roth, ishimaru and hennig 2013

The Core Microprocessor Component DiGeorge SyndromeCritical Region 8 (DGCR8) Is a Nonspecific RNA-bindingProtein*Received for publication,December 20, 2012, and in revised form, July 9, 2013 Published, JBC Papers in Press, July 26, 2013, DOI 10.1074/jbc.M112.446880

Braden M. Roth, Daniella Ishimaru, and Mirko Hennig1

From the Department of Biochemistry andMolecular Biology, Medical University of South Carolina,Charleston, South Carolina 29425

Background: Double-stranded RNA-binding domain-containing proteins play key roles in microRNA biogenesis.Results: The Microprocessor protein DGCR8 binds RNA targets nonspecifically.Conclusion: DGCR8 alone is not responsible for specific RNA target recognition by the Microprocessor.Significance: Faithful RNA processing is not causatively coupled to specific substrate binding properties of DGCR8.

MicroRNA (miRNA) biogenesis follows a conserved succes-sion of processing steps, beginning with the recognition and lib-eration of an miRNA-containing precursor miRNA hairpinfrom a large primary miRNA transcript (pri-miRNA) by theMicroprocessor, which consists of the nuclear RNase III Droshaand the double-stranded RNA-binding domain protein DGCR8(DiGeorge syndrome critical region protein 8). Current modelssuggest that specific recognition is driven by DGCR8 detectionof single-stranded elements of the pri-miRNA stem-loop followedby Drosha recruitment and pri-miRNA cleavage. Because count-lessRNAtranscripts feature single-stranded-dsRNAjunctions andDGCR8 can bind hundreds of mRNAs, we explored correlationsbetween RNA binding properties of DGCR8 and specific pri-miRNA substrate processing.We found that DGCR8 bound sin-gle-stranded, double-stranded, and random hairpin transcriptswith similar affinity. Further investigation of DGCR8/pri-mir-16 interactions by NMR detected intermediate exchangeregimes over a wide range of stoichiometric ratios. Diffusionanalysis of DGCR8/pri-mir-16 interactions by pulsed field gra-dient NMR lent further support to dynamic complex formationinvolving free components in exchange with complexes ofvarying stoichiometry, although in vitro processing assaysshowed exclusive cleavage of pri-mir-16 variants bearing single-stranded flanking regions. Our results indicate that DGCR8binds RNA nonspecifically. Therefore, a sequential model ofDGCR8 recognition followed byDrosha recruitment is unlikely.Known RNA substrate requirements are broad and include70-nucleotide hairpins with unpaired flanking regions. Thus,specific RNA processing is likely facilitated by preformedDGCR8-Drosha heterodimers that can discriminate betweenauthentic substrates and other hairpins.

miRNAs2 are a conserved class of ?22-nucleotide (nt) non-coding RNAs that regulate gene expression by base pairingwithtarget mRNA sequences. They play key roles in modulatingcellular and developmental processes (1, 2) and have beenimplicated in a broad range of metabolic, immunological, psy-chiatric, and cell cycle disorders (3–6). Mature miRNAs aregenerated through a conserved succession of processing steps.Typically, miRNA biogenesis begins in the nucleus with thegeneration of a long pri-miRNA transcript by RNA polymeraseII. A hairpin harboring the pri-miRNA is selected for process-ing by the nuclear RNase III Drosha, which liberates themiRNA-containing pre-miRNA. Exportin-5 transports thepre-miRNA to the cytoplasm where it undergoes a secondround of processing (7, 8). Another RNase III, Dicer, trims theterminal loop to generate a 21–24-nt RNA duplex bearing 2-nt3? overhangs (9).Processing of pri- and pre-miRNAs bymembers of the RNase

III family is facilitated by accessory proteins comprising multi-ple dsRBDs featuring characteristic ?-?-?-?-? motifs. Inhumans, DGCR8 is required for Drosha processing and con-tains tandem dsRBDs (10). DGCR8 interacts with ferric hemeas a cofactor to increase processing activity (11–13), andalthough the cooperative formation of oligomeric DGCR8/pri-miRNA assemblies has been described as another positivemodulator of pri-miRNA processing (14), lysine acetylationwithin the tandem dsRBDs decreases the affinity of DGCR8 forpri-miRNA thereby down-regulating processing (15).Two modes of pri-miRNA recognition have been presented.

In one model, the Microprocessor binds a large (?10-nt) ter-minal loop to position Drosha’s catalytic center roughly twohelical turns (?22-nt) from the stem-loop junction, resulting inthe liberation of the ?70-nt pre-miRNA (16). A more recentmodel suggests that the terminal loop is dispensable forMicro-processor recognition. Instead, DGCR8 recognizes the hair-pin’s basal ss-dsRNA junction and recruits Drosha for cleavage

* This work was supported by National Science Foundation Grants MCB0845512 and DBI 1126230 (to M. H.).

1 To whom correspondence should be addressed: Dept. of Biochemistry andMolecular Biology, Medical University of South Carolina, 70 President St.,P. O. Box 250509, Charleston, SC 29425. Tel.: 843-876-2382; Fax: 843-792-1627; E-mail: [email protected].

2 The abbreviations used are: miRNA, microRNA; pre-miRNA, precursormiRNA; pri-miRNA, primarymiRNA; PFG, pulsed field gradient; dsRBD, dou-ble-stranded RNA-binding domain; ssRNA, single-stranded RNA; nt, nucle-otide; BPPLED, bipolar pulse pair longitudinal eddy current delay; SEC, sizeexclusion chromatography; PDB, Protein Data Bank.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 288, NO. 37, pp. 26785–26799, September 13, 2013© 2013 by The American Society for Biochemistry and Molecular Biology, Inc. Published in the U.S.A.

SEPTEMBER 13, 2013•VOLUME 288 •NUMBER 37 JOURNAL OF BIOLOGICAL CHEMISTRY 26785

(17). However, the generality of this model is unclear becausefurther experiments revealed that the length and symmetry of5?- and 3?-nonstructured ssRNAs required for efficient pro-cessing are variable and pri-miRNA-dependent (18). MostCaenorhabditis elegans pri-miRNAs apparently lack determi-nants required for processing in human cells, yet one-fifth of allhuman pri-miRNAs lack primary sequence determinants suchas downstream SRp20 binding, basal UG, and the apical GUGmotifs, recently described by Bartel and co-workers (19). Col-lectively, this leaves 70-nt hairpins with unpaired flankingregions as the only feature and common denominator acrossMicroprocessor substrates.Although most dsRBDs have been reported to bind duplex

RNA as the name implies, unpaired ssRNA in loops, bulges, andmismatched pairs can also be recognized (20–23). This flexibil-ity in substrate recognition presents a challenge in identifyingboth protein and RNA features required for specific protein-RNA interaction in the absence of structural data. A recentgenome-wide view of DGCR8 function employing RNA cross-linking immunoprecipitation sequencing studies revealed amuch broader RNA target base than anticipated (24). Specifi-cally, these results showed that DGCR8 targets hundreds ofmRNAs in addition to the anticipated interactions with pri-miRNA substrates. Implications for DGCR8 function rangedfrom roles in small nucleolar RNA and long noncoding RNAstabilization to regulation of alternatively spliced isoforms.In this study, we investigated the RNA binding properties of

the integral Microprocessor component DGCR8. Along withssRNA and perfectly paired dsRNA substrates, we designed acomprehensive series of pri-mir-16-1 variants to elucidate therole of RNA secondary structure elements in DGCR8-RNA ingeneral as well as the initiating event of pri- to pre-miRNAprocessing. These mutants were also used to investigate the linkbetween recognition and processing of pri-miRNA substrates.Because detailed structural information on DGCR8?pri-miRNAcomplexes remains elusive, we utilized NMR spectroscopy toelucidate the interaction at single residue resolution and diffu-sion measurements to probe complex formation in solution.AlthoughDGCR8 in isolation did not discriminate between thebroad array of RNAs tested, functional processing assays con-firm the specificity of Microprocessor assemblies. Thus, ourresults point to an intricate relationship between DGCR8 andDrosha in which both proteins are required for binding andprocessing, rather than a hierarchical DGCR8-mediated pri-miRNA recognition and subsequent Drosha recruitmentmodel.

EXPERIMENTAL PROCEDURES

RNA Transcription and Purification—Pri-miRNA-contain-ing plasmids were purified with the Omega Plasmid Giga kit,linearized with the appropriate 3? restriction endonuclease,phenol/chloroform-extracted, and ethanol-precipitated. Theminimal pri-mir-16-1 construct includes 18 nt upstream of the5?- and 16 nt downstream of the 3?-cleavage site. Although twoadditional Gs were added to the 5? overhang for efficient tran-scription, an additional AUUU sequence was added to the 3?overhang after linearization employing SwaI (New EnglandBiolabs). In vitro reactions were performed in transcription

buffer (40mMTris, pH 8.0, 1mM spermidine, 10mMdithiothre-itol (DTT), 0.01%TritonX-100) containing 0.05?g/ml plasmidtemplate, 30 mMMgCl2, 1 ?M T7 RNA polymerase and 6.5 mMeach ATP, CTP, GTP, and UTP. Transcription reactions wereincubated at 37 °C for 4 h.Transcripts were purified as described previously (25).

