Professional Documents
Culture Documents
1771–1781, 2007
doi:10.1093/jxb/erm036 Advance Access publication 10 April, 2007
RESEARCH PAPER
Bo Zhou1, Yuhua Li1,*,†, Zhiru Xu1, Haifang Yan1,2, Shinichiro Homma2 and Saneyuki Kawabata2,†
1
College of Life Sciences, Northeast Forestry University, Harbin 150040, China
2
Graduate School of Agricultural and Life Sciences, University of Tokyo, Yayoi, Bunkyo-ku, Tokyo, 113-8657 Japan
Abstract Introduction
Ultraviolet A (UV-A)-mediated regulation of anthocyanin UV radiation from the sun induces various responses in
biosynthesis was investigated in swollen hypocotyls of higher plants. While the greatest portion of UV-B (280–
the red turnip ‘Tsuda’. The shaded swollen hypocotyls 320 nm) is absorbed by the ozone layer, UV-A (320–
which contained negligible anthocyanin were exposed 400 nm) penetrates the atmosphere to reach the earth
to artificial light sources including low fluence UV-B, surface. Typically, UV-A radiation in the temperate zone
UV-A, blue, red, far-red, red plus UV-A, far-red plus UV-A, is ;50 W m2, while UV-B is 2 W m2 during mid-
and blue plus red. Among these lights, only UV-A summer. High fluence UV-B causes serious damage to
induced anthocyanin biosynthesis and co-irradiation of DNA, membranes, and proteins. DNA is especially
red or far-red with UV-A did not affect the extent of UV-A- sensitive to high fluence UV-B, resulting in the formation
induced anthocyanin accumulation. The expression of of pyrimidine dimers (Taylor et al., 1997; Frohnmeyer
phenylalanine ammonia lyase (PAL; EC 4.3.1.5), chal- and Staiger, 2003). However, low fluence UV-B stim-
cone synthase (CHS; EC 2.3.1.74), flavanone 3-hydroxy- ulates distinct responses, such as the accumulation of
lase (F3H; EC 1.14.11.9), dihydroflavonol 4-reductase UV-absorbing pigments and expression of stress
(DFR; EC 1.1.1.219), and anthocyanidin synthase (ANS; response-related genes (Hahlbrock and Scheel, 1989; A-H
EC 1.14.11.19) genes was increased with time during
Mackerness et al., 2001; Brosché and Strid, 2003;
a 24 h exposure to UV-A. In contrast, irradiation with red,
Frohnmeyer and Staiger, 2003). These responses are
blue, UV-B, and a combination of blue with red failed to
considered to play a role as a protective mechanism
induce CHS expression. Microarray analysis showed
against potential damage by UV irradiation. Several
that only a few genes, including CHS and F3H, were
studies have suggested that these responses are mediated
induced significantly by UV-A, while a separate set of
many genes was induced by low fluence UV-B. The UV-
by a UV-B receptor (Kim et al. 1998; Boccalandro et al.,
A-specific induction of anthocyanin biosynthesis and 2001; Suesslin and Frohnmeyer, 2003).
the unique gene expression profile upon UV-A irradia- Since most of the UV-A from the sun penetrates the
tion as compared with blue and UV-B demonstrated that atmosphere, UV-A radiation at the earth’s surface is much
the observed induction of anthocyanin biosynthesis in stronger than other wavelengths of UV rays. Thus, UV-A
red turnips was mediated by a distinct UV-A-specific may have significant influences on plants. Unlike the
photoreceptor, but not by phytochromes, UV-A/blue responses to low fluence UV-B, however, such responses
photoreceptors, or UV-B photoreceptors. have not been well studied. Since UV-A is in a range of
long-wave UV and thus has lower energy, it is expected
Key words: Chalcone synthase, microarray, photoreceptor, to cause mild or negligible damage to plants directly,
SSH, subtractive library. but may induce protective mechanisms against more
ª The Author [2007]. Published by Oxford University Press [on behalf of the Society for Experimental Biology]. All rights reserved.
For Permissions, please e-mail: journals.permissions@oxfordjournals.org
1772 Zhou et al.
harmful short-wave UV as suggested for low fluence UV- part of the swollen tissue that appears above the ground originates
B. Low fluence UV-B and high fluence UV-A sometimes from the hypocotyls, which can be distinguished from the root by
the absence of the lateral roots. Approximately 1 month after
caused similar responses, such as necrosis observed in sowing, the outer tissue of the hypocotyls split and the inner tissue
UV-sensitive Arabidopsis mutants defective in succinic- initiates swelling. The anthocyanins accumulate in the epidermis of
semialdehyde dehydrogenase (Bouché et al., 2003). The this swelling tissue. The anthocyanin accumulation occurs only in
most common protective mechanism stimulated by low the epidermis of the swollen tissue that appears above the ground in
fluence UV-B radiation is the accumulation of anthocya- the ‘Tsuda’ turnip.
