Light-Dependent and Circadian Transcription Dynamics In Vivo Recorded with a Destabilized Luciferase Reporter in Neurospora

We show that firefly luciferase is a stable protein when expressed at 25°C in Neurospora, which limits its use as transcription reporter. We created a short-lived luciferase by fusing a PEST signal to its C-terminus (LUC-PEST) and applied the LUC-PEST reporter system to record in vivo transcription dynamics associated with the Neurospora circadian clock and its blue-light photosensory system over the course of several days. We show that the tool is suitable to faithfully monitor rapid, but also subtle changes in transcription in a medium to high throughput format.


Introduction
Firefly luciferase from Photinus pyralis is a highly sensitive reporter for monitoring gene expression activity in a variety of organisms [1,2,3,4,5]. In contrast to fluorescence, luciferase does not require light irradiation for excitation and is thus neither subject to photobleaching nor phototoxicity. Hence, luciferase is a suitable reporter for long term in vivo measurements that are extensively used to record and monitor circadian clock activity in bacteria, fungi, plants, flies and mammals [3,6,7,8,9].
Circadian clocks are self-sustained molecular oscillators that are dependent on interconnected transcriptional and posttranscriptional negative feedback loops [10]. In Neurospora, the circadian transcription factor WHITE COLLAR COMPLEX (WCC) controls rhythmic expression of FREQUENCY (FRQ), which in turn inhibits the WCC in a negative feedback loop [11,12,13,14]. The WCC is also directly activated by light via its LOVphotoreceptor domain [15]. In light-dark cycles the activity of the WCC is additionally attenuated by the LOV-photoreceptor VIVID [16,17,18], which directly interferes with light-dependent dimerization of the WCC via LOV-LOV interactions [15]. These feedback loops support rhythmic transcription of frq and other clock-controlled genes in constant darkness and also in light-dark cycles.
A Neurospora reporter assay based on a codon-optimized luciferase gene allows to record circadian genes expression rhythms in vivo over the course of several days [19]. The stability of firefly luciferase is temperature dependent [20,21]. In mammalian cells at 37uC luciferase is degraded with a half-time of about 3-4 h [22]. Addition of a PEST sequence significantly destabilizes luciferase [23], rendering it more suitable as a reporter for dynamic processes on a short time-scale.
We show here that luciferase is extremely stable (t 1/2 , 8 h) in Neurospora, which is generally cultured at 25uC. The pronounced stability of luciferase at 25uC limits its use as a reporter for dynamic transcription processes. Addition of a PEST sequence accelerates the turnover of luciferase in Neurospora by more than one order of magnitude. We show that the destabilized LUC-PEST protein (t 1/2 , 25 min) is a faithful reporter of promoter activity that readily uncovers rapid and considerably complex transcriptional dynamics.

Plasmid Construction and Neurospora Transformation
A Neuropsora codon optimized PEST sequence (synthetized by GenScript) was inserted in front of the luciferase STOP codon in pFH62 luc (pBM60-luc-trpC, [26]) via a genetically engineered BglII site. The frq promoter and the vvd promoter, respectively, were amplified by PCR and inserted into via BamH1/NotI and EcoRI/NotI, respectively, into pFH62 lucPEST. Neurospora conidia were transformed as described [27]. The his-3 locus was used for genomic integration of the plasmids by homologous recombination. Codon

RNA Analysis
RNA was prepared with peqGOLD TriFAST (peqLab, Erlangen, Germany) and reverse transcribed with the Maxima First Strand cDNA Synthesis Kit (Fermentas). Transcript levels were analyzed by quantitative real-time PCR in 96-well plates with the StepOnePlus Real-Time PCR System (Applied Biosystems). TaqMan Gene Expression Master Mix (Applied Biosystems) and was used with gene-specific primers and probes [13].

In vivo Luciferase Measurements
Sorbose medium containing 16 FGS (0.05% fructose, 0.05% glucose, 2% sorbose), 16 Vogels, 1% agarose, 10 ng/ml biotin and 150 mM firefly luciferin was used for the assessment of the luciferase rhythms. 96-well plates were inoculated with 3610 4 conidia per well and incubated in DD at 25uC. Bioluminescence was recorded in DD or LD at 25uC with an EnVision Xcite Multilabel Reader (Perkin Elmer). The light intensity used in LD cycles was 40 mE.
Light pulse assay ( Fig. 1C): liquid standard growth medium contained 2% glucose, 0.5% L-arginine, 16 Vogel's and 150 mM firefly luciferin. 96-well plates were inoculated with 1.5610 5 conidia per well and incubated in DD for 1 day at 25uC. Plates were then exposed to a 1 min light pulse (80 mE) before bioluminescence was recorded in DD at 25uC.

