A synthetic library of RNA control modules for predictable tuning of gene expression in yeast

Abstract

Advances in synthetic biology have resulted in the development of genetic tools that support the design of complex biological systems encoding desired functions. The majority of efforts have focused on the development of regulatory tools in bacteria, whereas fewer tools exist for the tuning of expression levels in eukaryotic organisms. Here, we describe a novel class of RNA-based control modules that provide predictable tuning of expression levels in the yeast Saccharomyces cerevisiae. A library of synthetic control modules that act through posttranscriptional RNase cleavage mechanisms was generated through an in vivo screen, in which structural engineering methods were applied to enhance the insulation and modularity of the resulting components. This new class of control elements can be combined with any promoter to support titration of regulatory strategies encoded in transcriptional regulators and thus more sophisticated control schemes. We applied these synthetic controllers to the systematic titration of flux through the ergosterol biosynthesis pathway, providing insight into endogenous control strategies and highlighting the utility of this control module library for manipulating and probing biological systems.

Figure 1

Genetic control elements based on Rnt1p hairpins. (A) Consensus elements of an Rnt1p hairpin. Color scheme is as follows: cleavage efficiency box (CEB), red; binding stability box (BSB), blue; initial binding and positioning box (IBPB), green. Black triangles represent location of cleavage sites. The clamp region is a synthetic sequence that acts to insulate and maintain the structure of the control element. (B) Schematic illustrating the mechanism by which Rnt1p hairpins act as gene control elements when placed in the 3′ UTR of a gene of interest (goi). Barrels represent protein molecules. (C) Sequences and structures of Rnt1p hairpin controls. (D) The transcript and protein levels associated with Rnt1p hairpins and their corresponding mutated tetraloop (CAUC) controls support that the observed gene regulatory activity is due to Rnt1p processing. Normalized protein expression levels are determined by measuring the median GFP levels from a cell population harboring the appropriate construct through flow cytometry analysis and values are reported relative to that from an identical construct lacking a hairpin module (no insert). Reported values and their error are calculated from the mean and standard deviation, respectively, from the three identically grown samples. Transcript levels are determined by measuring transcript levels of yEGFP3 and a housekeeping gene, ACT1, through qRT–PCR and normalizing the yEGFP3 levels with their corresponding ACT1 levels. Normalized transcript levels are reported relative to that from an identical construct lacking a hairpin module. Reported values and their error are calculated from the mean and standard deviation, respectively, from three identically prepared qRT–PCR reactions. Source data is available for this figure at www.nature.com/msb.

Figure 2

Design and in vivo screening of an Rnt1p cleavage library. (A) Sequence and structure of Rnt1p hairpin library containing the 12 randomized nucleotides in the CEB. (B) An in vivo, fluorescence-based screen of Rnt1p hairpin activity. The library pool is cloned through gap-repair into yeast, and clones are screened on a plate reader for sequences resulting in low fluorescence. (C) Sequences and structures of select library members highlight the diversity of the selected library sequences. The color scheme for hairpin sequences is described in Figure 1A.

Figure 3

In vivo characterization of the selected Rnt1p cleavage library. (A) The gene regulatory range of the Rnt1p library spans a broad range of protein expression levels. (B) The transcript and protein levels associated with all Rnt1p library members and their corresponding mutated tetraloop (CAUC) controls supports that the observed gene regulatory activity is due to Rnt1p processing. (C) Correlation analysis of protein and transcript levels from the Rnt1p hairpin library members supports a strong correlation between the two measures of gene regulatory activity. All normalized protein and transcript levels and their error are determined as described in Figure 1D. Source data is available for this figure at www.nature.com/msb.

Figure 4

Demonstration of functional modularity of the hairpin library in the context of a different genetic construct. (A) Correlation analysis of ymCherry protein and transcript levels from the Rnt1p hairpin library members supports a strong correlation between the two measures of gene regulatory activity. Normalized protein and transcript levels and their error are determined as described in Figure 1D, with the mean ymCherry fluorescence used for the protein level measurement. (B) Correlation analysis of ymCherry and yEGFP3 protein levels from the Rnt1p hairpin library members demonstrates a strong correlation between gene regulatory activities in different genetic contexts and preservation of library rank order. Red data point, C06. Source data is available for this figure at www.nature.com/msb.

