Volume 386, Issue 12

microRNAs (miRNAs) form an evolutionarily conserved and highly abundant class of non-coding RNAs that are 21–24 nucleotides (nt) in length. They are processed from double-stranded (ds) RNA precursors and sequence-specifically guide posttranscriptional gene silencing. The processing steps are facilitated by members of the RNAse III enzyme family, whereas gene silencing events are mediated by members of the highly conserved Argonaute (Ago) protein family. Initially discovered in Caenorhabditis elegans , in which they are essential for proper developmental timing, hundreds of miRNAs have been discovered to date in a variety of different organisms, including plants, flies and mammals. Expression profiling approaches demonstrated that miRNAs are specifically expressed not only during embryonic development, but also during cell differentiation and other cellular events such as hormonal signaling. Although miRNAs have been the object of extensive research in recent years, very little is known about their target mRNAs. Their identification along with a comprehensive description of the miRNA/target-mRNA interaction network will add a new level to our knowledge of gene regulation and will also provide new insights into the biology of so far poorly understood diseases, including various forms of cancer.

Small non-coding RNAs (sRNAs) have attracted considerable attention as an emerging class of gene expression regulators. In bacteria, a few regulatory RNA molecules have long been known, but the extent of their role in the cell was not fully appreciated until the recent discovery of hundreds of potential sRNA genes in the bacterium Escherichia coli . Orthologs of these E. coli sRNA genes, as well as unrelated sRNAs, were also found in other bacteria. Here we review the disparate experimental approaches used over the years to identify sRNA molecules and their genes in prokaryotes. These include genome-wide searches based on the biocomputational prediction of non-coding RNA genes, global detection of non-coding transcripts using microarrays, and shotgun cloning of small RNAs (RNomics). Other sRNAs were found by either co-purification with RNA-binding proteins, such as Hfq or CsrA/RsmA, or classical cloning of abundant small RNAs after size fractionation in polyacrylamide gels. In addition, bacterial genetics offers powerful tools that aid in the search for sRNAs that may play a critical role in the regulatory circuit of interest, for example, the response to stress or the adaptation to a change in nutrient availability. Many of the techniques discussed here have also been successfully applied to the discovery of eukaryotic and archaeal sRNAs.

In the cell, the protein synthetic machinery is a highly complex apparatus that offers many potential sites for functional interference and therefore represents a major target for antibiotics. The recent plethora of crystal structures of ribosomal subunits in complex with various antibiotics has provided unparalleled insight into their mode of interaction and inhibition. However, differences in the conformation, orientation and position of some of these drugs bound to ribosomal subunits of Deinococcus radiodurans (D50S) compared to Haloarcula marismortui (H50S) have raised questions regarding the species specificity of binding. Revisiting the structural data for the bacterial D50S-antibiotic complexes reveals that the mode of binding of the macrolides, ketolides, streptogramins and lincosamides is generally similar to that observed in the archaeal H50S structures. However, small discrepancies are observed, predominantly resulting from species-specific differences in the ribosomal proteins and rRNA constituting the drug-binding sites. Understanding how these small alterations at the binding site influence interaction with the drug will be essential for rational design of more potent inhibitors.

tRNase Z is the endoribonuclease that generates the mature 3′-end of tRNA molecules by removal of the 3′-trailer elements of precursor tRNAs. This enzyme has been characterized from representatives of all three domains of life (Bacteria, Archaea and Eukarya), as well as from mitochondria and chloroplasts. tRNase Z enzymes come in two forms: short versions (280–360 amino acids in length), present in all three kingdoms, and long versions (750–930 amino acids), present only in eukaryotes. The recently solved crystal structure of the bacterial tRNase Z provides the structural basis for the understanding of central functional elements. The substrate is recognized by an exosite that protrudes from the main protein body and consists of a metallo-β-lactamase domain. Cleavage of the precursor tRNA occurs at the binuclear zinc site located in the other subunit of the functional homodimer. The first gene of the tRNase Z family was cloned in 2002. Since then a comprehensive set of data has been acquired concerning this new enzyme, including detailed functional studies on purified recombinant enzymes, mutagenesis studies and finally the determination of the crystal structure of three bacterial enzymes. This review summarizes the current knowledge about these exciting enzymes.

