JBC Avanti Polar Lipids

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M611163200 on March 28, 2007

J. Biol. Chem., Vol. 282, Issue 20, 14960-14967, May 18, 2007
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
282/20/14960    most recent
M611163200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ramesh, A.
Right arrow Articles by Sacchettini, J. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ramesh, A.
Right arrow Articles by Sacchettini, J. C.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Crystal Structure of Rsr, an Ortholog of the Antigenic Ro Protein, Links Conformational Flexibility to RNA Binding Activity*Formula

Arati Ramesh{ddagger}§, Christos G. Savva||, Andreas Holzenburg||, and James C. Sacchettini{ddagger}§1

From the {ddagger}Department of Biochemistry and Biophysics, §Center for Structural Biology, Microscopy and Imaging Center, and ||Department of Biology, Texas A & M University, College Station, Texas 77843

Received for publication, December 5, 2006 , and in revised form, March 21, 2007.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 REFERENCES
 
Ro ribonucleoproteins are a class of antigenic ribonucleoproteins associated with rheumatic autoimmune diseases like systemic lupus erythematosus and Sjögrens syndrome in humans. Ro ribonucleoproteins are mostly composed of the 60-kDa Ro protein and small cytoplasmic RNAs, called Y RNAs, of unknown function. In eukaryotes, where Ro has been found to associate with damaged or mutant RNAs, it has been suggested that Ro may play a role in RNA quality control. In the radiation-resistant bacterium Deinococcus radiodurans and some eukaryotes, Ro has also been implicated in cell survival following UV damage. Here we present the first high resolution structure of a prokaryotic Ro ortholog, Rsr from D. radiodurans. The structure has been solved to 1.9 Å resolution and shows distinct differences when compared with the eukaryotic apo- and RNA-bound Ro structures. Rsr is composed of two domains: a helical RNA binding domain and a mixed "von Willebrand factor A-like" domain containing a divalent metal binding site. Although the individual domains of Rsr are similar to the eukaryotic Ro, significantly large differences are seen at the interface of the two domains. Since this interface communicates with the conserved central cavity of Ro, which is implicated in RNA binding, changes at this interface could potentially influence RNA binding by Ro. Although the apo-Rsr protein is monomeric, Rsr binds Y RNA to form multimers of ~12 molecules of a 1:1 Rsr-Y RNA complex. Rsr binds D. radiodurans Y RNA with low nanomolar affinity, comparable with previously characterized eukaryotic Ro orthologs.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 REFERENCES
 
Ro ribonucleoproteins, a class of RNA protein complexes, were first identified as being major targets of the immune system in patients suffering from rheumatic disorders like systemic lupus erythematosus and Sjögrens syndrome. The sera from these patients have an unusually high titer of autoantibodies that recognize Ro ribonucleoproteins (1). These ribonucleoproteins are composed of a 60-kDa protein called Ro and small noncoding RNAs called Y RNAs. Y RNAs are RNA polymerase III transcripts with an average size around 100 nucleotides and a yet unknown function. It has been shown that a subset of Ro ribonucleoproteins also contain other protein components like the La protein, which binds a number of nascent small RNAs, including Y RNAs (2) and possibly other proteins. In mice, the absence of Ro leads to the development of a lupus-like syndrome, characterized by autoantibodies against chromatin and ribosomes and glomerulonephritis and photosensitivity (3), leading to the proposal that the normal function of Ro may be important for the prevention of autoimmune reactions.

In Xenopus oocytes, Ro has been identified in complex with mutant pre-5 S rRNAs in addition to Y RNAs. These mutant pre-5 S rRNAs are characterized by additional nucleotides at the 3' end and one or more point mutations compared with the major oocyte 5 S rRNA sequence, which enable the mutants to adopt an alternate helical conformation (4). The mutant rRNAs, unfit to follow the pathway to maturation, eventually get degraded (5). This has led to the proposal that Ro may function as part of a novel quality control or discard pathway for 5 S rRNA production (3).

Recent studies on the eukaryotic ortholog of Ro from Xenopus laevis have shown that Ro is composed of two domains. One is an elliptical domain formed by many repeats of a pair of antiparallel {alpha}-helices called HEAT repeats (6). The second domain resembles the collagen-binding A domain of von Willebrand factor (vWFA domain)2 consisting of a beta-sheet sandwiched by multiple {alpha}-helices (7). The vWFA domain also includes a divalent cation-binding metal ion-dependent adhesion site (MIDAS) motif.

The structure of X. laevis Ro in complex with a fragment of Y RNA containing the Ro binding site shows that Ro binds Y RNAs at the outer surface of the helical domain. The positively charged central cavity surrounded largely by the elliptical domain, with the vWFA domain on one end, is large enough only to fit a single-stranded RNA. It has been proposed that this central cavity is the binding site for the single-stranded tails of mutant pre-5 S rRNAs (7). Mutational and structural analysis have also shown that at least part of the mutant pre-5 S rRNA binding site on X. laevis Ro overlaps the Y RNA binding site, thereby leading to a proposed role for Y RNAs in regulating the binding of Ro to other RNAs like the mutant pre-5 S rRNAs or other damaged RNAs (7, 8).

Although orthologs of Ro have now been identified in a variety of species, including all vertebrates, the nematode Caenorhabditis elegans, the unicellular eukaryote Clamydomonas reinhardtii, the cyanobacteria Nostoc punctiforme and Synechococcus, the eubacteria Deinococcus radiodurans and Mycobacterium smegmatis, and the mycobacteriophage Bxz1 (9), it is unclear why such a vastly different set of species have evolved to maintain the ortholog of the 60-kDa Ro and whether there is a common function for Ro in all of these organisms. Among the orthologs of Ro in prokaryotes, the best characterized is the Rsr (Ro-sixty-related) protein from D. radiodurans. D. radiodurans is a eubacterium with the ability to tolerate unusually high doses of ionizing radiation (10). Studies have shown that the Rsr protein contributes to the survival of D. radiodurans following unusually high doses of UV irradiation (11). D. radiodurans cells lacking Rsr are more sensitive to UV radiation than wild-type cells, and this phenotype can be partly complemented by Rsr expressed under the control of a heterologous promoter. Rsr, along with several small RNAs encoded upstream of rsr, accumulates following UV irradiation. These RNAs, of which at least one is predicted to assume a secondary structure that resembles a Y RNA, can be immunoprecipitated along with Rsr using anti-Rsr antibodies (11). These experiments along with studies on mammalian cells provide clear evidence that Ro plays a role in cell survival after UV irradiation (12). In humans, patients suffering from systemic autoimmune diseases often develop subcutaneous photosensitive lesions similar to that following UV damage (13).