Figure 5

In vitro characterization of the Rnt1p library supports the tuning of gene regulatory activity through modulation of cleavage rates. (A) Representative cleavage reaction assays and analyses by denaturing polyacrylamide gel electrophoresis on hairpins A01, A02 and C13. The top band corresponds to full-length RNA; the bottom band corresponds to the three cleavage products expected from Rnt1p processing. Owing to added sequences flanking the Rnt1p hairpin for insulation, the three cleavage products differ in size by 1 nt and cannot be resolved into individual bands under the assay conditions. RNA is added to the following final concentrations in each reaction (left to right; in μM): 0.2, 0.35, 0.5, 0.6–0.8. Reactions lacking Rnt1p are with 0.2 μM RNA. (B) Correlation analysis of relative cleavage rate (RCR) and normalized yEGFP3 transcript levels supports a strong correlation between cleavage rate and gene regulatory activity. Reported RCR values are determined from a Michaelis–Menten model parameter fit using Prism 5 (GraphPad) and standard error was calculated from the software. (C) Representative mobility shift assays and analyses by non-denaturing polyacrylamide gel electrophoresis on the mutated tetraloop (C13-GAAA) and C02. The top band corresponds to RNA–Rnt1p complexes; the bottom band corresponds to unbound RNA. Rnt1p is added to the following final concentrations in each reaction (left to right; in μM): 0, 0.42, 0.83, 1.25 and 1.66. (D) Correlation analysis of binding affinity (K_D) and normalized yEGFP3 transcript levels indicates a very weak correlation between binding affinity and gene regulatory activity. Reported K_D values are determined from a modified Scatchard model parameter fit using Prism 5 and standard error was calculated from the software. Source data is available for this figure at www.nature.com/msb.

Figure 6

Synthetic Rnt1p hairpins enable posttranscriptional control over endogenous ERG9 expression levels. (A) Simplified schematic of ergosterol biosynthesis from FPP showing key components for this work. Squalene is converted to ergosterol through 14 enzymatic steps. The dial highlights that ERG9 levels are tuned with the synthetic Rnt1p control modules. (B) Schematic of the construct and strategy used for introducing the synthetic Rnt1p control modules into the 3′ UTR of the endogenous ERG9 gene. The construct is designed to replace the native ERG9 3′ UTR with a synthetic 3′ UTR harboring an Rnt1p hairpin through homologous recombination between the integration cassette and chrVIII. The illustrated strategy maintains the native feedback regulation acting through transcriptional mechanisms to control ERG9 levels. (C) Correlation analysis of yEGFP3 and ERG9 transcript levels indicates that the synthetic Rnt1p hairpins maintain their gene regulatory activity in a different genetic context. Normalized ERG9 transcript levels and their error are determined as described in Figure 1D. Red data point, C06. (D) Correlation analysis of cellular growth rate and ERG9 transcript levels indicates that the titration of ERG9 levels results in two distinct phenotypic regimes—‘fast-growing’ and ‘slow-growing’. Growth rates are determined by measuring the OD₆₀₀ during a time course and fitting the data to an exponential growth curve using Prism 5 and standard error was calculated from the software. (E) Correlation analysis of relative ergosterol values (REVs) and ERG9 transcript levels indicates that ergosterol levels remain relatively consistent across varying ERG9 levels above a certain threshold value (∼40% normalized transcript levels). REVs are determined by extracting unsaponified sterols and measuring the absorbance of signature peaks associated with ergosterol in the UV spectrum. Reported REV values and their error are calculated from the mean and standard deviation from the three identical aliquots from sterol extractions, respectively. Black data point, wild-type yeast strain. (F) Correlation analysis of cellular growth rate and REV indicates that the two phenotypic measures of ERG9 levels are strongly correlated. Source data is available for this figure at www.nature.com/msb.