Integrative Transcript and Metabolite Analysis of Nutritionally
Enhanced DE-ETIOLATED1 Downregulated Tomato Fruit
Eugenia M.A. Enfissi,a,bFredy Barneche,c,dIkhlak Ahmed,cChristiane Lichtle ´,cChristopher Gerrish,a
Ryan P. McQuinn,eJames J. Giovannoni,e,fEnrique Lopez-Juez,a,bChris Bowler,cPeter M. Bramley,a,b
and Paul D. Frasera,b,1
aCentre for Systems and Synthetic Biology, University of London, Egham, Surrey TW20 0EX, United Kingdom
bSchool of Biological Sciences Royal Holloway, University of London, Egham, Surrey TW20 0EX, United Kingdom
cInstitut de Biologie de l’Ecole Normale Supe ´rieure, Centre National de la Recherche Scientifique,
Unite ´ Mixte de Recherche 8197, 75005 Paris, France
dStazione Zoologica "Anton Dohrn," Villa Comunale, I 80121 Naples, Italy
eU.S. Department of Agriculture, Agricultural Research Service, Plant Soil and Nutrition Laboratory, Ithaca, New York 14853
fBoyce Thompson Institute for Plant Research, Cornell University Campus, Ithaca, New York 14853
Fruit-specific downregulation of the DE-ETIOLATED1 (DET1) gene product results in tomato fruits (Solanum lycopersicum)
containing enhanced nutritional antioxidants, with no detrimental effects on yield. In an attempt to further our understand-
ing of how modulation of this gene leads to improved quality traits, detailed targeted and multilevel omic characterization
has been performed. Metabolite profiling revealed quantitative increases in carotenoid, tocopherol, phenylpropanoids,
flavonoids, and anthocyanidins. Qualitative differences could also be identified within the phenolics, including unique
formation in fruit pericarp tissues. These changes resulted in increased total antioxidant content both in the polar and
nonpolar fractions. Increased transcription of key biosynthetic genes is a likely mechanism producing elevated phenolic-
based metabolites. By contrast, high levels of isoprenoids do not appear to result from transcriptional regulation but are
more likely related to plastid-based parameters, such as increased plastid volume per cell. Parallel metabolomic and
transcriptomic analyses reveal the widespread effects of DET1 downregulation on diverse sectors of metabolism and sites
of synthesis. Correlation analysis of transcripts and metabolites independently indicated strong coresponses within and
between related pathways/processes. Interestingly, despite the fact that secondary metabolites were the most severely
affected in ripe tomato fruit, our integrative analyses suggest that the coordinated activation of core metabolic processes in
cell types amenable to plastid biogenesis is the main effect of DET1 loss of function.
Diets rich in fruits and vegetables have been associated with the
reduced incidence of chronic disease states (Key et al., 2002).
These findings have led many western governments to recom-
mend the consumption of five portions of fruits and vegetables
per day (Cooper, 2004). The health benefits conferred by certain
fruits and vegetables have been attributed to the presence of
health-promoting phytochemicals (more recently termed bioac-
tives). Carotenoids, flavonoids, phenylpropanoids, tocopherols,
and ascorbic acid (vitamin C) are all bioactives with potent
antioxidant properties. Ripetomato fruit (Solanumlycopersicum)
contain significant amounts of these compounds and are the
principal dietary source of the carotenoid lycopene in the human
diet (Giovannucci, 2002).
The enhancement of nutritional quality is an important objec-
tive of modern plant breeding. Conventional molecular breeding
and genetic modification (GM) technologies have been em-
ployed to generate better nutritional quality in crop plants,
particularly tomato. Traditional genetic engineering of the target
pathway has resulted in modest enhancement of specific me-
tabolites, such as lycopene (Fraser et al., 2002) or flavonoids
(Muir et al., 2001). Despite being more time consuming, labor
intensive, and not as precise, non-GM approaches, such as
marker-assisted screening, can be employed to achieve these
increases (Zamir, 2001). In this way, consumer concerns asso-
ciated with GM are avoided. However, more recently, genetic
engineeringapproaches involvingminipathwayreconstruction in
crop plants have resulted in dramatic increases in carotenoids,
Diretto et al., 2007). The potential of transcription factors to
modulate biochemical pathways has also been elegantly dem-
onstrated recently (Butelli et al., 2008; Luo et al., 2008). In most
cases, these approaches have focused on specific pathways to
deliver a defined end product. By contrast, the manipulation of
light signal transduction components (Liu et al., 2004; Davuluri,
et al., 2005) or photoreceptors (Giliberto et al., 2005) in to-
mato fruit has facilitated enhancement of multiple bioactives
1Address correspondence to email@example.com.
The author responsible for distribution of materials integral to the
findings presented in this article in accordance with the policy described
in the Instructions for Authors (www.plantcell.org) is: Paul D. Fraser
WOnline version contains Web-only data.
The Plant Cell, Vol. 22: 1190–1215, April 2010, www.plantcell.org ã 2010 American Society of Plant Biologists
simultaneously regardless of their formation by independent
biosynthetic pathways (e.g., carotenoids and flavonoids). A
disadvantage of manipulating components of the light signal
transduction pathway, such as DE-ETIOLATED1 (DET1; origi-
nally identified as HIGH PIGMENT [hp2]), UV-DAMAGED DNA
BINDING PROTEIN1 (DDB1; originally hp1), and CULLIN-4
(Wang et al., 2008) either through transgenic constitutive ex-
pression or via mutant alleles such as hp1w, hp2, hp2j, and hp2dg
is reduced fruit yield and loss of plant vigor (Davuluri et al., 2004).
However, the fruit-specific downregulation of endogenous DET1
expression is a good example of how light signal transduction
components can be manipulated for biotechnological benefit
more recent examples of genetic modification offer important
generic potential that is presently beyond the scope of conven-
Both the DET1 and DDB1 gene products are involved in the
suppression of light responses in the absence of light. Their
molecular function has been associated with chromatin remod-
eling (Benvenuto et al., 2002). Altered plastid biogenesis leading
to an increased plastid compartment per cell is also believed to
be a contributing factor to elevated chlorophyll and carotenoid
of plastid number per cell (Cookson et al., 2003; Liu et al., 2004)
as well as an abundance of differentially expressed transcripts
associated with plastid biogenesis (Kolotilin et al., 2007).
The fundamental characterization of DET1 clearly points to a
key role in core processes involved in plant development and
environmental adaptation. However, in ripe tomato fruit, the
downregulation of DET1 results in the simultaneous elevation of
vide an insight into the dynamic molecular events and metabolic
reprogramming leading to this DET1 fruit chemotype, we per-
formed integrative transcriptomic and metabolomic analyses.
Phenotypic Stability and Correlation with DET1
Downregulation during Fruit Development
Phenotypic inheritance has been shown for several DET1 down-
regulated events in the second generation (T2); these lines were
generated from segregating primary transformants (T1) (Davuluri
et al., 2005). Selected lines representing three different fruit
specific promoters (2A11, TFM7, and P119) have subsequently
been taken through a further three generations (T3 to T5) in this
study and their pheno/chemotypes evaluated. In comparison to
their wild-type background (T56), all genotypes exhibited dark-
green mature fruit and a more intense red internal coloration of
the ripe fruit. Determination of the carotenoid content found in
ripe fruit was consistently higher over the three generations (see
Supplemental Table 1 online). Therefore, stable phenotypic
inheritance has been achieved, leading to the designation of
physiological parameters (e.g., fruit yield, diameter, rate of
ripening, plant height, and growth rate) among the DET1 varie-
ties, in agreement with the determinations performed previously
on a wider range of events (Davuluri et al., 2005). Material from
these varieties has been used in this study for detailed charac-
terization using multiple omics-based approaches.
The promoters 2A11, TFM7, and P119 are known to act during
fruit development (Davuluri et al., 2005), but the precise timing
and strength by which they control DET1 expression remain
poorly characterized. To determine the quantitative spatial tim-
ing of DET1 downregulation by the three promoters, a develop-
mental series of fruit was generated and qRT-PCR used to
determine DET1 expression levels. Seven stages from immature
fruit to red-ripe as illustrated in Figure 1A have been analyzed.
The 2A11 promoter was found to be the weakest of the three
promoters; its effects were greatest in developmental stages 4
and 5 (Figure 1B) but represented only a 20% reduction in DET1
expression compared with the wild type. After stage 5, endog-
enous DET1 transcripts returned to wild-type levels. DET1 ex-
pression was similar under both P119 and TFM7 promoter
control. A reduction in the amount of DET1 transcript was initially
observed at stage 3 and progressed until stage 6 (mature green),
when an approximate 70% reduction in DET1 expression oc-
curred in both the P119 and TFM7 varieties (Figure 1B). Curi-
ously, in both cases, DET1 expression returned to wild-type
levels at the red-ripe stage.
In all cases, increases in chlorophyll, carotenoids, and phe-
and mimicked expression profiles until DET1 downregulation
was alleviated (Figures 1C to 1E). At this point, increased fruit
carotenoid, chlorophyll, and phenolic contents were maintained
even in the absence of DET1 downregulation. On the basis of the
expression and the concurrent appearance of the phenotype,
detailed characterization was initiated at the mature green
stages of fruit development (stage 5).
The Effect of DET1 Downregulation on Carotenoid
(Isoprenoid) and Phenolic Formation at the Metabolite
and Gene Expression Level
Carotenoids, xanthophylls, chlorophylls, and tocopherols were
profiled simultaneously during fruit development and ripening for
all three DET1 varieties. In comparison to fruits from the T56 wild
type, chlorophyll levels were elevated in all DET1 varieties up to
the breaker stage (Table 1). The 2A11 variety showed a signif-
icant 3-fold increase in chlorophyll at the mature green stage of
fruit development. In both the TFM7 and P119 varieties, greater
increases (9-fold) in chlorophyll were evident, and the total
carotenoid content of the fruit increased concurrently with chlo-
rophyll content. As a result, the carotenoid to chlorophyll ratio
(;3.0) between the DET1 varieties remained constant and
similar to the wild type. For comparison, two mutant hp2 alleles
in different backgrounds (Moneymaker and San Marzano) were
also analyzed. The total carotenoid content of the mature green
fruit were similar to the DET1 varieties in these mutants (Table 1).
However, the backgrounds for each of the hp alleles contained
higher carotenoid contents; therefore, the relative increases in
the P119 and TFM7 varieties were greater (e.g., 8-fold compared
Analysis of DET1 Downregulated Tomato Fruit1191
with 4-fold in the hp alleles).The effectsof DET1on the individual
carotenoids are also shown in Table 1. No change in the carot-
enoid composition was observed in the DET1 and hp mutant
varieties analyzed, with the relative increases among individual
carotenoids being similar.
At the breaker stage of fruit ripening, the 2A11, TFM7, and
P119 varieties showed 4-, 16-, and 13-fold increases, respec-
tively, in total carotenoid content compared with the T56 back-
alleles, which exhibited a 2-fold relative increase compared with
their wild-type backgrounds. Table 1 documents the levels of
individual carotenoids found in fruit at the breaker stage. The
carotenoids lutein and b-carotene were predominant in both the
T56 background and DET1 varieties. Of the carotenes, phytoene
was only found in the TFM7 and P119 varieties, and lycopene
was only present in the 2A11 and TFM7 varieties, whereas
Figure 1. Profile of Relative DET1 Expression, Total Chlorophyll, Carotenoid, and Phenolic Levels in the 2A11, TFM7, and P119 Genotypes during Fruit
Development and Ripening.
(A) Illustration of the designated fruit stages sampled, their approximate diameter, days after anthesis (dpa), and days postbreaker (dpb).
(B) Relative changes in DET1 expression (determined by qRT-PCR) among the DET1 downregulated varieties compared with their control (T56) at the
indicated stages of development and ripening. The dashed gray line designates a ratio of 1 (i.e., no change in expression).
(C) to (E) Relative changes in total chlorophyll, carotenoid, and phenolic contents compared with their controls. Phenolics are represented by the sum of
the phenylpropanoids, flavonoids, and anthocyanins analyzed. Typical determinations (mg/g DW) for total chlorophylls, carotenoid, and phenolics
found in the T56 at the various stages are 4.0, 1.1, and 1.0 at stage 3; 4.0, 1.0, and 0.9 at stage 4; 3.3, 0.8, and 0.63 at stage 5; 0.04, 0.02, and 0.2 at
stage 6; and 0, 3.0, and 0.4 at the red-ripe stage (stage 7). Biological replicates were performed in triplicate and the data presented as means 6 SD.
1192 The Plant Cell
neurosporene was detected in all three varieties, showing 3-, 17-,
and 8-fold increases in 2A11, TFM7, and P119, respectively. In
addition to carotenoids, the 2A11, TFM7, and P119 varieties
in the TFM7 variety. a-Tocopherol was the predominant tocoph-
erol but with g-tocopherol detectable and mimicking the relative
changes in a-tocopherol.
Lycopene was the major carotenoid found in all ripe tomato
fruit analyzed regardless of variety (Table 1). Increases in lyco-
pene ranged from 2-fold in 2A11 to 5-fold increases in the hp
mutant alleles. Phytoene and phytofluene both showed similar
the 2A11 and TFM7 varieties, b-carotene levels were increased
akin to the other carotenoids analyzed. However, in the P119
variety, the increase in b-carotene was significantly (Table 1)
greater (7-fold), reaching 2.0 mg/g dry weight (DW), and repre-
sented the equivalent of 3.5 times the recommended daily
allowance (RDA) of b-carotene (provitamin A) per tomato.
