8
DOI: 10.1002/cbic.201100487 MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation Yuan Fu, Sreenivasa Rao Ramisetty, Nejmun Hussain, and Anne M. Baranger* [a] Introduction Muscleblind-like (MBNL) proteins have been shown to be in- volved in regulation of the alternative splicing of pre-mRNA. [1–3] A toxic RNA gain-of-function model has suggested that the downregulation of muscleblind-like proteins by an expansion of repeated RNA sequences is one molecular cause of myoton- ic dystrophy (dystrophia myotonica, DM), which is a multisys- tem inherited disease characterized by progressive muscle wasting and weakness, cardiac conduction defects, and other neuromuscular problems. There are two types of DM, each as- sociated with a repeated sequence of RNA. Patients with myo- tonic dystrophy type 1 (DM1) have a poly(CTG) repeat in the 3-untranslated region of the dystrophia myotonic protein kinase (DMPK) gene. Patients with myotonic dystrophy type 2 (DM2) bear a poly(CCTG) repeat in the CCHC-type zinc finger 9 (ZNF9) gene. [1–5] There is considerable support for a toxic RNA gain-of-function of the poly(CUG) or poly(CCUG) transcripts, which tend to accumulate in nuclear foci, as the molecular basis of DM. [6–9] MBNL1, a muscleblind-like protein that is involved in regulat- ing pre-mRNA splicing, is strongly recruited into ribonuclear foci comprised of poly(CUG) or poly(CCUG) RNA in muscle tis- sues from patients. [6] It has been shown that loss of MBNL1 to the nuclear foci accounts for about 80–90 % of the total mis- splicing events observed in DM. [10] Among these missplicing events, at least four have been identified to be directly regulat- ed by MBNL1 through binding with specific regions of pre- mRNA. [11–19] MBNL1 is overexpressed in heart and skeletal muscle, which are the two tissues prominently affected in DM. [7, 20] MBNL1-deficient mice show similar symptoms as pa- tients with DM, and overexpression of MBNL1 reverses some symptoms of DM in a mouse poly(CUG) model. [20–23] MBNL1 contains four CCCH zinc finger domains, positioned as tandem pairs at the N terminus (ZnF1 and ZnF2) and the middle (ZnF3 and ZnF4) of the protein. ZnF1 and ZnF3 share a similar amino acid sequence, and ZnF2 and ZnF4 share similar sequences. In ZnF1 and ZnF3 domains, the Cys residues are found in CX 7 CX 6 CX 3 H sequences, and in CX 7 CX 4 CX 3 H sequences of the ZnF2 and ZnF4 domains. The RNA-binding activity of MBNL1 resides in the two tandem zinc finger domains (Scheme 1). [18, 19] The MBNL1 protein binds to RNA targets, including both helical and single-stranded RNA containing 5-YCGY-3sequen- ces. [11, 25] Structured helical sequences include the poly(CUG) repeat sequences of DM1, which form structures that resemble A-form RNA. [26, 27] Sequences with less stable secondary struc- tures include cardiac troponin T (cTNT), which is targeted by MBNL1 to regulate alternative splicing. MBNL1 can bind to a 32-mer segment of the cTNT pre-mRNA and promote the in- Scheme 1. Sequences of the four zinc fingers of MBNL1. The sites of muta- tions are marked with asterisks above the amino acids. Amino acids with side chains that interact with RNA in the structure of the complex [24] are marked above with triangles. Cys and His residues that chelate with zinc ion are in white on a black background. Positively charged amino acids are highlighted in white on a grey background. Aromatic amino acids are high- lighted in black on a grey background. Muscleblind-like proteins (MBNL) are RNA-binding proteins that bind to the poly(CUG) and poly(CCUG) sequences that are the causative agents of myotonic dystrophy. It has been sug- gested that as a result of binding to the repeating RNA sequences, MBNL1 is abnormally expressed and translocated, which leads to many of the misregulated events in myotonic dystrophy. In this work, steady-state fluorescence quenching experiments suggest that MBNL1 alters the structure of helical RNA targets upon binding, which may explain the selectivity of MBNL1 for less structured RNA sites. The removal of one pair of zinc fingers greatly impairs the binding affinity of MBNL1, which indicates that the two pairs of zinc fingers might possi- bly interact with RNA targets cooperatively. Alanine scanning mutagenesis results suggest that the binding energy may be distributed across the protein. Overall, the results presented here suggest that small molecules that stabilize the helical structure of poly(CUG) and poly(CCUG) RNAs will inhibit the formation of complexes with MBNL1. [a] Y. Fu, Dr. S. R. Ramisetty, N. Hussain, Prof. A. M. Baranger Department of Chemistry, University of Illinois 600 South Mathews Avenue, Urbana, IL 61801 (USA) E-mail : [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201100487. 112 # 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2012, 13, 112 – 119

MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation

  • Upload
    yuan-fu

  • View
    212

  • Download
    0

Embed Size (px)

Citation preview

Page 1: MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation

DOI: 10.1002/cbic.201100487

MBNL1–RNA Recognition: Contributions of MBNL1Sequence and RNA ConformationYuan Fu, Sreenivasa Rao Ramisetty, Nejmun Hussain, and Anne M. Baranger*[a]

Introduction

Muscleblind-like (MBNL) proteins have been shown to be in-volved in regulation of the alternative splicing of pre-mRNA.[1–3]

A toxic RNA gain-of-function model has suggested that thedownregulation of muscleblind-like proteins by an expansionof repeated RNA sequences is one molecular cause of myoton-ic dystrophy (dystrophia myotonica, DM), which is a multisys-tem inherited disease characterized by progressive musclewasting and weakness, cardiac conduction defects, and otherneuromuscular problems. There are two types of DM, each as-sociated with a repeated sequence of RNA. Patients with myo-tonic dystrophy type 1 (DM1) have a poly(CTG) repeat in the3’-untranslated region of the dystrophia myotonic proteinkinase (DMPK) gene. Patients with myotonic dystrophy type 2(DM2) bear a poly(CCTG) repeat in the CCHC-type zinc finger 9(ZNF9) gene.[1–5] There is considerable support for a toxic RNAgain-of-function of the poly(CUG) or poly(CCUG) transcripts,which tend to accumulate in nuclear foci, as the molecularbasis of DM.[6–9]

