RNA pseudoknots. Stability and loop size requirements.

The effects of ionic conditions, loop size and loop sequence on the formation of pseudoknots by RNA oligonucleotides have been investigated using biochemical and biophysical methods. An oligonucleotide with the sequence 5' GCGAUUUCUGACCGCUUUUUUGUCAG 3' and oligonucleotides with variations in the sequences of the two loop regions, denoted by bold face type, were studied. Each sequence with the potential to form a pseudoknot can also form two stable hairpins. The pseudoknot structure is stabilized relative to the hairpins by addition of Mg2+. Even in the presence of Mg2+, the pseudoknots formed by the sequences investigated are only marginally more stable (1.5 to 2 kcal mol-1 in free energy at 37 degrees C) than either of the constituent hairpins. The pseudoknot structure is the stable conformation in the presence of Mg2+ when the first loop region is at least three nucleotides and the second is at least four nucleotides. Further deletion of nucleotides from the loop regions stabilizes possible hairpin structures relative to the pseudoknot and equilibria among secondary and tertiary structures result.     

Structural changes in the 530 loop of Escherichia coli 16S rRNA in mutants with impaired translational fidelity.

The higher order structure of the functionally important 530 loop in Escherichia coli 16S rRNA was studied in mutants with single base changes at position 517, which significantly impair translational fidelity. The 530 loop has been proposed to interact with the EF-Tu-GTP-aatRNA ternary complex during decoding. The reactivity at G530, U531 and A532 to the chemical probes kethoxal, CMCT and DMS respectively was increased in the mutant 16S rRNA compared with the wild-type, suggesting a more open 530 loop structure in the mutant ribosomes. This was supported by oligonucleotide binding experiments in which probes complementary to positions 520-526 and 527-533, but not control probes, showed increased binding to the 517C mutant 70S ribosomes compared with the non-mutant control. Furthermore, enzymatic digestion of 70S ribosomes with RNase T1, specific for single-stranded RNA, substantially cleaved both wild-type and mutant rRNAs between G524 and C525, two of the nucleotides involved in the 530 loop pseudoknot. This site was also cleaved in the 517C mutant, but not wild-type rRNA, by RNase V1. Such a result is still consistent with a more open 530 loop structure in the mutant ribosomes, since RNase V1 can cut at appropriately stacked single-stranded regions of RNA. Together these data indicate that the 517C mutant rRNA has a rather extensively unfolded 530 loop structure. Less extensive structural changes were found in mutants 517A and 517U, which caused less misreading. A correlation between the structural changes in the 530 loop and impaired translational accuracy is proposed.     

Cleavage of a 23S rRNA pseudoknot by phenanthroline-Cu(II)

Studying the intricate folding of rRNA within the ribosome remains a complex problem. Phenanthroline-Cu(II) complexes cleave phosphodiester bonds in rRNA in specific regions, apparently especially where the rRNA structure is constrained in some fashion. We have introduced phenanthroline-copper complexes into 50S Escherichia coli ribosomal subunits and shown specific cleavages in the regions containing nucleotides 60-66 and 87-100. This specificity of cleavage is reduced when the ribosome is heated to 80 degrees C and reduced to background when the ribosomal proteins are extracted and the cleavage repeated on protein-free 23S rRNA. It has been suggested that nucleotides 60-66 and 87-95 in E.coli 23S rRNA are involved in a putative pseudoknot structure, which is supported by covariance data. The paired cleavages of nearly equal intensity of these two regions, when in the ribosome, may further support the existence of a pseudoknot structure in the 100 region of 23S rRNA.     

Stem-loop structures can effectively substitute for an RNA pseudoknot in -1 ribosomal frameshifting

-1 Programmed ribosomal frameshifting (PRF) in synthesizing the gag-pro precursor polyprotein of Simian retrovirus type-1 (SRV-1) is stimulated by a classical H-type pseudoknot which forms an extended triple helix involving base-base and base-sugar interactions between loop and stem nucleotides. Recently, we showed that mutation of bases involved in triple helix formation affected frameshifting, again emphasizing the role of the triple helix in -1 PRF. Here, we investigated the efficiency of hairpins of similar base pair composition as the SRV-1 gag-pro pseudoknot. Although not capable of triple helix formation they proved worthy stimulators of frameshifting. Subsequent investigation of ～30 different hairpin constructs revealed that next to thermodynamic stability, loop size and composition and stem irregularities can influence frameshifting. Interestingly, hairpins carrying the stable GAAA tetraloop were significantly less shifty than other hairpins, including those with a UUCG motif. The data are discussed in relation to natural shifty hairpins.     

