Biology Reference
In-Depth Information
positions diversified, the NDT library provided a significantly higher frequency of positive
variants (the protein function was epoxide hydrolysis by a hydrolase). This study
demonstrates that given limited library screening capabilities, functional redundancy
in the genetic code can be a limiting factor.
SCREENING AND SELECTION
The speed and accuracy of a screening platform typically dictate both the mutagenesis
strategy chosen, as well as the number of variants screened. Maximum attainable protein
library sizes are considered limited by cloning efficiency (ligation and transformation) at
B
10
10
members. For cell-free expression or in vitro display systems (see below), since
no transformation is required, protein libraries of up to 10
14
members can be constructed.
48
As outlined below, the throughput of screening systems requiring transformation can vary
from
10
9
10
9
, and perhaps 10
10
for genetic selections and phage display. The
universal theme in any protein screening assay is the retention of a direct link between the
expressed protein
10
3
to
B
B
'
'
s genetic code (i.e. the mutations conferring
function different from the parent). Thus, if a protein variant is not to be assayed in its host,
one must either keep track of the corresponding cell/plasmid, or physically link the protein
to its genetic information (display technologies).
s function and the protein
In general, the accuracy and quantitative nature of a protein assay vary inversely with the
throughput of the assay. This is particularly true for enzyme screening. For example, very
low-throughput analyses such as chromatography, NMR, or various spectroscopic methods
can provide quantitative information (concentration, structure, kinetics), while many
high-throughput methods that monitor cell growth or fluorescence usually provide
primarily qualitative information. Protein engineers are often willing to sacrifice data
accuracy and confidence in order to observe much larger numbers of clones, at the risk of
selecting false positives as well as missing positive clones.
31
An example scenario for engineering an enzyme
s activity toward a nonnative substrate
follows: a chromatography (HPLC) assay provides with high accuracy the amount of desired
compound produced by each enzyme variant (throughput of
'
10
2
clones per day per
instrument), a 384-well-microtiter plate spectrophotometric assay provides kinetic
information about each variant acting on a surrogate substrate (similar to but not exactly
the intended substrate) (throughput of
B
10
4
clones per day), and finally a growth selection
has been developed, in which a minimal level of the enzyme
B
s desired activity is required
for cell viability, but cell growth rate does not otherwise correlate well with activity
(
'
10
8
clones per day). Note that other screening technologies could easily substitute for
these three options in similar scenarios
B
e.g. FACS and phage display are similar to growth
selection in terms of throughput and information garnered.
If a good deal is known about the enzyme, perhaps a few attempts at site-directed
mutagenesis would be sufficient. Alternately,
n
active-site residues may be individually
mutated to all possible amino acids, resulting in 19
n
variants that could easily be tested
using HPLC for small values of
n
. The plate reader assay becomes preferable when
saturation mutagenesis is instead performed at many or all residues (individually) in the
enzyme, or other mutagenesis strategies result in library sizes greater than
3
1000. Even for
very large or intractable library sizes (from random mutagenesis, doped mutation
oligonucleotide amplification, or family shuffling), as long as the average mutation
frequency remains relatively low (1
B
3 amino acid substitutions per clone), or if the library
is a product of homologous DNA shuffling, the frequency of positive hits should remain
high enough to enable use of the plate reader assay, with the benefit of having quantitative
information about the relative activities of variants. Positive clones can later be verified
using the desired substrate and the HPLC assay. However, since the cell growth selection is
capable of quickly eliminating a large number of nonfunctional mutants, it may be
