MicroRNA and protein sequestration by non-coding RNAs (ncRNAs) has recently generated much interest. In the bacterial Csr/Rsm system, which is considered to be the most general global post-transcriptional regulatory system responsible for bacterial virulence, ncRNAs such as CsrB or RsmZ activate translation initiation by sequestering homodimeric CsrA-type proteins from the ribosome-binding site of a subset of messenger RNAs. However, the mechanism of ncRNA-mediated protein sequestration is not understood at the molecular level. Here we show for Pseudomonas fluorescens that RsmE protein dimers assemble sequentially, specifically and cooperatively onto the ncRNA RsmZ within a narrow affinity range. This assembly yields two different native ribonucleoprotein structures. Using a powerful combination of nuclear magnetic resonance and electron paramagnetic resonance spectroscopy we elucidate these 70-kilodalton solution structures, thereby revealing the molecular mechanism of the sequestration process and how RsmE binding protects the ncRNA from RNase E degradation. Overall, our findings suggest that RsmZ is well-tuned to sequester, store and release RsmE and therefore can be viewed as an ideal protein ‘sponge’.

The carbon storage regulator/regulator of secondary metabolism (Csr/Rsm) type of small non-coding RNAs (sRNAs) is widespread throughout bacteria and acts by sequestering the global translation repressor protein CsrA/RsmE from the ribosome binding site of a subset of mRNAs. Although we have previously described the molecular basis of a high affinity RNA target bound to RsmE, it remains unknown how other lower affinity targets are recognized by the same protein. Here, we have determined the nuclear magnetic resonance solution structures of five separate GGA binding motifs of the sRNA RsmZ of Pseudomonas fluorescens in complex with RsmE. The structures explain how the variation of sequence and structural context of the GGA binding motifs modulate the binding affinity for RsmE by five orders of magnitude (∼10 nM to ∼3 mM, K_d). Furthermore, we see that conformational adaptation of protein side-chains and RNA enable recognition of different RNA sequences by the same protein contributing to binding affinity without conferring specificity. Overall, our findings illustrate how the variability in the Csr/Rsm protein–RNA recognition allows a fine-tuning of the competition between mRNAs and sRNAs for the CsrA/RsmE protein.

High-resolution structural information on RNA and its functionally important complexes with proteins is dramatically underrepresented compared with proteins but is urgently needed for understanding cellular processes at the molecular and atomic level. Here we present an EPR-based protocol to help solving large RNA and protein–RNA complex structures in solution by providing long-range distance constraints between rigid fragments. Using enzymatic ligation of smaller RNA fragments, large doubly spin-labelled RNAs can be obtained permitting the acquisition of long distance distributions (>80 Å) within a large protein–RNA complex. Using a simple and fast calculation in torsion angle space of the spin-label distributions with the program CYANA, we can derive simple distance constraints between the spin labels and use them together with short-range distance restraints derived from NMR to determine the structure of a 70 kDa protein–RNA complex composed of three subcomplexes.