
could synthesize decaketide ( C20 ) through the double
262 云 南 植 物 研 究 30 卷
Fig . 7 Proposed biosynthesis of aromatic polyketides ( using ACT PKS-II as model system) . The hollow sphere indicates CLF; the black solid
sphere indicates KS with active cysteine; the hollow oval indicates ACP with PPT-arm ( Hertweck et al. , 2007) . A . PKS-I I primingmediated by
decarboxylation; B . PKS-II elongation . After each cycle, thepolyketide chain increases two carbon units with aketo group; C . PKS-I I termina-
tion . The formed aromatic polyketides still needs to befurther modified by post-PKS enzymes to yield the final products; D . PKS-II reaction mod-
el . The circled red dash line represents the cell . ACP with attached growing polyketide chain is accommodated in the binding channel of KS-CLF
heterodimer . The exact timing for thebinding of KR , ARO and CYC is not clear yet, but the interaction of theseproteinswithKS-CLF will modify
thebinding pocket and alter the structureof final product . The ninth carbon that is normally reduced by KR is labeled in blue
3623 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
mutation F116A?F109A in ACT-CLF , probably be-
cause these changes open the polyketide tunnel so that
intermediate chain can be extended further as in the
decaketide synthaseTCM PKS-II (C20) (Tang et al. ,
2003) .TheTCM-CLF could also be engineered intoan
enzyme capable of synthesizing both octaketides and
decaketides through thesinglemutationM120T (equals
to T112 in ACT-CLF) , which would constrict the tun-
nel and cause a growing polyketide to terminate at the
octaketide stage . However, the different results ob-
tained fromthe in vitro and invivo tests of themutants
indicated that the interaction between KS-CLF het-
erodimer and additional proteins like KR and CYC or
ARO may reshape the binding pocket to expose the ac-
tive keto groups in polyketide chain and initiate the cy-
clization to release the final products . In conclusion, it
seems that the chain length is determined by the entire
KS-CLF complex, with the binding pocket as the main
factor regulating the sizeof polyketides ( Fig . 7D) .
3 . 1 .5 Cyclization in PKS-II
How BPKS-II controls the folding pattern of the
nascent polyketide chain is another important question
(Fig . 7C and D) . With the same polyketide chain,
different cyclizations can result in production of differ-
ent polyketides . As discussed above, minimal PKS ap-
pears to have the necessary elements required for con-
trolling the chain length . However, minimal PKS by it-
self is unable toaccurately control the cyclization . Enz-
yme subunits, such as KR , ARO or CYC , are impor-
tant for the folding and cyclization process . KR is the
first enzyme to alter the structure of the nascent
polyketide chain prior to cyclization . Theproductionof
incorrectly foldedoctaketides SEK4 and SEK4b in ACT
PKS-II mutants with deficient KR confirmed its impor-
tance in polyketidestructures (Fu et al. , 1994) . CYC
functions in a chaperone-likemanner to help guide the
polyketide intermediates into the binding channel . It
should be pointed out that the cyclization of polyketide
is not just simply controlled by one or two enzymes,
but most likely determined by the whole BPKS-II com-
plex, including the KS-CLF heterodimer (Fig . 7D) .
3 . 2 ?Genetic engineering and combinatory biosyn-
thesis using BPKS-II
The heterologous expression systemCH999 is the
most widely used tool in BPKS- II genetic studies, in-
cluding domain inactivation, site-directed mutagenesis
and construction of hybrid enzymes . Numerous novel
aromatic polyketides have been produced by using this
system, including a few examples shown in Fig . 6 .
However, theyields of most new products arevery low .
The products are often with highly complex chemical
structures, which make it very complicated to separate
the desired polyketides . Becauseof the iterativenature
of the system, the production of metabolites in geneti-
cally engineered BPKS- II systems is not well con-
trolled . This leads to many shunt products that are not
specifically designed .
In Summary, BPKS-II represents another archi-
tecturally distinct group of PKS . Thephysical dissocia-
tion of subunit enzymes makes it relatively simple to
analyze, but the iterative nature complicates the mech-
anistic investigation . With earlier genetic studies and
more recent structural data, we now have a better pic-
ture regardinghow BPKS-II control product structure .
4 Fungal PKS (FPKS)
A large number of fungal secondary metabolites
arederived frompolyketides . Someof themare harmful
mycotoxins, such as zearalenone from Fusarium gra-
minearum ( Ciegler, 1975 ) , fumonisins from F . ver-
ticillioides ( Gibberella fujikuroi ) ( Bezuidenhout et al. ,
1988) , aflatoxins from Aspergillus parasiticus ( Nesbitt
et al. , 1962 ) , and T-toxins from Cochliobolus het-
erostrophus (Yang et al. , 1996) . Others are beneficial
metabolites, such as the HIV-1 integrase inhibitor eq-
uisetin from F . heterosporum (Sims et al. , 2005) , the
potent inhibitor for mammalian squalene synthase
squalestatin from Phoma sp . (Cox et al. , 2004) , and
the important cholesterol-lowering drug lovastatin?com-
pactin (MervacorTM ) from A. terreus (Shiao and Don,
1987; Kennedy et al. , 1999) .These fungal polyketides
play important roles in our daily life and represent a
rich resourceof natural products .
4 . 1 Classification of FPKS
Fungal polyketides are biosynthesized by modular
PKS, which resembles bacterial type I PKS . However,
462 云 南 植 物 研 究 30 卷
unlike BPKS-1 , FPKS typically only have a single
module that is iteratively used during the polyketide
chain assembly ( Fig . 8 ) . Therefore, the architecture
of FPKS is similar toBPKS-1 but the function is similar
to BPKS-II . FPKS can be subdivided into threegroups
based on the reduction level of the polyketide chain
( Bingle et al. , 1999; Nicholson et al. , 2001; Kroken
et al. , 2003 ) . The first group is non-reducing PKS
(NR-FPKS) . NR-FPKS synthesize aromatic polycyclic
compounds, such as tetrahydroxy naphthalene synthase
(THNS) in Colletotrichum lagenarium ( Fujii et al. ,
1999 ) and Wangiella dermatitidis ( Feng et al. ,
2001 ) . They have characteristic domain architecture
that has recently been redefined ( Crawford et al. ,
2006) . The domains consist of SAT (starter unit ACP
transacylase) , KS (β-ketoacylsynthase) , AT ( acyl
transferase) , PT ( product template) , ACP ( acyl carri-
er protein) , and TE?CLC ( thioesterase?Claisen-like
cyclase) . The SAT domain appears to be involved in
chain initiation control , while the PT domain to be in-
volved in chain length control ( Cox, 2007 ) . TE?CLC
is responsible for the chain cyclic-release and probably
also chain-length determination ( Fujii et al. , 2001;
Watanabe and Ebizuka, 2004 ) .
The second group is the partially-reducing PKS
(PR-FPKS) . PR-FPKS synthesize cyclic compounds
like 6-methylsalicylic acid ( 6-MSA ) . The compound
has an aromatic ringbut results fromapartially reduced
intermediate ( Beck et al. , 1990; Fujii et al. , 1996) .
6-MSA synthasehas domain architectureof KS-AT-DH-
Core-KR-ACP . It has twoβ-keto processing domains,
KR (β-ketoacyl reductase) and DH ( dehydratase) ,
which are selectively used during the biosynthesis to
produce“partially”reduced intermediates . The Core
domain is probably involved in subunit-subunit interac-
tion (Moriguchi et al. , 2006 ) .
The thirdgroup is the highly-reducing PKS (HR-
FPKS) . HR-FPKS synthesize non-aromatic compounds
like fumonisins, AAL-toxins, and lovastatin . The do-
main architecture resembles a typical module of BPKS-
I , consisting of KS-AT-DH-MT-ER-KR-ACP . Nota-
bly, it does not have SAT, PT , or TE?CLC, but con-
tains acomplete set ofβ-keto processing domains, KR ,
DH, ER ( enoylreductase) , and a methyltransferase
(MT) domain ( Cox, 2007) . The MT domain is dedi-
cated to catalyze the C-methylation using S-adenosyl
methionine (SAM) as substrate . This is very different
fromBPKS- I , which usually usemethylmalonyl-CoA as
substrate to introduce the C-methyl groups . Among the
threegroups, themechanismfor HR-PKS to control the
product structure is least understood . HR-FPKS with
identical domain architecturecan synthesize polyketides
with very different chain-lengths and reduction level
(Cox, 2007 ) . In addition, it is not clear how HR-
FPKS terminate and offload the linear polyketide
chain .
4 . 2 iGenetic manipulation and heterologous ex-
pression of FPKS genes
4 .2 .1 Fungal transformation
The development of convenient and efficient trans-
formation techniques is essential for genetically manipu-
lating FPKS genes for functional analysis as well as en-
gineered biosynthesis . Themost commonly used method
is the protoplast transformation . This method requires
cell wall degradation of germinated mycelia by lytic
enzymes and foreign DNA penetration through cell
membranes with the help of polyethylene glycol (PEG)
and CaCl2 ( Proctor et al. , 1999 ) . The method
requiresgoodquality protoplasts, which for somefungal
species are difficult to obtain . In addition, the trans-
formation efficiency is generally low, and there is a
high rate of undesired randomintegration and multiple
plasmid insertions, which complicate the subsequent
mutant analysis .
