application of enzymes and microorganisms for organic ......metabolic engineering: introducing new...

1

Application of Enzymes and Microorganisms for Organic

Synthesis

Jay Keasling

TaxolHO

O

OHCH3

O

O O

CH3 CH3O

O

CH3

OCH3O

H3CHN

O

HO

O O

2

Taxol is extracted from the Pacific Yew•The Pacific yew tree is an environmentally protected species and one of the slowest growing trees in the world.

•Isolation of the compound, which is contained in the bark, involves killing the tree.

•One 100-year old tree results in approximately 350 mg of taxol, just enough for one dose for a single cancer patient.

Total synthesis of taxol

3

Metabolic engineering: introducing new pathways in cells

Synthesis of complex molecules

• May require several enzymes from one or more organisms

• Expression of genes must be balanced– Underexpression of any one gene may limit flux

through the pathway and therefore product yields– Overexpression will lead to inefficiencies

• Precursors (from inside the cell or supplied from outside the cell) should not severely limit production of the desired product

4

Balancing enzymatic reactions in the cell

C

A

BX Y1 Z

Enzyme 1 Enzyme 3

Enzyme 2

Enzyme 4

Y2

Over-produced enzyme

Under-produced enzyme

What is metabolic engineering?• Metabolic engineering is a redirection of

enzymatically-catalyzed reactions for the production of a new compound or the degradation of a compound

– genetic modification of a single organism

– engineering a consortium of organisms

5

Why do metabolic engineering?• Production of novel compounds

– new biopolymers– antibiotics

• Production of existing compounds in better ways

• Bioremediation of recalcitrant compounds– pesticides/nerve agents– PCBs

HOO

OHCH3

O

O O

CH3 CH3O

O

CH3

OCH3O

H3CHN

O

HO

O

12

34 5

614

13

1211

109

8 715 16

17

18

19

20

O

Taxol has a diterpene

core

6

OPO(OH)OP(OH)2O

Geranylgeranyl diphosphate (GGPP)

Terpene metabolic pathways

OPO(OH)OP(OH)2O

Farnesyl diphosphate (FPP)

OPO(OH)OP(OH)2O

OPO(OH)OP(OH)2O

Dimethylallyl diphosphate(DMAPP)

Geranyl diphosphate (GPP)

OPO(OH)OP(OH)2O

Isopentenyl diphosphate (IPP)

Diterpenes

Monoterpenes

Sesquiterpenes

Carotenoids

In bacteria, IPP is produced via the non-mevalonate pathway

H3C CO2H

O

OHC

OH O OH

H3C

OH

OH OH

HO

1-deoxyxylulose-5-P (DXP)

OP(OH)2O OP(OH)2O

OP(OH)2O

Pyruvate glyceraldehyde-3-P

2-C-methyl-D-erythritol-4-P (MEP)

DXP synthase

DXP reductoisomerase

OPO(OH)OP(OH)2O

Isopentenyl diphosphate (IPP)

7

Metabolic engineering of carotenoid production

Pyr + G3P IPP

FPP

dxs dxrDMAPP

idi

0

5

1015

20

25

3035

Car

oten

oid

prod

uctio

n(μ

g/m

L/O

D60

0)

crtEBIdsx,idi

crtEBIdxs Carotenoids

crtEBI

crtEBI

dxs and dxr under Ptac control on high-copy cloning vectors

Pconst crtE crtI crtY

High-copyplasmid

Medium-copyplasmid

Ptac dxs

8

Overexpression of dxs from a high-copy plasmid with a strong promoter

01234567

0

Car

oten

oid

(μg/

ml)

Cell G

rowth (O

D570 )0

1234567

0.3 0.6 0.9IPTG concentration (mM)

Overexpression of dxs is a

metabolic burden on cell growth

Glc

G6P

Pyr

AcCoA

CIT

ICT

OGT

SUC

MAL

OAA

GOX

G3PDHAPDXP

Secretion ??

