Physical interaction between VIVID and white collar complex regulates photoadaptation in Neurospora.

Abstract

Photoadaptation, the ability to attenuate a light response on prolonged light exposure while remaining sensitive to escalating changes in light intensity, is essential for organisms to decipher time information appropriately, yet the underlying molecular mechanisms are poorly understood. In Neurospora crassa, VIVID (VVD), a small LOV domain containing blue-light photoreceptor protein, affects photoadaptation for most if not all light-responsive genes. We report that there is a physical interaction between VVD and the white collar complex (WCC), the primary blue-light photoreceptor and the transcription factor complex that initiates light-regulated transcriptional responses in Neurospora. Using two previously characterized VVD mutants, we show that the level of interaction is correlated with the level of WCC repression in constant light and that even light-insensitive VVD is sufficient partly to regulate photoadaptation in vivo. We provide evidence that a functional GFP-VVD fusion protein accumulates in the nucleus on light induction but that nuclear localization of VVD does not require light. Constitutively expressed VVD alone is sufficient to change the dynamics of photoadaptation. Thus, our results demonstrate a direct molecular connection between two of the most essential light signaling components in Neurospora, VVD and WCC, illuminating a previously uncharacterized process for light-sensitive eukaryotic cells.

Full text

Page 1

Physical interaction between VIVID and white collar
complex regulates photoadaptation in Neurospora
Chen-Hui Chena, Bradley S. DeMayb, Amy S. Gladfelterb, Jay C. Dunlapa,1, and Jennifer J. Lorosc,1
Departments ofaGenetics andcBiochemistry, Dartmouth Medical School, Hanover, NH 03755; andbDepartment of Biological Sciences, Dartmouth College,
Hanover, NH 03755
Contributed by Jay C. Dunlap, July 29, 2010 (sent for review July 7, 2010)
Photoadaptation, the ability to attenuate a light response on
prolonged light exposure while remaining sensitive to escalating
changes in light intensity, is essential for organisms to decipher time
informationappropriately,yettheunderlyingmolecularmechanisms
arepoorlyunderstood.InNeurosporacrassa,VIVID(VVD),asmallLOV
domain containing blue-light photoreceptor protein, affects photo-
adaptation for most if not all light-responsive genes. We report that
thereisaphysicalinteractionbetweenVVDandthewhitecollarcom-
plex (WCC), the primary blue-light photoreceptor and the transcrip-
tion factor complex that initiates light-regulated transcriptional
responses in Neurospora. Using two previously characterized VVD
mutants, we show that the level of interaction is correlated with
the level of WCC repression in constant light and that even light-
insensitive VVD is sufficient partly to regulate photoadaptation in
vivo. We provide evidence that a functional GFP-VVD fusion protein
accumulates in the nucleus on light induction but that nuclear locali-
zation of VVD does not require light. Constitutively expressed
VVD alone is sufficient to change the dynamics of photoadaptation.
Thus,ourresultsdemonstrateadirectmolecularconnectionbetween
two of the most essential light signaling components in Neurospora,
VVD and WCC, illuminating a previously uncharacterized process
for light-sensitive eukaryotic cells.
light|white collar-1|white collar-2|photoreceptor|LOV domain
L
presence or absence of light but be able to detectsubtle changesin
theintensity of light to elicit proper behavioral and developmental
responses. Thus, cells have evolved mechanisms to adapt to pro-
longed light exposure while retaining sensitivity to escalating
changes in light intensity at the same time. Light responses in the
filamentous fungus Neurospora crassa have been studied for sev-
eral decades because of their robustness and easily scored phe-
notypes(1–6).TheWHITECOLLAR(WC)-1andWC-2proteins
control most light responses in Neurospora (7, 8). WC-1, mainly
localized in the nucleus (9–12), is both a FAD-binding photore-
ceptor and a GATA zinc finger transcription factor (13, 14).