Another transformation method is restriction enz-
yme-mediated integration ( REMI ) (Lu et al. , 1994 ) .
The involvement of restriction enzymes can cause nicks
in fungal chromosome and facilitate the recognition of
the insertion sites . The transformation efficiency can
increase several folds compared with conventional pro-
toplast transformation . A thirdmethod is transposondi-
rected protoplast transformation ( Kempken and Kuck,
1998) . Transposons are ubiquitous genetic elements
that can move to new locations within their host ge-
nome . Thegeneinactivation caused by transposon inte-
gration is called gene tagging . Such generated new
5623 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
Fig . 8 Biosynthesis of typical fungal aromatic and reduced polyketides . The solid bold bond indicates the carbon unit derived from initial starter
unit acetate ( blue) or elongation unitmalonate ( red) . Thehashed bond indicatesthemethyl group derived from S-adenosylmethionine ( SAM ) . The
representative FPKS-I modular structures for 6-MSAS, LNKS and LDKS are shown; the cross mark indicates the inactivation of enzyme . The differ-
ent domains selected in each extended cycle are also pointed out . A . Proposed biosynthesis for 6-methylsalicylic acid ( Staunton and Weissman,
2001) ; B . Proposed pathway for production of lovastatin ( Kennedy et al. , 1999 ) . The two portions of lovastatin derived from independent FPKS-
I are presented separately: monacolin J from LNKS ( including the post-polyketide modification steps) and 2-methylbutyryl side chain fromLDKS
662 云 南 植 物 研 究 30 卷
phenotypes allow the recognition of disrupted target
gene, which can be identified by isolation of the trans-
poson with flanking regions fromthemutant genome .
A number of newer transformation methods have
also been tested . One is using electroporation, which
was successfully used in the transformation of Aspergil-
lus niger (Ozeki et al. , 1994) and inserted mutagene-
sis of Aspergillus fumigatus ( Brown et al. , 1998 ) .
This method does not require the preparation of proto-
plasts and is effective in nearly all cell types . Howev-
er, excessive exposureof live cells to electric field can
cause cell apoptosis and ion imbalances, which may
later lead to improper cell function and cell death, and
the percentage of ectopic integrations is significantly
higher than the protoplast method .
Another method is Agrobacterium-mediated trans-
formation . This method has been used for plant trans-
formation for decades, but is relatively new for fungal
transformation . It takes advantageof the natural patho-
genic activity of the soil-borne A. tumefaciens, which
is well known for its ability to transfer DNA into both
plants and fungi ( de Groot et al. , 1998) . This bacte-
rial pathogen causes tumorous growth on infected hosts
by inherent tumor inducing ( Ti ) plasmid directed in-
fection, which can insert specific transferred DNA ( T-
DNA) into the genome of host cells . Mutant libraries
have been constructed using this method (Michielse et
al. , 2005; Tucker and Orbach, 2007 ) . The results
indicated that Agrobacterium-mediated transformation
has higher throughput and homologous recombination
frequencies compared with the protoplast method .
In spiteof the availability of themany transforma-
tionmethods, the efficiency for fungal transformation is
generally low . This has been a major obstacle for suc-
cessful genetic manipulations . In addition, the rate of
homologous recombination is normally very low in fila-
mentous fungi . The target DNA can insert into the host
genome in a random manner, instead of integration
through predicted homologous recombination . This cer-
tainly complicates the subsequent mutants screening
and functional analysis . Even when the homologous
recombination has occurred, there are still rooms for
redundant DNA integration or gene loss at insertion
sites . Obtaining a true functional fungal mutant, such
as a PKS domain-replacement mutant, is not an easy
task (Yu et al. , 2005; Zhu et al. , 2006; Zhu et al. ,
2007) .
4 . 2 .2 Genetic manipulation
Among the three groups, the highly reduced
polyketides areparticularly challenging . These non-aro-
matic, often acyclic, metabolites arevery difficult to de-
tect and identify . The Du lab has been using sphinga-
nine-analog mycotoxins ( SAMT, including fumonisins
and AAL-toxins) as amodel systemto tackle thebiosyn-
thetic mechanismand manipulate the biosynthetic genes
(Zhu et al. , 2006; Du et al. , 2007; Zhu et al. ,
2007) . The PKS for fumosinins ( Fum1p, 2507 resi-
dues, encoded by FUM1) (Proctor et al. , 1999 ) and
the PKS for AAL-toxins (Alt1p, 2521 residues, en-
coded by ALT1 ) have an identical domain architec-
ture, KS-AT-DH-MT-ER-KR-ACP . However, Fum1p
synthesizes an 18-carbon chain, while Alt1p makes a
16-carbon chain . Interestingly, Fum1p and Alt1p do
not contain a thioesterase ( TE ) domain ( Hopwood,
1997) or a Claisen-like cyclase ( CLC) domain ( Fujii
et al. , 2001 ) . Thus, a novel mechanism is required
for the termination and offloading of the acyclic
polyketide chains (Du et al. , 2008) .
Using a genetic system developed for SAMT, we
generated several fumonisin FUM1 domain-replaced
mutants . The first was a F . verticillioides mutant with
the KS domain of T-toxin PKS1 from C . heterostrophus
(Kono and Daly, 1979; Yang et al. , 1996 ) replaced
the FUM1 KS ( Zhu et al. , 2006 ) . The mutant pro-
duced fumonisins, showing that the heterologous KS is
able to functionally interact with the six other domains
of Fum1p . We then replaced the KS domain of FUM1
with the KS domain of LovF , which encodes lovastatin
diketide synthase ( LDKS) responsible for the 2-meth-
ylbutyryl side chain of lovastatin in Aspergillus terreus
(Kennedy et al. , 1999 ) . This mutant produced four
new compounds, which are aromatic polyketides, di-
hydroisocoumarins ( Zhu et al. , 2007 ) . As discussed
above, the enzyme (NR-FPKS) architecture for fungal
aromatic ( non-reduced) polyketides is very different
fromthat (HR-FPKS) for fungal highly reduced ( non-
7623 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
aromatic) polyketides . The production of dihydroiso-
coumarins may suggest anewpolyketide chain releasing
mechanism, such as through inter-molecular ring for-
mation between two short ( C8 and C10 ) , linear
polyketide intermediates ( Zhu et al. , 2007 ) . These
results demonstrated that FPKS can be engineered to
produce new products . The dihydroisocoumarins are
known to have anti-malarial , antifungal , and anti-tu-
berculosis activities ( Kongsaeree et al. , 2003 ) and
had not been isolated from Fusarium . The new F . ver-
ticillioides strains do not produce any mycotoxins and,
instead, produce a group of active metabolites, which
could find application in biocontrol of fungal diseases .
4 . 2 .3 Heterologous expression of FPKS
Heterologous expression followed by biochemical
characterizationof theproduced enzymes is another way
to study FPKS function . 6-MSAS hasbeen expressed in
E. coli and S. cerevisiae ( Bedford et al. , 1995; Keal-
ey et al. , 1998) . Inorder for the FPKS to be function-
al , co-expressionof apromiscuous PPTasegene isoften
required . Unlike BPKS genes, FPKS genes usually
contain introns, which complicate the heterologous ex-
pression . To ensure the correct gene expression in the
heterologous host, either the full length cDNA of the
geneneeds to be obtained or the original gene must be
precisely spliced by a host strain to remove the introns .
Therefore, the commonly used E. coli andyeast systems
are not suitable for every FPKS . However, recent works
in the Tang lab showed that an intact and functional
NR-FPKS can be produced in E. coli ( Ma et al. ,
2007; Ma et al. , 2008) . They used PCR to amplify the
entire cDNA , and the primers were designed based on
information from sequence analysis . Interestingly, the
NR-FPKS were active when isolated from E. coli with-
out the co-expression of a promiscuous PPTase .
A number of filamentous fungi have been used for
heterologous expression of FPKS . The most successful
system is A. oryzae with the expression plasmid
pTAex3 . The atX gene from A. terreus, which en-
codes for the putative 6-MSAS, has been expressed in
A. oryzae ( Fujii et al. , 1996 ) . The 5 .5 kb atX
gene, including a70 bp intron, was cloned downstream
of the amyB promoter ( the starch-inducibleα-amylase
promoter from A. oryzae) in pTAex3 . The transformant
yielded a high level of 6-MSA , which proved the atX
gene encoding for a 6-MSAS analog . Later, the same
method was used for heterologous expression of PKS1
gene from Colletotrichumlagenariumin A. oryzae (Fu-
jii et al. , 1999) . The main compound produced by an
A. oryzae transformant was isolated and characterized to
be the acetate formof 1 , 3, 6, 8-tetrahydroxynaphthalene
(THN) , which unambiguously confirmed the THN syn-
thase encoded by the PKS1 gene .