MBNL1, a muscleblind-like protein that is involved in regulat-ing pre-mRNA splicing, is strongly recruited into ribonuclearfoci comprised of poly(CUG) or poly(CCUG) RNA in muscle tis-sues from patients.[6] It has been shown that loss of MBNL1 tothe nuclear foci accounts for about 80–90 % of the total mis-splicing events observed in DM.[10] Among these missplicingevents, at least four have been identified to be directly regulat-ed by MBNL1 through binding with specific regions of pre-mRNA.[11–19] MBNL1 is overexpressed in heart and skeletalmuscle, which are the two tissues prominently affected inDM.[7, 20] MBNL1-deficient mice show similar symptoms as pa-tients with DM, and overexpression of MBNL1 reverses somesymptoms of DM in a mouse poly(CUG) model.[20–23]

MBNL1 contains four CCCH zinc finger domains, positionedas tandem pairs at the N terminus (ZnF1 and ZnF2) and themiddle (ZnF3 and ZnF4) of the protein. ZnF1 and ZnF3 share a

similar amino acid sequence, and ZnF2 and ZnF4 share similarsequences. In ZnF1 and ZnF3 domains, the Cys residues arefound in CX7CX6CX3H sequences, and in CX7CX4CX3H sequencesof the ZnF2 and ZnF4 domains. The RNA-binding activity ofMBNL1 resides in the two tandem zinc finger domains(Scheme 1).[18, 19]

The MBNL1 protein binds to RNA targets, including bothhelical and single-stranded RNA containing 5’-YCGY-3’ sequen-ces.[11, 25] Structured helical sequences include the poly(CUG)repeat sequences of DM1, which form structures that resembleA-form RNA.[26, 27] Sequences with less stable secondary struc-tures include cardiac troponin T (cTNT), which is targeted byMBNL1 to regulate alternative splicing. MBNL1 can bind to a32-mer segment of the cTNT pre-mRNA and promote the in-

Scheme 1. Sequences of the four zinc fingers of MBNL1. The sites of muta-tions are marked with asterisks above the amino acids. Amino acids withside chains that interact with RNA in the structure of the complex[24] aremarked above with triangles. Cys and His residues that chelate with zinc ionare in white on a black background. Positively charged amino acids arehighlighted in white on a grey background. Aromatic amino acids are high-lighted in black on a grey background.

Muscleblind-like proteins (MBNL) are RNA-binding proteinsthat bind to the poly(CUG) and poly(CCUG) sequences that arethe causative agents of myotonic dystrophy. It has been sug-gested that as a result of binding to the repeating RNAsequences, MBNL1 is abnormally expressed and translocated,which leads to many of the misregulated events in myotonicdystrophy. In this work, steady-state fluorescence quenchingexperiments suggest that MBNL1 alters the structure of helicalRNA targets upon binding, which may explain the selectivity of

MBNL1 for less structured RNA sites. The removal of one pairof zinc fingers greatly impairs the binding affinity of MBNL1,which indicates that the two pairs of zinc fingers might possi-bly interact with RNA targets cooperatively. Alanine scanningmutagenesis results suggest that the binding energy may bedistributed across the protein. Overall, the results presentedhere suggest that small molecules that stabilize the helicalstructure of poly(CUG) and poly(CCUG) RNAs will inhibit theformation of complexes with MBNL1.

[a] Y. Fu, Dr. S. R. Ramisetty, N. Hussain, Prof. A. M. BarangerDepartment of Chemistry, University of Illinois600 South Mathews Avenue, Urbana, IL 61801 (USA)E-mail : [email protected]

Supporting information for this article is available on the WWW underhttp ://dx.doi.org/10.1002/cbic.201100487.

112 � 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2012, 13, 112 – 119

Page 2: MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation

clusion of exon 5 in cTNT.[16, 18, 24, 28] Here, we have used in vitrobinding assays and steady-state fluorescence measurementswith purified recombinant MBNL1 protein to probe the interac-tion of MBNL1 protein with RNAs of different structures. Ourdata suggest that that MBNL1 disrupts helical RNA structureupon binding and that the binding energy is distributedthroughout the two pairs of the zinc fingers of MBNL1.

Results

Analysis of the requirements of RNA secondary structure forMBNL1 binding

To investigate the binding of MBNL1 to RNA targets with dif-ferent structures, the oligonucleotide (CUG)12, which includes astabilizing GC at the end of the stem, was selected as a mimicof pathogenic, structured RNA in DM1 (Scheme 2). As a mimicof the natural target sites of MBNL1, a segment of cTNT com-prised of 18 nucleotides (cTNT18) was selected. Although themost stable secondary structure of cTNT18 that is predicted byUNAfold is shown in Scheme 2,[29, 30] the Tm was determined tobe 26�6 8C. Thus, the structure of this RNA is likely to be anequilibrating mixture of stem loop and single-stranded forms.In contrast, (CUG)12 forms a more stable structure with a Tm of71�2 8C. The homodimers formed by 5’-G(CUG)2C-2’ and 5’-(CUG)6-2’ were found to be A-form RNA helices in crystal struc-tures.[26, 27] All experiments reported here were carried out withMBNL1, which is comprised of amino acids 1–260 of MBNL,and binds to RNA with similar affinity as the full-length pro-tein.[18, 31] Apparent dissociation constants for MBNL1–RNAcomplexes were determined using gel electrophoresis mobilityshift assays. The apparent dissociation constants for binding to(CUG)12 are much larger than those for binding to cTNT18, andare similar to previously reported values.[18, 25]

Because most RNA targets of MBNL1 contain more than one5’-YGCY-3’ binding motif, we investigated the contribution ofeach of the 5’-YGCY-2’ motifs to complex stability. We com-pared the binding affinities of two cTNT18 variants in whicheither of the two 5’-GC-2’ sequences was replaced with a 5’-AA-2’ sequence (Figure 1). The mutations disrupt only onebinding motif, while leaving the other motif intact. Adenineswere chosen to replace GC because substitution with A is un-likely to introduce a new MBNL1 binding site. Modification atthe 5’-GC binding motif (cTNT18_AA1) resulted in an approxi-mately 6000-fold increase in the Kd value, while modification atthe second binding motif (cTNT18_AA2) resulted in an approxi-mately 20-fold increase in Kd (Table 1). These results suggestthat both 5’-YGCY-2’ motifs contribute to the interaction withMBNL1. The different apparent dissociation constants for thetwo modified cTNT18 RNAs suggest that the presence of 5’-YGCY-2’ is not the only factor that influences the binding affini-

Scheme 2. Most stable secondary structures predicted by UNAfold[29, 30] of(CUG)12, cTNT32, and cTNT18. The box in the cTNT32 RNA indicates the se-quence of cTNT18 RNA.