In this study, we report a high resolution structure of a prokaryotic ortholog of Ro, namely Rsr from D. radiodurans. Although the overall structure of Rsr resembles the architecture of X. laevis Ro, there are major structural differences between the two proteins. The most notable difference is at the interface of the two domains, where large movements are evident in the {alpha}-helices close to the interface. Movement of these helices results in an enlarged central cavity, suggesting that the structural flexibility of Ro proteins probably aids in switching between "open" and "closed" states, which may be required for entry, stabilization, and release of the RNA substrates at the conserved central cavity. We have also characterized the binding of Rsr to Deinococcus Y RNA in vitro, using a fluorescence-based assay. Our results reveal that Rsr binds Y RNA with at least low nanomolar affinity and forms a complex with 1:1 stoichiometry. Although the apo-Rsr protein is monomeric even at very high concentrations, electron microscopy and size exclusion chromatography show that upon binding Y RNA, Rsr forms multimers of ~12 molecules of the 1:1 Rsr-Y RNA complex.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS AND DISCUSSION
 REFERENCES
 
Construction of an Escherichia coli Overexpression Plasmid for Rsr and Purification of Rsr—As described in the supplemental information, the Rsr coding region was amplified from D. radiodurans genomic DNA and cloned into pET28b to construct pET28b-Rsr. The resultant construct was expressed in E. coli BL21(DE3). The overexpressed Rsr was purified using Ni2+ affinity chromatography. Following cleavage of the His6 tag using thrombin, the His6-tagged and untagged Rsr were separated on a second Ni2+ column. Finally, Rsr was dialyzed against 25 mM phosphate, pH 7.0, 150 mM sodium chloride, and 2 mM dithiothreitol and used for subsequent experiments.

Crystallization and X-ray Data Collection—Rod-shaped crystals for native Rsr were grown by the sitting drop method in Intelli plates, in a crystallization solution consisting of 0.1 M imidazole, pH 8.0, 12% polyethylene glycol 8000, and 0.2 M calcium acetate. Diffraction quality crystals were obtained by mixing 4 µl of crystallization solution with 2 µl of protein at a concentration of 400 µM and equilibrating against 100 µl of crystallization solution. The crystals were flash-frozen in liquid nitrogen using 25% glycerol in mother liquor as a cryoprotectant. The best crystal diffracted to 1.9 Å at beamline 23-ID at the Advanced Photon Source (Argonne National Laboratories), with cell dimensions of a = 103.56 Å, b = 87.61 Å, c = 70.62 Å, {alpha} = 90.0°, beta = 96.57°, and {gamma} = 90.0°.

Molecular replacement using full-length or individual domains of X. laevis Ro (Protein Data Bank code 1YVP or 1YVR) as a search model was not successful; hence, to obtain phase information for structure determination, selenomethionine-substituted Rsr was produced in E. coli B834 (DE3) cells as described in the supplemental information. The same purification protocol was followed as outlined for the native Rsr.

Crystals for selenomethionine-substituted Rsr were obtained in the same condition as that of native Rsr. The selenomethionine Rsr crystals were flash-frozen in liquid nitrogen using paratone as the cryoprotectant. The best crystal diffracted to 2.6 Å at beamline 23-ID, and the data were isomorphous to the native Rsr diffraction data, with cell dimensions of a = 105.41 Å, b = 89.39 Å, c = 70.55 Å, {alpha} = 90.0°, beta = 96.8°, and {gamma} = 90.0°. Data statistics are compiled in supplemental Table 1.

Structure Determination—X-ray diffraction data sets were integrated, scaled, and merged using HKL2000 (14). From the three-wavelength MAD data from selenomethionine Rsr crystals, 13 of 14 selenium sites were located using phenix.hyss (15). Initial phase calculation using data from 50 to 2.6 Å resolution, followed by density modification, was done using AutoSHARP (16). With these density-modified phases, the electron density map was calculated and initial model was built using TEXTAL (17). Despite the low resolution of the MAD data and poor quality of the phases, TEXTAL built a nearly complete initial model. Improvements in the model were done manually on XtalView (18) using the structure of X. laevis Ro as reference (Protein Data Bank code 1YVR). This improved model was refined against the native Rsr data first by rigid body refinement and subsequently by simulated annealing refinement using phenix.refine (15). AutoBUILD on the phenix suite was used to improve the model. Bias-minimized maps were created using the TB Bias Removal Server (19). Strong positive electron density was seen surrounded by the cation binding MIDAS motif (DXSXS... T... D), and Ca2+ was refined into this density, since the crystallization condition contains 200 mM calcium acetate. The final model was refined with crystallographic Rwork and Rfree of 22.3 and 26.2%, respectively. Refinement statistics are shown in Table 1. 99% of residues are in the allowed region of the Ramachandran plot with 1% in the generously allowed region. Electrostatic surface potential was calculated using APBS (20) and UCSF Chimera; all other figures were prepared using UCSF Chimera (21, 22).


View this table:
[in this window]
[in a new window]

  		TABLE 1
Refinement statistics for Rsr Highest resolution shell is shown in parentheses.

 
Preparation of D. radiodurans Y RNA (DrY RNA)—A 35-nucleotide-long fragment of DrY RNA (called DrY RNA35) encompassing the Ro binding conserved helical stem was purchased from Integrated DNA Technologies (5'-GCUCUGGGCCGAAGUGCCGAACCCUUGUCCAGAGC). mFOLD-predicted secondary structures (23) of the full-length DrY RNA and the 35-mer fragment DrY RNA35 are shown in supplemental Fig. 2. The RNA was resuspended in 25 mM Hepes, pH 7.0, 10 mM MgCl2, 0.5 mM EDTA, and 150 mM NaCl and annealed by heating at 85 °C for 5 min followed by cooling at room temperature for 45 min. Folded RNA was subjected to nondenaturing gel electrophoresis on a 15% TBE Ready Gel (Bio-Rad) to check for formation of alternate dimer conformations.

