The Plant Cell, Vol. 4, 1075-1087, September 1992 0 1992 American Society of Plant Physiologists
A Nove1 Circadian Phenotype Based on Firefly Luciferase
Expression in Transgenic Plants
Andrew J. Millar, Sharla R. Short, Nam-Hai Chua, and Steve A. Kay’I2
Laboratory of Plant Molecular Biology, The Rockefeller University, 1230 York Avenue, New York, New York 10021-6399
A 320-bp fragment of the Arabidopsis cab2 promoter is sufficient to mediate transcriptional regulation by both phyto-
chrome and the circadian clock. We fused this promoter fragment to the firefly luciferase (Luc) gene to create a real-time
reporter for regulated gene expression in intact plants. Cab2::Luc transcript accumulated in the expected patterns and
luciferase activity was closely correlated to cab2::Luc mRNA abundance in both etiolated and green seedlings. The con-
centration of the bulkof luciferase protein did not reflect these patterns but maintained a relatively constant level, implying
that a post-translational mechanism(s) leads to the high-amplitude regulation of luciferase activity. We used a low-light
video imaging system to establish that luciferase bioluminescence in vivo accurately reports the temporal and spatial
regulation of cab2 transcription in single seedlings. The unique qualities of the firefly luciferase system allowed us to
monitor regulated gene expression in real time in individual multicellular organisms. This noninvasive marker for tem-
poral regulation at the molecular level constitutes a circadian phenotype, which may be used to isolate mutants in the
Light is important to plant cells as a source of energy, as a
developmental signal, and as a timing cue. For example, in
dark-grown angiosperm seedlings, the light signal initiates a
developmental transition known as photomorphogenesis,
which permits the etiolated seedling to switch from heterotrophy
to autotrophy (Kendrick and Kronenberg, 1986). The best-
characterized plant photoreceptor, responsible for many
aspects of this transformation, is phytochrome (Quail, 1991).
Activation of phytochrome alone, by illumination of plants with
red light, initiates many of the photomorphogenic processes
from morphological changes (such as the inhibition of
hypocotyl elongation) to the molecular events underlying chlo-
roplast biogenesis (such as the expression of the cab genes,
which encode the chlorophyll alb binding proteins of the light-
The energy source for plants in nature, sunlight, varies in
intensity and in quality over the diurna1 cycle. If plants are ex-
perimentally removed from lighvdark (LD) cycles and placed
under constant environmental conditions, processes at many
levels continue to function rhythmically, with periodicities close
to 24 hr: this is the hallmark of regulation by the circadian clock
(Sweeney, 1987). This endogenous oscillator controls plant
functions ranging from stem elongation (Lecharny et al., 1990)
and the movements of petals (Engelmann and Johnsson, 1978)
Current address: NSF Center for Biological Timing, Department of
Biology, Gilmer Hall, University of Virginia, Charlottesville, VA 22901.
T o whom correspondence should be addressed.
to the rhythmic expression of several nuclear genes (Lumsden,
1991; reviewed in Kay and Millar, 1992). Some of the activities
controlled by the circadian clock are involved in photosynthe-
sis, and their regulation may be interpreted as a preparation
for photosynthetic activity during the day (Nagy et al., 1988a;
Adamska et al., 1991; Busheva et al., 1991), although this is
probably not the only adaptive significance of circadian con-
trol in higher plants. Indeed, circadian regulation in plants
shares many properties with the circadian rhythms of other
organisms, including temperature compensation of the period
in constant (“free-running”) conditions and phase shifting by
environmental stimuli (Edmunds, 1988). The principal reset-
ting stimulus for most circadian systems is light, which ensures
synchronization of the endogenous oscillator to the natural
dayhight cycle (entrainment). Given the similarities in circadian
regulation among diverse species, analysis of the molecular
organization of such a system should have broad significance.
Severa1 plant species have been of historical importance
in the study of circadian rhythms (Sweeney, 1987), although
more significant progress has occurred in Drosophila and Neu-
rospora, which are systems amenable to both classical and
molecular genetic studies of circadian rhythmicity. Genetic
analysis has defined several loci that are involved in regulat-
ing the periodicity of circadian rhythms, notably per in
Drosophila and ff9 in Neurospora, both of which have been
cloned (reviewed in Rosbash and Hall, 1989; Dunlap, 1990;
Hall, 1990). In addition, circadian control of transcription has
been described for various genes (Loros and Dunlap, 1991):
in Drosophila, this class includes the per gene (Hardin et al.,
The Plant Cell
1990). Recent experiments using fusion constructs in trans-
genic flies suggest that the control of per expression at the
transcriptional and/or post-transcriptional levels may be in-
volved in the mechanism of the circadian oscillator (Zwiebel
et al., 1991). Thus, analysis of the mechanisms of circadian
transcriptional regulation may throw new light on the mecha-
nism of the oscillator.
In higher plants, the cab genes have long provided a model
system to study phytochrome regulation (Nagy et al., 1988b),
but under various conditions the direct light regulation medi-
ated by this photoreceptor is modulated by the circadian clock
(Nagy et al., 1988a; Tavladoraki et al., 1989). Studies in etio-
lated pea (Gallagher et al., 1985) and tobacco (Paulsen and
Bogorad, 1988) that were transferred to constant white light
showed an initial, brief peak (at ca. 4 hr) and a second, broader
peak (ca. 20 hr) of cab gene transcription and mRNA abun-
dance, respectively. A very similar pattern was observed after
brief red light treatments of etiolated tobacco (Wehmeyer et
al., 1990) and bean (Tavladoraki et al., 1989). In light-grown
plants under LD conditions (reviewed in Kay and Millar, 1992),
the abundance of cab transcripts is generally highest in the
late morning and falls to a trough in the night. The rise in tran-
script abundance begins several hours before dawn, indicating
the light independence of this expression pattern. Rhythmic-
ity persists when green plants are transferred to constant light
(LL) or constant darkness (DD). Rapid reduction of peak lev-
els is observed in DD, although the rate of damping varies
among species. The majority of the analyses mentioned above
employed probes detecting the sum of all mRNAs from all
genes of the PSll cab type I family (for a review of the cab
gene family, see Green et al., 1991). This approach cannot dis-
cern any differences in regulation arnong family members,
which may be considerable (Millar and Kay, 1991). The iden-
tification of cis-acting elements from cab promoters, in wheat
(Nagy et al., 1988a; Fejes et al., 1990) and Arabidopsis (Millar
and Kay, 1991), now permits molecular analyses of the mech-
anisms mediating circadian responsive transcription of specific
One of the principal limitations of chronobiology at the mo-
lecular level has been the use of mRNA markers, which are
relatively time-consuming to assay and involve destruction of
the experimental material. Molecular genetic approaches
would be greatly expedited and classical genetics made pos-
sible if a facile, noninvasive assay for promoter activity were
employed. The luciferases are the most versatile class of nonin-
vasive reporter genes currently available (Alam and Cook,
1990). Luciferase coding regions from several organisms have
been used as reporter genes, although the firefly luciferase
has been the most common choice for the analysis of promoter
function in eukaryotic systems (Alam and Cook, 1990; Aflalo,
1991; Wood, 1991a). The firefly enzyme catalyzes the oxida-
tive decarboxylation of beetle luciferin using O2 and Mg2+-ATP
as substrates. In vitro at least, coenzyme A (COA) may also
participate in the reaction as a cofactor (Wood, 1991b). A pho-
ton is released at 560 nm in 90% of catalytic cycles: this light
emission can be quantified with high sensitivity, most com-
monly in a luminometer. In mammalian cell cultures, the half-life
of luciferase activity is approximately 3 hr (Nguyen et al., 1989;
Thompson et al., 1991), such that decreases in luciferase syn-
thesis over time, which would not be detectable using more
stable reporters such as CAT, may be recapitulated by lumines-
cence. Extremely sensitive video cameras are now available
to detect in vivo bioluminescence noninvasively (Wick, 1989;
Robinson, 1991); however, the unique ability of luciferase to
reveal the temporal regulation of transcription in a single in-
dividual has not yet been exploited in any multicellular
We show here that bioluminescence in vivo faithfully recapit-
ulates the phytochrome and circadian regulation of a cab
promoter fragment fused to the firefly luciferase gene in both
green and etiolated tobacco seedlings. Bioluminescence is
restricted to the cotyledons of cab2::Luc seedlings, confirm-
ing that luciferase accurately reports the spatial pattern of cab
expression. Furthermore, we are able to demonstrate that cir-
cadian regulation of basal cab transcription can occur in the
absence of phytochrome activation.