Briefly, the RNA was exchanged to buffer containing 10 mMNaPO4, pH 6.5, 25 mM KCl, 0.5 mM EDTA, and 25 ?M NaN3.Full-length transcripts were separated from the protein, unin-corporated nucleotides, and abortive transcripts by anionexchange chromatography (GE HiTrapQ HP).Protein Expression and Purification—N-terminal, tobacco

etch virus protease-cleavable His6-GB1/DGCR8?275, DGCR8PC,DGCR8core, and DGCR8RBD1 (residues 509–582) sequenceswere transformed into Escherichia coli BL21-CodonPlus(DE3)-RIPL competent cells, and expression was induced with0.5 mM isopropyl ?-D-thio-galactoside for 2–4 h at 37 °C(DGCR8?275 cultures were supplemented with 1 mM ?-amin-olevulinic acid) (13) and purified with a HisTrapFF affinity col-umn per the manufacturer’s instructions. Partially purifiedfusion proteins were cleaved with tobacco etch virus proteaseovernight at room temperature and reapplied to the HisTrapcolumn. The purified protein was then exchanged into bufferappropriate for NMR analysis or biochemical assays using aPD-10 desalting column. Cultures of E. coli BL21-CodonPlus(DE3)-RIPL cells expressing DGCR8core and DGCR8RBD1 (res-idues 509–582) were grown from a single colony in 2 ml ofLB/Amp100 at 37 °C overnight. These cultures were thenadapted to growth in 500 ml of 100% D2O-15N,13C-enrichedM9 minimal media (26) with a succession of small volume cul-tures grown at 37 °C with 200 rpm agitation. Expression wasinduced with 1 mM isopropyl ?-D-thio-galactoside for 4 h at37 °C and purified as described previously (27). Identicalbatches of DGCR8 protein variants were used in successivebinding and processing assays.Microfiltration Double-filter Binding Assays—Filter binding

was performed essentially as described in Arraiano et al. (28).Briefly, in vitro transcribed, purified RNAwas end-labeled with?-32P and then re-purified by urea-PAGE and gel extraction.The labeled RNA stocks were quantified and diluted to 30 ?M.[32P]RNAwas diluted to 200 pM (2? stock) withmicrofiltrationbuffer (20 mMTris, pH 7.5, 50 mMNaCl, 1 mMDTT). The RNAwas annealed by incubating at 65 °C for 5min and then ice for 3min. Purified protein for filter binding was prepared as 2?stocks by serial dilution in 1? microfiltration buffer. Equal vol-umes of 2? RNA and protein stocks were combined to yield50-?l binding reactions. The reactions were incubated at roomtemperature for 1 h.The binding reactions were filtered through nitrocellulose

and nylon membranes sequentially using the BioDot microfil-tration apparatus. Sample wells were washed with 2? 100 ?l ofmicrofiltration buffer before and after sample loading. Mem-branes were air-dried, and the intensity of the spots was quan-tified by a phosphorimager. The data were fit using a nonlinearleast sum of squares analysis with Equation 1,

FB ? FBmin ??FBmax?DGCR8?n?

?Kd ? ?DGCR8?n?(Eq. 1)

DGCR8 Binds RNA Nonspecifically

26786 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 288 •NUMBER 37 •SEPTEMBER 13, 2013

where FB is the fraction of RNA bound to protein; FBmin is theminimum fitted value of FB; FBmax is the maximum fitted valueof FB; [DGCR8] is the molar concentration of protein, and n isthe fitted value of the slope of the curve.Electrophoretic Mobility Shift Assay (EMSA)—Purified pri-

miRNA transcripts were ?-32P-end-labeled and purified asdescribed previously (29). A 20? pri-miRNA stock was heatedto 65 °C for 5 min and annealed on ice for 3 min. DGCR8 vari-ants were diluted to the appropriate concentration as 20?stocks with EMSA reaction buffer. EMSA samples were pre-pared in 20 ?l in reaction buffer containing 50 mM Tris-HCl,pH 7.5, 200 mM NaCl, 5% glycerol, 0.1% Triton X-100, and 0.2nM pri-miRNA. Reactions were incubated for 1 h at 37 °C andresolved on a 6% nondenaturing polyacrylamide gel in 1?TBE.RNA Secondary Structure Determination byNMR—AllNMR

spectra were recorded at 288 K or 298 K on Bruker Avance 800or 850 MHz spectrometer equipped with a TCI cryoprobe.NMR experiments were performed on samples of 500-?l vol-ume containing 0.1–1.6 mM of pri-mir-16 RNA variants. Datawere processed using nmrPipe (30) and analyzed using Sparky(31). One-dimensional imino proton spectra were acquiredusing a jump-return echo sequence (32). Imino proton reso-nances were assigned sequence-specifically from water flip-back two-dimensional NOESY spectra (?mix ? 200 ms) (33).Assignments were confirmed using a jump-return (32) 15N-HMQC (34). Elucidation of base pairing and secondary struc-ture was verified from the 2JNN coupling through hydrogenbonds in the TROSY-based HNN-COSY data (35, 36).The pri-mir-16lower RNA tertiary structure was predicted

using the MC-sym software (37). All information required togenerate and edit the scripts are found on line on theMC-pipe-line website. Briefly, all-atommodels for the 65-nucleotide pri-mir-16lower satisfying the NMR-derived secondary structurewere generated. 132 structures were clustered using MC-Sym,and five representative structures were selected. AMBERenergy minimization was employed to refine the final fiveatomistic three-dimensional structures as implemented inMC-sym.DGCR8 NMR Analysis—Backbone 1HN, 15N, 13CO, 13C?,

and 13C? chemical shift assignments for DGCR8core have beenobtained using a strategy based on three-dimensional triple res-onance experiments employing transverse relaxation-opti-mized spectroscopy (TROSY), and chemical shifts have beendeposited in the BioMagResBank under accession number17773 (27). Standard three-dimensional NMR techniques (38)were employed to obtain and confirm backbone assignmentsfor DGCR8RBD1 (residues 509–582) (39). Line width and cross-peak intensities of 1H,15N-TROSY experiments at varying pro-tein/RNA ratios were determined using Sparky, assumingGaussian line shapes (31).DGCR8core/pri-mir-16lower Titration—2H,15N-DGCR8core

and pri-mir-16lower were buffer-exchanged to NMR buffer (20mM NaPO4, pH 7.0, 150 mM KCl, 5 mM DTT) with the PD-10desalting column. Titrations were performed by the addition ofdiluted RNA to protein in NMR buffer. Samples were concen-tratedwith theAmicon concentrator (10KMWCO) in 500?l of90% H2O, 10% 2H2O. RNA was added to 250 ?M DGCR8core toreach protein/RNAmolar ratios of 48:1, 24:1, 12:1, 6:1, 3:1, 2:1,

1:1, and 1:2. Alternatively, diluted DGCR8core was added to 250?M pri-mir-16lower to reach RNA/protein molar ratios of 48:1,24:1, 12:1, 6:1, 3:1, 2:1, and 1:1, respectively. Titrations wereperformed in NMR buffer in the absence and presence of 5 mMMgCl2.Measurement of DGCR8-pri-miRNA Diffusion Constants by

PFG NMR—All PFG-NMR experiments were recorded at 298K on a Bruker Avance III spectrometer operating at 600 MHzand equipped with a 5-mm QXI quadruple resonance (1H,19F-13C,31P) room temperature probe with triple-axis shielded gra-dients. The bipolar pulse pair longitudinal eddy current delay(BPPLED) sequence (40, 41) was used, and one-dimensionalspectra were recorded after two-axis gradient encoding anddecoding separated by a diffusion delay of 100ms. Dephasing ofthe NMR signal in the BPPLED experiment was analyzedaccording to Equation 2 (40, 42),

A? g? ? A? g0?exp??D??H?Gmax?2?g2 ? g0

2??? ? ?/3 ? ? /2??

(Eq. 2)

where A(g) is the measured intensity of the NMR signal as afunction of the fractional gradient strength g; A(g0) is the peakintensity of the reference spectrum (g0 ? 2%); D is the transla-tional diffusion constant; ?H is the gyromagnetic ratio of a pro-ton (2.675197 ? 104 G?1?s?1); ? is the gradient duration; Gmaxis the combined maximum power (87.8 G?cm?1 at g ? 100%)offered by the x and z gradients;? is the time between gradients,and ? the recovery time. By taking the natural logarithms ofboth sides of Equation 2, a linear equation results, andD can becalculated from least squares linear fits. To exclude complica-tions with any exchangeable protons, the diffusion analysis ofdeuterated protein samples was performed by integrating overthe 0.35–1.05 ppm section of themethyl region of the spectrum(primarily CH2D and CHD2 isotopomers). RNA samples wereanalyzed by integrating over the 7.3–8.0 ppm section.The diffusion rate of a molecule depends on its size and

shape, its concentration, the temperature, and solvent viscosity.The Stokes-Einstein equation shows that a sphere’s diffusioncoefficient (D) is inversely related to the hydrodynamic radius(RH) and solvent viscosity (?) as shown in Equation 3,

D ? kBT/?6??RH? (Eq. 3)

where kB is the Boltzmann constant, and T is the temperature(298 K).Obtaining the hydrodynamic radius RH for a spherical parti-

cle using Equation 3 requires an accuratemeasure of the solventviscosity ?. All samples were dissolved in identical aqueous (90:10% H2O/D2O) buffer solutions containing 150 mM KCl. Usingviscosities of 1.097 and 0.8929 kg?cm?1?s?1 for a 100%D2O anda 100%H2O solution (43), the viscosity of a mixture of light andheavy water can be represented by a linear function of concen-tration (44) shown in Equation 4,

?0 ? x?H2O)?(H2O) ? (1 ? x)(D2O)?(D2O) (Eq. 4)

which yields 0.91331 kg cm?1 s?1 at 298 K. A salt effect correc-tion on viscosity was performed according to the Jones-DoleEquation 5 (45),



