The list of proteins with possible RNA chaperone activity is growing constantly. Proteins with different activities that support RNA folding are classified in this group. The definition of a protein with RNA chaperone activity is that the protein prevents RNA from misfolding by opening up misfolded structures. Proteins with RNA chaperone activity do not require ATP, which distinguishes them from RNA helicases, another group of proteins that facilitate RNA folding (e.g., Cyt-19) [2].

Proteins with RNA chaperone activity interact only transiently with RNA molecules and are supposed to be dispensable once the RNA has been folded correctly. This was shown for E. coli proteins S12 and StpA [3, 4]. A transient interaction and weak binding to RNA might be difficult to define because many of the identified RNA chaperones interact strongly with their target RNAs and are found in RNP complexes like ribosomal proteins, hnRNPs, La protein, and others. However, it has been demonstrated that a mutant StpA that shows stronger binding towards RNA shows decreased RNA chaperone activity suggesting that strong binding could also be detrimental to RNA folding [5]. In that way, proteins with RNA chaperone activity are also distinguished from “stabilizers” that are proteins that bind and stabilize an RNA structure and are required to stay bound in order to keep the RNA's native structure. Cyt-18, the tRNA synthetase from Neurospora crassa, is a “stabilizer” for the mitochondrial self-splicing group I intron: its presence is required to keep the native structure of the intron which otherwise unfolds readily.

In the growing database of “proteins with RNA chaperone activity”, there exists an increasing number of proteins that simply possess RNA annealing activity. A prominent and intensively studied member of this group is the bacterial host factor Hfq that showed annealing activity on random substrates. Hfq in addition is an RNA chaperone as it was further demonstrated that Hfq does possess unwinding activity upon its native substrates [6, 7].

In brief, the group of proteins with RNA chaperone activity includes proteins that, first, open up misfolded structures without requirement of ATP and that, second, are dispensable once the RNA has been folded.

RNA molecules are prone to misfold in vitro and are usually prevented from misfolding in vivo. RNA basically encounters two folding problems: a kinetic folding problem, where the RNA molecule has to surmount kinetic barriers during the search for its native structure. Secondly, RNA molecules meet a thermodynamic folding problem as the final native structure often has to compete with alternative folds that have similar energetic stabilities [1]. RNA folding is a hierarchical process, and first secondary structure elements have to form. Secondary structural elements form between regions within the RNA molecule that are in close proximity. They are A-form helices consisting of Watson-Crick base-pairs. Secondary structures are very stable. The stability of a base-pair depends on the stability of both of its neighbouring base-pairs. Already any RNA of a reasonable length is able to form alternative base-pairs leading to alternative helices that become folding traps.

Tertiary structures are higher order structures that are built by assembling the secondary structure elements into a more complex collapsed fold. They can also involve formation of helices. This is the case in pseudoknots where either a loop region interacts with a distant single stranded region or with another distant loop. Pseudoknots possess similar stability as secondary structures. But tertiary structural elements involve also other non-Watson-Crick interactions where for example, not only the Watson-Crick site of the nucleotide interacts with another nucleotide but also the Hoogsteen edge or the sugar edge of the nucleotide is involved in hydrogen bonding [8]. An often reoccurring tertiary structure motif is the A-minor interaction where an adenine interacts with the minor groove of the A-form helix [9]. Tertiary structures are often less stable and depend on the formation of secondary structures. Finally, monovalent or divalent metal ions play an important role in tertiary structure formation.

The first studies on RNA structure and folding were done in the 1960s with yeast tRNA molecules. Already then it was demonstrated that tRNAs are able to adopt two distinct conformations of which only one is the native structure which can be aminoacylated [10, 11].

The RNA folding problem becomes even more prominent in the case of large RNAs such as group I introns or in the context of large protein-ribonucleic acid complexes such as RNase P and the ribosome.

It was demonstrated that the self-splicing group I intron of the thymidylate synthase gene of phage T4 misfolds in the absence of translation: when the ribosome does not prevent base-pairing between exon and intron sequences, the intron is not able to fold correctly and cannot perform the splicing reaction [12]. A similar observation was made with the group I intron of Tetrahymena thermophila ribosomal RNA: a subset of molecules misfolds and accumulates into an inactive population [13]. Misfolding depends on exon sequences that form stable hairpins and intervene with 5′-splice-site formation.

In vivo, however, some group I introns require the assistance of proteins to splice efficiently and prevent misfolding. For an example, the Cyt-18 protein in Neurospora crassa mitochondira is a tRNA synthetase which stabilizes the P4-P6 domain of group I introns and recruits Cyt-19, an RNA helicase, which then unwinds folding traps and promotes splicing [2, 14].

For large RNP complexes such as the ribosome, a growing body of evidence suggests that several additional factors such as helicases exist that assist during the folding process in vivo.

The first viral RNA chaperone activities were reported in the early 1990s and showed that nucleocapsid protein 7 (Ncp7) of HIV increases hammerhead ribozyme cleavage significantly [20, 28]. Only a few years later, another virus-encoded protein with RNA chaperone activity has been identified, the HDV delta antigen, which was also monitored in the hammerhead ribozyme assay and furthermore shown to be dispensable after folding has occurred [29].