Cab2::Luc Fusions Reflect Phytochrome Regulation in
Etiolated Tobacco Seedlings
Karlin-Neumann et al. (1988) showed that the abundance of
the native cabe transcript is regulated by phytochrome in etio-
lated Arabidopsis seedlings. An and coworkers have also
shown that a small fragment of the cab3 promoter (positions
-111 to -21), which is 85% identical to the corresponding
region of cab2, mediates shoot-specific and white light-induc-
ible transcription in transgenic tobacco (Mitra et al., 1989).
These studies did not address the induction of circadian rhyth-
micity in etiolated seedlings, nor were the constructs in tobacco
tested with red light. Our earlier studies demonstrated that a
fragment of the Arabidopsis cab2 promoter from position -319
to +3 was sufficient to mediate circadian transcriptional regu-
lation in transgenic tobacco (Millar and Kay, 1991). We therefore
used the -319 cab2 promoter fragment fused to the firefly lu-
ciferase gene (Luc) to test the phytochrome regulation of this
promoter in etiolated tobacco plants and the involvement of
the circadian clock in this regulation.
Etiolated tobacco seedlings carrying the cab2::Luc fusion
gene were harvested at 4-hr intervals after a 2-min exposure
to red light. Figure 1A shows the results of a typical RNA gel
blot analysis of total RNA extracted from these plants. The in-
duction of the cab2::Luc transcript closely parallels that of the
endogenous tobacco cab gene family (Figure 1A). At this level
of time resolution, the initial peak, which occurs 4 hr after the
flash, is not distinguishable from that of the endogenous cab2
mRNA in Arabidopsis (Karlin-Neumann et al., 1988). The first,
Circadian-Regulated Bioluminescence 1077
0 Hours after Red 4 8 12 16 20 24 28
ngLuc 0.5 1 0 4 8 12 16 20 24 28
Figure 1. Analysis of cab2:±uc Expression in Etiolated Seedlings after
a Red Flash.
(A) RNA gel blot analysis. Total RNA prepared from plants harvested
at the times indicated following a 2-min red flash was hybridized to
coding region probes from Luc (LUC), the (3 subunit of mitochondrial
ATPase (BATP), and cab (CAB). Twenty micrograms of total RNA was
loaded in each lane.
(B) Protein gel blot analysis. Total protein prepared from plants har-
vested at the times indicated following a 2-min red flash was incubated
with luciferase antiserum. Eighty micrograms of total protein was loaded
in each lane. 05 and 1 refer to OS and 1 ng of purified luciferase (ng Luc).
narrow peak (4 hr) and the second, broader peak (16 hr) of
mRNA abundance are very similar to the induction pattern ob-
served by Wehmeyer et al. (1990) for mRNA of the tobacco
cab family. The first peak of cab2::Luc mRNA abundance, as
determined by densitometry, shows a fivefold amplitude (com-
paring 0 hr and 4 hr), as against 10- to 15-fold for the cab family,
and the second peak a fivefold to sevenfold increase (0 hr
versus 16 hr) for both Luc and cab. In contrast, the abundance
of mRNA encoding the P subunit of mitochondrial ATPase as-
sayed on the same blot declined slightly over the course of
the experiment. The utility of any reporter of circadian-regulated
transcription depends on the activity of the reporter being rel-
atively unstable. The rapid decrease in cab2::Luc transcript
abundance (fivefold to eightfold from 4 hr to 8 hr) demonstrates
that the fusion transcript is sufficiently unstable to follow the
fluctuations in transcription rate from the cab2 promoter.
We next tested the regulation of luciferase protein concen-
tration under the same conditions. Figure 16 shows the results
of a typical gel blot of total protein extracted from etiolated
cab2::Luc seedlings after a red light flash and probed with
anti-luciferase antiserum. At the level of luciferase protein, the
initial induction is not readily discernable. The second peak,
which coincides with the peak of cab2::Luc transcript, shows
only a slight decrease at 28 hr. Furthermore, the concentra-
tion of luciferase protein increases considerably less than the
cab2::Luc mRNA (approximately threefold, relative to lucifer-
ase standards on the same blot, data not shown), with a
maximum concentration of approximately 2 ng luciferase per
80 ug total protein. These results suggest that a post-
transcriptional mechanism may also affect the accumulation
of luciferase protein.
From these results, we conclude that the bulk luciferase pro-
tein pool does not reflect the abundance of cab2::Luc mRNA;
the use of bioluminescence as a marker, however, depends
on the activity of the luciferase enzyme rather than the total
protein concentration. We therefore used a luminometer to as-
say for luciferase activity in crude extracts from the tissue
samples used in Figure 1B. Figure 2A shows the results of these
in vitro luciferase assays. The luminometer data (in arbitrary
Light Units) has been calibrated using purified luciferase stan-
dards for ease of comparison to data from other luminometers.
In sharp contrast to the pattern of luciferase protein accumu-
lation, the specific luciferase activity follows the pattern of
mRNA induction closely. Both peaks of luciferase activity coin-
cide with the peaks of mRNA abundance (4 hr and 16 to 20
hr), and the amplitude of the first peak (fivefold from 0 hr to
4 hr) is faithfully maintained. The amplitude of the second
peak, however, is 15-fold (from 0 hr to 16 hr) at the level of ac-
tivity, whereas the abundance of cab2::Luc transcript increases
only fivefold to sevenfold. This amplification of the luciferase
response may be due to the relatively long period of lucifer-
ase accumulation during the rise of the second peak; the first
peak may be too brief to allow such amplification. Also, lucifer-
ase activity does not reflect the fivefold to eightfold decrease
in cab2::Luc transcript abundance from the fourth to the eighth
hour but rather shows only a slight decline in this interval. The
maximal activity, 16 hr after the flash, is equivalent to 2 to 3
pg purified luciferase per microgram of total protein, a con-
centration about 10-fold lower than that of luciferase protein
estimated from Figure 1B: this observation is consistent with
a post-translational regulation of luciferase activity in intact plant
tissue. The rapid decline in activity after the second peak
(sevenfold, from 20 hr to 28 hr) indicates that luciferase activ-
ity is sufficiently unstable to monitor both increases and
decreases in cab promoter activity over a timescale of several
Luciferase activity in dark control plants shows a fivefold fluc-
tuation with a peak that coincides with the second peak of red
light-induced activity but reaches only 20% of the induced level.
The pattern is consistent with a period length greater than 20
hr. We have observed the dark cycle in repeated experiments,
although neither the time of seed germination nor the time
of first luciferin application affects the phase of the rhythm (data
not shown). The peak of emission of firefly luciferase is 560 nm,
to which phytochrome is relatively insensitive, so that lucifer-
in application would not be expected to activate phytochrome
1078 The Plant Cell
12 1 6 20 24 28
12 1 6 20 24 28
z! 0.5 /MA"\
Time after Red Flash (hours)
16 20 24 28
Time after Red Flash (hours)
Figure 2. Luciferase Activity in Extracts of Etiolated Seedlings after
a Red Flash.
