Direct targeting of light signals to a promoter element-bound transcription factor.

Abstract

Light signals perceived by the phytochrome family of sensory photoreceptors are transduced to photoresponsive genes by an unknown mechanism. Here, we show that the basic helix-loop-helix transcription factor PIF3 binds specifically to a G-box DNA-sequence motif present in various light-regulated gene promoters, and that phytochrome B binds reversibly to G-box-bound PIF3 specifically upon light-triggered conversion of the photoreceptor to its biologically active conformer. We suggest that the phytochromes may function as integral light-switchable components of transcriptional regulator complexes, permitting continuous and immediate sensing of changes in this environmental signal directly at target gene promoters.

Full text

Page 1

DOI: 10.1126/science.288.5467.859
, 859 (2000);
288
Science
et al.Jaime F. Martínez-García,
Element-Bound Transcription Factor
Direct Targeting of Light Signals to a Promoter

www.sciencemag.org (this information is current as of February 3, 2009 ):
The following resources related to this article are available online at
http://www.sciencemag.org/cgi/content/full/288/5467/859
version of this article at:
including high-resolution figures, can be found in the online
Updated information and services,
http://www.sciencemag.org/cgi/content/full/288/5467/859#otherarticles
, 17 of which can be accessed for free:
cites 41 articles
This article
247 article(s) on the ISI Web of Science.
cited by
This article has been
http://www.sciencemag.org/cgi/content/full/288/5467/859#otherarticles
98 articles hosted by HighWire Press; see:
cited by
This article has been
http://www.sciencemag.org/cgi/collection/botany
Botany
:
subject collections
This article appears in the following
http://www.sciencemag.org/about/permissions.dtl
in whole or in part can be found at:
this article
permission to reproduce
of this article or about obtaining
reprints
Information about obtaining
registered trademark of AAAS.
is aScience2000 by the American Association for the Advancement of Science; all rights reserved. The title
CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.
(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the Science
on February 3, 2009
www.sciencemag.org
Downloaded from