Alternative splicing is a process by which multiple messenger RNAs (mRNAs) are generated from a single pre-mRNA, resulting in functionally distinct protein products. This is accomplished by the differential recognition of splice sites in the pre-mRNA, often regulated in a tissue- or development-specific manner. Alternative splicing constitutes not only an important mechanism in controlling gene expression in humans, but also an essential source for increasing proteome diversity. In this review we summarize the underlying mechanistic principles, focussing on the cis -acting regulatory elements. In particular, the role of short sequence repeats, which are often polymorphic, in splicing regulation is discussed.

The bacterial riboregulator 6S RNA was one of the first non-coding RNAs to be discovered in the late 1960s, but its cellular role remained enigmatic until the year 2000. 6S RNA, only recognized to be ubiquitous among bacteria in 2005, binds to RNA polymerase in a σ factor-dependent manner to repress transcription from a subgroup of promoters. The common feature of a double-stranded rod with a central bulge has led to the proposal that 6S RNA may mimic an open promoter complex.

Expression of many rhizobial small heat-shock genes is controlled by the ROSE element, a thermoresponsive structure in the 5′-untranslated region of the corresponding mRNAs. Using a bioinformatics approach, we found more than 20 new potential ROSE-like RNA thermometers upstream of small heat-shock genes in a wide variety of α- and γ-proteobacteria. Northern blot analyses revealed heat-inducible transcripts of the representative candidate Caulobacter crescentus CC2258 , Escherichia coli ibpA and Salmonella typhimurium ibpA genes. Typical σ 32 -type promoters were mapped upstream of the potential RNA thermometers by primer extension. Additional translational control was demonstrated in a lacZ reporter system and by site-directed mutagenesis. RNA secondary structure predictions strongly suggest that the Shine-Dalgarno sequence in the RNA thermometers is masked at low temperatures. Combining two regulatory modules, a σ 32 promoter and a ROSE-type RNA thermometer, provides a novel stringent mechanism to control expression of small heat-shock genes.

Post-transcriptional mechanisms contribute to the changes in gene expression induced by cell stress. The effect of UV-B light on mRNA degradation in HeLa cells was investigated using a transcriptional chase system to determine the decay kinetics of tet-off vector-derived mRNAs containing or lacking a destabilizing AU-rich element. Degradation of both mRNAs was strongly inhibited in cells exposed to UV-B light. Removal of the poly(A)-tail, considered a crucial step in mRNA degradation, was strikingly impaired. UV light also inhibited deadenylation and degradation of endogenous mRNA of the chemoattractant cytokine interleukin (IL)-8. Both effects occurred rapidly and independently of newly induced genes. Importantly, stabilization of IL-8 mRNA was accompanied by a strong increase in the duration of IL-8 protein formation. Furthermore, general inhibition of protein synthesis, a hallmark of the response to cell stress, required far higher doses of UV-B than inhibition of mRNA deadenylation and degradation. The difference in sensitivity of cells to these effects of UV-B light establishes a dose range in which mRNA stabilization can lead to dramatically enhanced expression of proteins derived from normally unstable mRNAs, such as those of inflammatory cytokines, growth factors and proto-oncogenes, and thereby have a major impact on the response to UV light.