Several HR-FPKS have also been expressed in fil-
amentous fungal hosts, such as the lovastatin LNKS
that was expressed in A. nidulans ( Kennedy et al. ,
1999) . The transformantswere selected through uridine
prototrophy andLNKS genewas inducedunder the con-
trol of alcA promoter ( the alcohol dehydrogenase pro-
moter from A. nidulans ) . The produced shunt
polyketides featured a shorter carbon chain and lower
degree of reduction than original dihydromonacolin L .
Further experiments proved that the aberrant behavior
of LNKS was not due to the heterologous backgroundof
host strain A. nidulans, but rather the lack of aspecif-
ic protein component LovC from the cluster to interact
with LNKS . The co-expressionof both LNKS and LovC
resulted in the production of authentic dihydromonaco-
lin L (Kennedy et al. , 1999 ) .
5 Type III PKS
Type III PKS ( PKS- III ) were initially thought to
be unique to plants but recently also found in certain
bacteria and fungi ( Staunton and Weissman, 2001;
Saxena et al. , 2003; Seshime et al. , 2005 ) . Fla-
vonoids, oneof the largest groups of plant metabolites,
are derived from polyketides . These compounds play
important roles in plant physiology, and some of them
also have biological activities beneficial to human
health, such as the stilbene-derived antioxidant res-
veratrol . Unlike the other two systems, PKS- III is a
unique single-protein system that does not have a do-
main or modular structure . The most distinct feature of
PKS-III is that it does not involve a carrier protein but
rather directly uses acyl CoAs as substrates . This single
protein complex can iteratively catalyze thebiosynthesis
862 云 南 植 物 研 究 30 卷
of polyketides through Claisen-type condensation .
Chalcone synthase ( CHS) was the first PKS-III
isolated from plants and is also the most extensively
studied PKS- III . It is a ubiquitous enzyme in plants
and consists of a super family of enzymes . To date,
Genebank has more than 700 CHS?STS-family genes
( Austin and Noel , 2003 ) . CHS catalyzes the first
committed step in flavonoid biosynthesis that forms
chalcone, which is a key intermediate of the pathway .
Chalcone is then modified by downstream enzymes in
various branching pathways to produce biologically ac-
tive flavonoids (Fig . 9) . PKS- III differs in the prefer-
ence for starter units, the number of condensation cy-
cles, and the ways to terminate and cyclize the chain .
For example, 2-pyrone synthase ( 2-PS) synthesizes 2-
methylpyrone by condensing acetyl-CoA with two ace-
tate units and cyclizes the carbon chain through lacton-
ization ( Jez et al. , 2000 ) ; the stilbene synthase
(STS) produces resveratrol almost in the same way as
CHS except that the intramolecular cyclization is
through an aldol condensation and accompanied by ad-
ditional decarboxylation of C1 carbon and dehydration
(Fig . 9) (Austin et al. , 2004) .
Both CHS and STS are homodimers with subunits
containing around 400 amino acids . The sequences of
commonly known STSs and CHSs are more than 60%
identical . Biochemical studies using site-directed mu-
tagenesis have identified the same essential active site
Fig . 9 Proposed biosynthetic pathway for chalcone and operating catalytic triadmachinery in active site cavity . The starter molecule is colored
in blue and the extender unit derived frommalonyl-CoA is representedwith a red bold bond . Theproposed mechanismsfor priming, decarboxy-
lation, Claisen condensation, elongation and cyclization are presented ( modified from (Austin and Noel , 2003; Austin et al., 2004) )
9623 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
Cys169 inCHS and STS, which corresponds to the cat-
alytic site cysteineused in normal KS for the condensa-
tion reaction ( Lanz et al. , 1991 ) . A hybrid PKS-III
was also constructed to get a CHS?STS recombinant,
which contained 107 amino acids of a CHS N-terminus
from Sinapis alba and 287 amino acids of a STS from
Arachis hypogaea ( Tropf et al. , 1994) . Although the
hybrid had no activity, three amino acid exchanges in
the CHS part (Q100E, V103M and V105R) were suf-
ficient to obtain low STS activity; and one additional
exchange (G23T) resulted in 20 - 25% of the STS ac-
tivity, which suggested an intimate evolutionary rela-
tionship between CHS and STS .
5 . 1 Structural studies of PKS-III
PKS-III is the best studied system in terms of its
three-dimensional structure . The Noel lab pioneered
this area . In 1999 , they obtained the first PKS struc-
ture, which was a 1 .56 ? resolution structure of CHS
from Medicago sativa (Ferrer et al. , 1999) . This was
a milestone in PKS- III research . Later, the 2 .05 ?
resolution crystal structureof 2-PS with the reaction in-
termediate acetoacetyl-CoA from Gerbera hybrida was
solved ( Jez et al. , 2000 ) . More recently, the 2 .1 ?
structure of STS from Pinussylvestrishas also been de-
termined (Austin et al. , 2004 ) . The determination of
the structures has led to the elucidation of the molecu-
lar basis of PKS-III biosynthesis and provided a frame-
work for genetic engineering of PKS-III genes to pro-
duce new products .
5 . 1 .1 CHS structure and biosynthetic mechanism
CHS uses 4-coumaroyl-CoA as the starter and
three malonyl-CoAs as the extender . The tetraketide
generated fromthe three cyclesof elongation is cyclized
by a Claisen condensation and aromatized to form the
ring system . The seriesof decarboxylation, condensati-
on, cyclization, and aromatization reactions are carried
out at a single activesite . Thestructureof CHS showed
that it forms a symmetric homodimer, with intertwined
N-terminal helix from each monomer and a tight loop
containingMet137 that acts as aknobwhich crosses the
interface and protrudes into thesurfaceof the adjoining
monomer ( Ferrer et al. , 1999 ) . Each monomer is
functionally independent . The active site is located at
the cleft between the upper and lower portions of each
monomer and is defined by threehighly conserved resi-
dues found in all known CHS-related enzymes . In this
catalytic triad, the central Cys164 serves as thenucleo-
phile to catalyze the condensation reaction; Asn336
plays a crucial role in the decarboxylation reaction by
providing ahydrogenbond to thecarbonyl oxygenof the
thioester, which can stabilize, with help of Phe215 ,
the enolate anion resulting from decarboxylationof mal-
onyl-CoA ; the protonated nitrogen on His303 couples
with Asn336 to forman oxyanion hole, which can pro-
vide an electron sink to stabilize the tetrahedral transi-
tion state formed during theCys164 activation and acetyl
carbanion derived from decarboxylation of malonyl-CoA
(Fig . 9) (Ferrer et al. , 1999; Jez et al. , 2000) .
Three interconnected cavities crosslink with these
four residues and form the active site architecture of
CHS . These cavities include a CoA-binding tunnel , a
coumaroyl-binding pocket, and a polyketide binding
pocket . There is alarge bi-lobed internal spaceof CHS
termed the initiation?elongation cavity (active site cavi-
ty) that covers the active site including catalytic triad .
One lobe at lower left of the CoA-binding tunnel′s end
is the coumaroyl-binding pocket . Several residues of
this pocket interact with the coumaroyl moiety of
polyketide chain and help the proper folding of the
chain . Another lobe adjacent to the active site is the
polyketide binding pocket which accommodates the
growing chain . The polyketide binding pocket is sepa-
rated with the coumaroyl-binding pocket by Phe265 ,
which is a steric gate to control the access of the cou-
maroyl-CoA . Some specific residues in the pocket are
potential hydrogen bond donors to interact with the
growingpolyketide chain and aid in the proper folding
of the intermediate for the cyclization reaction . The
overall active site cavity dictates how the linear
polyketide is folded and how the stereochemistry of the
cyclization reaction is controlled, as well as the choice
of starter molecule (Ferrer et al. , 1999) .
5 . 1 .2 Chain length determination in PKS-III
Although 2-PS from G. hybrida and CHS from
M. sativa share 74% amino acid sequence identity,
they use adifferent starter and catalyze different number
072 云 南 植 物 研 究 30 卷
Fig . 10 A . Proposed biosynthesis pathway for 2-methylpyrone . The starter molecule is colored in blue and the extender unit derived frommalo-
nyl-CoA iscolored in red bold bond; B . Aldol switch model with hydrogen bonding network in STS activesite cavity . Proposed cyclization mech-
anism in STS includes thioesterase- like hydrolysis, aldol condensation cyclization and subsequent decarboxylation and dehydration . The putative
hydrolytic water ( green) , starter unit ( blue) and extention units ( red bold bond) are highlighted (modified from ( Austin et al. , 2004 ) )
of condensation reactions ( Jez et al. , 2000 ) . 2-PS
synthesizes the triketide 2-methylpyrone froman acetyl-
CoA and two malonyl-CoAs (Fig . 10A) , whereas CHS
synthesizes the tetraketide chalcone from a 4-cou-
maroyl-CoA and three malonyl-CoAs . The 2 .05 ?
structureof 2-PS complexed with the reaction intermed-
1723 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
iate acetoacetyl-CoA was also determined ( Jez et al. ,
2000) . 2-PS and CHS share a common three-dimen-
sional fold, a set of conserved catalytic residues, and
similar CoA binding sites . However, the active site
cavity of 2-PS (274?3 ) is smaller than theone inCHS
(923 ?3 ) . Several bulky amino acid residues occupy
the 2-PS cavity that results in a smaller cavity, which
subsequently results in the useof a smaller starter unit
acetate in 2-PS . The structures imply that the size of
the activesitecavity controls startermolecule selectivity
and limits the chain lengthof products .Thishypothesis
is supported by experimental evidence fromsite-direct-
ed mutagenesis . Mutations in the active site cavity of
CHS ( T197L?G256L?S338I ) yielded an enzyme that
was functionally identical to 2-PS ( Jez et al. , 2000 ) .