A functional pseudoknot in 16S ribosomal RNA.

Several lines of evidence indicate that the universally conserved 530 loop of 16S ribosomal RNA plays a crucial role in translation, related to the binding of tRNA to the ribosomal A site. Based upon limited phylogenetic sequence variation, Woese and Gutell (1989) have proposed that residues 524-526 in the 530 hairpin loop are base paired with residues 505-507 in an adjoining bulge loop, suggesting that this region of 16S rRNA folds into a pseudoknot structure. Here, we demonstrate that Watson-Crick interactions between these nucleotides are essential for ribosomal function. Moreover, we find that certain mild perturbations of the structure, for example, creation of G-U wobble pairs, generate resistance to streptomycin, an antibiotic known to interfere with the decoding process. Chemical probing of mutant ribosomes from streptomycin-resistant cells shows that the mutant ribosomes have a reduced affinity for streptomycin, even though streptomycin is thought to interact with a site on the 30S subunit that is distinct from the 530 region. Data from earlier in vitro assembly studies suggest that the pseudoknot structure is stabilized by ribosomal protein S12, mutations in which have long been known to confer streptomycin resistance and dependence.     

Do mRNA and rRNA binding sites of E.coli ribosomal protein S15 share common structural determinants?

Escherichia coli ribosomal protein S15 recognizes two RNA targets: a three-way junction in 16S rRNA and a pseudoknot structure on its own mRNA. Binding to mRNA occurs when S15 is expressed in excess over its rRNA target, resulting in an inhibition of translation start. The sole apparent similarity between the rRNA and mRNA targets is the presence of a G-U/G-C motif that contributes only modestly to rRNA binding but is essential for mRNA. To get more information on the structural determinants used by S15 to bind its mRNA target as compared to its rRNA site, we used site-directed mutagenesis, substitution by nucleotide analogs, footprinting experiments on both RNA and protein, and graphic modeling. The size of the mRNA-binding site could be reduced to 45 nucleotides, without loss of affinity. This short RNA preferentially folds into a pseudoknot, the formation of which depends on magnesium concentration and temperature. The size of the loop L2 that bridges the two stems of the pseudoknot through the minor groove could not be reduced below nine nucleotides. Then we showed that the pseudoknot recognizes the same side of S15 as 16S rRNA, although shielding a smaller surface area. It turned out that the G-U/G-C motif is recognized from the minor groove in both cases, and that the G-C pair is recognized in a very similar manner. However, the wobble G-U pair of the mRNA is not directly contacted by S15, as in rRNA, but is most likely involved in building a precise conformation of the RNA, essential for binding. Otherwise, unique specific features are utilized, such as the three-way junction in the case of 16S rRNA and the looped out A(-46) for the mRNA pseudoknot.     

Selection of viral RNA-derived tRNA-like structures with improved valylation activities

The tRNA-like structure (TLS) of turnip yellow mosaic virus (TYMV) RNA was previously shown to be efficiently charged by yeast valyl-tRNA synthetase (ValRS). This RNA has a noncanonical structure at its 3'-terminus but mimics a tRNA L-shaped fold, including an anticodon loop containing the major identity nucleotides for valylation, and a pseudoknotted amino acid accepting domain. Here we describe an in vitro selection experiment aimed (i) to verify the completeness of the valine identity set, (ii) to elucidate the impact of the pseudoknot on valylation, and (iii) to investigate whether functional communication exists between the two distal anticodon and amino acid accepting domains. Valylatable variants were selected from a pool of 2 x 10(13) RNA molecules derived from the TYMV TLS randomized in the anticodon loop nucleotides and in the length (1-6 nucleotides) and sequence of the pseudoknot loop L1. After nine rounds of selection by aminoacylation, 42 have been isolated. Among them, 17 RNAs could be efficiently charged by yeast ValRS. Their sequence revealed strong conservation of the second and the third anticodon triplet positions (A(56), C(55)) and the very 3'-end loop nucleotide C(53). A large variability of the other nucleotides of the loop was observed and no wild-type sequence was recovered. The selected molecules presented pseudoknot domains with loop L1 varying in size from 3-6 nucleotides and some sequence conservation, but did neither reveal the wild-type combination. All selected variants are 5-50 times more efficiently valylated than the wild-type TLS, suggesting that the natural viral sequence has emerged from a combination of evolutionary pressures among which aminoacylation was not predominant. This is in line with the role of the TLS in viral replication.     