IPP GGPP

Carotenoids

(-)

9

Needs for Metabolic Engineering

• Optimal fluxes through the heterologous metabolic pathways

• Strict control over gene expression

• Consistent control of gene expression in all cells

• Minimal burden of the heterologous genes on the host

Some key problems in gene expression control

• Many expression vectors are unstable and have variable copy number in the host

• Many promoters do not allow tight and consistent control of gene expression

• There are few techniques to regulate expression of multiple heterologous genes

• Can we predict the levels of gene expression needed for flux redistribution?

10

Gene expression tools for metabolic engineering:

A Metabolic Engineering Toolbox

Gene expression tools for metabolic engineeringExpression

vector

11

Bacterial Artificial Chromosome (BAC)

oriV oriS parTn1000 flmccd

rep FIB rep FIA5 34 43 45 48 64 kb

Native F plasmid of Escherichia coli

EEE B P H

F plasmid

0

25

50

75

A partition element ensures that each daughter cell will receive a BAC

oriV oriS parccd

EE B P H43 45 48

Cell

Division

12

Specific replication origins time BAC replication with the cell cycle

oriV oriS

EE B P H43 45 48

parccd

A “Kill” element ensures that BAC-free daughter cells do not survive

oriV oriS parccd

EE B P H43 45 48

Cell

Division

13

BACs are stable indefinitely in the absence of selection pressure

0 40 80 120 1600

0.2

0.4

0.6

0.8

1

Culture time (generations)

Frac

tion

plas

mid

-bea

ring

ce

llsGene expression

induced Not induced

The auxiliary chromosomes have improved control of gene expression

Uninduced

Expres

sion

Induced

Expres

sion

Growth R

ate

of Host

BAC

High-CopyPlasmid

0.69 hr-1

0.53 hr-1

15 units

200 units

4,000 units

12,500 units

14

dxs and dxr under PBAD control on bacterial artificial chromosome

Pconst crtE crtI crtY

Bacterialartificial

chromosome

Medium-copyplasmid

PBAD

araC

dxs

Production of lycopene: dxs under control of PBAD on low-copy plasmid

Cel

l gro

wth

(OD

600n

m)

Lyc

open

e (m

g/l)

Arabinose concentration (mM)

0.0

2.0

4.0

6.0

8.0

10.0

0 0.013 0.133 1.33 13.3

15

Gene expression tools for metabolic engineeringReproducible

promotercontrol

inside

outside

P BADgfp

P Cara

C

araE

P araE

Green fluorescent protein

Experimental system

Chromosome

Plasmid

arabinose

arabinose

16

Expression of gfp from the arabinose-inducible promoter

100

1000

10000

100000

0.00001 0.0001 0.001 0.01 0.1 1 10

Arabinose (wt %)

Fluo

resc

ence

/OD

600

Desired gene expression in population

Induction/cell

Freq

uenc

yin

pop

ulat

ion

Induction/cellInduction/cell

100% of cells

are half induced

Gen

eex

pres

sion

Inducer concentration

17

‘All-or-None’ Gene Expression

Induction/cell

Freq

uenc

yin

pop

ulat

ion

Induction/cellInduction/cell

50% are fully induced

50% are uninduced

Gen

eex

pres

sion

Inducer concentration

Flow cytometric analysis

Fluorescence

Freq

uenc

y

Fluorescencedetector

FALS sensor

Laser

18

Native arabinose-inducible system gives rise to two populations

Increasing inducer concentration

Fluorescence intensity

Native arabinose-inducible system gives rise to two populations

Fluorescence intensity

Uninduced cellsInduced cells Tim

e afte

r inducti

on [h

]

01

23

46

Cel

l num

ber

19

Population dynamics as a function of arabinose concentration

0

20

40

60

80

100

0 2 4 6

Time [h]

Indu

ced

cells

[%]

0.0002%

0.002%

0.02%

0.2%

Experiment:steady-state induction levels as a function of

inducer concentration

0

20

40

60

80

100

Indu

ced

cells

[%]

