1. Cell Biology
  2. Genetics and Genomics

The nutrient-sensing GCN2 signaling pathway is essential for circadian clock function by regulating histone acetylation under amino acid starvation

  1. Xiao-Lan Liu
  2. Yulin Yang
  3. Yue Hu
  4. Jingjing Wu
  5. Chuqiao Han
  6. Qiaojia Lu
  7. Xihui Gan
  8. Shaohua Qi
  9. Jinhu Guo
  10. Qun He
  11. Yi Liu
  12. Xiao Liu  Is a corresponding author
  1. State Key Laboratory of Mycology, Institute of Microbiology, Chinese Academy of Sciences, China
  2. College of Life Sciences, University of the Chinese Academy of Sciences, China
  3. School of Life Sciences, Yunnan University, Kunming, China
  4. Key Laboratory of Gene Engineering of the Ministry of Education, State Key Laboratory of Biocontrol, School of Life Sciences, Sun Yat-sen University, China
  5. MOA Key Laboratory of Soil Microbiology, College of Biological Sciences, China Agricultural University, China
  6. Department of Physiology, University of Texas Southwestern Medical Center, United States
https://doi.org/10.7554/eLife.85241
Share this article
Cite this article
  1. Xiao-Lan Liu
  2. Yulin Yang
  3. Yue Hu
  4. Jingjing Wu
  5. Chuqiao Han
  6. Qiaojia Lu
  7. Xihui Gan
  8. Shaohua Qi
  9. Jinhu Guo
  10. Qun He
  11. Yi Liu
  12. Xiao Liu
(2023)
The nutrient-sensing GCN2 signaling pathway is essential for circadian clock function by regulating histone acetylation under amino acid starvation
eLife 12:e85241.
https://doi.org/10.7554/eLife.85241
  2. Download BibTeX
  3. Download .RIS

Abstract

Circadian clocks are evolved to adapt to the daily environmental changes under different conditions. The ability to maintain circadian clock functions in response to various stresses and perturbations is important for organismal fitness. Here, we show that the nutrient-sensing GCN2 signaling pathway is required for robust circadian clock function under amino acid starvation in Neurospora. The deletion of GCN2 pathway components disrupts rhythmic transcription of clock gene frq by suppressing WC complex binding at the frq promoter due to its reduced histone H3 acetylation levels. Under amino acid starvation, the activation of GCN2 kinase and its downstream transcription factor CPC-1 establish a proper chromatin state at the frq promoter by recruiting the histone acetyltransferase GCN-5. The arrhythmic phenotype of the GCN2 kinase mutants under amino acid starvation can be rescued by inhibiting histone deacetylation. Finally, genome-wide transcriptional analysis indicates that the GCN2 signaling pathway maintains robust rhythmic expression of metabolic genes under amino acid starvation. Together, these results uncover an essential role of the GCN2 signaling pathway in maintaining the robust circadian clock function in response to amino acid starvation, and demonstrate the importance of histone acetylation at the frq locus in rhythmic gene expression.

Editor's evaluation

This fundamental work is important in demonstrating that the general amino acid control response to amino acid limitation in Neurospora, which includes the key nutrient-controlled protein kinase Gcn2, is crucial to maintain circadian rhythmic cell growth and transcription of the FRQ gene, the master regulator of rhythmicity. There is an abundance of compelling evidence supporting the conclusions, with rigorous molecular and genetic assays of key mutants impaired for general amino acid control or transcriptional cofactors. The work will be of broad interest to geneticists and molecular biologists, and will be particularly valuable to researchers interested in circadian rhythm or nutrient control of gene expression.

https://doi.org/10.7554/eLife.85241.sa0

Introduction

Circadian clocks enable organisms to adapt to the daily environmental changes caused by the earth’s rotation (Bell-Pedersen et al., 2005; Dunlap and Loros, 2017; Johnson et al., 2017; Takahashi, 2017). Rhythmic gene expression allows different organisms to regulate their daily molecular, cellular, behavioral, and physiological activities. The ability to maintain circadian clock function in response to various stresses and perturbations is an important property of living systems (Bass, 2012; Hogenesch and Ueda, 2011). Although gene expression is sensitive to temperature changes, temperature compensation is a key feature of circadian clocks to maintain circadian period length at different physiological temperatures (Hu et al., 2021; Narasimamurthy and Virshup, 2021; Ode and Ueda, 2018). DNA damage and translation stress are known to reset the circadian clock through the checkpoint kinase 2 signaling pathway in Neurospora and the ATM signaling pathway in mammalian cells (Diernfellner et al., 2019; Oklejewicz et al., 2008; Pregueiro et al., 2006). Cellular redox balance, including oxidative stress, regulates the circadian clock by modulating CLOCK and NPAS2 activity (Nakahata et al., 2008; Rutter et al., 2001). Nutritional stress, such as a high-fat diet, can disrupt the oscillating metabolites and behavioral rhythms in mice (Eckel-Mahan et al., 2013; Kohsaka et al., 2007; Panda, 2016).

Starvation for all or certain amino acids leads to induced transcription followed by derepression of genes in many amino acid biosynthetic pathways, referred as general amino acid control in yeast and cross-pathway control (CPC) in Neurospora (Hinnebusch, 2005). General control nonderepressible 2 (GCN2) kinase, called CPC-3 in Neurospora, is a serine/threonine kinase that functions as an amino acid sensor (Battu et al., 2017; Efeyan et al., 2015; Sattlegger et al., 1998). GCN2 is activated by accumulated uncharged tRNAs when intracellular amino acids are limited (Ramirez et al., 1992; Wek et al., 1995). Activated GCN2 phosphorylates the α subunit of eukaryotic initiation factor 2 (eIF2α) to translationally repress protein synthesis (Lyu et al., 2021; Sonenberg and Hinnebusch, 2009). Meanwhile, it also upregulates the transcription activator GCN4, named CPC-1 in Neurospora, which activates amino acid biosynthetic and transport pathways (Ebbole et al., 1991; Hinnebusch, 2005). Recently, it has been shown that circadian clock control of the GCN2-mediated eIF2α phosphorylation is necessary for rhythmic translation initiation in Neurospora (Ding et al., 2021; Karki et al., 2020). Although amino acid starvation is known to activate the GCN2–GCN4 signaling pathway, how nutrient limitation, especially amino acid starvation, affects circadian clock is not known.

Despite evolutionary divergence in eukaryotes, circadian rhythms are controlled by the conserved transcription/translation-based negative feedback loops (Bell-Pedersen et al., 2005). Neurospora crassa has been established as one of the best studied model systems for analyzing the molecular mechanism of eukaryotic circadian clocks (Cha et al., 2015; Dunlap and Loros, 2017; Heintzen and Liu, 2007). In the Neurospora core circadian negative feedback loop, two PAS-domain-containing transcription factors, White Collar-1 (WC-1) and WC-2 form a complex (WCC) and bind to the C-box of the frq promoter to activate its transcription (Cheng et al., 2001b; Crosthwaite et al., 1997; Froehlich et al., 2002). FRQ protein is translated from frq mRNA in the cytosol and progressively phosphorylated at about 103 phosphorylation sites, which plays a major role in determining circadian periodicity by regulating the FRQ–CK1 interaction (Baker et al., 2009; Chen et al., 2023; Larrondo et al., 2015; Liu et al., 2019; Liu et al., 2000; Tang et al., 2009). To close the negative feedback loop, FRQ forms a complex with its partner FRQ-interacting RNA helicase (FRH) to inhibit the activity of the WCC by promoting WCC phosphorylation mediated by CK1 and CK2 (He et al., 2006; He and Liu, 2005; Schafmeier et al., 2005; Wang et al., 2019).

Chromatin structure and histone modification changes play important roles in regulating the transcription of circadian clock genes (Papazyan et al., 2016; Takahashi, 2017; Zhu and Belden, 2020). In mammals, rhythmic H3K4me3, H3K9ac, and H3K27ac modifications have been shown to be enriched at the promoter of clock genes and positively correlated with gene expression (Etchegaray et al., 2003; Katada and Sassone-Corsi, 2010; Koike et al., 2012). In Neurospora, rhythmic binding of WCC to the frq promoter is regulated by ATP-dependent chromatin remodeling factors, such as the SWI/SNF complex, the INO80 complex, the chromodomain helicase DNA-binding-1 (CHD1), and the Clock ATPase (CATP) (Belden et al., 2011; Cha et al., 2013; Gai et al., 2017; Wang et al., 2014). Histone chaperone FACT complex and histone modification enzymes, SET1, SET2, and RPD3 complex, have also been shown to affect frq transcription by regulating rhythmic histone compositions and modifications at the frq locus (Liu et al., 2017; Raduwan et al., 2013; Sun et al., 2016). However, it is still unknown how the chromatin structure is organized to allow rhythmic clock gene expression under nutrient limitation conditions.

In this study, we discovered that the disruption of the GCN2 (CPC-3) signaling pathway abolished robust circadian rhythms under amino acid starvation, which was important for rhythmic expression of metabolic genes. In the GCN2 signaling pathway mutants, amino acid starvation abolished the rhythmic binding of WC-2 at the frq promoter by decreasing its histone H3 acetylation levels. Amino acid starvation activated CPC-3 and CPC-1, which re-established a proper chromatin state at the frq promoter by recruiting the histone acetyltransferase GCN-5 to allow rhythmic frq expression. Furthermore, the inhibition of histone deacetylases could rescue the impaired circadian rhythm phenotypes under amino acid starvation, demonstrating the importance of rhythmic histone acetylation at the frq gene promoter for maintaining robust circadian rhythms of gene expression.

Results

CPC-3 and CPC-1 are required for robust circadian rhythms under amino acid starvation

To investigate whether the nutrient-sensing GCN2 pathway is involved in regulating circadian clock function under amino acid starvation, we created the cpc-3 and cpc-1 knockout mutants (see Materials and methods). In Neurospora, cpc-3 and cpc-1 encode for the GCN2 and GCN4 homolog, respectively. As expected, the CPC-3-mediated eIF2α phosphorylation and CPC-1 induction by 3-aminotriazole (3-AT) treatment were completely abolished in the cpc-3KO strain (Figure 1—figure supplement 1A). 3-AT is an inhibitor of the histidine synthesis enzyme encoded by his-3 which triggers the amino acid starvation response (Natarajan et al., 2001). On the other hand, CPC-1 expression and its induction by 3-AT were eliminated in the cpc-1KO mutant (Figure 1—figure supplement 1B). The circadian conidiation rhythms of the cpc-3KO and cpc-1KO mutants were examined by race tube assays. Under a normal growth condition (0 mM 3-AT), cpc-3 deletion had no effect on the circadian period but cpc-1 deletion led to the period lengthening of ~1.7 hr (Figure 1A). The long period of cpc-1KO strain could be rescued by the expression of Myc.CPC-1 (Figure 1—figure supplement 1C).

Figure 1 with 2 supplements see all
Download asset Open asset
CPC-3 and CPC-1 are required for circadian rhythm by regulating rhythmic frq transcription in response to amino acid starvation.

(A) Race tube assay showing that amino acid starvation (3-aminotriazole [3-AT] treatment) disrupted circadian conidiation rhythm of the cpc-3KO and cpc-1KO strains. 0 mM 3-AT is the normal growth condition. (B) Luciferase reporter assay showing that amino acid starvation disrupted rhythmic expression of frq promoter-driven luciferase of the cpc-3KO and cpc-1KO strains. A frq-luc transcriptional fusion construct was expressed in cpc-3KO and cpc-1KO strains grown on the fructose-glucose-sucrose FGS-Vogel’s medium with the indicated concentrations of 3-AT, and the luciferase signal was recorded using a LumiCycle in constant darkness (DD) for more than 7 days. Normalized data with the baseline luciferase signals subtracted are shown. (C) Western blot showing that amino acid starvation dampened rhythmic expression of FRQ protein of the cpc-3KO and cpc-1KO strains at the indicated time points in DD (n = 3; WT: p = 5.00E−08, cpc-3KO: p = 0.0016, cpc-1KO: p = 0.0004, RAIN; WT vs cpc-3KO: mesor p = 0.1421, amplitude p = 0.0774, phase p = 0.4319; WT vs cpc-1KO: mesor p = 0.0614, amplitude p = 0.1920, phase p = 0.4404, CircaCompare). The left panel showing that protein extracts were isolated from WT, cpc-3KO, and cpc-1KO strains grown in a circadian time course in DD and probed with FRQ antibody. The right panel showing that the densitometric analyses of the results of three independent experiments. (D) Northern blot showing that amino acid starvation dampened rhythmic expression of frq mRNA of the cpc-3KO and cpc-1KO strains at the indicated time points in DD (n = 3; WT: p = 6.56E−05, cpc-3KO: p = 0.0039, cpc-1KO: p > 0.05, RAIN; WT vs cpc-3KO: mesor p = 0.1153, amplitude p = 0.4316, phase p = 0.0788, CircaCompare). The densitometric analyses of the results from three independent experiments were shown on the right panel. Error bars indicate standard deviation (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001; Student’s t test was used.

Figure 1—source data 1

LumiCycle analysis dataset in Figure 1B.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig1-data1-v2.zip
Figure 1—source data 2

Scan of western blot probed for FRQ protein and quantification dataset in Figure 1C.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig1-data2-v2.zip
Figure 1—source data 3

Scan of Northern blot probed for frq mRNA and quantification dataset in Figure 1D.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig1-data3-v2.zip

To investigate how amino acid starvation affects circadian clock, we treated the wild-type (WT), cpc-3KO, and cpc-1KO strains with different concentrations of 3-AT. As shown in Figure 1A, although treatment with 3 or 4 mM 3-AT resulted in modest inhibition of the WT growth rate, the robust circadian conidiation rhythms were maintained. In the cpc-3KO and cpc-1KO strains, however, addition of 3 or 4 mM 3-AT resulted in severe inhibition of growth rates and the loss of circadian conidiation rhythms. To exclude the effect of 3-AT on other target genes, we examined the circadian rhythm of the his-3− strain, which contained a single mutation in the his-3 gene required for histidine synthesis and could not grow in the medium without histidine. Race tube assays showed that the his-3− strain grew normally and exhibited a robust circadian conidiation rhythm in the presence of histidine (1.0 × 10−2 mg/ml). Although addition of the same amount of histidine could rescue the growth of the cpc-3KO his-3− strain, it could not rescue its circadian conidiation rhythm (Figure 1—figure supplement 1D), indicating that CPC-3 is required for robust circadian rhythms under histidine starvation stress.

To confirm the loss of circadian rhythms at the molecular level, we introduced a frq promoter-driven luciferase reporter into the cpc-3KO and cpc-1KO strains. As shown in Figure 1B and Figure 1—figure supplement 2A, B, the robust circadian rhythms of luciferase activity seen in the WT strain were severely dampened or arrhythmic in the cpc-3KO and cpc-1KO strains upon 3 mM 3-AT treatment. Consistent with these results, western blot analysis showed that, after the initial light/dark transition, rhythms of FRQ levels and its phosphorylation were dampened in the cpc-3KO and cpc-1KO strains in the presence of 3-AT (WT: p = 5.00E−08, cpc-3KO: p = 0.0016, cpc-1KO: p = 0.0004) (Figure 1C and Figure 1—figure supplement 2C). The statistical tests of circadian rhythms were performed using a circadian statistical analysis tool CircaCompare (Parsons et al., 2020) (see Materials and methods). Northern blot analysis showed that the circadian rhythms of frq mRNA in the cpc-3KO and cpc-1KO strains were also dampened in the presence of 3-AT (WT: p = 6.56E−05, cpc-3KO: p = 0.0039, cpc-1KO: p > 0.05) and the levels of frq mRNA were constantly low in DD (Figure 1D and Figure 1—figure supplement 2D). Together, these results suggest that the GCN2 pathway is required for a functional clock by regulating rhythmic frq transcription in response to amino acid starvation.