Therefore, the RDA can be delivered in one P119 ripe tomato,
instead of requiring a person to eat three tomatoes with typical
b-carotene contents. Tocopherol contents in ripe fruit were
respectively. The increases in tocopherols mean that the RDA
for tocopherol can be achieved by consuming two P119 ripe
Table 1. Carotenoid, Chlorophyll, and Tocopherol Contents Found in the Transgenic DET1 Downregulated Varieties (2A11, TFM7, and P119) and
hp2 Mutant Alleles Compared with Their Wild-Type Backgrounds
DET1 Genotypes Mature Green
T56 2A11TFM7P119 MM MMhp2j
1.5 6 0.10
3.0 6 0.03
7.7 6 0.30
4.0 6 0.10
16.2 6 1.00
43.7 6 0.60
4.6 ± 0.3 ***
9.3 ± 0.3 ***
18.7 ± 0.7 ***
9.6 ± 1.6 *
42.2 ± 3.0 ***
141.2 ± 7.0 ***
11.0 ± 0.3 ***
29.5 ± 0.4 ***
52.4 ± 1.0 ***
29.0 ± 4.0 **
122.1 ± 5.0 ***
394.2 ± 12.0 ***
11.9 ± 0.5 ***
27.2 ± 1.0 ***
65.4 ± 0.6 ***
35.1 ± 2.5 ***
140.0 ± 0.7 ***
381.1 ± 2.0 ***
2.3 6 0.1
8.1 6 0.1
17.8 6 0.5
11.2 6 0.5
39.4 6 1.0
97.4 6 2.9
8.2 ± 0.2 ***
32.4 ± 0.6 ***
53.1 ± 0.6 ***
25.3 ± 1.0 ***
119.0 ± 1.6 ***
328.4 ± 4.0 ***
2.4 6 0.3
10.7 6 0.1
19.4 6 0.3
7.6 6 0.4
40.0 6 1.0
119.2 6 3.3
10.3 ± 0.2 **
36.4 ± 0.8 ***
73.5 ± 1.5 **
31.6 ± 0.1 ***
151.7 ± 2.2 ***
459.1 ± 12.2 **
DET1 Genotypes Breaker Fruit
(mg/g DW) T562A11TFM7 P119MMMMhp2j
1.3 6 0.02
0.9 6 0.02
2.1 6 0.60
9.7 6 0.60
2.04 6 0.03
14.6 6 2.7
25.0 6 2.0
19.2 6 0.6
1.4 6 0.3
5.8 ± 1.5 ***
7.7 ± 2.2 ***
28.2 ± 8.0 ***
0.1 ± 0.0 ***
6.20 ± 1.6 ***
50.0 ± 14.0 ***
176.9 ± 24.0 ***
37.0 ± 10.0 **
1.4 6 0.1
8.4 ± 1.7 *
34.3 ± 6.0 **
129.8 ± 24.0 **
10.2 ± 2.0 ***
32.5 ± 3.4 *
1.7 ± 0.01 **
217.1 ± 38.9 **
127.4 ± 22.5 ***
176.3 ± 14.0 *
9.7 ± 0.4 **
35.6 ± 2.3 **
36.4 ± 3.0 ***
83.8 ± 5.5 ***
16.0 ± 0.8 **
4.6 ± 0.4 **
182.9 ± 12.0 **
176.9 ± 24.0 ***
86.6 ± 0.04 **
10.0 ± 0.3 ***
61.8 ± 5.2 **
47.0 ± 1.3 ***
188.4 6 6.9
41.2 ± 1.3 **
0.4 6 0.002
337.3 ± 13.7 *
272.0 ± 23.0 **
257.4 ± 6.2 **
7.7 ± 4.0 ***
76.8 ± 1.3 ***
94.6 ± 1.5 ***
426.7 ± 11.0 ***
18.9 6 0.7
85.0 ± 6.2 **
1.6 6 0.2
5.2 6 0.13
714.0 ± 11.7 **
462.1 ± 5.0 **
433.6 ± 20.0 **
11.6 6 5.6
23.1 6 1.6
153.2 6 18.0
14.5 6 2.7
29.8 6 1.1
8.6 6 0.33
241.0 6 23.0
105.2 6 9.5
87.0 6 9.3
19. 2 6 1.3
27.4 6 0.5
250.0 6 7.7
57.7 6 2.0
42.6 6 0.20
6.1 6 0.23
24.4 6 0.3
442.6. 6 25.0
88.9 6 2.4
163.7 6 3.4
DET1 Genotypes Ripe Fruit
(mg/g DW) T562A11TFM7P119 MMMMhp2j
3.0 6 0.1
13.5 6 0.10
198.0 6 3.0
321.7 6 4.5
3.4 6 0.4
41.0 6 1.8
129.2 6 0.4
708.4 6 8.0
148.1 6 11.5
5.5 ± 0.9 *
24.8 ± 0.7 ***
28.2 6 8.0
725.2 ± 14.3 ***
144.0 ± 7.0 ***
315.6 ± 5.2 *
1659.0 ± 42.0 ***
112.4 6 15.3
9.5 ± 0.6 ***
46.4 ± 0.8 ***
851.1 ± 34.0 **
955.5 ± 17.0 ***
68.2 ± 3.2 ***
137.2 ± 1.9
2064.8 ± 44.5 ***
284.2 ± 6.1 ***
37.3 ± 09 ***
132.4 ± 3.0 ***
1455.6 ± 39.6 ***
1574.7 ± 13.5 ***
48.6 ± 6.7 ***
363.7 ± 3.2 ***
583.4 ± 0.8 ***
4179.5 ± 21.0 ***
505.1 ± 6.2 ***
15.2 ± 0.6 ***
70.8 ± 0.6 ***
188.4 6 6.9
1088.6 ± 14.0 ***
181.4 ± 4.0 ***
321.1 6 2.4
2773.4 ± 37.0 ***
403.8 ± 7.0 ***
6.7 6 0.2
19.4 6 0.2
250.0 6 7.7
347.3 6 4.6
117.0 6 1.5
292.2 6 3.2
1347.0 6 37.0
94.7 6 5.7
16.1 ± 0.2 ***
68.04 ± 0.8 ***
426.7 ± 11.0 ***
898.7 ± 11.0 ***
213.8 ± 7.9 ***
363.6.2 ± 4.7 ***
2861.5 ± 59.0 ***
258.3 ± 5.8 ***
19.3 6 0.3
351.0 6 18.0
483.5 6 20.7
130.7 6 1.0
302.6 6 5.0
1287.2 6 38.0
76.2 6 3.2
Tomato fruit at different developmental and ripening stages have been analyzed: mature green, breaker, and ripe. Carotenoid, chlorophyll, and
tocopherol contents are given as mg/g DW. Methods used for determinations are described in Methods. Three representative fruit from a minimum of
three plants were used for each developmental/ripening stage. The fruit were pooled and at least three determinations made per sample, making a
minimum of three biological and three technical replicates. The mean data are presented 6 SD. Student’s t tests were used to determine significant
differences between respective wild-type backgrounds and transgenic varieties. P < 0.05, P < 0.01, and P < 0.001 are indicated by *, **, and ***,
respectively. Values in bold indicate where significant differences have been found compared to the wild-type backgrounds. ND, not detected; CHL,
chlorophyll; CAR, carotenoid; MM, Moneymaker; SM, San Marzano.
Analysis of DET1 Downregulated Tomato Fruit1193
tomato fruits instead of six or seven T56 wild-type ripe toma-
To ascertain the causative effect of DET1 on carotenoid and
tocopherol levels, qRT-PCR was performed to determine if
transcriptional upregulation of key carotenoid and tocopherol
biosynthetic genes had occurred. The transcript levels for genes
encoding carotenoid and tocopherol biosynthetic enzymes
showed no consistent trend among the 2A11, TFM7, and P119
varieties (Figures 2A and 2B). In all three varieties, the majority of
the pathway transcripts were not significantly affected com-
pared with their control (T56) levels at the mature green stage of
fruit development (Figure 2A). Notable relative increases over
2-fold included PSY-2 in both the P119 and 2A11 varieties, and
GGPPS-2 and LCY-E in the P119 and TFM7 varieties, respec-
tively. However, reductions in transcript levels were more pre-
dominant. For example, DXS transcripts were reduced at least
in 2A11 and TFM7, whereas LCY-E was lower in both 2A11 and
P119 (Figure 2A). Transcripts for tocopherol biosynthetic en-
zymes were elevated significantly at the mature green stage, for
example, in P119, g-methyl tocopherol transferase (GMTT) and
geranylgeranyl pyrophosphate reductase (GGPR) were in-
creased 8- and 5-fold, respectively.
In ripe fruit, most transcripts for carotenoid biosynthetic en-
zymes were either decreased or showed very similar levels
relative to the control (T56) (Figure 2B). By contrast, several
isoprenoid related transcripts, such as GMTT and GGPPS-2,
were increased 5- and 10-fold in the P119 and 2A11 varieties,
respectively. In the 2A11 variety, increases over 2-fold were
observed for DXS, ZDS, and LCY-B transcripts. In spite of the
overall increases in carotenoids and isoprenoids, the most
striking observation from transcriptional analysis of the biosyn-
Phenylpropanoid, Flavonoid, and Anthocyanin Formation
Fruit samples were divided into skin and pericarp tissue and
analysis performed on each tissue type (Table 2). In the skin
tissue from control (T56) mature green fruit, the phenylpropa-
noids caffeic acid, p-coumaric, and chlorogenic acid were
detected, together with flavonoid derivatives of quercetin. Dra-
matic increases in the levels of caffeic acid and quercetin
derivatives were measured in the TFM7 and P119 varieties
compared with their control (T56) (Table 2). For example, in the
P119 variety, caffeic acid and quercetin showed a 13- and 5-fold
elevation, respectively, compared with the control (T56). The
(hp2j) mutant alleles (Table 2), although increases were not as
great (e.g., a 2.5- to 8-fold increase in caffeic acid content was
found in hp2j[Moneymaker]compared with 13-fold in P119). The
(Table 2). In all three varieties, naringenin-chalcone content was
(up to 4-fold). Chlorogenic acid contents were also elevated in all
varieties, most notably in the TFM7 variety where a 7.5-fold
was also elevated 13-fold in comparison to the control (T56).
Interestingly, the hp2jmutant alleles in both Moneymaker and
San Marzano backgrounds all demonstrated increases (up to
10-fold) for all phenolics tested except naringenin-chalcone
In the pericarp tissue from control mature green fruit, caffeic
acid was the only detectable phenolic, its content increasing
5- and 18-fold in the TFM7 and P119 varieties, respectively (Table
2). Quercetin derivatives were found in 2A11, TFM7, and P119,
whereas chlorogenic acid was unique to TFM7 and P119, and
with the hp2 alleles in the backgrounds tested (Moneymaker and
San Marzano), although the amounts determined were not as
substantial as those found with the transgenic DET1 genotypes
(Table 2). Ripe pericarp tissue from the control (T56) variety
contained all the phenylpropanoids targeted with the analytical
system used, but no flavonoids (e.g., quercetin derivatives and
naringenin-chalcone) (Table 2). Interestingly, in comparison to
the T56 background where flavonoids were absent in the peri-
carp tissue, the presence of flavonoids in thistissue was found in
both the TFM7 and P119 varieties. The hp2 mutant alleles also
contained flavonoids in the pericarp (Table 2). The caffeic,
p-coumaric, and chlorogenic acid contents of the P119 variety
were increased 3-fold in all cases. Chlorogenic acid content was
also elevated in the TFM7 variety. Incomparison, the hp2 mutant
alleles possessed similar increases in chlorogenic acid (3-fold),
only in the hp2jMoneymaker background.
The total content of anthocyanidins present in mature green
Table 2). Anthocyanidin content in the ripe fruit was increased
only in the DET1 varieties (3- to 5-fold; Table 2). Analysis of the
spectrometry revealed the sole presence of delphinidin-3-
(coumaroyl)-rutinoside-5-glucoside (M+H+919 m/z).
At the level of gene expression, several key transcripts
relating to biosynthetic genes involved in phenylpropanoid,
3A and 3B). In both the mature green and ripe fruit, the levels of
several transcripts encoding enzymatic steps within these phe-
nolic pathways were dramatically increased (e.g., CHS up to
37.5-fold), but there was no unifying trend between the different
varieties with respect to the biosynthetic steps they affect or the
evident. For example, CHS transcripts in mature green fruit
were increased 37-fold in P119 but only increased 2-fold in
TFM7 and were reduced in 2A11. By contrast, CHS transcripts
in the ripe stage increased 14-fold in TFM7 but increased only a
modest 3-fold in P119. Increased F3H and FLS transcripts were
found only in TFM7 at the ripe stage. Other notable increases
mainly occurred in the P119 variety and were related to
anthocyanidin biosynthetic transcripts (e.g., DFR, ANS, RT,
and 3-GT were induced 5- to 35-fold) (Figures 3A and 3B). In
spite of these inconsistencies, a general trend of increased
expression of genes involved in phenylpropanoid, flavonoid,
and anthocyanidin biosynthesis was nonetheless more evident
than it was for the carotenoid/isoprenoid genes (cf. Figures 2
1194 The Plant Cell
Total Antioxidant Activities
To assess whether the increased levels of carotenoids and
phenolics had resulted in elevated total antioxidant capacity in
analysis was performed (see Supplemental Figure 1 online).
Carotenoids are potent antioxidants that act in the nonpolar
(hydrophobic) phase, whereas phenolics exert theireffects in the
polar extracts. Therefore, both polar and nonpolar extracts were
subjected to TEAC assays. Increases in antioxidant capacity
were found in nonpolar extracts prepared from all three varieties
at the mature green stage (see Supplemental Figure 1A online)
mature green fruit extracts demonstrate modest 2-fold in-
creases, while extracts from the TFM7 and P119 varieties ap-
proach a 4-fold increase (see Supplemental Figure 1A online).
Nonpolar ripe fruit extracts from the TFM7 and P119 varieties
demonstrate increased (approaching 2-fold) antioxidant activity
(see Supplemental Figure 1B online). Polar extracts prepared
Figure 2. The Effect of DET1 Downregulation on Carotenoid/Isoprenoid-Related Gene Expression Levels in Mature Green and Ripe Fruit.
Pooled fruit originating from three plants per promoter (2A11, TFM7, and P119) were pulverized into a homogenous powder as described in Methods.
Total RNA was then extracted from an aliquot of this material. qRT-PCR was performed with gene-specific primers for (1) DXS, 1-deoxy-D-xylulose-5-
phosphate synthase; (2) GGPPS-1, 1-geranylgeranyl pyrophosphate synthase-1; (3) GGPPS-2; (4) PSY-1, phytoene synthase-1; (5) PSY-2; (6) PDS; (7)
ZDS, z-carotene desaturase; (8) CRTISO, carotene isomerase; (9) CYC-B, b-lycopene cyclase; (10) LCY-B, b-lycopene cyclase; (11) LCY-E, e-lycopene
cyclase; (12) GGPPR, geranylgeranyl pyrophosphate reductase; and (13) GMTT. The locations of these enzymatic steps on the pathway are shown by
the numbers superimposed into the circles with the gray backgrounds. The expression data shown have been normalized to the expression of actin.
Data are represented as relative levels found in the three varieties compared with the T56 wild type. Statistical determinations are shown as mean 6 SD
values, where n = 3 to 6. Student’s t tests illustrate statistically significant (*P < 0.05, **P < 0.01, and ***P < 0.001) differences from the wild-type levels.
The black bars of the histogram indicate the levels in 2A11, gray bars TFM7, and pale gray P119. The dashed line across each histogram indicates the
relative wild-type expression level. GGPP, geranylgeranyl pyrophosphate; PPP, phytyl pyrophosphate; IPP, isopentenyl pyrophosphate; DMAPP,
dimethylally pyrophosphate. ND, not detectable.
