Transcript Lecture 15, engineered biosynthesis

Reading
“Harnessing the biosynthetic code...” pp 63- 68
“Multiple genetic modifications of the erythromycin polyketide
synthase to produce a library of novel “un-natural” natural
products” pp 1846-1851
Polyketide Biosynthesis
Many bioactive natural products are polyketides, polymers of acetate
or other small, oxygenated organic molecules (like propionate, a C3)
Includes antibiotics (erythromycin, tetracycline), anticancer drugs
(daunomycin), immunosuppresants used after transplants (rapamycin)
polypropionate chain
(imaginary precursor)
the erythromycin aglycone core
(missing attached sugars)
Polyketide Biosynthetic Enzymes
Polyketides are made by polyketide synthases (PKS’s),
huge multi-functional enzymes that act like production lines
PKS proteins are organized into linear modules
In turn, each successive module:
- adds another unit to the chain (elongation)
- makes modifications to that piece of the chain
- hands the chain off to the next module
Function like big molecular assembly lines
Erythromycin Biosynthesis
The macrocyclic core of erythromycin is made by a 6-module
polyketide synthase called DEBS, producing the lactone 6-dEB
Complete synthase is 10,283 amino acids long !
- 3 huge subunits, each containing 2 modules
- each module has 3-6 domains, or catalytic sites
A Word About Precursors
either can be the
initial piece,
or primer
can be
decarboxylated
adds 2-carbon
units to skeleton
plus methyl groups
methyl can have either
stereochemistry
PKS enzymes: Modularity + Domains
Each module consists of a string of catalytic domains
- different domains carry out distinct types of reactions
EVERY module has 3 core domains
(1) ketosynthase (KS)
(2) acyl transferase (AT)
(3) acyl carrier protein (ACP)
together, add 1 block to the
growing chain
PKS enzymes I: 3 core domains
EVERY module has 3 core domains
(1) Ketosynthase (KS)
- Accepts polyketide from ACP domain of previous module
- Polyketide chain is bound via thioester to Cys -S(2) Acyl transferase (AT)
- Determines which extender unit gets incorporated next
(acetate C2, propionate C3, methyl malonate C4)
(3) Acyl carrier protein (ACP)
- Condenses chain w/ next extender unit, bound as a thioester
- Chain is attached via flexible phosphopantetheine linker,
ready for the hand-off to the next module
thioester
decarboxylation makes
the a-carbon of malonate
a good nucleophile
the flexible linker
phosphopantetheine
allows one ACP to pass
growing chain on to the
next module
decarboxylation makes
the a-carbon of malonate
a good nucleophile
the flexible linker
phosphopantetheine
allows one ACP to pass
growing chain on to the
next module
skeleton extended by 2-carbon backbone unit (plus -CH3)
Polyketide Biosynthetic Enzymes
Each cycle adds a 2-carbon extension to chain, introducing
a b-keto group and a possible side chain (depending on choice
of extender unit by AT domain)
Each b-keto group then undergoes none, some, or all in a series
of optional reduction steps
PKS enzymes II: optional domains
In addition to 3 core domains, each module contains 0-3
optional domains that determine how much the b-keto group
added by the previous module gets reduced
(4) Ketoreductase (KR) -
reduces
(5) Dehydratase (DH) -
reduces
(6) Enoyl reductase (ER) - reduces
Polyketide Biosynthetic Enzymes
In addition to 3 core domains, each module contains 0-3
optional domains that determine how much the b-keto group
added by the previous module gets reduced
A given module will have:
none
KR
KR + DH
KR + DH + ER
- optional domains
control the extent of
oxidation throughout
the mature polyketide
- reductions are done
as you go, not after the
chain is complete
Closing the Macrocycle Ring
Final cyclization is done by the terminal
thioesterase domain (TE)
- Catalyzes formation of the lactone
ring of erythromycin
This reaction also proceeds
spontaneously, but very slowly
Post-cyclization Modifications
Final modifications of 6-dEB are made by downstream enzymes
that oxidize C6 and glycosylate (add sugars to) C3 + C5
6-deoxy-erythronolide B
(6-dEB)
Erythromycin
Such post-PKS enzymes are often found in nearby gene clusters
Polyketide Diversity & Biosynthesis
The tremendous structural variation found among natural
polyketides stems from differences in:
(1) choice of starter unit (the “handle” at 1 end of the molecule)
(2) choice of extender units (structure + stereochemistry)
(3) overall chain length (# of modules)
(4) extent of b-keto modification (type of optional domains)
(5) regiospecific cyclizations (action of terminal TE domain)
(6) downstream (post-PKS) enzymatic modifications
[for example, adding sugars]
Engineered Biosynthesis
Knowing the function of domains from different modules,
and from entirely different organisms, can we use genetic tools
to engineer new PKS enzymes?
- How much can you alter the sequence of domains & modules
and still have a functional enzyme? Are domains really
independent of each other?
