Multiple circadian-regulated elements contribute to cycling period gene expression in Drosophila
1997; Springer Nature; Volume: 16; Issue: 16 Linguagem: Inglês
10.1093/emboj/16.16.5006
ISSN1460-2075
Autores Tópico(s)Light effects on plants
ResumoArticle15 August 1997free access Multiple circadian-regulated elements contribute to cycling period gene expression in Drosophila Ralf Stanewsky Ralf Stanewsky Department of Biology and National Science Foundation Center for Biological Timing, Brandeis University, Waltham, MA, 02254 USA Search for more papers by this author Creston F. Jamison Creston F. Jamison Department of Biology and National Science Foundation Center for Biological Timing, Brandeis University, Waltham, MA, 02254 USA Search for more papers by this author Jeffrey D. Plautz Jeffrey D. Plautz Department of Cell Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA, 92037 USA Search for more papers by this author Steve A. Kay Steve A. Kay Department of Cell Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA, 92037 USA Search for more papers by this author Jeffrey C. Hall Corresponding Author Jeffrey C. Hall Department of Biology and National Science Foundation Center for Biological Timing, Brandeis University, Waltham, MA, 02254 USA Search for more papers by this author Ralf Stanewsky Ralf Stanewsky Department of Biology and National Science Foundation Center for Biological Timing, Brandeis University, Waltham, MA, 02254 USA Search for more papers by this author Creston F. Jamison Creston F. Jamison Department of Biology and National Science Foundation Center for Biological Timing, Brandeis University, Waltham, MA, 02254 USA Search for more papers by this author Jeffrey D. Plautz Jeffrey D. Plautz Department of Cell Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA, 92037 USA Search for more papers by this author Steve A. Kay Steve A. Kay Department of Cell Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA, 92037 USA Search for more papers by this author Jeffrey C. Hall Corresponding Author Jeffrey C. Hall Department of Biology and National Science Foundation Center for Biological Timing, Brandeis University, Waltham, MA, 02254 USA Search for more papers by this author Author Information Ralf Stanewsky1,‡, Creston F. Jamison1,‡, Jeffrey D. Plautz2, Steve A. Kay2 and Jeffrey C. Hall 1 1Department of Biology and National Science Foundation Center for Biological Timing, Brandeis University, Waltham, MA, 02254 USA 2Department of Cell Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA, 92037 USA ‡R.Stanewsky and C.F.Jamison contributed equally to this work The EMBO Journal (1997)16:5006-5018https://doi.org/10.1093/emboj/16.16.5006 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info A new regulatory element necessary for the correct temporal expression of the period (per) gene was identified by monitoring real-time per expression in living individual flies carrying two different period–luciferase transgenes. luciferase RNA driven from only the per promoter was not sufficient to replicate the normal pattern of per RNA cycling; however, a per–luc fusion RNA driven from a transgene containing additional per sequences cycled identically to endogenous per. The results indicate the existence of at least two circadian-regulated elements—one within the promoter and one within the transcribed portion of the per gene. Phase and amplitude analysis of both per–luc transgenes revealed that normal per expression requires the regulation of these elements at distinct phases and suggests a mechanism by which biological clocks sustain high-amplitude feedback oscillations. Introduction Circadian rhythms are daily oscillations of biological processes that occur with a periodicity of ∼24 h. These oscillations persist under constant environmental conditions, indicating that they are controlled by an endogenous pacemaker known as a biological clock. A large body of evidence implicates both the period (per) and timeless (tim) genes of Drosophila melanogaster as central components of the circadian clock (reviewed by Sehgal et al., 1996). In constant darkness and temperature (DD), the null mutations per01 and tim01 each cause arrhythmic pupal