Page 2

A remarkable up-down asymmetry in the
nystagmus, together with the strong vestibular
input to up-BT cells, appears to suggest an
additional vestibular imbalance. The region
where we injected muscimol contained up-BT
neurons and no down eye movement cells. Be-
cause up-BT neurons received excitatory inputs
from the anterior semicircular canals, their inac-
tivation might be expected to result in downward
drift of the eye due to a decreased anterior canal
input. Instead, the eye drifted upward in a wide
oculomotor range, suggesting increased, rather
than decreased, signals from the anterior canal.
What is the neural mechanism that produces
thecentralvestibularimbalanceofanteriorcanal
dominance? Up-BT neurons probably project to
the flocculus, which has a unique connection
pattern with the vertical canal system. Only
anterior canal–related vestibular nucleus neu-
rons receive floccular inhibition (12). Therefore,
inactivation of the up-BT neurons may reduce
the activity of Purkinje cells, leading to disinhi-
bition of vestibular neurons that receive inputs
from the anterior canal. These vestibular neu-
rons then exhibit increased discharge. The ante-
rior canal input to the brainstem circuitry is
increased while the posterior canal input re-
mains unchanged. This idea is supported by a
similar downbeat nystagmus after floccular le-
sions (1). This experiment thus suggests that the
PMT-flocculus-vestibular nucleus pathway is
important in maintaining vestibular balance. It
should be noted, however, that the asymmetry
of nystagmus we observed does not necessarily
indicate an imbalance of vestibular inputs to the
neural integrator. The PMT-flocculus pathway
may be involved in an intrinsic mechanism of
the integrator that sets the neutral eye position.
It is known that the cerebellum is necessary
for normal operation of the brainstem neural
integrators (1, 3, 4). The cerebellum must ac-
quire oculomotor signals from the brainstem.
This study suggests that our up-BT neurons
relay eye position information to the flocculus.
Furthermore,theeffectofinactivationofup-BT
cells indicates their importance in the oculomo-
tor integration. The caudal pontine PMT area
may be a new component of the neural integra-
tion system for vertical, and perhaps horizontal,
eye movement, along with the midbrain inter-
stitial nucleus of Cajal, vestibular nuclei, and
nucleus prepositus hypoglossi.
References and Notes
1. D. S. Zee, A. Yamazaki, P. H. Butler, G. Gu ¨cer, J. Neu-
rophysiol. 46, 878 (1981).
2. D. A. Robinson, in Basic Mechanisms of Ocular Motil-
ity and Their Clinical Implications, G. Lennerstrand
and P. Bach-y-Rita, Eds. (Oxford Univ. Press, New
York, 1975), pp. 337–374.
3. D. A. Robinson, Brain Res. 71, 195 (1974).
4. B. Y. Kamath and E. L. Keller, Math. Biosci. 30, 341
(1976); D. S. Zee, R. J. Leigh, F. Mathieu-Millaire, Ann.
Neurol. 7, 37 (1980).
5. J. A. Bu ¨ttner-Ennever, A. K. E. Horn, K. Schmidtke, Rev.
Neurol. 145, 533 (1989); J. A. Bu ¨ttner-Ennever and
A. K. E. Horn, Ann. N.Y. Acad. Sci. 781, 532 (1996).
6. S. Nakao, I. Curthoys, C. H. Markham, Brain Res. 183,
291 (1980); G. Cheron, S. Saussez, N. Gerrits, J. Neu-
rophysiol. 74, 1367 (1995).
7. Five adult cats were prepared for recording of neuronal
activity in the brainstem. All experiments were done
with the permission of the Animal Experiment Commit-
tee of the University of Tsukuba, which is operated in
accordance with Japanese Governmental Law (no. 105).
A coil was implanted subconjunctivally under pentobar-
bital sodium anesthesia and aseptic conditions to mea-
sure eye movement by a magnetic-search coil tech-
nique (13). The tympanic bulla on each side was
opened, and silver-ball electrodes were implanted on
the round window to stimulate the vestibular nerve.
After recovery from surgery, each animal was trained to
accept restraining conditions without stress. A position
16.5° nose up from the stereotaxic horizontal was taken
as a zero vertical position. This was near the center of
the vertical oculomotor range. Glass-coated tungsten
electrodes were used for extracellular recordings. Care
was taken to ensure that the recording was from a cell
body and not from an axon. Only negative-positive
spikes with a duration (time to positive peak) of ?250
?s were regarded as action potentials of the soma.
These unit spikes could be recorded not only during
advancement but also during withdrawal of the elec-
trode, another indication of somatic recording. The MLF
was identified physiologically by recording monosynap-
tic volleys from the vestibular nerves and by monosyn-