Analysis of DET1 Downregulated Tomato Fruit1195
from all three varieties (2A11, TFM7, and P119) at both the
mature green and ripe fruit stages contained higher antioxidant
capacities (see Supplemental Figures 1C and 1D online). The
highest increases (2- and 4-fold in the mature green and ripe
extracts, respectively) belonged to the P119 variety.
Cellular Phenotypes of the DET1 Varieties
Previous studies have reported an increased plastid comple-
ment per cell as a characteristic feature of hp mutants (Cookson
et al., 2003; Liu et al., 2004; Kolotilin et al., 2007). Visualization of
cells from the DET1 varieties supported these findings (Figure
4A, i to iv). To quantify this observation, the cell indexes were
determined for all varieties using Nomarski microscopy. The cell
indexes represent the total plastid area per cell versus cellular
plan area, where the cellular plan area is the area of the cell’s
projected top view. To reduce heterogeneity among samples,
in a comparative manner with their respective controls. These
data confirmed that regardless of cell size and type, the number
of plastids, size of the plastids, and cell index were increased in
all DET1 varieties (Figures 4B to 4D). Collectively, the normalized
Table 2. Phenolic Contents of the DET1 Downregulated Varieties Determined Both in Skin and Pericarp Tissues at the Mature Green and Ripe
Stages of Fruit Development and Ripening; the Bottom Section Shows the Changes in Total Anthocyanin Content at Mature Green and Ripe Stages
Phenolics (mg/g DW)
DET1 Genotypes Mature Green Skin
T56 2A11TFM7 P119 MM MMhp2j
576 6 40
206 6 11
270 6 7
958 6 38
292 ± 14 ** 1,911 ± 178 ** 7,667 ± 460 ***
28 ± 1 ***
52 ± 2 ***
140 ± 2 ***228 6 33
743 ± 35 ** 2,766 ± 91 *** 4,462 ± 133 ***
1,535 6 57 12,839 ± 634 ***
64 6 8
552 6 50
3,684 ± 37 ***
3118 6 186
40 6 6
461 6 21
7,762 ± 593 **
367 ± 27 ***
502 ± 1 ***
134 ± 13 ***
271 ± 4 ***
114 6 27 ND
1,737 ± 62 ***
DET1 Genotypes Ripe Skin
Phenolics (mg/g DW) T562A11 TFM7 P119MMMMhp2j
Naringenin-chalcone 1,900 6 300
234 6 10
223 6 17
346 6 28
547 6 34
177 ± 5 **
176 ± 4 *
1,579 ± 24 *** 2,600 ± 63 ***
1,207 ± 34 *** 1,758 ± 50 ***
706 ± 110 **
173 ± 7 ***
153 ± 23 **
3,103 ± 98 ***
217 6 38
765 ± 42 ***
1,921 ± 37 *** 1,298 6 43
775 ± 215 ** 6,039 6 491
463 6 94
215 6 21
509 6 59
4,668 ± 774 **
390 ± 11 ***
749 ± 42 **
4,696 ± 375 **
452 ± 441 *** 12,386 6 1264 4,892 ± 15,93 ***
1307 6 38
288 6 12
247 6 5
1267 6 48
2,028 ± 38 ***
368 ± 19 **
1,,010 ± 59 ***
2897 ± 67 ***
331 ± 220 **
DET1 Genotypes Mature Green Pericarp
Phenolics (mg/g DW) T562A11 TFM7P119MM MMhp2j
Quercetin derivatives ND
347 6 25
235 ± 18 **
154 ± 22 ***
1,598 ± 104 ***
61 ± 3 ***
810 ± 437 ***
6,588 ± 85 ***
394 ± 1 ***
46 ± 2 ***
2,903 ± 33 *** ND
1041 6 80
4,710 ± 936 *
36 ± 6 ***
807 ± 78 *** ND
3224 6 68
4,611 ± 172 ***
175 ± 13 ***
905 ± 41 ***
DET1 Genotypes Ripe Pericarp
Phenolics (mg/g DW) T562A11 TFM7P119 MM MMhp2j
Quercetin derivatives ND
635 6 10
110 6 0
145 6 7
69 ± 2 ***
395 ± 3 ***
333 ± 82 *
341 ± 12 ***
1,686 ± 37 ***
296 ± 2 ***
417 ± 17 ***
945 ± 43 *** ND
50 ± 15 *** ND
386 6 26
2,938 ± 40 ***
391 ± 11 **
1,757 6 906
1,428 6 29
983 ± 17 ***
41 ± 8 *** 148 6 23 134 6 3
635 ± 6 ***
970 6 38
120 6 90
(mg/g DW)T56 2A11TlFM7P119 MMMMhp2j
163 6 7
90 6 6
307 ± 10 ***
272 ± 8 ***
767 ± 14 *** 1,323 ± 13 ***
212 ± 7 ***
239 6 3
204 6 39
697 ± 12 **
245 6 8
224 6 18
175 6 9
928 ± 18 ***
221 6 2
490 ± 21 ***
Values shown in bold are either significantly higher or significantly lower than those of the appropriate background strain. A minimum of three
biological and three technical replicate measurements were performed. The data are presented as means 6 SD. Student’s t tests were used to
determine significant differences between pairwise comparison between the wild-type (T56) and the transgenic varieties as well as the mutant hp2
alleles and their respective wild-type backgrounds; P < 0.05, P < 0.01, and P < 0.001 are indicated by *, **, and ***, respectively. Values in bold indicate
where significant differences have been found compared to the wild-type backgrounds. ND, not detected; MM, Moneymaker; SM, San Marzano.
1196 The Plant Cell
data (Figure 4D) indicated that the plastid area within TFM7 and
P119 cells at the mature green stage was increased ;3-fold,
whereas the increases in 2A11 were more modest (1.5-fold). To
further confirm these findings, a PCR-based assay was devel-
oped to determine the plastome-to-nuclear genome ratio. The
genes used in this assay were the large subunit of ribulose-1,5-
bisphosphate carboxylase/oxygenase (rbcL) for the plastome
and phytoene desaturase (PDS) for the nuclear genome. In the
TFM7 and P119 varieties, the increase in this ratio was 2.5-fold
relative to the T56 wild type (Figure 4E). Therefore, the PCR data
were consistent with the physical parameters of the plastid as
determined by microscopy. For comparison, the hp2Jmutant
allele in the Moneymaker background was analyzed in parallel
with the DET1 varieties (see Supplemental Figure 2 online). The
data sets were comparable to those of the DET1 varieties,
although the fold increases in plastid area per cell were greater
in the TFM7 and P119 DET1 varieties (e.g., 5-fold increases in
TFM7 compared with 3-fold in hp2J). These increases found in
chloroplast-containing tissues were also found in chromoplast-
containing tissues from ripe fruit.
Ultrastructural changes in the plastids of the DET1 varieties
to illustrate the findings from the DET1 varieties, Figure 4A (ii and
v) shows that the chloroplasts typically contain more membra-
nous structures and plastoglobuli in comparison to wild-type
controls. At the ripe stage of fruit development, the most striking
difference was the presence of more and larger plastoglobuli
(Figure 4A, iii and vi).
Metabolomic Analyses of DET1 Downregulated Varieties
Using a combination of analytical platforms, over 120 metabo-
lites were identified and quantified in a relative or absolute
manner. Multivariate principal component analysis (PCA) was
performed to calculate components and the loading contribu-
tions of each metabolite at the mature green and red-ripe stages
among the DET1 genotypes. Most of the variation (40 to 70%)
arose in the first and second components. Scatterplots of
components 1 and 2 showed the clearest grouping of genotypes
(Figure 5). Figure 5A illustrates that at the mature green stage of
fruit development, distinct separation of genotypes occurs with
tight clustering of the biological replicates. The P119 genotype
clusters furthest from the control (T56) in the positive sector of
PC-1, and between these two are the 2A11 and TFM7 geno-
types. Numerous metabolites had significant weightings, sug-
multiple metabolites. It was, however, observed that many of the
metabolites with the highest weightings belonged to the same
compound class. For example, in Figure 5A in the positive sector
of the PC-1 dimension, plastid isoprenoid related compounds,
suchaschlorophyll and carotenoids, were found.Intheopposite
negative sector, amino acids and organic acids were observed.
A similar pattern was found in ripe fruit (Figure 5C) wherein the
control (T56) clustered in the negative sector, with 2A11 closest
to the T56 background and TFM7 closer to the P119 cluster. The
isoprenoids lycopene, b-carotene, and a-tocopherol as well as
phenolics such as rutin were the metabolites with the highest
contributions to variation. A difference between the green and
ripe fruit was the greater separation in PC-2. This was due in part
to the weighting of the phytosterols (sitosterol, campesterol, and
stigmasterol) and amyrins in the negative sectors of dimension
PC-1 and PC-2. The initial targeted pathway analysis of these
genotypes performed in this study demonstrated that many
isoprenoids, phenylpropanoids, andflavonoidswere elevatedas
to ascertain whether similarities and differences in metabolites
between genotypes were due to intermediary metabolism, PCA
was performed again but without metabolites already predeter-
mined to be affected by DET1 manipulation. Figure 5B is the
scatterplot of PC-1 and -2 performed on metabolites from
mature green tissue after removal of those pathways known to
be altered, while Figure 5D is the same but with ripe fruit. It is
clear in comparison to Figures 5A and 5C that the degree of
clustering is less pronounced. The control (T56) and P119
genotypes again show the greatest separation predominantly
along the PC-1 dimension both at green and ripe stages. At the
mature green stage, the TFM7 and 2A11 genotypes cluster
between T56 and P119 genotyopes, but there is separation only
along the PC-2 dimension. In ripe material, 2A11 and TFM7
cluster between T56 and P119, but segregation is greatly re-
duced compared with Figure 5C. The metabolites contributing
most significantly to the differential clustering between geno-
types in the absence of isoprenoids and phenolics were amino
acids and organic acids (including ascorbic acid) and the Calvin
cycle intermediate sedoheptulose.
To investigate further the changes in the metabolomes of the
DET1 genotypes, metabolite changes relative to their control
(T56) levels were determined and statistical analysis performed
to assess the differences (see Supplemental Table 2 online). At
the mature green stage of fruit development, 29, 24, and 32% of
the total metabolites measured were upregulated in the 2A11,
TFM7, and P119 varieties, respectively, while 8, 22, and 29%
were downregulated. In the ripe fruit, 23, 24, and 42% of the
metabolites were upregulated and 17, 20, and 20% downregu-
lated in the 2A11, TFM7, and P119 varieties, respectively. The
changes in metabolites followed a similar trend among all three
DET1 genotypes and in most cases correlated with the degree of
DET1 downregulation conferred by the different promoters. The
changesto themetabolome werenotrestricted to oneorganelle,
with metabolites synthesized in the cytosol, mitochondria, and
plastid all being affected. To visually compare alterations in
relative changes in metabolite levels compared with their re-
spective controls were painted onto biochemical pathway dis-
changes for the P119 variety.
In mature green fruit, amino acids were reduced with the
exception of Ala and Lys, and reductions were typically ;2-, 4-,
and 10-fold in the 2A11, TFM7, and P119 varieties, respectively.
was also reduced in all DET1 genotypes compared with T56.
Sugars, polyols, and organic acids were not greatly altered,
although notable exceptions did occur. For example, the Calvin
cycle intermediate sedoheptulose phosphate was increased
significantly in the 2A11, TFM7, and P119 varieties (2-, 2-, and
10-fold, respectively). Among the organic acids, dehydroascorbic
Analysis of DET1 Downregulated Tomato Fruit1197
acid was elevated 3-, 5-, and 8-fold in the 2A11, TFM7, and P119
varieties, respectively. Further details of relative metabolite levels
in mature green fruit are presented in Supplemental Table 2 and
Supplemental Figures 3A and 4A online.
The amino acids present in the ripe fruit of the DET1 varieties
were at lower levels compared with those in T56 fruits, with the
exception of Ala, which showed up to 2-, 2-, and 10-fold
increases in the 2A11, TFM7, and P119 varieties, respectively.
In the case of organic acids, phytosterols, fatty acids, sugars,
polyols, phosphates, and N-containing compounds, it was dif-
ficult to deduce consistent trends among theDET1 varieties. The
most striking changes in the composition of the metabolites
occurred in the P119 DET1 variety. For example, all fatty acids
were increased 3- to 5-fold in this variety, but no changes in the
Figure 3. Changes in Gene Expression Levels of Some Key Phenylpropanoid and Flavonoid Biosynthetic Genes Resulting from DET1 Downregulation
at Both the Mature Green and Ripe Stages of Fruit Development and Ripening.
At each fruit stage, three individual fruit from three independent plants were pooled and pulverized into a homogenous powder as described in Methods.
Mature green fruit represented 37 to 40 days postanthesis (dpa) and ripe 5 days postbreaker (dpb). RNA was then extracted from an aliquot of this
material and three independent qRT-PCR determinations performed with gene-specific primers as detailed in Methods. Expression data were
normalized to the expression of actin. Data are presented as relative levels found in the three varieties compared with the T56 wild type. Statistical
determinations are provided as means 6 SD value where n = 3 to 6. Student’s t tests have been performed to illustrate statistically significant (*P < 0.05,
**P < 0.01, and ***P < 0.001) differences from the wild-type levels. The solid black bar represents 2A11, the gray bar TFM7, and light-gray bar P119. The
dashed line across each histogram indicates the relative wild-type expression level. PAL, phenylalanine ammonia lyase; CHS, chalcone synthase; CHI,
chalcone isomerize; F3H, flavanone-3-hydroxylase; F39H, flavonoid-39-hydroxylase; F3959H, flavonoid-3959-hydroxylase; FLS, flavonol synthase; DFR,
dihydroflavonol reductase; ANS, anthocyanidin synthase; 3-GT, flavonol-3-glucosyltransferase; RT, flavonol-3-glucoside-rhamnosyl transferase.
These abbreviations for the phenolic biosynthetic enzymes have been used to annotate the pathway illustrating their position in the pathway.#ND, not
detectable. The “#” indicates that the transcript was unique to the DET1 variety and could not be detected in the T56 background; an arbitrary value of
10 has been used in these cases.
1198 The Plant Cell
Figure 4. Changes in Plastid Parameters and Ultrastructure Resulting from DET1 Downregulation.
(A) Panels i and iv illustrate representative control (T56) and DET1 downregulated (TFM7) cells from mature green fruit viewed under Nomarski
microscopy. The solid bars indicate 100 mm. Panels ii, iii, v, and vi are transmission electron microscopy images, with the solid bars indicating 1 mm.
Panels ii and v are images of cells originating from mature green fruit from control (T56) and TFM7 downregulated varieties, respectively. Panels iii and vi
are images of cells originating from ripe fruit from control (T56) and TFM7 downregulated varieties, respectively. pl, plastoglobules; t, thylakoid
membranes; Cr, crystalline structures; mb, membraneous structures. Sections were prepared from three fruit from independent plants, and
representative sections have been illustrated.