A PEST Signal Destabilizes Luciferase in Neurospora
In order to decrease the half-life of firefly luciferase (LUC) a PEST-encoding DNA template [23] was codon optimized for Neurospora and fused to the 39 end of the codon-optimized luc ORF [19]. The luc-PEST fusion gene was expressed in Neurospora under the control of the frequency (frq) promoter (Fig. 1A). LUC-PEST was expressed at ,10-fold lower level than unmodified LUC (Fig. 1B  left), suggesting that the protein is unstable. To assess the destabilizing effect of the PEST-tag liquid cultures of strains expressing LUC and LUC-PEST, respectively, were treated with the translation inhibitor cycloheximide (CHX). Mycelia were then harvested over a time course of 8 h and luciferase activity (bioluminescence) was measured in vitro in protein extracts prepared from these samples. Under these conditions LUC-PEST was rapidly degraded with a half-time (t 1/2 ) of , 25 min while unmodified LUC was slowly degraded with an apparent t 1/2 .8 h, (Fig. 1B right). Thus, the PEST-tag accelerates the turnover kinetics of the protein by more than one order of magnitude.
The rapid degradation of LUC-PEST prompted us to investigate the response of the frq promoter to a short light pulse (80 mE, 1 min). frq mRNA peaked about 15 minutes after the light pulse and decreased to baseline levels after , 2 h (Fig. 1C, orange line). Bioluminescence supported by frq::luc-PEST reached its peak 40-60 minutes after the light pulse and then decreased to baseline levels after , 3.5 h. In contrast, the activity of the stable luciferase supported by frq::luc peaked after 2 h and levels did not drop significantly over the course of five hours. The data highlights the usefulness of both, stable and unstable reporters. The frq::luc-PEST reporter activity reflects with a rather short delay the transcriptional dynamics of the frq promoter, while the activity of frq::luc corresponds, due to the pronounced stability of LUC, to a temporal integration of the produced mRNA molecules.

Free-running frq Transcription Rhythms
We compared the circadian luciferase activity rhythms supported by the frq::luc and frq::luc-PEST reporters. The corresponding strains were cultured under restricted growth conditions (sorbose-containing solid growth medium), which allow life-cell measurements over extended time periods in a 96-well plate format. In constant darkness (DD) both reporters oscillated robustly with a free running period of , 23 h ( Fig. 2A). However, the amplitude (peak:trough ratio) supported by frq::luc-PEST was substantially higher than the amplitude of the frq::luc strain (6-fold versus 2.3-fold). Thus, the luc-PEST reporter appears to be more sensitive and may thus be suitable for recording of low amplitude promoter activity rhythms that would be masked by stable luciferase.

Entrained frq Transcription Rhythms
A key property of circadian clocks is their responsiveness to environmental cues such as light [28,29,30]. We recorded the activity of frq::luc-PEST and frq::luc strains grown under a 12 h light/12 h dark (LD) cycle (Fig. 2B). In this artificial square wave light/dark regime LUC-PEST activity reached a maximum rapidly after lights-on and then adapted within a few hours at a high steady-state levels. After light-to-dark transition LUC-PEST levels decreased quickly, reaching trough levels about 2-3 h after lights-off. The peak:trough expression ratio supported by frq::luc-PEST was , 20-fold. This activity profile of frq::luc-PEST corresponds closely to the previously reported temporal expression profile of frq RNA in light/dark cycles [31,32], indicating that the destabilized luciferase is a faithful reporter of the transcription dynamics of the frq promoter. In contrast, the activity of the stable luciferase encoded by frq::luc increased steadily throughout the light phase and then decreased during the dark phase. The peak:through ratio of the luc activity was about 2-fold. The temporal activity profile of frq::luc does not resemble the transcription dynamics the frq promoter.

Transition from Entrained to Constant Conditions
Transcript levels supported by the frq promoters are substantially higher in light than in the dark [33]. We recorded with frq::luc-PEST and frq::luc reporter strains the transition from light/ dark entrained conditions to constant darkness. Luciferase activity of the strains was measured for 4 days in 12 h/12 h LD cycles followed by 3 days in DD. The frq::luc-PEST reporter impressively revealed differences in the expression levels and amplitude of frq promoter driven expression under entrained and free running conditions (Fig. 2C). Thus, the luc-PEST reporter allows analyzing temporal expression profile of the frq promoter over an extended time period under varying conditions. In contrast, the rapid shutdown of frq transcription after transition from entrained to constant conditions was not resolved by the frq::luc strain (Fig. 2C). Due to the slow degradation of the stable LUC that had accumulated at high levels during the preceding light phase the free-running frq promoter oscillations in DD was almost completely masked.