CPC-3 and CPC-1 are required for rhythmic WCC binding in response to amino acid starvation

Since 3-AT treatment resulted in low frq mRNA levels in the cpc-3KO and cpc-1KO strains (Figure 1D), we first examined the protein levels of WC-1 and WC-2 and found that their levels were higher in the mutants than those in the WT strain at different time points (Figure 2A, B). We then performed WC-2 chromatin immunoprecipitation (ChIP) assays to examine whether the WCC binding to the frq promoter was affected. As shown in Figure 2—figure supplement 1A, WC-2 rhythmically bound to the frq C-box in the WT, cpc-3KO, and cpc-1KO strains without 3-AT treatment (WT: p = 2.60E−07, cpc-3KO: p = 0.0016, cpc-1KO: p = 1.06E−05). However, 3-AT treatment resulted in constant low levels of WC-2 binding to the frq C-box during a circadian cycle in both cpc-3KO and cpc-1KO strains (Figure 2C, D). These results indicate that the loss of circadian rhythms in the cpc-3KO and cpc-1KO strains under amino acid starvation is caused by loss of rhythmic frq transcription, due to impaired WCC binding at the frq promoter.

Figure 2 with 1 supplement see all
Download asset Open asset
CPC-3 and CPC-1 are required for rhythmic WCC binding in response to amino acid starvation.

Western blot assay showing that WCC protein levels were elevated in the cpc-3KO (A) and cpc-1KO (B) strains after 3 mM 3-aminotriazole (3-AT) treatment. Protein extracts were isolated from WT, cpc-3KO, and cpc-1KO strains grown in the indicated time points in DD and probed with WC-1 and WC-2 antibodies. Chromatin immunoprecipitation (ChIP) assay showing that amino acid starvation disrupted rhythmic WC-2 binding at the promoter of frq gene in the cpc-3KO (n = 3; WT: p = 1.84E−05, cpc-3KO: p > 0.05) (C) or cpc-1KO strains (n = 3; WT: p = 0.0025, cpc-1KO: p > 0.05) (D). Samples were grown for the indicated number of hours in DD prior to harvesting and processing for ChIP using WC-2 antibody. Occupancies were normalized by the ratio of ChIP to Input DNA. Error bars indicate standard deviation (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001; Student’s t test was used.

Figure 2—source data 1

Scan of western blot probed for WC-1 and WC-2 proteins in WT and cpc-3KO strains with 3-aminotriazole (3-AT) in Figure 2A.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig2-data1-v2.zip
Figure 2—source data 2

Scan of western blot probed for WC-1 and WC-2 proteins in WT and cpc-1KO strains with 3-aminotriazole (3-AT) in Figure 2B.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig2-data2-v2.zip
Figure 2—source data 3

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 2C.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig2-data3-v2.zip
Figure 2—source data 4

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 2D.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig2-data4-v2.zip

Because WCC phosphorylation inhibited its transcriptional activation activity (He et al., 2006; He and Liu, 2005; Lee et al., 2000; Schafmeier et al., 2005; Wang et al., 2019), we also examined WCC phosphorylation profiles and found that 3-AT treatment resulted in hypophosphorylation of WC-1 and WC-2 (which is usually associated with WCC activation) in the cpc-3KO and cpc-1KO strains (Figure 2—figure supplement 1B, C). Thus, their reduced WCC binding at the frq promoter is not caused by WCC hyperphosphorylation. It should be noted that the overall phosphorylation status of WCC does not always reflect its activity in driving frq transcription, which is possibly due to the unknown function of multiple key phosphosites on WCC (Wang et al., 2019; Zhou et al., 2018).

CPC-1 is required for the maintenance of chromatin structure in response to amino acid starvation

The low WC-2 binding at the frq promoter prompted us to examine the chromatin structure of the frq promoter. We first performed ChIP assay to examine the histone and its acetylation levels at the frq promoter in the WT strain. Although amino acid starvation had little effect on the histone H2B levels (Figure 3A), it led to significantly decreased histone H3 acetylation levels (H3 acetylated at the N-terminus) at the frq promoter at high concentrations of 3-AT (Figure 3B). These results suggest that amino acid starvation can affect frq transcription by reducing the histone acetylation levels at the frq promoter.

Figure 3
Download asset Open asset
CPC-1 is required for the maintenance of chromatin structure in response to amino acid starvation.

Chromatin immunoprecipitation (ChIP) assay showing that amino acid starvation slightly increased histone H2B levels (A) and dramatically decreased histone H3ac levels (B) at the promoter of frq gene in the WT strain at DD18 at the indicated concentration of 3-aminotriazole (3-AT). Relative H3ac levels were normalized with H2B levels. (C, D) ChIP assay showing that amino acid starvation slightly increased histone H2B levels (n = 3; WT: p = 3.85E−04, cpc-1KO: p = 0.0364, RAIN; WT vs cpc-1KO: mesor p = 0.0312, amplitude p = 0.2155, phase p = 0.2995, CircaCompare) (C) and dramatically decreased histone H3ac levels (n = 3; WT: p = 0.0168, cpc-1KO: p > 0.05) (D) at the promoter of frq gene in the cpc-1KO strain at the indicated time points in DD. Error bars indicate standard deviations (n = 3). *p < 0.05; ***p < 0.001; Student’s t test was used.

Figure 3—source data 1

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 3A.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig3-data1-v2.zip
Figure 3—source data 2

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 3B.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig3-data2-v2.zip
Figure 3—source data 3

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 3C.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig3-data3-v2.zip
Figure 3—source data 4

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 3D.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig3-data4-v2.zip

We then examined whether the CPC-3 and CPC-1 signaling pathway was involved in regulating chromatin structure at the frq promoter in response to amino acid starvation. Histone H2B and H3ac ChIP assays at different circadian time points showed that the relative histone H3ac levels were slightly decreased in the cpc-1KO strain compared to the WT strain under normal condition (Figure 2—figure supplement 1D). H2B levels were not markedly different between the WT and cpc-1KO strains in the presence of 3 mM 3-AT (Figure 3C). However, the relative histone H3ac levels were very different in these two strains in the presence of 3 mM 3-AT: it was rhythmic with a peak at DD18 in the WT strain but was dramatically reduced and arrhythmic in the cpc-1KO strain at different time points in DD (WT: p = 0.0168, cpc-1KO: p > 0.05) (Figure 3D). These results indicate that CPC-1 is required for maintaining the proper histone acetylation status at the frq promoter under amino acid starvation. The low H3ac levels at the frq promoter, which is critical for transcription activation, results in constant low WCC binding and arrhythmic frq transcription in the cpc-1KO strain.

CPC-1 recruits GCN-5 to activate frq transcription in response to amino acid starvation

To determine how CPC-1 is involved in regulating histone acetylation levels at the frq locus, we examined the occupancy of CPC-1 at the frq promoter by ChIP assays using CPC-1 antibody. As shown in Figure 4A and Figure 4—figure supplement 1A, CPC-1 was found to be rhythmically enriched at the frq promoter in the WT strain but not in the cpc-1KO strain under normal (WT: p = 8.37E−06, cpc-1KO: p > 0.05) or amino acid starvation (WT: p = 7.64E−05) conditions in DD, peaking at ~DD14, a time point corresponding to the peak of frq mRNA levels. Co-immunoprecipitation (Co-IP) assay showed that CPC-1 did not associate with WC-1 or WC-2 (Figure 4—figure supplement 1B), suggesting that CPC-1 and WCC bind independently to the frq promoter.

Figure 4 with 1 supplement see all
Download asset Open asset
CPC-1 recruits GCN-5 to activate frq transcription in response to amino acid starvation.

(A) Chromatin immunoprecipitation (ChIP) assay showing that CPC-1 rhythmically bound at the promoter of frq gene (n = 3; WT: p = 8.37E−06, cpc-1KO: p > 0.05). WT and cpc-1KO strains grown for the indicated number of hours in DD. Samples were crosslinked with formaldehyde and harvested for ChIP using CPC-1 antibody. CPC-1 ChIP occupancies were normalized by the ratio of ChIP to Input DNA. (B) Co-immunoprecipitation (Co-IP) assay showing that Flag.ADA-2 interacted with Myc.GCN-5 with or without 3 mM 3-aminotriazole (3-AT). Flag.ADA-2 and Myc.GCN-5 were co-expressed in the WT strain and immunoprecipitation was performed using Flag antibody. (C) Co-IP assay showing that Flag.ADA-2 interacted with Myc.CPC-1 with or without 3 mM 3-AT. Flag.ADA-2 and Myc.CPC-1 were co-expressed in the WT strain and immunoprecipitation was performed using Flag antibody. (D) ChIP assay showing that rhythmic GCN-5 binding at the promoter of frq gene was dampened in the cpc-1KO strain (n = 3; WT: p = 0.0006, cpc-1KO: p > 0.05). Samples were grown for the indicated number of hours in DD prior to harvesting and processing for ChIP as described in (A). Error bars indicate standard deviations (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001; Student’s t test was used.

Figure 4—source data 1

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 4A.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig4-data1-v2.zip
Figure 4—source data 2

Scan of western blot probed for Flag.ADA-2 and Myc.GCN-5 proteins in Figure 4B.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig4-data2-v2.zip
Figure 4—source data 3

Scan of western blot probed for Flag.ADA-2 and Myc.CPC-1 proteins in Figure 4C.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig4-data3-v2.zip
Figure 4—source data 4

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 4D.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig4-data4-v2.zip

How does the GCN2 signaling pathway regulate histone acetylation in response to amino acid starvation? The yeast GCN4 was previously shown to recruit the histone acetyltransferase GCN5 containing (Spt-Ada-Gcn5 acetyltransferase) SAGA complex to selective gene promoters, likely through its physical interaction with the ADA2 subunit (Barlev et al., 1995; Drysdale et al., 1998; Kuo et al., 2000). To test this possibility, we performed Co-IP assay to check the interaction between CPC-1 and the SAGA complex in Neurospora in strains expressing the epitope-tagged Neurospora SAGA homologs. As shown in Figure 4B, Myc-tagged GCN-5 was efficiently immunoprecipitated by the Flag-tagged ADA-2, indicating the existence of an SAGA complex in Neurospora. Although the Myc.GCN-5, MYC.CPC-1 or Flag.ADA-2 protein levels were repressed by 3 mM 3-AT treatment (potentially due to global translational inhibition by eIF2α phosphorylation) (Karki et al., 2020), the interaction between GCN-5 and ADA-2 was almost the same under either normal or amino acid starved conditions (IP was normalized with Input). Importantly, Myc.CPC-1 was also found to associate specifically with Flag.ADA-2 with/without 3-AT treatment (Figure 4C), suggesting that CPC-1 can recruit the SAGA complex to the frq promoter through its ADA-2 subunit to regulate histone acetylation levels at the frq locus under normal or amino acid starvation conditions. Furthermore, immunoprecipitation assays showed that WC-1 and WC-2 also interacted with Myc.GCN-5 (Figure 4—figure supplement 1C). These results suggest that CPC-1 can regulate histone acetylation by recruiting the SAGA complex to the frq promoter.

To further confirm if CPC-1 can recruit GCN5 to the frq promoter, we performed ChIP assay to examine the occupancy of GCN-5 at the frq promoter. As shown in Figure 4D, Myc-tagged GCN-5 rhythmically bound at the frq promoter in DD in the WT strain but its binding was constantly low and arrhythmic in the cpc-1KO strain (WT: p = 0.0006, cpc-1KO: p > 0.05), suggesting that CPC-1 recruits the GCN-5 containing SAGA complex to the frq promoter to allow rhythmic histone acetylation levels, which maintain rhythmic WC-2 binding and thus rhythmic frq transcription.

GCN-5 is required for rhythmic H3ac at the frq promoter

GCN-5 has been shown to regulate light induction and oxidative stress response in Neurospora (Grimaldi et al., 2006; Qi et al., 2018), but its role in the circadian clock remains unclear. To determine the function of GCN-5 in the circadian clock, we created the gcn-5 knockout mutant, and found that it exhibited slow growth and lacked a conidiation rhythm (Figure 5A). To determine its circadian clock at the molecular level, we introduced the FRQ-LUC reporter (luciferase fused at the C terminus of the FRQ protein) into the gcn-5KO strain (Larrondo et al., 2015). As shown in Figure 5B and Figure 5—figure supplement 1B, a robust circadian rhythm of luciferase activity was observed in the WT strain, but it was quickly dampened after 1 day in DD and became arrhythmic in the gcn-5KO strain. Western blot analysis showed that, after the initial light/dark transition, the rhythmic FRQ abundance and phosphorylation were significantly dampened in the gcn-5KO mutant (WT: p = 5.00E−8, gcn-5KO: p = 0.0016) (Figure 5C). Reverse Transcription Quantitative PCR RT-qPCR analysis showed that the circadian rhythms of frq mRNA were also severely dampened in the gcn-5KO strain without (WT: p = 8.37E−06, gcn-5KO: p > 0.05) or with 3 mM 3-AT (WT: p = 5.39E−12, gcn-5KO: p > 0.05) (Figure 5D, E). Together, these results indicate that GCN-5 is critical for maintaining the function of circadian clock.

Figure 5 with 1 supplement see all
Download asset Open asset
GCN-5 is required for rhythmic chromatin structure changes at the frq promoter.

(A) Race tube assay showing that the conidiation rhythm in gcn-5KO strain was lost compared with WT strain. (B) Luciferase assay showing that the luciferase activity rhythm was impaired in the gcn-5KO strain after 1 day transition from light to dark. A FRQ-LUC translational fusion construct was expressed in WT and gcn-5KO strains, and the luciferase signal was recorded in DD for more than 7 days. Normalized data with the baseline luciferase signals subtracted are shown. (C) Western blot assay showing that rhythmic expression of FRQ protein was dampened in the gcn-5KO strain (n = 3; WT: p = 5.00E−08, gcn-5KO: p = 0.0016, RAIN; WT vs gcn-5KO: mesor p = 0.1421, amplitude p = 0.0774, phase p = 0.4319, CircaCompare). RT-qPCR analysis showing that rhythmic expression of frq mRNA was dampened in the gcn-5KO strain without 3-aminotriazole (3-AT; n = 3; WT: p = 8.37E−06, gcn-5KO: p > 0.05) (D) or with 3-AT (n = 3; WT: p = 5.39E−12, gcn-5KO: p > 0.05) (E). (F) Chromatin immunoprecipitation (ChIP) assay showing decreased histone H3ac levels at the promoter of frq gene in the gcn-5KO strain at the indicated time points in DD (n = 3; WT: p = 8.10E−05, gcn-5KO: p > 0.05). Relative H3ac levels were normalized with H2B levels. (G) ChIP assay showing decreased WC-2 levels at the promoter of frq gene in the gcn-5KO strain at the indicated time points in DD (n = 3; WT: p = 0.0003, gcn-5KO: p = 0.0459, RAIN; WT vs gcn-5KO: mesor p = 0.2939, amplitude p = 0.0010, phase p = 0.6933, CircaCompare). Error bars indicate standard deviations (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001; Student’s t test was used.

Figure 5—source data 1

LumiCycle analysis dataset in Figure 5B.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig5-data1-v2.zip
Figure 5—source data 2

Scan of western blot probed for FRQ protein and quantification dataset in Figure 5C.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig5-data2-v2.zip
Figure 5—source data 3

RT-qPCR analysis dataset in Figure 5D.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig5-data3-v2.zip
Figure 5—source data 4

RT-qPCR analysis dataset in Figure 5E.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig5-data4-v2.zip
Figure 5—source data 5

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 5F.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig5-data5-v2.zip
Figure 5—source data 6

Chromatin immunoprecipitation (ChIP) analysis dataset in Figure 5G.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig5-data6-v2.zip

ChIP assay results showed that the H3ac levels were significantly decreased at the frq promoter, and their rhythmic occupancies were severely dampened in the gcn-5KO strain compared with the WT strain (WT: p = 8.10E−05, gcn-5KO: p > 0.05) (Figure 5F). Furthermore, the rhythmic WC-2 binding at the frq C-box of the WT strain was dramatically reduced in the gcn-5KO strain in DD (WT: p = 0.0003, gcn-5KO: p = 0.0459) (Figure 5G). These results indicate that GCN-5 is critical for circadian clock function by regulating rhythmic chromatin structure changes to allow rhythmic WC-2 binding at the frq promoter to drive rhythmic frq transcription.