Fluorescence Assays to Study Binding between Rsr and DrY RNA35—Intrinsic tryptophan fluorescence measurements were carried out on an ISS PC1 photon-counting spectrofluorometer with Ro at a concentration of 50 nM. The reaction buffer consisted of 25 mM Hepes, pH 7.0, 150 mM NaCl, and 10 mM MgCl2. RNA was titrated in to achieve final concentrations from 5.8 to 665 nM. Following each titration, the sample was allowed to equilibrate for 10 min, and fluorescence emission spectra (315–450 nm) were acquired upon excitation at 295 nm. Simultaneous titration of RNA to buffer alone was done, and the tryptophan emission data were corrected for fluorescence from RNA or buffer components. Control experiments and equations used to analyze the binding data are described in the supplemental information.

Size Exclusion Chromatography of Rsr-DrY RNA35—Rsr was purified as described above, on a Ni2+ affinity column. Rsr at 3 µM was mixed with a 5-fold molar excess of DrY RNA35 and dialyzed against buffer containing 25 mM phosphate, pH 7.0, 150 mM NaCl, and 2 mM dithiothreitol. Following dialysis, the sample was concentrated 10-fold and injected on a Superdex 200 analytical column equilibrated with the same buffer. All runs on the Superdex 200 column were performed at a flow rate of 1 ml/min. Identical elution profiles were obtained with the complex at concentrations as low as 1 µM. Different concentrations of protein alone (1–500 µM) and DrY RNA35 alone (2–590 µM) were injected in separate runs on the column. The column was calibrated with a mix of standard protein markers (catalog number 151-1901; Bio-Rad). A standard curve of elution volumes versus log10 of molecular weights was plotted and used to calculate the molecular weights of the Rsr, DrY RNA35, and Rsr-DrY RNA complex peaks. The ratio of absorbance at 260 nm to absorbance at 280 nm was measured for each fraction. Ratios greater than 1 were taken to be indicative of RNA. To detect protein in the fractions, the fractions were spotted on a nitrocellulose membrane, incubated successively with anti-His6 antibodies and anti-mouse IgG alkaline phosphatase, and stained with SigmaFASTTM 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium.

Electron Microscopy on Rsr-DrY RNA35 Complex—Rsr-DrY RNA35 (~3 µM in 25 mM phosphate, pH 7.0, 150 mM NaCl, 2 mM dithiothreitol) samples were prepared for electron microscopy according to Valentine et al. (24) and stained with 2% (w/v) aqueous uranyl acetate, pH 4.3. Duplicate grids were prepared similarly but stained with 2% (w/v) ammonium molybdate, pH 7.0. Samples were examined using a JEOL 1200EX transmission electron microscope operating at 100 kV. Micrographs were recorded at a calibrated magnification of x38,030. Selected micrographs were digitized using a Leafscan 45 microdensitometer at a 20-µm scan step corresponding to 5.26 Å/pixel at the specimen level. Particles were manually selected using the BOXER program from the EMAN (25) single particle analysis software package. Approximately 1,600 particles were chosen and further processed with IMAGIC-5 (26). Particle images were normalized, band pass-filtered, and subjected to reference-free classification. Nonredundant class averages were then chosen as references for multireference alignment (27). The process was iterated until stable classes were obtained based on visual inspection and intraclass variance. All class averages are displayed without imposing symmetry.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
REFERENCES
 
The Crystal Structure of D. radiodurans Rsr—The structure of Rsr was solved in a Ca2+-bound state, at a resolution of 1.89 Å. The final model was refined to an Rwork of 22.3% and an Rfree of 26.2% with excellent stereochemistry. X-ray data collection and structure refinement statistics are shown in Table 1 and supplemental Table 1. The refined model comprises residues 36–531 of the total 531 residues, one Ca2+ atom, and 400 molecules of water. Clear electron density was not visible for residues 1–35, which are presumably disordered in the crystals. Multiple sequence alignment of Ro from various organisms (supplemental Fig. 1) shows that this N-terminal stretch of 35 residues is highly variant in all Ro orthologs except for six conserved residues. In the crystal structure of Ro from X. laevis, this region is placed on the outer surface of the protein, away from the RNA binding sites, and forms two short beta-strands, which pack against a beta-strand of the C-terminal domain (7). The crystallographic asymmetric unit consists of one monomer of Rsr, and applying crystallographic symmetry does not generate any higher oligomer of Rsr with significant buried surface area. This, along with size exclusion studies performed on purified Rsr (see Fig. 4A), suggests that Rsr is a monomer.


Figure 1
View larger version (39K):
[in this window]
[in a new window]

  		FIGURE 1.
A, overall structure of D. radiodurans Rsr. Rsr is composed of two domains. The N-terminal ring domain (pink) consists of helices H1–H18 and one beta strand, beta1. The C-terminal vWFA domain (blue) consists of beta2–beta7, sandwiched between helices H19–H23. The calcium ion is shown as a red sphere. Conserved residues in the central cavity (Arg128, Lys172, Tyr173, Arg176, Glu247, Arg273, and Arg309) and on the outer surface of the ring domain (Arg116, Lys117, Arg151, Arg152, Lys187, His189, and Lys191) are shown as sticks. B, a coordination sphere around calcium (red sphere) bound to Rsr is shown. The calcium-coordinating residues (Ser369, Ser371, two waters, Glu464 (sticks)) along with Glu464' from the symmetry-related molecule (yellow stick) are shown. The second coordination sphere around calcium, consisting of MIDAS residues Asp367 and Asp462, which stabilize the water molecules, are shown as sticks.