(A) cab2::Luc seedlings.
(B) 35S::Luc seedlings.
Total protein prepared from plants harvested at the times indicated,
following a 2-min red flash (open symbols) or without light treatment
(filled symbols), was assayed for luciferase activity in a luminometer.
Data in (B) are from red-flashed seedlings. Raw luminometer data were
normalized using purified luciferase standards.
via luciferase luminescence. Etiolated seedlings were grown
in a constant-temperature incubator, as many circadian rhythms
are reset by temperature shifts (Edmunds, 1988) and cyclic
cab expression can be induced in etiolated seedlings by cy-
clic heat shocks (Kloppstech et al., 1991). The possibility of
other, uncontrolled environmental influences cannot be ex-
cluded. This fluctuation does not affect the kinetics or amplitude
of induction, as very similar patterns of luciferase activity are
induced by red flashes administered at different times of day
(data not shown). Figure 28 shows a high leve1 of luciferase
activity in extracts prepared from etiolated tobacco seedlings
bearing luciferase fused to the -343 cauliflower mosaic virus
(CaMV) 35s promoter. The activity in 35S::Luc seedlings rises
to a small peak at 8 to 12 hr and then falls threefold over the
course of the experiment. In agreement with previous reports
(Nagy et al., 1987), this pattern contrasts the accumulation of
activity in cab2::Luc seedlings. The levels of luciferase activ-
ity in Figures 2A and 2B are not directly comparable, as
luciferase is expressed in all organs of the CaMV 35S::Luc
seedlings, whereas cab2::Luc is active only in cotyledons (data
not shown; Nagy et al., 1986; Benfey et al., 1990).
The luciferase assays, although rapid and easy to perform,
are still destructive. We therefore used an extremely sensitive
video camera together with a photon-counting image proces-
sor (both from Hamamatsu Photonic Systems, Bridgewater,
NJ) to quantify luciferase bioluminescence noninvasively in
intact seedlings. Figure 3A shows the quantification of the Iu-
minescence in populations of etiolated cab2::Luc seedlings,
with or without a Bmin red flash. The pattern of luminescence
in vivo closely replicates the induction of the luciferase activ-
ity. This result implies that neither the access of the luciferin
substrate, administered by spraying, nor the availability of en-
dogenous ATP and O2 obscured the temporal pattern of
luminescence. Furthermore, a third peak of luminescence oc-
curs about 28 hr after the center of the second peak (20 hr),
showing the cyclic regulation of red light-induced transcrip-
tion. The third peak appears to be damped by 40% relative
to the second. Figure 38 shows the high luminescence levels
of CaMV 35S::Luc seedlings. As indicated by the in vitro as-
says, luciferase luminescence shows a small peak (less than
twofold in amplitude) after red light treatment, followed by a
reduction in activity. Depletion of endogenous substrate pools
apparently does not modify the temporal pattern of lumines-
cence significantly, even in these strongly expressing plants
during a prolonged time course. The luminescence of 35S::Luc
seedlings without light treatment was also relatively constant
(Figure 38). Furthermore, the images showed luminescence
from all organs of the CaMV 35S::Luc seedlings, whereas the
cab2::Luc plants emitted light only from the cotyledons (data
Cab2::Luc Fusions Reflect Circadian Regulation in
Green Tobacco Seedlings
We previously demonstrated circadian regulation conferred by
the -319 cab2 promoter in green tobacco seedlings when
fused to cat (Millar and Kay, 1991). We therefore investigated
the regulation of the cab2::Luc fusion transcript in plants grown
Circadian-Regulated Bioluminescence 1079
under light/dark cycles for 12 days by harvesting seedlings ev-
ery 4 hr from one 12-hr light/12-hr dark cycle (12L12D) followed
by 24 hr DD. Figure 4A shows the results of an RNA gel blot
from atypical experiment, probed with the Luc coding region.
ZT (hours) 03
6 9 12 15 18 21 24 27 30 33 36 39 42 45 48
. . | , t , . ; -"'
Time after Red Flash (hours)
24 36 48
0 12 24 36 48
Time after Red Flash (hours)
Figure 3. Luciferase Activity in Vivo in Etiolated Seedlings after a Red
(A) cab2::Luc seedlings.
(B) 35S::i.uc seedlings.
Seedlings growing in tissue culture dishes were imaged in a low-light
video system at the times indicated following a 2-min red flash (open
symbols) or without light treatment (filled symbols). Light emission was
quantified using image processing software. Exposure times were 10
min in (A) and 2 min in (B). Counts are not directly comparable be-
tween (A) and (B), as it was necessary to reduce the intensifier gain
ZT (hours) LUC Q 4 8 12 16 20 24 28 32 36 40 44 48
Figure 4. Analysis of cab2::Luc Expression in Green Seedlings.
(A) RNA gel blot analysis. Total RNA prepared from plants harvested
at the times indicated was hybridized to coding region probes from
Luc (LUC) and GUS. Twenty micrograms of total RNA was loaded in
(B) Protein gel blot analysis. Total protein prepared from plants har-
vested at the times indicated was incubated with luciferase antiserum.
Twenty micrograms of total protein was loaded in each lane. 1 ng LUC,
1 ng of purified luciferase.
Open box, light period; filled box, dark period.
The blot was rehybridized with a GUS probe to measure the
transcript produced from the CaMV 35S::GUS fusion present
on the transforming T-DNA. Times are expressed in hours as
Zeitgeber time (ZT), which is simply the time since the onset
of illumination (Zerr et al., 1990). Zeitgeber is a common term
used in chronobiology for the environmental signals that reset
the circadian clock. ZT is used to allow comparison of experi-
ments conducted under different LD regimes.
The circadian rhythm of cab2::Luc transcript abundance is
similar to that of the cao2::caf fusion reported previously (Millar
and Kay, 1991), with maxima at ZTO-3 in LD and ZT24-27 in
DD, minima at ZT15 and ZT42, and a 15-fold amplitude, as de-
termined by densitometry. The lower amplitude (15-fold versus
50-fold) of the cycle suggests that the cab2::Luc transcript may
be slightly more stable than the cab2::cat fusion transcript.
The peak in DD is damped by 30% relative to the peak in LD.
The 35S::GUS transcript shows a contrasting pattern of abun-
dance, with a diurnal fluctuation of low amplitude (twofold to
threefold) followed by a plateau in DD. Figure 4B shows the
results of a typical gel blot of total protein extracts derived from
plants harvested over the time course described for Figure 4A.
The concentration of luciferase protein fluctuates with only a
low amplitude at close to 1 ng of luciferase per 20 ng total
protein in both LD and DD. As in etiolated tissue (Figure 1B),
this suggests a post-transcriptional regulation of luciferase
The concentration of luciferase activity showed close cor-
respondence to mRNA abundance. Figure 5A shows the
luciferase activity in extracts of the samples used for pro-
tein gel blotting (Figure 4B) and in samples harvested
The Plant Cell
12 24 36 48
parison to cab. Figure 58 shows the luciferase activity from
extracts of 35S::Luc plants harvested simultaneously with
those described for Figure 5A; a threefold reduction from ZTO
to ZT12 is consistently observed in LD in repeated experiments
(data not shown). The abundance of the 35S::GUS transcript
(Figure 4A) follows avery similar pattern in LD, suggesting that
the 35s promoter may show a weak diurna1 rhythm in tran-
scriptional activity. The plateau of GUS mRNA in DD, however,
is not reflected in luciferase activity from 35S::Luc; rather, Iu-
ciferase activity in both DD and LL shows further fluctuations
from ZT24 to ZT40. The plateau of GUS mRNA has not been
consistently observed in experiments using this 35S::GUS ref-
erence construct (Millar and Kay, 1991; also, data not shown).