? ? ?0?1 ? A?c ? B?c?? (Eq. 5)

where c is the molar salt concentration; ?0 is the uncorrectedand? ? 0.9018 kg cm?1 s?1 the corrected solvent viscositywithA ? 0.0057 and B ? ?0.0147 for KCl (46).Fractional populations of monomer and dimer states can be

estimated using Equation 6,

Dobs ? (1 ? fD)DM ? fDDD (Eq. 6)

where DM and DD are the respective diffusion coefficients formonomer and dimer and fD is the molar fraction of the dimericspecies.Size Exclusion Chromatography (SEC)—SEC characteriza-

tion of 150 ?M DGCR8core was performed in NMR buffer (20mM NaPO4, pH 7.0, 150 mM KCl, 5 mM tris(2-carboxyethyl)phosphine) at 12 °C with a Superdex75 10/300 column at a flowrate of 0.5 ml/min. Conalbumin, ovalbumin, carbonic anhy-drase, GB1, and aprotinin were used to generate the standardcurve from which the apparent DGCR8core molecular weightwas estimated.pri-miRNA Processing Employing an Immunoprecipitated

Microprocessor—In vitro processing assays were performedessentially as described by Lee and Kim (29) with the followingmodifications: c-Myc-Drosha and FLAG/HA-DGCR8 (47)were co-transfected (Lipofectamine, Invitrogen) in the mam-malian cell line AD-293 (Stratagene), and Drosha/DGCR8 wasco-immunoprecipitated with either EZview anti-c-Myc affinitygel (Sigma) or agarose-conjugated anti-HA monoclonal anti-body (Sigma). Immunoprecipitated Microprocessor was incu-bated with 32P-pri-miRNA, in the presence of Mg2? andRNaseOUT (Invitrogen) for 90 min at 37 °C. RNA was purifiedby phenol/chloroform and resolved on a 12% denaturingPAGE. Gels were dried and analyzed by phosphorimaging.For Microprocessor reconstitution assays, the Drosha

immunoprecipitate was washed with a basic, high salt buffer toremove co-purified proteins. Stringent washes were performedwith 20 mM Tris, pH 8.0, 2.5 M NaCl, 5 mM EDTA, 1% TritonX-100, 0.1% SDS, 0.1% deoxycholic acid, and 1 M urea. A finalwash with 20 mM Tris, pH 8.0, 100 mM NaCl was performed.For the processing assays, c-Myc-Drosha was preincubatedwith recombinant DGCR8?275 in the presence of Mg2? andRNaseOUT (Invitrogen) for 10min at room temperature. Afteraddition of [32P]RNA, samples were incubated at 37 °C for 90min. RNA was purified, and samples were processed asdescribed above.DGCR8 and Drosha Co-immunoprecipitation—Whole

cell extracts of AD293 mock-transfected or transfected withc-Myc-Drosha/FLAG-DGCR8 were obtained as describedabove. Supernatant from lysed cells was incubated with eitherRNase inhibitor or RNase A/T1 mix (Ambion). After incuba-tion for 60 min at 4 °C, EZ view c-Myc beads (Sigma) wereadded and incubation proceeded for another 16 h at 4 °C. Beadswere washed with lysis buffer and resuspended in SDS-loadingbuffer. Proteins were subjected to SDS-PAGE, and Westernblot analysis was performed with anti-c-Myc or anti-FLAGantibodies (Sigma).

RESULTS

pri-miRNA Recognition by DGCR8—To investigate the sec-ondary structure elements required for pri-miRNA recognitionby DGCR8, we performed quantitative binding studies using aseries of pri-mir-16-1 variants. Along with the 104-nt hairpinthat includes single-stranded regions at both the base and ter-minal loop, we generated a terminal loop mutant (pri-mir-16?loop) by replacing 14 nucleotides of the loop with a G:C-C:G clamp and UUCG tetraloop. We also created a pri-mir-16mutant lacking single-stranded 5? and 3? overhangs (pri-mir-16?ss) and a pri-mir-16?? mutant that incorporates bothmodifications (Fig. 1A). Each pri-miRNA secondary structurewas confirmed byNMR spectroscopymonitoring imino protonresonances (Fig. 2).To complement the pri-mir-16 constructs, we produced

three recombinant DGCR8 variants to be tested in our bindingstudies, DGCR8?275, DGCR8PC, and DGCR8core (Fig. 1B).DGCR8?275 lacks the first 275 amino acids but retainsWWandheme-binding domains, two dsRBDs and the Drosha interac-tion/trimerization domain and is competent for both dimeriza-tion and pri-miRNAprocessing (13, 14, 48). DGCR8PC includesresidues 484–750 and represents the processing-competent(PC) core (48). DGCR8core (residues 493–706) lacks the Droshainteraction domain but retains the tandem dsRBDs and istherefore unable to participate in pri-miRNA processing but isRNA binding-competent (14, 49).DGCR8?275/pri-miRNA interactions were monitored by fil-

ter binding assays. As shown in Fig. 3B, DGCR8?275 boundpri-mir-16-1 with low nanomolar affinity (3.7 ? 0.9 nM), inqualitative agreement with previously published filter bindingexperiments (14). Surprisingly, DGCR8?275 bound pri-mir-16?loop, ?ss, and ?? constructs with similar affinities (Fig. 3,C–E, summarized in Table 1) and Hill coefficients, suggestingthat neither the ss-flanking regions nor the terminal loop aredeterminants for specific recognition by DGCR8.To test whether pri-mir-16 mutants contain enough ssRNA

in internal loops or bulges to attract DGCR8, we have modifiedthe pri-mir-16?? construct to produce a fully paired hairpin(pri-mir-16fp) by altering unpaired and bulged nucleotides toeffectively eliminate ssRNA andmaximize duplex stability (Fig.1A). DGCR8?275 was still able to bind pri-mir-16fpwith affinityand cooperativity equivalent to that of pri-mir-16-1,?loop,?ss,and ?? (Fig. 3F). To rule out the possibility that DGCR8?275

recognition of pri-mir-16-1 is driven by a sequence determi-nant common to all of our RNA substrates, we tested pri-mir-16x (Fig. 1A), a Quikfold (50)-predicted 82-nt hairpin co-tran-scribed with mir-16-1 and mir-15a on the DLEU2 lincRNA(long intergenic noncoding RNA) (51, 52). Located 397 ntupstream of mir-16-1, pri-mir-16x is overlooked by the Micro-processor. As shown in Fig. 3G, filter binding analysis revealedthat DGCR8?275 binds pri-mir-16x with no discernable loss ofaffinity or cooperativity relative to the other RNAs tested. Totest whether the lack of significant variation in our results couldbe explained by DGCR8?275 recognition of predominantlydsRNA elements, we produced an 87-nt ssRNA principallycomposed of nine tandemCCCCUAAArepeats (Fig. 1A). NMRanalysis of exchangeable imino protons confirmed that no sta-



ble secondary structure was formed in this transcript (data notshown). Finally, we also generated a pri-let-7b transcript toserve as a second bona fide pri-miRNA substrate. Again,DGCR8?275 bound both constructswithKd andHill values con-sistent with other RNA substrates (Fig. 3, H and I).Similarly, binding affinities and cooperativity measurements

remained consistentwhenwe analyzedDGCR8core and a subsetof the pri-miRNA variants, even though the affinities werereduced ?10-fold for DGCR8core, compared with DGCR8?275

(Table 1). Collectively, our filter binding experiments demon-strate that DGCR8 recognizes the tested RNA variants withhigh affinity and Hill coefficients ranging from 0.9 to 1.8. Noneof the DGCR8 constructs tested in filter binding demonstrateda pronounced tendency to bind cooperatively beyond a dimermodel.To directly detect oligomeric DGCR8-RNA complex forma-

tion, we performed electrophoretic mobility shift assays(EMSA) with DGCR8?275, DGCR8PC, or DGCR8core and the

FIGURE 1. Constructs used in DGCR8/pri-miRNA experiments. A, Mfold (73)-predicted secondary structures of in vitro transcripts. Dark boxes representsequencemutations; arrows indicate Drosha cleavage sites, and open boxes highlightmaturemiRNA sequences. The ssRNA construct is an unstructured 87-nttranscript containing nine repeats of the bracketed sequence. pri-mir-16x is an Mfold-predicted 82-nt hairpin located 397 nt upstream of pri-mir-16-1. B,full-length and truncated DGCR8 constructs are shown in the context of functionally significant domains.WW refers to the proline-bindingWWdomain; C352is a cysteine required for heme coordination; RBDs are RNA-binding motifs, and DI refers to the Drosha-interaction/trimerization domain.



pri-mir-16 transcript. As seen in Fig. 4A, progressively moretruncated variants of DGCR8 all populate similar heteroge-neous complexes with pri-mir-16-1. Additionally, DGCR8/pri-miRNA interactions reveal a “smearing” effect rather then dis-crete band formation, which could indicate that DGCR8 bindspri-miRNA transiently and with variable stoichiometry, sam-pling different regions of the substrate until excess proteineffectively saturates all potential binding sites on the RNA atany given time. EMSA titrations of DGCR8core into radiola-beled pri-mir-16-1 in the presence of varying amounts of non-specific competitor yeast tRNAPhe revealed that increasingamounts of tRNA increase the apparent Kd values for theDGCR8core/pri-mir-16-1 complex (Fig. 4B). Thus, on the basisof our EMSA results, DGCR8’s mode of binding resemblesother dsRBDs (such as protein kinase R (PKR) (53)) that inter-act with RNA through transient and/or noncooperative inter-actions featuring multiple, potentially overlapping RNA-bind-ing sites with comparable affinities. Taken together, ourbinding data imply that DGCR8 alone is not able to distinguishmiRNA-containing hairpins fromother RNA species but ratherretains high, indiscriminate affinity for RNAwithin its dsRBDs.DGCR8/pri-miRNA Binding Monitored by NMR—In the

interest of reducing overall molecular weight for NMR studies,we used a 65-nt version of pri-mir-16-1 (pri-mir-16lower, Fig.1A) that is correctly processed in vitro (data not shown) (17).UsingNMRspectroscopy,we have determined that the second-ary structure of pri-mir-16lower includes 5?- and 3?-ssRNAoverhangs, a 21-nt stem featuring tandemG-Amismatches anda C-A mismatch, and a stable 8-nt loop (Fig. 5).Our recent NMR analysis confirmed that our purified