0

2000

4000

6000

8000

-5 -4 -3 -2 -1 0 1Arabinose concentration [log %]

Fluo

rese

nce/

OD

20

inside

outside

P BADgfp

P Cara

C

araE

P araE

Stimulation of inducer transport gene expression

arabinose

arabinose

Green fluorescent protein

arabinose

inside

outside

P BADgfp P C

araC

araE

P con

arabinose

Decoupled transporter/reporter

system

21

Population analysis of E. coliexpressing gfp

Time a

fter i

nduction

[h]

01

34

6

Fluorescence intensity

Uninduced cellsInduced cells

2

Cel

l num

ber

Population analysis of E. coliexpressing gfp

Increasing inducer concentration

Fluorescence intensity

22

Experiment: Steady-state

induction levels as a function of

inducer concentration

0

20

40

60

80

100

Indu

ced

cells

[%]

0

2000

4000

6000

8000

-5 -4 -3 -2 -1 0 1Arabinose concentration [log %]

Fluo

rese

nce/

OD

Native system

Engineered system

inside

outside

P BADGen

e of

interest

P Cara

C

araE

P araE

Native system

Chromosome

Plasmid

Protein of interest

arabinose

arabinose

ON

OFF

23

inside

outside

P BAD

P Cara

Cara

EP con

Engineered system

Protein of interest

Gene o

f

interest

arabinose

arabinose

Low

arabinose

Medium

arabinosearabinose

High

Gene expression tools for metabolic engineering

Expressionof multiplegenes

24

Balancing enzymatic reactions in the cell

C

A

BX Y1 Z

Enzyme 1 Enzyme 3

Enzyme 2

Enzyme 4

Y2

gene 3

gene 4

gene 2

gene 1

Using individual control elements

C

A

BX Y1 Z

Enzyme 1 Enzyme 3

Enzyme 2

Enzyme 4

Y2

gene 4

P4

P1

gene 1P3

gene 3 P2

gene 2

25

Synthetic operons

C

A

BX Y1 Z

Enzyme 1 Enzyme 3

Enzyme 2

Enzyme 4

Y2

gene 3 gene 4

P

gene 2 gene 1

mRNA

DNA

C

A

BX Y1 Z

Enzyme 1 Enzyme 3

Enzyme 2

Enzyme 4

Y2

mRNA

RNase

26

The puf operon in Rhodobacterencodes a multi-subunit enzyme

Q B A L M X

? ?

RC

LHIMulti-subunitenzyme

The production of enzyme subunits is controlled by mRNA stability

DNA

mRNA

Black arrows indicate RNase E cleavages

27

mRNA is inactivated by a cleavage inside the coding region

endoribonuclease

5’ RBSmRNA

ribosome

cleavage site

exonuclease

Secondary structures in the mRNA protect natural mRNAs against

nucleases

RNase Eendonuclease

exonucleaseRBS

ribosome

28

tccatacgtcgacggtaccgtattttggatgataacgaggcgcaaaaaatgaggtatgcagctgccatggcataaaacctactattgctccgcgttttttac

A cassette system to design mRNA stability

lacZSal I Asp718Insertion of hairpin cassette

tccatacgtcgacttatctcgagtgagatattgttgacggtaccgtattttggatgataacgaggcgcaaaaaatgaggtatgcagctgaatagagctcactctataacaactgccatggcataaaacctactattgctccgcgttttttac

Transcription

gguaccguauuuuggaugauaacgaggcgcaaaaaug

ga

ac

c

c

c

c

c

u

uuu

uu

g

gg

g

g

g

a

a

a

a

a

u

u

u

u

u

ga

ggagtcgacttatctcgagtgagatattgttgacggtaccccgcctcagctgaatagagctcactctataacaactgccatggggc