(B) Increases in plastid number per cell found in the DET1 varieties as a function of cell area.
(C) Increases in total plastid area per cell regardless of cell area, found in the DET1 varieties compared with their control, are shown.
(D) Cell index calculated from the total plastid area per cell versus the plan area (the plan area being the area of the cell’s projection onto one plan) of the
cell for the DET1 downregulated varieties and their control background. These data demonstrate the increased plastid complement of the cell that
results from DET1 downregulation. Collectively, these microscopy data were compiled from three biological samples counting ;20 cells from three
areas of the pericarp (60 cells) per sample; the data are represented as means 6 SD.
(E) Data for a complementary genetic approach showing an increased copy number of the plastome per haploid nuclear genome in the DET1
downregulated varieties. These data are represented as means 6 SD calculated from three biological samples.
TFM7 and 2A11 varieties were observed. All polyols analyzed
were increased (2- to 8-fold) in the P119 variety, but no changes
were observed in the TFM7 or 2A11 varieties. Concerning
sugars, xylose, arabinose, rhamnose, glucose, gentobiose, and
sedoheptulose phosphate were all increased significantly (3-
fold) in the P119 variety, but only sedoheptulose phosphate
2-oxoglutarate, and dehydroascorbic acid all showed increased
(2-to 9-fold)levels compared withT56,butsuccinicacidwas the
only compound to have a consistent increase among all three
varieties. The levels of individual metabolites relative to their
Figure 5. PCA to Assess the Variance among the Metabolite Composition of the Different DET1 Varieties Compared with Their Background Geno-
Mature green fruit samples are shown in (A) and (B), while (C) and (D) are derived from ripe fruit. (A) and (C) were produced from data sets containing
metabolite variables determined both from targeted analysis (e.g., secondary metabolites) and more untargeted analysis (e.g., primary metabolites
predominantly), whereas (B) and (D) were obtained from untargeted analysis solely. Score and loadings plots were combined with solid symbols
indicating genotypes and open circles representing metabolite variables. Each symbol per genotype indicates a biological replicate. Dashed ellipses
have been overlaid to indicate the clustering of the specific genotypes. T56 clusters are shown as a solid red circle, P119 clusters are shown as solid
blue circles, TFM7 clusters are shown as solid orange circles, and 2A11 is shown as solid green circles. The percentage provided along the axis of (A) to
(D) for each component indicates the amount of variance they account for within the overall data set. Examples of key metabolites with contributions
affecting the dimension of the clustering have been annotated. at, a-tocopherol; a, amyrin; bc, b-carotene; c, campesterol; chl, chlorophyll; ci, citric
acid; dasc, dehydroascorbic acid; fa, fumaric acid; fr, fructose; gl, glycine; lu, lutein; L, lycopene; r, rutin; sd, sedoheptulose phosphate; st, stigmasterol;
si, sitosterol; ma, malic acid; va, valine; mt, mannitol; ne, neoxanthin.
1200 The Plant Cell
control values in T56 ripe fruit are detailed in Supplemental Table
2 online and for P119 are displayed visually in Figure 6B. The
Figures 3B and 4B online.
Interrogation of the metabolite data also revealed relative
changes in metabolites from mature green to ripe fruit. Again the
perturbations showed similar trends among the different varie-
ties, but the effects were greater in the P119 and TFM7 varieties.
The alterations were also not restricted to a specific class of
metabolites or site of synthesis (see Supplemental Table 3
online). Someof themoststriking increasesoccurred in ascorbic
acid and Gln levels, which exhibited 6- and 11-fold relative
increases in the P119 variety from mature green to ripe com-
pared with the control (T56). However, sedoheptulose phos-
phate was an example of a metabolite that showed a relative
decrease in all three varieties during ripening compared with the
Misregulated Transcriptional Networks Resulting from
Transcriptomic analysis using the TOM2 microarray was per-
formed on the DET1 varieties and their control (T56) using RNAs
extracted from fruit at the mature green, breaker, and red-ripe
stages. Validation of the array data was performed by qRT-PCR
for a small subset of transcripts exhibiting a 2- to >50-fold
upregulation in response to DET1 manipulation. These data
indicated a 20% coefficient of variance between the two ap-
proaches, indicating that global trends in transcript levels result-
ing fromDET1 downregulation can beascertained fromthe array
Over all three developmental stages, a major portion of tran-
scripts appeared to be misregulated, with upregulation of tran-
scripts being a dominating feature during fruit development (see
Supplemental Table 4 online). In the P119 and TFM7 varieties at
the mature green stage, 22 and 14% of transcripts were
upregulated, whereas 4 and 5% were downregulated. As pre-
the control of the strongest promoters (P119 and TFM7); thus,
relative global effects of DET1 downregulation on gene expres-
sion during fruit development, a set of transcripts common to all
developmental stages in both the P119 and TFM7 varieties were
selected on the basis of P values. This constituted 309 and 746
transcripts for P119 and TFM7 varieties, respectively (Figure 7).
Boxplot and analysis of frequency distribution for expression
values of all these selected transcripts revealed the overriding
trend for upregulation at all development/ripening stages in both
the P119 and TFM7 varieties (Figures 7A and 7B). The upregu-
lation of transcripts was quantitatively less in the TFM7 variety,
matching the comparative strengths of the two promoters.
Further comparison between the P119 and TFM7 varieties also
indicated other differences in the misregulation of transcripts.
For example, in the P119 variety, upregulation was greatest at
the mature green stageand lowest atthe ripe stage,whereas the
TFM7 variety possessed the greatest abundance of upregulated
transcripts at the breaker stage of development. These data
confirmed the global upregulation found when combining the
mean fold changes for all genes over mature green, breaker, and
red-ripe stages (see Supplemental Table 4 online).
The tendency of progressive reduction in gene expression
during ripening in P119 was also confirmed by self-organizing
maps (SOM) of the data (Figures 7C and 7D). This analysis
resulted in four main trends within the P119 data sets. The first
and largest group (140 transcripts) contained upregulated tran-
scripts at the mature green stage dissipating in intensity during
ripening. Group 2 (81 gene transcripts) was similar to group 1 but
displayed more pronounced downregulation between mature
green to breaker and breaker to ripe. Group 3 again showed the
same trend as groups 1 and 2 but instead a more gradual
reduction in upregulation. The last, group 4, showed patterns
that were different than the negative trends found in groups 1, 2,
and 3. In group 4, upregulation peaked at the breaker stage and
declined thereafter. Interestingly, the latter group represented
only 20% of the transcripts compared with ;80% in the other
groups. SOM analysis on the TFM7 transcriptomic data set
P119 variety. Two major groups, 1 and 4, contain roughly equal
numbers of genes and show completely opposite trends. Group
1 contains 303 genes and shows a general pattern of expression
gene transcripts) shows highest expression at the breaker stage
and a general increasing trend in expression during the ripening
process. To corroborate the SOM analysis, we performed cor-
respondence analysis(CA) ofgene expression data forP119 and
TFM7 (seeSupplemental Figures 5A and 5B online). CA does not
consider a predefined structure for the data set and hence can
represent continuous variation accurately. In close conformity to
the SOM results, CA also clustered gene transcripts together
belonging to the same SOM group. Within each group, we were
not able to find any overrepresentation of transcripts with a
common biological function.
Hierarchical clustering of the array data, displayed in Figures
three main clusters in TFM7. P119 cluster 1 formed with tran-
scripts upregulated at the breaker stage and downregulated at
the mature green stage. Cluster 2 is a heterogeneous cluster,
mostly consisting of transcripts upregulated at the mature green
stage and reduced through ripening. The TFM7 data sets could
be differentiated into two predominant clusters. First, those
transcriptsupregulated atthe breakerstage(cluster1)andthose
upregulated at the early mature green stage (cluster 2). A cluster
TFM7 (cluster 3).
Transcriptional Profiling Reveals Multiple Processes and
Pathways Perturbed in the DET1 Varieties
Expression data were functionally organized into pathways and
processes as detailed in the SolCyc database (http://solcyc.sgn.
at the mature green and red-ripe fruit stages are functionally
displayed in Supplemental Figures 6A and 6B online. These data
show good consistency both between varieties and stages of
fruit development/ripening. For example, at the mature green
developmental stage chlorophyll-, phytyl-PP biosynthesis-, and
Analysis of DET1 Downregulated Tomato Fruit1201
Figure 6. Metabolite Changes Occurring in Mature Green and Ripe Fruit as a Result of DET1 Downregulation under the Control of the P119 Promoter.
1202 The Plant Cell
photosynthesis-related transcripts encoding plastid proteins,
but also transcripts for glycolytic enzymes cytosol-based gly-
colysis, were all upregulated in all varieties. In red-ripe fruit, the
changes in the transcriptome were similar among varieties; for
example, transcripts involved in aerobic respiration/tricarboxylic
acid cycle cycle were upregulated ;25% and down ;13% in
the P119, TFM7, and 2A11 varieties. Thus, the qualitative trends
among transcripts appeared consistent. However, quantitative
differences among varieties were more pronounced, and this
finding is in good agreement with the targeted qRT-PCR data
sets described earlier. To elucidate the hierarchy of pathways
significantly affected by DET1 downregulation, the data sets for
P119 and TFM7 at each of the three developmental stages were
subjected to analysis using Plant MedGen Map (Joung et al.,
2009). More pathways were affected in the P119 variety in
comparison to TFM7. For example, at the mature green stage,
>20 pathways were significantly affected in P119 compared with
only four in TFM7 (see Supplemental Table 5 online). In the P119
variety, the overriding pathways affected were chloroplast-
related biochemical pathways and processes. For example, the
Calvin cycle, photorespiration, and gluconeogenesis were the
most affected pathways at the mature green and breaker stages
and persisted into the ripe stage. The extent of upregulation of
transcripts within these pathways is illustrated in Figure 8. These
pathways were also predominantly affected at the breaker stage
in TFM7 (see Supplemental Table 5 online). Within the TFM7 data
set, transcripts involved in stress-related pathways, such as
jasmonic acid biosynthesis, were also perturbed, with upregula-
tion typically observed.
Integrative Transcript and Metabolite Analysis
To provide insight into the regulatory infrastructure associated
with DET1 downregulation, the transcript and metabolite data
sets were integrated and correlation analysis performed as
described in Methods. These analyses were performed with all
three varieties at the mature green and ripe stages. In contrast
correlation analysis between transcripts revealed both positive
and negative correlations both at the mature green (see Supple-
mental Figure 7 online) and ripe (see Supplemental Figure 8
online) stages. A feature of the data was the blocks of strong
correlations, either positive or negative, among transcripts be-
longing to the same or related pathway or process. This feature
was exemplified by the positive correlations observed among
photosynthesis-related transcripts following DET1 downregula-
respectively). In fact, the positive correlations of multiple path-
ways/processes (e.g., ascorbate, chlorophyll, chorismate, and
starch biosynthesis) relating to photosynthetic processes were a
dominant feature of this analysis both at mature green and ripe
stages (see Supplemental Figures 7 and 8 online, respectively).
Correlation analyses between metabolites in response to
DET1 downregulation are illustrated in Supplemental Figure 9
online for the mature green fruit and Supplemental Figure 10
online for the ripe fruit. A similar effect to that found with the
transcripts was evident for metabolites with components of the
same or related pathways having matching coresponses. Fig-
ures 9C and 9D illustrate the coresponses within the isoprenoid
pathway. Interestingly, at the ripe fruit stage (Figure 9D), there
was a clear differential response between plastid-derived iso-
prenoids and extraplastidial synthesized isoprenoids. In the
mature green fruit, this distinction among different classes of
isoprenoids was not observed. In contrast with the transcript-to-
transcript and metabolite-to-metabolite correlations, significant
agreement in correlations between transcripts and metabolites
in the heat maps presented in Supplemental Figures 11A and
11B online. In this instance, both negative and positive correla-
tions between individual transcripts and metabolites existed, but
ripe stages (Figures 9E and 9F). The exceptions to this observa-
which at the mature green and ripe fruit stages, positive and
negativecorrelations predominated, respectively.To ensurethat
the lack of correlations between transcripts and metabolites was
not solely due to developmental timing, analysis was performed
between transcripts and metabolites determined at different
stages in thedevelopment offruit. Thesedata werein agreement
with the other analysis indicating reduced connectivity between
transcripts and metabolites but did support the findings at
mature green and ripe stages solely with regard to core pro-
Supplemental Figure 11C online illustrates poor coresponses
between processes occurring globally, the exception being
photosynthesis, in which negative correlations between tran-
scripts at the mature green stage and metabolites in the ripe fruit
can be seen. These data are in contrast with the positive
coresponses observed at the mature green stage solely, provid-
ing a logical biological confirmation to validate the data sets and
Figure 6. (continued).
The metabolomic data are displayed quantitatively over schematic representations of biochemical pathways produced with BioSynLab software (www.
biosynlab.com). False color scale is used to display the quantity of each metabolite in P119 relative to that in T56. Pale green indicates a significant
threefold increase, a 3- to 8-fold increase is green, and >8-fold is dark green. Gray indicates no significant change, whereas blue indicates that the
metabolite was not detected in the samples. White indicates that the compound cannot be detected using the analytical platforms available. Red
coloration has been used to represent decreased metabolite levels; dark red is below 8-fold, red is below 2- to 5-fold, and pale red is below 2-fold. Aco,
aconitic acid; L-Asc, ascorbic acid; citramal, citramalic acid; Cit, citric acid; dehydroasc, dehydroascorbic acid; Fum, fumaric acid; Mal, malic acid;
2-oxoglut, 2-oxoglutaric acid; Succ, succinic acid; Thre, threonic acid; 5HT, 5-hydroxytryptamine; 5-OxoPRO, 5-oxo-proline; Arab, arabinose; DXP,
deoxyxylulose-5-phosphate; F6P, fructose-6-phosphate; G6P, glucose-6-phosphate; 3-CaQuinic, 3-caffeoylquinic acid; CGA, chlorogenic acid; FPP,
farnesyl diphosphate; GPP, geranyl diphosphate.
Analysis of DET1 Downregulated Tomato Fruit1203
Figure 7. Transcriptional Misregulation Resulting from DET1 Downregulation.
The stages of fruit development at which the analysis were performed are indicated by MG (mature green), BR (breaker), and RR (red ripe).
(A) Boxplots for log2 ratio (transgenic versus T56) of microarray expression data. P119 is shown in red and TFM7 in blue. Boxes show center quartiles
(middle 50% of the data, whiskers extend to the most extreme data points that are no more than 1.5 times the interquartile range). The outliers are
shown as filled circles.