DGCR8core contains all of the secondary structure features ofthe reported crystal structure (Fig. 7A) (27, 49). Furthermore,we observed pronounced chemical shift differences between

the backbone amide 1H,15N-correlations of an isolated RBD1domain and those of the RBD1-RBD2-containing DGCR8coreconstruct (Fig. 6). This demonstrates that the two tandem RBDdomains interact with each other in solution confirming earlierreports based on the crystal structure (49) andMD simulations(54).To characterize the interactions between DGCR8 and pri-

miRNA at single residue resolution, an NMR titration studywas performed in which 1H,15N-TROSY spectra of 2H,15N-la-beled DGCR8core were monitored upon addition of increasing

FIGURE 2. NMR-based secondary structure confirmation of pri-mir-16-1variants. 800 MHz imino one-dimensional jump-return echo experimentswere collected at 298 K. A, pri-mir-16?? RNA. The spectrum is labeled withassignment information; numbering is consistent with full-length pri-mir-16-1 (Fig. 1A). The imino resonances labeled ?, ?, and ? are located in thenon-native G?:C-C:G? clamp and 5?-UUCG?-3? tetraloop capping pri-mir-16?? and pri-mir-16?loop, respectively, and connected using dashed drop-lines. B, pri-mir-16?ss. C, pri-mir-16?loop. D, pri-mir-16-1 RNA.

FIGURE 3. Filter binding measurement of DGCR8 binding affinity. A, rep-resentative rawdata fromdot-blot double filter binding experiments. Proteinconcentration increases from 2 pM to 8 ?M spanning two bracketed rows. B–I,quantified results of recombinant DGCR8?275/RNA binding, represented asthe fraction of bound RNA. All experiments were performed with 100 pM32P-5?-end-labeledRNA. Theaxesof all graphs are identical, and thevalues areshownalong the left and bottomof the figure. All curves havebeen calculatedfrom at least four independent experiments.



amounts of RNA. pri-mir-16lower was titrated into 250 ?M2H,15N-DGCR8core to produce protein/RNA solutions rangingfrom 48:1 to 1:2 molar ratios. We observed differential broad-ening of backbone amide 1H,15N correlations evident even atlow concentrations of pri-mir-16lower (Fig. 7). Because bindingand dissociation between DGCR8core and pri-mir-16loweroccurred on the intermediate time scale, 1H,15N correlationexperiments do not reveal discrete resonances for RNA-bound and free states commonly observed for systems in fastor slow exchange. Thus, our protein-RNA interface analysisfocused on differential broadening at a molar ratio wheremost protein resonances were still observed (24:1 ratio ofDGCR8core/pri-mir-16lower).Differences in amide proton LW (?LW) upon RNA binding

were calculated by subtracting the LWbound of a given residue inthe presence of sub-stoichiometric RNA from the LWfree

observed in the absence of RNA (Fig. 8A). A value of zero indi-cates no change in LW. The ?LW values for infinitely broad-ened resonances (Fig. 8A, gray bars) were calculated assuming aconservative LWbound of 50 Hz, which corresponds to the larg-est 1H-TROSY LWbound observable at 800 MHz. The residueswith the largest change in line width (most negative ?LW) val-ues map to the first and second helices of the each dsRBDmotif(RBD1:?1 and ?2, RBD2:?5 and ?6) (Figs. 7 and 8B). This pat-

tern of RNA-dependent line broadening correlates with previ-ously characterized canonical RBD interactions in whichapproximately one helical turn of a successive minor andmajorgroove dsRNA is occupied by the ?-helical regions of the ?-?-?-?-? dsRNA-binding domain (20, 55). However, no signifi-cant broadening is detected for either the ?2-?3 loop located inRBD1 or the ?5-?6 loop in RBD2, which one would predict tocontact an adjacentminor groove in a canonical dsRBD-dsRNAinteraction (Figs. 7 and 8B). Instead, noticeable broadeningoccurs within the ?4, ?7, and ?7 regions. For its part, ?7 hasbeen shown to stabilize theDGCR8core through extensive intra-molecular contacts with both dsRBDs (49, 54). Therefore, linebroadening in this region is not likely attributed to direct asso-ciation with pri-mir-16lower but is probably reporting on con-formational exchange on an intermediate time scale betweenbound and free DGCR8core forms. Ultimately, line broadeninganalysis of our DGCR8core/pri-mir-16lower titration suggestssome variation in archetypal dsRNA binding for both RBD sub-units of the DGCR8core where three structural elements of eachdsRBD, the N-terminal ?-helix, the loop between ?1 and ?2,and the C-terminal second helix, engage adjacent minorgrooves and the intervening groove of dsRNA (56).Finally, we tested whether protein resonances, which are

broadened beyond detection at substoichiometric RNA ratios,

TABLE 1Summary of filter binding experimentsCalculations were based on at least four independent experiments, and curves were fit using a nonlinear least sum of squares analysis with variable slope (see Equation 1 andsee under “Experimental Procedures”).

16-1 16?loop 16?ss 16?? 16fp 16x ssRNA let-7bDGCR8 Kd ? 3.7 ? 0.9 nM 2.7 ? 0.8 nM 6.2 ? 2.3 nM 7.1 ? 1.9 nM 2.7 ? 0.7 nM 1.3 ? 0.6 nM 4.6 ? 1.0 nM 4.2 ? 1.1 nM(?275) n ? 1.2 ? .05 1.1 ? .09 0.9 ?.07 0.9 ? .05 1.3 ? .01 1.3 ? .09 0.9 ? .05 1.0 ? .06

R2 ? 0.999 0.994 0.996 0.995 0.999 0.996 0.997 0.999DGCR8 Kd ? 73 ? 15 nM 87 ? 19 nM 77 ? 23 nM 74 ? 23 nM 115 ? 23 nM(core) n ? 1.5 ? .12 1.7 ? .15 1.5 ? .16 1.6 ? .18 1.1 ? .08

R2 ? 0.999 0.997 0.997 0.997 0.996

FIGURE 4. EMSA visualization of DGCR8-pri-mir-16-1 complex stoichiometries and apparent affinities. A, representative gel shifts of pri-mir-16-1 withDGCR8variants exhibit similar formationof higher order complexes.B, EMSA titrationofDGCR8core intopri-mir-16-1 RNA,with thebands corresponding to freepri-mir-16-1, multimeric DGCR8core?pri-mir-16-1 complexes, and DGCR8core concentration through the titration denoted. EMSA reaction buffers contained 40,20, and 4 ?g ml?1 of nonspecific competitor yeast tRNAPhe (Sigma), respectively.



would sharpen as a result of changes in exchange regimebrought about by the formation of specific, stable interactionsin the presence of excess pri-mirR16lower. Over the course ofthe titration and at the end point ratio 1:2 DGCR8core/pri-mir-16lower, the vast majority of assigned backbone amide reso-nanceswas broadened beyond detection. The remaining 30 res-idues experienced negligible chemical shift perturbations andmap primarily to the flexible linker between RBD1 and RBD2(residues Asp-579 to Glu-594), the N terminus, includingstrand ?1 (residues Gln-495 to Ile-505), and the C terminus ofthe protein (residues Ser-702 toVal-706) (Figs. 7 and 8C). Over-all, our findings indicate that DGCR8core exchanges betweenand among pri-mir-16lower-binding sites on the intermediatechemical shift time scale leading to differential line broadening.Such a model was further corroborated by a titration reveal-

ing global imino proton line broadening of pri-mir-16lower inthe presence of increasing amounts of DGCR8core (Fig. 9A).Divalent cations often stabilize important RNA tertiary struc-tures. Therefore, titrations were repeated in the presence of 5mM MgCl2. Imino proton chemical shift changes observed forpri-mir-16lower and larger RNA variants (Fig. 1A) in the pres-ence of Mg2?-cations were invariably small and did not exceed0.04 ppm (data not shown). Moreover, intermediate exchangekinetics prevailed in the presence of Mg2? cations (Fig. 9B).The NMR titration data also suggest that the structure ofDGCR8core may not be significantly perturbed by the pri-mir-16lower interaction because chemical shifts of resonancesobservable during the titration are mostly unchanged.DGCR8/pri-miRNA Binding Monitored by PFG-NMR—A

distinct advantage of PFG-NMR methods is that translationalself-diffusion coefficients can be determined under experimen-

tal conditions (protein and RNA concentration, temperature,and viscosity) identical to those used for collecting the NMRtitration data described above. Because diffusion measure-ments are sensitive to changes in molecular size and shape,these experiments provide a supplementary characterization ofDGCR8/pri-mir-16-1 association to complement our linebroadening analysis. The results of PFG-NMR experiments arereported in Table 2. A single (average) diffusion coefficient wasobserved in all experimental measurements of free DGCR8protein variants or pri-mir-16lower, as evident from linear cor-relations with R values ?0.996 (Fig. 10A).Possible explanations for a slower diffusion rate (8.7 ?

0.01 ? 10?11m2?s?1) of the 214-residue DGCR8core (largerapparent RH) in comparison with calculated values (9.8 ?10?11m2?s?1) include differences between the solution struc-ture and the crystal structure used for the prediction, intermo-lecular oligomerization, and/or the presence of unfolded pro-tein regions. In the case of intermolecular oligomerization, onewould expect the slower diffusion to be caused by partialdimerization. Assuming exchange averaging between a mono-mer and a dimer on the diffusion time scale, we calculated a 47%dimer population of the DGCR8core. In good agreement withresults previously published (49), our SEC analysis confirmedan estimated molecular mass of 31 kDa for the DGCR8core,significantly larger than the theoretical molecular mass of 24.2kDa (data not shown). The overall agreement of experimentaland calculated translational diffusion coefficients and hydrody-namic radii for the DGCR8 variants and pri-mir-16lower in iso-lation confirms that physically meaningful data could beobtained for macromolecular DGCR8?pri-mir-16 complexes.