Flaviviridae core proteins were also monitored and shown to possess RNA chaperone activity in a hammerhead cleavage assay and/or RNA strand annealing activities [30, 31]. Interestingly, many of the viral proteins show exceptional high degree of disorder. In the case of the Flaviviridae core proteins, it was also reported that heat denaturation still retained strand annealing activity suggesting that the disordered domains of the proteins are involved in chaperoning.

Nucleocapsid proteins from two members of the Coronaviridae family have been investigated and hammerhead ribozyme cleavage was shown to be enhanced in their presence [32, 33]. And again both nucleocapsid proteins show a high degree of disorder in in silico predictions.

A growing body of evidence suggests that there exist many more viral proteins that possess RNA chaperone activity. The list here is not complete but proteins that were shown to have only DNA annealing activities were left out in this list. The majority of these small viral proteins show strong propensity for disorder which suggests that disorder might be a mechanistic requirement for chaperoning.

Interestingly, studies on Nc proteins demonstrated that these proteins not only possess RNA chaperone activity in vitro but also are required for strand annealing and strand displacement activities on their target RNAs in vivo [33]. Recently, a specific template switching assay designed to study strand displacement in a retroviral-derived system demonstrated that nucleocapsid protein from Coronavirus shows RNA chaperone activity and most likely is an RNA chaperone in vivo [34].

The E. coli transcriptional regulator StpA, a 15 kD basic protein, was isolated as a repressor of a splicing-deficient group I intron in thymidylate synthase of phage T4 [4]. StpA was furthermore shown to possess strong RNA chaperone activity in vivo in the folding trap assay [35]. The protein was tested in vitro in a strand-annealing and strand-displacement assay and exhibited strong activities in both tests [36]. More detailed studies on StpA revealed that the protein binds transiently to RNA with a preference for unstructured regions and that binding to RNA is diminished in the presence of high ionic strength [5]. In an elaborate study applying in vivo DMS modifications to the RNA, in vivo folding of the group I intron was evaluated in the absence and presence of StpA [37]: Schroeder and coworkers demonstrated that StpA opens up tertiary interactions of the td group I intron. While the loosening effect is advantageous in wild type or misfolded introns, overexpression of StpA in the presence of introns that were already destabilized in their 3D structure was detrimental. Structure prediction of StpA suggested that this protein exhibits more than 70% disorder and it was suggested that this unfolded regions of StpA might play a role in chaperoning [38].

Ribosomal proteins are required within every cellular organism to build up the bacterial 70S or the eukaryal 80S ribosome. Many ribosomal proteins further regulate transcription or translation of their own operons. In addition, ribosomal proteins are also involved in various very different cellular processes and fulfill extraribosomal functions [39, 40]. Ribosomal proteins are highly conserved among various species and many ribosomal proteins have unusual long unstructured extensions that wind their way through the ribosome [41].

The first observation that ribosomal proteins are capable of chaperoning RNA folding came from the Belfort lab: screening for cellular factors that increase trans-splicing of the thymidylate synthase group I intron revealed that many ribosomal proteins possess chaperoning activity, with ribosomal protein S12 from the small ribosomal subunit having the strongest activity [3]. Furthermore, S12 significantly increased hammerhead ribozyme cleavage [3]. A systematic study on large ribosomal subunit proteins from E. coli showed that 1/3 of the tested proteins possesses strong RNA chaperone activity in vitro in the trans-splicing assay [21]. In addition, ribosomal protein L1 orthologs from eukarya, bacteria, and mesophilic archaea also exhibited strong trans-splicing and cis-splicing activities in vitro [42]. Although it makes sense that the RNA chaperone activity of ribosomal proteins could play a role during ribosome assembly, a definite proof for the requirement of this activity in vivo has not yet been provided. Recently, it was demonstrated that E.coli ribosomal proteins L15, L16, L18, and L19, that showed RNA chaperone activity in vitro, further possess protein chaperone activity comparable to other protein chaperones such as Hsp90 [43]. It was suggested that intrinsically unstructured domains of ribosomal proteins could play a role in chaperoning. The exact mechanism, however, still remains elusive (see Section 5).

Cold shock proteins (csps) are conserved throughout bacteria and plants. They are expressed during cold-shock, when misfolding of RNAs becomes a major problem for the organism and function as transcriptional antiterminator at low temperature. Many experiments that have been performed to describe chaperone activity of cold-shock proteins utilize DNA helices (and only sometimes in addition RNA duplexes) and refer to the activity as nucleic acid melting activity. However, it has to be mentioned that there are no elaborate studies on whether there is a difference between RNA duplex and DNA duplex melting and whether DNA melting activity automatically is the same as RNA melting.

E. coli contains nine members of the csp family and CspA, the major cold-shock protein and CspE were identified to interact non-specifically with RNA molecules and to possess nucleic acid melting activities [44–46].