eclosion and locomotor activity (Konopka and Benzer, 1971; Sehgal et al., 1994). The mRNA and protein gene products of both per and tim show daily oscillations in abundance, and these fluctuations are also abolished by either the per01 or tim01 mutation (reviewed by Sehgal et al., 1996). The molecular oscillations of both tim protein (TIM) and per protein (PER) appear similar. However, TIM is degraded rapidly in the presence of light; this observation suggests a mechanism which can explain how per and tim oscillations are reset by light (Hunter-Ensor et al., 1996; Lee et al., 1996; Myers et al., 1996; Zeng et al., 1996). The observation that per and tim mRNA cycle with a period and phase that is dependent upon the amino acid sequences of their coded proteins suggests that PER and TIM act in a negative feedback loop whereby they regulate their own transcription (Hardin et al., 1990; Marrus et al., 1996; Sehgal et al., 1996). Further evidence for the existence of this feedback loop comes from the examination of the cellular distribution of PER and TIM (Liu et al., 1992; Curtin et al., 1995; Hunter-Ensor et al., 1996). The timed nuclear localization of both PER and TIM is dependent on the presence of both proteins (Vosshall et al., 1994; Hunter-Ensor et al., 1996; Myers et al., 1996) and is preceded by the formation of a heterodimeric PER–TIM complex (Lee et al., 1996; Saez and Young, 1996; Zeng et al., 1996). Thus, it appears that per and tim form interdependent feedback loops where PER and TIM interact in vivo to translocate into the nucleus at the appropriate time of day and directly or indirectly repress per and tim transcription. Much of the evidence presented above suggests that per and tim comprise components of the biological clock. However, several inconsistencies exist that complicate this conclusion. First, assuming a simple negative feedback model for the oscillation of per and tim, one would anticipate that the lack of PER and TIM in the nucleus would lead to high per RNA levels. However, per RNA in a per01 mutant background shows relatively low levels of RNA when compared with per+ (Hardin et al., 1990; Van Gelder and Krasnow, 1996). This observation suggests that the mechanism regulating period gene expression is more complex than a simple negative feedback system. Second, when per mRNA is monitored in DD by RNase protections, the mRNA rhythm dampens appreciably by the fourth to fifth day of monitoring (Hardin et al., 1990). This observation is curious because Drosophila locomotor activity free-runs (remains cyclical) in DD for at least 3 weeks (e.g. Helfrich, 1986; Power et al., 1995). This suggests three possibilities: (i) per mRNA cycling is not responsible for the robustness of these activity rhythms; (ii) absence of a light–dark (LD) cycle reduces the amplitude of per mRNA cycling in many tissues but not in the behaviorally relevant clock cells; or (iii) the variation among the endogenous periods of individual flies (∼1–1.5 h) causes the flies to drift out of phase relative to one another and results in a dampened overall mRNA cycle. We developed a novel in vivo luciferase assay that allows us to measure real-time gene expression in individual flies (Brandes et al., 1996; Plautz et al., 1997). Several refinements to this assay now permit us to address questions concerning the continuity of rhythmic per expression in single flies under free-running conditions. In addition, the ability repeatedly to measure individual flies with high time resolution allows us to study the temporal regulation of per gene expression by comparing various period–luciferase fusion genes in a highly efficient and detailed manner. Here we report the expression analysis of two per–luc transgenes using both our novel in vivo luciferase assay and high time resolution RNA quantitations. One construct is a newly generated protein fusion (designated BG-luc), whereas the second is the per–luc only (plo) promoter fusion construct (Brandes et al., 1996). In vivo luciferase assays revealed no significant ultradian oscillations for per01 flies carrying the plo or BG-luc transgene. This indicates that the ultradian behavior observed in individual per01 