Computational Tools For The Synthetic Design Of Biochemical Pathways

A key promise of synthetic biology is the possibility to customize the metabolic system of microorganisms

for

the

commercial

production

of

a wide

range

of

high-value

biofuels

1,2

or natural products

. Pathways for the production of alcohols, biodiesels, polyketides and terpenoids have successfully been constructed by introducing combinations of parts from various origins

into

a bacterial

host

that

is

easy

to

cultivate

3–5

. Potentially, entire metabolic pathways can be (re)

designed in silico and implemented in specialized host organisms

. Successes obtained in pioneering work on the antimalarial drug artemisinin

11–15

suggest that such approaches can be very fruitful. A biosynthetic pathway towards this compound was successfully engineered

in Saccharomyces

cerevisiae

and

Escherichia

coli

16–18

(BOX 1), and this pathway has the potential to enable much more cost-effective production of this important drug compared to the costly and laborious process of harvesting it from the source plant Artemisia annua.

The experimental work involved in engineering a synthetic pathway is considerable, and even systematically

planned

experiments

are

usually

accompanied

by

much

trial

and

error.

When

conceiving

the

design

of

a

novel

biosynthetic

pathway

(FIG. 1),

the

synthetic

biologist

has

to

find

optimal

solutions

for

selecting

pathways,

enzymes

or

host

organisms

from

an

abundance

of

possibilities.

In

this

Review,

we

explore

how

the

use

of

powerful

computational

tools

(TABLE 1)

can

lead

to

better-informed

and

more

rapid

design

and

implementation

of

novel

pathways,

and

we

propose

ways

in

which

tools

from

different

fields

of

computation

can

be

linked

together

effectively.

We

discuss

the

different

1,6–10

methodologies for identifying all possible metabolic pathways that can lead to the synthesis of a compound of choice, and how to rank these pathways based on various criteria. Subsequently, we consider how flux balance analysis of pathways can be applied to identify the most suitable candidate host organisms. We also examine how to effectively search sequence databases to obtain a list of candidate parts (such as genes and operons) for the execution of each step in the proposed

pathway.

Finally,

we

discuss

how

computational

methods

can

aid

in refactoring

these

parts

and

integrating

them

into

well-designed

transcriptional

units

that

are

optimized

for

a specific

host

organism.

For specific case studies and more detailed explanations

on

the

inner

workings

of

each

of

the

computational

methods,

we

refer

the

reader

to

a range

of

excellent

specialist

reviews

that

have

been

published

recently

15,19–21

.

Prediction and prioritization of possible pathways

For compounds of biotechnological value, often only a single specific biosynthetic pathway has been characterized.

The

key

promise

of

the

synthetic

biology

approach

to

pathway

design

is,

however,

that

one

does

not

remain

limited

to

biosynthetic

routes

that

already

exist

in nature.

Instead,

realistic

biosynthetic

pathways

can,

for

instance,

be

constructed

from

first

principles

to

optimize

their

thermodynamic

efficiency.

During the past decade, a range of computational pathway prediction algorithms has been generated that can aid in pathway (re)design. Some predictors focus on changing existing pathways through making

ave been built to identify possible metabolic pathways from first principles

23,24

22

on the basis of possible biotransformations

between

chemical

structures.

More

recently,

several

algorithms

have

been

constructed

that

use

more

complex

search

heuristics

to

find

and

rank

all

possible

pathways

that

lead

to

a desired

end

compound

(FIG. 1; TABLE 1).

10

Software for metabolic pathway identification and ranking. One accessible and user-friendly system for pathway identification is From Metabolite to Metabolite (FMM), a freely available web service through which one can search possible pathways between known input and output compounds

. It combines the KEGG maps and KEGG LIGAND information to form combined pathway

maps,

identifies

the

corresponding

genes

and

organisms

and

gives

an

output

in which

different

pathways

can

be

...