Through their PER-ARNT-SIM (PAS) domains (15–17), WC-1
and WC-2, another GATA zinc finger transcription factor, form
thewhitecollarcomplex(WCC).Afterexposuretolight,theWCC
binds to the promoter of early light-responsive targets at defined
consensus sequences and activates transcription within 15 min or
less (14, 18). After prolonged light exposure in Neurospora, in-
duction of both early (average peak in steady-state RNA ∼30 min
afterexposuretolight)andlate(averagepeakinsteady-stateRNA
∼60 min after exposure to light) light responses exhibit an un-
ambiguous feature of photoadaptation, returning to their basal
level of expression after 1–4 h in constant light (7). To remain
sensitive to an additional increase in ambient light intensity,
Neurospora is able to adapt to various levels of light intensity (9,
19–21).
VIVID (VVD) is an additional blue-light photoreceptor that
functions downstream of the WCC to regulate negatively the light
responses initiated by the WCC (7, 9, 19–22). Loss-of-function
vvd mutants were originally identified by the unique appearance
ight is a strong environmental cue for many organisms across
different kingdoms. Organisms must not only be aware of the
of the bright orange conidia while grown in constant light (23),
presumably attributable to the persistent activation of carotenoid
biosynthesis(7).Inadditiontoitsfunctioninattenuatingthedirect
light response in Neurospora, VVD has been shown to take part in
regulating various circadian clock properties, most likely through
its effects on the WCC, including gating of light input to the clock
(21), maintenance of the clock during the light phase (24, 25), and
temperature compensation of the circadian phase (26). Despite
accumulating genetic and molecular evidence supporting the
connection between VVD and WCC, no mechanistic explanation
for the relationship has been proposed.
VVD is a small 21-kDa FAD-binding protein consisting of
a LOV domain and an N-terminal cap (21, 27), capable of un-
dergoing a photocycle in vitro (9, 27). Light induction of vvd gene
expression is robust and fast, with the peak in RNA expression
around 15 min and that in protein around 30 min after exposure to
light, and it is WCC-dependent (7, 9, 21). Once activated by blue
light, the formation of a covalent cysteinyl-flavin adduct within the
LOV domain induces an N-terminal conformational change
thoughttobeimportantforrepressorfunction:Apointmutationin
thehingeregionrenderstheproteinunabletochangeconformation
and causes excessive accumulation of carotenoid in constant light
(27). In light, VVD forms a rapidly exchanging dimer in solution,
suggesting the possibility that the LOV domain of VVD could in-
teract with other PAS domains, including those found in the WCC
proteins (28–30). Cell fractionation studies have shown VVD to be
exclusively localized in the cytosol (9), whereas most of the WCC is
nuclear (9–12), raising the question of how VVD communicates
with the WCC to affect various WCC-mediated activities.
Here,using the proteincross-linker 3,3′-dithiodipropionic aciddi
(N-hydroxysuccinimide ester) (DSP), we show that the WCC coim-
munoprecipitates with VVD under both low- and high-light con-
ditions and demonstrate that the level of interaction between VVD
and WCC is correlated with the level of WCC repression in con-
stant light. Unexpectedly, two previously characterized vvd mutant
alleles, vvdC71Sand vvdC108A, are not null for photoadaptation but
are partially functional in repressing light responses and, moreover,
aresufficienttoregulatereinitiationofthelightresponseinconstant
light with an increase in light intensity. In contrast to previous bio-
chemicalevidence,weshowbylive-cellmicroscopythatafunctional
GFP-VVD protein appears in the cytosol on light induction but
additionally accumulates in the nucleus, where most of the WCC
resides.However,nuclearlocalizationofVVDisindependentofthe
presence of light, as demonstrated by a constitutively expressed
GFP-VVD. Manipulating the dynamics of VVD expression offers
considerable support for thehypothesisthat thefunction ofVVD is
Author contributions: C.-H.C., A.S.G., B.S.D., J.C.D., and J.J.L. designed research; C.-H.C.
and B.S.D. performed research; C.-H.C., B.S.D., A.S.G., J.C.D., and J.J.L. analyzed data; and
C.-H.C., A.S.G., J.C.D., and J.J.L. wrote the paper.
The authors declare no conflict of interest.
1To whom correspondence may be addressed. E-mail: jay.c.dunlap@dartmouth.edu or
jennifer.loros@dartmouth.edu.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1011190107/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1011190107PNAS
| September 21, 2010
| vol. 107
| no. 38
| 16715–16720
PHYSIOLOGY

Page 2

WCC-dependent and that VVD protein expression alone is suffi-
cient to repress light responses in vivo.