aptic activation of secondary vestibular axons, which all
exhibitedinitialpositivity.Theexploredregionextended
rostrocaudally from 2.0 mm anterior to 1.5 mm poste-
rior of the rostral pole of the abducens nucleus and
ventrally to 3.5 mm from the floor of the fourth ven-
tricle. The lateral limit of the region extended from the
midline to about 1.5 mm. The tonic firing rate of a
cell was defined as an average rate for the fixation
period. The size of the burst component for a given
saccade was estimated by subtracting the eye posi-
tion–dependent component from the total number
of spikes [see (14)]. For natural vestibular stimulation,
the turntable was rotated sinusoidally in the light in
two mutually orthogonal planes approximately co-
planar with the two vertical canal pairs (45° away
from the pitch-and-roll planes). The two planes are
called the c-ac/i-pc plane and the i-ac/c-pc plane,
depending on their relation to the side of neurons
studied. Because the null position, an asymptote of
the exponential drift, appeared to vary from one slow
phase to another, we estimated the null and TC for
each slow phase. We first determined the null as a
position best linearizing the logarithm of eye dis-
placement from that position as a function of time.
At the termination of the experiments, some record-
ing sites were marked by making small electrolytic
lesions. The animals were then killed with a lethal
dose of pentobarbital sodium and perfused. In three
alert cats, muscimol, an inhibitory neurotransmitter,
?-aminobutyric acid type A receptor agonist (1.0 ?g
per microliter of saline, 0.2 to 0.6 ?l), was injected in
the MLF of the pons with an injection needle to
suppress spike activity of somata without affecting
that of MLF fibers.
8. W. Graf and K. Ezure, Exp. Brain Res. 63, 35 (1986);
R. A. McCrea, A. Strassman, S. M. Highstein, J. Comp.
Neurol. 264, 571 (1987).
9. S. C. Cannon and D. A. Robinson, J. Neurophysiol. 57,
1383 (1987).
10. E. Godaux, P. Mettens, G. Cheron, J. Physiol. London
472, 459 (1993); P. Mettens, E. Godaux, G. Cheron,
H. L. Galiana, J. Neurophysiol. 72, 785 (1994).
11. J. D. Crawford, W. Cadera, T. Vilis, Science 252,
1551(1991).
12. Y. Sato and T. Kawasaki, J. Neurophysiol.64, 551 (1990).
13. A. F. Fuchs and D. A. Robinson, J. Appl. Physiol. 21,
1068 (1966).
14. S. Chimoto, Y. Iwamoto, K. Yoshida, J. Neurophysiol.
81, 1199 (1999).
15. Supported by Core Research for Evolutionary Science
and Technology of Japan Science and Technology Cor-
poration. We thank S. Shoji for helpful comments.
22 November 1999; accepted 3 March 2000
Direct Targeting of Light Signals
to a Promoter Element–Bound
Transcription Factor
Jaime F. Martı ´nez-Garcı ´a, Enamul Huq, Peter H. Quail*
Light signals perceived by the phytochrome family of sensory photoreceptors
are transduced to photoresponsive genes by an unknown mechanism. Here, we
show that the basic helix-loop-helix transcription factor PIF3 binds specifically
to a G-box DNA-sequence motif present in various light-regulated gene pro-
moters, and that phytochrome B binds reversibly to G-box–bound PIF3 spe-
cifically upon light-triggered conversion of the photoreceptor to its biologically
active conformer. We suggest that the phytochromes may function as integral
light-switchable components of transcriptional regulator complexes, permit-
ting continuous and immediate sensing of changes in this environmental signal
directly at target gene promoters.
Plants use a set of sensory photoreceptors to
monitor the environment for informational light
signals (1). The phytochrome (phy) family,
comprising five members (phyA to phyE) in
Arabidopsis, track the red (R) and far red (FR)
light wavelengths by virtue of their capacity for
photoinduced, reversible switching between
two conformers: the R-absorbing, biologically
inactive Pr form and the FR-absorbing, biolog-
ically active Pfr form. Each phy molecule is a
dimer of subunits that consist of a ?125-kD
polypeptide with a covalently bound tetrapyr-
role chromophore that is autocatalytically at-
tached by the apoprotein (2). Light-driven Pfr
formation induces changes in the expression of
numerous genes underlying various aspects of
Department of Plant and Microbial Biology, University
of California, Berkeley, CA 94720, and U.S. Depart-
ment of Agriculture–Agricultural Research Service
Plant Gene Expression Center, 800 Buchanan Street,
Albany, CA 94710, USA.
*To whom correspondence should be addressed. E-
mail: quail@nature.berkeley.edu
R E P O R T S
www.sciencemag.org SCIENCEVOL 2885 MAY 2000
859
on February 3, 2009
www.sciencemag.org
Downloaded from