This demonstrates that the active site cavity volume in-
fluences the choiceof starter molecule and dictates the
length of the polyketide .
5 . 1 .3 Cyclization mechanisms in PKS-III
CHS and STS share 75% - 90% identical se-
quenceover their 400 residues, andboth enzymes cata-
lyzethesame iterative condensationof three acetyl units
to a starter molecule of 4-coumaroyl-CoA . However,
STS cyclizes the tetraketide intermediate via an in-
tramolecular C2 to C7 aldol condensation, whereas
CHS cyclizes the intermediate via an intramolecular C6
to C1 Claisen condensation . Recently, the 2 .1? crys-
tal structure of pinosylvin-forming STS from P . sylv-
estris has been solved andprovided insights into the cy-
clization mechanisms (Austin et al. , 2004 ) . As ex-
pected, STS has the sameoverall structural fold as ob-
served in CHS, except two areas . To determine wheth-
er any of thesestructural differences correlates with the
aldol cyclization specificity of STS, Austin and Noel
converted similar regions ( areas 1 - 4 ) in CHS to the
corresponding residues in STS through mutagenesis . In-
deed, the resulting 18xCHS mutant functionally resem-
bled wild-type STS, as reflected by both kinetic prop-
erties and product specificity ( Austin et al. , 2004 ) .
The available results from the mutant library revealed
that region B is unimportant for cyclization specificity,
but the buried loop (area 2 ) in region A is critical for
mediating the change fromaClaisen condensation to an
aldol cyclization . In contrast to the previously proposed
steric mechanism for STS and CHS functional diver-
gence, the finding of the aldol switch in STS suggests
that the cyclization specificity is determined by the ki-
netic competition between Claisen-mediated ring cy-
clization and thioesterase-catalyzed hydrolysis with fac-
ile aldol cyclization (Fig . 10B) . The latter process is
favored in STS by the emergence of a thioesterase-like
aldol switch, the hydrogen bonding network through
Glu192 and Thr132 activates a Ser338-positioned water
molecule as a nucleophile to catalyze basic hydrolysis
and lead to subsequent aldol cyclization .
In summary, the structural analysis coupled with
site directed mutagenesis established themechanismfor
chain length and cyclization specificity determination in
plant PKS- III . Using the catalytic triad as a condensa-
tion engine, all the reactions catalyzed by plant PKS-
III occur in the active site cavity . Thevolumeof cavity
dictates the length of the product, and the interactions
between cavity lining residues and intermediates decide
the folding pattern and cyclization pathway for final
polyketides .
5 . 2 Structure-directed mutagenesis in PKS-III
The availability of three-dimensional structures
has made it possible to carry out the structure-directed
mutagenesis in PKS- IIIs . This approach is not only
useful to confirm the proposed biosynthetic mechanism
but also to exploit the engineered biosynthesis of novel
polyketides . For example, Gly256 that resides on the
surfaceof the CHS active site and is in direct contact
with the polyketide has been targeted to probe the link
between cavity volume and polyketide chain-length .
The side chain volume of position 256 influences both
thenumber of condensation reactions and the conforma-
tionof the intermediates during the cyclization . Muta-
tions of G256A , G256V , G256L , and G256F reduced
the sizeof the active site cavity but without significant
alterations in the CHS backbones ( Jez et al. , 2001 ) .
The gatekeeper residue Phe215 , which is situated at
the active site entrance, has also been subjected to
site-directedmutagenesis to diversify CHS activity . The
F215S mutant preferentially accepted N-methylanthra-
niloyl-CoA to generate a novel alkaloid named N-meth-
272 云 南 植 物 研 究 30 卷
ylanthraniloyltriacetic acid lactone ( Jez et al. , 2002 ) .
This point mutation in the entrance of the CHS cavity
not only dramatically shifts the molecular selectivity of
the enzyme, but also changes the cyclization mode .
The S338V mutant and a triple mutant ( T197G?
G256L?S338T ) of CHS even produced octaketides
SEK4 and SEK4b from eight molecules of malonyl-
CoA , which are normally isolated fromPKS-II systems
(Abe et al. , 2006) . These results reaffirm the impor-
tance of lining residues around active site cavity region
and demonstrate the potential of structure-directed bio-
synthesis in PKS- III . More recently, astructure-guided
engineering of plant PKS-III has led to the production
of new polyketides (Abe et al. , 2007) .
5 . 3 PKS-IIIs from microorganisms
Traditionally , PKS- III hasbeen regarded as aplant
PKS system . Since 1999 , PKS- III-like enzymes have
been found in bacteria and fungi .The rppA gene isolat-
ed from Streptomyces griseus encodes for a 372-amino-
acid protein, which is significantly similar to CHS and
synthesizes 1, 3 , 6, 8-tetrahydroxynaphthalene ( THN)
(Funa et al. , 1999) . Another similar THN PKS- III was
found in Saccharopolyspora erythraea ( Cortes et al. ,
2002) . Two novel PKS-IIIs involved in the biosynthesis
of long-chain alpha-pyrones have been isolated from
Mycobacterium tuberculosis ( Saxena et al. , 2003 ) .
CHS-like genes arealsofound in the filamentous fungi ,
A. oryzae, N. crassa, and F . graminearum (Seshime
et al. , 2005 ) . Unlike the PKS- III in plants that share
as high as 60 - 95% similarity in sequences, the mi-
crobial PKS- III typically have only 25 - 50% similari-
ty . Microbial PKS-III also possesses the universal Cys-
His-Asn catalytic triad at similar postions identified in
CHS . More recently, structural studies of bacterial
PKS-III have also appeared in the literature (Sankara-
narayanan et al. , 2004 ) . The studies reaffirmed that
subtle changes in protein architecture can generateme-
tabolite diversity .
6 Final Remarks
Polyketides are probably the most important group
of natural products in terms of the significance in hu-
manmedicines . Thesemetabolites exhibit an enormous
rangeof functional and structural diversity . In the past
two decades, significant progress has been made in our
understanding of themolecular mechanisms for thebio-
synthesis . Numerous PKS have been identified from a
variety of biological sources . This process has been ac-
celerated in recent years largely due to the completion
of many genome sequencing projects . With rapidly in-
creasing numbers of PKS genes, there have been nu-
merous cases where theorganizations of the PKS do not
fall into any of the traditional classifications . Although
the PKS family has been classified into three major
groups, modular PKS-I includingmultimodular BPKS-I
and iterative unimodular FPKS, dissociate PKS-II , and
integrated complex PKS- III , many newly discovered
PKS exhibit mixed features of the three types (Shen,
2003; Muller, 2004 ) . It is now clear that the PKS
family isnot limited to theseveral classesof biosynthet-
ic systems . Numerous“non-classical”PKS genes have
emerged ( Jia et al. , 2006; Shao et al. , 2006) . More-
over, a largenumberof PKS exist as ahybridwith FAS
or nonribosomal peptide synthetases (NRPS) ( Du et
al. , 2000; Du et al. , 2003; Zhao et al. , 2006;
Gokhale et al. , 2007; Li et al. , 2008 ) . These fea-
tures have provided not only evidence for the catalytic
and mechanistic versatility of PKSs but also newoppor-
tunities for biosynthetic engineering to produce hybrid
metabolites ( Du et al. , 2001; Du and Shen, 2001;
Walsh, 2004 ) . The crystal structures of BPKS- I ,
PKS-II , and PKS-III havegreatly enhanced the under-
standing of the biosynthetic mechanisms at the molecu-
lar level . In some cases, the structures have also pro-
vided vital information for rational engineering of PKS
to produce novel polyketides .
In spiteof the enormous progress, much more re-
mains to be learned . It is still not clear what thediffer-
ences are in the overall structure of PKS or PKS com-
plex that distinguishes the characteristics and selectivi-
ties invariousgroupsof PKS .Themechanismbywhich
FPKS control the polyketide product structure remains
largely unknown . None of the fungal PKS has been
studied for the3-D structure . In theabsenceof structur-
al information, the genetic manipulation of FPKS has re-
cently shed light on themechanistic understandingof fun-
3723 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
gal polyketide biosynthesis and has initiated to produce
“unnatural”natural products in fungi (Zhu et al., 2007) .