Figure 1. Electrophoretic gel mobility shift assays of MBNL1 with cTNT18 and analogues. One representative gel is presented for each RNA target. The RNAstructures shown in the figures are the most stable secondary structures predicted by UNAfold.[29, 30] Sequences highlighted in boxes in the structures arespeculated binding sites. Positions of mutations in 5’-YGCY-2’ motifs in cTNT18_AA1 and cTNT18_AA2 are noted with arrows. In all gels the far right lanecontains RNA only.

ChemBioChem 2012, 13, 112 – 119 � 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chembiochem.org 113

MBNL1–RNA Recognition

Page 3: MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation

ty of MBNL1. However, it is worth noting that these modifica-tions not only disrupt the binding motif, but also change thefolding of the RNA structure. It is possible that the AA se-quence substituted for the more 3’-GC motif can pair with 5’-UU-2’ in the loop region and make the 5’-GC more accessibleto binding MBNL1.

To probe the ability of MBNL1 to bind 5’-YCGY-2’ target sitesin different structural contexts, the cTNT18 RNA was modifiedso that the two 5’-YCGY-2’ motifs would be separated(cTNT18m1) or closely associated (cTNT18m2) in the moststable secondary structures predicted by UNAfold. Similar dis-sociation constants were obtained for complexes formed withcTNT18 and cTNT18m1 RNA even though in cTNT18m1 thetwo 5’-YCGY-2’ motifs are forced apart by a double-strandedregion formed between the original loop region of cTNT18and additional RNA sequence at the 2’ end of cTNT18 RNA. Incontrast, when a stable stem region is located in the RNAstructure to bring the two 5’YGCY2’ motifs into close proximity,as in cTNT18m2 RNA, the complex is destabilized (Kd>

1000 nm). The high Tm values and low folding free energies ofthe cTNT18m1 and cTNT18m2 RNAs indicate that they foldinto stable secondary structures at the temperatures used forthe gel mobility shift assays.

Changes of RNA conformation upon binding MBNL1

To investigate conformational changes of the RNA upon bind-ing, we designed an RNA target (cTNT21, Figure 2 A) with a flu-orophore (FAM) and a quencher (Dabcyl) appended to the 5’and 3’ termini, respectively. Analysis of the structure with UNA-fold[29, 30] suggests that this RNA forms a duplex structure inwhich the quencher and fluorophore are in close proximity.Consistent with this structure, the Tm value of the labeled RNA(cTNT21*) was determined to be 52�3 8C (Figure S3 A in theSupporting Information). Because all the fluorescence measure-ments were performed at 20 8C, the labeled RNA moleculesshould form stable stem loop structures in these experiments.The affinity of cTNT21 for MBNL1 was 14�3 nm, which is ap-proximately tenfold weaker than that of cTNT18, presumablybecause of the two additional base pairs, which increase thesimilarity of this RNA target to (CUG)12.

The fluorescence intensity of cTNT21* should increase if thetwo strands are separated because quenching efficiency is di-rectly correlated to the distance between the quencher andfluorophore.[32] We observed a weak signal from the free RNAand a significant increase in fluorescence intensity upon addi-tion of protein (Figure 2 B). The increase in fluorescence inten-sity does not originate from the protein because a mixture ofMBNL1 and unlabeled RNA gave no fluorescence signal. As apositive control, antisense DNA was added to the labeled RNAsolution in 1:1 or 10:1 ratios, which resulted in a large increasein fluorescence intensity, presumably because the formation ofthe DNA–RNA duplex disrupts the RNA stem–loop structure. Inaddition, binding of MBNL1 to an RNA molecule labeled onlywith the fluorophore FAM resulted in only a small increase influorescence signal compared to the RNA labeled with boththe fluorophore and quencher. Together, these results suggestthat the conformation of the RNA molecule is changed uponprotein binding such that the 3’ and 5’ ends of the stem loopare separated.

The conformational changes of the RNA target upon MBNL1binding suggested by the fluorescence studies describedabove are supported by circular dichroism (CD) spectroscopy.The CD spectra of (CUG)12 cTNT21, and cTNT18 have maximaat 266, 273, and 275 nm, respectively. The CD spectrum of(CUG)12 is similar to those of (CUG)4 and (CUG)5 spectra thatwere reported previously.[18, 33] The (CUG)12 spectrum is consis-tent with A-form RNA, which has a maximum between 260and 270 nm.[34, 35] The maxima of cTNT18 and cTNT21 are atlonger wavelengths than that of (CUG)12, which is consistentwith these RNA sequences having reduced helical structurecompared to (CUG)12.

Upon binding of the RNA targets to MBNL1, the peak inten-sity of the (CUG)12 spectrum decreased by 23 %, that of cTNT21decreased by 14 %, and that of cTNT18 remained nearly un-changed. The decrease in peak intensity implies reduced basestacking in the bound structures of (CUG)12 and cTNT21.[36–38]

Table 1. RNA melting temperatures, free energies of folding, and equilib-rium dissociation constants of complexes formed with MBNL1.[a]

RNA name Tm [oC] DG [kcal mol�1] Kd [nm]

(CUG)12 71�2 �12.8�0.4 150�20cTNT18 26�6 n.d.[b] 1.3�0.2cTNT18_AA1 n.d. n.d. >6000cTNT18_AA2 n.d. n.d. 26�4cTNT18_m1 >89 n.d. 0.5�0.3cTNT18_m2 69�3 �7.8�0.3 >1000cTNT21* 52�3 �3.9�0.3 14�3

[a] Errors are the standard deviation of at least three independent meas-urements. [b] n.d. indicates that the free energy of folding could not beunambiguously determined from the absorbance versus temperaturecurve.