Page 3

plant photomorphogenesis, and promoter anal-
ysis has identified a number of cis-acting light-
responsive elements (LREs) and some cognate
DNA-binding proteins involved in regulating
expression (3).
Considerable progress has been made in
recent times toward identifying molecular com-
ponents potentially involved in early steps in
the signaling pathways linking the phyto-
chromes to photoresponsive genes. Evidence
from photoreceptor mutants in Arabidopsis in-
dicates that individual members of the phy fam-
ily have differential photosensory and/or phys-
iological functions in controlling development
(4–7), and genetic screens have identified sev-
eral loci specific to either phyA or phyB sig-
naling pathway segments (8–13). Molecular
cloning of two of these loci, FAR1 and SPA1,
specific to phyA signaling, has revealed that
they encode nuclear proteins (13, 14). Yeast
two-hybrid screening for phytochrome-interact-
ing proteins has identified PKS1, a cytoplasmic
protein (15), NDPK2, a nucleoside diphosphate
kinase (16), and PIF3, a nuclear-localized basic
helix-loop-helix (bHLH) protein (17). The
functionsofPKS1andNDPK2inphytochrome
signaling remain to be determined. However,
because PIF3 belongs to the bHLH superfamily
of transcription factors (18, 19), the possibility
of a direct signaling pathway from the photo-
receptor to target genes is suggested. This sug-
gestion is consistent with recent evidence that
phyA and phyB are induced to translocate from
the cytoplasm to the nucleus upon Pfr forma-
tion (20, 21). To explore this possibility, we
examined whether PIF3 has sequence-specific
DNA binding activity and, if so, whether phyB
would interact with DNA-bound PIF3.
Using a random binding site selection
(RBSS) procedure, we identified a palindromic
hexanucleotide DNA sequence—CACGTG,
known as a G-box motif (3, 22–26)—as the
core PIF3 target element (Fig. 1A). The speci-
ficity of this interaction was verified by electro-
phoretic mobility shift assay (EMSA), with the
use of a G-box containing probe (G-wt) repre-
sentative of those selected by RBSS (Fig. 1B),
and recombinant PIF3 synthesized in the TnT
in vitro transcription-translation system. Figure
1C shows that the low-mobility complex
formed in the presence of the PIF3 template–
programmed TnT reaction (lane 3) was effec-
tively competed by unlabeled G-wt probe
(lanes 4 to 6), but not an unlabeled mutant
probe (lanes 7 to 9) containing a single T to G
substitution in the G-box (G-mut, Fig. 1B). By
contrast, the higher mobility complex formed
by an endogenous TnT reaction component
(Fig. 1C, lane 2) was competed equally well
with G-wt and G-mut unlabeled probes. The
data show that PIF3 does indeed bind DNA,
withtargetsequencespecificitycharacteristicof
the bHLH family, and probably as a ho-
modimer, on the basis of the known structure of
DNA-bHLH protein complexes (27). The G-
box motif is found in a variety of light-regulat-
ed genes and has been implicated in the regu-
lation of some by functional assay (3, 22–26).
This motif is a representative of the more gen-
eral E-box motif, CANNTG, considered to be
the core consensus sequence for bHLH proteins
in nonplant systems (18, 19). The PIF3 bHLH
domain alone is sufficient for sequence-specific
binding to the G-box, similar to other bHLH
proteins (Fig. 1D) (19, 27).
Full-length, chromophore-conjugated phyB
interacts with PIF3 that is not bound to DNA
only upon conversion to the Pfr form (28). To
determine whether phyB would bind to PIF3
that had formed a complex with its DNA target
site, we performed EMSA with PIF3 and phyB
together (Fig. 2A). Neither the PHYB apopro-
tein nor photoactive phyB in either conformer
interacted directly with the DNA probe (Fig.
2B, lanes 3 to 5). Similarly, neither PHYB nor
the phyB Pr form (PrB) altered the mobility or
abundance of the PIF3-DNA complex when
added to that complex, indicating the absence
of any interaction (Fig. 2B, lanes 7 and 8). By
contrast, R irradiation of chromophore-conju-
gated phyB induced formation of a discrete,
lower mobility complex, presumably corre-
sponding to a ternary complex between PIF3,
phyB, and the DNA probe (Fig. 2B, lane 9).
The data indicate, therefore, that phyB does
indeed bind specifically to DNA-bound PIF3,
but only upon R light–induced conversion to
the Pfr form (PfrB). Figure 2B also shows that
phyB does not interact with the bHLH domain
of PIF3 (G:bhPIF3) when this isolated domain
is bound to its target sequence (lanes 10 to 13).
The data indicate, therefore, that the conformer-
specific recognition of PIF3 by phyB requires
molecular determinants outside the DNA-bind-
ing domain.
To determine whether the R light–induced
binding of phyB to DNA-bound PIF3 was re-
versible, we examined the effects of FR pulses
given after an initial R pulse on the ternary
complex detected by EMSA. The amount of
R-induced complex was rapidly reduced by
subsequent exposure to FR (Fig. 2D), indicat-
ing that the interaction triggered by Pfr forma-
Fig. 1. PIF3 is a se-
quence-specific
binding protein that tar-
gets G-box
through its bHLH do-
main. (A) Summary of
DNAsequencesselected
by random binding site
selection (RBSS)
with either E. coli–pro-
duced GST:PIF3:flag pro-
tein (top) or TnT-ex-
pressedHis6:PIF3protein
(bottom) (42). The con-
sensus-selected G-box is
indicated in bold type.
(B) Upper-strand nucle-
otide sequences of the
probes used for the ex-
perimentsin(C)and(D).
The G-box sequence is
highlighted in bold type.
G-wt was selected by
RBSS. G-mut contains a
pointmutationintheG-
box sequence (under-
lined). The nucleotides
added for probe labeling
are indicated in lower-
case. (C and D) Se-
quence specificity
PIF3 DNA-binding activ-
ity determined by com-
petitive EMSA. (C) PIF3
binding to labeled G-wt
probeiscompetedspecificallybycoldG-wtbutnotbycoldG-mut.AnonspecificcomplexformedbyaTnT
component in the absence of PIF3 (*) is competed equally by G-wt and G-mut probes. Lane 1, no protein;
lane 2, mock-translated TnT; lanes 3 to 9, PIF3. The binding complexes were competed by the addition of
none (lane 3), 5? (lanes 4 and 7), 25? (lanes 5 and 8), and 125? (lanes 6 and 9) molar excess of unlabeled
G-wt (lanes 4 to 6) or G-mut (lanes 7 to 9). (D) A truncated fragment of PIF3 containing only the bHLH
domain fused to GST (G:bhPIF3) binds to the labeled G-wt probe, and this binding is also competed by cold
G-wt,butnotbycoldG-mut.Lane1,PIF3;lanes2to8,G:bhPIF3.Thebindingcomplexeswerecompetedby
theadditionofnone(lane2),5?(lanes3and6),25?(lanes4and7),and125?(lanes5and8)molarexcess
of unlabeled G-wt (lanes 3 to 5) or G-mut (lanes 6 to 8). Proteins responsible for the binding complexes are
indicatedschematicallyontheright.Cross-hatchingindicatesbHLHdomain.GrayboxindicatesGSTprotein.
FP, free probe; (*) nonspecific binding complex.
DNA-
motifs
(45)
of
R E P O R T S
5 MAY 2000VOL 288SCIENCE www.sciencemag.org
860
on February 3, 2009
www.sciencemag.org
Downloaded from