Results
Physical Interaction Between VVD-V5 and WCC Revealed by Coim-
munoprecipitation After DSP Treatment. We hypothesized that a
central mechanism of photoadaptation could involve association
of VVD with other proteins, and therefore made strains with
atagged VVDtofacilitatebiochemicalapproaches.InFig.S1,our
data demonstrate that the ectopically expressed VVD appears
fully functional and the C-terminal V5 tag is not interfering with
biological functions of VVD in vivo. Several attempts using large-
scale preparations of purified VVD-V5 to identify the potential
interacting partners of VVD were not successful. Therefore, we
predicted that if VVD does interact with another protein to re-
press WCC activity, the interaction might be either weak or tran-
sient. Both VVD and WC-1 contain LOV domains, a variant PAS
domain with flavin-binding capabilities. WC-1 and WC-2 are PAS
domain-containing transcription factors, and formation of the
WCC heterodimer is mediated mainly through their shared PAS
domains (15–17). Given the similarity between LOV and PAS
domains, we hypothesized that VVD might be able to interact
directly with the WCC to repress its transcriptional activity and
that application of protein cross-linkers might be able to reveal
their relationship. We performed a coimmunoprecipitation (co-
IP) assay, with DSP treatment before cell lysis. DSP, a reversible
protein cross-linker, is widely used to “capture” weak or transient
protein-protein interactions (31–34). After the DSP treatment,
anti-V5 antibody-coated agarose was used to isolate VVD-V5 and
the coprecipitates were hybridized with either anti–WC-1 or anti–
WC-2antibodiestolookforinteraction.AsshowninFig.1A,when
we performed the co-IP assay following a dark-to-light transfer,
either with or without DSP treatment, the level of VVD-V5 in the
dark was too low to be detected even in the input lane. Following
exposure to light, VVD expression reached an intermediate level
15 min after exposure to white light (LL15) and peaked at LL30.
With DSP treatment, the amount of coimmunoprecipitated WC-1
and WC-2 appears to be proportional to the amount of VVD-V5
isolatedinthesolution,therebysubstantiatingthespecificityofthe
interaction. Notably, weak bands of WC-1 and WC-2, clearly ab-
sent in the wc-1, wc-2 double-knockout strain, can be seen on the
gel from a WT (no tag) strain in darkness (DD), indicating that
there is a low level of nonspecific binding between WCC and the
agarose used in the co-IP assay. Co-IP assays in the absence of
cross-linking with DSP treatment show a very weak but real in-
teraction that is above background (Fig. 1A). Next, we determined
what fraction of total WCC is bound to VVD-V5. As seen in Fig.
1A, a co-IP assay with anti-V5 antibody-coated agarose isolated
most of the VVD-V5 from the total input. In contrast, ≈3% of
WC-1 and WC-2 in the lysate coimmunoprecipitated with VVD-
V5, as determined by comparing a two-factor serial dilution of
total input (Fig. 1 B and C) to the amount in the 1× co-IP lane. To
confirm the above findings in the photoadaptation process, we
performed co-IP assays following a very low light (VLL, 3
μmol·m2·s)-to-LL (20 μmol·m2·s) transfer. Regardless of the
changes in light intensity, VVD-V5 retained the interaction with
theWCC, indicating a steady interaction between VVD and WCC
in the process of photoadaptation (Fig. 1D).
Level of InteractionBetween VVD andWCC Is Correlatedwith theLevel
of WCC Repression in Constant Light. To determine if the physical
interaction between VVD and the WCC is required for functional
photoadaptation, we asked whether two established loss-of-
function vvd mutants, vvd-v5C71Sand vvd-v5C108A, still retained
their capability to interact with the WCC. Previous studies have
shown that Cys71 is essential in controlling movements of the
N-terminal cap after light activation (27) and that Cys108 within
the LOV domain forms a cysteinyl-flavin adduct in response to
blue light and is required for VVD to undergo a complete pho-
tocycle in vitro (9). C71S and C108A amino acid changes in VVD
caused a vvd-null phenotype as assayed by coloration in constant
light (9, 27). After confirming the mutated sequences (Fig. 2A)
and transforming the mutated DNA into the Δvvd strain, we
found that neither allele could complement the photoadaptation
defect of Δvvd as judged from carotenoid levels, despite the ro-
bust light induction of their protein levels (Fig. 2B). Consistent
with the loss-of-function phenotype, the co-IP data indicate that
both mutant proteins display a considerable decrease in their ca-
pability to interact with the WCC relative to the WT VVD (Fig. 2
C–E), suggesting that their interaction is important for function
and that maximal interaction depends on a fully functional VVD.
