Circadian regulation of the CYP1A1 promoter in vitro

As demonstrated in the work of Ndikung et al. (2020), the CYP1A1 induction mediated by TCDD (2,3,7,8-Tetrachlorodibenzo-p-dioxin) exhibits a circadian oscillation. In this study, the model system and the circadian synchronization process have been characterized in depth. The objective of the present study is to elucidate the mechanism behind this observation. Consequently, the initial focus was directed towards the AHR-mediated promoter activity of CYP1A1. To this end, we determined the CYP1A1 induction upon treatment with TCDD over the course of 48 h in synchronized hTERT-HME1-PER2 (abbreviated as HME1) circadian reporter cells. This reporter cell system allows the visualization of the circadian rhythm by monitoring the oscillating LUCIFERASE expression, which is under control of the human PER2 promoter. First, hTERT-HME1-PER2 cells were treated with 1 µM dexamethasone for 1 h to synchronize all cells in the same circadian phase, or were left untreated for non-synchronized control. Luciferase activity was monitored for 60 h to ensure a stable synchronization over 48 h. To analyze the circadian rhythmicity, the data obtained were analyzed using the ChronAlyzer software (Ndikung et al. 2020), which normalizes the bioluminescence counts (Fig. S1a) and fits the series of measurements to an oscillatory curve. (Fig. 1a). As expected, the HME1 cells exhibited a circadian rhythm with a 24 h period for at least 2 days and thus were suitable for CYP1A1 induction assays.

Fig. 1

CYP1A1 induction upon TCDD exposure is regulated by circadian promoter activation. a) Representative curves of fitted bioluminescence measurements of non-synchronized (blue) and synchronized (red) circadian HME1 reporter cells, expressing LUCIFERASE under control of the human PER2 promoter. For the synchronization confluent cells were treated with 1 µM dexamethasone for 1 h. For the non-synchronized condition, the cells were cultured at a semi-confluent stage and left untreated. The cells were monitored for 60 h and the bioluminescence signal was recorded every 30 min and analyzed with the ChronAlyzer software. The non-synchronized cells display no rhythmic bio-luminescence signal, whereas synchronized cells show a circadian oscillation with a period of approximal 24 h. b) CYP1A1 induction at different time points in synchronized and non-synchronized HME1 cells upon treatment with 0.5 nM TCDD. At indicated timepoints, CYP1A1 mRNA after TCDD exposure was analyzed by RT-qPCR. Fold change was calculated based on the DMSO control of the respective timepoint. In synchronized cells, the TCDD-mediated CYP1A1 induction displays a circadian pattern peaking at 24 and 48 h. Each bar represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed. Statistical significance was determined using an unpaired, one-tailed t-test. c) De novo CYP1A1 induction at different timepoints in non-synchronized and synchronized HME1 cells upon treatment with 0.5 nM TCDD. At indicated timepoints, mRNA was extracted from the nucleus and analyzed by RT-qPCR. The fold change of the de novo CYP1A1 mRNA after TCDD exposure was calculated based on the DMSO control at the respective timepoint. The TCDD-mediated de novo CYP1A1 expression exhibits a circadian pattern only in the synchronized cells indicating that the promoter is circadian regulated. Each data point represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed. Statistical significance was determined by using an unpaired, one-tailed t-test

To study CYP1A1 induction, non-synchronized and synchronized HME1 cells were treated with 0.5 nM TCDD for 12, 24, 36 and 48 h, and mRNA levels were determined by RT-pPCR. Total CYP1A1 mRNA levels in non-synchronized cells did not show a circadian pattern, whereas in synchronized cells a typical circadian expression pattern was observed, with peaks at 24 h and 48 h, respectively (Fig. 1b, S1b). Notably, the mature mRNA showed a slight accumulation after 36 h, which partially masks circadian-driven reduction of CYP1A1 mRNA in synchronized cells. This finding might be attributable to an altered long half-life of CYP1A1 mRNA in breast cells, as was previously described when comparing liver and breast cells. (Lekas et al. 2000; Lo et al. 2017). To investigate whether this circadian expression pattern is due to the circadian activity of the CYP1A1 promoter, we analyzed the de novo CYP1A1 mRNA levels extracted from the nucleus. The determination of the de novo CYP1A1 mRNA levels in synchronized cells treated with TCDD showed a significant circadian pattern over time with clear induction peaks at 24/48 h, as well as Low induction at the level of non-synchronized cells at 12/36 h (Fig. 1c, S1c). These findings suggest that the transcriptional activity itself is subject to rhythmic regulation, thereby indicating that AHR-dependent regulation of CYP1A1 expression is under circadian control.

