Biomedical Engineering Reference
In-Depth Information
Box 6.3 Gridded (arrayed) hybridization reference libraries
Traditionally, library screening by hybridization
involves taking a plaque lift or colony blot, which
generates a replica of the distribution of clones on an
agar plate. However, an alternative is to individually
pick clones and arrange them on the membrane in
the form of a regular grid. Once a laborious process,
gridding or arraying has been considerably simplified
through the use of robotics. Machines can be
programmed to pick clones from microtitre dishes
and spot them onto membranes at a high density;
then the membrane can be hybridized with a
radioactive probe as normal. Using traditional
libraries, positive clones are detected by
autoradiography and the X-ray film must be aligned
with the original plates in order to identify the
corresponding plaques. With gridded libraries,
however, positive signals can be used to obtain sets
of coordinates, which then identify the corresponding
clone from the original microtitre dishes. Since
identical sets of membranes can be easily prepared,
duplicates can be distributed to other laboratories
for screening. These laboratories can then determine
the coordinates of their positive signals and order
the corresponding clone from the source laboratory.
Thus, one library can serve a number of different
users and all data can be centralized (Zehetner &
Lehrach 1994). Gridded libraries, while convenient
for screening and data sharing, are more expensive
to prepare than traditional libraries. Therefore,
they are often prepared for high-value libraries with
wide applications, such as genomic libraries cloned
in high-capacity P1, BAC or YAC vectors (Bentley
et al
. 1992) and also for valuable cDNA libraries
(Lennon & Lehrach 1991). It is possible to plate
libraries at a density of one clone per well, although
for PACs and BACs it is more common to pool clones
in a hierarchical manner, so that individual clones
may be identified by successive rounds of screening
on smaller subpools (e.g. Shepherd
et al
. 1994,
Shepherd & Smoller 1994).
is being used to isolate a homologous clone from
another species (e.g. see Old
et al
. 1982). Probes cor-
responding to a conserved functional domain of a
gene may also cross-hybridize with several different
clones in the same species at lower stringency, and
this can be exploited to identify members of a gene
family. The identification of the vertebrate
Hox
genes
provides an example in which cross-species hybrid-
ization was used to identify a family of related clones
(Levine
et al
. 1984). In this case a DNA sequence
was identified that was conserved between the
Drosophila
developmental genes
fushi tarazu
and
Antennapedia
. When this sequence, the homoeobox,
was used to screen a Southern blot of DNA from
other species, including frogs and humans, several
hybridizing bands were revealed. This led to the
isolation of a number of clones from vertebrate cDNA
libraries representing the large family of
Hox
genes
that play a central role in animal development.
Hybridization thus has the potential to isolate any
sequence from any library
if a probe is available
. If a
suitable DNA or RNA probe cannot be obtained from
an existing cloned DNA, an alternative strategy is to
make an oligonucleotide probe by chemical synthe-
sis. This requires some knowledge of the amino acid
sequence of the protein encoded by the target clone.
However, since the genetic code is degenerate (i.e.
most amino acids are specified by more than one
codon), degeneracy must be incorporated into probe
design so that a mixture of probes is made, at least
one variant of which will specifically match the tar-
get clone. Amino acid sequences known to include
methionine and tryptophan are particularly valu-
able because these amino acids are each specified by
a single codon, hence reducing the degeneracy of the
resulting probe. Thus, for example, the oligopeptide
His-Phe-Pro-Phe-Met may be identified and chosen
to provide a probe sequence, in which case 32 differ-
ent oligonucleotides would be required:
T
CA
T
TT
T
CCCTT
T
ATG
5
′
3
′
CC
A
C
G
These 32 different sequences do not have to be syn-
thesized individually because it is possible to perform