Since ADA-2 interacts with GCN-5 and it is a subunit of the SAGA complex, we also created the Neurospora ada-2 knockout mutant and examined the function of ADA-2 in the circadian clock. As expected, we found that the circadian phenotypes and frq expression of the ada-2KO strain were very similar to those of the gcn-5KO strain (Figure 5—figure supplement 1A–F). Together, these results demonstrate the importance of GCN-5 and ADA-2 in the Neurospora circadian clock function.

Elevated histone acetylation partially rescues circadian clock defects caused by amino acid starvation

Our results above suggest that amino acid starvation results in low histone acetylation levels at the frq promoter, which impairs the WC-2 binding and rhythmic frq transcription in the cpc-3KO and cpc-1KO mutants. To further confirm this conclusion, we hypothesized that the circadian clock defects under amino acid starvation should be rescued by increasing the histone acetylation levels. Trichostatin A (TSA) is a histone deacetylase (HDACs) inhibitor, which can inhibit HDACs activity and increase the histone acetylation levels in Neurospora (Selker, 1998). We treated the WT strain with different concentrations of 3-AT and TSA, and found that high concentrations of 3-AT lengthened the circadian period of the WT strain to ~24 hr (Figure 6A), but high concentrations of TSA resulted in a slight period shortening to 21 hr (Figure 6B). When TSA was used together with a high concentration of 3-AT (7.5 mM), the long period phenotype can be gradually rescued by increasing TSA concentrations (Figure 6C), which is consistent with our conclusion that the histone acetylation levels changes are responsible for the circadian clock defects caused by amino acid starvation.

Figure 6
Download asset Open asset
Elevated histone acetylation partially rescues impaired circadian rhythm caused by amino acid starvation stress.

(A) Race tube assay showing that high concentrations of 3-aminotriazole (3-AT) treatment elongated circadian conidiation period of WT strain. (B) Race tube assay showing that the high concentrations of Trichostatin A (TSA) treatment shortened circadian conidiation period of WT strain. (C) TSA treatment rescued prolonged circadian period of WT strain caused by 3-AT treatment. WT strain was grown on the race tube medium containing 7.5 mM 3-AT and indicated concentrations of TSA in DD. (D) Chromatin immunoprecipitation (ChIP) assay showing that H3ac levels were decreased in gcn-5KO strains and increased in hda-1KO strains at the promoter of frq gene. Error bars indicate standard deviations (n = 3). *p < 0.05; **p < 0.01; Student’s t test was used. (E) The hda-1KO strain exhibited near normal circadian period in the presence of 3-AT. hda-1KO strains were grown on the race tube medium containing the indicated concentrations of 3-AT in DD. (F) TSA treatment rescued the impaired circadian rhythm of cpc-3KO strain caused by 3-AT treatment. cpc-3KO strains were grown on the race tube medium containing 3 mM 3-AT and indicated concentrations of TSA in DD.

Figure 6—source data 1

RT-qPCR analysis dataset in Figure 6D.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig6-data1-v2.zip

GCN-5 is the major histone acetyltransferase responsible for histone acetylation at the frq locus in response to amino acid starvation. On the other hand, the histone deacetylase HDA-1 was previously reported as a major histone deacetylase that can antagonize and compete with GCN-5 for recruitment to promoters to deacetylate H3 (Islam et al., 2011; Vogelauer et al., 2000). ChIP assays showed that H3ac levels were indeed significantly decreased in the gcn-5KO strain and were markedly increased in the hda-1KO strain at the frq promoter region (Figure 6D), indicating that HDA-1 is responsible for histone deacetylation at the frq locus. Race tube assays showed that in contrast to period lengthening in the WT strain by 3-AT (Figure 6A), the hda-1KO strain exhibited nearly normal circadian period even in the presence of high 3-AT concentrations (Figure 6E), suggesting that reduced histone deacetylation can partially rescue the circadian clock defects in response to amino acid starvation. To further confirm our conclusion, we treated the cpc-3KO strains with different concentrations of TSA and found that the arrhythmic circadian conidiation rhythm caused by 3-AT treatment could be partially rescued by TSA treatments (Figure 6F). Together, these results strongly suggest that the amino acid starvation induced clock defects in the GCN2 signaling pathway mutants are largely due to the decreased histone acetylation at the frq promoter, which prevents efficient WC-2 binding and disrupts the rhythmic frq transcription.

Rhythmic expression of CPC-1 activated metabolic genes under amino acid starvation

Circadian clock has been shown to control metabolic processes and rhythmic transcription of metabolic genes (Baek et al., 2019; Hurley et al., 2014; Hurley et al., 2018). To determine the role of the GCN2 pathway in controlling gene expression under amino acid starvation, we performed RNA-seq experiments to analyze the genome-wide mRNA levels in the WT and cpc-1KO strains in the presence of 3-AT (12 mM). As shown in Figure 7A, 22.1% of genes were found to be upregulated and 11.2% of genes were downregulated in the WT strain after 3-AT treatment compared with normal condition. Specifically, those genes involved in the regulation of oxidoreductase and amino acid metabolism were particularly enriched and were mostly upregulated during the amino acid starvation (Figure 7B). However, the differential expression of the 148 upregulated and 127 downregulated genes found in the WT strain was abolished in the cpc-1KO strain after 3-AT treatment, suggesting that these genes were regulated by CPC-1 under amino acid starvation (Figure 7D and Figure 7—figure supplement 1A, B). Similar to those in Saccharomyces cerevisiae (Natarajan et al., 2001), genes of amino acid biosynthetic pathways, vitamin biosynthetic enzymes, peroxisomal components, and mitochondrial carrier proteins were also identified as CPC-1 targets. Among them, the genes involved in amino acid and vitamin metabolism were particularly enriched and were mostly upregulated during the amino acid starvation (Figure 7E). For example, amino acid synthesis genes his-3 (NCU03139), arg-1 (NCU02639), trp-3 (NCU08409), and ser-2 (NCU01439) were upregulated in the WT strain, but unchanged in the cpc-1KO strain under amino acid starvation (Figure 7A, C).

Figure 7 with 1 supplement see all
Download asset Open asset
Circadian clock control of CPC-1-activated metabolic genes under amino acid starvation.

(A) Comparison of the transcript expression profiles of the WT strains with and without 12 mM 3-aminotriazole (3-AT) treatment. (B) Gene functional enrichment analysis based on the mRNA levels changes for the up- and downregulated genes in the WT strains with and without 12 mM 3-AT treatment. (C) Comparison of the transcript expression profiles of the cpc-1KO strains with and without 12 mM 3-AT treatment. (D) Pie charts showing the overlaps of upregulated genes in the WT strain, but stable genes in the cpc-1KO strains after 12 mM 3-AT treatment. (E) Gene functional enrichment analysis based on the mRNA levels changes for the overlaps of upregulated genes in the WT strain, but stable genes in the cpc-1KO strains after 12 mM 3-AT treatment. (F) RT-qPCR analysis showing that amino acid synthetic genes his-3 (NCU03139) (n = 3; WT: p = 2.21E−05, cpc-1KO: p>0.05), arg-1 (NCU02639) (n = 3; WT: p = 0.0097, cpc-1KO: p > 0.05), and trp-3 (NCU08409) (n = 3; WT: p = 0.0009, cpc-1KO: p > 0.05) were activated by CPC-1 and were rhythmic expressed with 3 mM 3-AT treatment. The primers used for RT-qPCR are shown in Key Resources Table.

Figure 7—source data 1

Raw dataset of RNA-seq analysis in Figure 7A.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig7-data1-v2.zip
Figure 7—source data 2

Raw dataset of RNA-seq analysis in Figure 7C.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig7-data2-v2.zip
Figure 7—source data 3

RT-qPCR analysis dataset in Figure 7F.

https://cdn.elifesciences.org/articles/85241/elife-85241-fig7-data3-v2.zip

To examine whether these CPC-1 activated genes were controlled by circadian clock, we re-analyzed and compared our identified CPC-1 target genes with previously published RNA-seq data of rhythm samples (Hurley et al., 2014; Hurley et al., 2018). As shown in Supplementary file 1, we summarized the rhythmic expression of 79 upregulated and 67 downregulated CPC-1 target genes based on the eJTK Cycle results of Hurley et al., 2018 (its Supplemental Datasets 1 and 2). We re-analyzed the rhythmicity of CPC-1-targeted genes (148 upregulated and 127 downregulated) using CircaSingle (Parsons et al., 2020), and added the p-values in Supplementary file 1. There were 146 rhythmic genes based on the eJTK Cycle, and 132 rhythmic genes based on the CircaSingle. As expected, there were 106 overlapping genes between these two sets of data, confirming the results using eJTK Cycle. Thus, we performed further analysis based on the data from eJTK Cycle. There were about 146/275 (53%) of the CPC-1 up- and downregulated genes under amino acid starvation exhibiting rhythmicity, indicating a highly significant enrichment of CPC-1 regulated genes as clock-controlled genes (p = 3.341905e−06, hypergeometric distribution test). Furthermore, we performed ChIP experiments to examine whether CPC-1 directly activates the expression of amino acid synthetic genes. As shown in Figure 7—figure supplement 1C, CPC-1 was found to be constitutively enriched at the promoters of his-3 (NCU03139), trp-3 (NCU08409), and ser-2 (NCU01439) genes (WT (0 mM 3-AT): p > 0.05, WT (3 mM 3-AT): p > 0.05, cpc-1KO (0 mM 3-AT): p > 0.05), and the enrichment was enhanced by 3-AT treatment. Because the CPC-1 binding was not rhythmic, we performed RT-qPCR experiments to re-analyze whether those genes were rhythmically transcribed in DD. Consistent with the published RNA-seq analysis results, we found that frq and the amino acid synthetic genes such as his-3 (NCU03139) (WT: p = 2.21E−05, cpc-1KO: p > 0.05), arg-1 (NCU02639) (WT: p = 0.0097, cpc-1KO: p > 0.05), and trp-3 (NCU08409) (WT: p = 0.0009, cpc-1KO: p > 0.05) were rhythmically expressed in the WT but not cpc-1KO strain under amino acid starvation (Figure 7F). In addition, we also found that ser-2 (NCU01439) (WT: p = 2.25E−05, cpc-1KO: p > 0.05) exhibited rhythmic expression in the WT but not in the cpc-1KO strain under amino acid starvation (Figure 7—figure supplement 1D), even though it was not previously shown to be rhythmic in the published RNA-seq analysis study. These results suggest that many CPC-1-activated metabolic genes under amino acid starvation are regulated by circadian clock. Next, we performed RT-qPCR experiments to detect the mRNA levels of amino acid synthesis genes under normal condition (0 mM 3-AT), and found that arg-1 (NCU02639) (WT: p = 0.0353), trp-3 (NCU08409) (WT: p = 0.0436), and ser-2 (NCU01439) (WT: p = 0.0008) genes, but not the his-3 (NCU03139) (WT: p > 0.05) gene, were rhythmic in the WT strain in DD (Figure 7—figure supplement 1E). Together, these results suggest that the GCN2 signaling pathway functions to maintain the robust circadian clock and rhythmic expression of metabolic genes under amino acid starvation.

Discussion

Circadian clock drives robust rhythmic gene expression and activities in different environmental and nutritional conditions. Here, we demonstrate that the nutrient-sensing GCN2 pathway plays an unexpected role in maintaining the Neurospora circadian clock in response to amino acid starvation stress. Under normal conditions, CPC-1 is expressed at its basal levels in the WT or cpc-3KO strain, which is required for rhythmic expression of frq gene by recruiting the histone acetyltransferase GCN-5 containing SAGA complex to the frq promoter. However, amino acid starvation resulted in compact chromatin structure (due to decreased H3ac) in the frq promoter in the WT strain (Figure 3B), likely due to activation of the histone deacetylases or inhibition of histone acetyltransferases. The circadian clock-controlled CPC-3 and CPC-1 signaling pathway (Karki et al., 2020) is activated by amino acid starvation, and the elevated CPC-1 protein efficiently recruits the histone acetyltransferase GCN-5 containing SAGA complex to the frq promoter to increase the histone acetylation levels and loosen the chromatin structure, which permits rhythmic WC-2 binding. Therefore, under amino acid starvation, the disruption of the CPC-3 and CPC-1 signaling pathway results in decreased histone acetylation levels, reduced WC-2 binding at the frq promoter and the loss of robust rhythmic frq transcription (Figure 8).

Figure 8
Download asset Open asset
Model showing the role of CPC-3 and CPC-1 in maintaining the Neurospora circadian rhythm in response to amino acid starvation.

Under normal conditions, CPC-1 is expressed at its basal levels in the WT (Left) or cpc-3KO (Right) strain, which is required for rhythmic expression of frq gene by recruiting the histone acetyltransferase GCN-5 containing SAGA complex to the frq promoter. Under amino acid starvation conditions, the chromatin in the frq promoter of the WT strain is constitutively compacted (due to decreased H3ac), likely due to activation of histone deacetylases or inhibition of histone acetyltransferases. CPC-3 and CPC-1 signaling pathway was activated by amino acid starvation and the elevated CPC-1 protein would efficiently recruit the histone acetyltransferase GCN-5 containing SAGA complex to promote the histone acetylation levels, which permitted rhythmic WCC binding at the frq promoter (Left). Disruption of the CPC-3 and CPC-1 signaling pathway resulted in decreased histone acetylation levels of the frq gene promoter, reduced WCC binding and damped circadian oscillations in response to the amino acid starvation stress (Right).

Maintaining robust rhythmic gene expression and circadian activities in response to various environmental and nutritional stresses is important for the health or survival of different organisms. Circadian clock synchronizes metabolic processes and systemic metabolite levels, while nutrients and energy signals also feedback to circadian clocks to adapt their metabolic state (Bass, 2012; Hurley et al., 2014; Klemz et al., 2017; Reinke and Asher, 2019). Amino acid starvation is known to inhibit the global translation efficiency through activating GCN2-mediated eIF2α phosphorylation, which conserves energy and allows cells to reprogram gene expression to relieve stress damage. It was recently shown that circadian clock control of GCN2-mediated eIF2α phosphorylation was necessary for rhythmic translation initiation in Neurospora (Ding et al., 2021; Karki et al., 2020). However, it was previously unknown whether circadian clock would be affected by amino acid starvation stress. After activation of the GCN2-mediated eIF2α phosphorylation by amino acid starvation, a subset of transcripts containing overlapping upstream open reading frames (uORFs) in their 5′ untranslated region (5′ UTR) are efficiently translated, including the yeast transcription factor GCN4, the Neurospora CPC-1 and their mammalian ortholog ATF4, which activate the transcription of various amino acid biosynthetic genes (Hinnebusch, 1984; Paluh et al., 1988; Vattem and Wek, 2004). Our ChIP results showed that CPC-1 could rhythmically bind to the region close to C-box at the frq promoter to activate frq transcription (Figure 4A). WCC binding at the frq C-box region was slightly decreased under normal condition (Figure 2—figure supplement 1A), and dramatically decreased under amino acid starvation in the cpc-1KO strain (Figure 2D), suggesting that CPC-1 cooperates with WCC to promote frq transcription in response to amino acid starvation. Although CPC-1 rhythmically bound at the frq promoter, ChIP experiments showed that CPC-1 binding at several selected amino acid biosynthetic genes did not appear to be rhythmic (Figure 7—figure supplement 1C). However, our RT-qPCR results (Figure 7F and Figure 7—figure supplement 1) and the previous RNA-seq data showed that many CPC-1-targeted metabolic genes were rhythmic expressed (Hurley et al., 2018). It is possible that the binding of CPC-1 to these promoters is still rhythmic but with low amplitudes. As a result, the limited sensitivity of our ChIP assays failed to detect these rhythms. Alternatively, the rhythmic transcription of these genes might be controlled by the rhythmic transcriptional activation activity of CPC-1 rather than its binding. In addition, the rhythmic binding of WCC or WCC-controlled transcription factors (Hurley et al., 2014) might also contribute to their rhythmic transcription.