Circadian expression of the AHR co-activator SP1

Earlier studies in mice report a circadian expression of Ahr in several tissues, whereas in human breast cells the AHR expression seems not to be under circadian regulation (Richardson et al. 1998; Huang et al. 2002; Ndikung et al. 2020). To ascertain the extent to which the AHR pathway is subject to circadian regulation, HME1 cells were synchronized or left non-synchronized. The mRNA levels of AHR, as well as its co-factors HSP90, XAP2, and p23, and ARNT and the transcription factor SP1, along with its negative regulator AHRR, were determined over time (12, 24, 36, and 48 hours) in the presence of 0.5 nM TCDD. The non-synchronized cells displayed a constant expression level of all analyzed genes over 48 h (Fig. 2a, S2a; blue bars). Interestingly, we also found no circadian expression pattern for the common AHR co-factors in synchronized cells (Fig. 2a, S2a; red bars). However, with the exception of AHRR, a circadian expression pattern was observed, with Lower induction after TCDD exposure at 12/36 h and a peak induction at 24/48 h (see Fig. 2a).

Fig. 2

Circadian pattern in protein levels of the AHR cofactor SP1. a) To identify AHR co-factors with a circadian expression level, the mRNA fold change of non-synchronized and synchronized HME1 cells after a treatment with 0.5 nM TCDD was analyzed at the indicated timepoints after synchronization. The mRNA levels of AHR, ARNT, XAP2, SP1, HSP90, P23 and AHRR were analyzed by RT-qPCR. For each timepoint the fold change of the CYP1A1 expression was calculated based on the DMSO control of the respective timepoint. The TCDD mediated induction of AHR, ARNT, XAP2, SP1, HSP90, and P23 remained stable over time in both non-synchronized and synchronized cells. The induction of the AHR target gene AHRR displayed under synchronized conditions a circadian pattern with peaking after 24 h and 48 h. Each bar represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed. b) Representative Western blots from three independent experiments of the AHR co-factors AHR, ARNT, XAP2, SP1, HSP90 and p23 from HME1 cells after 12, 24, 36, and 48 h of synchronization. The protein levels of SP1 show a circadian pattern with peaking at 24 h and 48 h upon synchronization indicating that SP1 protein levels under circadian regulation. β-Actin and total protein stain served as loading control. The relative amounts of proteins were derived from the representative Western blot signal intensity and normalized to β-Actin. The quantification was conducted with Image Studio™ Lite Software (LI-COR Biosciences)

Since the mRNA levels of AHR co-factors are not circadian regulated, we examined the protein levels in synchronized cells of the previously analyzed co-factors by Western blotting. Therefore, we synchronized the HME1 cells and determined the protein levels at 12, 24, 36 and 48 h after synchronization (Fig. 2b). In addition, we quantified the relative signal intensity of each band to support the visual impression of the Western blots. Although, the mRNA levels of the AHR co-factors were not regulated in a circadian manner, we observed a circadian pattern of the protein levels of SP1, with a Low level at 12/36 h and a peak expression level at 24/48 h (Fig. 2b). The derived relative protein levels support this finding, as only SP1 protein exhibited a clear circadian pattern.

Taken together, we found no circadian expression of AHR co-factors at the mRNA level, but SP1 protein levels appeared to be circadian regulated.