(B) Frequency distribution of expression values for P119 (red) and TFM7 (blue) across all developmental and ripening stages. The x axis represents log2
expression ratio of the DET1 variety versus T56 (control), indicating the relative changes in transcripts occurring within the data sets. The y axis shows
the percentage of transcripts with a given relative expression value. An overwhelming proportion of expression values are above the T56 control level for
P119, but TFM7 shows significant downregulation.
(C) and (D) SOMs for TFM7 (C) and P119 (D). In each partition, the pattern reflects a general trend of expression gradient of the group across three
1204 The Plant Cell
Biotechnological Implications of the DET1 Downregulated
The targeted metabolite profiling approach used here has de-
termined the diversity and content of the health-related phyto-
chemicals present in DET1 downregulated fruit. The levels of the
individual antioxidants are in most cases comparable to those
achieved by other genetic engineering approaches (Verhoeyen
et al., 2002; Apel and Bock, 2009; Fraser et al., 2009). However,
the unique feature of the DET1 chemotype is the simultaneous
enhancement of multiple antioxidants that originate from di-
verse biochemical pathways and function in both the polar and
nonpolar cellular environments. To date, the only report dem-
onstrating simultaneous increases in multiple nutritional
components is the xenogenic pathway engineering of maize
(Zea mays; Naqvi et al., 2009), in which multiple gene products
sourced from bacteria were used. However, such an approach
has important regulatory restrictions and is less acceptable to
the consumer. In comparison, the DET1 chemotype has been
created using a cis-genic approach with a single endogenous
plant gene product.
The data presented here and reported previously (Davuluri
et al., 2005) also highlight that only small reductions at specific
developmental stages are required (or tolerated) to achieve
significant enhancement of these beneficial phytochemicals.
This suggests that exacerbation of DET1 downregulation would
have a detrimental effect on fruit viability in a manner akin to the
stronger mutations would be lethal to the plant. In fact, the DET1
GM varieties (e.g., 2A11, TFM7, and P119) do exhibit a higher
fold increase in antioxidant compounds compared with their
natural (non-GM) counterparts (e.g., the hp mutant alleles),
without the loss of plant vigor. Collectively, these data would
therefore question TILLING (Triques et al., 2007) or molecular-
assisted selection as effective approaches to deliver high nutri-
ent fruit because modulation of fruit-specific DET1 expression is
unlikely to be achieved by these approaches. Thus, the GM
approach adopted (Davuluri et al., 2005) represents the most
plausible strategy if optimal levels of the antioxidants in question
are the predominant criteria. Beyond tomato, the potential lethal-
utility in increasing multiple antioxidants in leafy vegetable crops.
The antioxidant assays performed demonstrate that the in-
creases in metabolites could be translated to increased antiox-
idant capacity of the fruit both in the polar and nonpolar phases.
The increases in antioxidant capacities were either comparable
or, if polar and nonpolar activities were combined, greater than
the polar activities reported as a result of high flavonoid or
hyperanthocyanidin production (Butelli et al., 2008; Luo et al.,
2008). Presumably, this suggests that either other polar antiox-
idants are elevated in the DET1 downregulated varieties or that
saturation of antioxidant activity can occur at high concentra-
tions in endogenous extracts.
The Effect of DET1 Downregulation on the Global
The DET1 gene product is predominantly involved in the trans-
adaptation of processes and metabolisms to the plants’ sur-
roundings, in this case light (Scha ¨fer and Bowler, 2002). The
effects of DET1 are implemented through its ability to modulate
the stability of transcription factors, some of which have multiple
potential gene targets (Osterlund et al., 2000; He et al., 2007).
in vivo (Benvenuto et al., 2002) and may therefore alter the
chromatin context of DNA, potentially exposing the DNA around
target genes to facilitate transcription. Therefore, it is not sur-
prising that the effects of perturbation of DET1 are widespread
throughout metabolism as witnessed from the metabolomic and
transcriptomic analyses herein performed. It can be envisaged
onto chromatin/DNA could result in (1) transcription of global
regulators and or transcription factors, (2) mediation via signal
transduction proteins, (3) release of attenuating transcription
factors or coordinated transcription of genes, and (4) direct
action on genes encoding biosynthetic pathway components. In
addition, the balance of the small molecules generated by these
changes may in turn further affect transcription of related genes.
The predominance of upregulated gene transcripts revealed
by transcript analysis suggests that DET1is anegative regulator.
The transcription of global regulators associated with ripening
(e.g., nor, rin, or CNR; Giovannoni, 2004) as well as those
by DET1 downregulation, suggesting that within the limits of the
TOM2 array the chemotype of DET1 downregulation was not
initiated by a limited number of cardinal transcription factors.
Instead, the correlation analysis of transcripts revealed blocks of
transcripts responding in a coordinated manner to DET1 down-
regulation. These blocks consisted of genes with related func-
tionality. In addition, a strong degree of connectivity was
observed among these pathway and process components.
Thus, coordinated transcriptional activation would appear to
Figure 7. (continued).
developmental stages with vertical bars showing the variance in the group at each stage. A gene is assigned to a single partition with similar groups
placed in nearby partitions.
(E) and (F) Heat maps of the gene expression data for TFM7 (E) and P119 (F). Each horizontal line represents the gene expression across the three
developmental stages. For each gene, log2 expression ratios of DET1 variety versus T56 are normalized across the three stages. Red indicates
upregulation and blue downregulation with respect to the T56 background. The z-scores have been used to indicate the deviation from normal
distribution (the distribution standard derivation) and calculated from the variable’s value minus the population’s mean divided by the SD of the
population. The vertical color bars next to gene trees indicate genes belonging to SOM classes in (C) and (D).
Analysis of DET1 Downregulated Tomato Fruit1205
play an important role in the implementation of the phenotypes
associated with DET1 downregulation. These findings are in
agreement with the genome-wide coexpression networks eluci-
dated from Arabidopsis thaliana data sets (Wei et al., 2006). In
these studies, the strongest transcriptional coordination exists
among and between core pathways such as those related to
photosynthesis (e.g., Calvin cycle and photorespiration). The
transcriptional coexpression analyses performed with the DET1
data sets corroborate these findings. Likewise, correlation anal-
ysis between metabolites of related pathways and among me-
tabolites of the same pathway exhibited significant coordination
Figure 8. The Effect of DET1 Downregulation on Photorespiration-Related Gene Expression Levels in P119 Mature Green Fruit.
(A) Pathways and processes involved in photorespiration; steps in the pathways where transcripts have been measured are numbered.
(B) Changes in gene expression levels relative to levels determined in control (T56 background) samples. Data from microarray analysis have been
used; the experimental design of these experiments is provided in Methods. The data are expressed as means 6 SD. Transcripts are labeled as follows,
1, ribulose-1,5-bisphosphate carboxylase small chain 3b; 2, glycolate oxidase; 3, hydroxypyruvate reductase; 4, glycine hydroxymethyl transferase; 5,
transketolase; 6, phosphoglycolate phosphatase; 7, photosystem II protein 16; 8, photosystem II 22-kD protein; 9, photosystem II 5-kD protein; 10,
photosystem II psbY; 11, photosystem II reaction center; 12, photosystem I subunit II; 13, photosystem I subunit III; 14, photosystem I subunit VI; 15,
photosystem I subunit X; 16, photosystem I subunit psaN. Pa, pheophytin; Pq, plastoquinone; Cyt, cytochrome bf complex; Pc, plastocyanin; Fd,
ferrodoxin; 3-PGA, 3-phosphoglycerate; GA-3-P, glyceraldehyde-3-phosphate. Stoichiometries of Calvin cycle components shown in parentheses.
1206 The Plant Cell
Figure 9. Selected Heat Maps Showing Intrapathway/Process Correlations between Photosynthesis-Related Transcripts and Isoprenoid Metabolites.
(A) and (B) Coresponses between photosynthesis-related transcripts in mature green and ripe fruit, respectively.
(C) and (D) Coresponses between isoprenoid metabolites in mature green and ripe fruit, respectively.
(E) and (F) Coresponses between isoprenoid metabolites and photosynthesis-related transcripts in mature green and ripe fruit, respectively. A false
color scale is used to indicate positive (green) and negative (red) correlations. Stringent cutoff coefficient values of either 0.8 or ?0.8 have been used
with a significance of P < 0.05. Pearson correlation coefficients (r) were calculated using data sets for all the DET1 genotypes, with triplicate biological
replication per genotype. Calculations were performed using BioSynLab software (www.Biosynlab.com)
evident for metabolites involved in core processes. Photosyn-
thesis-related transcripts also exhibited positive coresponses
with isoprenoid metabolites at the mature green developmental
stage. These data support the essential role of isoprenoids like
chlorophyll and xanthophylls in photosynthesis as well as phy-
tosterols in cell elongation and development. A complementary
set of coresponses were observed at the ripe stage in which
photosynthesis-related transcripts possessed a negative corre-
and physiological requirements of fruit during ripening with a
shutdown of photosynthesis, cell elongation, and accumulation
of chromoplast-associated pigments with no photosynthetic
functionality (Gillaspy et al., 1993). Despite the correlations
observed between photosynthesis-related transcripts and iso-
prenoids, a global lack of correlation between transcripts and
metabolites is evident. These findings do not support the tradi-
lation (protein levels), enzyme activity, and metabolite content.
Instead, it would appear that in the case of DET1, posttranscrip-
playing a key role in modulating transcription either. Perhaps the
of translation. More data collection at the different regulatory
levels will possibly elucidate such mechanisms. These data also
suggest that the isolated use of transcriptomic data for the
elucidation of regulatory networks is restrictive.
The Transcript and Metabolite Changes Associated with
DET1 Downregulation Reflect Its Role in Mediating
Responses to Light
The effect of light on plant development has been studied
extensively in many species (Terzaghi and Cashmore, 1995). A
number of pathways and processes are known to be transcrip-
tionally activated by light (Ghassemian et al., 2006). These
include photosynthesis, the Calvin cycle, chlorophyll, and flavo-
noid biosynthesis, as well as a number of stress-related pro-
cesses (Cominelli et al., 2008). In the DET1 downregulated fruit,
all these processes/pathways are upregulated and show a
positive inter- and intrapathway correlation. This supports pre-
vious transcript analysis performed on hp2 alleles (Kolotilinet al.,
2007). Therefore, in effect, it would appear that a pseudostate of
light hyperresponsiveness has been created in developing to-
mato fruit via DET1 downregulation, mimicking the molecular
responses to high light conditions, as already observed for
several Arabidopsis photomorphogenic mutants (including
det1-1) (Ma et al., 2003). This in turn has resulted in the coordi-
nated transcription of light-related biochemical pathways and
processes, in particular those associated with photosynthesis.
This shift in or enhancement of existing metabolic events is
perfectly plausible from a biochemical viewpoint because the
enhanced perception of light could signal a shift toward increas-
ing light use by the photosynthetic apparatus. The subsequent
products of the photosynthetic light reactions are ATP and
NADPH2, which sustain the Calvin cycle, facilitating the fixation
of CO2into sugar. The Calvin cycle in turn is a component of
photorespiration. These primary events mediated by DET1 also
provide an explanation as to why the DET1 downregulated
phenotype arises primarily in developing chloroplast-containing
tissues and why the timing of promoter control is crucial. For
DET1 to mediate its metabolic effects, the cell type should be
primed for core metabolic events (e.g., photosynthesis). Pre-
sumably this is also why such large differences in metabolite and
transcript profiles are found in spite of only subtle differences in
the developmental profiles of DET1 downregulation under the
control of the P119 and TFM7 promoters. It is therefore not so
surprising that ripening-specific downregulation of DET1 has no
phenotypic effect (Davuluri et al., 2005) because these special-
ized chromoplast-containing tissues are not designed for vege-
tative core metabolic events.
With the dampening effects of DET1 attenuated through
downregulation, it is expected that a number of stress-related
transcripts would also be upregulated, as the plant effectively
perceives greater light incidence, creating the potential for high
light stress. For example, the formation of phenolics and ascor-
bate along with transcriptional changes in lipoxygenase and
jasmonate biosynthesis pathways have all been associated with
light-related stress responses (Youssef et al., 2009).
Photomorphogenesis has been intensively studied (Ma et al.,
2001; Foo et al., 2006; Ghassemian et al., 2006; Lopez-Juez
et al., 2008) and specifically the role of DET1 in light adaptation
(Chory et al., 1989; Schroeder et al., 2002). The transcriptional
and metabolite changes determined in the mature green fruit of
the DET1 varieties were in general agreement with these studies.
Therefore, in developing tomato fruit, the changes occurring are
likely in part to reflect the function of DET1 in chloroplast-
containing tissues and its consequences.
How Does the Perturbation of DET1 Lead to Increased
Antioxidant Content in Tomato Fruit?
Integrative analysis of transcripts and metabolites has high-
lighted core intermediary processes and stress responses as the
progenitors of increased antioxidants in fruit. Interestingly, the
antioxidants in question are classical secondary metabolites in
tomato. Therefore, (1) how does DET1’s impact on core pro-
cesses affect secondary metabolites? And (2) why do its effects
persist after downregulation ceases?
It would appear that the flavonoids and anthocyanidin path-
ways are transcriptionally regulated at influential steps in the
pathway (e.g., CHS and DFR in the case of flavonoid and
anthocyanidin, respectively). In plant tissues, these compounds
have a protective role against the damaging effects of high light
incidence. Their upregulation has been demonstrated previously
in response to light and other stress conditions (Cominelli et al.,
2008; Lopez-Juez et al., 2008) as well as in Arabidopsis det1
mutants. In the particular case of CHS, transcriptional derepres-
sion could be shown using a CHS:GUS transcriptional fusion in a
det1-1 background (Chory and Peto, 1990). Therefore, consid-
of light hyperresponsiveness and abiotic stresses, it is logical
that the elevation of flavonoids and anthocycidins would arise.
Whether DET1 has a direct or indirect effect on transcriptional
upregulation of these pathways awaits elucidation, although the
altered transcript levels of several MYB and bHLH transcription
1208 The Plant Cell
factors, which are well-established players in the regulation of
flavonoid pathways (Allan et al., 2008), suggest that an indirect
effect is likely. Altogether, close to 30 putative transcription
factors were misregulated in P119 fruits, further suggesting an
indirect effect of DET1 ontranscription.It was alsonoted that the
gene encoding GIGANTEA, a nuclear protein associated with
light signaling in Arabidopsis (Huq et al., 2000), is downregulated
in the TFM7 and P119 varieties.
In contrast with phenolic biosynthetic pathways, none of the
transcripts encoding carotenoid and general isoprenoid biosyn-
thetic components are upregulated in the DET1 varieties, sug-
gesting that posttranscriptional regulation is operational.
