756 NATURE CHEMISTRY | VOL 6 | SEPTEMBER 2014 | www.nature.com/naturechemistry
news & views
C
hemists, biologists and engineers have
been developing high-throughput
screening (HTS) strategies for
synthetic molecules since the invention of
combinatorial chemistry by Geysen
1
and
Nielsen
2
during the 1980s and early 1990s.
All of these approaches have a similar
theme: (i) a randomized chemical library,
(ii) a molecular ‘bait-and-hook’ selection
strategy and (iii) an engineered screening
method to select for ‘hits’.
As well as a plethora of chemical systems
currently available, biological systems, for
example, yeast
3
and phage
4
display, have
also been co-opted and engineered for HTS
selection of protein binding molecules for
a diverse array of applications including
drug leads, novel enzymes and agents for
diagnostic assays. A popular HTS strategy
is the in vitro compartmentalization (IVC)
method
5
. Tis relies on the formation of
emulsion droplets in microfuidic systems
that capture molecular substrates so that
each droplet is efectively a single reaction
capsule. Te droplet formation can be
tuned — by adjusting the fow rates of
the emulsion components — so that a
substrate can be encased with a single
molecule of enzyme in a single droplet
(the content of droplets follows a Poisson
distribution). Screening for enzyme activity
can then be performed using modifed
fuorescence-activated cell sorting (FACS)
that can enable ultra-high-throughput
screening (UHTS). One of the attractive
benefts of IVC is the potential ability to
couple genotype and phenotype — that is,
to maintain the connection between the
genetic content (DNA sequence) and the
observable characteristics (for example,
enzyme activity). Tis is, however, a major
limitation of emulsion droplet systems
because once the emulsion is broken it
is difcult to track genotypes. Various
labelled DNA, afnity capture or proxy
barcoding strategies have been developed to
circumvent this limitation
6–9
. Briefy, these
strategies are secondary ‘selection’ methods
that indirectly report on the genotype of the
newly evolved protein.
Now, writing in Nature Chemistry,
Florian Hollfelder and co-workers have
described
10
a convenient and low-cost
approach that allows simultaneous
purifcation of both an enzyme and
the DNA that encodes it. Tey use this
system to perform a high-throughput
evolution of a phosphotriesterase enzyme
that can be used in the detoxifcation of
pesticides or nerve agents. Teir “all-
in-one” strategy (Fig. 1) relies on the
generation of “biomimetic gel-shell beads”
for IVC. In a similar fashion to other
droplet formation/FACS strategies, their
method frst generates emulsion droplets
so that each droplet encapsulates a single
bacterium carrying a variant of the enzyme
(at both the protein and DNA levels),
enzyme substrate, low-melting agarose,
lysis bufer and an alginate polyanion.
Lysis of the trapped bacteria followed by
solidifcation of the agarose in the presence
of poly(allylamine-hydrochloride) results
in the formation of a porous agarose gel-
bead containing the released enzyme, the
encoding DNA and the enzyme substrate.
Tus both the phenotype and encoding
genotype (an expression plasmid) are
trapped in a convenient bead-like substrate
amenable to a standard FACS process.
Afer identifying beads that contain
active enzymes, the encoding DNA can
ENZYME CATALYSIS
Evolution made easy
Directed evolution is a powerful tool for the development of improved enzyme catalysts. Now, a method that
enables an enzyme, its encoding DNA and a fluorescent reaction product to be encapsulated in a gel bead enables
the application of directed evolution in an ultra-high-throughput format.
Eugene J. H. Wee and Matt Trau
1
2
3 4
5
6
Expression
in E. coli
Droplet formation,
cell lysis
Catalysis
Selection by FACS
Gel-shell bead
formation
Genotype
recovery
Plasmid Enzyme Substrate (with tag) Product (with tag)
Figure 1 | Directed evolution of enzymes in gel-shell beads
10
. A library of enzymes is created by expression
within Escherichia coli carrying the encoding plasmid DNA. The bacteria are encapsulated in droplets so
that each droplet contains a single enzyme and the DNA that encodes it. Cell lysis releases the enzyme so
that it can be tested as a catalyst with the activity measured using fluorescence. Gel-shell bead formation
traps the enzyme with its encoding DNA and the fluorescent reaction product so that active enzymes can
be identified in a high-throughput fashion for analysis of the DNA sequence, and potentially for further
rounds of enzyme evolution.
© 2014 Macmillan Publishers Limited. All rights reserved