In organisms of all three domains of life, a plethora of sRNAs (small regulatory RNAs) exists in addition to the well-known RNAs such as rRNAs, tRNAs and mRNAs. Although sRNAs have been well studied in eukaryotes and in bacteria, the sRNA population in archaea has just recently been identified and only in a few archaeal species. In the present paper, we summarize our current knowledge about sRNAs and their function in the halophilic archaeon Haloferax volcanii. Using two different experimental approaches, 111 intergenic and 38 antisense sRNAs were identified, as well as 42 tRFs (tRNA-derived fragments). Observation of differential expression under various conditions suggests that these sRNAs might be active as regulators in gene expression like their bacterial and eukaryotic counterparts. The severe phenotypes observed upon deletion and overexpression of sRNA genes revealed that sRNAs are involved in, and important for, a variety of biological functions in H. volcanii and possibly other archaea. Investigation of the Haloferax Lsm protein suggests that this protein is involved in the archaeal sRNA pathway.
sRNAs (small regulatory RNAs)
Next to the well known RNA classes, tRNA, rRNA and mRNA, a new type of RNA has recently been identified: sRNAs [1,2]. A plethora of these sRNAs has been found in Bacteria and Eukarya as mediators of a number of cellular processes and, thanks to high-throughput techniques such as next-generation sequencing, more and more are being detected. In eukaryotes, several types of sRNAs have been identified, e.g. miRNAs (microRNAs), which are approx. 20–25 nt long and which are involved in the regulation of translation and mRNA degradation [3,4]. It has been estimated that more than 30% of eukaryotic genes are regulated by sRNAs and that more than thousands of sRNAs exist in eukaryotic cells. Furthermore, it has been shown that miRNAs have essential roles in development and differentiation. The major mode of interaction seems to be a more or less perfect binding to the mRNA 3′-end, thereby preventing translation or initiating mRNA degradation (Figure 1). sRNAs in bacteria cover a broader size range, from approx. 50 to approx. 400 nt. They have been shown to interact with the 5′-end of the mRNA, preventing ribosome binding to the mRNA, covering the initiation codon or the beginning of the coding sequence, altogether preventing translation and also in some cases initiating mRNA degradation (Figure 1) [5,6]. cis-Encoded sRNAs are fully complementary to their target RNA, whereas trans-encoded sRNAs are only partially complementary. Bacterial sRNAs specific for a single target have been found, as well as sRNAs acting on multiple targets [2,7,8].
sRNAs in three domains of life
Knowledge about this new class of RNAs in the third domain of life, the Archaea, lagged somewhat behind. The first sRNAs to be identified in archaea were small guide RNAs [named after their eukaryotic counterpart: snoRNAs (small nucleolar RNAs)], which are involved in modification of RNA nucleotides [9,10]. Subsequently, RNomics revealed the presence of additional sRNAs in two archaeal organisms: Sulfolobus solfataricus and Archaeoglobus fulgidus [11,12]. Since that time, sRNAs have been identified in several archaeal species [13–21]. In addition to the already identified snoRNAs, cis-encoded antisense RNAs and trans-encoded sRNAs were identified (for a review, see ). However, although an sRNA population was identified in archaea, it is not known yet whether these molecules are employed in regulation of gene expression and, if so, how they act. We chose the genetically tractable model species Haloferax volcanii for the characterization of the sRNA repertoire and the biological roles of sRNAs in archaea.
sRNAs in H. volcanii
H. volcanii belongs to the lineage of Euryarchaeota, grows optimally at approx. 2.1 M NaCl and, to cope with this salt concentration in the environment, raises the intracellular KCl concentration to similar values . We identified the sRNA population of H. volcanii using an RNomics approach as well as an HTS (high-throughput sequencing) approach (, and A. Jellen-Ritter, M. Dörr, J. Babski, J. Soppa and A. Marchfelder, unpublished work). In the RNomics approach, a cDNA library was generated from RNAs ranging in size from 130 to 460 nt, resulting in the identification of 21 intergenic sRNA genes, 18 antisense sRNA genes and 49 sRNA genes overlapping with ORFs (open reading frames) . The second approach analysed RNA ranging in size from 17 to 500 nt from six different RNA preparations, which were isolated from cells grown under three different conditions: normal (45°C, 18% salt), high-temperature (48°C, 18% salt) and low-salt (45°C, 11%) conditions, each at differential and stationary phase. Altogether, 111 intergenic RNAs, 38 cis-antisense RNAs and more than 200 sense RNAs were identified (A. Jellen-Ritter, M. Dörr, J. Babski, J. Soppa and A. Marchfelder, unpublished work). All sRNAs found with the RNomics approach were also present in the identified HTS pool, thus, together, both approaches identified 111 intergenic, 38 antisense and more than 200 sense sRNAs.
Recently, an additional type of sRNAs has been detected in eukaryotes: tRFs (tRNA-derived fragments) [24,25]. In human cells these fragments are second most abundant to miRNAs and are processed from mature tRNAs or precursor tRNAs. tRNA 5′-fragments, tRNA 3′-fragments and tRNA 3′-trailer sequences have been detected as stable molecules. Processing of these molecules seems to involve Dicer  as well as the tRNA 3′-processing endonuclease tRNase Z [24,25,27]. Recent studies suggest that these tRNA fragments are not random by-products of tRNA biogenesis and degradation, but are an abundant and novel class of short RNAs. They can be generated by different means and show specific expression patterns. Our HTS approach revealed the presence of 42 tRFs in H. volcanii, suggesting that these types of sRNAs are also active in archaea.