Modulators of circadian associated AHR activity

Except for SP1, none of the AHR co-factors displayed a circadian expression pattern. However, to determine if other co-factors influence the circadian modulation of CYP1A1 induction, we aimed in the next step to identify modulators of the circadian response to TCDD by siRNA-mediated knock down each co-factor individually. Nevertheless, a salient disadvantage of the HME1 cell system is the substantial difficulty in achieving effective transfection. Consequently, the cell system was changed, and the mechanistic issues were addressed using the M13SV1 human breast epithelial cell line. To ensure a proper circadian synchronization of the cellular system, the M13SV1 cells were transfected with circadian reporter plasmids, and circadian synchrony was monitored over the course of 48 hours (Fig. 3a, S3a). We found that the M13SV1 cells could be effectively synchronized. However, the synchrony was not as stable as in HME1 cells and almost Lost after a 48-hour period. Consequently, this cell system was considered to be suitable for our studies, but only for 36 hours after synchronization (circadian window, Fig. 3a). In order to ascertain whether the M13SV1 cells exhibit a rhythmic response within 36 hours, non-synchronous and synchronous cells were treated with 0.5 nM TCDD and CYP1A1 induction was determined after 12, 24 and 36 hours (Fig. 3b, S3b). In synchronous cells, a rhythmic CYP1A1 expression is observed, following the circadian rhythm, with an expression peak at 24 hours of TCDD treatment.

Fig. 3

Modulator of the circadian regulated AHR activity. a) Representative fitted curve of the circadian rhythm of M13SV1 cells transiently transfected with a PER2:LUCIFERASE (orange) or BMAL1:LUCIFERASE (blue) circadian reporter plasmid. M13SV1 were transfected via electroporation with the indicated reporter plasmids. 24 h after transfection confluent cells were synchronized by treatment with 1 µM dexamethasone for 1 h and subsequently monitored for 48h. The bioluminescence signal was recorded every 30 min and the obtained measurements was analyzed by the ChronAlyzer software. M13SV1 shows a stable synchronization with an anticyclic expression of the PER2 or BMAL1 control luciferase b) CYP1A1 induction at different time points in synchronized and non-synchronized M13SV1 cells upon treatment with 0.5 nM TCDD. CYP1A1 mRNA after TCDD exposure was analyzed by RT-qPCR at indicated timepoints. Fold change was calculated based on the DMSO control of the respective timepoint. In synchronized cells, the TCDD-mediated CYP1A1 induction displays a circadian pattern peaking at 24 h. Each bar represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed. Statistical significance was determined by using an unpaired, one-tailed t-test. c) To determine the knock down efficiency of the targeted genes, the mRNA levels of the indicated genes of interest (GOI) were analyzed by RT-qPCR. The knockdown level was calculated by comparing the GOI with the control siRNA transfected cells. Each bar represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed. d) Decrease and increase of CYP1A1 induction was determined in M13SV1 cells transfected with siRNA targeting the AHR or its co-factors ARNT, XAP2, HSP90 and p23. The transfected cells were synchronized with 1 µM dexamethasone or left non-synchronized and subsequently exposed to 0.5 nM TCDD for 24 h. CYP1A1 mRNA levels were determined by RT-qPCR and the change in induction was calculated for the siRNA-targeted gene compared to the control siRNA. Each bar represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed.

Since the M13SV1 cells were clearly appropriate for our research, we transfected them with siRNAs targeting AHR and its co-factors ARNT, XAP2, HSP90 and p23 (Fig. 3b, S3b). Following transfection, the cells were synchronized or left non-synchronized and exposed to DMSO (solvent) or 0.5 nM TCDD for 24 h. The CYP1A1 induction was analyzed by RT-qPCR determining the mRNA level of CYP1A1 for the different conditions (Fig. S3c). To identify co-factors affecting CYP1A1 induction in synchronized cells, the change in induction of control siRNA vs genes of interest (GOI) targeting siRNAs was calculated (Fig. 3c).

As expected, the cells displayed a noticeable induction of CYP1A1 when treated with TCDD. However, due to siRNA mediated repression of the essential elements of this transcriptional pathway, namely AHR, ARNT and XAP2, the observed induction exhibited a significant decrease.

Notably, comparing the observed effects in synchronized and non-synchronized cells, the reduced induction is more pronounced under synchronous conditions. It is reasonable that these co-factors hold a generic role in this pathway, resulting in a more pronounced reduction when the level of CYP1A1 mRNA is reduced to an equivalent level in both conditions. The partial depletion of HSP90 had no effect on the induction of CYP1A1.