When the hypocotyls started to swell, the upper part of
nins as observed in maize (Singh et al., 1999), rice the swollen tissue appearing above the ground was covered
(Reddy et al., 1994), apple fruits (Arakawa et al., 1985, with aluminium foil and overlaid with soil to prevent exposure to
1986; Arakawa, 1988; Dong et al., 1995), roses (Maekawa sunlight. Approximately 3 months after sowing, when the diameter
et al., 1980; Nakamura et al., 1980; Mor and Zieslin, of the swollen tissue reached 2–3 cm, the plants were subjected to
1990), kangaroo paw (Ben-Tal and King, 1997), apple light treatments as described below.
flowers (Dong et al., 1998), and Arabidopsis (Kubasek
Effects of light quality on the anthocyanin biosynthesis in
et al., 1992; Christie and Jenkins, 1996). Similarly, UV-A
0.4
Bootstrap clustering
To examine the differences in gene expression profiles between 0.3
the light treatments, probes which showed significant F statistics
at adjusted P <0.01 were subjected to multiscale bootstrapping
hierarchical clustering (Shimodaira, 2002, 2004) among arrays and 0.2
among probes. A package ‘pvclust’ (Suzuki and Shimodaira, 2006)
was used to perform hierarchical clustering and to calculate multiscale 0.1
BPs (bootstrap probability), and ‘scaleboot’ to calculate AU (approx- n.s. n.s. n.s. n.s. n.s.
n.s.
imately unbiased P-value). One thousand bootstrap iterations and
multiscales of 0.55–9.00 for clustering of probes and 10 000 0
Dark FR Red Blue UV-A low high UV-A UV-A Blue
iterations and multiscales of 0.11–9.00 for clustering of arrays were
UV-B UV-B +FR +Red +Red
used to calculate significance scores for the clusters.
Fig. 1. Effect of irradiation with different wavelengths of lights on
anthocyanin accumulation in the epidermis of the swollen hypocotyls of
3-month-old ‘Tsuda’ turnip. The swollen hypocotyls were exposed to
Results UV-B at 6 W m2, UV-A at 7 W m2, blue light at 12 W m2, red
light at 13 W m2, far-red light at 14 W m2, or a combination of red
UV-A-specific induction of anthocyanin biosynthesis or far-red (FR) with UV-A for 24 h, or collected without light exposure
(dark). Anthocyanin was extracted from the light-exposed part of the
in the peel of swollen hypocotyls epidermis of the swollen hypocotyls, and then the concentration was
To characterize the light-dependent regulation of anthocy- determined. Vertical bars indicate 6SE (n¼8). *** and n.s. indicate that
the difference in means between a treatment and the dark control is
anin biosynthesis in ‘Tsuda’, the effects of several wave- significant at P <0.001 and non-significant at P <0.05, respectively, by
lengths of light on anthocyanin biosynthesis in the Dunnet’s post hoc test using log-transformed values.
UV-A-specific induction of anthocyanins in Brassica rapa 1775
1.2
1
Anthocyanin (OD• cm-2)
0.8
0.6
0.4
0.2
n.s.
0
UVA - - + +
VIS - + - +
0.15 0.2
0.1
0.1
0.05 n.s.
n.s. n.s. n.s. n.s. n.s.
0 0
Dark Sun- FR Red Blue UV-A UV-B
0L 1L+23D 3L+21D 6L+18D 24L light
Fig. 5. Effect of the period and intensity of UV-A exposure on antho-
0.5
0.8
remains the possibility that failure of anthocyanin pro-
duction under blue or UV-B light could be due to
insufficient light intensity, or contamination of other
wavelengths of lights. Therefore, microarray analysis
0.6
(GSE6876) was conducted to see whether UV-A induces
the expression of a distinct gene set. A hierarchical
clustering of arrays using M values of the gene probes
Height
which showed significant changes by F-statistics at an
0.4
adjusted P-value of 0.01 was conducted (Fig. 8). The 1.00
dendrogram showed that all the microarrays of UV-B 0.90
0.97
treatment were grouped into one cluster, while other
0.2
microarrays were grouped into a distinct cluster. This 0.96
Blue1
Blue3
indicated that UV-B induced a distinct set of genes which
Blue2
1.00
were not induced by blue and UV-A treatments, and that 1.00 0.72
0.0
UVA1
UVA3
treatments. A hierarchical clustering was also conducted
UVB2
UVB1
UVB3
for microarray probes (Fig. 9). Several clusters were
identified by multiscale bootstrap clustering. Five groups
were selected by a criterion of AU >0.99, and the gene Fig. 8. Dendrogram of bootstrapping hierarchical clustering of micro-
arrays with M value correlation distance between each pair
expression profile for each group was shown in Fig. 9. of microarrays by average linkage. The epidermis of the swollen
Many genes showed up-regulation (Group #5) or down- hypocotyls of 3-month-old ‘Tsuda’ turnip plants was exposed to blue
regulation (Group #1) due to UV-B exposure. A smaller light at 12 W m2 (Blue), UV-A at 7 W m2 (UVA), or UV-B at 6 W
m2 (UVB), or kept in darkness (Dark). The M value was calculated as
set of genes showed up-regulation in response to UV-A the log2 ratio between blue and dark (Blue1, Blue2, Blue3), UV-A and
irradiation (Group #3). Genes for anthocyanin biosynthe- dark (UVA1, UVA2, UVA3), and UV-B and dark (UVB1, UVB2,
sis, including CHS and F3H, were grouped into this UVB3). The numbers followed by the treatment names are the identifier
of biological replicates. The values at the nodes indicate the approx-
group (Table 2), while PAL was grouped into Group imately unbiased P-value (AU) for the clusters.