References:
Abe I , MoritaH , Oguro S et al. , 2007 . Structure-based engineering of a
plant type II I polyketide synthase: formation of an unnatural nona-
ketide naphthopyrone [ J ] . J AmChemSoc, 129: 5976—5980
Abe #I , Watanabe T, Morita H et al. , 2006 . Engineered biosynthesis of
plant polyketides: manipulation of chalcone synthase [ J ] . Org Lett,
8 : 499—502
Apar #icio JF , Molnar I , Schwecke T et al. , 1996 . Organization of thebi-
osynthetic gene cluster for rapamycin in Streptomyces hygroscopicus:
analysis of the enzymatic domains in the modular polyketide synthase
[ J ] . Gene, 169: 9—16
Arth ?ur CJ , SzafranskaA , Evans SE et al. , 2005 . Self-malonylation is an
intrinsic property of a chemically synthesized type II polyketide syn-
thase acyl carrier protein [ J ] . Biochemistry, 44 : 15414—15421
Arth ?ur CJ , Szafranska AE , Long J et al. , 2006 . The malonyl transferase
activity of type II polyketide synthase acyl carrier proteins [ J ] . Chem
Biol , 13 : 587—596
Augu /st PR, Tang L , Yoon YJ et al. , 1998 . Biosynthesis of the ansamy-
cin antibiotic rifamycin: deductions from the molecular analysis of
the rif biosynthetic gene cluster of Amycolatopsis mediterranei S699
[ J ] . ChemBiol , 5 : 69—79
Aust ?in MB , Bowman ME, Ferrer JL et al. , 2004 . An aldol switch dis-
covered in stilbene synthases mediates cyclization specificity of type
III polyketide synthases [ J ] . Chem Biol , 11 : 1179—1194
Aust ?in MB , Noel JP, 2003 . The chalcone synthase superfamily of type III
polyketide synthases [ J ] . Nat Prod Rep, 20 : 79—110
Balt ?z RH , 2006 , Molecular engineering approaches to peptide, polyketide
and other antibiotics [ J ] . Nat Biotechnol , 24 : 1533—1540
Bao ?W, Sheldon PJ , Hutchinson CR, 1999 . Purification and properties of
the Streptomyces peucetius DpsC beta-ketoacyl : acyl carrier protein
synthase III that specifies the propionate-starter unit for type II
polyketide biosynthesis [ J ] . Biochemistry, 38 : 9752—9757
Bao ?W, Wendt-Pienkowski E , Hutchinson CR , 1998 . Reconstitution of
the iterative type II polyketide synthasefor tetracenomycin F2 biosyn-
thesis [ J ] . Biochemistry, 37 : 8132—8138
Beck ;J , Ripka S, Siegner A et al. , 1990 . Themultifunctional 6-methyl-
salicylic acid synthase gene of Penicillium patulum . I ts gene struc-
turerelative to that of other polyketide synthases [ J ] . Eur J Bio-
chem, 192 : 487—498
Bedf ?ord DJ , Schweizer E, Hopwood DA et al. , 1995 . Expression of a
functional fungal polyketide synthase in the bacterium Streptomyces
coelicolor A3 (2 ) [ J ] . J Bacteriol , 177: 4544—4548
Bezu idenhout SC , Gelderblom WCA , Gorst-Allman CP et al. , 1988 .
Structure elucidation of the fumonisins, mycotoxins from Fusarium
moniliforme [ J ] . J Chem Soc ChemCommun, 743—745
Bing ?le LE , Simpson TJ , Lazarus CM, 1999 . Ketosynthase domain probes
identify two subclasses of fungal polyketidesynthasegenes [ J ] . Fun-
gal Genet Biol , 26 : 209—223
Bisa ?ng C, LongPF , Cortes J et al. , 1999 . A chain initiation factor com-
mon to both modular and aromatic polyketide synthases [ J ] . Nature,
401: 502—505
Bohm ?I , Holzbaur IE , Hanefeld U et al. , 1998 . Engineering of a min-
imal modular polyketide synthase, and targeted alteration of theste-
reospecificity of polyketide chain extension [ J ] . Chem Biol , 5 :
407—412
Broa ?dhurst RW, Nietlispach D, Wheatcroft MP et al. , 2003 . The struc-
ture of docking domains in modular polyketide synthases [ J ] . Chem
Biol , 10 : 723—731
Brow ?n JS, Aufauvre-Brown A , HoldenDW, 1998 . Insertional mutagene-
sis of Aspergillus fumigatus [ J ] . Mol Gen Genet, 259 : 327—335
Caff ?rey P, Bevitt DJ , Staunton J , Leadlay PF , 1992 . Identification of
DEBS 1 , DEBS 2 and DEBS 3 , the multienzyme polypeptides of the
erythromycin-producing polyketide synthase from Saccharopolyspora
erythraea [ J ] . FEBS Lett, 304 : 225—228
Carr ?eras CW, Khosla C, 1998 . Purification and in vitro reconstitution of
the essential protein components of an aromatic polyketide synthase
[ J ] . Biochemistry, 37 : 2084—2088
Cieg ?ler A, 1975 . Mycotoxins: occurrence, chemistry, biological activity
[ J ] . Lloydia, 38 : 21—35
Cort ?es J , Haydock SF , Roberts GA et al. , 1990 . An unusually large
multifunctional polypeptide in the erythromycin-producing polyketide
synthase of Saccharopolyspora erythraea [ J ] . Nature, 348: 176—
178
Cortes J , Velasco J , Foster G et al. , 2002 . Identification and cloning of
a type III polyketide synthase required for diffusible pigment biosyn-
thesis in Saccharopolyspora erythraea [ J ] . Mol Microbiol , 44 :
1213—1224
Cox ?RJ , 2007 . Polyketides, proteins and genes in fungi : programmed
nano-machines begin to reveal their secrets [ J ] . Org Biomol Chem,
5 : 2010—2026
Cox RJ , GlodF , Hurley D et al. , 2004 . Rapid cloning and expression of
a fungal polyketidesynthasegene involved in squalestatin biosynthesis
[ J ] . ChemCommun ( Camb) , 2260—2261
Craw ?ford JM, Dancy BC , Hill EA et al. , 2006 . Identification of astarter
unit acyl-carrier protein transacylase domain in an iterative type I
polyketide synthase [ J ] . Proc Natl Acad Sci USA, 103: 16728—
16733
Crop ?p TA , Wilson DJ , Reynolds KA , 2000 . Identification of a cyclohex-
ylcarbonyl CoA biosynthetic gene cluster and application in the pro-
duction of doramectin [ J ] . Nat Biotechnol , 18 : 980—983
Crum ?p MP, Crosby J , Dempsey CE et al. , 1997 . Solution structure of
the actinorhodin polyketide synthase acyl carrier protein from Strep-
tomycescoelicolor A3 ( 2) [ J ] . Biochemistry, 36 : 6000—6008
de G ?root MJ , Bundock P, Hooykaas PJ et al. , 1998 . Agrobacterium tu-
mefaciens-mediated transformation of filamentous fungi [ J ] . Nat
Biotechnol , 16 : 839—842
Deng ?Z, Bai L , 2006 . Antibiotic biosynthetic pathwaysand pathway engi-
neering—a growing research field in China [ J ] . Nat Prod Rep,
23 : 811—827
Dona ?dio S, Katz L , 1992 . Organization of the enzymatic domains in the
472 云 南 植 物 研 究 30 卷
multifunctional polyketide synthase involved in erythromycin forma-
tion in Saccharopolyspora erythraea [ J ] . Gene, 111: 51—60
Dona +dio S, McAlpine JB , Sheldon PJ et al. , 1993 . An erythromycin an-
alog produced by reprogramming of polyketide synthesis [ J ] . Proc
Natl Acad Sci USA, 90 : 7119—7123
Dona +dio S, Monciardini P, Sosio M , 2007 . Polyketide synthases and
nonribosomal peptide synthetases: the emerging view from bacterial
genomics [ J ] . Nat Prod Rep, 24 : 1073—1109
Dona +dio S, Staver MJ , McAlpine J B et al. , 1991 . Modular organization
of genes required for complex polyketide biosynthesis [ J ] . Science,
252 : 675—679
Drei ?er J , Khosla C, 2000 . Mechanistic analysis of a type II polyketide
synthase . Role of conserved residues in the beta-ketoacyl synthase-
chain length factor heterodimer [ J ] . Biochemistry, 39 : 2088—2095
Drei ?er J , ShahAN , Khosla C, 1999 . Kinetic analysisof the actinorhodin
aromatic polyketidesynthase [ J ] . J Biol Chem, 274: 25108—25112
Du L A, Cheng Y , Ingenhorst G et al. , 2003 . Hybrid peptide-polyketide
natural products: Biosynthesis and prospects towards engineering
novel molecules [ A ] . In: Setlow JK ed, Genentic Engineering———
Principles and Methods, Vol 25 [M ] . New York: Kluwer Academ-
ic?Plenum Publishers, 227—267
Du L C, Sanchez C , Chen M et al. , 2000 . The biosynthetic gene cluster
for the antitumor drug bleomycin from Streptomyces verticillus
ATCC15003 supporting functional interactions between nonribosomal
peptide synthetases and a polyketide synthase [ J ] . Chem Biol , 7 :
623—642
Du L H, Sanchez C, Shen B, 2001 . Hybrid peptide-polyketide natural
products: Biosynthesis and prospects toward engineering novel mole-
cules [ J ] . Metabolic Engineering, 3 : 78—95
Du L ?, Shen B, 2001 . Biosynthesis of hybrid peptide-polyketide natural
products [ J ] . Curr Opin Drug Discov Devel , 4 : 215—228
Du L F, YuF , ZhuX et al. , 2007 . Biochemical and molecular analysis of
the biosynthesis of fumonisins [ A ] . In: Baerson SR ed,
Polyketides: Biosynthesis, Biological Activities and Genetic Engi-
neering, Vol ACS SymposiumSeries 955 [ M ] . Washington, DC :
American Chemical Society, 81—96
Du L ?, ZhuX , Gerber R et al. , 2008 . Biosynthesis of sphinganine-analog
mycotoxins [ J ] . J Ind Microbiol Biotechnol , in press
Dutt ?on CJ , Gibson SP, GoudieAC et al. , 1991 . Novel avermectins pro-
duced by mutational biosynthesis [ J ] . J Antibiot ( Tokyo) , 44 :
357—365
Feng 7B, WangX , Hauser M et al., 2001 .Molecular cloning and charac-
terization of WdPKS1 , agene involved in dihydroxynaphthalene mel-
anin biosynthesis and virulence in Wangiella ( Exophiala) dermatit-
idis [ J ] . Infect Immun, 69 : 1781—1794
Ferr ?er JL , J ez JM , Bowman ME et al. , 1999 . Structure of chalcone syn-
thase and the molecular basis of plant polyketide biosynthesis [ J ] .