Amino acid starvation was able to affect gene expression by regulating chromatin modifications. Mammalian transcription factor ATF2 was reported to promote the modification of the chromatin structure in response to amino acid starvation to enhance the transcription of numbers of amino acid-regulated genes (Bruhat et al., 2007). Amino acid starvation was also shown to induce reactivation of silenced transgenes and latent HIV-1 provirus by downregulation of histone deacetylase 4 in mammalian cells (Palmisano et al., 2012). Here, we found that amino acid starvation suppressed frq expression by decreasing the histone acetylation levels likely through activation of histone deacetylases or inhibition of histone acetyltransferases (Figure 3B). The activated GCN2 signaling pathway resulted in recruitment of the histone acetyltransferase SAGA complex to the frq promoter through its interaction with CPC-1 to re-establish a proper histone acetylation state to relieve the repression (Figure 3 and Figure 4). Although histone acetylation and deacetylation rhythms have been reported in mammalian cells (Papazyan et al., 2016; Takahashi, 2017), SAGA complex has also been reported to be involved in circadian regulation through interaction with the CLOCK complex in Drosophila (Mahesh et al., 2020), their function in circadian clock is unclear in Neurospora. Here we demonstrated the important role of GCN-5 in regulating the rhythmic histone acetylation and WC-2 binding at the frq promoter (Figure 5). Our study unveiled an unsuspected link between nutrient limitation and circadian clock function mediated by the GCN2 signaling pathway. These results provide key insights into the epigenetic regulatory mechanisms of circadian gene expression during amino acid starvation.

The nutrient-sensing GCN2 signaling pathway is highly conserved in eukaryotic cells from yeast to mammals. In the budding yeast S. cerevisiae and the filamentous fungus N. crassa, the GCN2 kinase responds to nutrient deprivation, whereas it phosphorylates eIF2α and upregulates the master transcription factors GCN4 or CPC-1, respectively, which binds target DNA as a dimer to activate amino acid biosynthetic genes (Ebbole et al., 1991; Hinnebusch, 2005; Hope and Struhl, 1987). In mammalian cells, several GCN2-related kinases can phosphorylate eIF2α in response to various stress conditions, triggering the integrated stress response (Costa-Mattioli and Walter, 2020; Donnelly et al., 2013). Interestingly, different from the normal circadian period of cpc-3KO strain in Neurospora without amino acid starvation (Figure 1A, Figure 1—figure supplement 1 and Figure 1—figure supplement 2), it was reported that GCN2 modulated circadian period by phosphorylation of eIF2α in mammals under normal condition, but it was unknown whether GCN2 was involved in circadian regulation of metabolism under nutrient limitation (Pathak et al., 2019). Our results suggest that the GCN2 signaling pathway is required for maintaining circadian clock under amino acid starvation, which is important for robust rhythmic expression of metabolic genes (Figure 7). Because GCN2 signaling pathway is important for nutrient sensing, it may also be important for nutritional compensation (Kelliher et al., 2023) and plays a role in maintaining the robustness of rhythms in a range of nutritional conditions. Time-restricted feeding prevents obesity and metabolic syndrome through circadian-related mechanisms (Chaix et al., 2019), but how eating pattern affects circadian regulation is unclear. The conservation of the GCN2 signaling pathway and our results here suggest that GCN2 may play an important role in mediating circadian regulation of metabolism during nutrient limitation caused by feeding restrictions in mammals. Together, our studies suggest a conserved role of the GCN2 signaling pathway in maintaining the robustness of circadian clock under nutrient starvation in eukaryotes.

Materials and methods

Key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Strain, strain background (Neurospora crassa)87-3 (ras-1bd, a)Dr.Yi Liu’s Laboratory
Strain, strain background (Neurospora crassa)301-6 (ras-1bd, his-3-, A)Dr.Yi Liu’s Laboratory
Strain, strain background (Neurospora crassa)ras-1bd;cpc-3KOFungal Genetics Stock CenterNCU01187
Strain, strain background (Neurospora crassa)ras-1bd;cpc-1KOFungal Genetics Stock CenterNCU04050
Strain, strain background (Neurospora crassa)ras-1bd;gcn-5KOFungal Genetics Stock CenterNCU10847
Strain, strain background (Neurospora crassa)ras-1bd;ada-2KOFungal Genetics Stock CenterNCU04459
Strain, strain background (Neurospora crassa)ras-1bd;hda-1KOFungal Genetics Stock CenterNCU01525
Strain, strain background (Neurospora crassa)ras-1bd;cpc-1KO, cpc-1-Myc.CPC-1Dr.Xiao Liu’s Laboratory
Strain, strain background (Neurospora crassa)301-6, cfp-Myc.CPC-1, cfp-Flag.ADA-2Dr.Xiao Liu’s Laboratory
Strain, strain background (Neurospora crassa)301-6, cfp-Myc.GCN-5, cfp-Flag.ADA-2Dr.Xiao Liu’s Laboratory
Strain, strain background (Neurospora crassa)frq-lucDr.Yi Liu’s Laboratory
Strain, strain background (Neurospora crassa)ras-1bd;cpc-3KO, frq-lucDr.Xiao Liu’s Laboratory
Strain, strain background (Neurospora crassa)ras-1bd;cpc-1KO, frq-lucDr.Xiao Liu’s Laboratory
Strain, strain background (Neurospora crassa)FRQ-LUCDr. Luis Larrondo’s Laboratory
Strain, strain background (Neurospora crassa)ras-1bd;gcn-5KO, FRQ-LUCDr.Xiao Liu’s Laboratory
Strain, strain background (Neurospora crassa)ras-1bd;ada-2KO, FRQ-LUCDr.Xiao Liu’s Laboratory
AntibodyRabbit polyclonal anti-Histone H2BAbcamCat# ab17901:3000
AntibodyRabbit polyclonal anti-Histone H3acMilliporeCat# 06-5991:3000
AntibodyMouse monoclonal anti-c-MycTransGenCat# HT1011:3000
AntibodyMouse monoclonal anti-FlagSigmaCat# F18041:3000
AntibodyRabbit polyclonal anti-P-eIF2αAbcamCat# ab321571:3000
AntibodyRabbit polyclonal anti-CPC-1Dr.Xiao Liu’s Laboratory1:3000
AntibodyRabbit polyclonal anti-FRQDr.Yi Liu’s Laboratory1:3000
AntibodyRabbit polyclonal anti-WC-1Dr.Yi Liu’s Laboratory1:4000
AntibodyRabbit polyclonal anti-WC-2Dr.Yi Liu’s Laboratory1:8000
Sequence-based reagentfrq-FDr.Xiao Liu’s LaboratoryRT-qPCRGCAGTGTCATTGACGACTTG
Sequence-based reagentfrq-RDr.Xiao Liu’s LaboratoryRT-qPCRCCTCCAACTCACGTTTCTTTC
Sequence-based reagenthis-3-FDr.Xiao Liu’s LaboratoryRT-qPCRCCTCGTTCGTCAAGCACATTA
Sequence-based reagenthis-3-RDr.Xiao Liu’s LaboratoryRT-qPCRCTCCTCAACCTTAGCCAACTG
Sequence-based reagenttrp-3-FDr.Xiao Liu’s LaboratoryRT-qPCRACCTATATCCTTCAGAACCAATACG
Sequence-based reagenttrp-3-RDr.Xiao Liu’s LaboratoryRT-qPCRGCTCGGTATCCTTCCAGTTG
Sequence-based reagentser-2-FDr.Xiao Liu’s LaboratoryRT-qPCRGCTGCTAACGGTGACTACTT
Sequence-based reagentser-2-RDr.Xiao Liu’s LaboratoryRT-qPCRGGTGAGGATGATGTTGTTGAG
Sequence-based reagentarg-1-FDr.Xiao Liu’s LaboratoryRT-qPCRCCCATCATTGCCCGTGCCC
Sequence-based reagentarg-1-RDr.Xiao Liu’s LaboratoryRT-qPCRTGACGACCCTGGAAGCGAG
Sequence-based reagentβ-tubulin-FDr.Xiao Liu’s LaboratoryRT-qPCRGCGTATCGGCGAGCAGTT
Sequence-based reagentβ-tubulin-RDr.Xiao Liu’s LaboratoryRT-qPCRCCTCACCAGTGTACCAATGCA
Sequence-based reagentfrq C-box-FDr.Xiao Liu’s LaboratoryChIP-qPCRGTCAAGCTCGTACCCACATC
Sequence-based reagentfrq C-box-RDr.Xiao Liu’s LaboratoryChIP-qPCRCCGAAAGTATCTTGAGCCTCC
Sequence-based reagentfrq promoter-FDr.Xiao Liu’s LaboratoryChIP-qPCRGTTGCCGTGACTCCCCCTTG
Sequence-based reagentfrq promoter-RDr.Xiao Liu’s LaboratoryChIP-qPCRCCGAAAGTATCTTGAGCCTCC
Sequence-based reagenthis-3 ChIP-FDr.Xiao Liu’s LaboratoryChIP-qPCRTTTTCATAAAGCCCGAGTCT
Sequence-based reagenthis-3 ChIP-RDr.Xiao Liu’s LaboratoryChIP-qPCRCAGGTATTGTGCTGTTCCCC
Sequence-based reagenttrp-3 ChIP-FDr.Xiao Liu’s LaboratoryChIP-qPCRAATCGGGTGAGTCAAAGGCG
Sequence-based reagenttrp-3 ChIP-RDr.Xiao Liu’s LaboratoryChIP-qPCRCGAGCAAGAGGGAGAGGTGT
Sequence-based reagentser-2 ChIP-FDr.Xiao Liu’s LaboratoryChIP-qPCRGGGACAAAAGCAGTGATTCTA
Sequence-based reagentser-2 ChIP-RDr.Xiao Liu’s LaboratoryChIP-qPCRCGATTTACATCCATCTGAGA
Sequence-based reagentfrq northern-FDr.Xiao Liu’s LaboratoryNorthern blotTAATACGACTCACTATAGGGCCTTCGTTGGATATCCATCATG
Sequence-based reagentfrq northern-RDr.Xiao Liu’s LaboratoryNorthern blotGAATTCTTGCAGGGAAGCCGG
Software, algorithmImageJhttps://imagej.nih.gov/ij/
Software, algorithmLumiCyclehttps://actimetrics.com/products/lumicycle/
Software, algorithmCircaComparehttps://github.com/RWParsons/circacompare/; Parsons et al., 2020

Strains, culture conditions, and race tube assay

Request a detailed protocol

The 87-3 (ras-1bd, a) and 301-6 (ras-1bd, his-3−, A) strain was used as the wild-type (WT) strain in this study. Knockout mutants were all generated based on the ras-1bd background (Belden et al., 2007). cpc-3KO (NCU01187), cpc-1KO (NCU04050), gcn-5KO (NCU10847), ada-2KO (NCU04459), and hda-1KO (NCU01525) strains were obtained from the Fungal Genetic Stock Center (FGSC) and were crossed with a ras-1bd strain to create the ras-1bd;cpc-3KO, ras-1bd;cpc-1KO, ras-1bd;gcn-5KO, ras-1bd;ada-2KO, and ras-1bd;hda-1KO strains (Colot et al., 2006). Constructs with the cpc-1 promoter driving expression of Myc.CPC-1 were introduced into the cpc-1KO strains at the his-3 (NCU03139) locus by homologous recombination. Constructs with the cfp promoter driving expression of Flag.ADA-2 were introduced into the 301-6, cfp-Myc.CPC-1 or 301-6, cfp-Myc.GCN-5 strains by random insertion with nourseothricin selection (He et al., 2020). Positive transformants were identified by western blot analyses, and homokaryon strains were isolated by microconidia purification with 5 µm filters.

Liquid cultures were grown in minimal media (1× Vogel’s, 2% glucose). For rhythmic experiments, Neurospora was cultured in petri dishes in liquid medium for 2 days. The Neurospora mats were cut into discs and transferred into medium-containing flasks and were harvested at the indicated time points.

The medium for race tube assay contained 1× Vogel’s salts, 0.1% glucose, 0.17% arginine, 50 ng/ml biotin, and 1.5% agar. After entrainment of 24 hr in the constant light condition, race tubes were transferred to constant darkness conditions and marked every 24 hr. The circadian period of the Neurospora strain could be calculated according to the ratio between the distance of marked conidia band positions and the distance of conidiation bands.

Luciferase reporter assay

Request a detailed protocol

The luciferase reporter assay was performed as reported previously (Gooch et al., 2008; Larrondo et al., 2015; Liu et al., 2017). The luciferase reporter construct (frq-luc) containing the luciferase gene under the control of the frq promoter, was introduced into 301-6, cpc-3KO or cpc-1KO strains by transformation. The luciferase reporter construct (FRQ-LUC) containing a luciferase fused to the C terminus of the FRQ protein, was introduced into gcn-5KO or ada-2KO strains by crossing. Firefly luciferin (final concentration of 50 μM) was added to autoclaved FGS-Vogel’s medium containing 1× FGS (0.05% fructose, 0.05% glucose, 2% sorbose), 1× Vogel’s medium, 50 μg/l biotin, and 1.8% agar. Conidia suspension was placed on autoclaved FGS-Vogel’s medium and grown in constant light overnight. The cultures were then transferred to constant darkness, and luminescence was recorded in real time using a LumiCycle after 1 day in DD. The data were then normalized with LumiCycle analysis software by subtracting the baseline luciferase signal, which increases as cell grows.

Protein analysis

Request a detailed protocol

Protein extraction, quantification, and western blot analyses were performed as previously described (Liu et al., 2017). Briefly, tissues were ground in liquid nitrogen with a mortar and pestle and suspended in ice-cold extraction buffer (50 mM Hydroxyethylpiperazine Ethane Sulfonic Acid HEPES (pH 7.4), 137 mM NaCl, 10% glycerol) with protease inhibitors (1 μg/ml Pepstatin A, 1 μg/ml Leupeptin, and 1 mM phenylmethylsulfonyl fluoride PMSF). After centrifugation, protein concentrations were measured using protein assay dye (Bio-Rad). For western blot analyses, equal amounts of total protein (40 μg) were loaded in each protein lane of 7.5% or 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) gels containing a ratio of 37.5:1 acrylamide/bisacrylamide. After electrophoresis, proteins were transferred onto PVDF membranes, and western blot analyses using FRQ, WC-1, WC-2, P-eIF2α (Abcam, ab32157), and CPC-1 antibodies were performed. Western blot signals were detected by X-ray films and scanned for quantification.

To detect the phosphorylation levels of WC-1 and WC-2, PPase inhibitors (25 mM NaF, 10 mM Na4P2O7·10H2O, 2 mM Na3VO4, 1 mM ethylenediaminetetraacetic acid EDTA) were made fresh and added to the protein extraction buffer. Proteins were loaded in each protein lane of 7.5% SDS–PAGE gels containing a ratio of 149:1 acrylamide/bisacrylamide.

RNA analysis

Request a detailed protocol

RNA was extracted with Trizol and further purified with 2.5 M LiCl as described previously (Liu et al., 2017). For Northern blot analysis, equal amounts of total RNA (20 μg) were loaded onto agarose gels. After electrophoresis, the RNA was transferred onto nitrocellulose membrane. The membrane was probed with [32P] UTP (PerkinElmer)-labeled RNA probes specific for frq. RNA probes were transcribed in vitro from PCR products by T7 RNA polymerase (Invitrogen, AM1314M) with the manufacturer’s protocol. The frq primers used for the template amplification are shown in Key Resources Table.