The low-light video imaging system confirmed that bio-
luminescence can act as a nove1 circadian phenotype. Figure
6A shows the quantification of luciferase luminescence in a
typical experiment, using tobacco seedlings grown under LD
for 9 days and imaged at the times shown. Figure 6B shows
the patterns of luminescence in seedlings that were transferred
to constant conditions at ZT12. The patterns of luminescence
in vivo closely match luciferase activity determined in extracts,
with peaks at ZT4,28, and 52 and troughs around ZT20 and
44. The amplitudes of cycles from ZT20 to ZT4 (or ZT28) are
greater than 10-fold in LD (or LL), whereas the damping of the
first peak leads to only a threefold amplitude in DD. The sec-
ond peak in DD is severely damped, such that the increase
in emission from ZT44 to ZT52 is not statistically significant.
The patterns of luminescence from 35S::Luc seedlings, im-
aged immediately before the cab2::Luc seedlings at each time
point, are shown in Figure 6C. The cycle in LD is consistently
observed in repeated experiments, with amplitudes between
twofold and threefold, consonant with the luciferase activity
datafrom extracts (Figure 58). The peaks of 35S::Luc activity,
however, precede those of cab2::Luc by approximately 4 hr.
Luminescence from 35S::Luc in DD shows a threefold to four-
fold decrease over the course of the experiment, possibly
14 1 6 i
Q l27 2ow
Figure 5. Luciferase Activity in Extracts of Green Seedlings.
(A) cab2::Luc seedlings.
(6) 35S::Luc seedlings.
Seedlings were transferred from LD (filled triangles) to LL (open
squares) or DD (filled squares). Total protein prepared from plants har-
vested at the times indicated was assayed for luciferase activity in a
luminometer. Raw luminometer data were normalized using purified
luciferase standards. Open box, light period; filled box, dark period.
simultaneously from plants transferred to LL instead of DD.
Luciferase activity follows a clear circadian cycle, with well-
defined peaks at ZT4 and ZT28, a minimum at ZT16-20 and
an amplitude greater than 10-fold in LD (ZT4 to ZT16). The first
peak of activity in DD is damped by 50% relative to LD, and
activity thereafter falls continuously from ZT28 to ZT48. In con-
trast, luciferase activity in LL rises sharply at ZT48, suggesting
a second peak 24 hr after the first. Thus, in green tissue also,
luciferase activity follows the cyclic accumulation of cab2::Luc
mRNAvery closely. However, the maximal levels of luciferase
activity are equivalent to 6 pg of purified luciferase per micro-
gram of protein, approximately 10-fold less than the protein
concentration estimated from Figure 48. Although the addi-
tion of extracts from untransformed plants has no effect on
the light emission of purified luciferase, the quantum yield of
luciferase synthesized in plants has not been directly mea-
sured. Nevertheless, these results imply that the pool of active
luciferase fluctuates independently of the bulk of luciferase
protein. The CaMV 35S::Luc lines were again used for com-
4 8 12 16 20 O
2 T (hours)
4 8 12
20 24 28 32 36 40 44 48 52 56 60
20 24 28 32 36 40 44 48 52 56 60
Figure 6. Luciferase Activity in Vivo in Green Seedlings.
indicating a slight substrate limitation in prolonged dark
Figure 7 shows a composite of bioluminescence image data
from the experiment described in Figures 6A and 6B: tissue
culture dishes containing 30 to 35 tobacco seedlings are
shown, the first dish during one trough and peak in LD and
the second dish over two cycles in DD. The high amplitude
of cycling in LD and the first peak in DD are clearly visible
by eye, comparing the left- and right-hand panels in the upper
two rows; the damping typical of DD conditions is reflected
in the decreasing luminescence of the panels descending the
right-hand column. We assessed the variability in lumines-
cence between individual seedlings by quantifying the intensity
of all the distinguishable single seedlings in images similar
to those shown in Figure 7. The standard error of the mean
(SE) for peak luminescence in LD, LL, or DD is 9 to 12% of
the mean, but a higher SE was observed in the trough (up to
20% of the mean). This variability is due in part to differences
in the time of seed germination: when the plates from DD were
closely examined, the seedlings that remained brightest in the
night phases were found to be significantly younger than the
rest of the population. Populations from which these younger
seedlings were removed before the experiment began gave
an SE for both peak and trough luminescence of 8 to 12% of
the mean (data not shown).
Cab2::Luc Fusions Reflect Organ-Specific Regulation
We tested the distribution of luminescence among the organs
of cab2::Luc and CaMV 35S::Luc seedlings to confirm the fi-
delity of the luciferase marker at a higher leve1 of spatial
resolution. The differences in expression pattern between the
35s and cab promoters are distinguishable in plates of seed-
lings imaged with a wide-angle lens; Figure 7 is typical of the
cab2::Luc seedlings. Figure 8 shows the patterns more clearly,
at a higher magnification. The cotyledon-specific expression
typical of cab2::Luc and the widespread expression of 35S::Luc
are consistent with datafrom other markers (Mitra et al., 1989;
Benfey et al., 1990; Chang and Walling, 1992). This result im-
plies that most, if not all, tissues are permeable to luciferin
at this stage in development, so the low luminescence inten-
sity observed in the roots of the cab2::Luc plant is due to low
(A) cab2::Luc and 35S::Luc seedlings in LD.
(E) cab2::Luc seedlings in constant conditions.
(C) 35S::Luc seedlings in constant conditions.
Seedlings growing in tissue culture dishes were transferred at ZT12
from LD to LL (open squares) or DD (filled squares), or maintained
in LD (triangles), and imaged in a low-light video system at the times
indicated. Filled triangles, cab; open triangles, 35s; open box, light
period; filled box, dark period. Light emission was quantified using
image processing software. Counts are not directly comparable be-
tween cab2::Luc and 35S::Luc, as it was necessary to reduce the
intensifier gain for35S::Luc.
The Plant Cell
•v-• - • . »
2T28 <D) * " ZT28 (DM
Figure 7. Images of Luminescence from cab2::Luc Seedlings in LD
Seedlings growing in tissue culture dishes were imaged for 7.5 min
in a low-light video system at the times indicated. Background counts
were removed by a noise filter after all images were collected. The
upper panels show seedlings at ZT 20 ("-4") and ZT4 in LD; the mid-
dle and lower panels show seedlings at ZT20, ZT28, ZT44, and ZT52
in a DD period that began at ZT12. The color scale on the right shows
the intensity of luminescence from dark blue (lowest) to white (highest).
luciferase expression. The luminescence per pixel over the
cotyledons of the 35S::Luc seedling is threefold to fourfold
greater than that of the root, whereas the cotyledons of the
cab2::Luc seedling emit 35-fold more than the root. Similar
expression patterns have been described for rbcS::Luc and
35S::i.uc using contact luminography (a plant pressed flat
against film) in older plants (Schneider et al., 1990; Quandt
et al., 1992).
We have demonstrated in this study that the firefly luciferase
gene allows noninvasive, real-time reporting of regulated gene
expression. This technique should be broadly applicable to
the analysis of transcriptional, post-transcriptional, or transla-
tional regulation in any sessile or immobilized organism that
is permeable to luciferin and transparent to green light. Fur-
thermore, the instability of luciferase activity allows the
bioluminescent marker to recapitulate both increases and
decreases in expression over a time course of several hours.
We exploited these characteristics to create a novel marker
for circadian regulation at the molecular level. Luciferase
Figure 8. Organ Specificity of Luminescence in cab2::Luc and
(A) Reference image captured under reflected light.
(B) Luminescence from the same field of view as (A).
Left-hand seedling, cab2::Luc; right-hand seedling, 35S::Luc. Back-
ground counts were removed by a noise filter after the image was
collected. The color scale on the right of (B) shows the intensity of
luminescence from dark blue (lowest) to white (highest).
luminescence in vivo reports many of the details of cab cy-
cling in etiolated and green, intact seedlings, including the
period and phase of free-running rhythms and the damping
of peak levels. We expect that this reporter will greatly facili-
tate studies of the factors influencing the rhythm of transcription
from the cab2 promoter.
