21.4 The Process
TILLING can be carried out using a range of
different technologies, but the central principle is
the same: the detection of rare mutations in
pooled DNA samples from large mutant popula-
tions in afirst round of screening, followed by
deconvolution of the pool to identify the indi-
vidual carrying the mutation. The majority of
current TILLING protocols utilise two different
and sequential techniques—one to screen the
pools and another to identify the mutant indi-
vidual. The definitive method for initial screening
uses single-strand-specific endonucleases that
cleave DNA at the site of mismatched nucleotides
in otherwise double-stranded DNA (Colbert et al.
2001 ). The mismatches form a tiny single-stran-
ded‘bulge’that is recognised and nicked by these
enzymes, and the resulting cleaved andfluores-
cently labelled PCR fragments are separated and
visualised. The process used forL. japonicushas
been described previously in detail for the initial
instrument employed to separate fragments, an
ABI 377 DNA analyzer (Perry et al. 2003 , 2005 ),
and also for the LI-COR®4300 DNA analyzer
(Perry et al. 2009 ). Both instruments use dena-
turing polyacrylamide gel electrophoresis for
fragment separation, and presently, most labora-
tories use this technique for TILLING.
A method that replaces the gel-based system
with fragment separation by capillary sequencing
(Le Signor et al. 2009 ) was developed by the John
Innes Genome Laboratory under the EU GLIP
project (Grain Legumes Integrated Project,http://
http://www.pcgin.org/GLIP/pubrep.pdf)..) This tech-
nique was adopted byRevGenUK and is currently
in use for all its populations. In this method, het-
eroduplexed PCR products from pooled DNA
samples are cleaved using purified celery juice
extract (CJE) and the resulting labelled fragments
separated on an Applied Biosystems™3730xl
DNA sequencer. The RevGenUK TILLING
pipeline utilises a number of resources to assist
with the design and implementation of the best
possible TILL for each gene. To be effective, the
genomic and coding or protein sequence of
the gene of interest needs to be known so that it
can be annotated correctly with all intron/exon
boundaries marked. To identify regions of the
gene that are likely to give the highest concen-
tration of deleterious mutations generated by
EMS-induced G/C to A/T transitions, the Web
software CODDLE (Codons Optimised to Dis-
cover Deleterious LEsions;http://www.proweb.
org/coddle) simplifies this selection procedure
(Colbert et al. 2001 ). The program automatically
searches homology databases using either sug-
gested or user-defined homology blocks (the latter
by means of the program Sorts Intolerant From
Tolerant (SIFT) to predict whether a substitution
affects protein function) (Kumar et al. 2009 ) and
identifies conserved regions important for protein
function. Taking this and the amino acid compo-
sition of the protein into account, CODDLE
identifies a region suitable for TILLING. The
CODDLE program can also design primer pairs
following standard quality criteria based on Pri-
mer3 (Untergrasser et al. 2012 ) and in addition
considers homology between similar genes.
However, where genes are interrupted by
numerous and/or large introns, CODDLE can
recommend a fragment unsuitable for our TILL-
ING protocol because it attempts to encompass
more than one exon and hence may contain an
unacceptably large proportion of non-coding
sequence. In addition, long tracts of homopoly-
mers are not excluded by CODDLE. These are
best avoided becauseTaqpolymerase slippage
can introduce mismatches in such regions. Con-
sequently, in some circumstances, CODDLE is
used for guidance only.
As part of our quality controls before TILL-
ING a gene, we generally test two primer pairs
for each fragment for their amplification effi-
ciency andfidelity. We select one pair and then
retest withfluorescently labelled versions using
unlabelled/labelled at a ratio of 3:2 to replicate
our TILL PCR conditions. Our current DNA
populations are mostly pooled eightfold; hence,
each TILL starts by using four to six eightfold
pooled 96-well plates. Products are tested to
determine how much of the PCR to use in the
subsequent CJE digest. After the digest step and
subsequent clean-ups, the samples are run on the232 T.L. Wang and F. Robson