Nat Struct Biol , 6 : 775—784
Fisc ?hbach MA , Walsh CT, 2006 . Assembly- line enzymology for
polyketide and nonribosomal Peptide antibiotics: logic, machinery,
and mechanisms [ J ] . Chem Rev, 106: 3468—3496
Fu H E, Hopwood DA , Khosla C, 1994 . Engineered biosynthesis of novel
polyketides: evidence for temporal , but not regiospecific, control of
cyclization of an aromatic polyketide precursor [ J ] . Chem Biol , 1 :
205—210
Fuji ?i I , Mori Y , WatanabeA et al. , 1999 . Heterologous expression and
product identification of Colletotrichumlagenariumpolyketidesynthase
encoded by thePKS1 gene involved inmelanin biosynthesis [ J ] . Bio-
sci Biotechnol Biochem, 63 : 1445—1452
Fuji ?i I , Ono Y , Tada H et al. , 1996 . Cloning of the polyketide synthase
gene atX from Aspergillus terreusand its identification as the 6-methyl-
salicylic acid synthasegene by heterologous expression [ J ] . Mol Gen
Genet, 253: 1—10
Fujii I , Watanabe A , Sankawa U et al. , 2001 . Identification of Claisen
cyclase domain in fungal polyketide synthase WA, a naphthopyrone
synthaseof Aspergillus nidulans [ J ] . Chem Biol , 8 : 189—197
FunaN, Ohnishi Y , Fujii I et al. , 1999 . A new pathway for polyketide
synthesis in microorganisms [ J ] . Nature, 400: 897—899
Gand ?echa AR , LargeSL , Cundliffe E, 1997 . Analysis of four tylosin bi-
osynthetic genes fromthe tylLM region of the Streptomycesfradiaege-
nome [ J ] . Gene, 184: 197—203
Gokh ?ale RS, Khosla C , 2000 . Roleof linkers in communication between
protein modules [ J ] . Curr Opin Chem Biol , 4 : 22—27
Gokh ?ale RS, Saxena P, ChopraT et al. , 2007 . Versatile polyketide enz-
ymatic machinery for the biosynthesis of complex mycobacterial lipids
[ J ] . Nat Prod Rep, 24 : 267—277
Gokh ?ale RS, Tsuji SY , Cane DE et al. , 1999 . Dissecting and exploiting
intermodular communication in polyketide synthases [ J ] . Science,
284: 482—485
Han ?L , Lobo S, Reynolds KA , 1998 . Characterization of beta-ketoacyl-
acyl carrier protein synthase III from Streptomyces glaucescens and its
role in initiation of fatty acid biosynthesis [ J ] . J Bacteriol , 180:
4481—4486
Hers ?hberger CL , 1996 . Metabolic engineering of polyketide biosynthesis
[ J ] . Curr Opin Biotechnol , 7 : 560—562
Hert ?weck C, Luzhetskyy A , Rebets Y et al., 2007 . Type II polyketide
synthases: gaining a deeper insight into enzymatic teamwork [ J ] .
Nat Prod Rep, 24 : 162—190
Hitchman TS, Crosby J , ByromKJ et al. , 1998 . Catalytic self-acylation
of type II polyketide synthase acyl carrier proteins [ J ] . Chem Biol ,
5 : 35—47
Hopw ?ood DA , 1997 . Genetic Contributions to Understanding Polyketide
Synthases [ J ] . Chem Rev, 97 : 2465—2498
J ez ?JM, Austin MB, Ferrer J et al. , 2000 . Structural control of
polyketide formation in plant-specific polyketide synthases [ J ] . Chem
Biol , 7 : 919—930
J ez ?JM, Bowman ME, Noel JP , 2001 . Structure-guided programming of
polyketide chain- length determination in chalcone synthase [ J ] . Bio-
chemistry, 40 : 14829—14838
J ez ?JM, Bowman ME , Noel JP, 2002 . Expanding the biosynthetic reper-
toireof plant type III polyketide synthases by altering starter molecule
specificity [ J ] . Proc Natl Acad Sci USA, 99 : 5319—5324
J ez ?JM, Ferrer JL , Bowman ME et al. , 2000 . Dissection of malonyl-co-
enzyme A decarboxylation from polyketide formation in the reaction
5723 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
mechanism of a plant polyketide synthase [ J ] . Biochemistry, 39 :
890—902
J ia XY , Tian ZH , Shao L et al. , 2006 . Genetic characterization of the
chlorothricin gene cluster as a model for spirotetronate antibiotic bio-
synthesis [ J ] . Chem Biol , 13 : 575—585
Kao $CM , Katz L , KhoslaC , 1994 . Engineered biosynthesis of a complete
macrolactone in a heterologous host [ J ] . Science, 265: 509—512
Kao !CM, Luo G, Katz L et al. , 1995 . Manipulation of macrolide ring
size by directed mutagenesis of amodular polyketide synthase [ J ] . J
AmChemSoc, 117: 9105—9106
Kao #CM , Pieper R, Cane DE et al. , 1996 . Evidencefor two catalytically
independent clusters of active sites in a functional modular polyketide
synthase [ J ] . Biochemistry, 35 : 12363—12368
Keal ?ey JT, Liu L , Santi DV et al. , 1998 . Production of apolyketidenat-
ural product in nonpolyketide-producing prokaryotic and eukaryotic
hosts [ J ] . Proc Natl Acad Sci USA , 95 : 505—509
Keat ?inge-Clay AT, Maltby DA , Medzihradszky KF et al. , 2004 . An an-
tibiotic factory caught in action [ J ] . Nat Struct Mol Biol , 11 : 888—
893
Keat ?inge-Clay AT, Stroud RM, 2006 . The structure of a ketoreductase
determines theorganization of the beta-carbon processing enzymes of
modular polyketide synthases [ J ] . Structure, 14 : 737—748
Kemp ?ken F , Kuck U , 1998 . Transposons in filamentous fungi—facts and
perspectives [ J ] . Bioessays, 20 : 652—659
Kenn /edy J , Auclair K , Kendrew SG et al. , 1999 . Modulation of
polyketide synthase activity by accessory proteins during lovastatin
biosynthesis [ J ] . Science, 284: 1368—1372
Khos &la C , Gokhale RS, Jacobsen JR et al. , 1999 . Tolerance and speci-
ficity of polyketide synthases [ J ] . Annu Rev Biochem, 68 : 219—
253
Khos &la C , McDaniel R , Ebert-KhoslaS et al. , 1993 . Genetic construc-
tion and functional analysis of hybrid polyketide synthases containing
heterologous acyl carrier proteins [ J ] . J Bacteriol , 175: 2197—
2204
Khos &la C, Tang Y , Chen AY et al. , 2007 . Structure and mechanism of
the 6-deoxyerythronolide B synthase [ J ] . Annu Rev Biochem, 76 :
195—221
Kim BES, Hopwood DA, Sherman DH , 1994 . Analysis of type II
polyketide beta-ketoacyl synthase specificity in Streptomycescoelicol-
or A3 ( 2) by trans complementation of actinorhodin synthasemuta-
nts [ J ] . J Bacteriol , 176 : 1801—1804
Kino ?shita K , Williard PG, Khosla C et al. , 2001 . Precursor-directed bi-
osynthesis of 16-membered macrolides by the erythromycin polyketide
synthase [ J ] . J AmChemSoc, 123: 2495—2502
Kong +saeree P , Prabpai S, Sriubolmas N et al., 2003 . Antimalarial di-
hydroisocoumarins produced by Geotrichum sp ., an endophytic fun-
gus of Crassocephalumcrepidioides [ J ] . J Nat Prod, 66 : 709—711
Kono >Y , Daly JM , 1979 . Characterization of the host-specific pathotoxin
produced by Helminthosporium maydis race T affecting corn with
Texas male sterile cytoplasm [ J ] . Bioorg Chem, 8 : 391—397
Kopp DF , Marahiel MA , 2007 . Macrocyclization strategies in polyketide
and nonribosomal peptide biosynthesis [ J ] . Nat Prod Rep, 24 :
735—749
Krok ?en S, Glass NL , Taylor JW et al. , 2003 . Phylogenomic analysis of
type I polyketide synthasegenes in pathogenic and saprobic ascomy-
cetes [ J ] . Proc Natl Acad Sci USA, 100: 15670—15675
