The Lotus japonicus Genome

population to screen for their own particular phe-
notypes, such as root architecture and, in particu-
lar, spontaneous nodulation (see Sect.21.5). A
collection of 80 Lotus ecotypes was also assem-
bled for use as an ecoTILLING population
(Fig.21.1a).
A summary of the types of mutation found by
TILLING across the Lotus populations is shown
in Table21.1. Not all mutations can generate a
phenotype. In our populations, 43 % of muta-
tions were in non-coding regions or were silent,
with an average of seven mutations identified for
each sequence targeted in the Gifu population
and 11 in the MG20 population, although the
latter sample is much smaller. Of note are the
4.9 % of mutations causing gene knockouts by
creating premature stop codons or missplicing at
splice–site junctions. Interestingly, there is a
difference between the two populations in the
ratio of mutations existing in heterozygous form

to those in homozygous ones. In initial obser-

vations of the Gifu population, the heterozygote-

to-homozygotes ratio was found to be 10:1,

indicating three germline cells contributed to

male and female gametes (Perry et al. 2009

Supplemental Fig. S1). If a single mutagenized

cell was carried through into the germline, one

would anticipate a 2:1 ratio as observed in the

original Arabidopsis population (Greene et al.

2003 ) and seen here in the MG20 population

(although in a much smaller sample than for the

Gifu population). Current data accumulated for

the Gifu population indicate a ratio of 5.5:1. This

is more in line with two germline cells in each

flower contributing to gametes. The reason for

the difference between the two ecotypes is not

obvious as one would expect germline genetics

of the two to be the same. It may indicate some

kind of selection in the generation of the original

Gifu population favouring heterozygotes.

Fig. 21.1 TheLotus japonicusplant populations used
for forward and reverse genetics. The initial populations
were developed in ecotype Gifu (a) (Perry et al. 2003 ,
2009 ) and new populations generated in MG20 in 2007

(b). The type of forward screen is included in italics. All

bar one (for isoflavonoids) of these screens was used

successfully to isolate mutants (see text for further details)

Table 21.1 Summary of mutations found by TILLING inLotus japonicus

No. of

genes

TILLed

No. of

fragments

TILLed

Mutation type Average

no. per

TILLING

fragment

Het:

Hom

Missense Premature ratio

Stop

Splice

Junction

Non-

coding

Silent

(no AA

change)

Total

Gifu 150 235 895 64 18 412 337 1726 7.4 5.5:1
MG20 20 20 140 13 1 23 46 223 11.2 2.1:1
Total 170 a 255 1,035 77 19 435 383 1,949
Data were taken from Table21.3, but includes embargoed genes. The ratio of the number of M 2 plants bearing mutations in heterozygous form to
homozygous ones is given in thefinal column.Notethe difference between the two populations in this ratio (see text for further comments). Note
also the mutations causing premature stop codons and splicing errors amount to 4.9 %, equivalent to the theoretical value (McCallum et al.a 2000 )
Some of the total of 163 genes have been TILLed in both populations

21 TILLING inLotus japonicus 231