Monday, June 28, 2010

A clever genetic strategy for the study of circadian output pathways
Circadian clocks control a large number of daily processes in most organisms. These endogenous cellular timekeepers regulate rhythms in gene expression, physiology and behaviour and enable organisms to anticipate predictable environmental variations.

Circadian clocks are composed of a central oscillator and two signaling pathways: input pathways convey external signals to the oscillator, so that it can be synchronized with the environment and output pathways allow the oscillator to temporally regulate diverse cellular processes.

The ascomycete Neurospora crassa, instrumental to the history of molecular biology (See The Almighty Fungi: The Revolutionary Neurospora crassa), has one of the best-understood circadian systems, in which a molecular negative feedback loop involving FREQUENCY and the WHITE COLLAR (WCC) complex lies at its core. Briefly, the WCC directly activates transcription of frq, and levels of FRQ protein start then to build up. FRQ inhibits its own expression by modulating the activity of the WCC, which results in the daily oscillation of both the frq mRNA and protein levels, and ultimately, in the rhythmic expression of a variety of clock-controlled genes (ccgs).

Despite the extensive knowledge accumulated on the molecular basis of some eukaryotic oscillators, including Neurospora’s, and the identification of a number of ccgs, little is known about the mechanisms that allow central oscillators to temporally control gene expression and the activity of different clock targets.

In this post, I will discuss a fascinating article from the Bell-Pedersen lab aiming at characterizing circadian output pathways in Neurospora.

A clever genetic selection

In order to identify players involved in output pathways in Neurospora, the authors focused on two well-characterized ccgs: ccg-1 and ccg-2. These genes depend on a functional FRQ-based oscillator to cycle and both peak on the subjective dawn. Clock-controlled gene-2 encodes a hydrophobin, a protein that coats the outer layer of asexual spores (conidiospores) and helps maintain the cell-surface hydrophobicity essential for their air dispersal. Loss of ccg-2 leads to a phenotype in which conidiospores “have a darkened, wetted appearance and mix readily with water” (the so-called Eas phenotype). On the other hand, the function of ccg-1 is unknown and no discernible phenotype is evident in the null strain.

The authors observed, just as it had already been reported, that on a WT strain, both ccg-1 and ccg-2 mRNA levels peak on the subjective morning and have their lowest levels in the subjective early evening. The authors noticed that in a strain carrying a frq null mutation (frq10), this pattern was altered, but in a different way for each gene. While ccg-1 mRNA levels appear to be high throughout the circadian day in the frq-less strain, ccg-2 mRNA levels were low at both time points evaluated (Figure 1). 

Figure 1. Differential clock regulation of ccg-1 and ccg-2. The mRNA levels of ccg-1 and ccg-2 were assayed by Northern blot analysis in FRQ+ and FRQ null strains at two different time points. Copyright 2004 by the Genetics Society of America

OK, but what does this have to do with trying to identify actors involved in output pathways?

The authors reasoned that mutations in genes involved in the circadian output pathways mediating ccg-1 and ccg-2 expression could be identified by finding mutant strains expressing low levels of ccg-1 mRNA or high levels of ccg-2 mRNAs at the subjective early evening in a frq-null background (see Figure 1).

Based on a previous report, the authors implemented a clever genetic selection to identify such strains. The promoter regions of these genes, containing their clock regulatory elements, were cloned in front of the mtr (methyl tryptophan resistance) gene. This gene is required for the uptake of certain amino acids. Loss of mtr function can be selected for on the basis of resistance to certain toxic amino acid analogs such as FPA. Gain of mtr function, on the other hand, can be selected for on a high-arginine (Arg)/low-tryptophan (Trp) medium in a trp-2 mutant background. Under these conditions, arginine blocks the basic amino acid transporter and the permease encoded by the mtr gene becomes essential, as no tryptophan can be synthesized without TRP-2.

The ccg-1::mtr and ccg-2::mtr fusions were then transformed into a frq10 mtr trp-2 strain, generating the CCG1M and CCG2M strains, respectively. As expected, the control (untransformed) strain grew in the presence of FPA, but failed to grow in High Arg/low Trp media, reflecting the MTR- phenotype.

Strain CCG1M grew in High Arg/low Trp media but not in the presence of FPA, suggesting high expression of the mtr transgene and the ability of this strain to import both Trp and FPA. On the contrary, strain CCG2M grew in the presence of FPA but not on High Arg/low Trp media, suggesting a low-level expression of the ccg-2::mtr fusion, as expected.

Identifying output pathway players

Having established this genetic selection system, the authors then set out to identify mutant strains in which the expression of the ccg-controlled mtr transgene is deregulated: they subjected CCG1M and CCG2M strains to UV mutagenesis in order to obtain CCG1M-derived strains that grow on FPA-containing media (that is, that exhibit low levels of the transgene) and CCG2M-derived strains that grow on High Arg/low Trp media (reflecting high expression levels of the transgene), as opposed to their corresponding non-mutagenized parental strains.