In addition, the specificity of the interaction is further supported
by Western blot analysis with either anti-PKC or antitubulin
antibodies (Fig. 2C). In the presence of DSP, PKC is not pulled
down in the VVD-V5 co-IP assays in light or dark. Tubulin dis-
plays a consistent band, indicating that it binds to the agarose
beads independent of VVD-V5 and light conditions.
Notably, both vvd-v5C71Sand vvd- v5C108Amutant proteins still
retain a reduced but significant level of WCC interaction (Fig. 2 D
and E). If the physical interaction between VVD and WCC is in-
deed essential for regulating photoadaptation, these two mutant
alleles might not be complete loss-of-function alleles, as suggested
by their inability to complement repression of carotenogenesis.
Therefore, we examined their photoadaptation defects at the mo-
lecular level to determine how the level of interaction correlated
with the ability to repress WCC-dependent light responses in vivo.
The expression of the early light response genes al-3 and sub-1 was
followed by quantitative PCR at time 0 and LL60 in both mutants
(Fig. 2 F and G). Consistent with our hypothesis that VVD-WCC
interaction is necessary for photoadaptation, we found that both
vvd-v5C71Sand vvd- v5C108Amutant proteins were partially func-
tional in their ability torepress light responses (Fig. 2 F and G) and
carotenoid induction (Fig. S2 A and B). We conclude that both
transgenic strains displayed mutant photoadaptation color pheno-
types resulting from excessive accumulation of carotenoid in con-
Fig. 1.
DSP treatment. (A) Anti-V5 co-IP assay with or without DSP treatment follow-
ing a time course up to 2 h. (B) Relative amount of WC-1 and WC-2 interacting
with VVD as determined by the anti-V5 co-IP assay. (C) Densitometric meas-
urements of WC-1 and WC-2 coimmunoprecipitated with the VVD-V5 (n = 5).
The data were quantified by comparison with the diluted inputs with the
closest image intensity, as shown in B. (D) Anti-V5 co-IP assay with DSP treat-
ment following a verylowlight(VLL,3 μmol·m2·s)toLL(20 μmol·m2·s)transfer.
NS,nonspecificbandontheWC-1blotasaloadingcontrol.AsterisksinAandB
highlight a nonspecific band close in size to the WC-2 signal. In all figures, LL
indicates a constant white light stimulus with a photon flux of 20 μmol·m2·s.
Physical interaction between VVD-V5 and WCC revealed by co-IP after
16716
| www.pnas.org/cgi/doi/10.1073/pnas.1011190107Chen et al.

Page 3

stant light over time (Fig. 2B). Intriguingly, the level of interaction
between VVD-V5 and WCC is inversely correlated with the func-
tional output of the WCC at LL60 (Fig. 2H), further supporting
the hypothesis that the physical interaction between VVD and
WCC is specific and determines the level of WCC repression.
Light-Insensitive VVD Is Sufficient to Regulate Photoadaptation in
Vivo. An important feature of photoadaptation is the sensitivity to
an increase in light intensity. Although mutant proteins encoded
by both vvd-v5C71Sand vvd-v5C108Awere functional at LL60 to re-
press the primary light response at least partially, we expected that
the ability to respond to increased light might require the VVD
protein to act as a photoreceptor and directly sense changes of
lightintensityindependentoftheWC-1photoreceptor.Therefore,
we asked whether the vvd mutants were capable of responding to
a light increase. Surprisingly, both vvd mutant proteins were still
functional, although less so than WT, in regulating photo-
adaptationinvivo,asindicatedbyfollowingthereactivationofal-3
and sub-1 expression (Fig. 3) after a low light-to-LL transfer. This
suggests that photoadaptation is regulated by VVD protein acting
as a negative feedback element on the WCC transcriptional ac-
tivity, with the positive element (WCC) but not VVD sensitive to
afurtherincreaseintheamountoflight.Intheabsenceofvvd(Fig.