Cold-shock proteins in higher plants are highly conserved. Glycine-rich and Zn-finger containing proteins from Arabidopsis thaliana have been monitored for their nucleic acid melting activity, and it was shown that GRP7 (glycine-rich protein) and CSDP1 (cold shock domain protein) possess RNA chaperone activity [47]. A recent study also demonstrated that out of six glycine-rich proteins in rice (Oryza sativa), which are likely to be involved in adaptation to cold-shock, three of them exhibit RNA- (and DNA-) melting activities suggesting that GRPs in plants fulfill a chaperoning role during low temperatures [48].

In E. coli translation, initiation factor 1 (IF1) is a small 71 amino-acid long peptide, which contains 5 rigid β-barrels and belongs to the OB (oligomer-binding)-fold proteins such as the cold shock proteins. N- and C-termini of IF1 are highly flexible. It was demonstrated that E. coli IF1 is capable of complementing for a cspB and cspC double knock-out in Bacillus subtilis suggesting that IF1 and csps have at least partially overlapping activities [49]. E. coli IF1 exhibits RNA chaperone activity in various assays including RNA annealing of complementary oligonucleotides, trans-splicing, in vivo folding trap assay, and transcription anti-termination in vivo and in vitro [25, 50].

Heteronuclear Ribonucleoproteins encompass a group of about 20 polypeptides that are predominantly nuclear in localization and are involved in RNA processing.

The first observation that hnRNPs possess RNA chaperone activity came in the early 1990s when fractionated HeLa nuclear extract was tested for annealing activity of an mRNA and its antisense partner. Three proteins, hnRNP A1, C1 and U, were identified and hnRNP A1 was further shown to enhance hammerhead ribozyme cleavage in vitro [16, 20]. Later, a detailed study on possible functions of Ro RNPs, which are Ro ribonucleoprotein complexes, composed of a small noncoding cytoplasmic RNA, termed Y RNA and its protein partners was conducted: besides the permanently associated proteins Ro60 and La, subpopulations of Ro-RNPs also contain hnRNP I and hnRNP K, both of which exhibited strong RNA chaperone activity in vitro in the trans-splicing and the cis-splicing assay [22]. hnRNP I is identical to poly-pyrimidine binding protein (PTB) isoform 4 and was identified as a splicing suppressor in mammalian cells [51]. It regulates cap-independent translation, localization of cytoplasmic RNAs, and poly-A-site cleavage [52]. PTB belongs to the IRES transacting factors (ITAFs), which are host factors (like La, hnRNP K, nucleolin, unr and many others) that interact with viral RNAs and induce conformational changes that then lead to translation initiation [53]. It was further reported that calcivirus replication requires PTB but only at lower or at higher temperatures than the permissive 37°C, suggesting a chaperoning role of PTB [54]. Members of the group of ITAFs have been implicated in RNA chaperoning like unr, a cold-shock domain containing protein [55], human La protein, and hnRNP K [22].

HnRNP K is also a multifunctional protein that is a transcriptional factor for c-myc and c-src [56–58]; it enhances splicing [59] and is a translational regulator [60].

La proteins primarily bind RNA polymerase III transcripts and protect them from nuclease attack [61]. They also interact with pre-tRNAs at their UUU-3′OH ends and facilitate their maturation. La contributes to assembly of RNP complexes by retaining RNAs in the nucleus. La is also involved in translation regulation. And human La was demonstrated to possess RNA chaperone activity in vitro in the cis-splicing assay and in vivo in the folding trap assay [22].

The bacterial protein Hfq was first discovered in the end 1960s as a host factor for bacteriophage Qβ replication [62]. The bacterial protein is a pleiotropic regulator for gene expression in bacteria. It interacts with many small RNAs and their mRNA targets and regulates posttranscriptional regulation of small noncoding RNAs such as DsrA, sodB, oxyS, rprA, and spot42 [63–67]. Hfq preferentially binds to A/U rich, unstructured regions.

Hfq encompasses an Sm-domain, which is highly conserved among various species and usually is found in eukaryotic spliceosomal RNPs. Crystallographic studies of Escherichia coli, Pseudomonas aeruginosa, and Staphylococcus aureus Hfq proteins showed that Hfq forms homohexameric ring structures with a central cationic pore that forms the RNA binding site [68–70].

Using strand annealing and strand melting assays to measure RNA chaperone activity of Hfq, only strand annealing activity was observed [27]. However, a detailed study using RNase footprinting on Hfq's interaction with its target RNAs sodB mRNA and the small noncoding regulator ryhB RNA demonstrated that Hfq indeed did loosen secondary structures within sodB mRNA that lead to binding of its regulatory RNA rhyB [7]. A similar observation was made using fluorescence labelled rpoS mRNA and dsrA small noncoding RNA [6]. By means of FRET, it was shown that Hfq induces annealing of dsrA to rpoS mRNA and prior to the annealing event Hfq disrupts rpoS secondary structure elements. Consequently, Hfq is entitled to be called RNA chaperone.