Frq promoter Dynamics in WT versus Dvivid
VVD determines the phase of the circadian clock by inhibiting the light-activated WCC in particular after light-to-dark transition of Neurospora [34]. We analyzed the dynamics of frq::luc-PEST expression in a Dvivid (Dvvd) strain under the same LD and DD regime as for WT. Following light-to-dark transitions, LUC-PEST activity decreased more slowly in Dvvd than in WT, which perpetuated in constant conditions into a 4 h phase delay (Fig. 2D). Hence, considerably small differences in circadian promoter activity are readily and reliably detected by the luc-PEST reporter. The high temporal resolution of the frq::luc-PEST reporter revealed in addition complex differences between Dvvd and WT in the adaptation dynamics of the frq promoter during the light phase in (Fig. 2D), which will be further investigated elsewhere.

Transcription Dynamics in Complex Light Regimes
The blue-light photoreceptor VVD regulates sensory adaptation of Neurospora to light by inhibiting the light-activated WCC [17,32,35,36], allowing Neurospora to respond to changes in light intensity over several orders of magnitude.
To analyze the temporal dynamics of the vvd promoter, we expressed vvd::luc-PEST and vvd::luc reporters in a WT background. Dark-grown conidial cultures of vvd::luc-PEST and vvd::luc were exposed in 96-well plates to consecutive 15 h light-intensity steps of 0.35 mE, 3.5 mE and 30 mE followed by a final 48 h period in the dark. As shown for vvd::luc-PEST, light responses of 20 individual cultures were highly reproducible (Fig. S1A, B). Consistent with vvd mRNA data obtained from a similar experimental setup [17], LUC-PEST activity increased sharply after each light step followed by an adaptation phase and approached a rather constant steady-state level that correlated with light intensity (Fig. 3A). Expression levels dropped rapidly after the subsequent light-to-dark transition and then oscillated at a low level in circadian fashion. The data demonstrates that Neurospora quickly responds to increases as well as decreases in light intensity, adjusting expression levels of vvd according to ambient light. In contrast, due to the stability of the luciferase, bioluminescence traces of a vvd::luc strain failed to resolve the induction and adaptation responses of the vvd promoter to the light intensity steps and to the final dark-transition (Fig. 3B).

Discussion
Destabilized luciferase reporters have been successfully used in a number of circadian as well as non-circadian studies [37,38,39,40,41]. In Neurospora, a codon-optimized luciferase has been used as transcriptional reporter of circadian promoter activity and as translational fusion with the FRQ protein to monitor clock protein rhythms [42]. We have shown here that in Neurospora firefly luciferase (LUC) activity is stable (t /1/2 ,8 h). The same protein in mammalian cells is turned over significantly faster (t /1/2 ,4 h). Since LUC is a thermo-labile protein, it is likely that the differences in LUC stability are due to the growth temperatures, i.e 25uC for Neurospora and 37uC for mammalian cells. The considerable stability of LUC at 25uC limits its use as transcription reporter, particularly as a reporter of rapid transcription dynamics, in Neurospora but presumably also in other The frq::luc-PEST reporter conveniently resolves the known differences in transcription dynamics between WT (blue) and Dvvd (red) strains. The frq::luc-PEST reporter was expressed in a Dvvd strain, which was exposed to the same light-dark regime as above. The frq::luc-PEST activity of WT grown in the same plate from part (C) is shown again for comparison (dotted line). Insert: magnification showing the ,4 h phase delay of Dvvd after release to DD [34]. doi:10.1371/journal.pone.0083660.g002 organisms that grow at lower temperatures. A PEST signal fused to LUC tremendously destabilizes the protein, triggering its degradation in Neurospora with a half-time of only ,25 min. The unstable luciferase is a faithful reporter of promoter activity, resolving rapid as well as small changes in gene transcription, which would be masked by a stable LUC reporter. The LUC-PEST reporter system in combination with growth restricted culture conditions makes it possible to record promoter activity in vivo over several days. In combination with a plate-reader equipped with an automated stacker, the system is suitable for medium to high throughput promoter analyses in 96-well plate formats. The high-density time course analyses are highly reproducible even under complex physiological regimes (e.g. light-dark cycles or light steps). The system is thus superior to conventional approaches based on labor intensive and rather error-prone quantification of RNA expression by qPCR or Northern analysis.