Sal I Asp718

A family of synthetic hairpins

acgucgacagguaccguauuuut1/2 = 2.6 min

pTC40

cc

c

c

uu

uu

g

g

g

a

a

u

gaccu

gggau

u

a

ga

gguaccguauuuu

t1/2 = 4.9 min

pHP14

cc

c

c

auu

g

g

gau

u

a

ga

gguaccguauuuu

u

u

ua

g

cc

a

a

g

u

cu u

a

u a

t1/2 = 2.1 min

pHP15t1/2 = 6.1 min

c

c

c

c

c

c

u

uuu

uu

g

gg

g

g

a

aaa

a

uu

u

u

a

ga

ggau

gguaccguauuuu

c gu a

pHP8

c

c

c

c

c

c

u

uauu

g

g ggg

g

g

a

a

aaa

a

uu

u

u

u

a

ga

gguaccguauuuu

t1/2 = 6.8 min

pHP10

cc

c

c

ua

uu

g

g

g

a

a

u

ccu

u

g ggg

a

a

aau

u

u

u

a

ga

gguaccguauuuu

t1/2 = 5.5 min

pHP9

c

c

c

c

c

c

u

uuu

uu

g

g ggg

g

g

a

a

aaa

a

uu

u

u

u

a

ga

gguaccguauuuu

t1/2 = 8.3 min

pHP4

t1/2 = 19.8 min

pHP17

ac

c

c

c

a

uu

g

g

gau

u

ga

gguaccguauuuu

g

c

ga

g

au

c

c

a

a

ug u

a

a t

uuu

uau acu ga

gc

t1/2 = 12.5 min

pHP16

c

c

c

c

c

c

u

uuu

uu

g

gg

g

g

g

g

a

a

a

aa

a

uu

c

u

u

a

ga

gguaccguauuuu

cuucu

g cagag

uu

u

a

29

A strategy to design mRNA stability?

cc

c

c

auu

g

g

gau

u

a

ga

gguaccguauuuu

u

uua

g

cc

aa

g

u

cuu

aua

c

c

c

c

c

c

u

uuu

uu

g

gg

g

g

a

aaa

a

uu

u

u

a

ga

ggau

gguaccguauuuu

c gu a

c

c

c

c

c

c

u

uuu

uu

g

g

g

g

a

gaa

a

uu

u

u

a

ga

ggcc

gguaccguauuuu

aa

g

caau

uac

aucgcag

u

0

5

10

15

20

5 10 15 20 25

mR

NA

hal

f-lif

e (m

in)

-ΔGfolding

(kcal/mol)

Degradation of multicistronic mRNA

30

Degradation of multicistronic mRNA

Constructs to test operons

pGL

pGH1E1H4L

pGH1E2H4L

gfp lacZ3'5'

gfp lacZ3'5'

RNase E site 1

gfp lacZ3'5'

RNase E site 2

H4H1

31

Relative protein levels/mRNA stabilities

0

1

2

3

4

pGL

pGH1E1H4LpGH1E2H4L

β-G

al/G

fp

Relative enzyme activities

Reversing gene order

pLG

pLH1E1H4G

pLH1E2H4G

lacZ gfp3'5'

lacZ gfp3'5'

RNase E site 2

lacZ gfp3'5'

RNase E site 1

H4H1

32

Relative protein levels/mRNA stabilities

0.1

1

10

100

1000

pLG

pLH1E1H4GpLH1E2H4G

β-G

al/G

fp

Relative enzyme activities

A synthetic operon for carotenoid production

Phytoene Lycopene β-Carotene

p70yHPxicrtY crtI 3'5'

RNase E site

HPxHP

CrtE CrtI CrtY

33

No hairpin 5’ of crtI

Phytoene Lycopene β-Carotene

p70yicrtY crtI

3'5'

RNase E site

HP

CrtE CrtI CrtY

Carotenoid Production

0

100

200

300

p70yi + pAC-PHYT

Rel

ativ

e Pr

oduc

tion

(Pea

k A

rea/

OD

600)

Phytoene

β-Carotene

Lycopene

34

crtY crtI3'5'

HP

Phytoene Lycopene β-CaroteneCrtE CrtI CrtY

3'5'