Page 4

tion was rapidly reversed by reconversion to Pr.
These data indicate that phyB recognition of
DNA-bound PIF3 requires maintenance of the
photoreceptor in the biologically active (Pfr)
form.
The G-box motif is neither present in all
light-regulated promoters, nor is it confined to
light-regulated genes. On the contrary, it is
found in a broad range of plant gene promoters
responsive to a diversity of nonlight-related
stimuli (3, 22–26, 29, 30). Moreover, most
studies aimed at identifying plant DNA-binding
proteins that recognize this motif report the
cloning of bZIP class factors rather than bHLH
proteins (22, 30). To address this apparent com-
plexity in relation to PIF3, we examined wheth-
erPIF3wascapableofrecognizingtheG-boxes
in photoresponsive genes in the context of their
native flanking sequences. This is pertinent be-
causethenucleotidesflankingthecorehexamer
E-box motif have been shown to influence the
specificity of bHLH family-member recogni-
tion of binding sites containing this core motif
(31, 32). Figure 3B shows that the G-box–
containing sequences from the promoters of
four light-regulated genes, RBCS-1A, CCA1,
LHY, and SPA1, all interacted effectively with
PIF3, despite deviations from the consensus
sequence of the PIF3 binding site (Fig. 1A) at
thepositionsflankingtheCACGTGhexanucle-
otide core in some cases (Fig. 3A). To deter-
mine whether PIF3 might recognize non–G-
box motifs in other functionally defined LREs
in photoresponsive genes, we examined PIF3
binding to the GT1, Z, and GATA motifs rep-
resenting consensus sequences from several
light-regulated promoters (33). PIF3 exhibited
no detectable interaction with these motifs (Fig.
3C), further verifying the sequence-specific na-
ture of the G-box recognition. Together the data
indicate that PIF3 is indeed capable of se-
quence-specific binding to the G-box–contain-
ing promoters of a variety of light-regulated
genes.
To determine whether PIF3 is necessary for
the phytochrome-regulated expression of these
genes, we examined the effect of continuous R
light (Rc) on their mRNA levels in wild-type
and PIF3-antisense (17) seedlings. The rapid
(within 1 hour) Rc-induced increase in expres-
sion of CCA1 and LHY was reduced in the
PIF3-antisense seedlings (Fig. 4, A and B). By
contrast, the similarly rapid increase in SPA1
expression was unaffected in the antisense
plants. Two more slowly induced genes also
showed no difference in expression between
wild-type and antisense plants. These were the
G-box–containing gene RBCS-1A (34) and
CHS (Fig. 4, A and B) for which there is no
evidence of a functionally active, fully palin-
dromic G-box in Arabidopsis (22, 35, 36). On
the other hand, the absence of the HY5 bZIP
protein in the hy5 null mutant (37) caused no
reduction in the photoresponsiveness of CCA1,
LHY, or SPA1, but markedly reduced the induc-
tion of CHS (Fig. 4, C and D). Together these
data suggest that there are multiple classes of
promoters in phytochrome-responsive genes:
G-box–containing promoters that require PIF3
for responsiveness; G-box–containing promot-
ers that do not require PIF3 for responsiveness,
despite their capacity to bind PIF3 in vitro; and
promoters lacking evidence of functionally ac-
tive G-boxes that do not require PIF3 for re-
sponsiveness, but nevertheless do require the
bZIP factor HY5, considered to be a G-box–
binding protein, for responsiveness.
On the basis of this pattern of expression
profiles, we suggest that a subclass of rapidly
induced genes, represented by CCA1 and LHY,
may be direct targets of phytochrome regula-
tion through binding of the photoreceptor to
PIF3,whichisinturnboundtoG-boxpromoter
elements. Other subclasses of phytochrome-re-
sponsive genes apparently have alternative re-
sponse pathways independent of PIF3. It is
intriguing that CCA1 and LHY encode similar
MYB-class proteins that have been implicated
inphytochrome-regulatedCABgeneexpression
and/or circadian clock regulation (38–40). It is
possible, therefore, that PIF3 represents the en-
try point for phytochrome regulation of the
plant circadian clock, as well as initiating one
branch of the phytochrome-induced gene ex-
pression cascade (41).
The data presented here and elsewhere (17,
20, 21, 28) suggest that the phytochromes may
integrate into, and function as photoswitchable
components of, transcription-regulator com-
plexes directly at target promoter sites after
light-induced translocation from cytoplasm to
nucleus (Fig. 5). The function of PIF3 in this
scheme would be to recruit phyB specifically to
the designated promoters. Regardless of the
biochemical basis of the ensuing signaling
transactions between phyB and the transcrip-
tional machinery, the data suggest that plants
have evolved a mechanism whereby an extra-
cellular signal can be monitored continuously
Fig. 2. PIF3 simulta-
neously binds G-box
DNA and the active
form of phyB (PfrB).
(A) Design of experi-
ments in (B). PHYB re-
fers to full-length phy-
tochrome B apopro-
tein. phyB refers to
photoactive
chrome B, after chro-
mophore attachment
to PHYB, depicted by
the small black rect-
angle (42). After coin-
cubation of proteins
withlabeled
probe, the
were given a pulse of
FR or R (46) and incu-
bated on ice in the
dark (Dk) for 2 addi-
tional hours
EMSA. (B) The binding
complex formed be-
tween PIF3 and the
G-wt probe is shifted
in the presence of R-
irradiated photoactive
phyB, and this super-
shifted complex is de-
pendent on full-length
PIF3. Lane 1, no pro-
tein; lane 2, mock-
translated TnT; lanes
3, 7, and 11, 2 ?l of
PHYB; lanes 4, 5, 8, 9,
12, and 13, 2 ?l of phyB; lanes 6 to 9, PIF3; lanes 10 to 13, G:bhPIF3. (C) Experimental design for
(D). Either phyB alone (lanes 3 and 4) or PIF3 and phyB together (lanes 5 to 12) were incubated for
3 hours in the dark (Dk) after being given a R and/or FR pulse, before G-box probe addition and
EMSA. After an initial R pulse (R) (lanes 3 and 5 to 8) a FR pulse was given either immediately [R ?
FR(0)], after 1 hour [R ? FR(1)], or 2 hours [R ? FR(2)] (lanes 6, 7, and 8, respectively). Conversely,
after an initial FR pulse (FR) (lanes 4 and 9 to 12), a R pulse was given immediately [FR ? R(0)],
after 1 hour [FR ? R(1)], or 2 hours [FR ? R(2)] (lanes 10, 11, and 12, respectively). (D) The
R-induced shift in the PIF3:G-wt probe binding complex caused by the presence of phyB is
photoreversible. Lane 1, mock-translated TnT; lanes 2 and 5 to 12, PIF3; lanes 3 to 12, 2 ?l of phyB.
Proteins responsible for the binding complexes are indicated to the right. FP, free probe; (*)
nonspecific binding complex; PfrB, biologically active form of phyB, formed by R pulse.
phyto-
G-wt
samples
before
R E P O R T S
www.sciencemag.org SCIENCEVOL 2885 MAY 2000
861
on February 3, 2009
www.sciencemag.org
Downloaded from