 
The overall structure of Rsr resembles a ring spanning ~76 Å in diameter and ~32 Å in thickness (Fig. 1A). Rsr possesses two distinct domains: an N-terminal domain, including residues 36–351, and a C-terminal domain, including residues 352–531. The N-terminal domain comprises HEAT repeats made up of numerous short helices (H1–H18), which form a ringlike arrangement, and one beta-strand (beta1). The ringlike arrangement of helices encloses a central cavity with dimensions of 18–20 Å and is lined by charged residues from the loops connecting successive helices from H2 to H18. The majority of residues pointing into this cavity (e.g. Arg128, Lys172, Tyr173, Arg176, Glu247, Arg273, and Arg309) are strictly conserved among the Ro orthologs. In Ro from X. laevis, this site has been implicated in binding single-stranded RNAs with these conserved residues directly interacting with an extended conformation of a single strand of RNA (7). Notably, on the outer surface of the helical ring domain, there is another cluster of positively charged residues (Arg116, Lys117, Arg151, Arg152, Lys187, His189, and Lys191), some of which are highly conserved in Ro orthologs. In X. laevis Ro, this site has been implicated in Y RNA binding (7). The N-terminal HEAT repeat domain of Rsr bears structural similarity to the HEAT repeat domains present in nuclear transport proteins like the importin-beta family.

The C-terminal domain of Rsr resembles a vWFA domain similar to that found in integrins and extracellular matrix proteins (28). The vWFA domain is characterized by a beta-sheet sandwiched between {alpha}-helices. In Rsr, the beta-sheet is formed by six parallel beta-strands (beta2–beta7) and is sandwiched by parallel helices H19, H20, and H23 on one side and helices H21 and H22 on the other. Interestingly, in Rsr, this C-terminal beta-sheet interacts with the beta-strand (beta1) from the N-terminal domain to form a seven-stranded sheet.

The C-terminal domain in Rsr also includes a divalent cation binding MIDAS motif. This motif consists of residues D367XSXS...T...D462, all of which are strictly conserved in all Ro orthologs. In Rsr, a peak of strong positive electron density corresponding to the 6{sigma} contour level in a difference density map was proximal to a site surrounded by the MIDAS residues. A Ca2+ ion was fit into this site based on the fact that the crystallization condition contains 200 mM calcium acetate (Fig. 1B). The Ca2+ ion is coordinated by O{gamma} of Ser369, O{gamma} of Ser371, two structured water molecules (HOH 3 and HOH 6), O{epsilon}2 of Glu464, and O{epsilon}1 and O{epsilon}2 of Glu464 from a crystallographically symmetry-related molecule. The distances between the Ca2+ ion and the ligands range from 2.3 to 2.6 Å. Two residues of the MIDAS motif, Asp367 and Asp462, form the second coordination sphere around the calcium ion and bind the metal-coordinating waters.

This feature of a metal-coordinating ligand belonging to a symmetry-related molecule has been observed in other crystal forms of the MIDAS motifs in vWFA domains and integrin I domains (29, 30), where it has been proposed that in physiological conditions, the coordinating ligand may belong to a binding partner of the protein. Although a disulfide bond between Cys374 of symmetry-related molecules of Rsr was observed, this is probably an artifact of crystallization where a high concentration of protein was used. In the RNA-bound structure of X. laevis Ro, the metal site is occupied by a Mg2+ ion, and the metal coordination is completed by an acetate molecule from the crystallization solution. It has been suggested that the acetate may mimic an aspartate or glutamate residue from a yet unidentified binding partner of Ro (7).

In both integrins and the cell adhesion proteins, the vWFA domain is the site for protein-protein interactions with a conserved cation-binding MIDAS site acting as a structural "glue," bringing protein ligands closer to the vWFA domain. In many cases, metal binding at the MIDAS site is also associated with large conformational changes at a distant site across the domain (31). Based on the similarity in the fold of the C-terminal domain of Rsr to other vWFA domains, it is possible that this conserved domain of Rsr along with the metal site helps to mediate interactions between Rsr and other proteins. Also, since Rsr has been shown to bind small RNAs that are damaged or targeted for degradation (11), it is possible that binding partners of Rsr include nucleases, helicases, and other RNA processing enzymes.

Rsr Binds Y RNA with Low Nanomolar Affinity—To characterize the binding of Rsr to D. radiodurans Y RNA, a fluorescence-based assay was used. Rsr has eight tryptophans, five of which are strictly conserved in all Ro homologs. Also, the structure of X. laevis Ro bound to RNA (7) suggests that one or more tryptophans are involved in stacking interactions with RNA. The RNA used in this experiment is a 35-mer fragment of Y RNA (called DrY RNA35), representing the Ro binding region (supplemental Fig. 2C). By monitoring the quenching of intrinsic tryptophan fluorescence upon titrating RNA, binding isotherms were obtained (Fig. 2). The minimum detectable amount of Rsr was found to be 25 nM; hence, the assay was done at a 50 nM concentration of Rsr. Relevant corrections for RNA fluorescence, photobleaching, and inner filter effect of RNA have been described under "Experimental Procedures." Using this assay, it was found that Rsr binds DrY RNA35 with a Kd of at least 16 ± 3 nM affinity and a stoichiometry of 1:1. Since this assay could not be performed at a lower concentration of Rsr because of the detection limit, the calculated Kd probably represents the apparent Kd. As expected for reactions involving stacking interactions between protein side chains and RNA bases, the reaction was markedly slower in the presence of higher salt concentrations (data not shown). Dissociation of RNAs from Ro at higher ionic strength has also been shown for human Ro ribonucleoprotein particles (32). Binding of Rsr to DrY RNA35 is comparable with that reported for X. laevis Ro and Y RNA interactions. Using an electrophoretic mobility shift assay, it has been shown that X. laevis Ro binds XlY RNA with ~5.2 nM affinity (8).


Figure 2
View larger version (29K):
[in this window]
[in a new window]

  		FIGURE 2.
Quenching of intrinsic tryptophan fluorescence of Rsr by DrY RNA35. Shown is a plot of fractional saturation ({theta}) versus DrY RNA35 concentration (µM) obtained from the emission at 335 nm. Error bars, S.D. of three independent samples. The equilibrium dissociation constant Kd is ~16 ± 3 nM, and the stoichiometry of binding is 1:1. Inset, fluorescence emission spectra (315–400 nm) obtained upon excitation at 295 nm. Quenching of fluorescence occurs with increasing concentrations of Y RNA.