3), there is no reactivation of light-induced gene expression, sup-
portingtheabsoluterequirementforNeurosporaVVDininitiating
gene expression in response to a change in light intensity.
In the absence of vvd, WC-1 becomes hyperphosphorylated in
constant light(19,21), which correlateswith theup-regulated light
responses.AlthoughtheWCC isinactivatedby phosphorylationin
response to light (18), the levels of WC-1 phosphorylation may
serve as an indicator for the functional output of the WCC. For
instance, no noticeable hyperphosphorylated WC-1 was observed
at LL60 in either the vvd-v5 or WT strain (Fig. S3), reflecting re-
pressed WCC activity in constant light. On the other hand, con-
sistent with the observation that both vvd-v5C71Sand vvd-v5C108A
are partially functional (Fig. 2 F and G), WC-1 phosphorylation at
LL60inthesetwostrainsappearedintermediatebetweentheΔvvd
andWTstrains(Fig.S3),supportingtheircapabilitytoaffectWCC
activity and repress light response.
GFP-VVD Is Functional in Vivo When Driven by Either the vvd or sod-1
Promoter. Thus far, the data support a functional and physical in-
teraction between VVD and WCC that is important for photo-
adaptation,sowenextdeterminedwhenandwherethisinteraction
may occur by localization of VVD in living cells. In previous
experiments using cell fractionation approaches to separate nuclei
from cytoplasm, VVD was seen exclusively in the cytosolic frac-
tions (9), whereas most of the WCC remained in the nucleus (9–
12). However, given that the size of VVD (21 kDa) is well below
the cutoff for possible passive diffusion through the nuclear pore,
whichappearstobe60kDaorabove(35–37),wehypothesizedthat
loss of VVD from the nucleus might be a technical artifact at-
tributable to the fact that small protein molecules might simply
leak out of the nucleus during the purification process.
To test whether VVD can enter the nucleus and repress WCC
activity, we tagged VVD-V5 with a GFP derivative (GFP-S65T,
SGPF) (38) at its N terminus and integrated the transgene (Fig.
S4A, gfp-vvd-v5), together with 3.5 kbp of its own promoter, into
the csr-1 locus in a vvd knockout strain. Additionally, we sought to
dissociate the regulation of VVD localization from its light acti-
vation.Recentstudies onthephototropinphotoreceptorsinplants
have implicated LOV domains in regulating subcellular localiza-
tion following light activation (39–41). The vvd promoter used to
Fig. 2.
with the level of WCC repression in constant light. (A) Sequence confirma-
tion of the vvd-v5C71Sand vvd-v5C108Aalleles. (B) Carotenoid accumulation as
a measure of photoadaptation defects in the vvd-v5C71Sand vvd-v5C108A
strains. Photoadaptation limits the expression of genes whose products are
involved in the carotenogenesis pathway. (Right) Light induction of the
VVD-V5 protein in either the vvd-v5C71Sor the vvd-v5C108Astrain. (C) Anti-V5
co-IP assay with the vvd-5, vvd-v5C71S, and vvd-v5C108Astrains. PKC and tu-
bulin are controls as described in the text. Densitometric measurements of
WC-1 (D) and WC-2 (E) coimmunoprecipitated with VVD-V5 (n = 3). The data
were normalized to the background signals in DD. Photoadaptation defects
were determined by measuring light induction of al-3 (F) and sub-1 (G) at
LL60 using RT quantitative PCR analysis (n = 3, mean values ± SE). Asterisks
indicate statistical significance when compared with the Δvvd strain at LL60
as determined by an unpaired t test. ***P < 0.001; **P < 0.01; *P < 0.05. (H)
Relative amount of WC-1 coimmunoprecipitated with VVD-V5 (filled squares)
vs. al-3 expression at LL60 (open squares) in respective strains (n = 3). LL
indicates constant white light stimulus with a photon flux of 20 μmol·m2·s.
The level of interaction between VVD and the WCC is correlated
Fig. 3.
vivo. Photoadaptation defects were determined by measuring induction of
al-3 (A) and sub-1 (B) following a very low light (VLL, 3 μmol·m2·s) to LL (20
μmol·m2·s) transfer using RT quantitative PCR (QPCR) analysis (n = 3, mean
values ± SE). Asterisks indicate statistical significance when compared with
the level of expression at VLL240 as determined by an unpaired t test. ***P <
0.001; **P < 0.01. NS, difference between the two groups is not significant.