Page 5

and directly by the control elements of target
genes, thereby potentially permitting almost in-
stantaneous modulation of transcription rates in
response to changes in signal content.
References and Notes
1. R. E. Kendrick and G. H. M. Kronenberg, Photomor-
phogenesis in Plants (Kluwer, Dordrecht, Netherlands,
ed. 2, 1994).
2. P. H. Quail, Plant Cell Environ. 20, 657 (1997).
3. W. B. Terzaghi and A. R. Cashmore, Annu. Rev. Plant
Physiol. Plant Mol. Biol 46, 445 (1995).
4. G. C. Whitelam and P. F. Devlin, Plant Cell Environ.
20, 752 (1997).
5. P. H. Quail, Philos. Trans. R. Soc. London 353, 1399
(1998).
6. G. C. Whitelam, S. Patel, P. F. Devlin, Philos. Trans. R.
Soc. London Ser. B 353, 1445 (1998).
7. X.-W. Deng and P. H. Quail, Semin. Cell Dev. Biol. 10,
121 (1999).
8. M.Ahmad,andA.R.Cashmore,PlantJ.10,1103(1996).
9. G. C. Whitelam et al., Plant Cell 5, 757 (1993).
10. D. Wagner, U. Hoecker, P. H. Quail, Plant Cell 9, 731
(1997).
11. U. Hoecker, Y. Xu, P. H. Quail, Plant Cell 10, 19 (1998).
12. M. S. Soh, S. H. Hong, H. Hanzawa, M. Furuya, H. G.
Nam, Plant J. 16, 411 (1998).
13. M. Hudson, C. Ringli, M. T. Boylan, P. H. Quail, Genes
Dev. 13, 2017 (1999).
14. U. Hoecker, J. M. Tepperman, P. H. Quail, Science 284,
496 (1999).
15. C. Fankhauser et al., Science 284, 1539 (1999).
16. G. Choi et al., Nature 401, 610 (1999).
17. M. Ni, J. M. Tepperman, P. H. Quail, Cell 95, 657 (1998).
18. W. R. Atchley and W. M. Fitch, Proc. Natl. Acad. Sci.
U.S.A. 94, 5172 (1997).
19. T. Littlewood and G. I. Evan, Helix-Loop-Helix Tran-
scription Factors (Oxford Univ. Press, New York, 1998).
20. R. Yamaguchi, M. Nakamura, N. Mochizuki, S. A. Kay,
A. Nagatani, J. Cell Biol. 145, 437 (1999)
21. S. Kircher et al., Plant Cell 11, 1445 (1999).
22. E. Menkens, U. Schindler, A. R. Cashmore, Trends
Biochem. Sci. 20, 506 (1995).
23. G. Arg?ello-Astorga and L. Herrera-Estrella, Annu.
Rev. Plant Physiol. Plant Mol. Biol. 49, 525 (1998).
24. F. Ishige, M. Takaichi, R. Foster, N.-H. Chua, K. Oeda,
Plant J. 18, 443 (1999).
25. R. G. K. Donald and A. R. Cashmore, EMBO J. 9, 1717
(1990).
26. S. Chattopadhyay, L.-H. Ang, P. Puente, X.-W. Deng,
N. Wei, Plant Cell 10, 673 (1998).
27. G. Patikoglou and S. K. Burley, Annu. Rev. Biophys.
Biomol. Struct. 26, 289 (1997).
28. M. Ni, J. M. Tepperman, P. H. Quail, Nature 400, 781
(1999).
29. M. E. Williams, R. Foster, N.-H. Chua, Plant Cell 4, 485
(1992).
30. X. Niu, L. Renshaw-Gegg, L. Miller, M. J. Guiltinan,
Plant Mol. Biol. 41, 1 (1999).
31. F. Fisher and C. R. Goding, EMBO J. 11, 4103 (1992).
32. T. Shimizu et al., EMBO J. 16, 4689 (1997).
33. P. Puente, N. Wei, X.-W. Deng, EMBO J. 15, 3732
(1996).
34. J. F. Martı ´nez-Garcı ´a and P. H. Quail, unpublished data.
35. E. Scha ¨fer, T. Kunkel, H. Frohnmeyer, Plant Cell Envi-
ron. 20, 722 (1997).
36. L.-H. Ang et al., Mol. Cell 1, 213 (1998).
37. T. Oyama, Y. Shimura, K. Okada, Genes Dev. 11, 2983
(1997).
38. Z.-Y. Wang et al., Plant Cell 9, 491 (1997).
39. Z.-Y. Wang and E. M. Tobin, Cell 93, 1207 (1998).
40. R. Schaffer et al., Cell 93, 1219 (1998).
41. D. E. Somers, P. F. Devlin, S. A. Kay, Science 282, 1488
(1998).
42. Binding reactions were performed with GST:PIF3:Flag,
His6:PIF3, G:bhPIF3, and phyB. The PIF3 coding se-
quence was amplified by polymerase chain reaction
(PCR). The Flag peptide sequence was added to the 3?
reverse primer, and the PCR product was cloned into
Eco RI–Not I–digested pGEX-4T-1 vector, to give pGPF.
GST:PIF3:flag protein was produced in Escherichia coli
Fig. 3. PIF3 binds to
G-box–containing pro-
moter sequences from
various light-regulated
genes. (A) Upper-strand
nucleotide sequences of
the probes used in the
EMSAexperimentsin(B)
and(C).Theprobescon-
tain G-box and sur-
rounding
from
RBCS-1A
CCA1 (38, 39), LHY (40),
and SPA1 (14) promot-
ers. The G-box sequence
ishighlightedinboldand
the coordinates are in
parentheses.(B)Compe-
tition of PIF3 binding to
G-wtbythedifferentG-
boxsequencesfromthefourlight-regulatedgenes.Lanes1and6,mock-translatedTnT;lanes2to5,7to16,
PIF3; lanes 3, 8, 11, and 14, 5? molar excess of cold probe; lanes 4, 9, 12, and 15, 25? molar excess of cold
probe; lanes 5, 10, 13, and 16, 125? molar excess of cold probe; lanes 3 to 5, RBCS-1A/G-box cold probe;
lanes8to10,CCA1/G-boxcoldprobe;lanes11to13,LHY/G-boxcoldprobe;lanes14to16,SPA1/G-boxcold
probe.(C)PIF3bindingtolabeledG-wtprobeiscompetedonlybytheLREthatcontainsaG-box(47).Lanes
1 to 9, PIF3; lanes 2, 4, 6, and 8, 25? molar excess of cold probe; lanes 3, 5, 7 and 9, 125? molar excess of
cold probe; lanes 2 and 3, GT1 cold probe; lanes 4 and 5, Z cold probe; lanes 6 and 7, G cold probe; lanes 8
and 9, GATA cold probe (47).
sequences
(25),
Fig. 4. The light-in-
duced expression of
CCA1 and LHY is re-
duced in the PIF3 Ara-
bidopsis antisense line
A22. (A) RNA blot
analysis of CCA1, LHY,
SPA1, and CHS mRNA
levels in the A22 line
(17) and its corre-
sponding
(No-0) in Rc for in-
creasing periods (48).
(B) Quantitative de-
termination of the rel-
ativelevels
transcripts shown in
(A) (49). (C) RNA blot
analysis of CCA1, LHY,
SPA1, and CHS mRNA
levels in a hy5 mutant
line [hy5-1 allele (37)]
and its corresponding
wild-type (Ler) (48).
(D) Quantitative determination of the relative levels of the transcripts shown in (C) (49).
wild type
of the
Fig. 5. Model depicting the proposed mechanism
of phyB regulation of gene expression. R-induced
conversion of phyB from its cytoplasmically lo-
calized, biologically inactive Pr form (PrB) to its
active Pfr form (PfrB) triggers translocation to the
nucleus (20, 21), where it binds to PIF3 that is
constitutively nuclear (17) and bound as a pre-
sumptive dimer to G-box motifs in target pro-
moters. Bound PfrB then activates (or represses)
transcription either directly, by functioning as
a coregulator in recruiting and/or biochemically
or allosterically modifying components of the
preinitiation complex (PIC) or associated factors
(solid arrowhead), or indirectly, by biochemically
or allosterically modifying the presumptive tran-
scriptional regulatory activity of PIF3, which then
in turn recruits coregulator or PIC components
(open arrowheads). Subsequent reconversion by
FR of bound PfrB to PrB causes rapid dissociation
of the photoreceptor from DNA-bound PIF3, dis-
rupting the enhanced (or reduced) transcriptional
activity of target genes. In the short term, subse-
quent reconversion by R of PrB to PfrB, either
before dissociation or nearby in the nucleoplasm,
would rapidly reestablish the previous enhanced
(or reduced) transcriptional state.
R E P O R T S
5 MAY 2000VOL 288 SCIENCEwww.sciencemag.org
862
on February 3, 2009
www.sciencemag.org
Downloaded from