 
Rsr Shows Large Domain Movements When Compared with Apo- or RNA-bound Ro from X. laevis—Consistent with our finding that molecular replacement using X. laevis Ro (Protein Data Bank codes 1YVP and 1YVR) with the diffraction data from Rsr was unsuccessful, the comparison showed very obvious structural differences between Rsr and the RNA-bound structure of X. laevis Ro (Fig. 3). The overall shape of the two proteins is significantly different, with Rsr forming a more rounded ring-shaped N-terminal domain (Fig. 3A) compared with the elliptical N-terminal domain of X. laevis Ro (Fig. 3B). The surface rendering of Rsr when colored by electrostatic potential (Fig. 3, A and B) shows two distinct patches of positive residues, one on the outer face of the ring domain corresponding to the Y RNA binding site of X. laevis Ro and the other lining the central cavity, which corresponds to the single-stranded RNA binding region in X. laevis Ro. Although the residues in these two sites are conserved, in X. laevis Ro, the distribution of the positive surface is more continuous between the two sites, unlike the separate clusters seen in Rsr. It is possible that these differences are due to the dissimilar conformations adopted by Rsr versus X. laevis Ro or even a result of alterations in the amino acid sequences between the two proteins in the regions apart from the RNA binding sites. Also, it is possible that the difference in charge distribution accounts for the variations in RNA sequences that bind Rsr versus X. laevis Ro. In both proteins, the vWFA domain is mostly negatively charged.

Superimposition of the entire proteins as rigid bodies, using COOT (33), demonstrated that the relative position of the two domains with respect to one another was quite different, with root mean square differences based on C{alpha} positions equal to 3.8 Å. The superimposition, using the Y RNA binding regions as reference, is shown in Fig. 3C. A much better fit was obtained when the domains were treated as two rigid bodies and superimposed on the corresponding domains from X. laevis Ro with root mean square differences of 2.8 Å for the N-terminal (Fig. 3D) and 2.1 Å for the C-terminal domains (supplemental Fig. 3B).

Superimposition of the vWFA domains of Rsr and X. laevis Ro (supplemental Fig. 3) shows no pronounced changes within the domain. One difference is in the loop region between residues Gln373 and Pro385. This loop is close to the metal binding site, and the changes seen could be a result of the presence of the larger Ca2+ ion in the metal site as compared with a Mg2+ ion.

However, superimposition of the N-terminal domains showed marked differences between the structures at the interface of the N- and C-terminal domains (Fig. 3, C and D). In Rsr, the stretch of residues from Gly52 to Ala58 that forms a beta-strand (beta1) is stabilized as a part of the seven-stranded beta-sheet from the vWFA domain (Fig. 3D), whereas the corresponding region in X. laevis Ro forms a loop. In X. laevis Ro, a conserved tyrosine residue (corresponding to Tyr55 in Rsr) in this loop has been shown to form the second coordination sphere around the MIDAS metal, forming a hydrogen bond with one of the metal-coordinating water molecules. In Rsr, this region is far from the metal site, with Tyr55 pointing away from the beta-sheet and forming hydrogen bonds with a water molecule and Glu498 (Fig. 3D, supplemental Fig. 3B).

Comparison of the N-terminal domains also shows a much larger central cavity in Rsr, with the four-helix bundle formed by H15–H18 being placed >15 Å away from helices H3 and H4 as compared with the structure of X. laevis Ro, where this distance is ~5 Å (Fig. 3D). This four-helix bundle undergoes >34° rotation away from the central cavity as compared with X. laevis Ro. The different placement of these helices in Rsr could potentially modulate RNA binding by influencing the size of the central cavity. Considering that the {alpha}-helices H17 and H18 are in close contact with the C-terminal domain, changes in the placement of these helices may get transmitted to the C-terminal vWFA domain. It is possible that this structure of Rsr with the four-helix bundle formed by H15–H18 placed away from the central cavity may represent an "open" state poised to allow entry or release of the RNA substrates from the central cavity. In the presence of the RNA substrate as seen in the X. laevis Ro bound to the single-stranded RNA (ssRNA) (7), the four-helix bundle may swing in toward the cavity, accompanied by the movement of the vWFA domain, which may further affect the downstream events of RNA processing.


Figure 3
View larger version (76K):
[in this window]
[in a new window]

  		FIGURE 3.
Comparison of structures of Rsr and X. laevis Ro. A, the electrostatic surface potential rendering of Rsr shows two distinct positively charged surfaces (blue) corresponding to the residues shown as sticks in Fig. 2A. The vWFA-like domain is negatively charged (red). B, the surface of X. laevis Ro shows a continuous patch of positive charge on the ring domain and negatively charged vWFA domain. C, superimposition of Rsr (pink and blue) and RNA-bound X. laevis Ro (yellow) using helical domains as reference shows displacement of vWFA domain and helices H15–H18. The direction of displacement is shown as curved arrows. Y RNA fragment (green), ssRNA (navy), and Mg2+ ion (magenta) bound to the X. laevis structure are shown. D, close-up of superimposed helical domains of Rsr (pink) and X. laevis Ro (yellow) shows that H15–H18 in Rsr are 15 Å away from H3-H4 (black line), whereas in X. laevis Ro, this distance is ~5Å.The beta1 region of Rsr forms an alternate loop conformation in X. laevis Ro (boxed region). The conserved tyrosines (Tyr55 in Rsr, Tyr47 in X. laevis Ro) are shown as sticks. Conserved glycine residues at positions of structural flexibility are shown as cyan spheres.