For RT-qPCR, the cultures of WT and cpc-1KO strains were collected at the indicated time points in constant darkness in liquid growth medium (1× Vogel’s, 2% glucose). RT-qPCR were performed as previously described (Cui et al., 2020). Each RNA sample (1 μg) was subjected to reverse transcription with HiScript II reverse transcriptase (Vazyme, R223), and then amplified by real-time PCR (Bio-Rad, CFX96). For RT-qPCR, primers target the coding genes of frq (NCU02265), his-3 (NCU03139), ser-2 (NCU01439), trp-3 (NCU08409), his-4 (NCU06974), arg-1 (NCU02639), and arg-10 (NCU08162) were designed, and the β-tubulin (NCU04054) was used as an internal control. The primers used for RT-qPCR are shown in Key Resources Table.

Generation of antiserum against CPC-1

Request a detailed protocol

Two CPC-1 peptides (SELDLLDFATFDGG and RDKPLPPIIVEDPS) were synthesized and used as the antigens to generate rabbit polyclonal antisera (ABclonal) as described previously (Cui et al., 2020; Zhou et al., 2013).

Co-IP analysis

Request a detailed protocol

Immunoprecipitation analyses were performed as previously described (Cao et al., 2018; Cheng et al., 2001b). Briefly, Neurospora proteins were extracted as described above. For each immunoprecipitation reaction, 2 mg protein and 2 μl c-Myc (TransGen, HT101), 2 μl Flag (Sigma, F1804), or 2 μl WC-2 antibody (Cheng et al., 2001a) were used. After incubation with antibody for 3 hr, 40 μl GammaBind G Sepharose beads (GE Healthcare, 17061801) were added, and samples were incubated for 1 hr. Immunoprecipitated proteins were washed three times using extraction buffer before western blot analysis. IP experiments were performed using cultures harvested in constant light.

ChIP analysis

Request a detailed protocol

ChIP assays were performed as previously described (Cui et al., 2020; Zhou et al., 2013) with 1 mg protein used for each immunoprecipitation reaction. The ChIP reaction was carried out with 2 μl WC-2 (Cheng et al., 2001a), CPC-1, H2B (Abcam, ab1790), or H3ac (Millipore, 06-599) antibody. Immunoprecipitated DNA was quantified by real-time qPCR. Occupancies were normalized by the ratio of ChIP to Input DNA. ChIP was performed using 2 μl c-Myc monoclonal antibody (TransGen, HT101) to examine occupancy of Myc.GCN-5. Occupancies of ChIP were normalized using IgG. The primers used for ChIP-qPCR are shown in Key Resources Table.

RNA-seq analysis

Request a detailed protocol

The WT and cpc-1KO strains were cultured with or without 12 mM 3-AT treatment in constant light. Total RNAs were extracted using Trizol reagents. Libraries were prepared according to the manufacturer’s instructions and analyzed using 150 bp paired-end Illumina sequencing (Annoroad Gene Technology, Beijing). After sequencing, the raw data were treated and mapped to the genome of Neurospora crassa and transformed into expression value. The gene expression levels were scored by fragments per kb per million (FPKM) method. The differences in gene expression between samples was compared by comparing FPKM values, and those with fold change more than 1 (False Discovery Rate FDR <0.1) were thought to be differentially expressed genes. The functional category enrichments, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes KEGG terms, were analyzed. The KEGG pathway enrichment was evaluated based on hypergeometric distribution, and the R package ‘ggplot2’ version 3.3.6 was used to visualize the enrichment results.

Quantification and statistical analysis

Request a detailed protocol

Quantification of western blot data were performed using Image J software. Error bars are standard deviations for ChIP assays from at least three independent technical experiments, and standard error of means for race tube assays from at least three independent biological experiments, unless otherwise indicated. Statistical significance was determined by Student’s t test. Statistical tests for the presence of rhythmicity and differences between two rhythms in parameters, including amplitude, phase, and mesor (the midline estimating statistic of rhythms) were analyzed using CircaSingle and CircaCompare (https://rwparsons.shinyapps.io/circacompare/) (Liu et al., 2021; Parsons et al., 2020). CircaCompare was used to compare the differences between two rhythmic datasets, and CircaSingle was used to re-analyze the rhythmic parameters of the RNA-seq results by eJTK Cycle. Amplitude refers to half of the difference between the peak and trough of a given response variable, phase refers to the time at which the response variable peaks, and mesor refers to the rhythm-adjusted mean level of a response variable around which a wave function oscillates. Statistical tests for the presence of rhythmicity and statistically significant differences between two groups of rhythmic parameters are indicated by p-values. A p-value <0.05 indicates the presence of rhythmicity, but a p-value >0.05 indicates the loss of rhythmicity. The results of the CircaSingle statistical tests are added in Supplementary file 1. The results of the CircaCompare statistical tests are summarized in Supplementary file 2.

Data and materials availability

Request a detailed protocol

All data generated or analyzed during this study are included in the manuscript and supporting files. Our generated RNA sequencing data have been deposited in GEO under accession code GSE220169. RNA sequencing data from Hurley et al., 2014 were previously deposited to the NCBI SRA under accession number SRP046458 (Hurley et al., 2014). Materials are available from the corresponding authors upon reasonable request.

Data availability

All data generated or analyzed during this study are included in the manuscript and supporting files. Our generated RNA sequencing data have been deposited in GEO under accession code GSE220169.

The following data sets were generated
    1. Liu XL
    2. Yang Y
    3. Liu X
    (2022) NCBI Gene Expression Omnibus
    ID GSE220169. The nutrient-sensing GCN2 signaling pathway is essential for circadian clock function by regulating histone acetylation under amino acid starvation.
    https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE220169
The following previously published data sets were used
    1. Hurley JM
    2. Dasgupta A
    3. Dunlap JC
    (2014) NCBI BioProject
    ID PRJNA250475. Analysis of clock-regulated genes in Neurospora reveals widespread posttranscriptional control of metabolic potential.
    https://www.ncbi.nlm.nih.gov/bioproject/PRJNA250475

References

Decision letter

  1. Alan G Hinnebusch
    Reviewing Editor; Eunice Kennedy Shriver National Institute of Child Health and Human Development, United States
  2. Kevin Struhl
    Senior Editor; Harvard Medical School, United States
  3. Joanna Chiu
    Reviewer
  4. Jay C Dunlap
    Reviewer; Geisel School of Medicine at Dartmouth, United States

Our editorial process produces two outputs: i) public reviews designed to be posted alongside the preprint for the benefit of readers; ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "The nutrient-sensing GCN2 signaling pathway is essential for circadian clock function by regulating histone acetylation under amino acid starvation" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Kevin Struhl as the Senior Editor. The following individuals involved in the review of your submission have agreed to reveal their identity: Joanna Chiu (Reviewer #2); Bin Wang and Jay C Dunlap (Reviewer #3, co-reviewers).

The reviewers have discussed their reviews with one another, and the Reviewing Editor Alan Hinnebusch has drafted this to help you prepare a revised submission.

Essential revisions:

1) Conduct statistical analysis to assess the significance of the circadian rhythms claimed for FRQ mRNA abundance, and for WC complex binding, H3Ac levels, GCN-5 binding, and CPC-1 binding at the FRQ promoter.

2) The data are not compelling that CPC-1 occupancy at FRQ is rhythmic. If the statistically significant rhythmic binding of CPC-1 is not demonstrated, then the molecular basis of the rhythmic recruitment of Gcn5 and H3Ac will be unclear. In that event, the authors may need to consider the model that CPC-1 assists constitutively in recruiting Gcn5 and attendant H3 acetylation, and that some other factor is involved in producing the circadian rhythms of these events, and in turn, the rhythmic binding of WC complex at the FRQ promoter.

3) It appears that the experiments to examine the involvement of GCN-5 and ADA-2 in the rhythmic transcription of FRQ were performed only in non-starvation conditions. If this was indeed the case, then at a minimum, key conclusions would have to be qualified. More likely, the experiments would have to be extended to include starvation conditions.

4) Provide clarification about why knock out of GCN5 reduces FRQ protein expression and diminishes its rhythm without corresponding effects on FRQ mRNA abundance (Figure 5C vs 5D). Are the data lacking in precision or accuracy, or is a more complex model required to explain these results?

5) The data provided to support the claim of rhythmic transcription of the four candidate amino acid biosynthetic genes in Figure 7F is not very compelling. If the statistical analysis fails to confirm their rhythmic transcription in WT, it might be possible to bolster their claim by evaluating published RNA-seq data (sampled every 2 hrs over 2 days, done in triplicate) for known CPC-1 target genes. Because these published experiments were not carried out in amino acid starvation conditions however, the measured expression of the biosynthetic gene transcripts might be dominated by transcription factors that bind constitutively rather than by CPC-1. In the absence of published data supporting rhythmic transcription of amino acid biosynthetic genes regulated by CPC-1, the authors should consider conducting additional RNA-Seq experiments at different time points in amino acid-starved cells. Alternatively, it might be possible to provide RT-qPCR data on additional candidate genes, coupled with statistical analysis of rhythmic transcription. Failing that, the authors would likely need to soften or eliminate claims of rhythmic transcription of these genes.

6) The proposed rhythmic transcription of amino acid biosynthetic genes would also predict rhythmic binding by CPC-1 in their promoters under starvation conditions. Conducting ChIP analysis of CPC-1 binding at a few candidate biosynthetic genes that show convincing rhythmic transcription would be highly desirable, and if observed, might compensate for less compelling transcript data. Such measurements would also be relevant to the question noted above of whether CPC-1 binding at the FRQ promoter itself is truly rhythmic.

7) It is necessary to add new text to address queries raised by Referees #2 and #3 about various findings that require additional clarification or qualification.

8) It is important to provide raw source data for all measurements.

Reviewer #1 (Recommendations for the authors):

i) Explain why knock out of GCN5 eliminates rhythmic FRQ expression but does not have a corresponding effect on FRQ mRNA abundance.

ii) Determine whether elimination of HDA1 restores H3ac and WCC binding at FRQ, and rhythmic growth, in 3AT-treated cpc-1 and gcn5 mutants.

iii) Provide stronger evidence for rhythmic transcription of aa biosynthetic genes.

iv) Possibly attempt to determine whether CPC-1 occupancy at aa biosynthetic genes (and at FRQ itself) is rhythmic.

Reviewer #2 (Recommendations for the authors):

1. The authors should use circadian statistics (e.g. CircaCompare; Parsons et al. 2020) to compute the phase and amplitude of the mRNA, DNA binding of WC complex, and H3Ac rhythms. This will allow them to compare between rhythms and provide statistical significance values, rather than just providing qualitative descriptions, e.g. "slightly decreased" in lines 189-190. This will be valuable when comparing rhythms between strains and between nutrient conditions.

2. Line 119: The authors indicate that the rhythmicity of histone acetylation at the frq gene promoter is lost in many of the mutants they tested, e.g. CPC-3 and CPC-1. However, it is not clear whether these mutants simply have a lower level of H3ac (Figure 3D), resulting in a lower level of WC binding to frq promoter (Figure 2C-D), and consequently lower level of frq mRNA (Figure 1D). When the levels of these measures are low, it is difficult to observe any rhythms. Furthermore, is it really necessary to emphasize the maintenance of rhythms? When clock mRNA levels are too low, clocks are normally disrupted anyway.

3. The authors propose that GCN5 works downstream of CPC-1 and CPC-3. If that is the case, why does frq mRNA in gcn-5 KO strain show dampened but consistently high level (Figure 5D), rather than a consistently low level as in the case of cpc-1 and cpc-3 ko strains? Can the authors comment on this? Also, it appears that H3ac and WC binding to frq promoter is lower compared to WT (Figure 5E-F). Can the authors explain why frq mRNA level is higher given H3ac and WC binding to frq promoter is lower?

4. There is no data in this study showing that the GCN2 pathway is activated in amino acid-starved conditions, only that it is required to maintain robust frq and conidiation rhythms. Can the authors clarify how they are defining "activation of the GCN2 pathway" in this study? For example, is it recruitment of GCN-5 and SAGA complex to frq promoter?

5. The experiments to examine the involvement of GCN-5 and ADA-2 were performed in normal conditions (no amino acid starvation). Unlike cpc-1 and cpc-3 KO strains, gcn-5 and ada-2 KO strains showed severely disrupted frq rhythms, suggesting they are normally required for robust circadian rhythms. If GCN-5 and the SAGA complex are normally involved in regulating H3ac rhythms in the frq loci, how does the GCN2 pathway modulates the activity of GCN-5 and SAGA complex in conditions of amino acid starvation? Are the interactions between GCN2/4 with GCN-5 and SAGA complex different in normal vs amino acid starved conditions? The authors should clarify their model.

6. Given that the GCN2 pathway is important for nutrient sensing, the authors should not disregard the alternative hypothesis that the GCN2 pathway may be important for nutrient compensation and plays a role in maintaining the robustness of rhythms in a range of nutrient conditions.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "The nutrient-sensing GCN2 signaling pathway is essential for circadian clock function by regulating histone acetylation under amino acid starvation" for further consideration by eLife. Your revised article has been evaluated by Kevin Struhl (Senior Editor) and a Reviewing Editor, Alan Hinnebusch.

It appears that you have satisfied some, but not all, of the major issues raised about the previous version of the paper. It was recognized that you provided new evidence indicating that knockout of GCN5 eliminates rhythmic FRQ expression with a corresponding effect on FRQ mRNA abundance. While you also provided new ChIP evidence suggesting that CPC-1 occupancy at FRQ is rhythmic; it was provided only to the referees. You performed statistical analysis of rhythmicity using the CircaCompare tool; however, the application of tests and the resulting statistics have not been presented in a transparent or comprehensive manner. It also appears that you incorrectly cited published evidence regarding the rhythmicity of amino acid biosynthetic gene expression. These and a few other issues are fully described below:

1. The description of the CircaCompare statistical tests of rhythmicity is inadequate and needs to be better explained, providing a brief "layman's" description in the Methods of the meaning of the statistical parameters and how the tests were applied to their data. In addition, the results need to be documented completely for all experiments in which rhythmicity was concluded to exist for any parameter. While the authors added time-series statistics in certain figure legends, they did not indicate the p-values for rhythmicity, and the asterisks in figure panels seem to refer to t-tests rather than to RAIN/CircaCompare statistics. To increase transparency, the results of the CircaCompare statistical tests could be summarized in an additional supplementary table.

2. Discrepancies remain regarding whether transcription of amino acid (AA) biosynthetic genes is rhythmic as, contrary to the statement in the Response to Reviewers, the publication by Hurley et al. 2014 does not show ser-2 (NCU01439), trp-3 (NCU08409) or arg-1 (NCU02639) to be rhythmic. The authors should note that the original data were re-analyzed using better statistical tools, described in Suppl Dataset S1 of Hurley et al. (2018), and in the re-analyzed data, ser-2(NCU01439) is not rhythmic at the RNA level, although it is rhythmic at the protein level; trp-3(NCU08409) is rhythmic at the RNA level, and arg-1 (NCU02639)- is rhythmic at the RNA level. The authors need to carefully examine the results of Hurley et al. (2018) and provide an accurate summary of the presence or absence of rhythmicity in the expression of all known CPC-1-induced target genes, possibly with an additional supplementary table that lists the key statistical criteria for the assessments. They should also explain to the reviewers why they incorrectly cited the findings of Hurley et al. (2014) in the revised manuscript.

3. The new ChIP data for CPC-1 binding at the AA genes should be included as a supplementary figure with the addition of the appropriate explanatory text in Results and Discussion, including the suggestion made in their Response to Reviewers of rhythmic binding of WCC or WCC-controlled transcription factors, at the AA genes and whether any evidence exists in the literature for the latter.

4. Clarify whether the authors are interpreting their data to indicate that GCN5/SAGA perform the same functions at FRQ promoter in both non-starved and starved cells and are simply being additionally mobilized by CPC-1 in starved cells, possibly by altering the model in Figure 8.

Reviewer #1 (Recommendations for the authors):

The authors appear to have satisfied some, but not all, of the major issues raised about the previous version of the paper. They have provided new evidence indicating that knock out of GCN5 eliminates rhythmic FRQ expression with a corresponding effect on FRQ mRNA abundance. They also provided new ChIP evidence suggesting that CPC-1 occupancy at FRQ is rhythmic. However, the following issues seem to require additional attention and adequate responses from the authors.