flies is not correlated with molecular rhythmicity of per (reviewed by Dowse and Ringo, 1992). Our single-fly assays also indicated that expression of both per–luc reporter transgenes dampens to arrhythmicity in DD conditions. This finding, which differs from our earlier report (Plautz et al., 1997), suggests that the molecular dampening in DD (cf. Hardin et al., 1990) is due to dampening of per gene expression in individual flies and not exclusively to asynchrony among flies. Detailed comparisons of the amplitudes and phases associated with the two per–luc constructs revealed differences between the per-promoted reporter RNAs and that transcribed from the endogenous per gene. Furthermore, in the per01 genetic background, the relative level of the plo transgene's RNA is high whereas the BG-luc transgene's RNA level is low. Taken together, our observations reveal the existence of a circadian-regulated element within the transcribed portion of the period gene. The significant findings of this study are presented in a model which suggests a mechanism by which high-amplitude feedback oscillations are sustained. Results Real-time measurement of period gene expression in individual flies To elucidate the mechanisms regulating temporal period gene expression, two different per–luc constructs were designed by fusing genomic regions of the period gene to coding sequences from the firefly luciferase gene. One construct (plo) contains only the per promoter fused to a luciferase cDNA (Brandes et al., 1996). The second construct contains the same per and luc sequence as plo, plus additional per genomic DNA sufficient to encode the N-terminal two-thirds of PER (designated BG-luc) (Figure 1A). Figure 1B shows an average plot of a BG-luc line tested in LD in a per+ genetic background. All BG-luc flies showed a similar temporal expression pattern. Figure 1E shows a representative example of a single BG-luc fly's bioluminescence record and indicates that our improved assay conditions allow very clean single fly oscillations (see Materials and methods). These clean oscillations are also evident in flies from the plo lines (Figure 1C). Figure 1.BG-luc and plo temporal bioluminescence in LD. (A) plo and BG-luc constructs are shown along with the per genomic region for reference. The plo construct contains per genomic DNA from −4.2 kb to +32 bp fused to the 1.8 kb luciferase cDNA (Brandes et al., 1996). The BG-luc construct contains per genomic DNA from −4.2 to +5.6 kb (Citri et al., 1987) fused to the luciferase cDNA. Closed bars indicate untranslated exons and open bars indicate translated portions of exons. luciferase cDNA sequences are indicated. (B) Comparison of averaged per+ BG-luc flies (n = 71) and averaged per01 BG-luc flies (n = 37) in LD. per+ BG-luc flies show rhythmicity with ∼2.5-fold amplitude and the phase of its peak at ZT19 (see Table I). Noticeable downward trend and amplitude reduction of the recordings are due to an artifact of the luciferase assay and can be removed (see Figure 4B–D). A secondary peak of per transcription cannot be discerned for the BG-luc strain (as opposed to the plo strain; see C). Low-amplitude oscillations of per01 BG-luc are most likely to be due to BG-luc interaction with TIM (see Results). Note the intermediate level of bioluminescence in the per01 background when compared with per+. (C) Comparison of averaged per+ plo1b-1 flies (n = 67) and averaged per01 plo1b-1 flies (n = 36) in LD. per+ plo flies also show an amplitude of ∼2.5-fold; the apparent lower amplitude of plo compared with BG-luc oscillations is due to the higher expression level of BG-luc (see D). In contrast to BG-luc, the phase of the plo peak occurs at ZT17 (see Table I). In addition, a secondary shoulder is evident just when the lights come on (see also Brandes et al., 1996). per01 flies show no evidence of rhythmicity. Note the high level of bioluminescence in the per01 background when compared with per+. (D) The per+ BG-luc and per+ plo1b-1 averages were normalized and plotted together to reveal an ∼2 h phase difference. Since the BG-luc bioluminescence levels are ∼5-fold higher than plo levels, the first evening peak of both data sets was