FIGURE5.NMR iminoassignments and secondary structuredeterminationofpri-mir-16lower.A, sequential imino-iminoprotonNOEassignmentpaths areshown by different colors for A-form helical stems I (black), II (magenta), and III (cyan). The spectrum is labeled with assignment information. B, NMR-basedsecondary structure of the lower stem of pri-mir-16lower. The secondary structure representation was generated using PseudoViewer (74); pri-mir-16lower

numbering is consistent with full-length pri-mir-16-1 (Fig. 1).



For the DGCR8core?pri-mir-16lower complex, diffusion ratesof the protein in the presence of excess RNA (1:2 molar ratio)revealed a deviation from linearity (R ? 0.990) (Fig. 10B), sug-gesting a polydisperse system described by Equation 2. Linearcorrelations obtained after subtracting the contributions fromthe slowly diffusing complex (i.e. elimination of last five points)or the faster diffusing component (i.e. elimination of first fivepoints) yielded upper and lower boundaries of D ? 10.1 ?1.02 ? 10?11 m2?s?1 with (R ? 0.999) and D ? 6.4 ? 0.15 ?10?11m2?s?1 (R ? 0.999). Extrapolated to infinite dilution,Stokes radii of RH ? 24.1 ? 2.6 Å and 38.1 ? 0.9 Å can thus beestimated for the smallest and largest diffusing spherical parti-cles for the DGCR8core?pri-mir-16lower complex exchanging onthe diffusion time scale. The smallest diffusing molecule islikely unbound DGCR8core given the close agreement with thepredicted RHcalc value of 24.7 Å.Recombinant DGCR8?275 Is an Active Microprocessor

Component—We performed in vitro Microprocessor assaysthat mimic the first step in miRNA biogenesis to examine therelationship between observed DGCR8/pri-miRNA bindingpatterns and processing. When incubated with immunopre-

cipitated Microprocessor, the pri-mir-16-1 and 16?loops wereefficiently processed to release respective 65- and 59-nt pre-miRNAs aswell as the 20-nt 5? and 3? reaction by-products (Fig.11A). However, pri-mir-16 variants lacking ss-flanking RNA(16?ss and 16??) were not cleaved by the Microprocessor, inagreement with previously published work (14, 17, 18). In addi-tion, no cleavage products were observed when pri-mir-16xwas used in the processing assay. These data demonstrate thatonly a distinct subgroup, including pri-mir-16-1 and 16?loop,is able to productively engage theMicroprocessor, even thoughall RNA constructs display similar binding by DGCR8.To address whether DGCR8’s RNA specificity determinant could

reside inthetruncatedportionoftheDGCR8?275constructs,wehaveperformed in vitro processing with a DGCR8?275-reconstitutedMicroprocessor. c-Myc-Drosha alone did not efficiently processpri-mir-16-1, 16?ss, or 16?loop (Fig. 11B). However, when incu-bated with purified recombinant DGCR8?275, 5? cleavage prod-ucts were observed for pri-mir-16-1 and 16?loop but not 16?ss,confirming that recombinant DGCR8?275 is a properly folded,active component of a highly specific in vitro pri-miRNAprocess-ing complex.

FIGURE 6.Chemical shift perturbation analysis of DGCR8RBD1 andDGCR8core constructs.A, color-coded representation of combined backbone amide 1HN

and 15N chemical shift differences of DGCR8RBD1 (residues 509–582) and DGCR8core (residues 493–706). Backbone ribbon color of the DGCR8 crystal structure(PDB code 2YT4) varies between white (?? ? 0 ppm) and red (?? ?0.2 ppm). Residues Val-581 and Lys-582 (?? ? 0.6 ppm) at the very C-terminal end of theRBD1 construct have beenomitted for clarity. Side chains of residueswith?? ?0.2 ppmare shown. The location of a putative salt bridge involvingAsp-549 andLys-659 (italic) inferred from the distances observed in the x-ray structure is indicated using arrows. B, backbone amide 1HN and 15N chemical shift differencesof DGCR8RBD1 (residues 509–582) in comparison with DGCR8core (residues 493–706) as a function of residue number. Secondary structure elements arenumbered and highlighted using gray boxes.



Taken together, our binding and processing data stronglysuggest that DGCR8 alone is not the specificity determinant forrecognition of bona fide miRNA-containing hairpins. Instead,DGCR8 works in concert with Drosha to promote specific rec-ognition and processing of pri-miRNA transcripts.DGCR8-Drosha Co-immunoprecipitation—In the assump-

tion that the DGCR8-Drosha complex provides the specificityfor pri-miRNA recognition, it is necessary that DGCR8 andDrosha can form heterodimers prior to pri-miRNA binding. Toprobe for the prerequisite heterodimer, a co-immunoprecipita-tion assay was performed in the absence or presence of RNases.Western blot analysis of Drosha-immunoprecipitated com-plexes revealed the presence of DGCR8 in extracts treated witheither RNase inhibitor or RNase A/T1 mix (data not shown).This observation confirms earlier reports by Han et al. (10) andsuggests that association of Drosha and DGCR8 does notdepend on the availability of RNA substrate molecules.

DISCUSSION

Accessory dsRNA-binding proteins involved in small RNApathways have been shown to be required for efficient andaccuratemodulation of specific RNA species, when part ofmul-

tiprotein complexes (10, 57, 58). Despite the importance andcharacterized functions of endonuclease-associated dsRBD-containing proteins in miRNA biogenesis, few reports havedescribed the detailed mechanistic and structural nature oftheir interactions with RNase IIIs or specific RNA substrates(59).Previous results from filter binding experiments form the

basis for a recognitionmodel in which highly cooperative bind-ing by DGCR8 is the distinguishing characteristic of specificpri-miRNA recognition (14). Although our Kd values are inagreement with those previously reported for filter binding (13,49), we have been unable to demonstrate conclusive deviationsin binding cooperativity among various RNA substrates. Ourdirect, EMSA-based observation of higher order complex for-mation involving DGCR8core, which lacks a previously identi-fied trimerization domain, also questions a generalized modelof pri-miRNA recognition involving stepwise, cooperativebinding mediated by the DGCR8 trimerization domain (14).We employed concentrations of 100 pM of end-labeled RNA

and a low salt buffer (50 mM NaCl) in filter binding assays andobserved tighter affinity and lower Hill coefficients for theheme-bound DGCR8?275/pri-mir-16-1 interaction when com-

FIGURE 7. DGCR8core?pri-mir-16lower complex formation monitored by NMR. A, traces taken through the TROSY cross-peaks of Arg-522 (located in ?1),Glu-540 (?2-?3 loop), Lys-590 (?2-?3 interdomain loop)/Met-705 (C terminus), His-625 (?6), and Gly-646 (?5-?6 loop) during the titration at various 2H,15N-DGCR8core:pri-mir-16lower stoichiometric ratios ranging from 48:1 (black) to 1:2 (red). B, 800MHz 1H,15N TROSY spectra of 2H,15N-DGCR8core recorded at 25 °C inthe absence of pri-mir-16lower (cyan contours) and at protein:RNA stoichiometric ratios of 48:1 (single black contour) and 1:2 (red contours). Dashed-horizontallines indicate traces shown in Awith cross-peak positions circled and assignments given.



pared with previously published studies by Faller et al. (13, 14)using unspecified concentrations of body-labeled pri-mir21and pri-mir30a in complex with DGCR8?275 in a reactionbuffer containing 85 mM NaCl. We observed reduced affinityfor the DGCR8core/pri-mir-16-1 interaction using EMSA in ahigh salt buffer (150mMKCl, data not shown). Thus, the appar-ent Kd values for DGCR8 binding to pri-mir-16-1 generallydecrease with increasing salt concentration.Hill coefficients often do not provide accurate estimates of

the number of binding sites (60), particularly when multisitecomplex formations are involved. In general, nitrocellulose fil-ter binding describes only macroscopic association parametersthat are composite averages of individual binding sites andcooperativity constants. Here, one must assume a relationshipbetween the number of ligands bound and complex retentionby the nitrocellulose membrane to accurately describe theinteraction. Senear et al. (61) have concluded that it is impossi-

ble to correctly resolve microscopic binding and cooperativityconstants of amultisite complex by filter binding. The apparentsteepness of dose-response curves and the corresponding Hillcoefficients (n) depend on invariant 32P-labeled RNA concen-tration below the Kd. When RNA concentrations approach theKd value, the resulting best fits to the Hill equation no longerprovide accurate information on either Kd or n (60, 62).Dimerization at higher DGCR8 concentration can potentiallyalter binding characteristics that would further complicate reli-able curve fitting. Using SEC and NMR diffusion measure-ments, we and others (49) have detected a dimerization ten-dency for DGCR8 variants. Finally, because we investigatedDGCR8-pri-mir-16-1 and pri-let-7b complex formation, wecannot rule out the possibility that the highly cooperative asso-ciation previously observed could be a specific feature ofDGCR8 binding to pri-mir30a and pri-mir21 studied by Falleret al. (13, 14). Although processed by theMicroprocessor, alter-

FIGURE 8. 1H,15N-TROSY resonance broadening map for binding of DGCR8core to pri-mir-16lower. A, changes in 1H LW (?LW) of 250 ?M2H,15N-labeled