1. To support the conclusions of rhythmic FRQ mRNA abundance, WCC complex binding, H3Ac levels, GCN-5 binding, and CPC-1 binding at the FRQ promoter, and of expression levels of amino acid biosynthetic genes in WT cells, they claim to have conducted a statistical analysis using CircaCompare. However, I could not find even a cursory explanation of this tool and the meaning of the different parameters it evaluates in Methods, nor a description of how it was being applied in comparing results for WT and mutants. In addition, it is generally unclear what data have been analyzed with CircaCompare and what the outcomes were. In particular, the legends often state that data were judged to be arhythmic in a cpc mutant without indicating whether it was significantly rhythmic in WT. What is needed is a table that lists clearly all the CircaCompare tests that were conducted and the P-values for all of the parameters that were analyzed, with a final column indicating whether the data are rhythmic or arhythmic and from what results this assessment was made. In short, for each experiment, it is necessary to provide justification using CircaCompare that the data are significantly rhythmic in WT, not only that they are arhythmic in mutants, and to make the CircaCompare tests and results completely transparent, accessible, and understandable to a general audience.

2. The aforementioned shortcoming is particularly notable regarding the expression of amino acid (AA) biosynthetic genes in WT cells. It is unclear whether the CircaCompare analysis actually confirms rhythmic mRNA expression for these genes in WT cells, both for the group of genes they examined by RT-PCR and those interrogated using published RNA-seq data. It is also unclear why they are showing the published RNA-seq results for only 4 genes out of a presumably much larger number of AA biosynthetic genes known to be induced by CPC-1 in starved cells. Did many other AA genes not show rhythmic transcription? It would useful to have a table summarizing the CircaCompare analysis of the RNA-seq data for all known CPC-1 target genes.

3. This last issue of whether AA biosynthetic gene transcription is truly rhythmic was exacerbated by their new CPC-1 ChIP data (presented only for referees) indicating constitutive CPC-1 occupancy at all four AA biosynthetic genes they tested, in contrast to its apparently rhythmic binding at the FRQ gene. They suggest that the AA biosynthetic genes might exhibit rhythmic binding of WCC, or some other transcriptional activator; however, is there any evidence that WCC binds to these AA genes, or that any other activator besides WCC binds to promoters rhythmically? I remain skeptical of the claim of rhythmic transcription of AA biosynthetic genes that are induced by CPC-1 in starved cells.

Reviewer #2 (Recommendations for the authors):

The authors have adequately addressed my comments and concerns from the previous manuscript version, in some cases with additional experiments. I greatly appreciate their effort. I think this study is a nice contribution to the field and will inform on the metabolic regulation of circadian rhythms at the chromatin level. One minor comment that will help clarify this complex model for readers.

In the previous version, I commented:

"The experiments to examine the involvement of GCN-5 and ADA-2 were performed in normal conditions (no amino acid starvation). Unlike cpc-1 and cpc-3 KO strains, gcn-5 and ada-2 KO strains showed severely disrupted frq rhythms in normal nutrient conditions, suggesting they are normally already required for robust circadian rhythms. If GCN-5 and the SAGA complex are normally involved in regulating H3ac rhythms in the frq loci, how does GCN2 pathway modulates the activity of GCN-5 and SAGA complex in conditions of amino acid starvation? Are the interactions between GCN2/4 with GCN-5 and SAGA complex different in normal vs amino acid starved conditions? "

Based on the additional experiments and their text edit, there appears to be no difference in the activity of GCN-5 and SAGA complex in normal vs amino acid-starved conditions. It's doing what it normally does in normal conditions even when it is in amino acid-starved conditions. Its activity is however more necessary in amino acid-starved conditions because the chromatin at frq promoter is constitutively compacted in amino acid-starved conditions. So since CPC-3 and CPC-1 are necessary to activate the GCN-5/ADA-2 complex and since an amino acid-starved condition activates the GCN2 pathway and induced CPC-1 expression, the arrhythmic phenotype in circadian rhythm is more severe in CPC-3 and CPC-1 KO mutants in amino acid starved condition when compared to normal condition.

I wonder if the model figure (Figure 8) would be more clear to readers if it includes the normal scenario instead of just having the amino acid-starved conditions. Besides that, I think the manuscript is much approved.

Reviewer #3 (Recommendations for the authors):

There are frequent and numerous issues with English usage that should be corrected. There are too many of these to mark each one, and they begin with problems in the Abstract.

Here are a few examples just from the Introduction:

"Circadian clocks are evolved to adapt to the daily environmental changes caused by the earth rotation". Should be earth's rotation "The ability to maintain circadian clock.

54 function in response to various stress and perturbations is an important property of living systems.

55 (Bass, 2012; Hogenesch and Ueda, 2011). Although gene expression is sensitive to temperature.

56 changes, temperature compensation is a key feature of circadian clocks that maintain circadian.

57 period length".

Should read "in response to various stresses" and "compensation is a key feature of circadian clocks that maintains".

Lines 128 and 136, the same sentence appears twice: "3-AT is an inhibitor…".

Lines 168, 170, 171, and 177, cpc-1 KO to cpc-1KO; line 364, "related echanisms(Chaix,"

to "related mechanisms (Chaix,"; line 381, "nourseothricin selection(L." to "nourseothricin selection (L.". Line 460, "using 150bp" to "using 150 bp".

Line 174 could add the citation to Wang et al., 2019 here as was done in citing the same conclusion on lines 95/96.

Line 230 should be "we crossed ras-1[bd] to the gcn-5KO strain obtained from the FGSC" or something to that effect. Likewise in the Materials and methods and elsewhere, "bd" should be ras-1[bd] or ras-1bd. Likewise on line 247 with ada-2 where ras-1[bd] was crossed into the knockout strain.

More generally it is important that strains always be referred to by their correct and full genotypes unless a clear method of abbreviations is stated early in the main text and invariably applied throughout. This is not trivial, especially with chromatin-modifying proteins. For instance, Belden (PLoS Gen 2011) found a synthetic lethality between ∆chd1 and ras-1[bd].

In the model (Figure 8), formation of the CPC-1 oligomer was not confirmed experimentally, and this could be mentioned.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "The nutrient-sensing GCN2 signaling pathway is essential for circadian clock function by regulating histone acetylation under amino acid starvation" for further consideration by eLife. Your revised article has been evaluated by Kevin Struhl (Senior Editor) and a Reviewing Editor.

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

1. In the new Supplementary file S1. The authors have provided lists of 148 cpc-1 up-regulated genes and 127 cpc-1 down-regulated genes, of which 79 up-regulated genes apparently show rhythmic expression and 67 down-regulated genes apparently show rhythmic expression. They have not however provided the statistics of the CircaCompare tests of rhythmicity for the WT strain, which needs to be added for each gene listed in the new file S1. In addition, they should provide new text in the RESULTS explaining that ~53% of both the up- and down-regulated genes exhibit rhythmicity and indicate that this represents a highly significant enrichment of all cpc-1 regulated genes as judged by the hypergeometric distribution test, providing the P-value. The file S1 should also include a notes sheet fully explaining the data provided there.

2. A notes sheet should also be added to the new Supplementary file S2 in which the results in each sheet should be linked to data in the corresponding figures, and also stipulating the difference between results listed in different sheets for FRQ versus frq (Protein vs. mRNA?). In short, each Supplementary file should be understandable as a stand-alone document.

3. There are also some grammatical errors in the revised text that should be corrected.

https://doi.org/10.7554/eLife.85241.sa1

Author response

Essential Revisions:

1) Conduct statistical analysis to assess the significance of the circadian rhythms claimed for FRQ mRNA abundance, and for WC complex binding, H3Ac levels, GCN-5 binding, and CPC-1 binding at the FRQ promoter.

We revised as suggested by reviewer 2 using CircaCompare.

2) The data are not compelling that CPC-1 occupancy at FRQ is rhythmic. If the statistically significant rhythmic binding of CPC-1 is not demonstrated, then the molecular basis of the rhythmic recruitment of Gcn5 and H3Ac will be unclear. In that event, the authors may need to consider the model that CPC-1 assists constitutively in recruiting Gcn5 and attendant H3 acetylation, and that some other factor is involved in producing the circadian rhythms of these events, and in turn, the rhythmic binding of WC complex at the FRQ promoter.

To address this concern, we now performed ChIP assays using the CPC-1 antibody instead of Myc antibody. As shown in the revised Figure 4A, the CPC-1 enrichment at DD14/DD18 was significantly higher than that of DD10/DD22. In addition, we also analyzed the circadian rhythm of CPC-1 binding at the frq promoter with CircaCompare, which showed that the CPC-1 occupancy at the frq promoter was rhythmic.

3) It appears that the experiments to examine the involvement of GCN-5 and ADA-2 in the rhythmic transcription of FRQ were performed only in non-starvation conditions. If this was indeed the case, then at a minimum, key conclusions would have to be qualified. More likely, the experiments would have to be extended to include starvation conditions.

Thanks for the suggestion. We have now examined the rhythmic transcription of frq in both amino acid non-starvation and starvation conditions in the gcn-5KO (Figure 5D and 5E) and ada-2KO mutants (Figure 5—figure supplement 1E and 1F). The results were found to be similar with/without amino acid starvation in the gcn-5KO and ada-2KO mutants.

4) Provide clarification about why knock out of GCN5 reduces FRQ protein expression and diminishes its rhythm without corresponding effects on FRQ mRNA abundance (Figure 5C vs 5D). Are the data lacking in precision or accuracy, or is a more complex model required to explain these results?

To address this concern, we performed RT-qPCR analysis to compare the rhythms of frq mRNA in the WT and gcn-5KO mutant. The results showed that the frq mRNA were decreased in the gcn-5KO mutant with or without 3-AT, especially at DD16h and DD40h, when frq mRNA peaked in WT (Figure 5D and 5E). Thus, the frq mRNA and FRQ protein results are consistent with each other.

5) The data provided to support the claim of rhythmic transcription of the four candidate amino acid biosynthetic genes in Figure 7F is not very compelling. If the statistical analysis fails to confirm their rhythmic transcription in WT, it might be possible to bolster their claim by evaluating published RNA-seq data (sampled every 2 hrs over 2 days, done in triplicate) for known CPC-1 target genes. Because these published experiments were not carried out in amino acid starvation conditions however, the measured expression of the biosynthetic gene transcripts might be dominated by transcription factors that bind constitutively rather than by CPC-1. In the absence of published data supporting rhythmic transcription of amino acid biosynthetic genes regulated by CPC-1, the authors should consider conducting additional RNA-Seq experiments at different time points in amino acid-starved cells. Alternatively, it might be possible to provide RT-qPCR data on additional candidate genes, coupled with statistical analysis of rhythmic transcription. Failing that, the authors would likely need to soften or eliminate claims of rhythmic transcription of these genes.

The rhythmic expression of these candidate amino acid biosynthetic genes are supported by the following evidence. First, we analyzed the published rhythmic RNA-seq data from Dr. Jay Dunlap’s lab (Hurley JM et al., 2014, PMID: 25362047) and found that these amino acid biosynthetic gene (ser-2, trp-3, arg-1) were expressed rhythmically. Our analysis data is now presented in Figure 7—figure supplement 1C. Second, we also performed RT-qPCR analysis to examine the relative mRNA levels of his-3, ser-2, trp-3, and three additional amino acid biosynthetic genes (his-4, arg-1 and arg-10) at different time points (sampled every 6 hrs over 30 hrs). We found that these amino acid biosynthetic genes were rhythmic expressed in the WT but not in the cpc-1KO strain with/without 3-AT treatment (Figure 7 F and Figure 7—figure supplement 1D and 1E).

6) The proposed rhythmic transcription of amino acid biosynthetic genes would also predict rhythmic binding by CPC-1 in their promoters under starvation conditions. Conducting ChIP analysis of CPC-1 binding at a few candidate biosynthetic genes that show convincing rhythmic transcription would be highly desirable, and if observed, might compensate for less compelling transcript data. Such measurements would also be relevant to the question noted above of whether CPC-1 binding at the FRQ promoter itself is truly rhythmic.

To address this concern, we performed ChIP using our CPC-1 antibody to analyze CPC-1 occupancy at the frq promoter and a few amino acid biosynthetic genes including his-3, ser-2 and trp-3. As the results shown in Figure 4A, CPC-1 binding at the frq promoter was rhythmic (measured with CircaCompare). However, the CPC-1 occupancy at the his-3, ser-2 and trp-3 promoters was not rhythmic (measured with CircaCompare). These results suggested that CPC-1 binding is rhythmic at the frq promoter but not at the amino acid biosynthetic genes. Thus, the rhythmic transcription of amino acid biosynthetic genes might be controlled by rhythmic binding of WCC or WCC-controlled transcription factors but not by rhythmic CPC-1 binding.

Author response image 1
Download asset Open asset

7) It is necessary to add new text to address queries raised by Referees #2 and #3 about various findings that require additional clarification or qualification.

Revised as suggested. We have edited the text and revised the figures in the revised manuscript.

8) It is important to provide raw source data for all measurements.

Revised as suggested. Raw source data of measurements are now shown in the revised folder of source data.

Reviewer #1 (Recommendations for the authors):

i) Explain why knock out of GCN5 eliminates rhythmic FRQ expression but does not have a corresponding effect on FRQ mRNA abundance.

Please see response to the Essential revisions point 4.

ii) Determine whether elimination of HDA1 restores H3ac and WCC binding at FRQ, and rhythmic growth, in 3AT-treated cpc-1 and gcn5 mutants.

Because TSA treatment partially rescued the rhythmicity of cpc-3KO strain after 3-AT treatment (Figure 6F), we created hda-1KOcpc-3KO double mutant and performed race tube assay with addition of different concentrations of 3-AT. The results showed that the hda-1KOcpc-3KO mutant was sensitive to 3-AT, becoming arhythmic with 2 or 3 mM 3-AT, which is similar to cpc-3KO strain (data shown in Author response image 2). These results suggested that in addition to HDA-1, other HDACs are also involved in restoring H3ac and WCC binding at the frq promoter in the cpc-3 KO, cpc-1 KO or gcn-5 KO mutants under amino acid starvation conditions. We are planning to further investigate this mechanism in the future.

Author response image 2
Download asset Open asset

iii) Provide stronger evidence for rhythmic transcription of aa biosynthetic genes.

Please see response to Essential revisions point 5.

iv) Possibly attempt to determine whether CPC-1 occupancy at aa biosynthetic genes (and at FRQ itself) is rhythmic.

Please see response to Essential revisions point 6.

Reviewer #2 (Recommendations for the authors):

1. The authors should use circadian statistics (e.g. CircaCompare; Parsons et al. 2020) to compute the phase and amplitude of the mRNA, DNA binding of WC complex, and H3Ac rhythms. This will allow them to compare between rhythms and provide statistical significance values, rather than just providing qualitative descriptions, e.g. "slightly decreased" in lines 189-190. This will be valuable when comparing rhythms between strains and between nutrient conditions.

As suggested, we used CircaCompare to analyze our data.

2. Line 119: The authors indicate that the rhythmicity of histone acetylation at the frq gene promoter is lost in many of the mutants they tested, e.g. CPC-3 and CPC-1. However, it is not clear whether these mutants simply have a lower level of H3ac (Figure 3D), resulting in a lower level of WC binding to frq promoter (Figure 2C-D), and consequently lower level of frq mRNA (Figure 1D). When the levels of these measures are low, it is difficult to observe any rhythms. Furthermore, is it really necessary to emphasize the maintenance of rhythms? When clock mRNA levels are too low, clocks are normally disrupted anyway.

Thanks for the suggestions. We re-analyzed the data using CircaCompare and revised the manuscript.