normalized to 1. Quantifying the absolute level of RNA expression from both BG-luc and plo transgenes revealed that the ∼5-fold bioluminescence difference is due to an ∼5-fold difference in RNA levels (data not shown). Note the amplitude similarity and phase difference. (E) Example of a single per+ BG-luc fly from the average in (B). (F) Comparison of per+ BG-luc (n = 14) and tim01 BG-luc (n = 10) fly averages in LD. Flies are heterozygous for the transgene. Note the absence of the low-amplitude per01 BG-luc oscillations in tim01 BG-luc. Bioluminescence was measured in counts per second. ZT indicates Zeitgeber time. Open bars indicate lights on and closed bars indicate lights off. Download figure Download PowerPoint Figure 2.Temporal RNA expression pattern of BS-CAT and BG-luc transgenes compared with per RNA cycling. (A) RNase protection experiment, with 1 h time resolution, of flies carrying the BS-CAT transgene. Band intensities of the protected CAT fragments were quantified and standardized to the rp49 signal. The average pattern of per expression (resulting from five independent experiments in which the per transcript was tracked: cf. Brandes et al., 1996) is plotted in the same graph; for both transcripts, the maximum expression value was set to 1. CAT reporter RNA shows an earlier rise in abundance and also cycled with a lower amplitude (∼3-fold) compared with the per transcript (∼15-fold). The amplitudes were calculated after dividing the peak values of expression (between ZT8 and ZT15 for BS-CAT and ZT13 and ZT16 for per) by the trough values (ZT22 to ZT4 and ZT23 to ZT4, respectively). In a different BS-CAT transgenic line, the same earlier onset of transcription was observed and the amplitude was also reduced (∼4-fold). (B) RNase protection experiment, with 1 h time resolution, involving BG-luc transgenic flies. Endogenous per and BG-luc reporter RNAs were assayed simultaneously using the luc and per 5/6 probes (Materials and methods). Band intensities were plotted after standardization to the rp49 control and setting the absolute peak value for each cycling RNA to 1. In the lane corresponding to ZT13, much less RNA was loaded, explaining the dip in the rising portion of both RNA curves. The amplitudes (14-fold) were calculated as described above using the peak values from ZT15 to ZT18 and the trough values from ZT1 to ZT5 for both RNAs. The white and black portions of the bar represent times when the lights were on or off, respectively. Download figure Download PowerPoint Figure 3.RNA expression levels of reporter and per RNA, in per+ and per01 genetic backgrounds, of plo and BG-luc transgenic flies. (A) RNA expression of flies from the plo1a-1 strain was measured in a per+ and per01 background in a 2 h time resolution RNase protection experiment using the per 2/3 and luc probes (Materials and methods). Protected luciferase and per RNA fragments in both backgrounds were separated on the same gel to allow a direct comparison between transcription levels in both genetic backgrounds. (B) Quantification of the data shown in (A) after standardizing the band intensities of the protected fragment against the rp49 control and setting the maximum expression values of each transcript to 1. In a per+ background, luciferase and per RNAs cycled with their characteristically different phases and amplitudes (see also Figure 2 and Brandes et al., 1996). In a per01 background, the level of per RNA is only at 20% of its peak level in per+ (here: ZT14 and ZT16), whereas the level of luciferase RNA in per01 is at 81% of its maximal level in per+ (which was at ZT9, ZT11, ZT13, ZT14 and ZT15 in this experiment). Similar results were obtained in two different experiments with 2 h time resolution (using the plo1 strain) where per01 RNA was ∼50% less abundant than two per+ control RNAs from plo1 flies collected at ZT8 and ZT10. In the same experiments, the average luciferase RNA abundance in per01 was at the same level of the luciferase RNA isolated from the controls flies at ZT8 and ZT10. Recall that per expression is still rising between ZT8 and ZT10, whereas luciferase expression is reaching its maximum level at that time. (C) Same experiment as