Equilibrium unfolding pathway of an H-type RNA pseudoknot which promotes programmed -1 ribosomal frameshifting.

The equilibrium unfolding pathway of a 41-nucleotide frameshifting RNA pseudoknot from the gag-pro junction of mouse intracisternal A-type particles (mIAP), an endogenous retrovirus, has been determined through analysis of dual optical wavelength, equilibrium thermal melting profiles and differential scanning calorimetry. The mIAP pseudoknot is an H-type pseudoknot proposed to have structural features in common with the gag-pro frameshifting pseudoknots from simian retrovirus-1 (SRV-1) and mouse mammary tumor virus (MMTV). In particular, the mIAP pseudoknot is proposed to contain an unpaired adenosine base at the junction of the two helical stems (A15), as well as one in the middle of stem 2 (A35). A mutational analysis of stem 1 hairpins and compensatory base-pair substitutions incorporated into helical stem 2 was used to assign optical melting transitions to molecular unfolding events. The optical melting profile of the wild-type RNA is most simply described by four sequential two-state unfolding transitions. Stem 2 melts first in two closely coupled low-enthalpy transitions at low tmin which the stem 3' to A35, unfolds first, followed by unfolding of the remainder of the helical stem. The third unfolding transition is associated with some type of stacking interactions in the stem 1 hairpin loop not present in the pseudoknot. The fourth transition is assigned to unfolding of stem 1. In all RNAs investigated, DeltaHvH approximately DeltaHcal, suggesting that DeltaCpfor unfolding is small. A35 has the thermodynamic properties expected for an extrahelical, unpaired nucleotide. Deletion of A15 destabilizes the stem 2 unfolding transition in the context of both the wild-type and DeltaA35 mutant RNAs only slightly, by DeltaDeltaG degrees approximately 1 kcal mol-1(at 37 degrees C). The DeltaA15 RNA is considerably more susceptible to thermal denaturation in the presence of moderate urea concentrations than is the wild-type RNA, further evidence of a detectable global destabilization of the molecule. Interestingly, substitution of the nine loop 2 nucleotides with uridine residues induces a more pronounced destabilization of the molecule (DeltaDeltaG degrees approximately 2.0 kcal mol-1), a long-range, non-nearest neighbor effect. These findings provide the thermodynamic basis with which to further refine the relationship between efficient ribosomal frameshifting and pseudoknot structure and stability.     

A pseudoknot-compatible universal site is located in the large ribosomal RNA in the peptidyltransferase center

The RNA secondary structure is not confined to a system of the hairpins and can contain pseudoknots as well as topologically equivalent slipped-loop structure (SLS) conformations. A specific primary structure that directs folding to the pseudoknot or SLS is called SL-palindrome (SLP). Using a computer program for searching the SLP in the genomic sequences, 419 primary structures of large ribosomal RNAs from different kingdoms (prokaryota, eukaryota, archaebacteria) as well as plastids and mitochondria were analyzed. A universal site was found in the peptidyltransferase center (PTC) capable of folding to a pseudoknot of 48 nucleotides in length. Phylogenetic conservation of its helices (concurrent replacements with no violation of base pairing, covariation) has been demonstrated. We suggest the reversible folding-unfolding of the pseudoknot for certain stages of the ribosome functioning.     

Mutational analysis of the pseudoknot in the tRNA-like structure of turnip yellow mosaic virus RNA. Aminoacylation efficiency and RNA pseudoknot stability.

Site-directed mutations were introduced in the connecting loops and one of the two stem regions of the RNA pseudoknot in the tRNA-like structure of turnip yellow mosaic virus RNA. The kinetic parameters of valylation for each mutated RNA were determined in a cell-free extract from wheat germ. Structure mapping was performed on most mutants with enzymic probes, like RNase T1, nuclease S1 and cobra venom ribonuclease. An insertion of four A residues in the four-membered connecting loop L1 that crosses the deep groove of the pseudoknot reduces aminoacylation efficiency. Deletions up to three nucleotides do not affect aminoacylation or RNA pseudoknot formation. Deletion of the entire loop abolishes aminoacylation. Although elimination of the pseudoknot is presumed, this could not be demonstrated. Unlike the mutations in loop L1, all mutations in the three-membered connecting loop L2 that crosses the shallow groove of the RNA pseudoknot decrease the aminoacylation efficiency considerably. Nonetheless, the RNA pseudoknot is still present in most mutated RNAs. These results indicate that a number of mutations can be introduced in both loops without abolishing aminoacylation. Results obtained with the introduction of mismatches and A.U base-pairs in stem S1 of the pseudoknot, containing three G.C base-pairs in wild-type RNA, indicate that the pseudoknot is only marginally stable. Our estimation of the gain of free energy due to the pseudoknot formation is at most 2.0 kcal/mol. The pseudoknot structure can, however, be stabilized upon binding the valyl-tRNA synthetase.     