The expression of more than 50 sRNA genes under various conditions was characterized using Northern blots and DNA microarray analyses ([19,28], and A. Jellen-Ritter, M. Dörr, J. Babski, J. Soppa and A. Marchfelder, unpublished work). The expression of most sRNAs was confirmed, and several sRNA genes were shown to be differentially expressed e.g. in response to growth phase, temperature or salt concentration (, A. Jellen-Ritter, M. Dörr, J. Babski, J. Soppa and A. Marchfelder, unpublished work). In addition, to investigate the biological function of the sRNAs, deletion mutants were generated for the intergenic sRNA genes and overexpression constructs were made containing the genes for antisense sRNAs and tRFs. With two exceptions, all deletion mutants were viable, indicating that only a minority of sRNAs are essential under all conditions. However, nine sRNA-deletion mutants exhibited severe growth phenotypes under specific conditions (, and A. Jellen-Ritter, M. Dörr, J. Babski, J. Soppa and A. Marchfelder, unpublished work), e.g. one sRNA-deletion strain is unable to grow at high temperatures, whereas another one can barely grow in low salt concentrations, again others have growth defects on different carbon sources. Overexpression of the genes for antisense sRNAs and tRFs likewise showed various phenotypes under different conditions. These results reveal that archaeal sRNAs are involved in a variety of biological functions, ranging from stress response over metabolic regulation to behaviour.
Involvement of the archaeal Lsm protein in the sRNA pathway
In bacteria, several sRNAs require the protein Hfq to mediate the interaction between sRNA and target RNA [29,30]. The bacterial Hfq protein belongs to the Sm/Lsm protein family, which has representatives in all three domains of life . Eukaryotes have several Sm and Lsm proteins involved, e.g., in mRNA splicing, histone maturation, telomere maintenance and mRNA degradation. In archaea, three different types of Lsm proteins have been identified, termed Lsm1, Lsm2 and Lsm3, and usually an archaeal organism encodes one or two Lsm proteins [32–34]. In addition, an Hfq protein has been identified in Methanocaldococcus jannaschii [35,36]. Although the structure and RNA-binding activities of the archaeal Lsm proteins have been investigated [33,37–41], the biological function of these proteins is not known to date. To unravel the potential involvement of the archaeal Lsm protein in the sRNA pathway and to identify the biological functions of this protein, we characterized the H. volcanii Lsm protein. Haloferax encodes only a single Lsm protein, an Lsm1 subtype, allowing uncomplicated genetic analysis of Lsm function .
Eukaryotes contain several different Lsm proteins which have been shown to form heteroheptameric complexes ; in contrast, the bacterial Hfq protein complex consists of six Hfq protein monomers . To characterize the nature of the archaeal Lsm protein complex, we first expressed the lsm gene in Escherichia coli, yielding a pure recombinant protein fraction . Investigation of the complex formation of the halophilic Lsm revealed that the protein forms homoheptameric complexes, similar to its bacterial counterpart .
Upon deletion of the hfq gene, E. coli shows pleiotropic physiological effects, which is consistent with the observation that the bacterial Hfq protein is involved in many cellular processes [31,45]. To unravel the biological function of the archaeal Lsm protein, a Haloferax lsm deletion strain (Δlsm) was generated. The Lsm protein is encoded together with the ribosomal protein L37e on a dicistronic operon; both frames even overlap by four nucleotides. Therefore care was taken to delete only the Lsm frame, leaving the L37e frame intact. The deletion mutant is viable, but shows growth defects when compared with the wild-type strain under various conditions, suggesting that the archaeal Lsm protein is also involved in many cellular processes .
Co-purification of interaction partners
Bacterial, as well as eukaryotic, Lsm and Hfq proteins have been shown to interact with a plethora of protein and RNA-interaction partners [46,47]. To identify the interaction partners of the archaeal Lsm protein, we expressed a Lsm–FLAG fusion protein in Haloferax cells. Subsequently, the fusion protein was purified together with all molecules attached. MS of the co-purified proteins revealed more than 30 interacting proteins belonging to different functional classes . Similar observations had been made with the bacterial Hfq protein and the eukaryotic Lsm proteins. Both also bind to several other proteins, confirming their role in a plethora of cellular pathways.
Co-purified RNAs were reverse-transcribed into cDNAs which were subsequently used as probe for microarrays. These microarray experiments revealed that several sRNAs were co-purified with the Lsm protein, suggesting that the archaeal Lsm protein, like its bacterial counterpart, might also be involved in the sRNA pathway .
Recent results clearly show the presence of an sRNA population in archaea, and, although the function of these sRNAs could yet not be clarified, data from sRNA-gene-deletion mutants and overexpression of sRNA genes, as well as the interaction with the Lsm protein, suggest that these sRNAs might be active as regulators in gene expression like their bacterial and eukaryotic counterparts. Future research will focus on the identification of the sRNA targets and on unravelling the molecular mechanisms underlying the archaeal sRNA pathway.
Molecular Biology of Archaea II: A Biochemical Society Focused Meeting held at Robinson College, Cambridge, U.K., 16–18 August 2010. Organized and Edited by Stephen Bell (Oxford, U.K.) and Finn Werner (University College London, U.K.).
This work was supported by the Deutsche Forschungsgemeinschaft in the frame of the priority programme ‘Sensory and Regulatory RNAs in Prokaryotes’ [grant number SPP1258].