in (A) performed with BG-luc transgenic flies using the per 5/6 and luc probes. (D) Quantification of data shown in (C), after standardizing against rp49 and setting the absolute peak expression value of each transcript to 1. For per, the exon 6 band is graphed, although exon 5 gave very similar results. In a per+ background, BG-luc and per RNAs cycled with nearly identical phase and amplitude (see also Figure 2B). In a per01 background, both transcripts are expressed at a constitutively low level (per, 33%; BG-luc, 40.1%) when compared with their peak times of expression in a per+ background (here ZT15 and ZT18 for both transcripts). The white and black portions of the bar represent 12 h time segments when the lights were on or off respectively. Download figure Download PowerPoint To compare the oscillations of BG-luc and plo transgenes, both data sets were normalized and plotted together (Figure 1D). In both cases, the amplitudes are ∼2.5-fold; however, the phase of BG-luc is ∼2 h later than plo (Figure 1D; see below for quantitative analysis). Additionally, a secondary bioluminescence peak is observed at the dark to light transition in the plo (Figure 1C; Brandes et al., 1996) but not in the BG-luc transgenics. Comparison of BG-luc and plo flies in a per01 genetic background also revealed differences. The per01 plo average shows no evidence of cycling and maintains high levels of bioluminescence relative to the plo oscillations in a per+ genetic background (Figure 1C). In contrast, the per01 BG-luc average exhibits low-amplitude oscillations and maintains intermediate levels of bioluminescence when compared with BG-luc in a per+ genetic background (Figure 1B). These low-amplitude oscillations appear to be driven by protein interactions between BG-luc and TIM, the latter of which shows minor oscillations in a per01 mutant background (Zeng et al., 1996; Dembinska et al., 1997; see below). Analysis of individual flies reveals significant differences between BG-luc and plo expression The principal advantage of monitoring period gene expression by means of an in vivo reporter is the ability to monitor large numbers of individual flies from which one can make statistical conclusions. To examine whether the differences between plo and BG-luc bioluminescence are statistically significant, each bioluminescence record was subjected to a quantitative cosine fit analysis (see Materials and methods; also see Plautz et al., 1997). From this analysis, amplitude, period and phase estimates were determined from each fly's record (Table I). The presence of a secondary bioluminescent peak in the plo transgenics prompted us to search for rhythmicity in two different control transgenes—luciferase driven from a heat-shock promoter (hsp-luc; Lockett et al., 1992) and luciferase driven from a P-element promoter. Lines from both transgenes showed low-amplitude 12 or 24 h oscillations (see, for example, Figure 4A). Additionally, alterations in the testing conditions (e.g. monitoring flies that carried only one copy of the transgene) resulted in the appearance of secondary peaks for BG-luc (data not shown). Thus, all luciferase transgenes, irrespective of their promoter, showed evidence of low-amplitude oscillations. The widespread persistence of these oscillations indicates that they are unrelated to per gene expression and are a consequence of monitoring luciferase bioluminescence (see Discussion). Figure 4.Downward trend removal from luciferase bioluminescence. (A) Average of per+ hsp-luc flies (n = 54) from three different lines with different genomic insert locations. In addition to the downward trend, low-amplitude, 12 h oscillations can be discerned (see Results). Note how the morning hsp-luc peaks coincide with the secondary peaks of the plo strains (see C). (B–D) The downward trend in the per+ BG-luc (B), per+ plo (C) and hsp-luc (D) have been removed to reveal the true cycling of per-luc expression. The hsp-luc trend-adjusted average reveals the relatively flat expression expected for flies tested at constant temperature. The same data manipulation has successfully removed both the amplitude reduction and downward trend of both per–luc constructs. See