Light-insensitive VVD is sufficient to regulate photoadaptation in
Chen et al.PNAS
| September 21, 2010
| vol. 107
| no. 38
| 16717
PHYSIOLOGY

Page 4

drive the gfp-vvd-v5 allele was replaced with the sod-1 promoter,
which constitutively drives gene expression independent of the
light conditions (7) (Fig. S4B, gfp-vvd-v5cx). Western blot analysis
with anti-V5 and anti-GFP antibodies confirmed the identity and
expression kinetics of GFP-conjugated VVD-V5 proteins when
driven by either vvd or sod-1 promoters (Fig. S4C). Phenotype
analysis indicates that the ectopically expressed GFP-VVD-V5
driven by the vvd promoter is sufficient to restore the photo-
adaptation defects of Δvvd in vivo, as assayed by coloration (Fig.
S4D). In contrast, the transgenic strain with GFP-VVD-V5 driven
by the sod-1 promoter displayed a partially complemented pheno-
type (Fig. S4D), most likely reflecting the relatively lower level of
GFP-VVD expression in constant light compared with the endog-
enous vvd promoter allele (compare LL60 samples in Fig. S4C).
Moreover, GFP-VVD is not only phenotypically functional but
retainsinteractionwithWC-1andWC-2inconstantlight(Fig.S4E).
Consistent with the complemented phenotype on solid medium
slants, the photoadaptation defects of Δvvd were complemented by
both GFP-VVD alleles at themolecularlevel as well(Fig. S4F and
G).Atthelevelofourassays,theincreasedsize ofVVD(21kDa+
27 kDa of GFP) to 48 kDa does not appear to affect VVD’s in-
teraction with the WCC or its biological functions in vivo, and it
behaves like its WT counterpart in the process of photoadaptation.
GFP-VVDAccumulatesintheNucleusonLightInductioninLiveCellsand
NuclearLocalizationofGFP-VVDDoesNotRequireLightActivation.We
followed the dynamics of GFP-VVD-V5 on light induction in
vegetative hyphae by live-cell imaging. Consistent with the West-
ern blot analysis (Fig. S4C), light treatment robustly induced GFP
signals in the gfp-vvd-v5 strain at LL60 (Fig. 4A). Next, we asked if
the expression of GFP-VVD-V5 colocalized with the nuclear
Hoechst stain after paraformaldehyde fixation. In agreement with
our hypothesis but somewhat surprisingly, given previous work,
GFP-VVD-V5 was clearly concentrated in the nucleus on light
induction (Fig. 4A and Fig. S5A), although GFP-VVD is also seen
in the cytoplasm. Consistent with the Western blot analysis (Fig.
S4C), GFP-VVD under sod-1 regulation is seen in both dark- and
light-grown hyphae and appears to be expressed at a lower level
thanunderitsownpromoterinLL.Unlikephototropinproteinsin
Arabidopsis, nuclear localization of the constitutively expressed
GFP-VVD-V5 protein showed no apparent difference in locali-
zation before and after light exposure (Fig. 4B and Fig. S5B),
thereby suggesting that light-induced conformational changes in
VVD protein are not necessary for the nuclear enrichment of
VVD. A WT strain without a GFP tag was examined in parallel to
control for nonspecific autofluorescent signals in Neurospora (Fig.
4C). Highly expressed GFP alone has been previously shown to be
cytoplasmic in Neurospora (42).