Untying the FIV frameshifting pseudoknot structure by MS3D

The structure of the putative feline immunodeficiency virus (FIV) ribosomal frameshifting pseudoknot (PK) has been investigated by a mass spectrometric three-dimensional (MS3D) approach, which involves the application of established solvent-accessibility probes and chemical crosslinkers with detection by electrospray ionization (ESI) Fourier transform mass spectrometry (FTMS). Regardless of their size, probed substrates can be treated with ribonucleases and analyzed by ESI-FTMS to obtain the correct position of chemically modified nucleotides. Protection maps and distance information can be utilized to generate 3D models using the constraint satisfaction algorithm provided by MC-SYM and the energy minimization modules included in CNS. Control experiments were performed on a mutant of mouse mammary tumor virus pseudoknot (VPK), for which an NMR structure is available. Comparison between the MS3D model and the high-resolution structure provided a approximately 3A root-mean-square deviation calculated from all the atoms present in double-stranded regions. Applied to FIV-PK, the MS3D approach confirmed that the selected sequence could fold into an actual pseudoknot, supporting the sequence alignment predictions. Characteristic features of H-type pseudoknots were recognized immediately, but a putative A13-U30 pair was not observed at the stem junction, making FIV-PK resemble VPK more closely than the initially suggested simian retrovirus type-1 pseudoknot. In our model, the unpaired U30 protrudes into the medium, while the hinging A13 assumes a stacked conformation that enables the stems to form a approximately 60 degrees bend and relieve the strain caused by a short loop 1. The model provided the basis to explain the different alkylation patterns observed in the absence and presence of Mg(2+), suggesting the possible formation of a specific metal-binding site between loop 1 and stem 2. This instance illustrates how the MS3D model of FIV-PK can be utilized effectively to generate hypotheses and support functional observations in the absence of a high-resolution structure.     

An NMR and mutational study of the pseudoknot within the gene 32 mRNA of bacteriophage T2: insights into a family of structurally related RNA pseudoknots.

NMR methods were used to investigate a series of mutants of the pseudoknot within the gene 32 messenger RNA of bacteriophage T2, for the purpose of investigating the range of sequences, stem and loop lengths that can form a similar pseudoknot structure. This information is of particular relevance since the T2 pseudoknot has been considered a representative of a large family of RNA pseudoknots related by a common structural motif, previously referred to as 'common pseudoknot motif 1' or CPK1. In the work presented here, a mutated sequence with the potential to form a pseudoknot with a 6 bp stem2 was shown to adopt a pseudoknot structure similar to that of the wild-type sequence. This result is significant in that it demonstrates that pseudoknots with 6 bp in stem2 and a single nucleotide in loop1 are indeed feasible. Mutated sequences with the potential to form pseudoknots with either 5 or 8 bp in stem2 yielded NMR spectra that could not confirm the formation of a pseudoknot structure. Replacing the adenosine nucleotide in loop1 of the wild-type pseudoknot with any one of G, C or U did not significantly alter the pseudoknot structure. Taken together, the results of this study provide support for the existence of a family of similarly structured pseudoknots with two coaxially stacked stems, either 6 or 7 bp in stem2, and a single nucleotide in loop1. This family includes many of the pseudoknots predicted to occur downstream of the frameshift or readthrough sites in a significant number of viral RNAs.     

Crystal structure of a luteoviral RNA pseudoknot and model for a minimal ribosomal frameshifting motif

To understand the role of structural elements of RNA pseudoknots in controlling the extent of -1-type ribosomal frameshifting, we determined the crystal structure of a high-efficiency frameshifting mutant of the pseudoknot from potato leaf roll virus (PLRV). Correlations of the structure with available in vitro frameshifting data for PLRV pseudoknot mutants implicate sequence and length of a stem-loop linker as modulators of frameshifting efficiency. Although the sequences and overall structures of the RNA pseudoknots from PLRV and beet western yellow virus (BWYV) are similar, nucleotide deletions in the linker and adjacent minor groove loop abolish frameshifting only with the latter. Conversely, mutant PLRV pseudoknots with up to four nucleotides deleted in this region exhibit nearly wild-type frameshifting efficiencies. The crystal structure helps rationalize the different tolerances for deletions in the PLRV and BWYV RNAs, and we have used it to build a three-dimensional model of the PRLV pseudoknot with a four-nucleotide deletion. The resulting structure defines a minimal RNA pseudoknot motif composed of 22 nucleotides capable of stimulating -1-type ribosomal frameshifts.     

