Effect of seed layer removal on germination and on BFB
seedling transmission. Confirmation of the location of A. citrulli
in seeds was sought by removing seed layers and assaying the
seeds for BFB seedling transmission. We assumed that A. citrulli
cells associated with each tissue would be eliminated as the seed
layers were removed, thereby reducing BFB seedling transmission.
The mean germination percentages for seeds with only the
testae removed (78.5 and 75.7% for pericarp- and pistil-inoculated
lots, respectively) were not significantly different to intact
seeds (76.5 and 81.0% for pericarp and pistil-inoculated seeds,
respectively) (Fig. 1A and B). However, the germination percentages
for seeds from which the testae and PE layers were removed
(45.5 and 51.0% for pericarp- and pistil-inoculated lots, respectively)
were significantly lower than for intact seeds and seeds
from which only the testae were removed (Fig. 1C). With regards
to BFB seedling transmission, intact seeds from pericarp- and
pistil-inoculated lots yielded seedlings with 76.3 and 75.0% BFB
incidence, respectively (Fig. 2A). Testa removal did not have a
significant effect on BFB seedling transmission for either seed
inoculation method (P = 0.74). In contrast, the removal of the
testae and PE layers from pericarp-inoculated seeds resulted in a
statistically significant reduction in BFB seedling transmission
(14.8%) compared with intact pericarp-inoculated seeds (76.5%)
(P < 0.0001) and seeds with the testae removed (85%) (P =
0.003). This was not the case for pistil-inoculated lots, for which
testa and PE layer removal did not have a significant effect (P =
0.89) on BFB seedling transmission (72%) relative to intact seeds
or seeds with testae removed (Fig. 2). Finally, testa and PE layer
removal from pericarp-inoculated seeds resulted in a statistically
significant (P = 0.003) reduction in BFB seedling transmission
relative to pistil-inoculated seeds treated in the same manner (Fig.
2). All seedlings that displayed putative BFB symptoms were
confirmed as infected with A. citrulli by bacterial recovery on
semiselective media and subsequent species-specific real-time
PCR assays (data not shown)