Transcriptional responses to fatty acid are coordinated by combinatorial control

Authors


Abstract

In transcriptional regulatory networks, the coincident binding of a combination of factors to regulate a gene implies the existence of complex mechanisms to control both the gene expression profile and specificity of the response. Unraveling this complexity is a major challenge to biologists. Here, a novel network topology-based clustering approach was applied to condition-specific genome-wide chromatin localization and expression data to characterize a dynamic transcriptional regulatory network responsive to the fatty acid oleate. A network of four (predicted) regulators of the response (Oaf1p, Pip2p, Adr1p and Oaf3p) was investigated. By analyzing trends in the network structure, we found that two groups of multi-input motifs form in response to oleate, each controlling distinct functional classes of genes. This functionality is contributed in part by Oaf1p, which is a component of both types of multi-input motifs and has two different regulatory activities depending on its binding context. The dynamic cooperation between Oaf1p and Pip2p appears to temporally synchronize the two different responses. Together, these data suggest a network mechanism involving dynamic combinatorial control for coordinating transcriptional responses.

Synopsis

The direct regulation of a class of genes by a combination of factors is represented by multi-input motifs in transcriptional regulatory networks. Condition-specific formation of these motifs can control the transcription profiles of the target genes (i.e. the speed, stability or duration of the responses). Additionally, comprehensive analysis of yeast regulators and their targets suggests that combinatorial control involving these motifs is likely a prevalent mechanism to control specificity of transcriptional responses. Given the importance of this aspect of regulation, we aimed to identify combinatorial control mechanisms and characterize their properties.

We comprehensively identified multi-input motifs in a network of four yeast regulators and their targets that formed in response to an environmental stimulus. By identifying functionally relevant trends in the network structure, we characterized novel properties of these motifs. We found that multiple-related multi-input motifs can form in response to a stimulus, each with different regulatory mechanisms and outputs. In this context, a single factor can divergently regulate and temporally synchronize different responses to the same stimulus through its involvement in multiple multi-input motifs. The analysis and the results are summarized below.

The response studied was that of yeast to the fatty acid oleate, which is very well suited to the study of condition-specific multi-input motifs because many genes upregulated by fatty acids are conditionally controlled by one of two multi-input motifs, targeted by either Oaf1p and Pip2p (OP), or by Oaf1p, Pip2p and Adr1p (AOP). In addition, there are a variety of other responses to fatty acids including transient upregulation of oxidative stress response genes and a corresponding downregulation of general stress response genes. These responses are likely related to oleate-induced uncoupling of the respiratory chain, but the mechanisms of regulation have not yet been elucidated.

To characterize the response, we first comprehensively analyzed the targets of the three oleate-responsive factors discussed above and Oaf3p, a fourth uncharacterized factor implicated in the response, by large-scale genome localization analysis both in the presence and absence of oleate. The results were represented graphically as protein–DNA interaction networks. To characterize the structure of the networks, we used a straightforward statistical analysis of network motifs that form in response to oleate exposure that can easily be applied to the study of other transcriptional responses. Targets were sorted based on their network topology and significantly overrepresented multi-input motifs were identified using the cumulative distribution function (CDF) (see Figure 1). Next, we characterized the type of genes targeted by each multi-input motif and the influence of each network factor on each cluster. To do this, we overlaid oleate-specific large-scale data sets onto the network and identified significant overrepresentation of gene attributes in each cluster using CDF. These data sets included gene ontology annotations, transcription factor binding site motifs and time-course gene expression profiles. In addition to these data sets, we generated and overlaid microarray data measuring the response to the deletion of each of the four factors in the presence of oleate.

The results supported data in the literature and revealed new insight into the coordinate network function. In the presence of oleate, the network became larger and more cooperative. This was primarily due to the formation of three significantly overrepresented multi-input motifs represented by the AOY cluster (targeted by Adr1p, Oaf1p and Oaf3p), the AOPY cluster (targeted by all four factors) and the OPY cluster (targeted by Oaf1p, Pip2p and Oaf3p). Genes related to peroxisomes that are upregulated by oleate were enriched in the OPY and AOPY clusters. Further analysis suggested that in addition to the known positive regulators of these genes (Oaf1p, Adr1p and Pip2p), Oaf3p weakly negatively regulates these genes in response to oleate. The third overrepresented cluster, the AOY cluster, is enriched for general stress response genes that are transiently downregulated by oleate. Interestingly, Oaf1p is a negative regulator of this cluster (in contrast to its positive effect on the OPY and AOPY clusters). The control of this dual function appears to be exerted by the binding context as OPY and AOPY clusters are enriched for oleate response elements (bound by Oaf1p/Pip2p heterodimers), but not the AOY cluster. These data suggest that the network controls multiple responses through the formation of multiple-related multi-input motifs. The regulatory mechanisms here appear to be complex as each involves a combination of both positive and negative regulators.

Although the chromatin localization data were collected at two time points of oleate induction (0 and 5 h), several of the data sets used for the analysis are more dynamic, including two different time-course microarray data sets, and analysis of protein levels of selected targets from the three most significantly overrepresented clusters at 0, 5 and 20 h. These data enabled generation of a dynamic model of the response to characterized network further. All available data were used to inform the model at the 5 h time point and the time course data sets were used to predict the network structure before and after the 5 h time point. The model revealed that the two responses controlled by the network are temporally regulated, while the general stress response is immediate and transient, the peroxisome response is delayed and long lasting. The data suggest a regulatory mechanism whereby Oaf1p is a negative regulator of AOY genes early in the response to oleate, but as Pip2p levels increase by feedforward activation, increasingly more Oaf1p molecules heterodimerize with Pip2p molecules, drawing Oaf1p molecules away from AOY targets to OPY and AOPY targets, where it positively activates transcription as a heterodimer with Pip2p. The analysis suggests that involvement of a regulator in multiple multi-input motifs can differently control and synchronize two different responses.

In summary, this study suggests that multi-input motifs control not only the specificity of transcriptional responses, but the formation of multiple-related multi-input motifs can divergently control multiple responses to the same environment. In addition, temporal regulation of the activities of a multi-functional transcription factor has the potential to add complexity to the expression profiles of target genes and temporally synchronize these multiple responses with each other.

Ancillary