Abstract
Biological functions can often be realized by multiple network architectures, that require different numbers of molecular species. Here, we focus on the example of negative vs. positive gene regulation. While these distinct architectures can both implement the same input-output relation, they incur different component usage. We analyze how the choice of either form of regulation affects global transcriptional crosstalk, when genes are differentially expressed. Previous works proposed that the form of regulation correlates with gene usage: positive if the gene is frequently expressed and negative when infrequently expressed. This implies that in the absence of regulation genes are typically found in their unrequired activity state, hence regulatory intervention is often necessary. We find that such excess use of regulation usually entails high crosstalk levels. Instead, we propose that crosstalk reduction could be facilitated by the opposite parsimonious strategy, where genes are in their frequently required activity state, when not employing any form of regulation. Our study demonstrates the pervasive impact of a new type of protein production cost which is typically overlooked: that of regulatory interference caused by the presence of excess DNA-binding proteins in the cellular medium. This regulatory cost is predicted to increase with the number of molecular species and interactions involved. Hence, it only becomes apparent, when multiple transcription factors and multiple genes are simultaneously considered.
Author Summary Gene regulation can take two basic forms: positive - namely, the gene is inactive unless a regulator activates it; or negative - such that it is inactive unless a regulator suppresses its activity. A long-standing question is why some genes are positively, while others are negatively regulated. Savageau’s demand rule proposed, that the form of regulation correlates with the gene’s usage: positive if it is frequently active and negative when infrequently active. This implies, however, that in the absence of regulation genes are typically found in their least required activity state, hence regulatory intervention is often needed. In this work, we studied how the form of regulation - positive or negative - affects crosstalk levels, using a mathematical model for global crosstalk in a many-genes many regulators setup. By crosstalk, we mean that genes are occasionally found in an undesired regulatory state, because they interact with one of the many foreign regulators available. We find that crosstalk usually increases with the availability of regulators. Crosstalk can be reduced by the oppo-site parsimonious strategy, where genes are in their frequently required activity state, when not employing any form of regulation. This study demonstrates that crosstalk is a new type of global constraint on gene regulatory networks.