Page 6

(BL21 cells) transformed with pGPF, and the fusion
protein was purified by using sequentially both Gluta-
thione Sepharose 4B (Pharmacia Biotech) and anti-Flag
M2affinitygel(EastmanKodak)beads.Thisproteinwas
used only for RBSS experiments (Fig. 1A). His6:PIF3,
G:bhPIF3, and phyB proteins were produced by tran-
scription and translation (TnT) systems (Promega). PIF3
coding sequence was amplified by PCR and cloned into
pRSETb vector (Invitrogen). The resulting fusion protein
contains a His6-tag at the NH2-terminal end and was
used throughout the work (referred to as PIF3 in the
figures). G:bhPIF3 corresponds to GST (glutathione S-
transferase) (Fig. 1D, gray box) fused to the PIF3 bHLH
domain (from residue 340 to 397; cross-hatched box).
The PIF3 bHLH domain was amplified by PCR and
cloned in pGEX-4T-1. The resulting coding region was
amplified with oligonucleotides that added the T7 pro-
moter sequence upstream of the first ATG, and the PCR
product was directly used as a template in the TnT
reaction. The full-length Arabidopsis PHYB apoprotein
was produced by TnT reaction, and the chromophore
was autocatalytically attached (28).
43. The binding reactions were performed essentially as
described [ J. F. Martı ´nez-Garcı ´a and P. H. Quail, Plant
J. 18, 173 (1999)] with modifications. The binding
buffer was supplemented with 10% glycerol and
0.05% NP-40, and nonspecific competitor used was
50 ng of poly(dI-dC). poly(dI-dC) per reaction. The
binding complexes were resolved by EMSAs in 4%
polyacrylamide gel in 0.5? tris-borate EDTA buffer at
room temperature (90 min at 10 V cm?1), and the
gels were dried and autoradiographed.
44. F. R. Canto ´n and P. H. Quail, Plant Physiol. 121, 1207
(1999).
45. RBSS was performed as described [T. K. Blackwell and H.
Weintraub, Science 250, 1104 (1990)] with modifica-
tions. We synthesized 60-base oligonucleotides of
which the middle 12 bases consisted of random se-
quences (5?-GTCTGTCTGGATCCGAGGTGAGTA-N12-
ACGTCTTCCGAAGCTTACGTCGCG-3?). Two 20-base
oligonucleotides were also synthesized as forward (5?-
GTCTGTCTGGATCCGAGGTG-3?) and reverse (5?-CGC-
GACGTAAGCTTCGGAAG-3?) primers. The stringency
of RBSS was increased by increasing the amount of
nonspecific competitor (50, 100, 200, 400, and 500 ng,
fromfirsttofifthcycles,respectively)andbydecreasing
the amount of protein [2, 2, 1, 1, and 1 ?l of TnT-
expressed PIF3 (42), or 500, 500, 500, 250, and 250 ng
of E. coli–purified GST:PIF3:flag protein (42), from first
to fifth cycle, respectively] and the amount of labeled
DNA probe [90,000 cpm (?106cpm ?g?1) for the first
round; 20,000 cpm of high-specific activity probe from
second to fifth cycle] in the binding reaction (43). After
five rounds of selection, the retarded DNA was ampli-
fied by PCR, digested with Bam HI–Hind III, and cloned
into pBluescript. Individual clones were randomly se-
lected and sequenced. The sequences were aligned cen-
tered around the identified G-box motif.
46. Light sources are described in (28). Pulses were 2 min of
FR (88 ?mol m?2s?1) or R light (88 ?mol m?2s?1).
47. The selected LREs used [GT1, 4? (5?-TGTGTGGTTA-
ATATG-3?); Z, 2? (5?-ATCTATTCGTATACGTGTCAC-
3?); G, 4? (5?-TGACACGTGGCA-3?); and GATA, 4?
(5?-AAGATAAGATT-3?)] have been described else-
where (33).
48. After germination, seedlings were grown in the dark at
22°C for 4 days. Material was harvested at various
times after 0, 1, 2, 3, and 4 hours of exposure to
continuous red light (Rc; 20 ?mol m?2s?1). Total RNA
was isolated with the RNeasy Plant Miniprep kit (Qia-
gen). For RNA analyses, 5 ?g of total RNA were loaded
per lane, and then transferred to MSI Nylon mem-
branes.ThemembraneswerehybridizedinChurchbuff-
er at 65°C overnight with random primer–labeled frag-
ments (CCA1, LHY, CHS, SPA1). CCA1 and LHY probes
were amplified by PCR from Arabidopsis DNA with the
use of specific primers, cloned, and confirmed by se-
quencing. The CCA1 probe (amplified with the primers
5?-GCAGCTGCTAGTGCTTGGTGGGCT-3? and 5?-TCA-
TGTGGAAGCTTGAGTTTCCAA-3?)correspondedtopo-
sitions 2082 to 3010 in the 3? region of the main open
reading frame (ORF) (38, 39); the LHY probe (amplified
with the primers 5?-CATGCTGCAGCTACATTCGCT-
GCT-3? and 5?-TCATGTAGAAGCTTCTCCTTCCAATCG-
3?) corresponded to positions 1271 to 2275 in the 3?
region of the main ORF (40). Southern blot analysis
showed no detectable cross-hybridization between
CCA1 and LHY probes under the washing conditions
used (34). The SPA1 (14) and CHS [R. L. Feinbaum and
F. M. Ausubel, Mol. Cell. Biol. 8, 1985 (1988)] probes
have been described elsewhere.