Materials and methods for the trend-adjusting procedure. Bioluminescence was measured in counts per second. ZT indicates Zeitgeber time. Open bars indicate lights on and closed bars indicate lights off. Download figure Download PowerPoint Table 1. Quantitative analysis of plo and BG-luc bioluminescent oscillations Transgenic line LD DD # Rhy/tested (% rhythmic) Mean period ± SEM Mean Rel-Amp Error ± SEM Mean phase ZT ± SEM # Rhy/tested (% rhythmic) Mean period ± SEM Mean Rel-Amp Error ± SEM per+ BG-luc 71/72 (99) 24.4 ± 0.0 0.20 ± 0.00 19.3 ± 0.1 53/61 (87) 25.1 ± 0.1 0.48 ± 0.01 per+ BG-luc3 15/16 (94) 24.2 ± 0.1 0.25 ± 0.02 20.2 ± 0.2 7/14 (50) 23.8 ± 0.3 0.58 ± 0.03 per+ BG-luc42 15/15 (100) 23.9 ± 0.1 0.31 ± 0.03 21.5 ± 0.5 10/15 (67) 25.0 ± 0.4 0.45 ± 0.02 per+ BG-luc56 32/32 (100) 24.4 ± 0.1 0.22 ± 0.01 19.4 ± 0.2 14/28 (50) 25.0 ± 0.4 0.54 ± 0.02 per+ BG-luc57 35/35 (100) 24.4 ± 0.1 0.23 ± 0.01 19.5 ± 0.1 21/32 (66) 25.8 ± 0.2 0.53 ± 0.02 per+ plo1b-1 67/67 (100) 24.4 ± 0.0 0.25 ± 0.01 17.1 ± 0.2 58/60 (97) 23.6 ± 0.1 0.45 ± 0.01 per+ plo1a-1 32/32 (100) 24.6 ± 0.1 0.38 ± 0.02 17.1 ± 0.4 17/31 (55) 24.0 ± 0.3 0.54 ± 0.03 per+ plo2a-2 15/16 (94) 24.2 ± 0.1 0.22 ± 0.01 18.1 ± 0.4 15/15 (100) 22.8 ± 0.2 0.33 ± 0.03 per+ plo2a-1 14/14 (100) 24.6 ± 0.1 0.30 ± 0.01 16.8 ± 0.3 per+ plo2b-3 10/10 (100) 24.3 ± 0.2 0.26 ± 0.02 17.6 ± 0.6 per+ plo3b-1 31/31 (100) 24.1 ± 0.1 0.22 ± 0.01 19.2 ± 0.3 30/31 (97) 23.2 ± 0.2 0.46 ± 0.01 per01 BG-luc 27/37 (73) 25.1 ± 0.5 0.52 ± 0.02 20.5 ± 0.8 0/39 (0) per01 BG-luc3 12/15 (80) 24.2 ± 0.2 0.54 ± 0.02 20.6 ± 1.0 1/15 (7) 10.9 0.61 per01 BG-luc57 14/14 (100) 24.5 ± 0.2 0.36 ± 0.03 20.3 ± 0.7 1/16 (6) 31.3 0.52 per01 plo1b-1 0/36 (0) 0/33 (0) per01 plo1a-1 1/15 (7) 22.8 0.61 0/13 (0) tim01 BG-luc 4/23 (17) 24.9 ± 1.0 0.62 ± 0.03 tim01 plo2b-3 1/14 (7) 25.0 0.68 In order to remove the low-amplitude oscillations from the quantitative analysis, we developed a rhythmic cut-off by quantitatively analyzing bioluminescence fluctuations from three hsp-luc strains. Assuming hsp-luc expresses constitutively at constant temperature, any resulting cycling should not be due to gene transcription. Thus, the empirically determined cut-off for rhythmicity [relative amplitude errors (Rel-Amp errors) 99% of all per+ BG-luc and per+ plo flies tested in LD were rhythmic and showed periods in the circadian range (22–26 h; Table I). In agreement with the average plot (Figure 1D), quantitatively determined phases for the plotted lines showed a 2 h delay: Zeitgeber time 19.3 (ZT19.3) for BG-luc and ZT17.1 for plo (Table I; ZT0 = lights on, ZT12 = lights off). Even though both transgenes exhibited phase variation between lines, the average of the mean phases from all the BG-luc lines was ZT20.0 compared with ZT17.6 for the plo lines. Statistical tests indicate that these phase differences are significant (P < 0.01). Individual per01 flies do not exhibit per ultradian oscillations Period analysis of single per01 plo flies revealed no significant ultradian or circadian oscillations in these individuals (Table I). One fly showed a 22.8 h period (Table I); however, using the assigned rhythmic cut-off, one would expect 5% of the tested flies to show rhythms (see Materials and methods). This result is significant considering the ultradian rhythms frequently seen in per01 locomotor activity records (Dowse and Ringo, 1992). Our results indicate that the behavioral rhythmicity observed for single per01 flies is unrelated to rhythmic expression from the per promoter. As seen for the average plot of per01 BG-luc (Figure 1B), low-amplitude oscillations were detected from the quantitative analysis of individual flies (Table I). These oscillations appear to be due to the BG-luc fusion's interactions with TIM protein. PER and TIM proteins interact via the PAS protein dimerization domain that is fully included in the BG-luc fusion protein (Gekakis et al., 1995; Saez and Young, 1996). TIM is known to cycle in LD conditions in a per01 mutant but not in DD (Hunter-Ensor et al., 1996; Myers et al., 1996; Zeng et al., 1996). Indeed, the low-amplitude bioluminescence cycling observed for per01 BG-luc in LD is absent in DD (compare Figures 1B and 