Heuristic RNA pseudoknot prediction including intramolecular kissing hairpins

Pseudoknots are an essential feature of RNA tertiary structures. Simple H-type pseudoknots have been studied extensively in terms of biological functions, computational prediction, and energy models. Intramolecular kissing hairpins are a more complex and biologically important type of pseudoknot in which two hairpin loops form base pairs. They are hard to predict using free energy minimization due to high computational requirements. Heuristic methods that allow arbitrary pseudoknots strongly depend on the quality of energy parameters, which are not yet available for complex pseudoknots. We present an extension of the heuristic pseudoknot prediction algorithm DotKnot, which covers H-type pseudoknots and intramolecular kissing hairpins. Our framework allows for easy integration of advanced H-type pseudoknot energy models. For a test set of RNA sequences containing kissing hairpins and other types of pseudoknot structures, DotKnot outperforms competing methods from the literature. DotKnot is available as a web server under http://dotknot.csse.uwa.edu.au.     

A procedure for RNA pseudoknot prediction

The RNA pseudoknot has been proposed as a significant structuralmotif in a wide range of biological processes of RNAs. A pseudoknotinvolves intramolecular pairing of bases in a hairpin loop withbases outside the stem of the loop to form a second stem andloop region. In this study, we propose a method for searchingand predicting pseudoknots that are likely to have functionalmeaning. In our procedure, the orthodox hairpin structure involvedin the pseudoknot is required to be both statistically significantand relatively stable to the others in the sequence. The basesoutside the stem of the hairpin loop in the predicted pseudoknotare not entangled with any formation of a highly stable secondarystructure in the sequence. Also, the predicted pseudoknot issignificantly more stable than those that can be formed froma large set of scrambled sequences under the assumption thatthe energy contribution from a pseudoknot is proportional tothe size of second loop region and planar energy contributionfrom second stem region. A number of functional pseudoknotsthat have been reported before can be identified and predictedfrom their sequences by our method.     

Essential structural features in the Schizosaccharomyces pombe pre-rRNA 5' external transcribed spacer

The proximal region in the 5' external transcribed spacer (5'ETS) of the genes encoding ribosomal RNAs in Schizosaccharomyces pombe was examined with respect to structural features which underlie rRNA maturation. Computer analyses and partial digestion with nuclease probes indicate a crucifix-like structure composed primarily of three extended hairpins which are more highly ordered than previously proposed in Saccharomyces cerevisiae. A re-evaluation of the same region in S. cerevisiae indicates a conserved core structure, including the U3 snoRNA binding site within this higher-order structure. The sequences encoding the individual hairpins were deleted by PCR-mediated mutagenesis and the mutant rDNAs were expressed in vivo to determine the effect of these features on rRNA maturation. Quantitative hybridization analyses indicate that the first hairpin only has modest effects on 18 S rRNA maturation, but the other two regions are critical and no mature 18 S rRNA was observed. When smaller changes were systematically introduced into the critical regions, strong correlations were observed with known or putative events in rRNA maturation. Changes associated with an intermediate cleavage site in helix II and with the putative U3 snoRNA binding site were again critical to 18 S rRNA production. In each case, the effects were sequence dependent and not simply the result of disrupted structure. Further analyses of the 5.8 S rRNA indicate that the large ribosomal subunit RNA can be properly processed in each case but the efficiency is reduced by as much as 60 %, an observation which provides new evidence of interdependency in the maturation process. The results illustrate that rRNA processing is more critically dependent on the 5'ETS than previously believed.     

Pseudoknot-dependent read-through of retroviral gag termination codons: importance of sequences in the spacer and loop 2.