DGCR8core in the presence of 10.4 ?M pri-mir-16lower versus the amino acid sequence of DGCR8core.?LW values were calculated by subtracting the LWbound ofa given residue from the LWfree.?LW for infinitely broadened resonances (gray bars) were calculated assuming a conservative LWbound of 50 Hz correspondingto the largest 1H-TROSY LWbound observable at 800MHz. Secondary structure elements (?1–7 and?1–7) as determined by chemical shift index using TALOS?(75) are identified below the residue number. B, color-codedworm representation of DGCR8core/pri-mir-16lower titration at a 24:1 protein/RNA ratio. Secondarystructure elements arenumberedand indicated.Worm thicknessof theDGCR8core crystal structure (PDBcode2YT4) variesbetween2.00 (?LW? ?38.3Hz, red)and 0.25 (?LW ? 9.2 Hz, blue). Missing crystal residues were modeled and energy-minimized in Chimera (76). C, DGCR8core/pri-mir-16lower titration at a 1:2protein/RNA ratio. Individual RBDs are numbered; N and C termini are shown and indicated. The 39 visible TROSY peaks remaining in the presence of excessRNApredominantlymap to theN andC termini, and flexible linker regions of theDGCR8core crystal structure (PDB code 2YT4) and aremapped andhighlightedin blue on the protein sequence (A).



native biogenesis pathways for pri-mir21 independent ofDGCR8 have been suggested (63).UV cross-linking experiments in the presence of varying

amounts of competitors have previously been employed toestablish relative affinities of DGCR8 to pri-miRNA (17).Although DGCR8 could be cross-linked with decreasing effi-ciency to diverse targets such as pri-mir-16-1, a small interfer-ing (si)RNA duplex and ssRNA, the ability to compete forDGCR8 binding revealed some advantage for ss-dsRNA junc-tions. Furthermore, a length dependence where shorter RNAconstructs competed gradually less efficiently with the longerpri-miRNAwas observed.OurDGCR8core/pri-mir-16-1 EMSAand increasing apparent Kd values in the presence of varyingamounts of nonspecific competitor yeast tRNAPhe suggest that

DGCR8core/RNA interactions are nonspecific. Consistent witha nonspecific binding mode, the results obtained from compe-tition experiments may have uncovered a correlation betweenefficient competition and the number of nonspecific bindingsites provided by the various competitors.To overcome the binding assay limitations described

above and to obtain independent information about theDGCR8core?pri-mir-16 complex equilibrium in solution, weconducted NMR titration and diffusionmeasurements. Resultsfrom NMR titrations strongly suggest that the DGCR8core?pri-mir-16lower complex represents a nonspecific interaction. Theobserved line broadening in the absence of chemical shift per-turbations indicates the formation of transient, nonspecificcomplexes (64). Results obtained from diffusionmeasurementsfurther corroborated nonspecific DGCR8core/pri-mir-16lowerbindingmodels as we observed interconverting freeDGCR8coreand complexes with hydrodynamic radii of ?38.1 Å.These findings are consistentwith reports of other dsRBD-con-

taining proteins involved in small RNA pathways. For example,dsRBDs of both HYL1 (HYPONASTIC LEAVES 1) and PKR thatexhibit noncanonical binding and extensive line broadening havebeen shown to bind RNAwithout apparent specificity (55, 65, 66).Similarly, we have been unable to detect by NMR nucleotide-spe-cific changes to RNA targets complexed with DGCR8core. Analo-gous to the protein spectra, pri-mir-16 imino-proton peaks werebroadened globally and severely. This result suggests that DGCR8dsRBDs simultaneously bind overlapping interfaces of the RNA inamanner resemblingnonspecific PKR?dsRNAcomplex formation(66). Finally, theDicer accessorydsRBD-containingprotein-trans-activating response RNA-binding protein and two homologs,including PKR, weremost recently shown to diffuse along dsRNAsubstrates, which is clearly an activity conceivably adopted byDGCR8 to anchor Drosha-DGCR8 heterodimers (67).Although our nonspecific DGCR8/pri-miRNA interaction

model differs from the proposed cooperative recognitionmodel (14), several common observations have been described.Two studies have identified that helix 2 in each ?-?-?-?-?dsRNA-binding domain ofDGCR8 is important for pri-miRNA

FIGURE 9. pri-mir-16lower?DGCR8core complex formation monitored byNMR. A, 850 MHz imino one-dimensional jump-return echo experiment ofpri-mir-16lower collectedat 298K (black) labeledwith assignment information;numbering is consistent with full-length pri-mir-16-1 (Fig. 1). RNA/proteinmolar ratios ranging from 48:1 (orange) to 1:1 (gray) were achieved by addingdilute aliquots of DGCR8core to a 250 ?M pri-mirlower sample in NMR buffer. B,same as above with 5 mM MgCl2 added to the NMR buffer.

TABLE 2Summary of least squares regressions and experimental and theoretical translational self-diffusion constants and hydrodynamic radii

DGCR8RBD1 DGCR8core DGCR8PC pri-mir-16lowerDGCR8core/pri-mir-16lower

complexComplexlower limit

Complexupper limit

Slope ?18.34 ?11.768 ?11.21 ?11.24 ?9.541 ?13.78 ?8.641ln(I/I0) versus (g2-g02)a R ? 0.997 R ?0.998 R ? 0.997 R ?0.996 R ? 0.990 R ? 0.999 R ? 0.999D 13.561 8.657 8.246 8.267 7.017 10.132 6.35610?11 M2 s?b ? 0.044 ? 0.100 ? 0.040 ? 0.121 ? 0.223 ? 1.024 ? 0.153D calc 13.83 9.815 8.71610?11 M2 s?1c ? 0.098RHexp 17.9 28.0 29.4 34.5 24.1 38.1Åd ? 1.0 ? 0.3 ? 0.1 ? 1.1 ? 2.6 ? 0.9RHcalc 17.5 24.7Åe

a Slopes were derived from linear least squares fits using three replicate measurements (30 data points total for the 1st 5 columns; 15 data points were used to determine up-per and lower boundaries, 6th and 7th columns). Data analysis was performed using Prism 4.0 software.

b Translational diffusion constants derived from Equation 2 at 298 K in aqueous (90:10% H2O/D2O) buffer solutions containing 135 mM KCl. Three replicate experimentswere performed, and the resulting average values for D with corresponding standard deviations are reported.

c Theoretical translational diffusion constants were calculated from atomic level structures available for DGCR8RBD1 and DGCR8core using HYDROPRO (PDB accessioncodes 1X47 and 2YT4, respectively) (71, 72). All-atom models for the 65 nucleotide pri-mir-16lower were generated using MC-Sym from sequence and NMR-based second-ary structure (37). 132 structures were clustered using MC-Sym, and five representative structures displaying various bend angles were analyzed using HYDROPRO. Theresulting average value for D with corresponding standard deviation is reported.

d Experimental hydrodynamic radii assuming spherical shapes derived from Equation 3. Three replicate experiments were performed, and the resulting average values for RHwith corresponding standard deviations are reported.

e Theoretical hydrodynamic radii calculated from atomic level structures available for DGCR8RBD1 and DGCR8core using HYDROPRO (PDB accession codes 1X47 and2YT4, respectively).



recognition (15, 49). In good agreement, our NMR titrationanalysis recognizes the first and second helices of each dsRBDmotif (RBD1:?1 and ?2, RBD2:?5 and ?6) as primary pri-mir-16-1 interaction interfaces. The binding affinity of DGCR8 forthe 160-nt P4–P6 domain of theTetrahymena ribozyme is sim-ilar to the ones observed for bona fide pri-miRNA substrates(14). Likewise, we cannot distinguish ssRNA, fully paired (pri-mir-16fp) or random hairpin (pri-mir-16x) interactions involv-ing DGCR8 from pri-miRNA-derived substrates based on Kdvalues. Previous studies have highlighted the importance ofoligomeric (DGCR8)n?pri-miRNA complexes using filter bind-ing assays, SEC, and low resolution electron tomography (14).Our NMR titration and diffusion analysis directly probe solu-tion equilibria and offer further support for oligomeric, albeitnonspecific, binding of DGCR8 to RNA substrates. Further-

more, we were able to visualize DGCR8?pri-miRNA complexeswith variable stoichiometry using EMSA.DGCR8 alone shows no preference for a particular RNA sub-

strate, yet we and others have demonstrated through in vitroprocessing assays that the Microprocessor is highly selectivewhen cleaving its RNA targets (17, 18). Thus, we conclude thatDGCR8-RNA substrate binding is not the specificity determi-nant of pri- to pre-miRNA processing, and a model of stepwisepri-miRNA recognition and Drosha recruitment by DGCR8seems unlikely. Because known requirements common to allMicroprocessor substrates remain broad and include 70-nthairpins with unpaired flanking regions (19), the functionalimplication of our finding is that the concerted action of aDGCR8?Drosha complex is required for effective recognition ofmiRNA-containing transcripts. Consistently, recent studies

FIGURE10.Pulsed fieldgradientNMRandDGCR8-pri-mir-16-1 complexmodels.A,natural log scale of signal integrals of BPPLED-PFG 1HNMRexperimentsas a function of gradient amplitude squared and linear data fits. Diffusion of individual protein and RNA components is as follows: DGCR8RBD1 (red), DGCR8core

(blue), pri-mir-16lower (green), andDGCR8PC (black).B,diffusionof 1:2molar ratioDGCR8core?pri-mir-16lower complexes as follows: average (dashed red line), lowerlimit (solid blue line), and upper limit (dashed black line). The inset (G2-G0

2 ? 0.00 to 0.05, ln(1/I0)? 0.0 to?0.5) in the upper right corner shows the best fit for theupper complex limit (dashed black line).