Figure 2. A) Schematic illustration of the structure of the cTNT21 RNA la-beled with the fluorophore (FAM) and quencher (Dabcyl). B) Quantificationof fluorescence emission intensity at 520 nm under different conditions;cTNT21* is labeled with FAM and Dabcyl ; cTNT21f is labeled with FAM only.Concentrations: MBNL1 (1.5 mm), RNA (20 nm).

114 www.chembiochem.org � 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2012, 13, 112 – 119

A. M. Baranger et al.

Page 4: MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation

Upon binding MBNL1, the wavelength maximum was red-shift-ed by 3 nm for (CUG)12, 3 nm for cTNT18, and 5 nm forcTNT21, which is consistent with changes in RNA structure oc-curring upon MBNL1 binding. A decrease in peak intensity andshifting towards longer wavelengths upon MBNL1 binding waspreviously observed with (CUG)4 RNA.[18] This type of change isalso typical when ssDNA-binding proteins bind DNA and RNAtargets.[36, 39] These data support a model in which the struc-tures of (CUG)12 and cTNT21 are altered towards single-strand-ed upon binding MBNL1, while that of cTNT18 undergoesstructural changes. The CD spectra of the bound cTNT18p andcTNT18 RNAs are nearly identical, although those of the freeRNA differ, suggesting that the structures of these RNA se-quences become similar upon binding MBNL1. (Figure 3)

Evaluation of the RNA binding affinities of truncated andmutated MBNL1 proteins

To investigate whether pairs of the zinc finger motifs found inMBNL1 are able to bind RNA, we created ZnF12, which is com-prised of the two tandem zinc finger motifs at the N terminusof the protein, and ZnF34, which is comprised of the more C-terminal tandem zinc finger motifs. ICP-MS analysis suggeststhat the truncated proteins have a similar zinc inclusion rateper zinc finger compared to MBNL1N (Figure S4). The com-plexes formed with both CUG12 and cTNT18 RNAs are destabi-lized by truncation of either pair of zinc fingers (Table 2). No

binding of either construct was observed to (CUG)12 RNA,while binding to cTNT18 RNA of either ZnF12 or ZnF34 wasobserved only at the highest protein concentrations (Kd>3750for GST-ZnF12; Kd>5000 for GST-ZnF34; see Figure S4).

To probe the contribution of individual amino acids to therecognition of RNA by MBNL1 we carried out site-directed mu-tagenesis involving ten amino acids (Scheme 1). Conserved ar-omatic amino acid residues in all four motifs were selected.Phe202 and Tyr236 were shown in the crystal structure of thecomplex formed between zinc fingers 3 and 4 and CGCUGU tointeract with G and C bases in the RNA target.[24] Phe36 andTyr68 are placed at similar positions in zinc fingers 1 and 2. Inaddition, similarly placed aromatic amino acids in the CCCHmotifs of the TIS11d protein intercalate between UU and AUdiribonucleotides.[40] Two arginine residues (Arg195 andArg201) that interact with RNA in the crystal structure weremutated.[24]

Each selected amino acid was replaced with alanine to avoidintroducing other functional groups or a bias for a particularpeptide structure or solvent exposure. The dissociation con-stants of mutant proteins with (CUG)12 and cTNT18 RNAs aregiven in Table 2 (for representative gel pictures see Figure S5).The binding affinities of the mutated proteins for cTNT18 werenearly identical. The mutations resulted in up to a fourfolddestabilization of the complexes formed with (CUG)12 with thelargest destabilization observed for Phe54Ala, Tyr68Ala, andArg201Ala mutations. These modest destabilizations show thatnone of these individual amino acids are essential to binding.

Discussion

Truncation of MBNL1 destabilizes complexes with RNA

The weak binding of ZnF12 and ZnF34 to RNA suggests thatone tandem pair of zinc fingers can bind RNA, but with muchlower affinity than MBL1N. There is conflicting data in the liter-ature about whether tandem pairs of zinc fingers can bindRNA. These inconsistencies may result from different buffersbeing used during binding measurements.[24, 31] Because ourgoal is to compare the binding affinities of truncated and full

Figure 3. Circular dichroism spectra of A) (CUG)12 RNA free (square) andbound to MBNL1 (triangle) and B) Free cTNT18 (&) and cTNT21 (&) free andcTNT18 (~) and cTNT21 (~) bound to MBNL1N. Concentrations: RNA(2.5 mm), protein (12 mm).

Table 2. Equilibrium dissociation constants of wild-type and mutantMBNL1 proteins with (CUG)12 and cTNT18 RNA sequences.[a]

Protein (CUG)12

[nm]cTNT18[nm]

Protein (CUG)12

[nm]cTNT18[nm]

wild-type 150�20 1.3�0.2 Phe22Ala 270�50 0.6�0.4Phe36Ala 390�20 1.4�0.1 Phe54Ala 510�50 2�1Tyr68Ala 540�50 1.6�0.2 Tyr188Ala 160�10 1.3�0.4Arg195Ala 180�50 1.0�0.4 Arg201Ala 500�100 0.6�0.2Phe202Ala 220�50 1.3�0.3 Arg231Ala 190�30 1.3�0.2Tyr236Ala 70�20 0.9�0.3 wild-type-

GST160�30 4�2

ZnF12-GST

n.b.[b] >3750 ZnF34-GST n.b. >5000

[a] Errors are the standard deviation of at least three independent meas-urements. [b] n.b. indicates that no binding was observed in the gel mo-bility shift assay.

ChemBioChem 2012, 13, 112 – 119 � 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chembiochem.org 115

MBNL1–RNA Recognition

Page 5: MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation

length MBNL1N, binding measurements were performed usingthe same conditions for the truncated and full-length proteins.The data imply that the two pairs of zinc fingers may coopera-tively interact with RNA substrates. However, more experi-ments need to be carried out in the future to confirm the co-operativity between the zinc fingers. The small complex desta-bilization caused by substitution of individual amino acids withalanine may indicate that the binding energy is distributedthrough large portions of the zinc finger regions. The similarityof the effects of MBNL1 truncation and mutation on the stabili-ty of complexes formed with (CUG)12 and cTNT18 RNAs sup-ports similar recognition of structured and unstructured RNAtargets by MBNL1.