5A). In addition, the BG-luc transgene's RNA in a per01 genetic background shows no evidence of cycling (Figure 3C and D). Lastly, when the BG-luc transgene was placed in the tim01 background, the low-amplitude cycling disappeared (Figure 1F). These findings demonstrate that the BG-luc's low-amplitude oscillations in a per01 genetic background are not due to rhythmic gene transcription but appear to be a consequence of light-driven TIM cycling and its interaction with the BG-luc fusion protein. Figure 5.BG-luc and plo temporal bioluminescence in DD. (A) Comparison of averaged per+ BG-luc flies (n = 71) and averaged per01 BG-luc flies (n = 37) in DD. As opposed to LD, per01 flies no longer show minor oscillations (Figure 1B). Although per+ flies show cycling, the amplitude dampens much faster (almost to the point of arrhythmicity). This dampening corresponds well with the dampening observed in RNase protections (Hardin et al., 1990, 1994). (B) Comparison of averaged per+ plo1b-1 flies (n = 58) and averaged per01 plo1b-1 flies (n = 33) in DD. As seen in LD, per01 shows no rhythmicity whereas per+ shows cycling; however, the amplitude dampens as seen in the BG-luc (see A). (C and D) Examples of a single per+ BG-luc fly (C) and a single per+ plo1b-1 fly (D) tested in DD. Note that the same degree of dampening is seen for both individuals and for the average. This result suggests that the majority of the dampening observed for the average in DD is not due to individual fly asynchrony but to dampening per expression within each fly (see Results). (E and F) The data from the two flies in (C) and (D) were trend-adjusted (E and F respectfully) to remove the dampening due to the luciferase assay artifact (see Materials and methods). This manipulation reveals the degree to which per expression dampens in DD. The resulting curves may be compared with the LD averages (Figure 4B and C) to reveal the dampening due to DD alone. Bioluminescence was measured in counts per second. One day of LD is included in each plot. The x-axis indicates hours in darkness. Open bars indicate lights on, closed bars indicate subjective night and stippled bars indicate subjective day. Download figure Download PowerPoint Normal period RNA expression requires two circadian-regulated elements Since the BG-luc transgene encodes a PER–LUC fusion protein and the plo transgene contains no PER amino acids, the differences observed between plo and BG-luc bioluminescence reflect regulation of either per–luc RNA or PER–LUC protein. To resolve the origin of these differences, the temporal RNA expression from both constructs was analyzed by RNase protections. The plo transgene RNA shows both an amplitude reduction and an advanced phase when compared with endogenous per (Figure 3A and B). This observation suggests that either the per promoter lacks specific regulatory sequences necessary for proper RNA cycling or luciferase sequences have an aberrant effect on RNA cycling. To distinguish between these possibilities, temporal RNA expression was analyzed from both a BS-CAT and the BG-luc construct. The BS-CAT construct contains the same per sequences present in plo but is fused to the bacterial CAT gene (Hardin et al., 1992). Our high time resolution RNase protections showed the CAT reporter RNA to exhibit a phase advance and amplitude reduction when compared with endogenous per RNA (Figure 2A), just as is seen for plo RNA (Figure 3A and B). Expression of both reporter RNAs started to increase strongly after ZT4, whereas per expression increased 2–4 h later (Figures 2A and 3A and B). Endogenous per RNA cycled with an ∼15-fold amplitude (average of five experiments: see legend for amplitude calculations). In contrast, the amplitude of plo and BS-CAT RNA cycling was reduced ∼3-fold (5-fold for plo, average of five experiments; 3.5-fold for BS-CAT, average of two experiments; for each construct two independent transgene insertion lines were tested). The plo lines showed an average RNA peak phase of ZT12.2 (SEM = 0.3; six experiments) for the reporter transcript and ZT13.9 (SEM = 0.2) for the endogenous per transcript (see Materials and methods). Temporal RNA expression analy
Referência(s)