Circadian-Regulated Bioluminescence 1083
A video imaging system is not absolutely required for the
noninvasive recording of bioluminescence, as a photomultiplier
tube could perform the same task (as in Knight et al., 1991).
The spatial resolution of the camera, however, allows us to as-
say luciferase activity simultaneously in up to severa1 hundred
individual seedlings (data not shown). Variability of lumines-
cence is accessible to analysis, even to the level of single,
variant individuals in a population: putative mutant seedlings
may readily be recovered and grown to seed. The luciferase
marker therefore constitutes a circadian phenotype, which
should permit a direct genetic screen for clock mutants at the
molecular level, without resorting to the brute force approaches
used to isolate all currently known clock mutants. Alternatively,
organ-leve1 resolution of a few seedlings (Figure 8) and cellu-
lar resolution (Gallie et al., 1989) can be achieved using
appropriate optics. In this manner, we may investigate the tis-
sue focus of the circadian clock and its characteristics at the
cellular level or screen for mutants defective in the spatial regu-
lation of cab gene expression.
We used the in vivo luminescence assay to confirm and ex-
tend our previous analysis of the circadian regulation of the
-319 promoter fragment of Arabidopsis cab2. In green tissue,
the unambiguous circadian regulation of cab2::Luc mRNA,
luciferase activity, and luminescence in vivo is in close agree-
ment with our previous data on cab2::cat fusions (Millar and
Kay, 1991). The damping in DD is of similar magnitude and
transcriptional cycling is maintained in LL, as it is from the
wheat cabl promoter (Nagy et al., 1988a). In etiolated tissue,
the expression patterns we observed are in agreement with
previous data for tobacco (Wehmeyer et al., 1990). However,
the length of the delay between the inductive red flash and
the peaks of luminescence and also the period lengths be-
tween peaks deviated considerably from 24 hr. The first peak
of luminescence may represent an acute response to the in-
ductive red flash, independent of the circadian system. The
subsequent fall in cab expression and the appearance of fur-
ther peaks, in contrast, may be regulated by the circadian clock.
Results from studies of rbcS transcript levels support this hy-
pothesis: rbcS is regulated by a circadian clock in green pea
plants (Kloppstech, 1985), and rbcS induction by red light is
transient in etiolated seedlings (Gallagher et al., 1985). Tomato
and tobacco plants, however, show neither the cycling in green
tissue (Giuliano et al., 1988; Paulsen and Bogorad, 1988;
Piechulla, 1988) nor the transient induction in etiolated tissue
(Wehmeyer et al., 1990). The approximately 30-hr period we
observed between the second and third peaks in etiolated tis-
sue may be interpreted as a transient (a cycle of deviant period,
commonly observed immediately after transfer to free-running
conditions; Sweeney, 1987) oras evidence for a difference in
period length between etiolated and green tissue. Assuming
the clock controlling the second peak of the induced rhythm
is similar to that in green tissue, we may infer that the red light
pulse sets the clock to about ZT6.
The sensitivity of the luciferase system has allowed us to
detect rhythmic activity of the cab promoter in populations of
etiolated plants grown without red light treatment (Figures 2A
and 3A). This result implies that cab gene expression in these
plants is not only rhythmic but that members of the population
are in the same phase, without activation of the phytochrome
system. Tavladoraki et al. (1989) have suggested that levels
of translatable cab mRNA in etiolated bean seedlings may be
regulated by a circadian rhythm, although the experimental
variability of their assay system prevented a firm conclusion.
Neither seed germination nor the first application of luciferin
determined the phase of the dark cycle of luciferase activity
in preliminary experiments (data not shown), but the possibil-
ity of uncontrolled environmental influences currently unknown
to us cannot be excluded. Alternatively, these data may be in-
terpreted as evidence for phase determination during seed
development; rhythmicity is not unknown in seed, as circa-
dian rhythms of gas exchange have been observed even in
dry onion seed (Bryant, 1972). Circadian changes in the spec-
tral properties of phytochrome and in the responsiveness of
hypocotyl unhooking to red light have also been described in
etiolated zucchini seedlings (Horwitz and Epel, 1978). We did
not pursue this line of investigation, as the kinetics and ampli-
tude of the red light response are not affected by the phase
of the red light treatment relative to the dark rhythm (data not
shown). Possible explanations for the latter observation include
the hypothesis that saturation of the phytochrome system ei-
ther resets the dark cycle to a constant phase or uncouples
the cab promoter from it, or that the high induced activity masks
the low level of the dark rhythm while both continue to oper-
ate. The first hypothesis is attractive in that it does not require
two circadian clocks, controlling cab expression in different
conditions, and postulates only an extreme form of the known,
phytochrome-mediated phase-shifting (Tavladoraki et al., 1989).
Luciferase enzyme activity, whether it is assayed in extracts
or in vivo, follows the pattern of cab2::Luc mRNA abundance
closely (comparing Figure 1A with 2A and 3A, and Figure 4A
with 5A and 6A). The major difference between the mRNA and
activity patterns is in the relative heights of the first (4 hr) and
second (16 to 20 hr) peaks of red light-induced activity. Whereas
mRNA abundance reaches comparable levels in these peaks,
luciferase activity is higher in the second peak. The prolonged
high abundance of cab2::Luc mRNA in the second peak may
allow the unstable luciferase activity to rise further than in the
brief, first peak. The use of in vivo bioluminescence as a marker
for circadian regulation depends crucially upon the stability
of both the luciferase transcript and the luciferase enzyme ac-
tivity. Our data are not sufficient to determine accurately the
half-life of the cab2::Luc mRNA, the luciferase protein, or the
luciferase enzyme activityin vivo. The cab2::Luc fusion tran-
script is unstable enough to fall dramatically after accumulation
to high levels, both after red light induction and in diurna1 and
circadian cycles. At the level of protein concentration, in con-
trast, the amplitude of regulation is diminished in both
conditions. This dichotomy indicates that translational or post-
translational mechanism(s) mask the fluctuations of mRNA
abundance. The luciferase protein may be sufficiently stable
in vivo to allow the protein pool to accumulate far beyond the
level of the de novo synthesis that is limited to periods of high
The posttranslational mechanism(s) responsible for the
1084 The Plant Cell
fluctuations in luciferase activity that we observe despite the
relative constancy of the bulk protein concentration is unknown
at present. Severa1 examples of enzymatic activities that fluc-
tuate independently of the enzyme abundance have been
described in plants (for example, Nimmo et al., 1987). Bac-
teria1 luciferase is inactivated in vivo (Tu and Hastings, 1975)
in a process (or processes) dependent on the presence of the
aldehyde substrate (Baldwin et al., 1978); the inactive enzyme
may accumulate to high levels (Mitchell and Hastings, 1970).
Consistent with such a mechanism, the high-amplitude varia-
tions in luciferase activity observed in Figures 2 and 5 are
dependent upon exposure of the plants to luciferin prior to
harvesting (data not shown). The stability of firefly luciferase
activity in the absence of substrate restricts its utility as a
reporter gene in higher plants (Quandt et al., 1992). Further
investigation of the mechanisms responsible for the apparent
changes in specific luciferase activity will be best approached
in a system more amenable to biochemical feeding studies
than intact seedlings.