Notwithstanding, the upregulation of transcripts encoding to-
copherol biosynthetic steps does suggest that transcriptional
regulation is involved in this related branch of the pathway. The
abundance of transcripts encoding plastid-related proteins and
the confirmatory determination of an increased plastid compart-
ment per cell are the likely explanation for increased carotenoid
formation in these varieties. Presumably the increase in plastid
area per cell creates a collective increase in biosynthetic capac-
ity. The reason for an increased plastid cellular compartment is
unlikely to be a specific effect, but rather the coordinated effects
of DET1 on plastid-related events and the need for increased
plastid biogenesis per se. For example, the expression of neither
transcription factors involved in plastid biogenesis (e.g., GLK-
like; Fitter et al., 2002; Yasumura et al., 2005) nor structural
proteins involved in plastid biogenesis (e.g., ribosomal proteins)
or plastid division (FtsZ and MinD) (Aldridge et al., 2005) were
data contradict findings with the natural alleles and may reflect
with aspects of vegetative development. The effects on core
metabolic processes are, however, common to the DET1 vari-
eties, hp alleles (Kolotilin et al., 2007), and photomorphogenesis
studies in Arabidopsis (Ghassemian et al., 2006). Collectively,
these data could suggest that it is the initiation of core metabolic
processes that drive subsequent plastid biogenesis to provide a
defined cell compartment for these processes. The division or
premature differentiation (Fraser et al., 2007; Maass et al., 2009)
of plastids does not appear to be dramatically affected by DET1
modulation, as increased plastid numbers with a reduced area
were not evident and detectable levels of chromoplast-specific
carotenes were not found.
Another explanation for global increases in antioxidants could
be that increased metabolic activity within the cell leads to the
To dissipate the detrimental effects of these molecules, the
synthesis of protective antioxidants, such as carotenoids, to-
copherols, phenolics, and ascorbate, would be advantageous
(Foyer and Noctor, 2005). Therefore, the increases in antioxi-
dants observed could simply result from the cells’ natural pro-
tective mechanisms against free radical imbalances. The
increases in some of the antioxidants, such as carotenoids,
could also have secondary benefits in maintaining levels of other
related compounds. For example, the increased pool of carot-
enoids couldwellprovideprotection againstchlorophylldamage
and contribute to the persistence and increased levels of chlo-
rophyll. Elevation of ROS can also lead to the activation of stress
responses (Foyer and Noctor, 2005). Several of the stress-
related pathways misregulated by DET1 have been associated
with ROS generation (e.g., jasmonate formation and lipoxyge-
nase activation) (Vanderauwera et al., 2005). Interestingly, the
activity of lipoxygenases initiate the formation of jasmonates
(Wasternack, 2006), and the latter have also been implicated in
altering chloroplast parameters within the cell as well as the
levels of anthocyanidins and other antioxidants (Jung, 2004;
stress-related phytohormone abscisic acid in the hp3 mutant,
affected in zeaxanthin epoxidase (Zep), has been shown to alter
the total chloroplast area per cell, which presumably leads to the
be informative to ascertain in this case if similar perturbations in
metabolism arise and, if so, to correlate them with those ob-
served with DET1 downregulation.
To use the effect of DET1 to generate antioxidants, this study
and previous studies (Davuluri et al., 2005) indicate that down-
regulation in developing fruit tissues with active chloroplast
proliferation is the key. However, the effects remain or are
exacerbated in ripe fruit after DET1 expression has ceased.
Presumably this arises from the metabolic changes established
at the earlier developmental stages. It is feasible that through
increasing core metabolic events in developing fruit, a greater
pool of primary metabolites for use in secondary metabolic
pathways could exist. In addition, more reductant may be
generated to drive associated secondary pathways. For exam-
ple, glyceraldehdye-3-phosphate is a component of the Calvin
cycle and a precursor of the plastidial isoprenoid biosynthetic
pathway. Our metabolite profiling illustrated the large increases
insedoheptulose phosphateoccurring inthispathway.Through-
out isoprenoid biosynthesis there are also numerous steps that
use reductant in their catalysis. It is also feasible that the
increased plastid area improves sequestration of metabolic
end products eliminating inhibitory feedback of the biosynthetic
pathway. In support of this, the misregulation of transcripts did
reveal that certain core intermediary metabolic processes per-
sisted into ripening fruit in comparison with their controls (e.g.,
photosynthesis and the Calvin cycle). Such an extension of their
developmental effectiveness could also play a role in delivering
more usable precursors for metabolism in the ripening fruit.
In conclusion, the integrative transcript and metabolite ap-
proaches used have proven invaluable in providing insight into
the changes in the cell’s regulatory infrastructure and reprogram-
ming of metabolism associated with DET1 downregulation. The
sequence of events leading to the simultaneous enhancement of
in a systematic manner using conventional biochemical ap-
proaches. These approaches will permit the elucidation of further
Generation and Preparation of Plant Material
Tomato (Solanum lycopersicum) varieties used included the T56 indus-
trial processing genotype as the background for the P119, TFM7, and
2A11 DET1 downregulated genotypes (Davuluri et al., 2005). The hp2j
Analysis of DET1 Downregulated Tomato Fruit1209
mutant allele in the Moneymaker variety background and the hp2 mutant
in the San Marzano background were also used. Details of the pheno-
types and recommendations for growth can be found on the Tomato
Genetic Resource Centre (TGRC) at www.tgrc.ucdavis.edu. Seeds for
obtained from TGRC. Tomato plants were glasshouse grown from seed
under supplementary lighting. Routinely, six plants per genotype were
grown in a randomized manner concurrently with their respective back-
grounds. Figure 1 illustrates the designations used to defined the stages
of fruit development and ripening used in this study. The breaker stage
and ripe fruit as 7 d postbreaker.
Measurement of Gene Expression by Real-Time qRT-PCR
at a similar timein the day. Tissue was prepared by cutting the fruitin half,
the seeds and jelly removed, and then frozen immediately in liquid
nitrogen. These fruit samples were pooled and broken down in a mortar
and pestle. Complete homogenization into a powder was then performed
using a freezer mill (CertiPrep/6750 apparatus; SPEX). For quantitative
real-time RT-PCR determinations, RNA was extracted from 100 mg of
milled tissue, while RNAs for microarray analysis were extracted from the
same material but in large quantities as described in subsequent sec-
tions. The RNeasy reagents and protocol (Qiagen), including on-column
time RT-PCR kit (Qiagen) was used to determine gene expression levels.
Determinations used 25 ng of RNA, and primers were added to provide a
were performed on a Rotor-Gene 3000 thermocycler (Corbett Research).
Thermocycling conditions were 30 min at 508C for reverse transcription,
15 min at 958C, followed by 30 to 40 cycles of 15 s at 948C, 30 s at 568C,
and 30 s at 728C. Sequencing of PCR products and melt curve analysis
verified their specificity. For quantification, calibration curves were run
simultaneously with experimental samples, and Ct calculations were
performed by Rotor-Gene software (Corbett Research). The actin gene
served as reference for normalization. Primers for real-time qRT-PCR
were designed using Primer3 software (http://biotools.umassmed.edu/
bioapps/primer3_www.cgi). The primers used for the PCR amplification
are provided in Supplemental Table 6 online.
Determination of Plastome: Genome Ratios by Real-Time PCR
The gene encoding the large subunit of ribulose-1,5-bis-phosphate
carboxylase/oxygenase (rbcL) was used as a plastome marker, while
the PDS gene was used as the nuclear genome marker. A PCR-based
assay was developed to determine the ratios between the two and thus
provide a rapid quantitativeindicationof changesin theplastid content of
the cell. Typically, three fruit from independent plants of the same
genotype were sampled as described in the previous section. DNA was
extracted from 100 mg of powdered sample using the DNeasy reagents
and protocol (Qiagen). The QuantiTect SYBR GreenPCR kit (Qiagen) was
used to determine the number of copies of each gene per 20 ng of DNA in
each sample. Primers were added to a final concentration of 0.3 mM in a
reaction volume of 20 mL. Reactions were performed on a Rotor-Gene
3000 thermocycler (Qiagen). Thermocycling conditions were 958C for 15
min followed by 40 cycles of 15 s at 948C, 30 s at 508C, and 15 s at 728C.
Sequencing of PCR products plus melt curve analysis verified the
reactions’ specificity. For quantification, calibration curves were run
simultaneously with experimental samples, and actin was used as a
reference gene for normalization of the data. The primers used for the
PCR amplification are provided in Supplemental Table 6 online.
Transcriptional profiling was performed using the TOM2 long oligonucle-
cornell.edu). Biological replication for transcriptome analysis was created
using a minimum population of six plants per genotype further sorted into
pairs of plants. From each plant within a given pair, three fruit were
minimum of three such pools were generated for each genotype from fruit
harvested at the mature green, breaker, and ripe stages. Pericarp tissue
(including the peel) was prepared from fruit using a freezer mill as
described earlier. The expression profiling experiments were designed
the respective T56 fruit stage as the control. Data were generated from a
minimum of three biological replicates, each with a duplicate dye swap
and creating a total of between six and eight hybridizations.
Protocols used for RNA extraction, cDNA synthesis and labeling,
microarray hybridizations, and scanning are all available online at the
Tomato Functional Genomics Database website (www.ted.bti.cornell.
edu). In brief, total RNA extraction was performed from frozen powdered
pericarp tissue (;5.0 g) as described by Griffiths et al. (1999). cDNA
synthesis and labeling with Cy3 and Cy5 (Cy Dye Postlabeling Reactive
Dye Pack; GE Amersham) was performed indirectly using the amino-ally
method (Invitrogen SuperScript indirect cDNA labeling kit) according to
Alba et al. (2004). Hybridizations of Cy-labeled cDNA targets to the TOM2
oligoarray were performed at 428C for 14 to 16 h in the dark. Following
hybridization, slides were washed twice in 13 SSC containing 0.2% (w/v)
SDS at 438C for 3 min and then washed in 0.13 SSC containing 0.2%
(w/v) SDS at room temperature for 3 min. Slides were finally dipped in
0.13 SSC containing no SDS to remove excess SDS before washing
three more times for 3 min each in 0.13 SSC solution. Scanning of arrays
was performed on a two-channel confocal microarray scanner (ScanAr-
ray5000; GSI Lumonics) with ScanArray software v3.1 (Packard Biochip
Technologies). Scans were performed at a resolution of 10 mm. Excita-
tion/emission settings were 543/570 nm and 633/670 nm for the Cy3 and
5 fluors, respectively. Numerical images were generated from the micro-
array images using ImaGene software (version 5.5; BioDiscovery). Those
quality were flagged by the software and not included in subsequent
statistical analysis. Normalization used the print-tip strategy and was
applied to ratio values for each array using the marray package in
Bioconductor (Yang et al., 2002). Differentially expressed genes were
identified using patterns from gene expression (Grant et al., 2005). Genes
with false discovery rates of <0.1 and fold change $2-fold were identified
as differentially expressed genes. Functional annotations were assigned
according to Alba et al. (2005).
Determination of Plastid Parameters in Intact Cells
Tomato fruit tissues were harvested and fixed immediately in 3.5% (v/v)
glutaraldehyde solution for 1 hin the dark (Cookson et al., 2003). Pericarp
cells were prepared from green fruit by heating at 608C for 15 min in a
0.1 M Na2-EDTA solution and ripening fruit by storage in Na2-EDTA at
room temperature. Nomarski differential interference contrast opticson a
Nikon Optiphot microscope in conjunction with Lucia image software
were used to determine plastid parameters per cell. Total plastid area per
cell and plastid number per cell were determined. Cell index parameters
were calculated as a measure of the total plastid compartment size in
relation to cell size, calculated as total plastid area per cell/cell plan area.
The cell plan area is the area of the top view of the cell.
Tissue for electron microscopy was prepared from fruit derived from
three plants of identical genotype. Pericarp tissue was fixed in 4% (v/v)
1210 The Plant Cell
formaldehyde, 3% (v/v) glutaraldehyde buffered in 0.1 PIPES, pH 7.2, for
1.5 h at 48C, and then rinsed in buffer. The same buffer containing 1%
(w/v) OsO4was used to fix the tissues at 48C overnight. Samples were
repeatedly washed with aqueous ethanol (10 to 50% [v/v]) and finally
dehydrated with ethanol washes. Embedding used Spurr’s resin and
sectioning was performed. Sections were stained with ethanol-saturated
uranyl acetate and Reynold’s lead stain. Transmission electron micros-
copy was performed on an EM109 (Zeiss).
Preparation of Tomato Tissue
The tomato fruit tissue used for metabolite analysis was derived from the
same batch as that used for molecular analysis and thus prepared in an
identical manner, with the exception that the material was freeze-dried.
Metabolite analysis was performed on a minimum of three (up to six)
biological replicates, with up to four technical replicates.
Carotenoid and Isoprenoid Analysis
The extraction, HPLC separation, photodiode array (PDA) detection, and
quantification of carotenes, carotenoids, xanthophylls, tocopherols, and
quinones has been described in detail previously (Fraser et al., 2000). A
10- to 20-mg aliquot of ground homogeneous fruit tissue was extracted
with chloroform and methanol (2.5:1 by volume) after mixing and incu-
A partition was formed after mixing by centrifugation. The organic
hypophase was removed and the aqueous phase reextracted with
chloroform (2.5 by volume). HPLC separations were performed using a
C30reverse-phase column (250 3 4.6 mm) purchased from YMC. The
mobile phases used were methanol (A), water/methanol (20/80 by vol-
ume) containing 0.2% ammonium acetate (B), and tert-methyl butyl ether
(C). The gradient used was 95% A:5% B, isocratically for 12 min and then
B:65% C over 30 min was performed. A Waters Alliance model 2695
continuously from 220 to 700 nm with an online PDA (Waters 966).
Identification was performed by cochromatography and comparison of
spectral properties with authentic standards and reference spectra
(Britton, 2004). Annotated chromatograms of this HPLC system have
been provided elsewhere (Fraser et al., 2007). Quantitative determination
of carotenoids was performed by comparison with dose–response
curves (0.2 to 1.0 mg) constructed from authentic standards. Purchasing,
preparation, and characterization of authentic standards have been
described previously (Fraser et al., 2000; Long et al., 2006). Total
carotenoids along with chlorophyll A and B were determined spectro-
photometrically as described by Wellburn (1994).
Flavonoids and Phenylpropanoid Analysis
Procedures used to extract, separate, detect, and quantify phenylpro-
panoids and flavonoids from freeze-dried tomato material have been
standard (10 mg). The mixture was heated to 908C for 60 min after cooling
the suspension centrifuged at 3000g for 5 min the supernatant removed
and extracts passed through a 0.2-mm syringe fitting filter before HPLC
analysis. An Agilent 1100 HPLC solvent delivery system was used to
separate phenylpropanoids and flavonoids on a reverse-phase C18
column (250 mm 34.6 mm; 5 mm; Hichrom). The mobile phase consisted
of (A) 2% water in methanol acidified with 0.015% HCl by volume and (B)
acetonitrile. The initial gradient conditions used were 95% A, 5% B for 10
min, followed by a linear gradient to 50% B over 30 min. An online PDA
enabled detection and identification from characteristic UV/Vis spectra.