Considering the established role of p23 as a co-chaperone within this specific pathway, it was hypothesized that its repression would yield a comparable outcome to that of HSP90 on CYP1A1 induction. Surprisingly, the induction of CYP1A1 was found to be twofold higher in cells that had been synchronized and exhibited reduced p23 expression. In contrast, in non-synchronized cells, siRNA-mediated p23 repression resulted in a decrease in CYP1A1 induction. These results indicate that p23 may act as a distinctive negative regulator of circadian controlled CYP1A1 induction.

SP1 and p23 modulate circadian associated AHR activity

Prior experiments revealed two candidate genes that may impact the circadian modulated regulation of CYP1A1 induction. One promising candidate was the transcription factor SP1, which exhibited a circadian pattern at protein levels peaking at the same time point as CYP1A1. The second one was the co-chaperone p23 that uniquely regulated circadian modulated CYP1A1 induction, although the p23 expression levels were not circadian regulated. To investigate how these two candidate genes modulate CYP1A1 induction upon TCDD exposure in synchronized cells, we transiently transfected M13SV1 cells with either siRNA targeting p23 and SP1 or with a plasmid for ectopic expression of p23-GFP and SP1 (Fig. S4a, S4b). The transfected cells were subsequently synchronized and exposed to increasing doses of TCDD as indicated.

Remarkably, not only the reduction of p23 expression by siRNA knock down resulted in an enhanced stimulation of CYP1A1 expression upon TCDD treatment in a dose-dependent manner (Fig. 4a). Vice versa, the overexpression of p23 by transfection with a p23 expression vector led to a diminished dose response (Fig. 4b). Interestingly, our results for SP1 were reversed. The repression of SP1 led to a reduced dose response (Fig. 4c), while overexpression resulted in an increased dose response (Fig. 4d). A detailed analysis of mRNA expression revealed that the dynamic range of cellular response is modulated by alterations in p23 expression. A repression of p23 expended the dynamic response range whereas the overexpression of p23 reduced the dynamic response range (Fig. 4e and f). Interestingly, we demonstrated in the earlier study, that circadian synchronized cells exhibit a higher dynamic response range in comparison to non-synchronized cells (Ndikung et al. 2020). The role of SP1 in the circadian modulation of CYP1A1 appears to be similar, albeit with a significantly lower effect level (Fig. 4g and h). These findings suggest that SP1 and p23 are counter-regulators of circadian-driven CYP1A1 induction upon exposure to AHR ligands, such as TCDD.