49. Relative levels of transcripts were normalized to 18S
ribosomal RNA levels (44) after PhosphorImager
Storm 860 (Molecular Dynamics) quantification.
50. We thank Y. Kang for technical assistance; M. Ni
for A22 seeds and original PIF3 clones; C. Fair-
child for the phycocyanobilin; U. Hoecker for SPA1
cDNA; N. Wei for the CHS probe; E. Monte and M.
Rodrı ´guez-Concepcio ´n for support and discus-
sion; all the lab members for discussion and sup-
port; and the Arabidopsis Biological Resource Cen-
ter (Columbus, Ohio) for providing hy5 (hy5-1
allele) seeds. Supported by grants from the U.S.
Department of Energy Basic Energy Sciences (DE-
FG03-87ER13742) and U.S. Department of Agricul-
ture Current Research Information Service (5335-
21000-010-00D).
13 December 1999; accepted 25 February 2000
Template Boundary in a Yeast
Telomerase Specified by RNA
Structure
Yehuda Tzfati, Tracy B. Fulton, Jagoree Roy,
Elizabeth H. Blackburn*
The telomerase ribonucleoprotein has a phylogenetically divergent RNA sub-
unit, which contains a short template for telomeric DNA synthesis. To under-
stand how telomerase RNA participates in mechanistic aspects of telomere
synthesis,westudiedaconservedsecondarystructureadjacenttothetemplate.
Disruption of this structure caused DNA synthesis to proceed beyond the
normal template boundary, resulting in altered telomere sequences, telomere
shortening, and cellular growth defects. Compensatory mutations restored
normal telomerase function. Thus, the RNA structure, rather than its sequence,
specifies the template boundary. This study reveals a specific function for an
RNA structure in the enzymatic action of telomerase.
Telomerase, a ribonucleoprotein reverse tran-
scriptase (RT), replenishes telomeric DNA that
would otherwise be lost with each round of
eukaryoticDNAreplication(1).Thetelomerase
complex contains an RNA subunit (TER), a
catalytic RT protein (TERT), and several addi-
tional protein components (2). Telomerase is
activated in most human cancers, and its ectop-
ic expression can greatly extend the life-span of
normal human cells in culture (3).
Telomerase RNAs are extremely divergent
in sequence and vary in length from 146 nucle-
otides(nt)intheciliateTetrahymenaparavorax
(4) to 1544 nt in the budding yeast Candida
albicans (5). Unlike other RTs, which perform
extensive genome copying, telomerase copies
only a small portion (termed the “template”) of
an intrinsic RNA moiety (6). This feature al-
lows telomerase to synthesize onto telomeres a
species-specific, 5- to 26-base-long repeated
sequence(7).Howtelomerasespecifiesitstem-
plate boundaries (where DNA synthesis ini-
tiates and where it ends on the TER sequence)
is not understood.
Nontemplate regions have been previously
shown to be required for telomerase activity (8,
9) and ribonucleoprotein (RNP) assembly (9,
10). To further investigate the participation of
telomerase RNA in the enzymatic function of
telomerase, we searched for conserved se-
quences and structural elements in budding
yeast telomerase RNAs. We cloned and ana-
lyzed TER genes from four Kluyveromyces
species closely related to K. lactis (11). The
mature RNAs ranged in length from 930 nt in
K. aestuarii to 1320 nt in K. dobzhanskii. Se-
quence identity between any given pair of
genes ranged from insignificant to about 70%
overall identity. The computer program mfold
(12) predicted extensive secondary structures
for these RNA sequences, including a common
feature shared by all five TERs: base pairing of
the sequence immediately upstream of the tem-
plate (pairing element B) (Fig. 1A) with a
sequence 200 to 350 nt further upstream (pair-
ing element A), located near the 5? end of the
RNA. The region between the pairing elements
(indicated by the dashed line in Fig. 1A and the
dashed loop in Fig. 1B) was shown previously
to be dispensable in K. lactis (9). The proximity
of this conserved putative pairing region to
the 5? end of the template led us to hypoth-
esize that its function is to limit DNA syn-
thesis, thereby defining the downstream
boundary of the template.
To test this hypothesis, we constructed a
series of mutations in the putative pairing re-
gion of the K. lactis TER gene (Fig. 1, C and
D). We replaced the wild-type TER gene in K.
lactis with the mutant genes by a vector-shuf-
Department of Microbiology and Immunology, Uni-
versity of California, San Francisco, San Francisco, CA
94143–0414, USA.
*To whom correspondence should be addressed. E-
mail: telomer@itsa.ucsf.edu
R E P O R T S
www.sciencemag.org SCIENCE VOL 2885 MAY 2000
863
on February 3, 2009
www.sciencemag.org
Downloaded from