Biology Reference
In-Depth Information
fragments are assembled), fragment sizes isolated from DNAse I treatment for reassembly can
vary from as low as 10
50 bp to greater than 1 kb. Random point mutations tend to occur at
low rates during recombination even with a high-fidelity polymerase, and researchers will often
intentionally employ error-prone PCR during PCR-based gene recombination to further
diversify their library. In a noteworthy, early example of DNA shuffling, a family of 20 human
interferon-alpha genes was shuffled followed by selection of antiviral and antiproliferation
activities in murine cells, resulting in variants having 285 000-fold increased activity.
25
The best
chimeras were composed of up to five parental genes and contained no random point
mutations. In another example, shuffling of 26 homologous protease genes generated many
chimeric proteases that were significantly improved over any of the parental enzymes for four
different properties assayed.
26
A number of other PCR-based and homology-dependent protocols have since been developed
which accomplish essentially the same as DNA shuffling, and many of these are summarized in
a methods volume.
1
Examples include the staggered extension process,
27
random-priming
in vitro recombination,
28
and random chimeragenesis on transient templates.
29
In general
these methods work well for DNA sequences sharing at least 65% identity, although higher
sequence identity more readily yields a greater number of crossovers.
In vitro recombination methods are also often used in directed evolution, even when the
only genetic diversity is introduced by random mutagenesis of a single parent gene. Here,
one or more rounds of mutagenesis and screening to isolate improved variants results in a
handful of mutant genes, each carrying a different set of point mutations. By shuffling these
highly identical mutant DNA sequences, one can readily obtain a library containing all
combinations of point mutations. Beneficial mutations can be combined and may show
additive effects, while any potentially deleterious mutations that have accumulated will be
eliminated by
'
'
with the wild-type sequence.
In Vivo Homologous Recombination
Whereas gene libraries generated using PCR are most
commonly ligated into an expression vector and transformed into
E. coli
(not necessarily
for expression and screening, but at least for plasmid library recovery), a more recently
developed technique that does not require ligation takes advantage of yeast homologous
recombination.
30
33
The general approach involves cotransforming in yeast a linear vector
expressing the target gene with linear, homologous DNA fragments. In vivo homologous
recombination between these genes results in a library of mutants cloned into the vector. The
homologous DNA fragments providing sequence diversity can be PCR products of in vitro
recombination, a family of homologues, or even a library of synthetic oligonucleotides.
Nonhomologous Recombination
Often it is desirable to generate protein fusions or
higher-order chimeras by combining two or more genes that are either not homologues or
do not share sufficient sequence similarity for the above methods. One approach is rational
design by assembling protein domains or modules, as described above (fusion or crossover
positions are selected and the corresponding chimera is cloned and expressed). Typically,
however, the choice of positions to insert protein domains, modules, or nonhomologous
crossover points is not obvious, or individually constructed chimeras do not work. A variety
of nonhomologous recombination methods have therefore been developed for the
generation of chimeric libraries which can subsequently be screened for function. The
methods of
back-crossing
29
(ITCHY),
34
ITCHY
'
incremental truncation for the creation of hybrid enzymes
'
),
35
and
combined with DNA shuffling (
'
SCRATCHY
'
'
sequence homology-independent
(SHIPREC)
36
involve in vitro construction of gene fusions by
ligating libraries of nonhomologous genes which have been randomly truncated. More
recently, alternate techniques have been developed which utilize more advanced molecular
biology tools, in efforts to enhance crossovers.
37,38
protein recombination
'
Whereas these previous methods involve randomizing the locations of nonhomologous
sequence crossovers, parent protein structural information can also be used to guide the design
