Sizes of Long RNA Molecules Are Determined by the Branching Patterns of Their Secondary Structures.

Long RNA molecules are at the core of gene regulation across all kingdoms of life, while also serving as genomes in RNA viruses. Few studies have addressed the basic physical properties of long single-stranded RNAs. Long RNAs with nonrepeating sequences usually adopt highly ramified secondary structures and are better described as branched polymers. To test whether a branched polymer model can estimate the overall sizes of large RNAs, we employed fluorescence correlation spectroscopy to examine the hydrodynamic radii of a broad spectrum of biologically important RNAs, ranging from viral genomes to long noncoding regulatory RNAs. The relative sizes of long RNAs measured at low ionic strength correspond well to those predicted by two theoretical approaches that treat the effective branching associated with secondary structure formation-one employing the Kramers theorem for calculating radii of gyration, and the other featuring the metric of maximum ladder distance. Upon addition of multivalent cations, most RNAs are found to be compacted as compared with their original, low ionic-strength sizes. These results suggest that sizes of long RNA molecules are determined by the branching pattern of their secondary structures. We also experimentally validate the proposed computational approaches for estimating hydrodynamic radii of single-stranded RNAs, which use generic RNA structure prediction tools and thus can be universally applied to a wide range of long RNAs.

A Prüfer-Sequence Based Algorithm for Calculating the Size of Ideal Randomly Branched Polymers.

Branched polymers can be represented as tree graphs. A one-to-one correspondence exists between a tree graph comprised of N labeled vertices and a sequence of N - 2 integers, known as the Prüfer sequence. Permutations of this sequence yield sequences corresponding to tree graphs with the same vertex-degree distribution but (generally) different branching patterns. Repeatedly shuffling the Prüfer sequence we have generated large ensembles of random tree graphs, all with the same degree distributions. We also present and apply an efficient algorithm to determine graph distances directly from their Prüfer sequences. From the (Prüfer sequence derived) graph distances, 3D size metrics, e.g., the polymer's radius of gyration, Rg, and average end-to-end distance, were then calculated using several different theoretical approaches. Applying our method to ideal randomly branched polymers of different vertex-degree distributions, all their 3D size measures are found to obey the usual N(1/4) scaling law. Among the branched polymers analyzed are RNA molecules comprised of equal proportions of the four-randomly distributed-nucleotides. Prior to Prüfer shuffling, the vertices of their representative tree graphs, these "random-sequence" RNAs exhibit an Rg ∼ N(1/3) scaling.

Role of RNA Branchedness in the Competition for Viral Capsid Proteins.

To optimize binding-and packaging-by their capsid proteins (CP), single-stranded (ss) RNA viral genomes often have local secondary/tertiary structures with high CP affinity, with these "packaging signals" serving as heterogeneous nucleation sites for the formation of capsids. Under typical in vitro self-assembly conditions, however, and in particular for the case of many ssRNA viruses whose CP have cationic N-termini, the adsorption of CP by RNA is nonspecific because the CP concentration exceeds the largest dissociation constant for CP-RNA binding. Consequently, the RNA is saturated by bound protein before lateral interactions between CP drive the homogeneous nucleation of capsids. But, before capsids are formed, the binding of protein remains reversible and introduction of another RNA species-with a different length and/or sequence-is found experimentally to result in significant redistribution of protein. Here we argue that, for a given RNA mass, the sequence with the highest affinity for protein is the one with the most compact secondary structure arising from self-complementarity; similarly, a long RNA steals protein from an equal mass of shorter ones. In both cases, it is the lateral attractions between bound proteins that determines the relative CP affinities of the RNA templates, even though the individual binding sites are identical. We demonstrate this with Monte Carlo simulations, generalizing the Rosenbluth method for excluded-volume polymers to include branching of the polymers and their reversible binding by protein.

Viral RNAs are unusually compact.