 
To understand the basis of such flexibility in the structure of Rsr, a detailed analysis of its sequence was done. Rsr has a distribution of highly conserved glycine residues close to the regions that involve structural changes (Fig. 3D). One example is the Gly51–Gly52 motif preceding the beta-strand (beta1), which is present in an alternate loop conformation in X. laevis Ro. Another Gly144–Gly145 motif precedes the region connecting helices H6 and H7. This region corresponds to a loop in X. laevis Ro that becomes ordered upon Y RNA binding (7). Especially notable are Gly267 (conserved proline or glycine at this position), which precedes the four-helix bundle (H15–H18), and the Gly-X-X-Gly-X-X-Gly (positions 276–282) motif that is present on the C terminus of helix H15. Glycine-rich motifs in other proteins have been shown to be the site of structural flexibility (34, 35). The positions of these glycines in Rsr possibly represent hinge regions that allow helices H15–H18 to undergo displacement away from the central cavity. The N-terminal domain of Rsr is composed of HEAT repeats very similar to that found in some nuclear transport proteins like importin-beta. Various crystallographic, biochemical, and molecular dynamics simulation studies strongly support a "spring-loaded" model for ligand binding, which involves the movement of some helices with respect to the others with a critical role for glycine and proline residues as hinge residues (36). Our proposal that the structure of Rsr represents the "open" form and the conformational differences we see at the RNA binding central cavity agree well with this previously described mechanism for ligand binding by HEAT repeat domains in nuclear transport proteins. To confirm this prediction, extensive mutational analysis of these hinge glycine residues is required, along with reliable assays that could help measure the flexibility of specific regions of the protein in solution. Fluorescence-based assays to measure conformational changes are sensitive, but the ability to introduce probes into desired locations within the protein is crucial for the outcome of these experiments and often the most challenging part. In Rsr, the presence of multiple reactive groups along with the presence of an extensive and buried interface adds yet another level of complexity. Attempts to study RNA binding by the N-terminal ring domain alone in the absence of the vWFA domain have been unsuccessful. Expression of the N terminus domain in E. coli results in insoluble aggregates of the protein that do not respond to various refolding protocols.3

Potential Implications of the Rsr Structure on Its Interactions with Different RNA Substrates—To understand the features of Rsr that allow its interaction with different RNA substrates, we compared the structure of Rsr with the X. laevis Ro bound to Y RNA and ssRNA (Protein Data Bank code 1YVP). At the Y RNA binding site, residues that interact with the nucleotides G1 to A10 in the first strand of X. laevis Y RNA are structurally conserved between the two proteins. His189, Tyr179, Arg148, Arg151, and Tyr146 in Rsr are present in similar conformations as in X. laevis Ro and possibly interact with Y RNA in the same manner as seen in X. laevis Ro. Some residues that interact with Y RNA in X. laevis Ro are not conserved, like residues Lys187 and Asp141. Although the Y RNA substrates of Ro from different species are predicted to form identical secondary structures (36), there are subtle differences in their sequence even within the conserved Ro binding region (supplemental Fig. 2). It is possible that the nonconserved positions at the Y RNA binding site of Ro account for the diversity in the Ro binding regions of Y RNAs from different species.

Ro in X. laevis has also been shown to bind ssRNAs at the central cavity (7, 8). A comparison of this site with Rsr shows that most residues that interact with ssRNA in X. laevis Ro are conserved (Fig. 3D). Mutagenesis in X. laevis Ro has identified the conserved residues corresponding to Lys172 and Arg176 to be important in binding ssRNA (7). These, along with residues Thr123 and Arg174, interact through hydrogen bonds with the phosphate backbone of the ssRNA. Significantly, residues of X. laevis Ro corresponding to the conserved residues Arg273, Arg296, and Arg309 interact with the ssRNA bases but are not critical for binding the ssRNA. In Rsr, these residues lie on the helices H15 and H16, which are displaced away to form a significantly larger central cavity. This suggests that Rsr may bind a variety of RNA substrates in the central cavity by changing the size of the cavity and using the conserved residues in the flexible helices to interact with these RNAs, whereas the primary residues forming the primary, most stabilizing interactions with the RNA come from the more rigid regions of the protein.

Previous studies in various organisms (5, 8, 12) have shown that besides Y RNAs, Ro also binds snU2 RNAs and pre-5 S rRNAs. Also, in D. radiodurans, at least four different RNAs that associate with Rsr accumulate post-UV irradiation, of which only one resembles a Y RNA (11). It is possible that recognition of such a variety of RNAs by Ro is a result of structural adaptation of the protein, made possible by a series of highly conserved motifs that serve as flexible joints in the protein backbone.

The structure of Rsr reveals many conserved glycine-rich motifs, and the positions of these motifs coincide with regions of Rsr that show a marked difference from the previously described structure of Ro from X. laevis. These large structural differences are most pronounced in the region surrounding the central cavity that has been shown to bind ssRNA in X. laevis. The outward movement of helices at the interface of the two domains results in a significantly larger central cavity. Other proteins that possess a repeated antiparallel helix-turn-helix or HEAT repeat domain similar to the N-terminal domain of Rsr have been shown to be highly flexible, and it has been suggested that flexibility may be an intrinsic feature of such domains (37). Also, one of the major changes seen in the Rsr structure is the switch in the conserved beta1 region between a beta-strand and the loop conformation seen in X. laevis Ro. Based on the observation that a conserved tyrosine residue from this region forms the second coordination site around the MIDAS metal in the vWFA domain of X. laevis Ro but is placed away from the site in Rsr, we speculate that the vWFA domain may play a role in mediating structural changes via this region, all the way to the RNA-binding central cavity.


Figure 4
View larger version (33K):
[in this window]
[in a new window]

  		FIGURE 4.
A, size exclusion profiles from a Superdex-200 column. A plot of absorbance at 280 nm versus elution volume shows the elution characteristics of Rsr alone (blue), DrY RNA35 alone (red), the Rsr-DrY RNA35 complex and excess unbound DrY RNA35 (green), and Rsr-DrY RNA35 complex and excess unbound DrY RNA35 in buffer supplemented with 10 mM metal chelator EDTA (yellow). A black asterisk marks the peak corresponding to multimers of Rsr-DrY RNA35 complex. The calibration curve is shown as an inset. B, electron microscopic analysis of single Rsr-DrY RNA35 particles. Shown are representative class averages displaying no symmetry (top row) and 2-fold (middle row) and 3- and 4-fold (bottom row, left to right) rotational symmetry. Scale bar, 20 nm.