MBNL1 alters the structure of helical RNA targets

The data presented here suggest that RNA targets with highhelicity, such as poly(CUG), are shifted towards single-strandedstructures upon binding MBNL1. Thus, when MBNL1 binds toRNA that has a significant secondary structure, the associationof the RNA with MBNL1 is in competition with the formationof the RNA secondary structure. As a result the binding affinityof structured RNA targets for MBNL1 is reduced compared toless structured RNA targets.

These observations may explain the affinities of MBNL1 forRNA targets with different secondary structures. For example,in poly(CUG) repeats, although the presence of the U–U mis-match lowers the foldingenergy of RNAs, the secondarystructure is stable. As a resultthe dissociation constant of thecomplex formed with (CUG)12

RNA is two orders of magnitudegreater than that of the com-plex formed with cTNT18 RNA,a nearly single-stranded targetsite. The difference in bindingfree energy of MBNL1 for(CUG)12 and cTNT18 RNA is simi-lar to the calculated energeticpenalty for disrupting the sec-ondary structure of (CUG)12 atan MBNL1 binding site.[41–43]

Consistent with this hypothesis,the dissociation constant of thecomplex formed with cTNT21RNA is in between those of thecomplexes formed with (CUG)12

and cTNT18 RNA. When the sta-bility of the secondary structureof the RNA substrate is in-creased, for example cTNT18m2in Figure 1, the protein nolonger binds to the RNA be-cause the protein–RNA interac-tion cannot compete with theformation of the RNA secondary

structure. Thus, there is an inverse relationship between thestability of the RNA secondary structure and that of the com-plex.

RNA secondary structure may play a role in regulation ofalternative splicing by MBNL1

The preference of MBNL1 for RNA targets with little secondarystructure may be important for MBNL1 function. MBNL1 recog-nizes 5’-YGCY-2’ (Y = C/U) motifs in pre-mRNA,[11] but suchmotifs are expected to have a high occurrence in pre-mRNAsequences. If each of the four bases were to have a 25 %chance of occupying a certain position on the pre-mRNA se-quence, the 5’-YGCY-2’ motif would be found every 64 nucleo-tides on average. Indeed, a sequence search by Berglund andco-workers revealed hundreds of potential MBNL1 bindingmotifs in pre-mRNAs that are misregulated in DM.[11] However,it may be possible to rule out some of these potential bindingsites by considering the preference of MBNL1 for RNA sub-strates with low folding energy. We carried out a slidingwindow analysis of pre-mRNA folding on the two best studiedMBNL1 binding sites in the human genome, intron 4 of cTNTpre-mRNA and intron 11 of insulin receptor pre-mRNA.[16, 18, 44]

As expected, there is a relatively low mRNA free folding energyat previously identified binding sites, which are highlighted inblack in Figure 4.[16, 18, 44]

Figure 4. Sliding window analysis of the free energy of folding pre-mRNA. A) Human cTNT; B) Human insulin re-ceptor. Identified binding sites are in black.

116 www.chembiochem.org � 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2012, 13, 112 – 119

A. M. Baranger et al.

Page 6: MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation

The data presented here suggest that MBNL1 may be a newexample of an RNA chaperone.[45, 46] It could possibly regulatealternative splicing by helping the pre-mRNAs fold into confor-mations favorable for splicing without consumption of ATP.One of the best studied RNA and DNA chaperones is the HIV-1nucleocapsid (NC) protein, which uses CCHC zinc fingers tobind RNA.[47–49] Like MBNL1, the NC protein binds with higheraffinity to single-stranded than double-stranded nucleic acids.Binding of the NC protein to hairpin DNA sequences shifts theequilibrium towards a partial unwinding of the helix.[50] Al-though it has been proposed that MBNL1 could block the spli-ceosome assembly to promote the exclusion of exons,[18] it isstill unknown how MBNL1 promotes the inclusion of certainexons. The conformational change of RNA induced by MBNL1may play an important role in these activities.

New perspectives for drug development for myotonicdystrophy

Our results provide new insights for destabilization of the in-teractions of MBNL1 with poly(CUG) or poly(CCUG) RNA se-quences.[51–56] First, our data suggest that binding of MBNL1 tocTNT21* results in at least a partial dissociation of the duplexstructure. Therefore, to bind to an RNA target, MBNL1 mayneed to overcome an energy barrier for melting the RNA, andany molecule that stabilizes poly(CUG) or poly(CCUG) in aduplex structure could inhibit binding of MBNL1. Second, theresults of the experiments performed with truncated MBNL1proteins suggest that for an inhibitor that targets the protein,blocking the RNA binding of one pair of zinc fingers may besufficient to destabilize the complex because the protein withonly one pair of zinc fingers binds RNA with low binding affini-ty. These findings suggest approaches for drug developmentand may contribute to an understanding of molecules identi-fied from screening approaches.

Experimental Section

Expression vectors: The expression vector pGEX-6p-1/MBNL1Nwas obtained from Maurice S. Swanson (University of Florida, Col-lege of Medicine, Gainesville, FL).[19] The expressed protein is com-prised of amino acids 1–260 from human MBNL1 and a His6 tag atthe C-terminus. The expression vectors for ZnF12 and ZnF34 trun-cated proteins were constructed by amplifying the appropriatesequences of pGEX-6p-1/MBNL1N using the following primers:ZnF12: 5’-CGGCG GGATC CATGG CTGTT AGTGT CA-3’, 5’-CGGCCGCTCG AGGAT TACTC GTCCA TT-3’; ZnF34: 5’-CGGGA TCCAT GTTAATGCGA ACAGA C-3’, 5’-CCGCT CGAGC TGGTA TTGGG-3’.

The amplified sequences were subcloned into pGEX-6p-1 using theBamHI and XhoI (Invitrogen) restriction sites. The correct sequen-ces were confirmed by DNA sequencing.