FIGURE 11. In vitro pri-miRNA processing assays. A, processing of uniformly labeled pri-miRNAs in the absence (Input) or in the presence (c-myc-IP) ofimmunoprecipitatedMicroprocessor. Assayswere performed as described under “Experimental Procedures.” Sizes of oligonucleotideDNAmarkers are shownon the left. 5?, 3?, andpre-miRNAprocessingproducts are indicated on the right. B,DGCR8?275-dependent pri-miRNAprocessing. RecombinantDGCR8?275wasincubated with stringently washed IP-Drosha. 32P-5?-End-labeled pri-mir-16-1 variants were incubated in the absence (?) or presence (?) of Microprocessorcomponents. Images in A and B are representative of at least two independent experiments.



revealed that associationwith theHIV-1 LTR requires the pres-ence of bothMicroprocessor components, Drosha and DGCR8(68). An alternative scenario, where Drosha can distinguish itstarget among nonspecific DGCR8?pri-miRNA complexes can-not be strictly ruled out based on our results.Importantly, stable association ofDGCR8 andDrosha occurs

independently of RNA (10). Presumably, this protein/proteininteraction leads to enhancedRNAbinding specificity byway ofconformationalmodulation ofDGCR8’s dsRBDs, by steric limi-tation of binding opportunities to constrained regions of theRNA hairpin, or a combination of both scenarios. The DGCR8/RNA interaction and other examples (55, 67, 69, 70) suggestthat dsRBDs associated with miRNA biogenesis are generallynonspecific binders. Although this form of association mightseem less than satisfying, nonspecific and/or loose binding canbe an important characteristic for an accessory protein. Thus,the biological implication is that nonspecific associations cangreatly enhance the RNA binding flexibility of a protein com-plex, which is in good agreement with reports demonstratingDGCR8’s involvement in controlling the fate of diverse classesof RNA such asmRNAs, small nucleolar RNAs, long noncodingRNAs, and pri-miRNAs (24). For theMicroprocessor in partic-ular, the ability to bind a broad range of loosely related hairpinstructures could be critical to carry out its universal function.

Acknowledgments—We thank members of the Hennig laboratory forstimulating discussions and comments on the manuscript. Weacknowledge the support of the Hollings Marine Laboratory NMRFacility for this work.

REFERENCES1. Bartel, D. P. (2009) MicroRNAs: target recognition and regulatory func-

tions. Cell 136, 215–2332. Voinnet, O. (2009) Origin, biogenesis, and activity of plant microRNAs.

Cell 136, 669–6873. Dreyer, J. L. (2010) New insights into the roles of microRNAs in drug

addiction and neuroplasticity. Genome Med. 2, 924. Farazi, T. A., Spitzer, J. I., Morozov, P., and Tuschl, T. (2011) miRNAs in

human cancer. J. Pathol. 223, 102–1155. Ferland-McCollough, D., Ozanne, S. E., Siddle, K., Willis, A. E., and Bush-

ell, M. (2010) The involvement of microRNAs in type 2 diabetes. Biochem.Soc. Trans. 38, 1565–1570

6. Wittmann, J., and Jack, H. M. (2011) microRNAs in rheumatoid arthritis:midget RNAs with a giant impact. Ann. Rheum. Dis. 70, Suppl. 1, 92–96

7. Lee, S. J., Jiko, C., Yamashita, E., and Tsukihara, T. (2011) Selective nuclearexport mechanism of small RNAs. Curr. Opin. Struct. Biol. 21, 101–108

8. Yi, R., Qin, Y., Macara, I. G., and Cullen, B. R. (2003) Exportin-5 mediatesthe nuclear export of pre-microRNAs and short hairpin RNAs.Genes Dev.17, 3011–3016

9. Lee, Y., Ahn, C., Han, J., Choi, H., Kim, J., Yim, J., Lee, J., Provost, P.,Rådmark, O., Kim, S., and Kim, V. N. (2003) The nuclear RNase III Droshainitiates microRNA processing. Nature 425, 415–419

10. Han, J., Lee, Y., Yeom, K. H., Kim, Y. K., Jin, H., and Kim, V. N. (2004) TheDrosha-DGCR8 complex in primary microRNA processing. Genes Dev.18, 3016–3027

11. Barr, I., Smith, A. T., Chen, Y., Senturia, R., Burstyn, J. N., and Guo, F.(2012) Ferric, not ferrous, heme activates RNA-binding protein DGCR8for primary microRNA processing. Proc. Natl. Acad. Sci. U.S.A. 109,1919–1924

12. Barr, I., Smith, A. T., Senturia, R., Chen, Y., Scheidemantle, B. D., Burstyn,J. N., and Guo, F. (2011) DiGeorge critical region 8 (DGCR8) is a double-cysteine-ligated heme protein. J. Biol. Chem. 286, 16716–16725

13. Faller, M., Matsunaga, M., Yin, S., Loo, J. A., and Guo, F. (2007) Heme isinvolved in microRNA processing. Nat. Struct. Mol. Biol. 14, 23–29

14. Faller, M., Toso, D., Matsunaga, M., Atanasov, I., Senturia, R., Chen, Y.,Zhou, Z. H., and Guo, F. (2010) DGCR8 recognizes primary transcripts ofmicroRNAs through highly cooperative binding and formation of higher-order structures. RNA 16, 1570–1583

15. Wada, T., Kikuchi, J., and Furukawa, Y. (2012) Histone deacetylase 1 en-hances microRNA processing via deacetylation of DGCR8. EMBO Rep.13, 142–149

16. Zeng, Y., Yi, R., and Cullen, B. R. (2005) Recognition and cleavage ofprimary microRNA precursors by the nuclear processing enzyme Drosha.EMBO J. 24, 138–148

17. Han, J., Lee, Y., Yeom, K. H., Nam, J. W., Heo, I., Rhee, J. K., Sohn, S. Y.,Cho, Y., Zhang, B. T., and Kim, V. N. (2006) Molecular basis for therecognition of primary microRNAs by the Drosha-DGCR8 complex. Cell125, 887–901

18. Zeng, Y., and Cullen, B. R. (2005) Efficient processing of primary mi-croRNA hairpins by Drosha requires flanking nonstructured RNA se-quences. J. Biol. Chem. 280, 27595–27603

19. Auyeung, V. C., Ulitsky, I., McGeary, S. E., and Bartel, D. P. (2013) Beyondsecondary structure: primary-sequence determinants license pri-miRNAhairpins for processing. Cell 152, 844–858

20. Ramos, A., Grunert, S., Adams, J., Micklem, D. R., Proctor, M. R., Freund,S., Bycroft, M., St Johnston, D., and Varani, G. (2000) RNA recognition bya Staufen double-stranded RNA-binding domain. EMBO J. 19, 997–1009

21. Stefl, R., Xu, M., Skrisovska, L., Emeson, R. B., and Allain, F. H. (2006)Structure and specific RNA binding of ADAR2 double-stranded RNAbinding motifs. Structure 14, 345–355

22. Wu, H., Henras, A., Chanfreau, G., and Feigon, J. (2004) Structural basisfor recognition of the AGNN tetraloop RNA fold by the double-strandedRNA-binding domain of Rnt1p RNase III. Proc. Natl. Acad. Sci. U.S.A.101, 8307–8312

23. Stefl, R., Oberstrass, F. C., Hood, J. L., Jourdan, M., Zimmermann, M.,Skrisovska, L., Maris, C., Peng, L., Hofr, C., Emeson, R. B., and Allain, F. H.(2010) The solution structure of the ADAR2 dsRBM-RNA complex re-veals a sequence-specific readout of the minor groove. Cell 143, 225–237

24. Macias, S., Plass, M., Stajuda, A., Michlewski, G., Eyras, E., and Caceres,J. F. (2012) DGCR8 HITS-CLIP reveals novel functions for the Micropro-cessor. Nat. Struct. Mol. Biol. 19, 760–766

25. Roth, B. M., and Hennig, M. (2011) in RNA Structure Determination byNMR: Combining Labeling and Pulse Techniques (Dingley, A. J., and Pas-cal, S. M., eds) pp. 205–228, IOS Press, Washington, D. C.

26. Sambrook, J., Fritsch, E. F., and Maniatis, T. (eds) (1989)Molecular Clon-ing, Vol. 3, 2nd Ed., Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY

27. Roth, B. M., and Hennig, M. (2012) Backbone 1H(N), 13C, and 15N reso-nance assignments of the tandem RNA-binding domains of humanDGCR8. Biomol. NMR Assign, in press

28. Arraiano, C. M., Barbas, A., and Amblar, M. (2008) Characterizing ribo-nucleases in vitro examples of synergies between biochemical and struc-tural analysis. Methods Enzymol. 447, 131–160

29. Lee, Y., and Kim, V. N. (2007) In vitro and in vivo assays for the activity ofDrosha complex. Methods Enzymol. 427, 89–106

30. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A.(1995) NMRPipe: a multidimensional spectral processing system based onUNIX pipes. J. Biomol. NMR 6, 277–293

31. Goddard, T. D., and Kneller, D. G. (2004) SPARKY 3, University of Cali-fornia, San Francisco, San Francisco, CA

32. Sklenar, V., and Bax, A. (1987) Spin-echo water suppression for the gen-eration of pure-phase two-dimensional NMR spectra. J. Magn. Reson. 74,469–479

33. Lippens, G., Dhalluin, C., and Wieruszeski, J. M. (1995) Use of a waterflip-back pulse in the homonuclear NOESY experiment. J. Biomol. NMR 5,327–331

34. Bax, A., Griffey, R. H., and Hawkins, B. L. (1983) Correlation of proton andnitrogen-15 chemical shifts by multiple quantum NMR. J. Magn. Reson.55, 301–315

35. Dingley, A. J., and Grzesiek, S. (1998) Direct observation of hydrogen



bonds in nucleic acid base pairs by internucleotide (2)J(NN) couplings.J. Am. Chem. Soc. 120, 8293–8297