A majority of viruses are composed of long single-stranded genomic RNA molecules encapsulated by protein shells with diameters of just a few tens of nanometers. We examine the extent to which these viral RNAs have evolved to be physically compact molecules to facilitate encapsulation. Measurements of equal-length viral, non-viral, coding and non-coding RNAs show viral RNAs to have among the smallest sizes in solution, i.e., the highest gel-electrophoretic mobilities and the smallest hydrodynamic radii. Using graph-theoretical analyses we demonstrate that their sizes correlate with the compactness of branching patterns in predicted secondary structure ensembles. The density of branching is determined by the number and relative positions of 3-helix junctions, and is highly sensitive to the presence of rare higher-order junctions with 4 or more helices. Compact branching arises from a preponderance of base pairing between nucleotides close to each other in the primary sequence. The density of branching represents a degree of freedom optimized by viral RNA genomes in response to the evolutionary pressure to be packaged reliably. Several families of viruses are analyzed to delineate the effects of capsid geometry, size and charge stabilization on the selective pressure for RNA compactness. Compact branching has important implications for RNA folding and viral assembly.

Characterization of Viral Capsid Protein Self-Assembly around Short Single-Stranded RNA.

For many viruses, the packaging of a single-stranded RNA (ss-RNA) genome is spontaneous, driven by capsid protein-capsid protein (CP) and CP-RNA interactions. Furthermore, for some multipartite ss-RNA viruses, copackaging of two or more RNA molecules is a common strategy. Here we focus on RNA copackaging in vitro by using cowpea chlorotic mottle virus (CCMV) CP and an RNA molecule that is short (500 nucleotides (nts)) compared to the lengths (≈3000 nts) packaged in wild-type virions. We show that the degree of cooperativity of virus assembly depends not only on the relative strength of the CP-CP and CP-RNA interactions but also on the RNA being short: a 500-nt RNA molecule cannot form a capsid by itself, so its packaging requires the aggregation of multiple CP-RNA complexes. By using fluorescence correlation spectroscopy (FCS), we show that at neutral pH and sufficiently low concentrations RNA and CP form complexes that are smaller than the wild-type capsid and that four 500-nt RNAs are packaged into virus-like particles (VLPs) only upon lowering the pH. Further, a variety of bulk-solution techniques confirm that fully ordered VLPs are formed only upon acidification. On the basis of these results, we argue that the observed high degree of cooperativity involves equilibrium between multiple CP/RNA complexes.

Reply to the comment by S. Harvey on "entropy, energy, and bending of DNA in viral capsids".

The comment by Stephen Harvey in this issue of the Biophysical Journal concludes with two statements regarding my recent letter about DNA packaging into viral capsids. Harvey agrees with my interpretation of the origin of the large confinement entropy predicted by the molecular-dynamics simulations of his group, and its sensitive dependence on the molecular parameters of their wormlike chain model of double-stranded DNA. On the other hand, he doubts my assertion that the confinement entropy is already included in the interstrand repulsion free energy derived from osmotic stress measurements, which constitutes the major contribution to the packaging free energy used in recent continuum theories of this process. Harvey suggests instead that the confinement entropy should be added to this free energy as a separate term (using, for instance, the method described in my letter). I will argue that this addition is redundant, and, in a brief discussion of continuum theories, will also discuss his comments as relates to the work of other researchers.

Entropy, energy, and bending of DNA in viral capsids.

Inspired by novel single-molecule and bulk solution measurements, the physics underlying the forces and pressures involved in DNA packaging into bacteriophage capsids became the focus of numerous recent theoretical models. These fall into two general categories: Continuum-elastic theories (CT), and simulation studies-mostly of the molecular dynamics (MD) genre. Both types of models account for the dependence of the force, and hence the packaging free energy (ΔF), on the loaded DNA length, but differ markedly in interpreting their origin. While DNA confinement entropy is a dominant contribution to ΔF in the MD simulations, in the CT theories this role is fulfilled by interstrand repulsion, and there is no explicit entropy term. The goal of this letter is to resolve this apparent contradiction, elucidate the origin of the entropic term in the MD simulations, and point out its tacit presence in the CT treatments.