Constitutively Expressed VVD Is Sufficient to Change the Dynamics of
Photoadaptation.Morethan300genesarelight-responsiveinNeu-
rospora (7). To test whether the gene repression aspect of pho-
toadaptation is dependent on light-induced genes in addition to
vvd, we asked whether constitutively expressed VVD (driven by the
sod-1 promoter as shown in Fig. S4C) is sufficient to attenuate
WCC-dependent light responses at an early time point when light-
responsive genes would not yet be fully translated (Fig. 5A). Com-
pared with VVD driven by its own promoter (gfp-vvd-v5), we found
that constitutively expressed VVD (gfp-vvd-v5cx) reduced the level
of carotenoid production in hyphal mats in liquid culture (Fig. 5B
and Fig. S2B), suggesting that VVD expressed in the dark is
repressing WCC-driven carotenogenesis. Dark-expressed VVD ef-
fectivelyrepressedacutelightresponsesatLL15ofthreeearlylight-
responsive genes, al-3, sub-1, and frq (Fig. 5 C–E), suggesting that
expressionof VVD aloneis sufficient to change the dynamics of the
photoadaptation response, independent of other light-responsive
gene products. Notably, although light-insensitive VVD remained
functional in repressing WCC target genes (Fig. 3) inconstantlight,
the constitutively expressed VVD failed to affect the basal expres-
sion of al-3, sub-1, and frq (Fig. 5 C–E) in the dark, indicating that
therepressorfunctionofVVDisrestrictedandspecifictoresponses
mediated by light-activated WCC.
Discussion
In the filamentous fungus Neurospora, photoadaptation can be de-
scribed as a series of molecular events that includes the initial at-
Fig. 4.
and the nuclear localization of GFP-VVD does not require light activation.
(A) Localization of GFP-VVD-V5 before and after light treatment. Mycelia
samples for live-cell imaging were collected before (DD) and after light
treatment for 60 min. For fixed-cell images, the mycelia were fixed in
paraformaldehyde for 7 min and then incubated with Hoechst dye for an-
other 30 min before imaging. (B) Localization of GFP-VVD-V5cx(constitu-
tively expressed) before and after light treatment. (C) WT strain under the
same light conditions. LL indicates constant white light (photon flux of
20 μmol·m2·s). White arrows highlight representative nuclei. (Scale bar:
5 μm.) All images were acquired using identical light exposure settings,
deconvolved, and displayed using identical linear contrast enhancement.
GFP-VVD accumulates in the nucleus on light induction in live cells,
16718
| www.pnas.org/cgi/doi/10.1073/pnas.1011190107Chen et al.

Page 5

tenuationorrepressionofthewaveofgeneexpressioninresponseto
a light stimulus and, if the light remains, a reinitiation of gene ex-
pression following an increase in the light intensity. Considerable
evidence suggests thatVVD regulates thetranscriptionalactivity of
the WCC in mediating photoadaptation in constant light as well as
several other properties associated with the circadian clock under
free-running conditions (7, 21, 24, 26). Despite the genetic re-
lationship between VVD and the WCC, biochemical studies ex-
amining their physical interaction and subcellular localization have
not shown direct interactions or colocalization. Therefore, the
communicationbetweenVVDandWCChasbeenlargelyviewedas
indirect.Throughco-IPanalysisafterproteincross-linkingandlive-
cell imaging of GFP-VVD, we find a specific physical interaction
between VVD and WCC and, additionally, that both are presentin
thenucleusonlightinduction.Thus,ourfindingsofferamechanistic
explanation for photoadaptation in Neurospora and possibly other
light-sensitive eukaryotic cells (Fig. 6).
Inadditiontoseveralsequenceandfunctionalhomologsinother
fungi,theLOV/PASdomainfoundinVVDandWC-1iswidespread
and functionally conserved in different kingdoms (27, 43–48). For
instance,LOVdomainsfromphototropinsandFKF1inArabidopsis
are still able to sense blue light when incorporated in Neurospora
photoreceptor proteins (49). Interestingly, phototropins (PHOT1
and PHOT2) in Arabidopsis are also involved in sensing and
responding to changes of light intensity (50), but their subcellular
localization on light activation is complex. A small fraction of
PHOT1 is released to the cytoplasm from the plasma membrane
(40),whereasPHOT2becomesassociatedwiththeGolgiapparatus
(39, 41). How these light-induced changes in localization affect re-
ceptor signaling is not completely understood (51). Phototropins
may therefore behave in a manner similar to VVD and WCC,
whereby communication is mediated by a weak and possibly tran-
sient physical interaction after light activation.