Authentic standards were used to confirm the identity of the phenyl-
propanoids and flavonoids. Relative quantification was carried out by
comparison of integrated peak areas with the internal standard at the
lmax of the phenylpropanoids and flavonoids detected. Total anthocy-
anins were extracted and quantified as described by Martin et al. (1985).
Identification of the predominant anthocyanin was performed by liquid
chromatography–electrospray ionization–mass spectrometry using iden-
tical conditions to those described above withtheinclusion offormic acid
instead of HCl in the mobile phase.
Total phytosterols were routinely extracted from ground tomato powder
prepared in the freezer mill as described above. b-Cholestanol was
added asan internal standard(20 mg)and the suspensionsaponifiedwith
5% KOH in 80% ethanol (5 mL) for 1 h at 908C. When cooled, this solution
was neutralized with concentrated sulphuric acid (;200 mL). The phyto-
sterols were extracted with hexane (2 3 3 mL), adding water (5 mL) to
create a partition. The pooled hexane extracts were dried under nitrogen
and the residue redissolved in acetonitrile (20 mL) and derivatized by
adding 20 mL N-methyltrimethylsilytrifluoroacetaminde (Sigma-Aldrich).
This mixture was incubated at 378C for 1 h. Separation and detection of
phytosterols were performed with a Varian CP3380 gas chromatogram
containing a flame ionization detector and fitted with an autosampler
CP8410, and separation was performed on a CP-SIL8 column (Varian).
The temperature program used consisted of 1408C initial conditions
increasing to 2408C by 27 min at 48C per min and holding for 2 min. These
conditions were then held for another 18 min. The injector (Varian 1177)
temperature was set at 2608C. Quantification was performed from inte-
grated peak areas relative to a known amount of the internal standard.
Extraction, Derivatization, and Gas Chromatography–Mass
Spectrometry Analysis of Metabolites
Methanol (1 mL; HPLC grade) and the internal standard ribitol (20 mg/mL)
were added to freeze-dried (20 to 25 mg) tomato powder. The material
and the suspension was mixed vigorously and then incubated at room
temperature for 30 min with continuous agitation. To remove cell debris,
samples were centrifuged at 12,000 rpm for 2 min, the resulting super-
natant was removed, and from each sample an aliquot (100 mL) removed
and dried completely under nitrogen gas. At this point, the extracts could
be derivatized immediately or stored without degradation at 2208C.
(Sigma-Aldrich) prepared at a concentration of 20 mg/mL in pyridine.
After incubation in screw-capped tubes at 378C for 2 h, N-methyltrime-
thylsilytrifluoroacetaminde from Macherey Nagel was added (85 mL) and
the samples incubated for a further 1 h at 378C before analysis. Gas
chromatography–mass spectrometry analysis was performed on an
Agilent HP6890 gas chromatograph with a 5973MSD. Typically, samples
(1 mL) were injected with a split/splitless injector at 2908C with a 20:1 split
and repeated on a 200:1 split for sugar quantification. Retention time
locking to the internal standard was used. The gas chromatography oven
was held for 4 min at 708C before ramping at 58C/min to 3108C. This final
temperature was held for a further 10 min, making a total time of 60 min.
The interface with the MS was set at 2908C and MS performed in full scan
mode using 70 eV EI+ and scanned from 10 to 800 D. To identify
chromatogramcomponentsfoundin the tomato profiles, amassspectral
(MS) library was constructed from in-house standards as well as the NIST
automated MS deconvolution and identification system. A retention time
calibration was performed on all standards to facilitate the determination
Analysis of DET1 Downregulated Tomato Fruit1211
of retention indices (RIs). Using the retention indices and MS, identifica-
tion was performed by comparison with the MS library. Quantification
was achieved using Chemstation (Agilent) software facilitating integrated
peak areas for specific compound targets (qualifier ions) relative to the
ribitol internal standard peak.
Total Antioxidant Activity Assays
A nonpolar extract was generated using the carotenoid/isoprenoid ex-
dried tomato tissue used was 10 mg. Polar extracts were produced using
the method described for phenylpropanoid and flavonoid extraction, with
was added, the sample was not heated, and incubation time was
extended to 2 h at room temperature in the dark. TEAC assays were
ethylbenzothiazoline-6-sulphonic acid) (ABTS; purchased from Sigma-
Aldrich), free radical cation, ABTS+. To generate this reagent, an ABTS
stock solution (7.0 mM) was mixed with 2.45 mM dipotassium peroidi-
sulfate (Sigma-Aldrich) in a 1:1 ratio. The mixture was left in the dark for
12 h at room temperature. This ABTS-free radical cation solution formed
was diluted in ethanol, typically 8 mL ABTS+in ethanol (1 mL). An
absorbance of 0.70 Abs at 734 nm should result. Sample (10 mL) was
added to 1 mL ABTS+solution and the Abs 734 nm recorded. Results
were compared with the ability of Trolox (a water-soluble vitamin E
analog) to quench the ABTS+radical. Results were expressed as a TEAC
in mmol of Trolox per gram of DW.
Data Processing and Statistical Treatment
technical replicates unless stated otherwise. Metabolite levels from the
different technology platforms were combined. PCA analysis was per-
formed on these data matrixes. SPSS software version 12.01 (SPSS) and
SIMCA-P+ (Umetrics) were used to carry out and display clusters derived
from PCA analysis. Student’s t tests were used to determine significant
differences between pairwise comparisons among the different DET1
varieties with their control (T56). Where appropriate, P < 0.05, P < 0.01,
means, and standard deviations were all calculated using GraphPad
Prism software (GraphPad Software) or Excel (Microsoft) embedded
algorithms. The overlaying ofmetabolite dataover biochemicalpathways
was performed using BioSynLab software (www.biosynlab.com) from
csv Excel files.
The mean and P values for 2A11, TFM7, and P119 transcriptome data
sets were calculated from all the expression ratios including three
biological and two technical replicates, each with a dye-swap, and
redundantsignals, and theunique transcripts having P values <0.05 were
retained for the final analysis. Boxplot and frequency of distribution
analysis was performed on a set of transcripts common to all develop-
mental stages in both the P119 and TFM7 varieties, selected on the basis
of P values (309 and 746 transcripts for P119 and TFM7, respectively).
The boxplots, histograms, and multivariate analysis (CA and SOM) of
gene expression data were performed within the R statistical program-
ming environment (www.r-project.org) using packages MADE4 available
from Bioconductor. The hierarchy of significantly affected pathways in
transcriptomic data was established using Plant MetGenMap (Joung
et al., 2009). For each transcriptome analysis, the plate IDs available for
the entire set of transcripts data with P value <0.05 has been usd. The
software uses hypergeometric test to check the significance of pathway
changes, performed on a set of pathways simultaneously using P values
that correspond to raw P values corrected by false discovery rate.
Unprocessed microarray data are available at www.mycib.ac.uk/zope/
Correlation analysis between metabolites and transcripts (variables)
was performed by calculating Pearson’s correlation coefficients. Bio-
ware programs were used to calculate correlation coefficients. A P value
<0.05 was used to judge the significance of the correlation coefficients.
Positive and negative correlations have been represented as heat maps
using a false color scaling. To obtain greater robustness of significance
correlation analysis, data sets for all the DET1 genotypes have been used
with triplicate biological replication per genotype. Transcripts common to
Set 1 online.
The microarray and metabolomic data sets are available through the
BiotechnologyandBiological ScienceResearch Council (BBSRC)Centre
for Plant Integrative Biology at www.mycib.ac.uk/zope/TomSysBio. The
accession numbers and properties of the genes used in microarray and
correlation analysis are provided in Supplemental Data Set 1 online.
Supplemental Table 6 online contains the accession numbers for the
genes used in qRT-PCR analysis.
The following materials are available in the online version of this article.
Supplemental Figure 1. Determination of Both Lipophilic and Hy-
drophilic Antioxidant Activity in Both Mature Green and Ripe Fruit.
Supplemental Figure 2. Altered Plastid Parameters Resulting from
the hp2jMutant Allele.
Supplemental Figure 3. Changes in Metabolites Occurring in Mature
Green and Ripe Fruit as a Result of DET1 Downregulation under the
Control of the TFM7 Promoter.
Supplemental Figure 4. Changes in Metabolites Occurring in Mature
Green and Ripe Fruit as a Result of DET1 Downregulation under the
Control of the 2A11 Promoter.
Supplemental Figure 5. Correspondence Analysis for P119 and
Supplemental Figure 6. Overview of Relative Changes in Gene
Supplemental Figure 7. Heat Map Illustrating the Changes Occurring
in Gene-to-Gene Correlations Resulting from DET1 Downregulation at
the Mature Green Stage of Fruit Development.
Supplemental Figure 8. Heat Map Illustrating the Changes Occurring
in Gene-to-Gene Correlations Resulting from DET1 Downregulation at
the Stage of Ripe Fruit.
Supplemental Figure 9. Heat Map Illustrating the Changes Occurring
in Gene-to-Gene Correlations Resulting from DET1 Downregulation at
the Mature Green Stage of Fruit Development.
Supplemental Figure 10. Heat Map Illustrating the Changes Occur-
ring in Gene-to-Gene Correlations Resulting from DET1 Downregula-
tion at the Ripe Stage of Fruit Development.
Supplemental Figure 11. Heat Map Displaying Transcript and Me-
tabolite Correlations Associated with the Relative Changes Resulting
from DET1 Downregulation Compared with the Wild-Type (T56)
Control at the Mature Green Fruit Developmental Stage, and the Ripe
Fruit Stage, as well as an Analysis of Correlations between Mature
Green Transcripts and Ripe Fruit Metabolites.
Supplemental Table 1. The Stability of the Increased Carotenoid
Phenotype in DET1 Lines over the T3 to T5 Generations.
1212The Plant Cell
Supplemental Table 2. Metabolite Levels of DET1 Varieties Relative
to Their Respective Controls at the Mature Green and Ripe Fruit
Supplemental Table 3. Changes in Relative Metabolite Levels Found
in DET1 Varieties during Fruit Development and Ripening.
Supplemental Table 4. Overview of Global Changes in P119 and
TFM7 Transcript Levels during Ripening.
Supplemental Table 5. Hierarchy of Pathways Significantly Affected
in P119 and TFM7 DET1 Varieties.
Supplemental Table 6. Primer Sequences used for Quantitative
Real-Time RT-PCR and PCR.
Supplemental Data Set 1. Gene Identifiers for Transcripts Used in
This work was supported in part by EU-FP6 EU-SOL to P.M.B., P.D.F.,
and C.B., a PhD studentship from ANR-07-BLAN-0216 from the French
Agence Nationale de la Recherche to I.A., long-term fellowship
LT00299/2005 from the Human Frontier Science Program to F.B., ANR-
BBSRC SysBio BB/F005644/1 to P.M.B., P.D.F., and C.B., Agronano-
tech Fondo per gli Investimenti della Ricerca di Base and GenoPom
from Ministero dell’Istruzione, dell’Universita ` e della Ricerca to C.B., and
EU-FP7 METAPRO 244348 to P.D.F. The tomato microarray expression
profiling analysis was supported by the USDA–Agricultural Research
Service and Grant 0501778 and 0606595 from the National Science
Foundation Plant Genome Research Program to J.J.G. We thank Tom
Wells for bioinformatic assistance and help, John Halket for assistance
with metabolomic analysis, and Gita Patel and Marta Staff for assistance
in gathering cellular data. P.D.F. and E.M.A.E. also thank Lee Sweetlove
for valuable discussions and advice on aspects of primary metabolism.
We thank the TGRC for the supply of hp mutant seeds.
Received January 5, 2010; revised March 23, 2010; accepted April 6,
2010; published April 30, 2010.
Alba, R., et al. (2004). ESTs, cDNA microarrays and gene expression
profiling: tools for dissecting plant physiology and development. Plant
J. 39: 697–714.
Alba, R., Payton, P., Fei, Z., McQuinn, R., Debbie, P., Martin, G.,
Tanksley, S., and Giovannoni, J. (2005). Transcriptome and selected
fruit metabolite analysis reveal multiple points of ethylene regulatory
control during tomato fruit development. Plant Cell 17: 2954–2965.
Aldridge, C., Maple, J., and Moller, S.G. (2005). The molecular biology
of plastid division in higher plants. J. Exp. Bot. 56: 1061–1077.
Allan, A.C., Hellens, R.P., and Laing, W.A. (2008). MYB transcription
factors that colour our fruit. Trends Plant Sci. 13: 99–102.
Apel, W., and Bock, R. (2009). Enhancement of carotenoid biosynthesis
in transplastomic tomatoes by induced lycopene to Provitamin A
conversion. Plant Physiol. 151: 59–66.
Benvenuto, G., Formiggini, F., Laflamme, P., Malakhov, M., and
Bowler, C. (2002). The photomorphogenesis regulator DET1 binds
the amino-terminal tail of histone H2B in a nucleosome context. Curr.
Biol. 12: 1529–1534.
Butelli, E., Titta, L., Giorgio, M., Mock, H.-P., Matros, A., Peterek, S.,
Schijlen, E.G.W.M., Hall, R.D., Bovy, A.G., Luo, J., and Martin, C.A.
(2008). Enrichment of tomato fruit with health-promoting anthocyanins
by expression of select transcription factors. Nat. Biotechnol. 26:
Chory, J., and Peto, C.A. (1990). Mutations in the DET1 gene affect
cell-type-specific expression of light-regulated genes and chloroplast
development in Arabidopsis. Proc. Natl. Acad. Sci. USA 87: 8776–
Chory, J., Peto, C.A., Feinbaum, R., Pratt, L., and Ausubel, F. (1989).
Arabidopsis thaliana mutant that develops as a light-grown plant in the
absence of light. Cell 58: 991–999.
Cominelli, E., Gusmaroli, G., Allgera, D., Galbiati, M., Wade, H.K.,
Jenkins, G.I., and Tonelli, C. (2008). Expression analysis of antho-
cyanin regulatory genes in response to different light qualities in
Arabidopsis thaliana. J. Plant Physiol. 165: 886–894.
Cookson, P.J., Kiano, J.W., Shipton, C.A., Fraser, P.D., Roemer, S.,
Schuch, W., Bramley, P.M., and Pyke, K.A. (2003). Increases in cell
elongation, plastid compartment size and phytoene synthase activity
underlie the phenotype of the high pigment-1 mutant of tomato.
Planta 217: 896–903.
