A mRNA's translation rate is controlled by several sequence determinants, including the presence of RNA structures within the N-terminal regions of its coding sequences. However, the physical rules that govern when such mRNA structures will inhibit translation remain unclear. Here, we introduced systematically designed RNA hairpins into the N-terminal coding region of a reporter protein with steadily increasing distances from the start codon, followed by characterization of their mRNA and expression levels in Escherichia coli. We found that the mRNAs’ translation rates were repressed, by up to 530-fold, when mRNA structures overlapped with the ribosome's footprint. In contrast, when the mRNA structure was located outside the ribosome's footprint, translation was repressed by <2-fold. By combining our measurements with biophysical modeling, we determined that the ribosomal footprint extends 13 nucleotides into the N-terminal coding region and, when a mRNA structure overlaps or partially overlaps with the ribosomal footprint, the free energy to unfold only the overlapping structure controlled the extent of translation repression. Overall, our results provide precise quantification of the rules governing translation initiation at N-terminal coding regions, improving the predictive design of post-transcriptional regulatory elements that regulate translation rate.

No abstract

No abstract

mRNA degradation is a central process that affects all gene expression levels, and yet, the determinants that control mRNA decay rates remain poorly characterized. Here, we applied a synthetic biology, learn-by-design approach to elucidate the sequence and structural determinants that control mRNA stability in bacterial operons. We designed, constructed, and characterized 82 operons in Escherichia coli, systematically varying RNase binding site characteristics, translation initiation rates, and transcriptional terminator efficiencies in the 5′ untranslated region (UTR), intergenic, and 3′ UTR regions, followed by measuring their mRNA levels using reverse transcription quantitative polymerase chain reaction (RT-qPCR) assays during exponential growth. We show that introducing long single-stranded RNA into 5′ UTRs reduced mRNA levels by up to 9.4-fold and that lowering translation rates reduced mRNA levels by up to 11.8-fold. We also found that RNase binding sites in intergenic regions had much lower effects on mRNA levels. Surprisingly, changing the transcriptional termination efficiency or introducing long single-stranded RNA into 3′ UTRs had no effect on upstream mRNA levels. From these measurements, we developed and validated biophysical models of ribosome protection and RNase activity with excellent quantitative agreement. We also formulated design rules to rationally control a mRNA’s stability, facilitating the automated design of engineered genetic systems with desired functionalities.

The homodimeric type I reaction center in heliobacteria is arguably the simplest known pigment-protein complex capable of conducting (bacterio)chlorophyll-based conversion of light into chemical energy. Despite its structural simplicity, the thermodynamics of the electron transfer cofactors on the acceptor side have not been fully investigated. In this work, we measured the midpoint potential of the terminal [4Fe-4S](2+/1+) cluster (FX) in reaction centers from Heliobacterium modesticaldum. The FX cluster was titrated chemically and monitored by (i) the decrease in the level of stable P800 photobleaching by optical spectroscopy, (ii) the loss of the light-induced g ≈ 2 radical from P800(+•) following a single-turnover flash, (iii) the increase in the low-field resonance at 140 mT attributed to the S = (3)/2 ground spin state of FX(-), and (iv) the loss of the spin-correlated P800(+) FX(-) radical pair following a single-turnover flash. These four techniques led to similar estimations of the midpoint potential for FX of -502 ± 3 mV (n = 0.99), -496 ± 2 mV (n = 0.99), -517 ± 10 mV (n = 0.65), and -501 ± 4 mV (n = 0.84), respectively, with a consensus value of -504 ± 10 mV (converging to n = 1). Under conditions in which FX is reduced, the long-lived (∼15 ms) P800(+) FX(-) state is replaced by a rapidly recombining (∼15 ns) P800(+)A0(-) state, as shown by ultrafast optical experiments. There was no evidence of the presence of a P800(+) A1(-) spin-correlated radical pair by electron paramagnetic resonance (EPR) under these conditions. The midpoint potentials of the two [4Fe-4S](2+/1+) clusters in the low-molecular mass ferredoxins were found to be -480 ± 11 mV/-524 ± 13 mV for PshBI, -453 ± 6 mV/-527 ± 6 mV for PshBII, and -452 ± 5 mV/-533 ± 8 mV for HM1_2505 as determined by EPR spectroscopy. FX is therefore suitably poised to reduce one [4Fe-4S](2+/1+) cluster in these mobile electron carriers. Using the measured midpoint potential of FX and a quasi-equilibrium model of charge recombination, the midpoint potential of A0 was estimated to be -854 mV at room temperature. The midpoint potentials of A0 and FX are therefore 150-200 mV less reducing than their respective counterparts in Photosystem I of cyanobacteria and plants. This places the redox potential of the FX cluster in heliobacteria approximately equipotential to the highest-potential iron-sulfur cluster (FA) in Photosystem I, consistent with its assignment as the terminal electron acceptor.