The role of other components that VVD might use to repress
WCC transcriptional activity remains to be explored. Notably,
a few genes other than vvd have been implicated in regulating
photoadaptation in Neurospora. For instance, previous work has
demonstrated that changes in PKC activity result in altered light
responses in constant light (52, 53). However, we were unable to
establish a physical interaction between PKC and VVD or WCC
(Fig. S6). Additionally, although there is an inverse correlation
between WC-1 phosphorylation and VVD activity in constant
light, our screenings with a large collection of kinase and phos-
phataseknockout strains(Table S1) failedtoidentifyany vivid-like
strains at the phenotypic level. However, using a genetic screen,
Navarro-Sampedro et al. (54)reported several interesting mutants
specifically deficient in regulating photoadaptation of con-10 and
con-6 (i.e., not all light-responsive genes are affected). Further
characterization of these mutants might reveal previously unde-
scribed components involved in mediating photoadaptation in
a gene-specific manner. However, given that most if not all light-
responsive genes are affected in the Δvvd strain and that, to date,
there are no reports of any mutants that display photoadaptation
defectsclosetothoseoftheΔvvd,amorecomprehensivegeneticor
interactome screening may be required to identify any previously
undescribed participating components. Meanwhile, our live-cell
imaging data demonstrating VVD’s presence in the nucleus are
in agreement with the nuclear localization of the WCC and the
interaction results. A study examining GFP-conjugated WC-2 re-
vealedanunexpectednucleocytoplasmic shuttling that occursover
a very rapid time frame (within minutes) (10). Because the study
examined samples in constant conditions, it is unclear how light
might affect the shuttling of WCC and whether the shuttling plays
any roles in regulating light responses. Taken together, our data
reveal a direct molecular connection between two of the most es-
sential light signaling components in Neurospora and provide an
important perspective on photoadaptation regulated by proteins
containing the widespread LOV/PAS domains.
Materials and Methods
Strains. TheWTstrainusedhereisOR74A.Allsingle-knockoutstrainscamefrom
the Neurospora knockout project (55) and have been deposited in the Fungal
Genetics Stock Center (www.fgsc.net). The identity of csr-1 knock-in strains
used in this study was examined by PCR analysis (30 cycles) for the absence of
thecsr-1ORFinthegenome;allappeartobehomokaryons,asshowninFig.S7.
NCU (Neurospora crassa unfinished) strain numbers are from Neurospora an-
notation (www.broad.mit.edu/annotation/genome/neurospora/home.html).
Culture Conditions and Light Treatment. Frozen conidia were inoculated onto
a minimal Vogel slant (56) 1 wk in advance of experiments to generate fresh
conidia. On day 0, to form a mycelial layer, 107conidia suspended in sterile
water were inoculated into a 10-cm Petri dish with 20 mL of Bird medium
(57) containing 2%(wt/vol) glucose. After 24 h of incubation in DD at 25 °C,
a mycelial plug was cut with a no. 4 cork borer (8-mm diameter) and
transferred into a 125-mL flask with 50 mL of Bird medium containing 2%
(wt/vol) glucose. All procedures were performed in a low red-light envi-
ronment to avoid any possible light-stimulating effects. After another 24 h
of culturing with constant shaking (125 rpm) in DD at 25 °C, the flasks were
Fig. 5.
photoadaptation. (A) Schematic depicting the light induction kinetics of GFP-
VVD-V5 in either the gfp-vvd-v5 or gfp-vvd-v5cxstrain (constitutively ex-
pressed). (B) Photoadaptation phenotype of the gfp-vvd-v5 and gfp-vvd-v5cx
strains at LL60. Photoadaptation kinetics were determined by measuring
light induction of al-3 (C), sub-1 (D), and frq (E) at LL15 using RT quantitative
PCR (QPCR) analysis (n = 3, mean values ± SE). Asterisks indicate statistical
significance as determined by an unpaired t test. ***P < 0.001; **P < 0.01. NS,
difference between two groups is not significant. LL indicates constant white
light stimulus with a photon flux of 20 μmol·m2·s.
Constitutively expressed VVD is sufficient to change the dynamics of
Fig. 6.
transiently binds to the promoter of light-responsive genes to activate
transcription, including vvd. The induced VVD protein accumulates in the
nucleus and physically interacts with WCC to regulate photoadaptation by
repressing WCC activity in constant light. The kinetics of photoadaptation
are predominantly regulated by the amount of VVD protein in the system.
Components capable of sensing light directly through a chromophore are
marked with dashed lines.
Photoadaptation in Neurospora. After light activation, the WCC
Chen et al.PNAS
| September 21, 2010
| vol. 107
| no. 38
| 16719
PHYSIOLOGY