3'5'

A strong hairpin 5’ of crtI

Phytoene Lycopene β-Carotene

p70yHP4icrtY crtI 3'5'

RNase E site

HP4HP

CrtE CrtI CrtY

35

Carotenoid Production

0

100

200

300

p70yi + pAC-PHYT

p70yHP4i +pAC-PHYT

Rel

ativ

e Pr

oduc

tion

(Pea

k A

rea/

OD

600)

Phytoene

β-Carotene

Lycopene

crtY crtI3'5'

HP

Phytoene Lycopene β-CaroteneCrtE CrtI CrtY

3'5'

5' 3'

HP4

36

A stronger hairpin 5’ of crtI

Phytoene Lycopene β-Carotene

p70yHP16icrtY crtI 3'5'

RNase E site

HP16HP

CrtE CrtI CrtY

Carotenoid Production

0

100

200

300

p70yi + pAC-PHYT

p70yHP4i +pAC-PHYT

p70yHP16i +pAC-PHYT

Rel

ativ

e Pr

oduc

tion

(Pea

k A

rea/

OD

600)

Phytoene

β-Carotene

Lycopene

37

crtY crtI3'5'

HP

Phytoene Lycopene β-Carotene

CrtE CrtI CrtY

3'5'

HP16

5' 3'

A weak hairpin 5’ of crtI

Phytoene Lycopene β-Carotene

p70yHP17icrtY crtI 3'5'

RNase E site

HP17HP

CrtE CrtI CrtY

38

Carotenoid Production

0

100

200

300

p70yi + pAC-PHYT

p70yHP4i +pAC-PHYT

p70yHP16i +pAC-PHYT

p70yHP17i +pAC-PHYT

Rel

ativ

e Pr

oduc

tion

(Pea

k A

rea/

OD

600)

Phytoene

β-Carotene

Lycopene

crtY crtI3'5'

HP

Phytoene Lycopene β-CaroteneCrtE CrtI CrtY

3'5'

HP17

5' 3'

39

Gene expression tools for metabolic engineering

Metabolic models toquantify

flux

How do you coordinate

the expression of multiple

genes?

40

Mass Balance on Cellular MetabolitesdXdt

= S ν - b where

X = Concentration of metabolitesS = Stoichiometric matrix

-> known enzymatic reactionsb = Uptake, secretion, and biomass synthesis

-> known from cell compositionν = Reaction flux vector

-> unknown

Mass Balance on Cellular Metabolites

dXdt

= S ν - b

GenomicsProteomicsPhysiomics

41

=S ν bnumber of

fluxes(495)

number of metabolites

(289)>

Objective Function: Z = ΣConstraints:

LowerBound

UpperBound< νi < i = 1, 2, ...

S ν = b1.2.

Solve for fluxes using linear optimization

At steady-state:

Linear Optimization:

ci νii

Predicted fluxes for growth on glucose and acetate

Experimental DataWalsh & Koshland (1985)

Model Predictions

7.260.0

2.91

CELL CONSTITUENTS

0.92

4.62

4.62

0.0

4.62

1.14

3.77

4.64

4.62

1.75

1.73

FATTY ACIDS

ACETATE

PYR

CIT

CELLCONSTITUENTS

GLX

OA

ISOCIT

SUCC

7.00GLUCOSE

0.81

5.00

5.00

1.13

3.87

0.00.0

3.87

3.87

1.78 0.81

5.0

5.00

0.81

MAL

AcCoA

PEP

-KGαFUM

42

Some environmental examples of metabolic engineering

Application of metabolic engineering to biodesulfurization of fossil fuels

• Dibenzothiophene (DBT) is typical of the organic sulfur compounds found in fossil fuels.

