Limits...
Balancing noise and plasticity in eukaryotic gene expression.

Bajić D, Poyatos JF - BMC Genomics (2012)

Bottom Line: This additionally implies that genome neighboring organization -as modifier- appears only effective in highly plastic genes.In this class, we confirm bidirectional promoters (bipromoters) as a configuration capable to reduce coupling by abating noise but also reveal an important trade-off, since bipromoters also decrease plasticity.This presents ultimately a paradox between intergenic distances and modulation, with short intergenic distances both associated and disassociated to noise at different plasticity levels.

View Article: PubMed Central - HTML - PubMed

Affiliation: Logic of Genomic Systems Laboratory, Spanish National Biotechnology Centre, Consejo Superior de Investigaciones Científicas-CSIC, Madrid, Spain. jpoyatos@cnb.csic.es

ABSTRACT

Background: Coupling the control of expression stochasticity (noise) to the ability of expression change (plasticity) can alter gene function and influence adaptation. A number of factors, such as transcription re-initiation, strong chromatin regulation or genome neighboring organization, underlie this coupling. However, these factors do not necessarily combine in equivalent ways and strengths in all genes. Can we identify then alternative architectures that modulate in distinct ways the linkage of noise and plasticity?

Results: Here we first show that strong chromatin regulation, commonly viewed as a source of coupling, can lead to plasticity without noise. The nature of this regulation is relevant too, with plastic but noiseless genes being subjected to general activators whereas plastic and noisy genes experience more specific repression. Contrarily, in genes exhibiting poor transcriptional control, it is translational efficiency what separates noise from plasticity, a pattern related to transcript length. This additionally implies that genome neighboring organization -as modifier- appears only effective in highly plastic genes. In this class, we confirm bidirectional promoters (bipromoters) as a configuration capable to reduce coupling by abating noise but also reveal an important trade-off, since bipromoters also decrease plasticity. This presents ultimately a paradox between intergenic distances and modulation, with short intergenic distances both associated and disassociated to noise at different plasticity levels.

Conclusions: Balancing the coupling among different types of expression variability appears as a potential shaping force of genome regulation and organization. This is reflected in the use of different control strategies at genes with different sets of functional constraints.

Show MeSH
Distinct chromatin regulation strategies to achieve noisy or quiet plasticity. A) Each dot represents the mean effect in the expression of a set of genes in a subclass (HNHP, x coordinate; LNHP, y coordinate; normalized by effect in HP class) when a particular regulator is mutated[18]. A ratio >1 thus implies that the corresponding subclass is more strongly influenced by certain regulator than the full HP group. A strong negative correlation is found indicating that many regulators are highly specific to either HNHP or LNHP genes. This confirms that these groups are enriched by complementary functional classes (stress and growth related genes, respectively) which are generally regulated in opposite sense[12,13,20]. Dot colors denote the dominant effect of the regulator on the HP class (blue; regulator is mostly repressing expression, red; regulator is commonly activating) while sizes describe the strength of the dominant effect; e.g., LNHP genes are frequently affected by strong chromatin activators. B) We examined in detail the effects on LNHP genes (box in A). Except rsc30 (a regulator of ribosomal proteins[21]) all these mutations involved TAF1, which is part of the general transcription factor TFIID[22,23]. This essential factor regulates ∼90% of the genes in the genome, not including most of HNHP (which are regulated by SAGA) but including almost all LP genes (see main text). Nevertheless, we observed that all these mutations affected significantly more strongly LNHP than LP genes [K-S tests with FDR-corrected -log(p-value)’s shown at the right].
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3539894&req=5