3. The authors propose that GCN5 works downstream of CPC-1 and CPC-3. If that is the case, why does frq mRNA in gcn-5 KO strain show dampened but consistently high level (Figure 5D), rather than a consistently low level as in the case of cpc-1 and cpc-3 ko strains? Can the authors comment on this? Also, it appears that H3ac and WC binding to frq promoter is lower compared to WT (Figure 5E-F). Can the authors explain why frq mRNA level is higher given H3ac and WC binding to frq promoter is lower?

Please see response to Essential revisions point 4.

4. There is no data in this study showing that the GCN2 pathway is activated in amino acid-starved conditions, only that it is required to maintain robust frq and conidiation rhythms. Can the authors clarify how they are defining "activation of the GCN2 pathway" in this study? For example, is it recruitment of GCN-5 and SAGA complex to frq promoter?

Thanks for the question. CPC-3, the GCN2 homolog in Neurospora, is the only eIF2α kinase responsible for eIF2α phosphorylation at serine 51(Karki S et al. 2020, PMID: 32355000). As shown in the revised Figure 1—figure supplement 1A, the eIF2α phosphorylation and CPC-1 were induced by 3-AT treatment in the WT but not in the cpc-3KO strain. These results demonstrate that the GCN2 pathway is activated by amino acid starvation, and as a result, the CPC-1 expression is activated to recruit the SAGA complex to the frq promoter.

5. The experiments to examine the involvement of GCN-5 and ADA-2 were performed in normal conditions (no amino acid starvation). Unlike cpc-1 and cpc-3 KO strains, gcn-5 and ada-2 KO strains showed severely disrupted frq rhythms, suggesting they are normally required for robust circadian rhythms. If GCN-5 and the SAGA complex are normally involved in regulating H3ac rhythms in the frq loci, how does the GCN2 pathway modulates the activity of GCN-5 and SAGA complex in conditions of amino acid starvation? Are the interactions between GCN2/4 with GCN-5 and SAGA complex different in normal vs amino acid starved conditions? The authors should clarify their model.

As mentioned above, our data suggested that GCN-5 and ADA-2 are required for robust circadian rhythms under normal conditions. As suggested, we did detect dampened rhythmic expression of frq in the gcn-5KO and ada-2KO strains under amino acid starvation (Figure 5D and 5E and Figure 5—figure supplement 1E and 1F). We also performed Co-IP to compare the difference of interactions between CPC-1 with ADA-2 and GCN5 with ADA-2 under normal and amino acid starved conditions. The results showed that although the Myc.GCN-5, MYC.CPC-1 or Flag.ADA-2 protein level was repressed by 3 mM 3-AT treatment (likely due to global translational inhibition by induced eIF2α phosphorylation) (Karki S et al. 2020, PMID: 32355000), the interactions between CPC-1 with ADA-2 and GCN-5 with ADA-2 were almost the same under normal and amino acid starved conditions (IP was normalized with Input) (Figure 4B and 4C). These results indicated that amino acid starved conditions had little impact on the protein interactions between CPC-1 with GCN-5 and SAGA complex.

In our model, we proposed that amino acid starvation resulted in compact chromatin structure (due to decreased H3ac) in the frq promoter in the WT strain (Figure 3B), likely due to activation of histone deacetylases or inhibition of histone acetyltransferases. Amino acid starvation activates GCN2 pathway and induces CPC-1 expression. The induced CPC-1 can recruit GCN5-containing SAGA complex to the frq promoter to loosen the chromatin structure, promoting frq rhythmic transcription under starvation conditions. However, in the cpc-3KO mutants, CPC-1 could not effectively recruit GCN5 containing SAGA complex to frq promoter, resulting in arrhythmic frq transcription. We have now clarified our model in the revised discussion.

6. Given that the GCN2 pathway is important for nutrient sensing, the authors should not disregard the alternative hypothesis that the GCN2 pathway may be important for nutrient compensation and plays a role in maintaining the robustness of rhythms in a range of nutrient conditions.

Thanks for the suggestion. We now discussed the alternative hypothesis in the revised manuscript. “Because GCN2 signaling pathway is important for nutrient sensing, it may be important for nutrient compensation and plays a role in maintaining the robustness of rhythms in a range of nutrient conditions”.

Reviewer #3 (Recommendations for the authors):

There are frequent and numerous issues with English usage that should be corrected. There are too many of these to mark each one, and they begin with problems in the Abstract.

Here are a few examples just from the Introduction:

"Circadian clocks are evolved to adapt to the daily environmental changes caused by the earth rotation". Should be earth's rotation "The ability to maintain circadian clock.

54 function in response to various stress and perturbations is an important property of living systems.

55 (Bass, 2012; Hogenesch & Ueda, 2011). Although gene expression is sensitive to temperature.

56 changes, temperature compensation is a key feature of circadian clocks that maintain circadian.

57 period length".

Should read "in response to various stresses" and "compensation is a key feature of circadian clocks that maintains".

Thanks for the suggestions. The manuscript is revised as suggested.

Lines 128 and 136, the same sentence appears twice: "3-AT is an inhibitor…".

Lines 168, 170, 171, and 177, cpc-1 KO to cpc-1KO; line 364, "related echanisms(Chaix,"

to "related mechanisms (Chaix,"; line 381, "nourseothricin selection(L." to "nourseothricin selection (L.". Line 460, "using 150bp" to "using 150 bp".

Line 174 could add the citation to Wang et al., 2019 here as was done in citing the same conclusion on lines 95/96.

Line 230 should be "we crossed ras-1[bd] to the gcn-5KO strain obtained from the FGSC" or something to that effect. Likewise in the Materials and methods and elsewhere, "bd" should be ras-1[bd] or ras-1bd. Likewise on line 247 with ada-2 where ras-1[bd] was crossed into the knockout strain.

More generally it is important that strains always be referred to by their correct and full genotypes unless a clear method of abbreviations is stated early in the main text and invariably applied throughout. This is not trivial, especially with chromatin-modifying proteins. For instance, Belden (PLoS Gen 2011) found a synthetic lethality between ∆chd1 and ras-1[bd].

Thanks for the suggestions. The manuscript is revised as suggested. The ras-1 bd background was revised in the Materials and methods, and stated in the beginning of the main text “we created cpc-3 and cpc-1 knockout mutants (See Materials and methods)”.

In the model (Figure 8), formation of the CPC-1 oligomer was not confirmed experimentally, and this could be mentioned.

GCN4/CPC-1 was previously reported to bind DNA in the form of dimers (Hope IA, et al., 1987, EMBO J, PMID: 3678204). We revised our model in Figure 8 and edited the description in the text, “the GCN2 kinase responds to nutrient deprivation, whereas it phosphorylates eIF2α and upregulates the master transcription factors GCN4 or CPC-1, respectively, which binds target DNA as a dimer to activate amino acid biosynthetic genes”.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

It appears that you have satisfied some, but not all, of the major issues raised about the previous version of the paper. It was recognized that you provided new evidence indicating that knockout of GCN5 eliminates rhythmic FRQ expression with a corresponding effect on FRQ mRNA abundance. While you also provided new ChIP evidence suggesting that CPC-1 occupancy at FRQ is rhythmic; it was provided only to the referees. You performed statistical analysis of rhythmicity using the CircaCompare tool; however, the application of tests and the resulting statistics have not been presented in a transparent or comprehensive manner. It also appears that you incorrectly cited published evidence regarding the rhythmicity of amino acid biosynthetic gene expression. These and a few other issues are fully described below:

Thanks for the comments to our revised manuscript. We have now revised the manuscript to address all raised concerns.

1. The description of the CircaCompare statistical tests of rhythmicity is inadequate and needs to be better explained, providing a brief "layman's" description in the Methods of the meaning of the statistical parameters and how the tests were applied to their data. In addition, the results need to be documented completely for all experiments in which rhythmicity was concluded to exist for any parameter. While the authors added time-series statistics in certain figure legends, they did not indicate the p-values for rhythmicity, and the asterisks in figure panels seem to refer to t-tests rather than to RAIN/CircaCompare statistics. To increase transparency, the results of the CircaCompare statistical tests could be summarized in an additional supplementary table.

We have now revised the manuscript as suggested. Please see our response to the Reviewer #1.

2. Discrepancies remain regarding whether transcription of amino acid (AA) biosynthetic genes is rhythmic as, contrary to the statement in the Response to Reviewers, the publication by Hurley et al. 2014 does not show ser-2 (NCU01439), trp-3 (NCU08409) or arg-1 (NCU02639) to be rhythmic. The authors should note that the original data were re-analyzed using better statistical tools, described in Suppl Dataset S1 of Hurley et al. (2018), and in the re-analyzed data, ser-2(NCU01439) is not rhythmic at the RNA level, although it is rhythmic at the protein level; trp-3(NCU08409) is rhythmic at the RNA level, and arg-1 (NCU02639)- is rhythmic at the RNA level. The authors need to carefully examine the results of Hurley et al. (2018) and provide an accurate summary of the presence or absence of rhythmicity in the expression of all known CPC-1-induced target genes, possibly with an additional supplementary table that lists the key statistical criteria for the assessments. They should also explain to the reviewers why they incorrectly cited the findings of Hurley et al. (2014) in the revised manuscript.

Thanks for the correction and suggestions. We have revised the manuscript as suggested. Please see our response to the Reviewer #1.

3. The new ChIP data for CPC-1 binding at the AA genes should be included as a supplementary figure with the addition of the appropriate explanatory text in Results and Discussion, including the suggestion made in their Response to Reviewers of rhythmic binding of WCC or WCC-controlled transcription factors, at the AA genes and whether any evidence exists in the literature for the latter.

Thanks for the suggestions. We have revised the manuscript as suggested. Please see our response to the Reviewer #1.

4. Clarify whether the authors are interpreting their data to indicate that GCN5/SAGA perform the same functions at FRQ promoter in both non-starved and starved cells and are simply being additionally mobilized by CPC-1 in starved cells, possibly by altering the model in Figure 8.

We have revised the Figure 8 and manuscript as suggested. Please see our response to the Reviewer #2.

Reviewer #1 (Recommendations for the authors):

The authors appear to have satisfied some, but not all, of the major issues raised about the previous version of the paper. They have provided new evidence indicating that knock out of GCN5 eliminates rhythmic FRQ expression with a corresponding effect on FRQ mRNA abundance. They also provided new ChIP evidence suggesting that CPC-1 occupancy at FRQ is rhythmic. However, the following issues seem to require additional attention and adequate responses from the authors.

1. To support the conclusions of rhythmic FRQ mRNA abundance, WCC complex binding, H3Ac levels, GCN-5 binding, and CPC-1 binding at the FRQ promoter, and of expression levels of amino acid biosynthetic genes in WT cells, they claim to have conducted a statistical analysis using CircaCompare. However, I could not find even a cursory explanation of this tool and the meaning of the different parameters it evaluates in Methods, nor a description of how it was being applied in comparing results for WT and mutants. In addition, it is generally unclear what data have been analyzed with CircaCompare and what the outcomes were. In particular, the legends often state that data were judged to be arhythmic in a cpc mutant without indicating whether it was significantly rhythmic in WT. What is needed is a table that lists clearly all the CircaCompare tests that were conducted and the P-values for all of the parameters that were analyzed, with a final column indicating whether the data are rhythmic or arhythmic and from what results this assessment was made. In short, for each experiment, it is necessary to provide justification using CircaCompare that the data are significantly rhythmic in WT, not only that they are arhythmic in mutants, and to make the CircaCompare tests and results completely transparent, accessible, and understandable to a general audience.

Thanks for the comments and suggestions. We have now revised the manuscript as suggested. We have added in the Methods details of the CircaCompare statistical analysis described in previously published papers (X. Liu et al., 2021; Parsons et al., 2020). The results of the CircaCompare statistical tests were summarized in the Supplementary file 2.

2. The aforementioned shortcoming is particularly notable regarding the expression of amino acid (AA) biosynthetic genes in WT cells. It is unclear whether the CircaCompare analysis actually confirms rhythmic mRNA expression for these genes in WT cells, both for the group of genes they examined by RT-PCR and those interrogated using published RNA-seq data. It is also unclear why they are showing the published RNA-seq results for only 4 genes out of a presumably much larger number of AA biosynthetic genes known to be induced by CPC-1 in starved cells. Did many other AA genes not show rhythmic transcription? It would useful to have a table summarizing the CircaCompare analysis of the RNA-seq data for all known CPC-1 target genes.

Thanks for the suggestions. In the previous manuscript, we only cited the Hurley et al. (2014) paper, because we downloaded and re-analyzed the RNA-seq data in that study. However, the data were re-analyzed using improved statistical methods described in the Supplemental Dataset 1 (https://doi.org/10.17632/r68j3rnxhw.1) of Hurley et al. (2018). We have now cited both papers in our manuscript.

In our study, there were 148 up-regulated and 127 down-regulated genes found in the WT strain after 3-AT treatment but their regulation were abolished in the cpc-1KO strain (Figure 7D and Figure 7—figure supplement 1A), suggesting that these genes were regulated by CPC-1 under amino acid starvation. As shown in the Supplementary file 1, we summarized the rhythmic expression of 79 up-regulated and 67 down-regulated CPC-1 target genes based on the results of Hurley et al. (2018) (its Supplemental Dataset 1). We also performed RT-qPCR experiments to re-analyze the expression rhythms for some of those genes and found that the rhythmic expression of his-3 (NCU03139), trp-3 (NCU08409) and arg-1 (NCU02639) in the WT was abolished in the cpc-1KO strain under amino acid starvation, which is consistent with the published RNA-seq analysis results. In addition, we found that ser-2 (NCU01439) was also rhythmic in the WT but not in the cpc-1KO strain under amino acid starvation, even though it was not previously shown to be rhythmic in the published RNA-seq analysis study. These results suggest that many CPC-1 activated metabolic genes under amino acid starvation are regulated by circadian clock.

3. This last issue of whether AA biosynthetic gene transcription is truly rhythmic was exacerbated by their new CPC-1 ChIP data (presented only for referees) indicating constitutive CPC-1 occupancy at all four AA biosynthetic genes they tested, in contrast to its apparently rhythmic binding at the FRQ gene. They suggest that the AA biosynthetic genes might exhibit rhythmic binding of WCC, or some other transcriptional activator; however, is there any evidence that WCC binds to these AA genes, or that any other activator besides WCC binds to promoters rhythmically? I remain skeptical of the claim of rhythmic transcription of AA biosynthetic genes that are induced by CPC-1 in starved cells.

Thanks for the concern of this review. Our results and those of Hurley et al. (2018) demonstrated that CPC-1 is required for the rhythmic expression of many CPC-1 targeted metabolic genes. We agree with this reviewer that the exact mechanism for how this regulation is achieved is still unclear since our ChIP experiments showed that the CPC-1 binding at several selected amino acid biosynthetic genes was not rhythmic (Figure 7—figure supplement 1C). There are several possibilities. First, the rhythms of CPC-1 binding at these amino acid biosynthetic genes have very low amplitude and could not be detected by our ChIP assays. Second, the rhythmic transcription of these genes could be controlled by rhythmic binding of WCC or WCC-controlled transcription factors (Hurley et al. 2014). Indeed, the ChIP-seq analysis by Hurley et al. (2014) showed that WC-2 is associated at the promoter of arg-1 (NCU02639) (https://www.pnas.org/doi/suppl/10.1073/pnas.1418963111/suppl_file/pnas.1418963111.sd11.xlsx). Third, although the binding of CPC-1 is arrhythmic, its activity in activating transcription can still be rhythmic, which could be due to rhythmic posttranslational modification/association of other factors.