Retroviruses whose gag and pol genes are in the same reading frame depend upon approximately 5% read-through of the gag UAG termination codon to make the gag-pol polyprotein. For murine leukemia virus, this read-through is dependent on a pseudoknot located eight nucleotides 3' of the UAG. Other retroviruses whose gag and pol genes are in the same frame can potentially form similar pseudoknots 3' of their UAG codons. Beyond the similar secondary structures, there is strong sequence conservation in the spacer region and in loop 2 of the pseudoknots. The detrimental effects of substitutions of several of these conserved spacer and loop 2 nucleotides in the murine leukemia virus sequence show their importance for the read-through process. The importance of specific nucleotides in loop 2 of the pseudoknot contrasts with the flexibility of sequence in loop 2 of the most intensively studied frameshift-promoting pseudoknot which occurs in infectious bronchitis virus. Two nucleotides in loop 2 of the murine leukemia virus pseudoknot, which were shown to be important by mutagenic analysis, display hypersensitivity to the single-strand specific nuclease, S1. They are likely to be particularly accessible or are in an unusually reactive conformation.     

Binding of DNA hairpins to an assembler-strand as part of a primordial translation device

A crucial event in the process leading to the origin of life is the emergence of a simple translation device. To approach experimental realization of this device the binding ability of short DNA hairpins to complementary oligonucleotides fixed on a solid support was investigated. The binding is achieved by base pairing between the loop nucleotides of the hairpins containing different numbers of adenosine residues and oligothymidylates covalently linked to cellulose. The loop has to consist of at least five nucleotides to achieve binding. The exact number of established base pairs was determined in two ways. First, the elution temperatures of hairpins and those of oligoadenylates which had the length of the loop were compared. Secondly, the architecture of the loop was analyzed by means of the single-strand-specific nuclease from mung bean acting as structural probe. Onlyn-2 of n loop nucleotides of a hairpin are able to form base pairs. Therefore, a strong evidence for the formation of a triplet of base pairs between primeval tRNA and mRNA sufficient to stabilize the complex enzyme-free is given.     

Solution structure of a luteoviral P1-P2 frameshifting mRNA pseudoknot

A hairpin-type messenger RNA pseudoknot from pea enation mosaic virus RNA1 (PEMV-1) regulates the efficiency of programmed -1 ribosomal frameshifting. The solution structure and 15N relaxation rates reveal that the PEMV-1 pseudoknot is a compact-folded structure composed almost entirely of RNA triple helix. A three nucleotide reverse turn in loop 1 positions a protonated cytidine, C(10), in the correct orientation to form an A((n-1)).C(+).G-C(n) major groove base quadruple, like that found in the beet western yellows virus pseudoknot and the hepatitis delta virus ribozyme, despite distinct structural contexts. A novel loop 2-loop 1 A.U Hoogsteen base-pair stacks on the C(10)(+).G(28) base-pair of the A(12).C(10)(+).G(28)-C(13) quadruple and forms a wedge between the pseudoknot stems stabilizing a bent and over-rotated global conformation. Substitution of key nucleotides that stabilize the unique conformation of the PEMV-1 pseudoknot greatly reduces ribosomal frameshifting efficacy.     

DNA hairpin loops in solution. Correlation between primary structure, thermostability and reactivity with single-strand-specific nuclease from mung bean.

Hairpin structures formed by seven DNA inverted repeats have been studied by PAGE, UV(CD)-spectroscopy and nuclease cleavage. The hairpins consisted of (CG)3 stems and loops of 2, 3 and 4 residues. Thermal stabilities (Tm) have been determined in low and high ionic strength buffers, where the hairpins were structured in the B- and Z-DNA form respectively. The thermodynamic parameters of hairpin formation have been obtained by a two-state analysis of the hairpin-coil transitions. It is found that, on increasing the number of bases in the loop from 2 to 3 and 4, the Tms of the B-hairpins decrease, whereas the Tms of the same hairpins in the Z-form increase. This confirms previous evidence (1,2) that in a hairpin molecule the size and structure of the loop are modulated by the conformation of the helical stem. Moreover, B-hairpins with loops comprising 2, 3 and 4 bases have been digested with the single-strand-specific nuclease from mung bean. In our experimental conditions (0 degrees C) the nuclease preferentially cleaves the unbonded nucleotides of the loops. However, the rates of loop hydrolysis, which roughly follow a first-order kinetics, markedly depend on the size of the loop. At a ratio of 3 enzyme units/micrograms DNA, the half-lives of hairpins which are expected to form loops of 4, 3 and 2 residues are 90, 145 and 440 minutes respectively. Thermostability and enzymatic digestion data suggest that two-membered loops can be formed in B-hairpins but not in Z-hairpins.