The authors indeed were able to obtain FPA-sensitive CCG2M mutant strains, suggesting a high-level expression of the mtr transgene. These strains, however, were sterile in crosses and were not characterized further.

The really interesting results come from the CCG1M mutant strains, though. A number of FPA-resistant CCG1M mutant strains where obtained and they displayed a low-level expression of the mtr transgene, as suspected. Further, they exhibited an altered expression pattern of the endogenous ccg-1 gene. The endogenous ccg-1 mRNA levels where assayed at the subjective early evening, which as mentioned, is a time of day when cycling ccg-1 levels are normally low in wild-type strains and high in frq10 strains (Figure 1). The authors identified mutant strains (COP strains) in which the endogenous ccg-1 mRNA levels where reduced to around the same low levels found in the WT strain and chose these for further study (see Figure 2).

Figure 2. Endogenous ccg-1 mRNA levels in selected mutant strains at the subjective early evening. COP strains correspond to the mutant strains derived from CCG1M. Copyright 2004 by the Genetics Society of America

Among these strains, two phenotypic groups could be observed. One group was similar to the parental CCG1M strain and the other one had a phenotype resembling the ccg-2 loss-of-function phenotype (an Eas-like phenotype). The latter is remarkable, considering that it was obtained through a screen aimed at identifying strains with a dysregulated ccg-1 expression pattern. The strains in this group indeed had low endogenous levels of ccg-2 mRNA, suggesting that a single mutation could affect the expression of not only ccg-1 but other ccgs. Indeed, the authors showed that in these strains, the effects on ccg-1 and ccg-2 expression appeared to be the result of a single mutation: the Eas-like phenotype always segregated with the low-level ccg-1 mRNA one.

The authors mention that one mechanism that could explain the effects of this single mutation on the expression of these two ccgs, is that it “simultaneously increases the activity of a repressor of ccg-1 and decreases activity of an activator of ccg-2”, in the output pathways. One could argue though, that the mutations observed in the CCG1M mutant strains displaying the Eas-like phenotype could be in fact affecting the central oscillator rather than output pathways. This could also explain why the expression of different clock controlled genes appears to be deregulated.

The band (bd) mutation in N. crassa allows clear visualization of the circadianly regulated process of asexual spore formation (conidiation) and evaluation of this phenotype has been historically used as an indirect measurement of the status of the FRQ-based oscillator (for an example, see here)

In order to examine if the FRQ-based oscillator is affected in these group of mutant strains, they were crossed to a bd frq+ strain and the conidiation rhythms assayed. Note that this assay cannot be done in the original frq10 CCG1M mutant strains, as strains bearing the frq10 allele are arrhythmic.

The authors observed a period defect in some of the mutant strains (a shorter period in conidiation as compared to the control strain) and also examined the circadian rhythms in ccg-1 and ccg-2 mRNA levels, noticing that they did not display a proper and typical circadian expression pattern.

The results from the conidiation rhythm assays are consistent with the mutations affecting either output pathways controlling conidiation or the central oscillator itself. The results from the mRNA levels could also point to a defect in the central oscillator.

This “dispute” could easily be settled by measuring frq mRNA levels on a circadian time course and evaluating if it is cycling properly in these strains. This would suggest that the FRQ-based oscillator is unaltered. The authors, however, did not test this and instead argue that the lack of correlation between the conidiation rhythm assays (in which some of the strains had a period defect, but are still rhythmic) and the molecular assays (in which these same strains displayed somewhat arrhythmic ccg-1 and ccg-2 mRNA levels), suggest that the mutations must be affecting output pathways rather than the oscillator itself.

Interestingly, the authors also identified a strain in which the expression of ccg-1 but not of ccg-2 was altered, suggesting that at least in the case of that strain, the mutation is affecting a regulatory element downstream of the oscillator.

In conclusion, the authors were able to find mutant strains that appear to be affected in circadian output pathways. Output pathways remain poorly characterized in the Neurospora circadian system (and in the other circadian model systems, as well), and through a clever genetic approach the authors were able to identify mutant strains that can help address the attractive question of how the information is passed from the oscillator to the gene and gene products that ultimately participate in the overt rhythms present in most organisms.

What genes are affected in these mutant strains? Is really the FRQ-based oscillator intact in these strains? Stay tuned for when I blog about the next chapter in this fascinating story.


Vitalini MW, Morgan LW, March IJ, & Bell-Pedersen D (2004). A genetic selection for circadian output pathway mutations in Neurospora crassa. Genetics, 167 (1), 119-29 PMID: 15166141