#5 (see Supplementary Table S1 at JXB online). Other
genes of flavonoid biosynthesis spotted on the microarray
did not show significant changes, and were therefore
excluded. UV-A treatment, indicating that UV-A at 7 W m2 did
The list of gene probes which showed significant not induce general protective mechanisms as observed for
changes due to light exposure is shown in Supplementary UV-B in ‘Tsuda’ swollen hypocotyls. The damage due to
Table S1 at JXB online. Group #5 genes included genes UV-A exposure is unlikely to be as severe as that caused
related to ethylene biosynthesis, reactive oxygen species by low fluence UV-B, even if the intensity of UV-A was
reactions such as glutathione S-transferase and superoxide equivalent to that of UV-B. The results did not support the
dismutase, and carbohydrate metabolism such as hypothesis that UV-A would induce similar protective
chitinase and glyceraldehyde-3-phosphate dehydrogenase. responses against potential damage upon UV exposure as
Down-regulated genes included genes related to tissue does the low fluence UV-B. UV-A intensity of sunlight
or cell growth such as extensin, xyloglucan endotransgly- reached 50 W m2 during mid-summer in the temperate
cosylase/hydrolase, and aquaporin, and to glucosynolate zone. Such a high intensity of UV-A may induce
metabolism, such as cytochrome P450 (CYP83A1), protective mechanisms as observed for low fluence UV-B.
methylthioalkylmalate synthase, myrosinase-binding pro- Instead of inducing expression changes in a robust set
teins, and myrosinase-associated proteins. of genes related to the stress response, UV-A induced
anthocyanin biosynthesis. The anthocyanin biosynthetic
pathway involves several enzymes including PAL, CHS,
Discussion CHI, F3H, DFR, ANS, and glycosyl- or acyl-transferases.
Genes encoding these enzymes have been isolated and
UV-A did not induce a broad range of stress their developmental and tissue-specific expression patterns
responses as observed for low fluence UV-B have been investigated (Koes et al., 1994; Holton and
Low fluence UV-B was found to stimulate the transcript Cornish, 1995; Weisshaar and Jenkins, 1998). CHS plays
levels of a robust set of genes involved in stress responses a key role in the regulation of anthocyanin biosynthesis in
(Casati and Walbot, 2003, 2004; Ulm et al., 2004). In the many plant species (Jenkins et al., 2001; Wade et al.,
present experiment, UV-B irradiation altered the expres- 2001). The microarray analysis and subsequent northern
sion of many genes that are involved in stress responses. blot hybridization demonstrated the stimulation of PAL,
In contrast, such responses were not observed for the CHS, F3H, DFR, and ANS expression by UV-A.
1778 Zhou et al.
Group #1 Group #2 Group #3
4
2
2
0
0
-2
-2
-2
-4
-4
-4
Log2(Cy3/Cy5)
Blue UVA UVB Blue UVA UVB Blue UVA UVB
4 Group #4 Group #5
4
2
2
0
-2
-4
-4
Blue UVA UVB Blue UVA UVB
Fig. 9. M values of probes in each of the significant clusters at P >0.99 based on bootstrapping hierarchical clustering with M value correlation
distance between each pair of probes by average linkage. The epidermis of the swollen hypocotyls of 3-month-old ‘Tsuda’ turnip plants was exposed
to blue light at 12 W m2 (Blue), UV-A at 7 W m2, or UV-B at 6 W m2, or kept in darkness for 24 h. Microarray hybridization was performed
between blue/dark (Blue), UV-A/dark (UVA), or UV-B/dark (UVB). The y-axis indicates M values calculated as the log2 ratio of light treatment
signals to the dark control signals. Values are means of three microarray replicates. Bootstrap clustering with 1000 iterations and AU (approximately
unbiased P-value) threshold of 0.99 identified five major clusters (Group #1–5), and several subclusters (not shown).