Fig. 4

SP1 and p23 modulate circadian associated AHR activity: a) Suppression of p23 increases the M13SV1 cell response to TCDD. M13SV1 cells were transiently transfected with scramble (SCR) or p23 siRNA. After 48 h of transfection, the cells were synchronized and treated for 24 h with different concentrations of TCDD (0, 0.1, 0.25, 0.5, 1 and 2 nM). The CYP1A1 mRNA expression was determined by RT-qPCR and the fold change for each TCDD concentration was calculated based on the respective DMSO control. Each point represents the mean ± SD of three independent experiments. b) Overexpression of p23 decreases the M13SV1 cell response to TCDD. M13SV1 cells were transfected via electroporation with empty vector or p23-GFP plasmid. After 48 h of transfection, they were synchronized and treated for 24 h with different concentrations of TCDD (0, 0.1, 0.25, 0.5, 1 and 2 nM). The CYP1A1 mRNA expression was determined by RT-qPCR and the fold change for each TCDD concentration was calculated based on the respective DMSO control. Each point represents the mean ± SD of three independent experiments. c) Suppression of SP1 decreases the M13SV1 cell response to TCDD. M13SV1 cells were transfected with scramble (SCR) or SP1 siRNA. After 48 h of transfection, they were synchronized and treated for 24 h with different concentrations of TCDD (0, 0.1, 0.25, 0.5, 1 and 2 nM). The CYP1A1 mRNA expression was determined by RT-qPCR and the fold change for each TCDD concentration was calculated based on the respective DMSO control. Each point represents the mean ± SD of three independent experiments. d) Overexpression of SP1 enhances the M13SV1 cell response to TCDD. M13SV1 cells were transfected via electroporation with empty vector or SP1 overexpression plasmid. After 48 h of transfection, they were synchronized and treated for 24 h with different concentrations of TCDD (0, 0.1, 0.25, 0.5, 1 and 2 nM). The CYP1A1 mRNA expression was determined by RT-qPCR and the fold change for each TCDD concentration was calculated based on the respective DMSO control. Each point represents the mean ± SD of three independent experiments. e) Relative CYP1A1 mRNA expression in TCDD treated synchronized cells transfected with SCR or p23 siRNA. 48 h after transfection, M13SV1 cells were synchronized with dexamethasone (1 µM) for 1 h and subsequently treated with different TCDD concentration (0, 0.1, 0.25, 0.5, 1 and 2 nM) for 24 h. The mRNA levels of CYP1A1 were analyzed by RT-qPCR and normalized to the endogenous B2M. Each bar represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed. f) Relative CYP1A1 mRNA expression in TCDD treated synchronized cells transfected with empty vector or p23-GFP plasmid DNA. 48 h after transfection, M13SV1 cells were synchronized with dexamethasone (1 µM) for 1 h and subsequently treated with different TCDD concentration (0, 0.1, 0.25, 0.5, 1 and 2 nM) for 24 h. The mRNA levels of CYP1A1 were analyzed by RT-qPCR and normalized to the endogenous B2M. Each bar represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed. g) Relative CYP1A1 mRNA expression in TCDD treated synchronized cells transfected with SCR or SP1 siRNA. 48 h after transfection, M13SV1 cells were synchronized with dexamethasone (1 µM) for 1 h and subsequently treated with different TCDD concentration (0, 0.1, 0.25, 0.5, 1 and 2 nM) for 24 h. The mRNA levels of CYP1A1 were analyzed by RT-qPCR and normalized to the endogenous B2M. Each bar represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed. h) Relative CYP1A1 mRNA expression in TCDD treated synchronized cells transfected with empty vector or SP1 plasmid DNA. 48 h after transfection, M13SV1 cells were synchronized with dexamethasone (1 µM) for 1 h and subsequently treated with different TCDD concentration (0, 0.1, 0.25, 0.5, 1 and 2 nM) for 24 h. The mRNA levels of CYP1A1 were analyzed by RT-qPCR and normalized to the endogenous B2M. Each bar represents the mean ± SD of three independent experiments. The individual data points from the single experiments are displayed

To ensure that the repression or overexpression of SP1 and p23 did not interfere with the circadian synchrony per se, we transfected the circadian reporter cells, hTERT-HME1-PER2, with the candidates and monitored the circadian rhythm for 84 h. As both candidates have not been described as circadian regulators before, we did not expect any impairment of circadian synchrony. Indeed, the analysis of the bioluminescent signal of the transiently transfected reporter plasmids confirmed that there were no changes in the circadian rhythm following the repression or overexpression of SP1 and p23 (Fig. S4c).

Finally, to exclude the possibility that the repression or overexpression of p23 or SP1 interferes per se with the dose-dependent CYP1A1 induction, non-synchronous cells were treated with 0.5 nM and 2 nM TCDD for 24 h, and the CYP1A1 mRNA induction was determined by qPCR. In contrast to the synchronous cell, no significant alterations in CYP1A1 expression were observed due to suppression or overexpression of SP1 and p23 (Fig. S4d).

AHR interaction with the clock core complex BMAL1/CLOCK

In our previous experiments, we demonstrated the circadian modulated regulation of AHR target gene induction and revealed SP1 and p23 as central regulators of the circadian-driven transcriptional activity of AHR upon TCDD treatment. However, a direct link of the AHR pathway and circadian machinery is still missing, since AHR and its major co-factors seem not to be direct clock control genes in human breast cells. Thus, we hypothesized that the regulation is achieved by a protein-protein interaction between the clock machinery and the AHR.