Earlier publications (Subramani and DeLuca, 1988; Koncz
et al., 1990) identified the access of luciferin to luciferase in
living tissue as a potential problem for the firefly luciferase
reporter system, as beetle luciferin is negatively charged at
neutra1 pH, and also because the native firefly luciferase is
targeted to the peroxisome (Gould et al., 1990). The close cor-
respondence between the patterns of luciferase activity
measured in extracts and in vivo implies that neither the avail-
ability of luciferin nor the endogenous pools of ATP and O2
(and possibly COA; Wood, 1991b) obscured the cyclic pattern
of luminescence in our experiments with cab2::Luc. In studies
using older plant material and promoters expressed in more
diverse cell types than cab, substrate limitation may influence
the patterns of luminescence observed in vivo (Barnes, 1990;
Schneider et al., 1990).
We have recently extended our analysis of circadian regu-
iation using the firefly luciferase system to transgenic
Arabidopsis, with results similar to those described above for
tobacco. We are now pursuing a molecular and genetic ap-
proach to isolate nove1 mutants in the plant circadian system.
Such mutants will be essential in elucidating not only the mo-
lecular architecture of this biological clock and its interaction
with other receptors (notably phytochrome) but also the role
of this system in the regulation of metabolism and develop-
ment in higher plants.
Tobacco plants (var. SR1) were maintained in sterile culture on MS
(Murashige and Skoog, 1962) medium to provide leaves for transfor-
mation experiments. Transformation was performed according to
standard techniques (Horsch et al., 1988). Regenerated kanamycin-
resistant T1 plants (primary transformants) carrying cab2::Luc or
35S::Luc fusions were grown up in a greenhouse to allow seed set.
T2 seed of strongly expressing T1 lines were grown in sterile culture
on MS medium containing kanamycin, at 2PC, for 6 to 10 days in 16L8D,
then for 2 days in 12L12D (green tissue) or for 7 days, wrapped in
aluminium foil, in aconstant-temperature incubator in a dark room (etio-
lated tissue). Red light treatments were for 2 min using the source
described (Nagyet al., 1986). Three to four independently transformed
lines were analyzed for each construct in each experiment. Plants used
for protein or luciferase extracts and for in vivo imaging were sprayed
with 5 mM o-luciferin (Analytical Bioluminescence Laboratories, San
Diego, CA) in 0.010/0 Triton X-100 three times before the start of the
experimental time course and once approximately 30 min prior to each
harvesting or imaging time point.
Clones and Construction
The construction of the -319 to +3 cab2 promoter fragment fused
to the wheat cabl untranslated RNA leader has been described (Millar
and Kay, 1991). This promoter-leader fusion was ligated, at the Sal1
site used previously in cab2:cat fusions, to the firefly luciferase cod-
ing region from plasmid pJD261, kindly provided by Dr. Jeff deWet
(Stanford University) followed by the poly-A addition sequence from
pea nbcS-E9. The cab2::Luc::ES fusion was then transferred to the poly-
linker of the binary vector pMON721 (Monsanto Corporation, St. Louis,
MO) containing a CaMV 35S::GUS::rbcS-3C fusion described previ-
ously (Millar and Kay, 1991). The Luc coding region was also inserted
into the CaMV 35s expression vector pMON530E9 (Cuozzoet al., 1987)
to create the 35S::Luc fusion.
RNA Extraction and Analysis
; s :
Total RNA was prepared from whole seedlings and analyzed in RNA
gel blots as described previously (Nagy et al., 1988~). The probes em-
ployed were from the coding regions of the p subunit of tobacco
mitochondrial ATPase (Boutry and Chua, 1985), Arabidopsis cabl
(Leutweiler et al., 1986), and firefly luciferase (de Wet et al., 1987).
Autoradiograms of RNA gel blots were quantified using a densitome-
ter (Helena Laboratories, Beaumont, TX). Given the fluctuations of the
35S::GUS fusion and P-ATPase transcripts, the amplitude of cab2::Luc
transcript regulation was calculated from the absolute density of the
relevant cab2::Luc bands alone.
Protein Extraction and Analysis
Total protein was extracted from whole seedlings in boiling 2% SDS
extraction buffer and analyzed in protein gel blds as described (Harlowe
and Lane, 1988). In some experiments, 1 x Cell Lysis Reagent (CLR,
containing 1% Triton X-100; Promega) was used with very similar results.
Protein content of the samples was determined using the DC protein
assay kit (Bio-Rad, Richmond, CA), according to the manufacturer's
instructions. Proteins were transferred to supported nitrocellulose mem-
branes (BA-S; Schleicher & Schuell, Keene, NH), and antibody binding
was visualized using the Protc-Blot kit (alkaline phosphatase/NBT reac-
tion) as recommended by the manufacturer (Promega). The luciferase
antiserum was raised from rabbits immunized with purified firefly Iu-
ciferase (United States Biochemical Corp. or Boehringer Mannheim).
Circadian-Regulated Bioluminescence 1085
Luciferase Extraction and Analysis
Total protein was extracted from whole seedlings ground in liquid nitro-
gen, using a 2:l volume ratio of ice cold 1 x CLR, supplied as part
of a luciferase assay kit (Promega). This method extracts luciferase
protein as efficiently as boiling the tissue powder in 2% SDS buffer,
as judged by protein gel blots (see above). The extracts were prepared
in batches of no more than 20, were cleared by two 5-min centrifuga-
tions at lO,OOOg, at 4 O C , and were kept on ice at all times as the luciferase
activity proved unstable in extracts. Protein content of the samples was
determined using the DC protein assay kit (Bio-Rad), according to the
manufacturer’s instructions. Luciferase activity was measured for 10
2-sec intervals in a luminometer (model No. ILA911; Tropix, Bedford,
MA), using 10 pL of extract in 50 pL of Luciferase Assay Reagent
(Promega) according to the manufacturer‘s recommendations. Extracts
of 35S::Luc plants were diluted 10-fold in 1 x CLR containing 1 mglmL
bovine serum albumin (BSA) and assayed as described. Luminescence
from the highest 2-sec count was normalized using a standard curve
prepared with seria1 dilutions of purified firefly luciferase (Boehringer
Mannheim Biochemicals, Indianapolis, IN) in 1 x CLR containing
1 mglmL BSA. In this system, 100 fg of luciferase is routinely detect-
able (100 RLUl2 sec, threefold above background), luminescence is
linearly related to luciferase concentration over at least four orders of
magnitude, and the addition of extracts from etiolated or green, un-
transformed tobacco plants has no effect on the assay (data not shown).
In Vivo lmaging of Luciferase Bioluminescence
Plant materials were grown and sprayed with luciferin as described
above. Seedlings were imaged at 2OoC in tissue culture dishes using
intensified cameras (VIM) and photon-counting image processors
(ARGUS-50 or ARGUS-100) purchased from Hamamatsu Photonic Sys-
tems (Bridgewater, NJ). Exposure times were 7.5 min for green and
10 min for etiolated cab2::Luc seedlings, and 2 min for all CaMV
35S::Luc seedlings. Due to the strong expression of 35S::Luc seed-
lings and limited intrafield dynamic range of the camera, it was
necessary to reduce the gain of the intensifier for the images of these
plants. Therefore, the counts for cab::Luc and 35S::Luc plants are not
directly comparable. Luminescence was quantified from images cap-
tured in centroid-processing mode (this records a single pixel of unit
depth for each photoelectron detected by the camera) and corrected
for background counts from thermal electrons. Figure 8 was taken using
a macio lens, with a 10-min exposure. The images in Figures 7 and
8 were compiled in the ARGUS image processor, passed through a
noise filter to remove background counts, and captured by 35-mm pho-
tography of the imaging monitor.
We are grateful to Dr. Jay Dunlap for helpful discussion in the early
part of this work, to Drs. Keith Wood (Promega) and Kazuyuki Hiratsuka
for helpful discussions and critical reading of the manuscript, to Ellen
Leheny and Stanley Sotnikov for their assistance in plant transforma-
tion and maintenance, to Arnold Hinton for photography, and to Dr.