For site-directed mutagenesis, the pGEX-6p-1/MBNL1N vector wasamplified using the following primers: Phe22Ala: 5’-GTATG TAGAGAGGCC CAGAG GG-3’, 5’-GTCCC CCTCT GGGCC TCTCT AC-3’;Phe36Ala: 5’-GGAAT GTAAA GCTGC ACATC CTTCG-3’, 5’-CGAAGGATGT GCAGC TTTAC ATTCC-3’; Phe54Ala: 5’-ATCGC CTGCG CTGATTCATT-3’, 5’-AATGA ATCAG CGCAG GCGAT-3’; Tyr68Ala: 5’-GAACTGCAAA GCTCT TCATC CA-3’, 5’-TGGAT GAAGA GCTTT GCAGT TC-3’;

Tyr188Ala: 5’-TATGT CGAGA GGCCC AACGT-3’, 5’-ACGTT GGGCCTCTCG ACATA-3’; Arg195Ala: 5’-AATTG CAACG CAGGA GAAAA TG-3’, 5’-CATTT TCTCC TGCGT TGCAA TT-3’; Arg201Ala: 5’-GAAAATGATT GTGCG TTTGC TC-3’, 5’-GAGCA AACGC ACAAT CATTT TC-3’;Phe202Ala: 5’-TGATT GTCGG GCTGC TCATC C-3’, 5’-AGGAT GAGCAGCCCG ACAAT CA-3’; Arg231Ala: 5’-GGAGA TGCTC TGCGG AAAAGTG-3’, 5’-CACTT TTCCG CAGAG CATCT CC-3’; Tyr236Ala: 5’-AAGTGCAAAG CCTTT CATCC CC-3’, 5’-GGGGA TGAAA GGCTT TGCAC TT-3’.

The PCR products were treated with DpnI (Invitrogen), transformedinto XL1-blue competent cells (Stratagene), and isolated using QIA-quick Spin Miniprep Kit (Qiagen). The correct sequences were con-firmed by DNA sequencing.

Recombinant protein preparation and purification: Proteins wereexpressed in BL21-CodonPlus(DE3)-RP cells (Stratagene) in LBmedia and induced with IPTG (1 mm) at OD600 0.6. The cultureswere grown for an additional 2 h at 37 8C after induction. The cellswere collected by centrifugation, resuspended in lysis buffer(Tris·HCl (25 mm), NaCl (0.5 m), imidazole (10 mm), BME (2 mm), 5 %glycerol, 0.1 % Triton X-100, lysozyme (2 mg mL�1), PMSF (0.1 mm),pepstatin (1 mm), leupeptin (1 mm), pH 8), and lysed by ultrasonica-tion. The cell lysate was centrifuged (13 000 g, 10 min), and the su-pernatant was collected and filtered through a 45 mm Millex filter(Millipore). The sample was incubated with Ni–NTA-agarose(Qiagen) for 1 h at 4 8C, washed with buffer (Tris·HCl (25 mm), NaCl(0.5 m), imidazole (20 mm) and 0.1 % Triton X-100, pH 8), andeluted with an elution buffer (Tris·HCl (25 mm), NaCl (0.5 m), imida-zole (250 mm), and 0.1 %Triton X-100, pH 8). The eluent was col-lected and incubated with glutathione-Sepharose 4B (GE Health-care) for 1 h at 4 8C. After washing with a buffer containing Tris·HCl(25 mm), NaCl (300 mm), BME (5 mm) and 0.1 % Triton X-100, pH 8,the beads were collected and incubated with PreScission protease(GE Healthcare) in the washing buffer overnight at 4 8C. The pro-tein was collected in the flow-through of the column and was con-centrated with a Microcon centrifugal filter 3000 MWCO (Millipore).

For fluorescence measurements, ICP-MS analysis and electrophoret-ic mobility shift assays with truncated proteins, the eluent contain-ing the GST fusion protein was dialyzed against a storage buffer(Tris·HCl (20 mm), NaCl (100 mm), MgCl2 (5 mm), 1 mm BME,pH 7.5). The GST fusion proteins were directly used in these assays.The Kd of the complex formed with the GST fusion of MBNL1 withRNA is within error of that of MBNL1N (Figure S3).

For all proteins, the molecular weights were confirmed by MALDImass spectrometry, purity was determined by SDS-PAGE, and theconcentrations of protein samples were determined from Bradfordassays (Bio-Rad). ICP-MS analyses of the proteins were performedby Microanalytical Lab (University of Illinois at Urbana-Champaign).

Electrophoretic mobility shift assays: All RNA oligonucleotideswere purchased from IDT, and were purified by HPLC. RNA was la-beled with [g-32P]ATP using T4 polynucleotide kinase (Invitrogen).Radiolabeled RNAs were heated to 95 8C for 2 min and then placedon ice for 20 min. RNA samples were diluted to 0.2 nm with buffercontaining Tris·HCl (27 mm), NaCl (66 mm), and MgCl2 (6.7 mm),pH 7.5. Proteins were serially diluted in buffer containing Tris·HCl(20 mm), NaCl (175 mm), MgCl2 (5 mm), BME (1.25 mm), 12.5 %glycerol, BSA (2 mg mL�1), and heparin (0.1 mg mL�1), pH 7.5. Non-GST-tagged protein was used in all the assays except the assayswith truncated proteins, in which GST tag was not removed, inorder to improve the yield of proteins. Protein and RNA solutionswere mixed in a 1:1 ratio by volume and incubated at room tem-perature for 20 min. The samples were loaded on a 6 % polyacryl-amide gel (acrylamide/bisacrylamide, 37:1). The gel was run at

ChemBioChem 2012, 13, 112 – 119 � 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chembiochem.org 117

MBNL1–RNA Recognition

Page 7: MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation

360 V in 0.5 X Tris-borate buffer (pH 8.3) for 30 min and visualizedon a Molecular Dynamics Storm phosphorimager. The apparent Kd

values were obtained by fitting fraction RNA bound versus proteinconcentration using the following Equation (1):

fraction RNA bound ¼ 1=ð1þ K d=½protein�totalÞ ð1Þ

All measurements were performed with greater than tenfoldexcess of protein over RNA so that [protein] would be approxi-mately equal to [protein]total. For representative gel pictures, seeFigure S1.