 
Rsr-DrY RNA35 Complex Forms Large Multimers in Solution—To further characterize the Rsr-DrY RNA35 complex, we performed size exclusion chromatography using a Superdex 200 column. Rsr forms multimers upon binding DrY RNA35. The multimers are of an average size of >700 kDa (Fig. 4A) and are composed of ~12 molecules of Rsr-DrY RNA. Consistent with the tryptophan fluorescence assays described above, the ratio of absorbance 260/280 nm of the multimers corresponds to a 1:1 complex of Rsr-Y RNA. Although the multimer formation occurs in the presence of both DrY RNA and DrY RNA35, Rsr alone and DrY RNA35 alone even over a wide range of concentrations (1 to ~500 µM) did not show formation of multimers and eluted at the expected monomeric sizes (representative curves are shown in Fig. 4A). Similar experiments performed in the presence of saturating amounts of magnesium ion (data not shown) or with an excess of metal chelator EDTA (Fig. 4A) showed identical multimer formation, suggesting that the formation of multimers is independent of the presence of metal ions at the vWFA domain of Rsr. Our observations suggest that although the multimers are stable in solution, the equilibrium shifts more toward smaller oligomers (e.g. a 1:1 complex) when the size exclusion column is run at flow rates lower than 0.5 ml/min. Dissociation of the multimer also occurs at higher concentrations of salt (data not shown), which is presumably due to disruption of ionic interactions between Rsr and DrY RNA as seen in the fluorescence assays.

To confirm that the formation of the 700-kDa multimer is based on specific interactions and not a consequence of nonspecific aggregation, the fractions corresponding to multimers were analyzed by electron microscopy. Electron micrographs confirmed the presence of particles displaying symmetrical patterns in projection. Representative class averages obtained from 1600 particles (Fig. 4B) suggest 2–4-fold rotational symmetry residing within the multimers measuring up to 180 Å along the long axis. In addition, some classes did not reveal any rotational symmetry in projection. The observed heterogeneity within the Rsr-DrY RNA population has so far hampered three-dimensional reconstruction efforts.

Since the formation of multimers occurs in the presence of both the full-length Y RNA and the conserved region of Y RNA and the fluorescence assays suggest a 1:1 complex between Rsr and Y RNA, it is likely that the multimer formation does not involve additional RNA-RNA or RNA-protein interactions. This suggests that upon Y RNA binding, Rsr adopts an alternate conformation, which promotes oligomerization. This feature of forming large, ordered multimers in the presence of Y RNA has not been reported yet for other Ro orthologs. Formation of higher order oligomers upon binding single-stranded DNA has been observed for the human RecQ helicase (38). The RecQ-single-stranded DNA complex forms oligomeric ringlike structures surrounding a central pore, the site where annealing of complementary strands of DNA could take place. This higher order oligomerization in the presence of single-stranded DNA has been shown to regulate the enzyme, switching the activity from DNA unwinding to DNA strand annealing. Although the RNA binding properties of Ro have been extensively characterized (8), it is not known what events take place after Ro binds Y RNA and what triggers those events to take place. It is possible that formation of these multimers acts as a trigger to set off the downstream processing events, which in the case of D. radiodurans are likely to be the removal of UV-damaged RNA.

In summary, we show that Rsr possesses a high degree of flexibility at the interface of its two domains. The rearrangement of helices at this interface results in alterations in the size of the central RNA binding cavity. This high degree of flexibility may be possible due to conserved glycine-rich motifs and may possibly be required by Ro orthologs to structurally adapt to different RNA substrates to be bound at the central cavity and for facilitating the entry and release of RNA substrates. Identification of different substrates that bind at the central cavity of Rsr is in progress and would greatly improve our understanding of the mode of RNA binding by Ro as well as the function of Ro in prokaryotes. Although little is known about the downstream events following Ro-Y RNA binding, our finding that Rsr interacts with Y RNA to form large multimers suggests one possible trigger for the recognition of this complex by downstream processing events.


   FOOTNOTES
 
The atomic coordinates and structure factors (code 2NVO) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).

* This work was supported by Robert A. Welch Foundation Grant A1295 (to J. C. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

Formula The on-line version of this article (available at http://www.jbc.org) contains supplemental Table 1 and Figs. 1–3. Back

1 To whom correspondence should be addressed: Dept. of Biochemistry and Biophysics, Texas A&M University, College Station, TX 77843-2128. Tel.: 979-862-7637; Fax: 979-845-9274; E-mail: sacchett{at}tamu.edu.

2 The abbreviations used are: vWFA, collagen-binding A domain of von Willebrand factor; MIDAS, metal ion-dependent adhesion site; DrY RNA, D. radiodurans Y RNA; ssRNA, single-stranded RNA. Back

3 A. Ramesh and J. C. Sacchettini, unpublished data. Back


   ACKNOWLEDGMENTS
 
We thank Peter Cornish, Suzanne Stammler, and Dr. David P. Giedroc for very helpful discussions.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 REFERENCES
 