• DBT is recalcitrant to hydrodesulfurization.• Used extensively in biodesulfurization studies

S

DBT (dibenzothiophene)

43

Rhodococcus erythropolis IGTS8 desulfurization pathway

S S

O

O2 O2

DBT (dibenzothiophene)

DszC,FMNH2

Dsz C,FMNH2 S

O O

O2 DszA,FMNH2

S

O

O

HOHO

Dsz B

HBP (2-hydroxybiphenyl)

+SO32-

DBT-5-oxide DBT-5,5-dioxide

2-(2' hydroxyphenol)benzene sulfinate

Role of reducing equivalents

• DszA & DszC require reduced flavin (FMNH2).

NADH NAD

FMN FMNH2

O2 H2O

DBT DBTO

N

N

OHOH

OH

H O

NH

ON

H

OPO3 DSZ

FMN reductase

44

Comparison of degradation rates

Organism Rate(mg/hr/g dcw)

Source

E. coli pRED/pDSZ

51 This work

Rhodococcus erythropolis IGTS8

3 Gallardo (1997)

Rhodococcus erythropolis H-2

5 Oshiro (1996)

Relieving bottlenecks in the desulfurization pathway

S S

O

O2 O2

DBT (dibenzothiophene)

Dsz C,FMNH2

Dsz C,FMNH2 S

O O

O2 Dsz A,FMNH2

S

O

O

HOHO

Dsz B

HBP (2-hydroxybiphenyl)

+SO32-

DBT-5-oxide DBT-5,5-dioxide

2-(2' hydroxyphenol)benzene sulfinate

45

Overlapping reading frames in dszABmay limit flux through the pathway

dszB dszCdszA

DBT DBTO2HBPS HBP

dszB dszCdszA

DBT DBTO2 HBPS HBP

Aerobic precipitation of cadmium

Cys

Cysteinedesulfhydrase

NH3+ pyruvate+

Cd2+ CdS

S2-

46

Native cysteine biosynthesis pathway

SO42-

Serineacetyltransferase

(SAT)

Cysteine

(-)

Serine +Acetyl-CoA

O-Acetyl-Serine

(+)

S(II)

Deregulated cysteine biosynthesis pathway

SO42-

Serineacetyltransferase

(SAT)

Cysteine

Serine +Acetyl-CoA

O-Acetyl-Serine

(+)

S(II)

47

Cysteine production is inhibited by cadmium

0

100

200

300

400

500

0 50 100 150

Cd (μM)

Cys

tein

e ( μ

M)

Control

Cys production(SAT*)

Expression of cysE* and cysDfor aerobic cadmium precipitation

CysCysE*

S2-

Cd2+

CdS

cysE*Pcon

PyruvateNH3

CysD

cysDPtrclacIQ

48

Precipitation of CdS by the engineered strain

Control Engineered strain

CdS

Cadmium precipitates on the cell wall as cadmium sulfide

50 10

Energy(keV)

Cou

nts

200

150

100

50

0

Cu

Os

Os

Cl

Cl

S

Cou

nts

200

150

100

50

0

S

S

Cu

Cu

Os

Os

Cd

Cd

Cd

Ca

CaCa

Cl

Cl

Cu

0 5 10

Energy (keV)

a b

c d

CuCu

49

Acknowledgements

Funding– National Science Foundation– Office of Naval Research– DARPA– Department of Energy– National Institutes of Health– Maxygen– Merck

Graduate StudentsJaya PramanikDoug BoleschRobert PapeNatalya EliashbergStephen Van DienSalomeh KeyhaniTrent A. Carrier

Kristala JonesR. Brent NielsenPiper TrelstadClifford WangIlana AldorDavid ReichmuthJessica Hittle

Andrew WalkerNichole GoedenNeil RenningerChristina SmolkeSundiep TeharaStacie CowanCynthia Gong

Michel MaharbizDavid Lubertozzi Katherine McMahonDoug PiteraKai WangSydnor WithersLance KizerBrian FrushourBrian Pfleger

Post-docsSusan T. SharfsteinWubin PanSang-Weon BangEric GilbertArtem KhlebnikovYet-Pole ISeon-Won KimVincent MartinGuangyi Wang