Can we now custom-tailor new polyketides, built to order, by
putting together the correct sequence of domains into a
recombinant PKS enzyme?
Deletion of Modules from DEBS
add TE =
more product
Deletion of Modules from DEBS
Results showed that domains from 1 module could be fused to
domains from another module and produce a functional PKS
- TE domain is “flexible” enough to recognize much shorter
chains than its normal substrate
- TE domain improves enzyme turnover (rate of production)
Deletion of Modules from DEBS
Results showed that domains from 1 module could be fused to
domains from another module and produce a functional PKS
- TE domain is “flexible” enough to recognize much shorter
chains than its normal substrate
- Domains can be rearranged without loss of activity
Next: can you tack on domains from other modular enzymes?
Module Swapping
Replaced DEBS “loading” module (which uses propionate) with
Non-Ribosomal Peptide loading domain of rifamycin synthase
NRPS loading domain uses benzoic acid as a
starting block to prime rifamycin synthesis, not propionate
- the fusion protein incorporated benzoate into the expected
derivative of 6-dEB, w/ benzene ring in place of ethyl chain
Engineering 6-dEB Derivatives
Alter the domains in DEBS Module 2, which controls blue area:
6-dEB
1
2
3
(1) replace DEBS AT domain w/ the AT from
rapamycin PKS module 2, which uses
malonyl-CoA instead of methyl-malonyl
- as predicted, product is missing the methyl
group normally found at this position
Engineering 6-dEB Derivatives
Alter the domains in DEBS Module 2, which controls blue area:
6-dEB
1
2
3
(2) replace KR domain w/ rap KR/DH
from module 4, to reduce the -OH
- alcohol moiety replaced w/ alkene carbon
Engineering 6-dEB Derivatives
Alter the domains in DEBS Module 2, which controls blue area:
6-dEB
1
2
3
(3) replace KR domain w/ rap KR/DH/ER
from module 4, to fully reduce the alkene
- alcohol in 6-dEB replaced w/ alkane carbon
in the engineered product
Engineering 6-dEB Derivatives
1
2
3
4
5
(4) combinatorial replacement:
- replace AT domain w/ rap AT from module 2
- replace KR domain w/ rap KR/DH from module 4
- product is missing the methyl group and has the alkene
Engineering 6-dEB Derivatives
1
2
3
4
5
(5) combinatorial replacement:
- replace AT domain w/ rap AT from module 2
- replace KR domain w/ rap KR/DH/ER from module 4
- product is missing the methyl group and has the alkane
Library of “unnatural”
natural products made
by combinatorial
biosynthesis
By subbing 5 alternative
cassettes into a scaffold
of 6 modules, produced
>100 macrolides
Most could be converted
to erythromycin analogs
by post-PKS modifier
enzymes
McDaniel et al. 1999, PNAS 1846-1851
Problems for Genetic Manipulation
The DEBS PKS is made by Saccharopolyspora erthraea, which
is not a genetically well-understood or easily cultured bacteria
- genes are typically cloned onto plasmids in E. coli,
DNA circles made for easy insertion of pieces of DNA,
that will express cloned protein in bacterial cultures
Problems for working with PKS genes:
(1) DEBS proteins are so huge, they don’t always fold correctly
in E. coli
(2) E. coli lacks the appropriate accessory enzymes
- missing metabolic precursors (2-methyl-malonate)
- no downstream modifying enzymes (glycosylases)
Problems for Genetic Manipulation
How do you perform complex genetic rearrangements and end
up with an easily cultured organism to grow in mass quantities?
Is there a way to incorporate genes from bacteria that have never
been cultivated or studied genetically?
- Many bacteria, especially marine species and
“extreme-ophiles”, cannot be cultured
Host Issues: Solution 1
1994 (Science 265: 509-512)
- Make plasmids in E. coli, where genetic manipulation is easy
- Then move plasmids into genetically less tractable host,
a strain of the PK-producing Streptomyces coelicolor
- Not nearly as handy as E. coli, but possesses the accessory
enzymes and precursors needed for macrolactone biosynthesis
- Produces sizeable amounts of polyketides according to the
PKS sequences found on the engineered plasmids
Host Issues: Solution 2
1995 (Nature 378: 263-266)
- Cell-free expression system: purify high-mol. weight
PKS enzymes from S. coelicolor homogenates
- Shown in vitro that the enzymes carry out polyketide synthesis
- In fact, proteins were able to incorporate various unnatural
substrates into polyketide chains, suggesting great flexibility
of these enzymes for substrate recognition
Host Issues: Solution 3
2001 (Science 291: 1790-1792)
- Made a metabolically engineered strain of E. coli that can
express and fold PKS proteins + produce correct precursors
- Move PKS genes off of plasmids, onto the E. coli chromosome
- Engineered E. coli produced amounts of 6-dEB comparable
to the native bacterium
- Shows sophistication of genetic control, flexibility of natural
biosynthetic pathways to rational manipulation