  1. von Muhlen, C. A., and Tan, E. M. (1995) Semin. Arthritis Rheum. 24, 323-358[CrossRef][Medline] [Order article via Infotrieve]
  2. Wolin, S. L., and Cedervall, T. (2002) Annu. Rev. Biochem. 71, 375-403[CrossRef][Medline] [Order article via Infotrieve]
  3. Xue, D., Shi, H., Smith, J. D., Chen, X., Noe, D. A., Cedervall, T., Yang, D. D., Eynon, E., Brash, D. E., Kashgarian, M., Flavell, R. A., and Wolin, S. L. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 7503-7508[Abstract/Free Full Text]
  4. Shi, H., O'Brien, C. A., Van Horn, D. J., and Wolin, S. L. (1996) RNA 2, 769-784[Abstract]
  5. O'Brien, C. A., and Wolin, S. L. (1994) Genes Dev. 8, 2891-2903[Abstract/Free Full Text]
  6. Andrade, M. A., and Bork, P. (1995) Nat. Genet. 11, 115-116[CrossRef][Medline] [Order article via Infotrieve]
  7. Stein, A. J., Fuchs, G., Fu, C., Wolin, S. L., and Reinisch, K. M. (2005) Cell 121, 529-539[CrossRef][Medline] [Order article via Infotrieve]
  8. Fuchs, G., Stein, A. J., Fu, C., Reinisch, K. M., and Wolin, S. L. (2006) Nat. Struct. Mol. Biol. 13, 1002-1009[CrossRef][Medline] [Order article via Infotrieve]
  9. Chen, X., and Wolin, S. L. (2004) J. Mol. Med. 82, 232-239[CrossRef][Medline] [Order article via Infotrieve]
  10. Cox, M. M., and Battista, J. R. (2005) Nat. Rev. Microbiol 3, 882-892[CrossRef][Medline] [Order article via Infotrieve]
  11. Chen, X., Quinn, A. M., and Wolin, S. L. (2000) Genes Dev. 14, 777-782[Abstract/Free Full Text]
  12. Chen, X., Smith, J. D., Shi, H., Yang, D. D., Flavell, R. A., and Wolin, S. L. (2003) Curr. Biol. 13, 2206-2211[CrossRef][Medline] [Order article via Infotrieve]
  13. Tebbe, B., and Orfanos, C. E. (1997) Lupus 6, 96-104[Abstract/Free Full Text]
  14. Otwinowski, Z., and Minor, W. (1997) Methods Enzymol. 276, 307-326
  15. Adams, P. D., Gopal, K., Grosse-Kunstleve, R. W., Hung, L. W., Ioerger, T. R., McCoy, A. J., Moriarty, N. W., Pai, R. K., Read, R. J., Romo, T. D., Sacchettini, J. C., Sauter, N. K., Storoni, L. C., and Terwilliger, T. C. (2004) J. Synchrotron Radiat. 11, 53-55[Medline] [Order article via Infotrieve]
  16. Cowtan, K. D., and Main, P. (1996) Acta Crystallogr. Sect. D Biol. Crystallogr. 52, 43-48[CrossRef][Medline] [Order article via Infotrieve]
  17. Ioerger, T. R., and Sacchettini, J. C. (2003) Methods Enzymol. 374, 244-270[Medline] [Order article via Infotrieve]
  18. McRee, D. E. (1999) J. Struct. Biol. 125, 156-165[CrossRef][Medline] [Order article via Infotrieve]
  19. Reddy, V., Swanson, S. M., Segelke, B., Kantardjieff, K. A., Sacchettini, J. C., and Rupp, B. (2003) Acta Crystallogr. Sect. D Biol. Crystallogr. 59, 2200-2210[CrossRef][Medline] [Order article via Infotrieve]
  20. Baker, N. A., Sept, D., Joseph, S., Holst, M. J., and McCammon, J. A. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 10037-10041[Abstract/Free Full Text]
  21. Sanner, M. F., Olson, A. J., and Spehner, J. C. (1996) Biopolymers 38, 305-320[CrossRef][Medline] [Order article via Infotrieve]
  22. Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C., and Ferrin, T. E. (2004) J. Comput. Chem. 25, 1605-1612[CrossRef][Medline] [Order article via Infotrieve]
  23. Zuker, M. (2003) Nucleic Acids Res. 31, 3406-3415[Abstract/Free Full Text]
  24. Valentine, R. C., Shapiro, B. M., and Stadtman, E. R. (1968) Biochemistry 7, 2143-2152[CrossRef][Medline] [Order article via Infotrieve]
  25. Ludtke, S. J., Baldwin, P. R., and Chiu, W. (1999) J. Struct. Biol. 128, 82-97[CrossRef][Medline] [Order article via Infotrieve]
  26. van Heel, M., Harauz, G., Orlova, E. V., Schmidt, R., and Schatz, M. (1996) J. Struct. Biol. 116, 17-24[CrossRef][Medline] [Order article via Infotrieve]
  27. Van Heel, M. (1989) Optik 82, 114-126
  28. Whittaker, C. A., and Hynes, R. O. (2002) Mol. Biol. Cell 13, 3369-3387[Abstract/Free Full Text]
  29. Lacy, D. B., Wigelsworth, D. J., Scobie, H. M., Young, J. A., and Collier, R. J. (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 6367-6372[Abstract/Free Full Text]
  30. Lee, J. O., Rieu, P., Arnaout, M. A., and Liddington, R. (1995) Cell 80, 631-638[CrossRef][Medline] [Order article via Infotrieve]
  31. Shimaoka, M., Takagi, J., and Springer, T. A. (2002) Annu. Rev. Biophys. Biomol. Struct. 31, 485-516[CrossRef][Medline] [Order article via Infotrieve]
  32. Boire, G., and Craft, J. (1990) J. Clin. Invest. 85, 1182-1190[Medline] [Order article via Infotrieve]
  33. Krissinel, E., and Henrick, K. (2004) Acta Crystallogr. Sect. D Biol. Crystallogr. 60, 2256-2268[CrossRef][Medline] [Order article via Infotrieve]
  34. Jacob, J., Duclohier, H., and Cafiso, D. S. (1999) Biophys. J. 76, 1367-1376[Medline] [Order article via Infotrieve]
  35. Rosenhouse-Dantsker, A., and Logothetis, D. E. (2006) Biophys. J. 91, 2860-2873[CrossRef][Medline] [Order article via Infotrieve]
  36. Farris, A. D., Koelsch, G., Pruijn, G. J., van Venrooij, W. J., and Harley, J. B. (1999) Nucleic Acids Res. 27, 1070-1078[Abstract/Free Full Text]
  37. Conti, E., Muller, C. W., and Stewart, M. (2006) Curr. Opin. Struct. Biol. 16, 237-244[CrossRef][Medline] [Order article via Infotrieve]
  38. Muzzolini, L., Beuron, F., Patwardhan, A., Popuri, V., Cui, S., Niccolini, B., Rappas, M., Freemont, P. S., and Vindigni, A. (2007) PLoS Biol. 5, e20[CrossRef][Medline] [Order article via Infotrieve]

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?



This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
282/20/14960    most recent
M611163200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ramesh, A.
Right arrow Articles by Sacchettini, J. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ramesh, A.
Right arrow Articles by Sacchettini, J. C