Kuhs ?toss S, Huber M, Turner JR et al. , 1996 . Production of a novel
polyketide through the construction of a hybrid polyketide synthase
[ J ] . Gene, 183: 231—236
Lai ?JR , Koglin A , Walsh CT , 2006 . Carrier protein structure and recog-
nition in polyketide and nonribosomal peptide biosynthesis [ J ] . Bio-
chemistry, 45 : 14869—14879
Lamb ?alot RH , Gehring AM , Flugel RS et al. , 1996 . A new enzyme su-
perfamily-thephosphopantetheinyl transferases [ J ] . Chem Biol , 3 :
923—936
Lanz T, Tropf S, Marner FJ et al. , 1991 . The role of cysteines in
polyketide synthases . Site-directed mutagenesis of resveratrol and
chalcone synthases, two key enzymes in different plant-specific path-
ways [ J ] . J Biol Chem, 266 : 9971—9976
Li L ?, DengW, Song J et al. , 2008 . Characterization of the saframycin A
gene cluster from Streptomyces lavendulae NRRL 11002 revealing a
nonribosomal peptide synthetase system for assembling the unusual
tetrapeptidyl skeleton in an iterative manner [ J ] . J Bacteriol , 190:
251—263
Lu S , Lyngholm L , Yang G et al., 1994 . Tagged mutations at the Tox1
locus of Cochliobolus heterostrophus by restriction enzyme-mediated
integration [ J ] . Proc Natl Acad Sci USA , 91 : 12649—12653
Ma S ?M, Zhan J , Watanabe K et al. , 2007 . Enzymatic synthesis of aro-
matic polyketides using PKS4 from Gibberella fujikuroi [ J ] . J Am
ChemSoc, 129: 10642—10643
Ma S ?M, Zhan J , XieX et al. , 2008 . Redirecting the cyclization steps of
fungal polyketide synthase [ J ] . J AmChemSoc, 130: 38—39
MacN ?eil DJ , Occi JL , Gewain KM et al. , 1992 . Complex organization of
the Streptomyces avermitilis genes encoding the avermectin polyketide
synthase [ J ] . Gene, 115: 119—125
Maie ?r T, Jenni S, Ban N , 2006 . Architecture of mammalian fatty acid
synthase at 4 . 5 A resolution [ J ] . Science, 311: 1258—1262
Mars ?den AF , Wilkinson B, Cortes J et al., 1998 . Engineering broader
specificity into an antibiotic-producing polyketide synthase [ J ] . Sci-
ence, 279 : 199—202
Math ?aru AL , Cox RJ , Crosby J et al. , 1998 . MCAT is not required for
in vitro polyketide synthesis in aminimal actinorhodin polyketidesyn-
thase from Streptomyces coelicolor [ J ] . Chem Biol , 5 : 699—711
McDaniel R , Ebert-Khosla S, Fu H et al. , 1994 . Engineered biosynthe-
sis of novel polyketides: influence of a downstream enzyme on the
catalytic specificity of a minimal aromatic polyketide synthase [ J ] .
Proc Natl Acad Sci USA, 91 : 11542—11546
McDa ?niel R , Ebert-KhoslaS, HopwoodDA et al. , 1993 . Engineered bi-
osynthesis of novel polyketides [ J ] . Science, 262: 1546—1550
McDa ?niel R , Ebert-KhoslaS, HopwoodDA et al. , 1993 . Engineered bi-
osynthesis of novel polyketides: manipulation and analysis of an aro-
matic polyketide synthasewith unproven catalytic specificities [ J ] .
J AmChemSoc, 115: 11671—11675
McDa ?niel R , Ebert-Khosla S, Hopwood DA et al. , 1994 . Engineered
672 云 南 植 物 研 究 30 卷
Biosynthesis of Novel Polyketides: actVII and actIV Genes Encode
Aromatase and Cyclase Enzymes, Respectively [ J ] . J Am Chem
Soc, 116: 10855—10859
McDa @niel R , Ebert-Khosla S, Hopwood DA et al. , 1995 . Rational de-
sign of aromatic polyketidenatural products by recombinant assembly
of enzymatic subunits [ J ] . Nature, 375: 549—554
Mead 3ows ES, Khosla C, 2001 . In vitro reconstitution and analysis of the
chain initiating enzymes of the R1128 polyketide synthase [ J ] . Bio-
chemistry, 40 : 14855—14861
Menz .ella HG, ReevesCD, 2007 . Combinatorial biosynthesis for drug de-
velopment [ J ] . Curr Opin Microbiol , 10 : 238—245
Meur +er G, Gerlitz M, Wendt- Pienkowski E et al. , 1997 . Iterative type II
polyketide synthases, cyclases and ketoreductases exhibit context-de-
pendent behavior in the biosynthesis of linear and angular deca-
polyketides [ J ] . Chem Biol , 4 : 433—443
Mich ielse CB , Hooykaas PJ , van denHondel CA et al. , 2005 . Agrobac-
terium-mediated transformation as a tool for functional genomics in
fungi [ J ] . Curr Genet, 48 : 1—17
Moor e BS, Hertweck C, 2002 . Biosynthesisand attachment of novel bac-
terial polyketide synthase starter units [ J ] . Nat Prod Rep, 19 : 70—
99
Mori ?guchi T, Ebizuka Y , Fujii I , 2006 . Analysis of subunit interactions
in the iterative type I polyketidesynthaseATX from Aspergillus terreus
[ J ] . Chembiochem, 7 : 1869—1874
Mull ?er R , 2004 . Don′t classify polyketide synthases [ J ] . Chem Biol ,
11 : 4—6
Nesb &itt BF , O′Kelly J , Sargeant K et al. , 1962 . Aspergillus flavus and
turkey X disease . Toxic metabolites of Aspergillus flavus [ J ] . Na-
ture, 195 : 1062—1063
Nich ?olson TP, Rudd BA , Dawson M et al. , 2001 . Design and utility of
oligonucleotide gene probes for fungal polyketide synthases [ J ] .
Chem Biol , 8 : 157—178
Okam ?oto S, Sakurada M , Kubo Y et al. , 2001 . Inhibitory effect of
aflastatin A on melanin biosynthesis by Colletotrichum lagenarium
[ J ] . Microbiology, 147: 2623—2628
Oliy ?nyk M, Brown MJ , Cortes J et al. , 1996 . A hybrid modular
polyketide synthase obtained by domain swapping [ J ] . Chem Biol ,
3 : 833—839
Ozek &i K , Kyoya F , HizumeK et al. , 1994 . Transformation of intact As-
pergillus niger by electroporation [ J ] . Biosci Biotechnol Biochem,
58 : 2224—2227
Pan ?H , Tsai S, Meadows ES et al. , 2002 . Crystal structure of theprim-
ing beta-ketosynthase from the R1128 polyketidebiosynthetic pathway
[ J ] . Structure ( Camb) , 10 : 1559—1568
Pfei ?fer BA , Admiraal SJ , Gramajo H et al. , 2001 . Biosynthesis of com-
plex polyketides in a metabolically engineered strain of E. coli [ J ] .
Science, 291: 1790—1792
Piel ?J , Hertweck C , Shipley PR et al. , 2000 . Cloning, sequencing and
analysis of theenterocin biosynthesis gene cluster from the marine iso-
late‘ Streptomyces maritimus’: evidencefor thederailment of an aro-
matic polyketide synthase [ J ] . Chem Biol , 7 : 943—955
Piep ?er R, Luo G, Cane DE et al. , 1995 . Cell-free synthesis of
polyketides by recombinant erythromycin polyketide synthases [ J ] .
Nature, 378: 263—266
Proc ?tor RH , DesjardinsAE , Plattner RD et al. , 1999 . A polyketidesyn-
thase gene required for biosynthesis of fumonisin mycotoxins in Gib-
berella fujikuroi mating population A [ J ] . Fungal Genet Biol , 27 :
100—112
Quad ?ri LE , Weinreb PH , Lei M et al. , 1998 . Characterization of Sfp, a
Bacillus subtilis phosphopantetheinyl transferase for peptidyl carrier
protein domains in peptide synthetases [ J ] . Biochemistry, 37 :
1585—1595
Rang ?anathan A , Timoney M, Bycroft M et al. , 1999 . Knowledge-based
design of bimodular and trimodular polyketide synthases based on do-
main and module swaps: a route to simple statin analogues [ J ] .