Thermal denaturation studies: The melting curves (absorbanceversus temperature) of the RNAs were measured on a ShimadzuUV2450 spectrophotometer equipped with a temperature control-ler. The absorbance of each RNA sample (RNA (~3 mm) in buffer[Tris·HCl (27 mm), NaCl (66 mm), and MgCl2 (6.7 mm), pH 7.5]) wasmonitored at 260 nm from 10–90 8C at a ramp rate of 0.5 8C min�1.The melting temperature (Tm) of each sample was determinedfrom the maximum point of the first derivative of the meltingcurve determined using LabSolutions–Tm Analysis version 1.0. Freeenergy values were determined from the melting curve using Melt-win 3.5. (http://www.meltwin.com/).

Circular dichroism experiments: CD spectra were recorded on aJasco J-700 spectropolarimeter. Protein or RNA samples were dis-solved in buffer (Tris·HCl (27 mm), NaCl (66 mm), and MgCl2

(6.7 mm), pH 7.5, 250 mL). The final concentration of protein was12 mm and of RNA was 2.5 mm. Before taking measurements, RNAsamples were heated at 95 8C for 5 min and then cooled on ice for20 min. The CD spectra were monitored from 250 to 300 nm at arate of 200 nm min�1. Data points were acquired every 0.5 nm. Thespectra were not evaluated at wavelengths below 250 because ofthe strong CD spectrum of the protein in this region. The spectraobtained were averages of 50 accumulations.

Steady-state fluorescence emission measurements: Fluorescein-labeled RNA oligonucleotides were purchased from IDT, and werepurified by HPLC. All the fluorescence measurements were ob-tained using the MBNL1N-GST fusion, which was shown to havesimilar binding affinity as the non-fusion protein MBNL1N (Fig-ure S3). Steady-state fluorescence spectra were recorded on a Fluo-roMax-2 spectrometer equipped with a 150 W xenon lamp andmodified Czerny–Turner spectrometers in both the excitation andemission position utilizing Datamax spectroscopy software. The in-strument was interfaced with a Neslab RTE-111 temperature con-troller with a remote sensor. The excitation and emission slits were4 nm and the excitation wavelength was 485 nm. All fluorescencescans were recorded at 20 8C using 1 nm wavelength increments.Samples prepared for fluorescence measurements were dissolvedin a buffer containing Tris·HCl (20 mm, pH 7.5), NaCl (100 mm),MgCl2 (5 mm) and BME (1 mm).

Sliding window analysis of mRNA folding energy: The RNA fold-ing energy was calculated using the software RNAstructure 4.6,using the default setting: 37 8C, 1 mm RNA, and 29 nt windowlength.

Acknowledgements

We thank Dr. Madhavaiah Chandra for assistance with PCR andDr. Maurice S. Swanson for the MBNL1N construct. This work wassupported by the National Institutes of Health (RO1AR058361s).

Keywords: fluorescence quenching · helix destabilization ·myotonic dystrophy · protein–RNA complexes · RNA

[1] J. D. Amack, M. S. Mahadevan, Dev. Biol. 2004, 265, 294.[2] A. Mankodi, C. A. Thornton, Curr. Opin. Neurol. 2002, 15, 545.[3] L. P. Ranum, T. A. Cooper, Annu. Rev. Neurosci. 2006, 29, 259.[4] C. L. Liquori, K. Ricker, M. L. Moseley, J. F. Jacobsen, W. Kress, S. L. Naylor,

J. W. Day, L. P. Ranum, Science 2001, 293, 864.[5] M. S. Mahadevan, R. S. Yadava, Q. Yu, S. Balijepalli, C. D. Frenzel-McCar-

dell, T. D. Bourne, L. H. Phillips, Nat. Genet. 2006, 38, 1066.[6] M. Fardaei, K. Larkin, J. D. Brook, M. G. Hamshere, Nucleic Acids Res.

2001, 29, 2766.[7] M. Fardaei, M. T. Rogers, H. M. Thorpe, K. Larkin, M. G. Hamshere, P. S.

Harper, J. D. Brook, Hum. Mol. Genet. 2002, 11, 805.[8] A. Mankodi, X. Lin, B. C. Blaxall, M. S. Swanson, C. A. Thornton, Circ. Res.

2005, 97, 1152.[9] K. L. Taneja, M. McCurrach, M. Schalling, D. Housman, R. H. Singer, J. Cell

Biol. 1995, 128, 995.[10] H. Du, M. S. Cline, R. J. Osborne, D. L. Tuttle, T. A. Clark, J. P. Donohue,

M. P. Hall, L. Shiue, M. S. Swanson, C. A. Thornton, M. Ares, Jr. , Nat.Struct. Mol. Biol. 2010, 17, 187.

[11] E. S. Goers, J. Purcell, R. B. Voelker, D. P. Gates, J. A. Berglund, NucleicAcids Res. 2010, 38, 2467.

[12] S. Hino, S. Kondo, H. Sekiya, A. Saito, S. Kanemoto, T. Murakami, K. Chi-hara, Y. Aoki, M. Nakamori, M. P. Takahashi, K. Imaizumi, Hum. Mol.Genet. 2007, 16, 2834.

[13] H. Jiang, A. Mankodi, M. S. Swanson, R. T. Moxley, C. A. Thornton, Hum.Mol. Genet. 2004, 13, 3079.

[14] Y. Kino, C. Washizu, Y. Oma, H. Onishi, Y. Nezu, N. Sasagawa, N. Nukina,S. Ishiura, Nucleic Acids Res. 2009, 37, 6477.

[15] W. Dansithong, S. Paul, L. Comai, S. Reddy, J. Biol. Chem. 2005, 280,5773.

[16] T. H. Ho, B. N. Charlet, M. G. Poulos, G. Singh, M. S. Swanson, T. A.Cooper, EMBO J. 2004, 23, 3103.

[17] M. Pascual, M. Vicente, L. Monferrer, R. Artero, Differentiation 2006, 74,65.

[18] M. B. Warf, J. A. Berglund, RNA 2007, 13, 2238.[19] Y. Yuan, S. A. Compton, K. Sobczak, M. G. Stenberg, C. A. Thornton, J. D.

Griffith, M. S. Swanson, Nucleic Acids Res. 2007, 35, 5474.[20] R. N. Kanadia, C. R. Urbinati, V. J. Crusselle, D. Luo, Y. J. Lee, J. K. Harrison,

S. P. Oh, M. S. Swanson, Gene Expression Patterns 2003, 3, 459.[21] M. Hao, K. Akrami, K. Wei, C. De Diego, N. Che, J. H. Ku, J. Tidball, M. C.