In order to address this, interaction studies were conducted by immunoprecipitating endogenous AHR of circadian synchronized HME1 cells treated with DMSO or TCDD and comparing it with non-synchronous cells also treated with DMSO and TCDD. In detail, HME1 cells were synchronized with 1 µM dexamethasone for 1 h or left non-synchronized, cultivated for 23 h and subsequently exposed to 2 nM TCDD for additional 1 h. This sequential treatment scenario ensures optimal conditions for interaction studies that favor the detection of directly induced complexes and minimize possible secondary, compensatory effects as well as steady-state formation that may occurs in long-term treated cells. The interaction partners of AHR were identified and analyzed by Western blotting. As expected, a clear increased interaction between AHR and ARNT could be observed in TCDD-treated cells (Fig. 5a). This finding served as a positive control for synchronized and non-synchronized conditions to ensure that the precipitated AHR complex was active and reflected the expected results. The observation of p23 not binding to AHR, even under the control conditions, is not unexpected, given that p23 is only bound to HSP90 in the cytosolic complex. In the case of HSP90, a strong interaction could not be observed under all conditions. However, the TCDD-treated synchronized cells showed a modest increase in the interaction of AHR and HSP90.

Fig. 5

The clock core complex BMAL1/CLOCK directly interacts with the AHR/ARNT complex. HME1 cells were synchronized with 1 µM dexamethasone for 1 h and after 23 h cultivation exposed to 2 nM TCDD for additional 1 h. a) Immunoprecipitation of endogenous AHR. AHR was immunoprecipitated from whole-cell lysates and AHR, ARNT, BMAL1, CLOCK, SP1, p23, HSP90 and GAPDH (as control) were detected on Western blots. Representative Western blots are shown from independent experiments. b) Immunoprecipitation of endogenous BMAL1. BMAL1 was immunoprecipitated from whole-cell lysates and BMAL1, CLOCK, AHR, ARNT, SP1, p23, HSP90 and GAPDH (as control) were detected on Western blots. c) Immunoprecipitation of endogenous CLOCK. CLOCK was immunoprecipitated from whole-cell lysates and CLOCK, BMAL1, AHR, ARNT, SP1, HSP90 and GAPDH (as control) were detected on Western blots. Representative Western blots are shown from independent experiments. For all interaction studies, an antibody mock control was included to determine unspecific bands from the antibody solution.

Interestingly, we identified a very weak interaction between AHR and the core clock protein BMAL1 and its primary partner CLOCK in synchronized cells.

Nevertheless, the interaction between BMAL1/CLOCK does not appear to be particularly strong, and in the case of BMAL1, does not depend on the treatment with TCDD. The situation is different with CLOCK, where the interaction seems to disappear upon TCDD treatment. It is important to note that the interactions in the non-synchronized cells are even weaker than those in the synchronized cells. In particular for CLOCK, we could not find a stronger interaction with AHR when treated with DMSO.

In order to conduct a more profound analysis of the interaction complex and to verify the findings of the AHR interaction study, a reverse study was conducted with BMAL1. To this end, HME1 cells were synchronized or left non-synchronized and treated as previously described. Subsequently, an immunoprecipitation of the endogenous BMAL1 was performed. As interaction control for an active complex, we detected CLOCK, the major co-factor of BMAL1 (Fig. 5b). Indeed, AHR and ARNT were clearly identified in this complex in circadian synchronized cells, however, with no interaction difference, particular with ARNT, between DMSO and TCDD treated cells. It is important to note that this interaction between BMAL1 and AHR/ARNT is not detectable in non-synchronized cells. Interestingly, in this setup we could also identify SP1 as a player in this network, while no interaction with p23 or HSP90 could be detected.

To further elucidate the composition of the circadian complex, we finally conducted a CLOCK immunoprecipitation and confirmed the presence of the relevant players (Fig. 5c). The interaction of major co-factor BMAL1 served again as control for an active circadian complex. Furthermore, a clear interaction was observed between CLOCK and AHR, as well as CLOCK and SP1, while no interaction was identified with ARNT. In accordance with the findings of previous interaction studies of AHR and BMAL1, the interaction of CLOCK with AHR and ARNT was found to be significantly weaker in non-synchronized cells. As previously noted, no TCDD-dependent change was observed in the composition of these complexes. It is noteworthy that we observe a clear interaction of CLOCK with HSP90 in synchronized cells.

Taken together, these findings indicate a potential interaction between an active circadian BMAL1/CLOCK complex and the AHR/ARNT complex, particularly in synchronized cells. This interaction could function as a direct connection between the clock machinery and the AHR pathway. The BMAL1/CLOCK core complex appears to function as a molecular bridge between the circadian-regulated SP1 transcription factor and the AHR/ARNT response complex, thereby maintaining the circadian expression of target genes such as CYP1A1.