Richard Pine for the use of the densitometer. A genomic clone of the
cab2 gene was kindly provided by Dr. Elaine Tobin, and the firefly Iu-
ciferase cDNA by Dr. Jeff deWet. We are also grateful to Dr. Robert
Wick and Masafumi Oshiro of Hamamatsu Photonic Systems, Inc.,
for introducing the authors (A.J.M. and S.A.K.) to photon-counting im-
aging at their facilities. This work was supported by National lnstitutes
of Health Grant No. GM 44640 and a grant from the Human Frontier
Science Program to N.-H.C. and National Science Foundation Grant
No. MCB-9216399 to S.A.K. A.J.M. is supported by a William O. Eaker
Fellowship, through a grant from the Mellon Foundation. S.A.K. is sup-
ported by a grant from the W. M. Keck Foundation.
Received July 13, 1992; accepted July 17, 1992.
Adamska, I., Scheel, B., and Kloppstech, K. (1991). Circadian oscil-
lations of nuclear-encoded chloroplast proteins in pea (Pisum
sativum). Plant MOI. Biol. 17, 1055-1065.
Aflalo, C. (1991). Biologically localized firefly luciferase: A tool to study
cellular processes. Int. Rev. Cytol. 130, 269-323.
Alam, J., and Cook, J.L. (1990). Reporter genes: Applications to the
study of mammalian gene transcription. Anal. Biochem. 188,
Baldwin, T.O., Nicoli, M.Z., and Reeve, C.A. (1978). Bacterial lucifer-
ase: Proteolysis in vitro and inactivation in vivo. In Fogarty Conference
on Limited Proteolysis in Microorganisms, H. Holzer and G. Cohen,
eds (Bethesda, MD: National lnstitutes of Health), pp. 217-223.
Barnes, W.M. (1990). Variable patterns of expression of luciferase in
transgenic tobacco leaves. Proc. Natl. Acad. Sci. USA 87,9183-9187.
Benfey, P.N., Ren, L., and Chua, N A (1990). Tissue-specific ex-
pression from CaMV 35s enhancer subdomains in early stages of
plant development. EMBO J. 9, 1677-1684.
Boutry, M., and Chua, N.-H. (1985). A nuclear gene encoding the
p subunit of the mitochondrial ATPase in N. plumbaginifolia. EMBO
J. 4, 2159-2165.
Bryant, T.R. (1972). Gas exchange in dry seeds: Circadian rhythmic-
ity in the absence of DNA replication, transcription, and translation.
Science 178, 634-636.
Busheva, M., Garab, G., Liker, E., Tóth, i ! . , Szbll, M., and Nagy,
F. (1991). Diurna1 fluctuations in the content and functional proper-
ties of the light harvesting chlorophyll alb complex in thylakoid
membranes. Correlation with the diurna1 rhythm of the mRNA level.
Plant Physiol. 95, 997-1003.
Chang, Y.C., and Walling, L.L. (1992). Spatial and temporal expres-
sion of Cab mRNAs in cotyledons in the developing soybean
seedling. Planta 186, 262-272.
Cuouo, M., OConnell, K., Kaniewski, W., Fang, I?.-X., Chua,
N.-H., and Tumer, N. (1987). Vira1 protection in transgenic plants
expressing the cucumber mosaic virus coat protein or its antisense
RNA. Biotechnology 6, 549-557.
de Wet, J.R., Wood, K.V., DeLuca, M., Helinski, D.R., and
Subramani, S. (1987). Firefly luciferase gene: Structure and expres-
sion in mammalian cells. MOI. Cell. Biol. 7, 725-737.
Dunlap, J.C. (1990). Closely watched clocks: Molecular analysis of cir-
cadian rhythms in Neurospora and Drosophila. Trends Genet. 6,
Edmunds, L.N., ed (1988). Cellular and Molecular Bases of Biologi-
cal Clocks (New York: Springer-Verlag).
The Plant Cell
Engelmann, W., and Johnsson, A. (1978). Attenuation of the peta1
movement rhythm in Kalanchoe with light pulses. Physiol. Plant.
Fejes, E., Pay, A., Kanevsky, I., Szell, M., Adam, E., Kay, S., and
Nagy, F . (1990). A 268-bp upstream sequence mediates the circa-
dian clock-regulated transcription of the wheat Cab-1 gene in
transgenic plants. Plant MOI. Biol. 15, 921-932.
Gallagher, T.F., Jenkins, G.I., and Ellis, R.J. (1985). Rapid modula-
tion of transcription of nuclear genes encoding chloroplast proteins
by light. FEBS Lett. 186, 241-245.
Gallie, D.R., Lucas, W.J., and Walbot, V. (1989). Visualizing mRNA
expression in plant protoplasts: Factors influencing efficient mRNA
uptake and translation. Plant Cell 1 , 301-311.
Giuliano, G., Hoffman, N.E., Ko, K., Scolnik, P.A., and Cashmore,
AR. (1988). A light-entrained circadian clock controls transcription
of severa1 plant genes. EMBO J. 7 , 3635-3642.
Gould, S.J., Keller, G.-A., Schneider, M., Howell, S.H., Garrad, L.J.,
Goodman, J.M., Distel, B., Tabak, H., and Subramani, S. (1990).
Peroxisomal protein import is conserved between yeast, plants, in-
sects and mammals. EMBO J. 9, 85-90.
Green, B.R., Pichersky, E., and Kloppstech, K. (1991). Chlorophyll
a/b-binding proteins: An extended family. Trends Biochem. Sci. 16,
Hall, J.C. (1990). Genetics of circadian rhythms. Annu. Rev. Genet.
Hardin, P.E., Hall, J.C., and Rosbash, M. (1990). Feedback of the
Dmsophila period gene product on circadian cycling of its messenger
RNA levels. Nature 343, 536-540.
Harlowe, E., and Lane, D. (1988). Antibodies (Cold Spring Harbor,
NY: Cold Spring Harbor Laboratory).
Horsch, R., Fry, J., Hoffman, N., Neidermeyer, J., Rogers, S., and
Fraley, R. (1988). Leaf disc transformation. In Plant Molecular Biol-
ogy Manual, A5, S. Gelvin and R. Schilperoort, eds (Dordrecht, The
Netherlands: Kluwer Academic), pp. 1-23.
Horwitz, B.A., and Epel, B.L. (1978). Circadian changes in activity
of the far-red form of phytochrome: Physiological and in vivo spec-
trophotometric studies. Plant Sci. Lett. 13, 9-14.
Karlin-Neumann, G.A., Sun, L., and Tobin, E.M. (1988). Expression
of light-harvesting chlorophyll alb-protein genes is phytochrome-
regulated in etiolated Arabidopsis thaliana seedlings. Plant Phys-
iol. 88, 1323-1331.
Kay, S.A., and Millar, A.J. (1992). Circadian regulated Cab gene tran-
scription in higher plants. In The Molecular Biology of Circadian
Rhythms, M. Young, ed (New York: Marcel Dekker), pp. 73-89.
Kendrick, R.E., and Kmnenberg, G.H.M., eds (1986). Photomorpho-
genesis in Plants (Dordrecht, The Netherlands: Martinus Nijhoff).
Kloppstech, K. (1985). Diurnal and circadian rhythmicity in the ex-
pression of light-induced plant nuclear messenger RNAs. Planta
Kloppstech, K., Otto, B., and Sierralta, W . (1991). Cyclic tempera-
ture treatments of dark-grown pea seedlings induce a rise in specific
transcript levels of light-regulated genes related to photomorpho-
genesis. MOI. Gen. Genet. 225, 468-473.
Knight, M.R., Campbell, A.K., Smith, S.M., andTrewavas, A.J. (1991).
Transgenic plant aequorin reports the effects of touch and cold-shock
and elicitors on cytoplasmic calcium. Nature 352, 524-526.