Table 2. The list of probes in Group #3 in Fig. 9 Although UV-A at 7 W m2 did not induce the same
The epidermis of the swollen hypocotyls of 3-month-old ‘Tsuda’ turnip protective mechanisms as UV-B, this intensity was
plants was exposed to blue light at 12 W m2 (Blue), UV-A at 7 W sufficient to induce anthocyanin production, which may
m2, or UV-B at 6 W m2, or kept in darkness for 24 h. Microarray
hybridization was performed between blue/dark (Blue), UV-A/dark contribute to the tolerance to UV irradiation. Anthocya-
(UVA), or UV-B/dark (UVB). M values are the log2 ratio of light nins in the outer layer of the tissue not only can absorb
treatment signals to the dark control signals. Means of three microarray UV to protect the inner layer of the tissue, but also serve
replicates are indicated. The top hit results of the blast search using the
‘RefSeq’ database with an E-value threshold of E-15 are indicated as a scavenger of reactive oxygen species (Gould et al.,
(Annotation/top blast hit). 2002). The production of anthocyanins by UV-A exposure
may have contributed to the protection of the plant tissue
Probe M value Annotation/top
ID blast hit from the potential damage by UV absorption, and this
Blue UVA UVB may be responsible for the observation that UV-A did not
have a strong impact on gene expression profiles as
BR3105D05 0.317 2.042 0.946 Myrosinase-associated
protein, putative observed for low fluence UV-B response.
(AT4G01130)
BR3105D04 0.759 3.100 1.545 CHS (chalcone
synthase) Involvement of UV-A-specific photoreceptor
BR3105H01 0.459 2.072 1.292 CHS (chalcone At present, three types of photoreceptors have been
synthase)
BR8871 1.647 3.857 2.069 CHS (chalcone characterized: (i) phytochromes which perceive red or far-
synthase) red light (Quail, 1994); (ii) UV-A/blue photoreceptors
BR8691 0.886 2.016 0.634 CHS (chalcone including cryptochrome (Ahmad and Cashmore, 1993;
synthase)
BR9398 0.391 2.719 1.156 CHS (chalcone Lin, 2000) and phototropin (Briggs and Christie, 2002);
synthase) and (iii) the UV-B photoreceptor. Although the primary
BR10900 1.103 3.190 1.899 Invertase/pectin UV-A-responsive photoreceptor known is the UV-A/blue
methylesterase inhibitor
family protein photoreceptor, all three types of these photoreceptors
(AT5G62350) could potentially sense the UV-A signal. The measure-
BR8668 0.245 1.791 1.299 F3H (naringenin ment of absorption spectra of phytochrome suggested that
3-dioxygenase/flavanone
3-hydroxylase) it could also sense the blue/UV-A signal (Mancinelli,
BR3101E06 0.107 1.008 0.765 RNA-binding/nucleic 1986). In fact, Pratt and Briggs (1966) reported in vivo
acid-binding conversion of phytochrome by blue light. Devlin and Kay
(AT3G10845)
BR8879 0.849 1.692 1.496 Unknown protein (2000) proposed that phytochrome A acts as a photorecep-
(AT3G24100) tor in blue light input in circadian photoperception. For
BR10780 0.314 1.341 0.345 Unknown protein white cabbage seedlings, phytochrome alone appeared to
(AT3G54260)
be involved in anthocyanin production at a low fluence
UV-A-specific induction of anthocyanins in Brassica rapa 1779
rate of UV (Lercari et al., 1989). In addition, as the UV-B profiles of gene expression under irradiation by UV-A,
photoreceptor molecule has not been isolated, it has not low fluence UV-B, and blue light indicated the existence
been characterized yet. of separate regulatory mechanisms stimulated by these
The studies using artificial light sources demonstrated light sources.
the absence of red, far-red, blue, and UV-B induction of Additional experiments using seedlings provided further
CHS and anthocyanin production. It was also indicated evidence for a UV-A-specific photoreceptor. The light
that anthocyanin production induced by the UV-A lamp responses of turnip seedlings have been extensively
was not induced by possible contaminant UV-B. Whereas studied before. Anthocyanin biosynthesis in the hypo-
one-third the intensity of UV-A light failed to induce cotyls of turnip seedlings was shown to be induced by
anthocyanin production, UV-A filtered with a glass plate, blue, red, and far-red (Siegelman and Hendricks, 1957), and
which completely absorbs UV-B (<320 nm), could induce controlled by the red/far-red reversible reaction of phyto-
anthocyanin production. An LED light of 360 nm UV-A, chrome with co-action of blue light (Grill, 1965). The
which did not contain UV-B, induced anthocyanin pro- present experiment using turnip seedlings showed that the
duction (data not shown). The effect of longer wavelength upper part of the seedling hypocotyls displayed far-red-