We have now revised the manuscript to address this concern. In the results, we described “We performed ChIP experiments to examine whether CPC-1 directly activates the expression of amino acid synthetic genes. As shown in Figure 7—figure supplement 1C, CPC-1 was found to be constitutively enriched at the promoters of his-3 (NCU03139), trp-3 (NCU08409) and ser-2 (NCU01439) genes, and the enrichment was enhanced by 3-AT treatment”. In the discussion, we added that “Although CPC-1 rhythmically binds at the frq promoter, ChIP experiments showed that CPC-1 binding at several selected amino acid biosynthetic genes did not appear to be rhythmic (Figure 7—figure supplement 1C). However, our RT-qPCR results (Figure 7F and Figure 7—figure supplement 1) and the previous RNA-seq data showed that many CPC-1 targeted metabolic genes were rhythmic expressed (Hurley et al., 2018). It is possible that the binding of CPC-1 on these promoters are still rhythmic but with low amplitudes. As a result, the limited sensitivity of our ChIP assays failed to detected these rhythms. Alternatively, the rhythmic transcription of these genes might be controlled by rhythmic transcriptional activation activity of CPC-1 rather than its binding. In addition, the rhythmic binding of WCC or WCC-controlled transcription factors (Hurley et al., 2014) might also contribute to their rhythmic transcription.”

Reviewer #2 (Recommendations for the authors):

The authors have adequately addressed my comments and concerns from the previous manuscript version, in some cases with additional experiments. I greatly appreciate their effort. I think this study is a nice contribution to the field and will inform on the metabolic regulation of circadian rhythms at the chromatin level. One minor comment that will help clarify this complex model for readers.

In the previous version, I commented:

"The experiments to examine the involvement of GCN-5 and ADA-2 were performed in normal conditions (no amino acid starvation). Unlike cpc-1 and cpc-3 KO strains, gcn-5 and ada-2 KO strains showed severely disrupted frq rhythms in normal nutrient conditions, suggesting they are normally already required for robust circadian rhythms. If GCN-5 and the SAGA complex are normally involved in regulating H3ac rhythms in the frq loci, how does GCN2 pathway modulates the activity of GCN-5 and SAGA complex in conditions of amino acid starvation? Are the interactions between GCN2/4 with GCN-5 and SAGA complex different in normal vs amino acid starved conditions? "

Based on the additional experiments and their text edit, there appears to be no difference in the activity of GCN-5 and SAGA complex in normal vs amino acid-starved conditions. It's doing what it normally does in normal conditions even when it is in amino acid-starved conditions. Its activity is however more necessary in amino acid-starved conditions because the chromatin at frq promoter is constitutively compacted in amino acid-starved conditions. So since CPC-3 and CPC-1 are necessary to activate the GCN-5/ADA-2 complex and since an amino acid-starved condition activates the GCN2 pathway and induced CPC-1 expression, the arrhythmic phenotype in circadian rhythm is more severe in CPC-3 and CPC-1 KO mutants in amino acid starved condition when compared to normal condition.

I wonder if the model figure (Figure 8) would be more clear to readers if it includes the normal scenario instead of just having the amino acid-starved conditions. Besides that, I think the manuscript is much approved.

Thanks for the comments and suggestions. It’s correct that the interaction between CPC-1 with ADA-2 and GCN5 is not changed by amino acid starvation. However, the activity of GCN-5 and SAGA complex is more important in the cpc-3KO and cpc-1KO mutants compared to the WT strain under amino acid starved conditions. Because the chromatin structure at the frq promoter is constitutively compacted by amino acid starvation, CPC-1 protein level is a limiting factor for efficiently recruiting GCN-5 and SAGA complex to loosen the chromatin to drive rhythmic frq expression under amino acid starved conditions. As suggested, we have now revised the model in Figure 8 and edited the figure legend and discussion to make it more clear to readers.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

1. In the new Supplementary file S1. The authors have provided lists of 148 cpc-1 up-regulated genes and 127 cpc-1 down-regulated genes, of which 79 up-regulated genes apparently show rhythmic expression and 67 down-regulated genes apparently show rhythmic expression. They have not however provided the statistics of the CircaCompare tests of rhythmicity for the WT strain, which needs to be added for each gene listed in the new file S1. In addition, they should provide new text in the RESULTS explaining that ~53% of both the up- and down-regulated genes exhibit rhythmicity and indicate that this represents a highly significant enrichment of all cpc-1 regulated genes as judged by the hypergeometric distribution test, providing the P-value. The file S1 should also include a notes sheet fully explaining the data provided there.

Thanks for the comments to our revised manuscript. We have now revised the manuscript to address the concerns.

First, it should be noted that the published RNA-seq data were first generated by Hurley et al. in 2014 and were re-analyzed by them using the improved statistical tool eJTK Cycle in 2018. The P-values of eJTK Cycle analysis (from the Supplemental Datasets 1 and 2 of Hurley et al., 2018) has now been added in our revised Supplementary file 1. CircaCompare is used to compare the differences between two rhythmic datasets and CircaSingle (Parsons et al., 2020) is used to analyze the rhythmic parameters of a single group. As suggested by the reviewer, we re-analyzed the rhythmicity of CPC-1 targeted genes (148 up-regulated and 127 down-regulated) using CircaSingle and added the P-values in the revised Supplementary file 1. There were 146 rhythmic genes based on the eJTK Cycle and 132 rhythmic genes based on the CircaSingle. There were 106 overlapping genes between the two sets of data, confirming the results using eJTK Cycle. Thus, we performed further analysis based on the data from eJTK Cycle.

Second, we have now added the description of the rhythmicity of CPC-1 targeted genes in the Results “There were 146/275 (53%) of the CPC-1 up- and down-regulated genes under amino acid starvation exhibiting rhythmicity, indicating a highly significant enrichment of CPC-1 regulated genes as clock-controlled genes (p = 3.341905e-06, hypergeometric distribution test)”. The rhythmic genes and all detected genes of RNA-seq data from Hurley et al. 2018 were added in the revised Supplementary file 1, which was used for the hypergeometric distribution test.

Third, we have now revised the Supplementary file 1 by adding a notes sheet.

2. A notes sheet should also be added to the new Supplementary file S2 in which the results in each sheet should be linked to data in the corresponding figures, and also stipulating the difference between results listed in different sheets for FRQ versus frq (Protein vs. mRNA?). In short, each Supplementary file should be understandable as a stand-alone document.

Thanks for the suggestions. We have now added a notes sheet for Supplementary file 2 and stipulated the information for FRQ protein and frq mRNA.

3. There are also some grammatical errors in the revised text that should be corrected.

Thanks for the suggestion. We have carefully checked and improved the English in the revised manuscript.

https://doi.org/10.7554/eLife.85241.sa2

Article and author information

Author details

  1. Xiao-Lan Liu

    State Key Laboratory of Mycology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Validation, Investigation, Writing - original draft, Writing - review and editing
    Contributed equally with
    Yulin Yang
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-1755-3387

  2. Yulin Yang

    1. State Key Laboratory of Mycology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
    2. College of Life Sciences, University of the Chinese Academy of Sciences, Beijing, China
    Contribution
    Conceptualization, Formal analysis, Validation, Investigation
    Contributed equally with
    Xiao-Lan Liu
    Competing interests
    No competing interests declared

  3. Yue Hu

    State Key Laboratory of Mycology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
    Contribution
    Formal analysis, Validation, Investigation
    Competing interests
    No competing interests declared

  4. Jingjing Wu

    1. State Key Laboratory of Mycology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
    2. College of Life Sciences, University of the Chinese Academy of Sciences, Beijing, China
    Contribution
    Formal analysis, Validation, Investigation
    Competing interests
    No competing interests declared

  5. Chuqiao Han

    1. State Key Laboratory of Mycology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
    2. School of Life Sciences, Yunnan University, Kunming, Yunnan, China
    Contribution
    Formal analysis, Validation, Investigation
    Competing interests
    No competing interests declared

  6. Qiaojia Lu

    1. State Key Laboratory of Mycology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
    2. College of Life Sciences, University of the Chinese Academy of Sciences, Beijing, China
    Contribution
    Formal analysis, Validation, Investigation
    Competing interests
    No competing interests declared

  7. Xihui Gan

    Key Laboratory of Gene Engineering of the Ministry of Education, State Key Laboratory of Biocontrol, School of Life Sciences, Sun Yat-sen University, Guangzhou, China
    Contribution
    Investigation
    Competing interests
    No competing interests declared

  8. Shaohua Qi

    MOA Key Laboratory of Soil Microbiology, College of Biological Sciences, China Agricultural University, Beijing, China
    Contribution
    Investigation
    Competing interests
    No competing interests declared

  9. Jinhu Guo

    Key Laboratory of Gene Engineering of the Ministry of Education, State Key Laboratory of Biocontrol, School of Life Sciences, Sun Yat-sen University, Guangzhou, China
    Contribution
    Resources, Formal analysis
    Competing interests
    No competing interests declared

  10. Qun He

    MOA Key Laboratory of Soil Microbiology, College of Biological Sciences, China Agricultural University, Beijing, China
    Contribution
    Resources, Formal analysis, Funding acquisition
    Competing interests
    No competing interests declared

  11. Yi Liu

    Department of Physiology, University of Texas Southwestern Medical Center, Dallas, United States
    Contribution
    Resources, Formal analysis, Funding acquisition, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8801-9317

  12. Xiao Liu

    1. State Key Laboratory of Mycology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
    2. College of Life Sciences, University of the Chinese Academy of Sciences, Beijing, China
    Contribution
    Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Writing - original draft, Writing - review and editing
    For correspondence
    liux@im.ac.cn
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6053-132X

Funding

National Natural Science Foundation of China (32170092)

  • Xiao Liu

National Natural Science Foundation of China (31970079)

  • Xiao Liu

National Key Research and Development Program of China (2021YFA0911300)

  • Xiao Liu

Strategic Priority Research Program of the Chinese Academy of Sciences (XDA28030402)

  • Xiao Liu

Beijing Natural Science Foundation (5202020)

  • Xiao Liu

CAS Interdisciplinary Innovation Team

  • Xiao Liu

National Natural Science Foundation of China (32200056)

  • Xiao-Lan Liu

National Key Research and Development Program of China (2018YFA0900500)

  • Qun He

National Natural Science Foundation of China (32170560)

  • Qun He

National Institutes of Health (R35 GM118118)

  • Yi Liu

Welch Foundation (I-1560)

  • Yi Liu

The funders had no role in study design, data collection, and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Dr. Luis Larrondo for providing the translational fused FRQ-LUC plasmid and strain, and Dr. Shaojie Li for providing Neurospora mutant strains. We thank Dr. Jay Dunlap and Dr. Jennifer Hurley for providing Neurospora RNA-seq information. We also thank Dr. Linqi Wang, Dr. Wenbing Yin, and members of our lab for assistance and discussions. This work was supported by grants from National Natural Science Foundation of China (32170092, 31970079), National Key Research and Development Program of China (2021YFA0911300), Strategic Priority Research Program of the Chinese Academy of Sciences (XDA28030402), Beijing Natural Science Foundation (5202020), and CAS Interdisciplinary Innovation Team to Xiao Liu; National Natural Science Foundation of China (32200056) to Xiaolan Liu; National Key Research and Development Program of China (2018YFA0900500) and National Natural Science Foundation of China (32170560) to Qun He; National Institutes of Health (R35 GM118118) and the Welch Foundation (I-1560) to Yi Liu. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Senior Editor

  1. Kevin Struhl, Harvard Medical School, United States

Reviewing Editor

  1. Alan G Hinnebusch, Eunice Kennedy Shriver National Institute of Child Health and Human Development, United States

Reviewers

  1. Joanna Chiu
  2. Jay C Dunlap, Geisel School of Medicine at Dartmouth, United States

Publication history

  1. Received: November 29, 2022
  2. Preprint posted: December 12, 2022 (view preprint)
  3. Accepted: April 20, 2023
  4. Accepted Manuscript published: April 21, 2023 (version 1)
  5. Version of Record published: May 17, 2023 (version 2)

Copyright

© 2023, Liu, Yang et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 436
    Page views
  • 132
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Xiao-Lan Liu
  2. Yulin Yang
  3. Yue Hu
  4. Jingjing Wu
  5. Chuqiao Han
  6. Qiaojia Lu
  7. Xihui Gan
  8. Shaohua Qi
  9. Jinhu Guo
  10. Qun He
  11. Yi Liu
  12. Xiao Liu
(2023)
The nutrient-sensing GCN2 signaling pathway is essential for circadian clock function by regulating histone acetylation under amino acid starvation
eLife 12:e85241.
https://doi.org/10.7554/eLife.85241

Categories and tags

Research organism

Further reading

    1. Cell Biology
    2. Chromosomes and Gene Expression

    Cohesin-independent STAG proteins interact with RNA and R-loops and promote complex loading

    Hayley Porter, Yang Li ... Suzana Hadjur
    Research Article Updated

    Most studies of cohesin function consider the Stromalin Antigen (STAG/SA) proteins as core complex members given their ubiquitous interaction with the cohesin ring. Here, we provide functional data to support the notion that the SA subunit is not a mere passenger in this structure, but instead plays a key role in the localization of cohesin to diverse biological processes and promotes loading of the complex at these sites. We show that in cells acutely depleted for RAD21, SA proteins remain bound to chromatin, cluster in 3D and interact with CTCF, as well as with a wide range of RNA binding proteins involved in multiple RNA processing mechanisms. Accordingly, SA proteins interact with RNA, and R-loops, even in the absence of cohesin. Our results place SA1 on chromatin upstream of the cohesin ring and reveal a role for SA1 in cohesin loading which is independent of NIPBL, the canonical cohesin loader. We propose that SA1 takes advantage of structural R-loop platforms to link cohesin loading and chromatin structure with diverse functions. Since SA proteins are pan-cancer targets, and R-loops play an increasingly prevalent role in cancer biology, our results have important implications for the mechanistic understanding of SA proteins in cancer and disease.

    1. Cell Biology

    Identification of paired-related Homeobox Protein 1 as a key mesenchymal transcription factor in pulmonary fibrosis

    Emmeline Marchal-Duval, Méline Homps-Legrand ... Arnaud A Mailleux
    Research Article

    Matrix remodeling is a salient feature of idiopathic pulmonary fibrosis (IPF). Targeting cells driving matrix remodeling could be a promising avenue for IPF treatment. Analysis of transcriptomic database identified the mesenchymal transcription factor PRRX1 as upregulated in IPF. PRRX1, strongly expressed by lung fibroblasts, was regulated by a TGF-b/PGE2 balance in vitro in control and IPF human lung fibroblasts, while IPF fibroblast-derived matrix increased PRRX1 expression in a PDGFR dependent manner in control ones. PRRX1 inhibition decreased human lung fibroblast proliferation by downregulating the expression of S phase cyclins. PRRX1 inhibition also impacted TGF-β driven myofibroblastic differentiation by inhibiting SMAD2/3 phosphorylation through phosphatase PPM1A upregulation and TGFBR2 downregulation, leading to TGF-β response global decrease. Finally, targeted inhibition of Prrx1 attenuated fibrotic remodeling in vivo with intra-tracheal antisense oligonucleotides in bleomycin mouse model of lung fibrosis and ex vivo using human and mouse precision-cut lung slices. Our results identified PRRX1 as a key mesenchymal transcription factor during lung fibrogenesis.

    1. Cell Biology
    2. Neuroscience

    APOE expression and secretion are modulated by mitochondrial dysfunction

    Meghan E Wynne, Oluwaseun Ogunbona ... Victor Faundez
    Research Article Updated

    Mitochondria influence cellular function through both cell-autonomous and non-cell autonomous mechanisms, such as production of paracrine and endocrine factors. Here, we demonstrate that mitochondrial regulation of the secretome is more extensive than previously appreciated, as both genetic and pharmacological disruption of the electron transport chain caused upregulation of the Alzheimer’s disease risk factor apolipoprotein E (APOE) and other secretome components. Indirect disruption of the electron transport chain by gene editing of SLC25A mitochondrial membrane transporters as well as direct genetic and pharmacological disruption of either complexes I, III, or the copper-containing complex IV of the electron transport chain elicited upregulation of APOE transcript, protein, and secretion, up to 49-fold. These APOE phenotypes were robustly expressed in diverse cell types and iPSC-derived human astrocytes as part of an inflammatory gene expression program. Moreover, age- and genotype-dependent decline in brain levels of respiratory complex I preceded an increase in APOE in the 5xFAD mouse model. We propose that mitochondria act as novel upstream regulators of APOE-dependent cellular processes in health and disease.