Cooper, D.A. (2004). Carotenoids in health and disease: Recent scien-
tific evaluations, research recommendations and the consumer. J.
Nutr. 134: 221–224.
Davuluri, G.R., Tuinen, A., Mustilli, A.C., Manfredonia, A., Newman,
R., Burgess, D., Brummell, D.A., King, S.R., Palys, J., Uhlig, J.,
Pennings, H.M.J., and Bowler, C. (2004). Manipulation of DET1
expression in tomato results in photomorphogenic phenotypes
caused by post-transcriptional gene silencing. Plant J. 40: 344–354.
Davuluri, G.R., et al. (2005). Fruit-specific RNAi-mediated suppression
of DET1 enhances tomato nutritional quality. Nat. Biotechnol. 23:
Diretto, G., Al-Babili, S., Tavazza, R., Papacchioli, V., Beyer, P., and
Giuliano, G. (2007). Metabolic engineering of potato carotenoid
content through tuber-specific overexpression of a bacterial mini-
pathway. PLoS One 2: e350.
Fitter, D.W., Martin, D.J., Copley, M., Scotland, R.W., and Langdale,
J.A. (2002). GLK gene pairs regulate chloroplast development in
diverse plant species. Plant J. 31: 713–727.
Foo, E., Ross, J.J., Davies, N.W., Ried, J.B., and Weller, J.L. (2006). A
role for ethylene in the phytochrome-mediated control of vegetative
development. Plant J. 46: 911–921.
Foyer, C.H., and Noctor, G. (2005). Redox homeostasis and antioxi-
dant signaling: A metabolic interface between stress perception and
physiological responses. Plant Cell 17: 1866–1877.
Fraser, P.D., Enfissi, E.M.A., and Bramley, P.M. (2009). Genetic
engineering of carotenoid formation in tomato fruit and the potential
application of systems and synthetic biology approaches. Arch.
Biochem. Biophys. 482: 196–204.
Fraser, P.D., Enfissi, E.M.A., Halket, J.H., Truesdale, M.R., Yu, D.M.,
Gerrish, C., and Bramley, P.M. (2007). Manipulation of phytoene
levels in tomato fruit: Effects on isoprenoids, plastids and intermediary
metabolism. Plant Cell 19: 3194–3211.
Fraser, P.D., Pinto, M.E., Holloway, D.E., and Bramley, P.M. (2000).
Application of high-performance liquid chromatography with photo-
diode array detection to the metabolite profiling of plant isoprenoids.
Plant J. 24: 551–558.
Fraser, P.D., Roemer, S., Shipton, C.A., Mills, P.B., Kiano, J.W.,
Misawa, N., Drake, R.G., Schuch, W., and Bramley, P.M. (2002).
Evaluation of transgenic tomato plants expressing an additional
phytoene synthase in a fruit-specific manner. Proc. Natl. Acad. Sci.
USA 99: 1092–1097.
Galpaz, N., Wang, Q., Menda, N., Zamir, D., and Hirschberg, J.
Analysis of DET1 Downregulated Tomato Fruit1213
(2008). Abscisic acid deficiency in the tomato mutant high-pigment3
leading to increased plastid number and higher fruit lycopene content.
Plant J. 53: 717–730.
Ghassemian, M., Lutes, J., Tepperman, J.M., Chang, H.-S., Zhu, T.,
Wang, X., Quail, P.H., and Lange, M.B. (2006). Integrative analysis of
transcript and metabolite profiling data sets to evaluate the regulation
of biochemical pathways during photomorphogenesis. Arch. Bio-
chem. Biophys. 448: 45–59.
Giliberto, L., Perrotta, G., Weller, J.L., Fraser, P.D., Bramley, P.M.,
Fiore, A., Tavazza, M., and Giuliano, G. (2005). Manipulation of the
blue light photoreceptor cryptochrome2 in tomato affects vegetative
development, flowering time and fruit antioxidant content. Plant
Physiol. 137: 199–208.
Gillaspy, G., Ben-David, H., and Gruissem, W. (1993). Fruits: A
developmental perspective. Plant Cell 5: 1439–1451.
Giovannoni, J. (2004). Genetic regulation of fruit development and
ripening. Plant Cell 16: S170–S180.
Giovannucci, E. (2002). Lycopene and prostate cancer risk. Method-
ological considerations in the epidemiologic literature. Pure Appl.
Chem. 74: 1427–1434.
Grant, G.R., Liu, J., and Stoeckert, C.J., Jr. (2005). A practical false
discovery rate approach to identifying patterns of differential expres-
sion in microarray data. Bioinformatics 11: 2684–2690.
Griffiths, A., Barry, C.S., Alpuche-Solils, A., and Grierson, D. (1999).
Ethylene and developmental signals regulate expression of lipoxyge-
nase genes during tomato fruit ripening. J. Exp. Bot. 50: 793–798.
He, K., Stolc, V., Lee, H., Figueroa, P., Gao, Y., Tongprasit, W., Zhao,
H., Lee, I., and Deng, X.W. (2007). Analysis of transcription factor
HY5 genomic binding sites revealed its hierarchical role in light
regulation of development. Plant Cell 19: 731–749.
Huq, E., Tepperman, J.M., and Quail, P. (2000). GIGANTEA is a
nuclear protein involved in phytochrome signalling in Arabidopsis.
Proc. Natl. Acad. Sci. USA 97: 9789–9794.
Joung, J.-G., Corbett, A.M., Moore-Fellman, S., Tieman, D.M., Klee,
H.J., Giovannoni, J.J., and Fei, Z. (2009). Plant MetGenMAP: An
integrative analysis system for plant systems biology. Plant Physiol.
Jung, S. (2004). Effect of chlorophyll reduction in Arabidopsis thaliana
by methyl jasmonate or norflurazon on antioxidant systems. Plant
Physiol. Biochem. 42: 225–231.
Key, T.J., Allen, N.E., Spencer, E.A., and Travis, R.C. (2002). The
effect of diet on risk of cancer. Lancet 360: 861–868.
Kolotilin, I., Koltai, H., Tadmor, Y., Bar-Or, C., Reuveni, M., Meir, A.,
Nahon, S., Shlomo, H., Chen, L., and Levin, I. (2007). Transcriptional
profiling of high pigment-2dgtomato mutant links early plastid bio-
genesis with its overproduction of phytonutrients. Plant Physiol. 145:
Liu, Y., Roof, S., Ye, Z., Barry, C., van Tuinen, A., Verbalov, J.,
Bowler, C., and Giovannoni, J. (2004). Manipulation of light signal
transduction as a means of modifying fruit nutritional quality in tomato.
Proc. Natl. Acad. Sci. USA 101: 9897–9902.
Long, M., Millar, D.J., Kimura, Y., Donovan, G., Rees, J., Fraser,
P.D., Bramley, P.M., and Bolwell, G.P. (2006). Metabolite profiling of
carotenoid and phenolic pathways in mutant and transgenic lines of
tomato: Identification of a high antioxidant fruit line. Phytochemistry
Lopez-Juez, E., Dillon, E., Magyar, Z., Khan, S., Hazeldine, S., de
Pager, S.M., Murria, J.A.H., Beemster, G.T.S., Bogre, L., and
Shanahan, H. (2008). Distinct light-initiated gene expression and
cell cycle programs in the shoot apex and cotyledons of Arabidopsis.
Plant Cell 20: 947–968.
Luo, J., Butelli, E., Hill, L., Parr, A., Niggeweg, R., Bailey, P.,
Weisshaar, B., and Martin, C.A. (2008). AtMYB12 regulates caffeoyl
quinic acid and flavonol synthesis in tomato: Expression in fruit results
in very high levels of both types of polyphenol. Plant J. 56: 316–326.
Ma, L., Li, J., Qu, L., Hager, J., Chen, Z., Zhao, H., and Deng, X.W.
(2001). Light control of Arabidopsis development entails coordinated
regulation of genome expression and cellular pathways. Plant Cell 13:
Ma, L., Zhao, H., and Deng, X.W. (2003). Analysis of the mutational
effects of the COP/DET/FUS loci on genome expression profiles
reveals their overlapping yet not identical roles in regulating Arabi-
dopsis seedling development. Development 130: 969–981.
Maass, D., Arango, J., Wu ¨st, F., Beyer, P., and Welsch, R. (2009).
Carotenoid crystal formation in Arabidopsis and carrot roots caused
by increased phytoene synthase protein levels. PLoS One 4: e6373.
Martin, C., Carpenter, R., Sommer, H., Saedler, H., and Coen, E.S.
(1985). Molecular analysis of instability in flower pigmentation of
Antirrhinum majus, following isolation of the pallida locus by transpo-
son tagging. EMBO J. 4: 1625–1630.
Muir, S.R., Collins, G.J., Robinson, S., Hughes, S., Bovy, A., De Vos,
C.H.S., van Tunen, A.J., and Verhoeyen, M.E. (2001). Overexpres-
sion of petunia chalcone isomerase in tomato results in fruit contain-
ing increased levels of flavonols. Nat. Biotechnol. 19: 470–474.
Naqvi, S., Zhu, C., Farre, G., Ramessar, K., Bassie, L., Breitenbach,
J., Perez-Conesa, D., Ros, G., Sandmann, G., Capell, T., and
Christou, P. (2009). Transgenic multivitamin corn through biofortifi-
cation of endosperm with three vitamins representing three distinct
metabolic pathways. Proc. Natl. Acad. Sci. USA 106: 7762–7767.
Osterlund, M.T., Hardtke, C.S., Wei, N., and Deng, X.W. (2000).
Targeted destabilization of HY5 during light-regulated development of
Arabidopsis. Nature 405: 462–466.
Re, R., Pellergrini, N., Protrggente, A., Pannala, A., Yang, M., and
Rice-Evans, C.A. (1999). Antioxidant activity applying an improved
ABTS radical cation decolorization assay. Free Radic. Biol. Med. 26:
Sasaki-Sekimoto, Y., et al. (2005). Coordinated activation of meta-
bolic pathways for antioxidants and defence compounds by jasm-
onates and their roles in stress tolerance in Arabidopsis. Plant J. 44:
Scha ¨fer, E., and Bowler, C. (2002). Phytochrome-mediated photo-
perception and signal transduction in higher plants. EMBO Rep. 3:
Schroeder, D.F., Gahrtz, M., Maxwell, B.B., Cook, R.K., Kan, J.M.,
Alonso, J.M., Ecker, J.R., and Chory, J. (2002). De-etiolated 1 and
damaged DNA binding protein 1 interact to regulate Arabidopsis
photomorphogenesis. Curr. Biol. 12: 1462–1472.
Terzaghi, W., and Cashmore, A. (1995). Light-regulated transcription.
Annu. Rev. Plant Physiol. Plant Mol. Biol. 46: 445–474.
Triques, K., Sturbios, B., Gallais, S., Dalmais, M., Chauvin, S., Clepet,
C., Aubourg, S., Rameau, C., Caboche, M., and Bendahmane, A.
(2007). Characterisation of Arabidopsis thaliana mismatch specific
endonucleases: application to mutation discovery by TILLING in pea.
Plant J. 51: 1116–1125.
Vanderauwera, S., Zimmermann, P., Rombauts, S., Vandenabeele,
S., Langebartels, C., Gruissem, W., Inze, D., and Van Breusegem,
F. (2005). Genome-wide analysis of hydrogen peroxide-regulated
gene expression in Arabidopsis reveals a high light-induced tran-
scriptional cluster involved in anthocyanin biosynthesis. Plant Physiol.
Verhoeyen, M.E., Bovy, A., Collins, G., Robinson, S., De Vos, C.H.R.,
and Colliver, S. (2002). Increasing antioxidant levels in tomatoes
through modification of the flavonoid biosynthetic pathway. J. Exp.
Bot. 53: 2099–2106.
Wang, S.H., Liu, J.K., Feng, Y.Y., Niu, X.L., Giovannoni, J., and Liu,
Y.S. (2008). Altered plastid levels and potential for improved fruit
1214 The Plant Cell
content by down-regulation of the tomato DDB1-interacting protein
CUL4. Plant J. 55: 89–103.
Wasternack, C. (2006). Jasmonates; An update on biosynthesis, signal
transduction and action in plant stress response, growth and devel-
opment. Ann. Bot. (Lond.) 100: 681–697.
Wei, H., Persson, S., Mehta, T., Srinivasasainagendra, V., Chen, L.,
Page, G.P., Somerville, C.R., and Loraine, A. (2006). Transcriptional
coordination of the metabolic network in Arabidopsis. Plant Physiol.
Wellburn, A.R. (1994). The spectral determination of chlorophyll A and
chlorophyll B as well as total carotenoids, using various solvents and
with spectrophotomers of different resolution. J. Plant Physiol. 144:
Yang, Y.H., Dudoit, S., Luu, P., Lin, D.M., Peng, V., Ngai, J., and
Speed, T.P. (2002). Normalisation for cDNA microarray data: A robust
composite method addressing single and multiple slide systematic
variation. Nucleic Acids Res. 30: e15.
Yasumura, Y., Moylan, E.C., and Langdale, J.A. (2005). A conserved
transcription factor mediates nuclear control of organelle biogenesis
in anciently diverged land plants. Plant Cell 17: 1894–1907.
Ye, X.D., Al-Babili, S., Kloti, S., Zhang, J., Lucca, P., Beyer, P., and
Potrykus, I. (2000). Engineering the provitamin A (beta-carotene)
biosynthetic pathway into (carotenoid-free) rice endosperm. Science
Youssef, A., Laizet, Y., Block, M.A., Marechal, E., Alcaraz, J.-P.,
Larson, T.R., Pontier, D., Gaffe, J., and Kuntz, M. (2009). Plant lipid-
associated fibrillin proteins condition jasmonate production under
photosynthetic stress. Plant J. 61: 436–445.
Zamir, D. (2001). Improving plant breeding with exotic genetic libraries.
Nat. Rev. Genet. 2: 983–989.
Analysis of DET1 Downregulated Tomato Fruit1215
This information is current as of June 6, 2010 Download full-text
2010;22;1190-1215; originally published online Apr 30, 2010;
McQuinn, James J. Giovannoni, Enrique Lopez-Juez, Chris Bowler, Peter M. Bramley and Paul D.
Eugenia M.A. Enfissi, Fredy Barneche, Ikhlak Ahmed, Christiane Lichtlé, Christopher Gerrish, Ryan P.
Downregulated Tomato Fruit
DE-ETIOLATED1 Integrative Transcript and Metabolite Analysis of Nutritionally Enhanced
This article cites 64 articles, 26 of which you can access for free at:
THE PLANT CELL
Sign up for eTOCs for
Sign up for CiteTrack Alerts for
is available at:
The Plant Cell
Subscription information for
ADVANCING THE SCIENCE OF PLANT BIOLOGY
© American Society of Plant Biologists