Figure 2: Distinct chromatin regulation strategies to achieve noisy or quiet plasticity. A) Each dot represents the mean effect in the expression of a set of genes in a subclass (HNHP, x coordinate; LNHP, y coordinate; normalized by effect in HP class) when a particular regulator is mutated[18]. A ratio >1 thus implies that the corresponding subclass is more strongly influenced by certain regulator than the full HP group. A strong negative correlation is found indicating that many regulators are highly specific to either HNHP or LNHP genes. This confirms that these groups are enriched by complementary functional classes (stress and growth related genes, respectively) which are generally regulated in opposite sense[12,13,20]. Dot colors denote the dominant effect of the regulator on the HP class (blue; regulator is mostly repressing expression, red; regulator is commonly activating) while sizes describe the strength of the dominant effect; e.g., LNHP genes are frequently affected by strong chromatin activators. B) We examined in detail the effects on LNHP genes (box in A). Except rsc30 (a regulator of ribosomal proteins[21]) all these mutations involved TAF1, which is part of the general transcription factor TFIID[22,23]. This essential factor regulates ∼90% of the genes in the genome, not including most of HNHP (which are regulated by SAGA) but including almost all LP genes (see main text). Nevertheless, we observed that all these mutations affected significantly more strongly LNHP than LP genes [K-S tests with FDR-corrected -log(p-value)’s shown at the right].

Mentions: To further appreciate what determines the coupling (or uncoupling) of noise with plasticity, we inspected potential differences in the type of chromatin regulation. We computed the mean effect in expression of a compendium of mutations in regulators[18] (CRE score before represents a subset, see Methods) on plastic genes. This analysis highlighted a strong anti-correlation between the effect of perturbations in low-noise high-plasticity genes (LNHP, see Methods for definition of these classes) and high-noise high-plasticity genes (HNHP, ρ=−0.83,p<2.2×10−16n=170, Figure2A; this correlation is much stronger than the expected baseline correlation, Additional file1: Figure S5). This confirms mechanistically a complementary program of regulation between these two groups of genes (enriched by growth –ribosomal– and stress genes, respectively[12,20]), to be added to the previously observed distinctions in promoter nucleosome occupancy (Additional file1: Figure S6), and histone modification enrichment[13].


Balancing noise and plasticity in eukaryotic gene expression.

Bajić D, Poyatos JF - BMC Genomics (2012)

Distinct chromatin regulation strategies to achieve noisy or quiet plasticity. A) Each dot represents the mean effect in the expression of a set of genes in a subclass (HNHP, x coordinate; LNHP, y coordinate; normalized by effect in HP class) when a particular regulator is mutated[18]. A ratio >1 thus implies that the corresponding subclass is more strongly influenced by certain regulator than the full HP group. A strong negative correlation is found indicating that many regulators are highly specific to either HNHP or LNHP genes. This confirms that these groups are enriched by complementary functional classes (stress and growth related genes, respectively) which are generally regulated in opposite sense[12,13,20]. Dot colors denote the dominant effect of the regulator on the HP class (blue; regulator is mostly repressing expression, red; regulator is commonly activating) while sizes describe the strength of the dominant effect; e.g., LNHP genes are frequently affected by strong chromatin activators. B) We examined in detail the effects on LNHP genes (box in A). Except rsc30 (a regulator of ribosomal proteins[21]) all these mutations involved TAF1, which is part of the general transcription factor TFIID[22,23]. This essential factor regulates ∼90% of the genes in the genome, not including most of HNHP (which are regulated by SAGA) but including almost all LP genes (see main text). Nevertheless, we observed that all these mutations affected significantly more strongly LNHP than LP genes [K-S tests with FDR-corrected -log(p-value)’s shown at the right].
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3539894&req=5