36. Pervushin, K., Ono, A., Fernandez, C., Szyperski, T., Kainosho, M., andWuthrich, K. (1998) NMR scaler couplings across Watson-Crick base pairhydrogen bonds in DNA observed by transverse relaxation optimizedspectroscopy. Proc. Natl. Acad. Sci. U.S.A. 95, 14147–14151

37. Parisien, M., and Major, F. (2008) The MC-Fold and MC-Sym pipelineinfers RNA structure from sequence data. Nature 452, 51–55

38. Sattler, M., Schleucher, J., and Griesinger, C. (1999) Heteronuclear multi-dimensional NMR experiments for the structure determination of pro-teins in solution employing pulsed field gradients. Prog. Nucl. Magn.Reson. Spectrosc. 34, 93–158

39. Wostenberg, C., Quarles, K. A., and Showalter, S. A. (2010) Dynamicorigins of differential RNA binding function in two dsRBDs from themiRNA “Microprocessor” complex. Biochemistry 49, 10728–10736

40. Chou, J. J., Baber, J. L., and Bax, A. (2004) Characterization of phospholipidmixed micelles by translational diffusion. J. Biomol. NMR 29, 299–308

41. Wu, D. H., Chen, A. D., and Johnson, C. S. (1995) An improved diffusion-ordered spectroscopy experiment incorporating bipolar-gradient pulses.J. Magn. Reson. A 115, 260–264

42. Stejskal, E. O., and Tanner, J. E. (1965) Spin diffusion measurements: Spinechoes in the presence of a time-dependent field gradient. J. Chem. Phys.42, 288–292

43. Harris, K. R., and Woolf, L. A. (2004) Temperature and volume depen-dence of the viscosity of water and heavy water at low temperatures.J. Chem. Eng. Data 49, 1064–1069

44. Kestin, J., Imaishi, N., Nott, S. H., Nieuwoudt, J. C., and Sengers, J. V.(1985) Viscosity of light and heavy water and their mixtures. Phys A: Stat.Theo. Phys. 134, 38–58

45. Jones, G., and Dole, M. (1929) The viscosity of aqueous solutions of strongelectrolytes with special reference to barium chloride. J. Am. Chem. Soc.51, 2950–2964

46. Dordick, R., Korson, L., and Drost-Hansen, W. (1979) High-precision vis-cosity measurements: I. Aqueous solutions of alkali chlorides. J. ColloidInterface Sci. 72, 206–214

47. Landthaler, M., Yalcin, A., and Tuschl, T. (2004) The human DiGeorgesyndrome critical region gene 8 and its D. melanogaster homolog are re-quired for miRNA biogenesis. Curr. Biol. 14, 2162–2167

48. Yeom, K. H., Lee, Y., Han, J., Suh, M. R., and Kim, V. N. (2006) Character-ization of DGCR8/Pasha, the essential cofactor for Drosha in primarymiRNA processing. Nucleic Acids Res. 34, 4622–4629

49. Sohn, S. Y., Bae, W. J., Kim, J. J., Yeom, K. H., Kim, V. N., and Cho, Y. (2007)Crystal structure of human DGCR8 core. Nat. Struct. Mol. Biol. 14,847–853

50. Markham, N. R., and Zuker, M. (2005) DINAMelt web server for nucleicacid melting prediction. Nucleic Acids Res. 33,W577–W581

51. Corcoran, D. L., Pandit, K. V., Gordon, B., Bhattacharjee, A., Kaminski, N.,and Benos, P. V. (2009) Features of mammalian microRNA promotersemerge from polymerase II chromatin immunoprecipitation data. PLoS One4, e5279

52. Corcoran, M. M., Hammarsund, M., Zhu, C., Lerner, M., Kapanadze, B.,Wilson, B., Larsson, C., Forsberg, L., Ibbotson, R. E., Einhorn, S., Oscier,D. G., Grander, D., and Sangfelt, O. (2004) DLEU2 encodes an antisenseRNA for the putative bicistronic RFP2/LEU5 gene in humans and mouse.Genes Chromosomes Cancer 40, 285–297

53. Williams, B. R. (1999) PKR: a sentinel kinase for cellular stress. Oncogene18, 6112–6120

54. Wostenberg, C., Noid, W. G., and Showalter, S. A. (2010) MD simulationsof the dsRBP DGCR8 reveal correlated motions that may aid pri-miRNAbinding. Biophys. J. 99, 248–256

55. Yang, S. W., Chen, H. Y., Yang, J., Machida, S., Chua, N. H., and Yuan, Y. A.(2010) Structure of Arabidopsis HYPONASTIC LEAVES1 and its molec-ular implications for miRNA processing. Structure 18, 594–605

56. Tian, B., Bevilacqua, P. C., Diegelman-Parente, A., and Mathews, M. B.(2004) The double-stranded-RNA-binding motif: interference and much

more. Nat. Rev. Mol. Cell Biol. 5, 1013–102357. Gregory, R. I., Yan, K. P., Amuthan, G., Chendrimada, T., Doratotaj, B.,

Cooch, N., and Shiekhattar, R. (2004) The Microprocessor complex me-diates the genesis of microRNAs. Nature 432, 235–240

58. Haase, A. D., Jaskiewicz, L., Zhang, H., Laine, S., Sack, R., Gatignol, A., andFilipowicz, W. (2005) TRBP, a regulator of cellular PKR and HIV-1 virusexpression, interacts with Dicer and functions in RNA silencing. EMBORep. 6, 961–967

59. Nowotny, M., and Yang, W. (2009) Structural and functional modules inRNA interference. Curr. Opin. Struct. Biol. 19, 286–293

60. Weiss, J. N. (1997) The Hill equation revisited: uses and misuses. FASEB J.11, 835–841

61. Senear, D. F., Brenowitz, M., Shea, M. A., and Ackers, G. K. (1986) Ener-getics of cooperative protein-DNA interactions: comparison betweenquantitative deoxyribonuclease footprint titration and filter binding. Bio-chemistry 25, 7344–7354

62. Hall, K. B., and Kranz, J. K. (1999) Nitrocellulose filter binding for deter-mination of dissociation constants. Methods Mol. Biol. 118, 105–114

63. Yi, R., Pasolli, H. A., Landthaler, M., Hafner, M., Ojo, T., Sheridan, R.,Sander, C., O’Carroll, D., Stoffel, M., Tuschl, T., and Fuchs, E. (2009)DGCR8-dependent microRNA biogenesis is essential for skin develop-ment. Proc. Natl. Acad. Sci. U.S.A. 106, 498–502

64. Marintchev, A., Frueh, D., and Wagner, G. (2007) NMR methods forstudying protein-protein interactions involved in translation initiation.Methods Enzymol. 430, 283–331

65. Rasia, R. M., Mateos, J., Bologna, N. G., Burdisso, P., Imbert, L., Palatnik, J. F.,and Boisbouvier, J. (2010) Structure and RNA interactions of the plant mi-croRNA processing-associated protein HYL1. Biochemistry 49, 8237–8239

66. Ucci, J. W., Kobayashi, Y., Choi, G., Alexandrescu, A. T., and Cole, J. L.(2007) Mechanism of interaction of the double-stranded RNA (dsRNA)-binding domain of protein kinase R with short dsRNA sequences. Bio-chemistry 46, 55–65

67. Koh, H. R., Kidwell, M. A., Ragunathan, K., Doudna, J. A., and Myong, S.(2013) ATP-independent diffusion of double-stranded RNA-binding pro-teins. Proc. Natl. Acad. Sci. U.S.A. 110, 151–156

68. Wagschal, A., Rousset, E., Basavarajaiah, P., Contreras, X., Harwig, A., Lau-rent-Chabalier, S.,Nakamura,M.,Chen,X., Zhang,K.,Meziane,O., Boyer, F.,Parrinello,H., Berkhout, B., Terzian,C., Benkirane,M., andKiernan,R. (2012)Microprocessor, Setx,Xrn2, andRrp6 co-operate to inducepremature termi-nation of transcription by RNAPII. Cell 150, 1147–1157

69. Parker, G. S., Eckert, D. M., and Bass, B. L. (2006) RDE-4 preferentiallybinds long dsRNA and its dimerization is necessary for cleavage of dsRNAto siRNA. RNA 12, 807–818

70. Parker, G. S., Maity, T. S., and Bass, B. L. (2008) dsRNA binding propertiesof RDE-4 and TRBP reflect their distinct roles in RNAi. J. Mol. Biol. 384,967–979

71. Fernandes, M. X., Bernado, P., Pons, M., and García de la Torre, J. (2001)An analytical solution to the problem of the orientation of rigid particlesby planar obstacles. Application to membrane systems and to the calcu-lation of dipolar couplings in protein NMR spectroscopy. J. Am. Chem.Soc. 123, 12037–12047

72. García De La Torre, J., Huertas, M. L., and Carrasco, B. (2000) Calculationof hydrodynamic properties of globular proteins from their atomic-levelstructure. Biophys. J. 78, 719–730

73. Zuker, M. (2003) Mfold web server for nucleic acid folding and hybridiza-tion prediction. Nucleic Acids Res. 31, 3406–3415

74. Han, K., Lee, Y., and Kim, W. (2002) PseudoViewer: automatic visualiza-tion of RNA pseudoknots. Bioinformatics 18, S321–S328

75. Shen, Y., Delaglio, F., Cornilescu, G., and Bax, A. (2009) TALOS?: a hy-brid method for predicting protein backbone torsion angles from NMRchemical shifts. J. Biomol. NMR 44, 213–223

76. Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt,D. M., Meng, E. C., and Ferrin, T. E. (2004) UCSF Chimera–a visualizationsystem for exploratory research and analysis. J. Comput. Chem. 25,1605–1612



roth, ishimaru and hennig 2013

Documents