Biomedical Engineering Reference
In-Depth Information
the identification of differentially expressed cDNAs
that are also moderately abundant (e.g. Dworkin
& Dawid 1980). Let us consider, for example, the
isolation of cDNAs derived from mRNAs which are
abundant in the gastrula embryo of the frog
Xenopus
but which are absent, or present at low abundance,
in the egg. A cDNA library is prepared from gastrula
mRNA. Replica filters carrying identical sets of
recombinant clones are then prepared. One of these
filters is then probed with
32
P-labelled mRNA (or
cDNA) from gastrula embryos and one with
32
P-
labelled mRNA (or cDNA) from the egg. Some
colonies will give a positive signal with both probes;
these represent cDNAs derived from mRNA types
that are abundant at both stages of development.
Some colonies will not give a positive signal with
either probe; these correspond to mRNA types
present at undetectably low abundance in both
tissues. This is a feature of using
complex probes
,
which are derived from mRNA populations rather
than single molecules: only abundant or moder-
ately abundant sequences in the probe carry a
significant proportion of the label and are effective
in hybridization. Importantly, some colonies give a
positive signal with the gastrula probe, but not
with the egg probe. These can be visually identified
and should correspond to differentially expressed
sequences.
A recent resurgence in the popularity of differen-
tial screening has come about through the develop-
ment of DNA microarrays (Schena
et al
. 1995). In
this technique, cDNA clones are transferred to a
miniature solid support in a dense grid pattern and
screened simultaneously with complex probes from
two sources, which are labelled with different fluoro-
phores. Clones that are expressed in both tissues will
fluoresce in a colour that represents a mixture of
fluorophores, while differentially expressed clones
will fluoresce in a colour closer to the pure signal
of one or other of the probes. A similar technique
involves the use of DNA chips containing densely
arrayed oligonucleotides. These methods are com-
pared in Box 6.5.
clones by removing sequences that are common to
two sources. This is called a
subtracted cDNA library
and should greatly assist the isolation of rare cDNAs.
If we use the same example as above, the aim of the
experiment would be to generate a library enriched
for cDNAs derived from gastrula-specific mRNAs.
This would be achieved by hybridizing first-strand
cDNAs prepared from gastrula mRNA with a large
excess of mRNA from
Xenopus
oocytes. If this driver
population is labelled in some way, allowing it to be
removed from the mixed population, only gastrula-
specific cDNAs would remain behind. A suitable
labelling method would be to add biotin to all the
oocyte mRNA, allowing oocyte/gastrula RNA/cDNA
hybrids as well as excess oocyte mRNA to be sub-
tracted by binding to streptavidin, for which biotin
has great affinity (Duguid
et al
. 1988, Rubinstein
et al
. 1990). Highly enriched libraries can be pre-
pared by several rounds of extraction with driver
mRNA, resulting in highly enriched subtracted
libraries (reviewed by Sagerstrom
et al
. 1997).
An example of subtractive cDNA cloning and
differential screening is provided by Nedivi
et al
.
(1993). These investigators were interested in the
isolation of rat cDNAs that are induced in a particular
region of the brain (the dentate gyrus (DG)) known
to be involved in learning and memory. The induc-
ing stimulus was kainate, a glutamate analogue
that induces seizures and memory-related synaptic
changes. Poly(A)
+
RNA was extracted from the DG
of kainate-treated animals and used for first-strand
cDNA synthesis. Ubiquitous sequences present in
the activated DG cDNA preparation were hybridized
with an excess of poly(A)
+
RNA from total unin-
duced rat brain. This RNA had previously been
biotinylated (using a photobiotinylation procedure)
and so hybrids and excess RNA could be removed
using a streptavidin extraction method (Sive & St
John 1988). The unhybridized cDNA was then con-
verted into double-stranded form by conventional
methods and used to construct the subtracted cDNA
library in
ZAP. This subtracted library was differ-
entially screened using radiolabelled cDNA from
activated and non-activated DG as the differential
probes. A large number of activated DG clones were
isolated, of which 52 were partially sequenced. One-
third of these clones corresponded to known genes;
the remainder were new.
λ
Enrichment for differences - subtractive cloning
An alternative to differential screening is to generate
a library that is enriched in differentially expressed