Figure 2: Distinct chromatin regulation strategies to achieve noisy or quiet plasticity. A) Each dot represents the mean effect in the expression of a set of genes in a subclass (HNHP, x coordinate; LNHP, y coordinate; normalized by effect in HP class) when a particular regulator is mutated[18]. A ratio >1 thus implies that the corresponding subclass is more strongly influenced by certain regulator than the full HP group. A strong negative correlation is found indicating that many regulators are highly specific to either HNHP or LNHP genes. This confirms that these groups are enriched by complementary functional classes (stress and growth related genes, respectively) which are generally regulated in opposite sense[12,13,20]. Dot colors denote the dominant effect of the regulator on the HP class (blue; regulator is mostly repressing expression, red; regulator is commonly activating) while sizes describe the strength of the dominant effect; e.g., LNHP genes are frequently affected by strong chromatin activators. B) We examined in detail the effects on LNHP genes (box in A). Except rsc30 (a regulator of ribosomal proteins[21]) all these mutations involved TAF1, which is part of the general transcription factor TFIID[22,23]. This essential factor regulates ∼90% of the genes in the genome, not including most of HNHP (which are regulated by SAGA) but including almost all LP genes (see main text). Nevertheless, we observed that all these mutations affected significantly more strongly LNHP than LP genes [K-S tests with FDR-corrected -log(p-value)’s shown at the right].
Mentions: To further appreciate what determines the coupling (or uncoupling) of noise with plasticity, we inspected potential differences in the type of chromatin regulation. We computed the mean effect in expression of a compendium of mutations in regulators[18] (CRE score before represents a subset, see Methods) on plastic genes. This analysis highlighted a strong anti-correlation between the effect of perturbations in low-noise high-plasticity genes (LNHP, see Methods for definition of these classes) and high-noise high-plasticity genes (HNHP, ρ=−0.83,p<2.2×10−16n=170, Figure2A; this correlation is much stronger than the expected baseline correlation, Additional file1: Figure S5). This confirms mechanistically a complementary program of regulation between these two groups of genes (enriched by growth –ribosomal– and stress genes, respectively[12,20]), to be added to the previously observed distinctions in promoter nucleosome occupancy (Additional file1: Figure S6), and histone modification enrichment[13].

Bottom Line: This additionally implies that genome neighboring organization -as modifier- appears only effective in highly plastic genes.In this class, we confirm bidirectional promoters (bipromoters) as a configuration capable to reduce coupling by abating noise but also reveal an important trade-off, since bipromoters also decrease plasticity.This presents ultimately a paradox between intergenic distances and modulation, with short intergenic distances both associated and disassociated to noise at different plasticity levels.

View Article: PubMed Central - HTML - PubMed

Affiliation: Logic of Genomic Systems Laboratory, Spanish National Biotechnology Centre, Consejo Superior de Investigaciones Científicas-CSIC, Madrid, Spain. jpoyatos@cnb.csic.es

ABSTRACT

Background: Coupling the control of expression stochasticity (noise) to the ability of expression change (plasticity) can alter gene function and influence adaptation. A number of factors, such as transcription re-initiation, strong chromatin regulation or genome neighboring organization, underlie this coupling. However, these factors do not necessarily combine in equivalent ways and strengths in all genes. Can we identify then alternative architectures that modulate in distinct ways the linkage of noise and plasticity?

Results: Here we first show that strong chromatin regulation, commonly viewed as a source of coupling, can lead to plasticity without noise. The nature of this regulation is relevant too, with plastic but noiseless genes being subjected to general activators whereas plastic and noisy genes experience more specific repression. Contrarily, in genes exhibiting poor transcriptional control, it is translational efficiency what separates noise from plasticity, a pattern related to transcript length. This additionally implies that genome neighboring organization -as modifier- appears only effective in highly plastic genes. In this class, we confirm bidirectional promoters (bipromoters) as a configuration capable to reduce coupling by abating noise but also reveal an important trade-off, since bipromoters also decrease plasticity. This presents ultimately a paradox between intergenic distances and modulation, with short intergenic distances both associated and disassociated to noise at different plasticity levels.

Conclusions: Balancing the coupling among different types of expression variability appears as a potential shaping force of genome regulation and organization. This is reflected in the use of different control strategies at genes with different sets of functional constraints.

Show MeSH