Chem Biol , 6 : 731—741
Rowe ?CJ , Cortes J , Gaisser S et al. , 1998 . Construction of new vectors
for high- level expression in actinomycetes [ J ] . Gene, 216: 215—
223
Ruan ?X , PeredaA , Stassi DL et al. , 1997 . Acyltransferase domain sub-
stitutions in erythromycin polyketide synthase yield novel erythromy-
cin derivatives [ J ] . J Bacteriol , 179: 6416—6425
Sank ?aranarayanan R , Saxena P, Marathe UB et al. , 2004 . A novel tun-
nel in mycobacterial type II I polyketide synthase reveals the structural
basis for generating diverse metabolites [ J ] . Nat Struct Mol Biol ,
11 : 894—900
Saxena P, Yadav G, Mohanty D et al. , 2003 . A new family of type III
polyketide synthases in Mycobacterium tuberculosis [ J ] . J Biol
Chem, 278 : 44780—44790
Sesh ?imeY , Juvvadi PR , Fujii I et al. , 2005 . Discovery of anovel super-
family of type III polyketidesynthases in Aspergillusoryzae [ J ] . Bio-
chem Biophys Res Commun, 331: 253—260
Shao ?L , Qu XD, Jia XY et al. , 2006 . Cloning and characterization of a
bacterial iterative type I polyketide synthase gene encoding the 6-
methylsalicyclic acid synthase [ J ] . Biochem Biophys Res Commun,
345: 133—139
Shen ?B, 2000 . Biosynthesis of aromatic polyketides . Biosynthesis: Aro-
matic Polyketides, Isoprenoids [ J ] . Alkaloids, 209 : 1—51
Shen ?B , 2003 . Polyketide biosynthesis beyond the type I , I I and III
polyketide synthase paradigms [ J ] . Curr Opin Chem Biol , 7 :
285—295
Sher ?manDH , Smith JL , 2006 . Clearing the skies over modular polyketide
synthases [ J ] . ACS Chem Biol , 1 : 505—509
Shiao MS, Don HS, 1987 . Biosynthesis of mevinolin, a hypocholester-
olemic fungal metabolite, in Aspergillus terreus [ J ] . Proc Natl Sci
Counc Repub China B , 11 : 223—231
Sims JW, Fillmore JP, Warner DD et al. , 2005 . Equisetin biosynthesis
in Fusariumheterosporum [ J ] . ChemCommun ( Camb) , 186—188
Stas ?si DL , Kakavas SJ , Reynolds KA et al. , 1998 . Ethyl-substituted
erythromycin derivatives produced by directed metabolic engineering
[ J ] . Proc Natl Acad Sci USA , 95 : 7305—7309
Stau ?nton J , Weissman KJ , 2001 . Polyketide biosynthesis: a millennium
review [ J ] . Nat Prod Rep, 18 : 380—416
Tang L , Shah S, Chung L et al. , 2000 .Cloning and heterologous expres-
7723 期 ZHU Xiang-Cheng et al . : Biosynthesis and Genetic Engineering of Polyketides
sion of the epothilonegene cluster [ J ] . Science, 287: 640—642
Tang 1Y , Kim CY , Mathews II et al. , 2006 . The 2 . 7-Angstrom crystal
structureof a 194-kDa homodimeric fragment of the 6-deoxyerythro-
nolide B synthase [ J ] . Proc Natl Acad Sci USA, 103 : 11124—
11129
Tang 5Y , Lee TS, Kobayashi S et al. , 2003 . Ketosynthases in the initia-
tion and elongationmodulesof aromatic polyketidesynthases haveor-
thogonal acyl carrier protein specificity [ J ] . Biochemistry, 42 :
6588—6595
Tang 2Y , Tsai SC , Khosla C , 2003 . Polyketide chain length control by
chain length factor [ J ] . J AmChem Soc, 125: 12708—12709
Trop ?f S, Lanz T , Rensing SA et al. , 1994 . Evidence that stilbene syn-
thases have developed from chalcone synthases several times in the
courseof evolution [ J ] . J Mol Evol , 38 : 610—618
Tsai !SC , Miercke LJ , Krucinski J et al. , 2001 . Crystal structure of the
macrocycle-forming thioesterase domain of the erythromycin polyketide
synthase: versatility from aunique substrate channel [ J ] . Proc Natl
Acad Sci USA , 98 : 14808—14813
Tuck &er SL , Orbach MJ , 2007 . Agrobacterium-mediated transformation to
create an insertion library in Magnaporthe grisea [ J ] . Methods Mol
Biol , 354: 57—68
Van #Lanen SG, Shen B, 2008 . Advances in polyketide synthase structure
and function [ J ] . Curr Opin Drug Discov Devel , 11 : 186—195
Wals &h CT, 2004 . Polyketide and nonribosomal peptide antibiotics: modu-
larity and versatility [ J ] . Science, 303: 1805—1810
Wals &h CT, Gehring AM, Weinreb PH et al. , 1997 . Post-translational
modification of polyketide and nonribosomal peptide synthases [ J ] .
Curr Opin Chem Biol , 1 : 309—315
Wata +nabe A , Ebizuka Y , 2004 . Unprecedented mechanism of chain
length determination in fungal aromatic polyketide synthases [ J ] .
Chem Biol , 11 : 1101—1106
Watt ?anachaisaereekul S, Lantz AE, Nielsen ML et al. , 2007 . Optimiza-
tion of heterologous production of the polyketide 6-MSA in Saccharo-
myces cerevisiae [ J ] . Biotechnol Bioeng, 97 : 893—900
Weis &smanKJ , 2006 . Single amino acid substitutionsalter the efficiency of
docking in modular polyketide biosynthesis [ J ] . Chembiochem, 7 :
1334—1342
Weis &sman KJ , 2006 . The structural basis for docking in modular
polyketide biosynthesis [ J ] . Chembiochem, 7 : 485—494
Weis ?sman KJ , Hong H , Popovic B et al. , 2006 . Evidencefor aprotein-
protein interaction motif on an acyl carrier protein domain from a
modular polyketide synthase [ J ] . Chem Biol , 13 : 625—636
Weis ?smanKJ , Leadlay PF , 2005 . Combinatorial biosynthesis of reduced
polyketides [ J ] . Nat Rev Microbiol , 3 : 925—936
Xian ?g L , Moore BS, 2003 . Characterization of benzoyl coenzyme A bio-
synthesis genes in the enterocin-producing bacterium“ Streptomyces
maritimus”[ J ] . J Bacteriol , 185 : 399—404
Xue ?Y , ZhaoL , Liu HW et al. , 1998 . A gene cluster for macrolide anti-
biotic biosynthesis in Streptomyces venezuelae: architecture of meta-
bolic diversity [ J ] . Proc Natl Acad Sci USA, 95 : 12111—12116
YangG, Rose MS, Turgeon BG et al. , 1996 . A polyketide synthase is
required for fungal virulence and production of thepolyketideT-toxin
[ J ] . Plant Cell , 8 : 2139—2150
Ye J ?, Dickens ML , Plater R et al. , 1994 . Isolation and sequence analy-
sis of polyketide synthase genes from the daunomycin-producing
Streptomyces sp . strain C5 [ J ] . J Bacteriol , 176: 6270—6280
Yu F ?, Zhu X , Du L , 2005 . Developing a genetic system for functional
manipulations of FUM1 , apolyketide synthasegenefor the biosyn-
thesis of fumonisins in Fusariumverticillioides [ J ] . FEMS Microbiol
Lett, 248: 257—264
Yu T ?W, Shen Y , Doi-Katayama Y et al. , 1999 . Direct evidence that the
rifamycin polyketidesynthase assemblespolyketide chains processive-
ly [ J ] . Proc Natl Acad Sci USA , 96 : 9051—9056
Zhao ?C , J u J , Christenson SD et al. , 2006 . Utilization of the me-
thoxymalonyl-acyl carrier protein biosynthesis locus for cloning the
oxazolomycin biosynthetic gene cluster from Streptomyces albus
JA3453 [ J ] . J Bacteriol , 188: 4142—4147
Zhou ?P, FlorovaG, Reynolds KA , 1999 . Polyketide synthase acyl carrier
protein ( ACP) asa substrate and a catalyst for malonyl ACP biosyn-
thesis [ J ] . Chem Biol , 6 : 577—584
Zhu X , Yu F , Bojja RS et al. , 2006 . Functional replacement of the ke-
tosynthase domain of FUM1 for the biosynthesis of fumonisins, a
group of fungal reduced polyketides [ J ] . J Ind Microbiol Biotechnol ,
33 : 859—868
Zhu ?X , Yu F , Li XC et al. , 2007 . Production of dihydroisocoumarins in
Fusariumverticillioides by swapping ketosynthase domain of the fun-
gal iterative polyketide synthase Fum1pwith that of lovastatin diketide
synthase [ J ] . J AmChemSoc, 129 : 36—37
872 云 南 植 物 研 究 30 卷