Koncz, C., Langridge, W.H.R., Olsson, O., Schell, J., and Szalay,
A.A. (1990). Bacterial and firefly luciferase genes in transgenic plants:
Advantages and disadvantages of a reporter gene. Dev. Genet. 11,
Lecharny, A., Tremolières, A., and Wagner, E. (1990). Correlation
between the endogenous circadian rhythmicity in growth rate and
fluctuations in oleic acid content in expanding stems of Chenopo-
dium rubrum L. Planta 182, 211-215.
Leutweiler, L.S., Meyerowitz, E.M., and Tobin, E.M. (1986). Struc-
ture and expression of three light-harvesting chlorophyll alb-binding
protein genes in Arabidopsis thaliana. Nucl. Acids Res. 14,
Loros, J.J., and Dunlap, J.C. (1991). Neurospora crassa clockcontrolled
genes are regulated at the leve1 of transcription. MOI. Cell. Biol. 11,
Lumsden, P . J . (1991). Circadian rhythms and phytochrome. Annu. Rev.
Plant Physiol. Plant MOI. Biol. 42, 351-371.
Millar, A.J., and Kay, S.A. (1991). Circadian control of cab gene tran-
scription and mRNA accumulation in Arabidopsis. Plant Cell 3,
Mitchell, G.W., and Hastings, J.W. (1970). Light-induced biolumines-
cence. lsolation and characterization of a specific protein involved
in the absorption and delayed emission of light. Biochemistry 9,
Mitra, A., Choi, H.K., and An, G. (1989). Structural and functional
analyses of Arabidopsis thaliana chlorophyll alb-binding protein (cab)
promoters. Plant MOI. Biol. 12, 169-179.
Murashige, T., and Skoog, F . (1962). A revised medium for rapid growth
and bioassays with tobacco tissue. Physiol. Plant. 15, 493-497.
Nagy, F., Kay, S.A., Boutry, M., Hsu, M.-Y., and Chua, N.-H. (1986).
Phytochrome-controlled expression of a wheat Cab gene in trans-
genic tobacco seedlings. EM60 J. 6, 1119-1124.
Nagy, F . , Boutry, M., Hsu, M.-Y., Wong, M., and Chua, N.-H. (1987).
The 5’-proximal region of the wheat Cab-i gene contains a 268-bp
enhancer-like sequence for phytochrome response. EMBO J. 9,
Nagy, F., Kay, S.A., and Chua, N.-H. (1988a). A circadian clock regu-
lates transcription of the wheat Cab-1 gene. Genes Dev. 2,376-382.
Nagy, F., Kay, S.A., and Chua, N.-H. (1988b). Gene regulation by
phytochrome. Trends Genet. 4, 37-42.
Nagy, F., Kay, S.A., and Chua, N.-H. (1988~). Analysis of gene ex-
pression in transgenic plants. In Plant Molecular Biology Manual,
S. Gelvin and R. Schilperoort, eds (Dordrecht, The Netherlands:
Kluwer Academic Publishers), pp. 1-29.
Nguyen, V . T . , Morange, M., and Bensaude, O. (1989). Protein denatu-
ration during heat shock and related stress. J. Biol. Chem. 264,
Nimmo, G.A., Wilkins, M.B., Fewson, C.A., and Nimmo, H.G. (1987).
Persistem circadian rhythms in the phosphorylation state of phos-
phoenolpyruvate carboxylase from Sryophyllum fedtschenkoi leaves
and in its sensitivity to inhibition by malate. Planta 170, 408-415.
Paulsen, H., and Bogorad, L. (1988). Diurnal and circadian rhythms
in the accumulation and synthesis of mRNA for the light-harvesting
chlorophyll alb-binding protein in tobacco. Plant Physiol. 88,
Piechulla, B. (1988). Plastid and nuclear mRNA fluctuations in tomato
leaves-diurna1 and circadian rhythms during extended dark and
light periods. Plant MOI. Biol. 11, 345-353.
Circadian-Regulated Bioluminescence 1087
Quail, P.H. (1991). Phytochrome: A light-activated molecular switch
that regulates plant gene expression. Annu. Rev. Genet. 25,389-409.
Quandt, H.J., Broer, I., and Pühler, A. (1992). Tissue-specific activ-
ity and light-dependent regulation of a soybean rbcS promoter in
transgenic tobacco plants monitored with the firefly luciferase gene.
Plant Sci. 82, 59-70.
Roblnson, C. (1991). Capturing the light and capturing the market.
Trends Biotech. 9, 72-76.
Rosbash, M., and Hall, J.C. (1989). The molecular biology of circa-
dian rhythms. Neuron 3, 387-398.
Schneider, M., Ow, D.W., and Howell, S.H. (1990). The in vivo pat-
tern of firefly luciferase expression in transgenic plants. Plant MOI.
Biol. 14, 935-947.
Subramani, S., and DeLuca, M. (1988). Applications of firefly lucifer-
ase as a reporter gene. In Genetic Engineering: Principles and
Methods, Vol 10, J.K. Setlow, ed. (New York: Plenum Press), pp.
Sweeney, B.M. (1987). Rhythmic Phenomena in Plants (San Diego:
Tavladoraki, P . , Kloppstech, K., and Argyroudi-Akoyunoglou, J.
(1989). Circadian rhythm in the expression of the mRNA coding for
the apoprotein of the light-harvesting complex of photosystem (I.
Phytochrome control and persistent far red reversibility. Plant
Physiol. 90, 665-672.
Thompson, J.F., Hayes, L.S., and Lloyd, D.B. (1991). Modulation of
firefly luciferase stability and impact on studies of gene regulation.
Gene 103, 171-177.
Tu, S.-C., and Hastings, J.W. (1975). Photoexcited bacterial lumines-
cence. Spectral properties and mechanistic implication of a reduced
flavine-like prosthetic group associated with photoexcitable lucifer-
ase. Biochemistry 14, 1975-1980.
Wehmeyer, B., Cashmore, A.R., and Schafer, E. (1990). Photocon-
trol of the expression of genes encoding chlorophyll alb binding
proteins and small subunit of ribulose-l,5-bisphosphate carboxylase
in etiolated seedlings of Lycopersicon esculentum (L.) and Nicotiana
tabacum (L.). Plant Physiol. 93, 990-997.
Wick, R.A. (1989). Photon counting imaging: Applications in biomed-
ical research. BioTechniques 7 , 262-268.
Wood, K.V. (1991a). Recent advances and prospects for use of beetle
luciferases as genetic reporters. In Bioluminescence and Chemilu-
minescence: Current Status, P.E. Stanley and J. Kricka, eds
(Chichester, UK John Wiley & Sons), pp. 543-546.
Wood, K.V. (1991b). The origin of beetle luciferases. In Biolumines-
cence and Chemiluminescence: Current Status, P.E. Stanley and
J. Kricka, eds (Chichester, UK: John Wiley & Sons), pp. 11-13.
Zerr, D.M., Hall, J.C., Rosbash, M., and Siwicki, K.K. (1990). Circa-
dian fluctuations o f period protein immunoreactivity in the CNS and
the visual system of Drosophila. J. Neurosci. 10, 2749-2762.
Zwiebel, L.J., Hardin, P.E., Liu, X., Hall, J.C., andRosbash, M. (1991).
A post-transcriptional mechanism contributes to circadian cycling
of a per-P-galactosidase fusion protein. Proc. Natl. Acad. Sci. USA
A. J. Millar, S. R. Short, N. H. Chua and S. A. Kay
A Novel Circadian Phenotype Based on Firefly Luciferase Expression in Transgenic Plants
This information is current as of July 10, 2011
Sign up for eTOCs at:
Sign up for CiteTrack Alerts at:
is available at:
The Plant Cell
Subscription Information for
ADVANCING THE SCIENCE OF PLANT BIOLOGY
© American Society of Plant Biologists