Recent studies on the chemical alkaline degradation of ferredoxins have contributed to the hypothesis that linear three-iron centres are commonly observed as degradation intermediates of iron-sulfur clusters. In this work we assess the validity of this hypothesis. We studied different proteins containing iron-sulfur clusters, iron-sulfur centres and di-iron centres with respect to their chemical degradation kinetics at high pH, in the presence and absence of exogenous sulfide, to investigate the possible formation of linear three-iron centres during protein unfolding. Our spectroscopic and kinetic data show that in these different proteins visible absorption bands at 530 and 620 nm are formed that are identical to those suggested to arise from linear three-iron centres. Iron release and protein unfolding kinetics show that these bands result from the formation of iron sulfides at pH 10, produced by the degradation of the iron centres, and not from rearrangements leading to linear three-iron centres. Thus, at this point any relevant functional role of linear three-iron centres as cluster degradation intermediates in iron-sulfur proteins remains elusive.

MCoTI-I and MCoTI-II from the seeds of Momordica cochinchinensis are inhibitors of trypsin-like proteases and the only known members of the large family of squash inhibitors that are cyclic and contain an additional loop connecting the amino- and the carboxy-terminus. To investigate the contribution of macrocycle formation to biological activity, we synthesized a set of open-chain variants of MCoTI-II that lack the cyclization loop and contain various natural and non-natural amino acid substitutions in the reactive-site loop. Upon replacement of P1 lysine residue #10 within the open-chain variant of MCoTI-II by the non-natural isosteric nucleo amino acid AlaG [β-(guanin-9-yl)-L-alanine], a conformationally restricted arginine mimetic, residual inhibitory activity was detected, albeit reduced by four orders of magnitude. While the cyclic inhibitors MCoTI-I and MCoTI-II were found to be very potent trypsin inhibitors, with picomolar inhibition constants, the open-chain variants displayed an approximately 10-fold lower affinity. These data suggest that the formation of a circular backbone in the MCoTI squash inhibitors results in enhanced affinity and therefore is a determinant of biological activity.

The protease Pim1/LON, a member of the AAA+ family of homo-oligomeric ATP-dependent proteases, is responsible for the degradation of soluble proteins in the mitochondrial matrix. To establish the molecular parameters required for the specific recognition and proteolysis of substrate proteins by Pim1, we analyzed the in organello degradation of imported reporter proteins containing different structural properties. The amino acid composition at the amino-terminal end had no major effect on the proteolysis reaction. However, proteins with an amino-terminal extension of less than 60 amino acids in front of a stably folded reporter domain were completely resistant to proteolysis by Pim1. Substrate proteins with a longer amino-terminal extension showed incomplete proteolysis, resulting in the generation of a defined degradation fragment. We conclude that Pim1-mediated protein degradation is processive and is initiated from an unstructured amino-terminal segment. Resistance to degradation and fragment formation was abolished if the folding state of the reporter domain was destabilized, indicating that Pim1 is not able to unravel folded proteins for proteolysis. We propose that the requirement for an exposed, large, non-native protein segment, in combination with a limited unfolding capability, accounts for the selectivity of the protease Pim1 for damaged or misfolded polypeptides.

Serpins appear to be ubiquitous in eukaryotes, except fungi, and are also present in some bacteria, archaea and viruses. Inhibitory serpins with a glutamine as the reactive-center P1 residue have been identified exclusively in a few plant species. Unique serpins with a reactive center sequence of three Gln residues at P3-P1 or P2-P1′ were isolated from barley and wheat grain, respectively. Barley BSZ3 was an irreversible inhibitor of chymotrypsin, with a second-order association rate constant for complex formation k a ′ of the order of 10 4  M -1  s -1 ; however, only a minor fraction of the serpin molecules reacted with chymotrypsin, with the majority insensitive to cleavage in the reactive center loop. Wheat WSZ3 was cleaved specifically at P8 Thr and was not an inhibitor of chymotrypsin. These reactive-center loops may have evolved conformations that are optimal as inhibitory baits for proeinases that specifically degrade storage prolamins containing Gln-rich repetitive sequences, most likely for digestive proteinases of insect pests or fungal pathogens that infect cereals. An assembled full-length amino acid sequence of a serpin expressed in cotton boll fiber (GaZ1) included conserved regions essential for serpin-proteinase interaction, suggesting inhibitory capacity at a putative reactive center P2-P2′ with a sequence of four Gln residues.