Graves, P. B. Shieh, F. Chen, Dev. Dyn. 2008, 237, 403.[22] R. N. Kanadia, J. Shin, Y. Yuan, S. G. Beattie, T. M. Wheeler, C. A. Thornton,

M. S. Swanson, Proc. Natl. Acad. Sci. USA 2006, 103, 11748.[23] A. Matynia, C. H. Ng, W. Dansithong, A. Chiang, A. J. Silva, S. Reddy,

PLoS One 2010, 5, e9857.[24] M. Teplova, D. J. Patel, Nat. Struct. Mol. Biol. 2008, 15, 1343.[25] D. M. Cass, R. Hotchko, P. Barber, K. Jones, D. P. Gates, J. A. Berglund,

BMC Mol. Biol. 2011, 12.[26] B. H. Mooers, J. S. Logue, J. A. Berglund, Proc. Natl. Acad. Sci. USA 2005,

102, 16626.[27] A. Kiliszek, R. Kierzek, W. J. Krzyzosiak, W. Rypniewski, Nucleic Acids Res.

2009, 37, 4149.[28] M. B. Warf, J. V. Diegel, P. H. von Hippel, J. A. Berglund, Proc. Natl. Acad.

Sci. USA 2009, 106, 9203.[29] N. R. Markham, M. Zuker, Nucleic Acids Res. 2005, 33, W577.[30] N. R. Markham, M. Zuker, Methods Mol. Biol. 2008, 453, 3.[31] Y. Kino, D. Mori, Y. Oma, Y. Takeshita, N. Sasagawa, S. Ishiura, Hum. Mol.

Genet. 2004, 13, 495.[32] J. R. Lakowicz in Principles of Fluorescence Spectroscopy : Quenching of

Fluorescence, Springer, New York, 2006, p. 278.[33] P. Pinheiro, G. Scarlett, A. Rodgers, P. M. Rodger, A. Murray, T. Brown,

S. F. Newbury, J. A. McClellan, J. Biol. Chem. 2002, 277, 35183.[34] D. M. Gray, J. J. Liu, R. L. Ratliff, F. S. Allen, Biopolymers 1981, 20, 1337.[35] D. M. Gray, S.-H. Hung, K. H. Johnson, Methods Enzymol. 1995, 246, 19.[36] B. L. Mark, D. Gray, Biopolymers 1997, 42, 337.[37] R. Tan, A. D. Frankel, Biochemistry 1992, 31, 10288.

118 www.chembiochem.org � 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2012, 13, 112 – 119

A. M. Baranger et al.

Page 8: MBNL1–RNA Recognition: Contributions of MBNL1 Sequence and RNA Conformation

[38] T. J. Daly, J. R. Rusche, T. E. Maione, A. D. Frankel, Biochemistry 1990, 29,9791.

[39] Circular Dichroism : Principles and Applications, 2nd ed. (Eds. : N. Berova,K. Nakanishi, R. W. Woody), Wiley-VCH, New York, 2000.

[40] B. P. Hudson, M. A. Martinez-Yamout, H. J. Dyson, P. E. Wright, Nat.Struct. Mol. Biol. 2004, 11, 257.

[41] A. R. Davis, B. M. Znosko, Biochemistry 2007, 46, 13425.[42] D. H. Mathews, J. Savina, M. Zuker, D. H. Turner, J. Mol. Biol. 1999, 288,

911.[43] T. Xia, J. SantaLucia, M. E. Burkard, R. Kierzek, S. J. Schroeder, X. Jiao, C.

Cox, D. H. Turner, Biochemistry 1998, 37, 14719.[44] S. Sen, I. Talukdar, Y. Liu, J. Tam, S. Reddy, N. J. Webster, J. Biol. Chem.

2010, 285, 25426.[45] L. Rajkowitsch, D. Chen, S. Stampfl, K. Semrad, C. Waldsich, O. Mayer,

M. F. Jantsch, R. Konrat, U. Blasi, R. Schroeder, RNA Biol. 2007, 4, 118.[46] K. Semrad, Biochem. Res. Internat. 2011, 2011, 1.[47] G. Cosa, Y. Zeng, H. W. Liu, C. F. Landes, D. E. Makarove, K. Musier-For-

syth, P. F. Barbara, J. Phys. Chem. B 2006, 110, 2419.[48] J. Azoulay, J.-P. Clamme, J.-L. Darlix, B. P. Roques, Y. M�ly, J. Mol. Biol.

2003, 326, 691.

[49] M. K. Hong, E. J. Harbron, D. B. O’Connor, J. Guo, P. F. Barbara, J. G.Levin, K. Musier-Forsyth, J. Mol. Biol. 2003, 325, 1.

[50] G. Cosa, E. J. Harbron, Y. Zeng, H. W. Liu, D. B. O’Connor, C. Eta-Hosoka-wa, K. Musier-Forsyth, P. F. Barbara, Biophys. J. 2004, 87, 2759.

[51] P. C. Gareiss, K. Sobczak, B. R. McNaughton, P. B. Palde, C. A. Thornton,B. L. Miller, J. Am. Chem. Soc. 2008, 130, 16254.

[52] J. F. Arambula, S. R. Ramisetty, A. M. Baranger, S. C. Zimmerman, Proc.Natl. Acad. Sci. USA 2009, 106, 16068.

[53] M. D. Disney, M. M. Lee, A. Pushechnikov, J. L. Childs-Disney, ChemBio-Chem 2010, 11, 375.

[54] M. M. Lee, A. Pushechnikov, M. D. Disney, ACS Chem. Biol. 2009, 4, 345.[55] A. Pushechnikov, M. M. Lee, J. L. Childs-Disney, K. Sobczak, J. M. French,

C. A. Thornton, M. D. Disney, J. Am. Chem. Soc. 2009, 131, 9767.[56] M. B. Warf, M. Nakamori, C. M. Matthys, C. A. Thornton, J. A. Berglund,

Proc. Natl. Acad. Sci. USA 2009, 106, 18551.

Received: August 1, 2011

Published online on November 22, 2011

ChemBioChem 2012, 13, 112 – 119 � 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chembiochem.org 119

MBNL1–RNA Recognition