Theory and simulations of adhesion receptor dimerization on membrane surfaces.

The equilibrium constants of trans and cis dimerization of membrane bound (2D) and freely moving (3D) adhesion receptors are expressed and compared using elementary statistical-thermodynamics. Both processes are mediated by the binding of extracellular subdomains whose range of motion in the 2D environment is reduced upon dimerization, defining a thin reaction shell where dimer formation and dissociation take place. We show that the ratio between the 2D and 3D equilibrium constants can be expressed as a product of individual factors describing, respectively, the spatial ranges of motions of the adhesive domains, and their rotational freedom within the reaction shell. The results predicted by the theory are compared to those obtained from a novel, to our knowledge, dynamical simulations methodology, whereby pairs of receptors perform realistic translational, internal, and rotational motions in 2D and 3D. We use cadherins as our model system. The theory and simulations explain how the strength of cis and trans interactions of adhesive receptors are affected both by their presence in the constrained intermembrane space and by the 2D environment of membrane surfaces. Our work provides fundamental insights as to the mechanism of lateral clustering of adhesion receptors after cell-cell contact and, more generally, to the formation of lateral microclusters of proteins on cell surfaces.

The size of RNA as an ideal branched polymer.

Because of the branching arising from partial self-complementarity, long single-stranded (ss) RNA molecules are significantly more compact than linear arrangements (e.g., denatured states) of the same sequence of monomers. To elucidate the dependence of compactness on the nature and extent of branching, we represent ssRNA secondary structures as tree graphs which we treat as ideal branched polymers, and use a theorem of Kramers for evaluating their root-mean-square radius of gyration, ̂R(g)=√R(g)(2). We consider two sets of sequences--random and viral--with nucleotide sequence lengths (N) ranging from 100 to 10,000. The RNAs of icosahedral viruses are shown to be more compact (i.e., to have smaller ̂R(g)) than the random RNAs. For the random sequences we find that ̂R(g) varies as N(1/3). These results are contrasted with the scaling of ̂R(g) for ideal randomly branched polymers (N(1/4)), and with that from recent modeling of (relatively short, N ≤ 161) RNA tertiary structures (N(2/5)).

Transforming binding affinities from three dimensions to two with application to cadherin clustering.

Membrane-bound receptors often form large assemblies resulting from binding to soluble ligands, cell-surface molecules on other cells and extracellular matrix proteins. For example, the association of membrane proteins with proteins on different cells (trans-interactions) can drive the oligomerization of proteins on the same cell (cis-interactions). A central problem in understanding the molecular basis of such phenomena is that equilibrium constants are generally measured in three-dimensional solution and are thus difficult to relate to the two-dimensional environment of a membrane surface. Here we present a theoretical treatment that converts three-dimensional affinities to two dimensions, accounting directly for the structure and dynamics of the membrane-bound molecules. Using a multiscale simulation approach, we apply the theory to explain the formation of ordered, junction-like clusters by classical cadherin adhesion proteins. The approach features atomic-scale molecular dynamics simulations to determine interdomain flexibility, Monte Carlo simulations of multidomain motion and lattice simulations of junction formation. A finding of general relevance is that changes in interdomain motion on trans-binding have a crucial role in driving the lateral, cis-, clustering of adhesion receptors.

A sequential folding model predicts length-independent secondary structure properties of long ssRNA.

We introduce a simple model for folding random-sequence RNA molecules, arguing that it provides a direct route to predicting and rationalizing several average properties of RNA secondary structures. The first folding step involves identifying the longest possible duplex, thereby dividing the molecule into a pair of daughter loops. Successive steps involve identifying similarly the longest duplex in each new pair of daughter loops, with this process proceeding sequentially until the loops are too small for a viable duplex to form. Approximate analytical solutions are found for the average fraction of paired bases, the average duplex length, and the average loop size, all of which are shown to be independent of sequence length for long enough molecules. Numerical solutions to the model provide estimates for these average secondary structure properties that agree well with those obtained from more sophisticated folding algorithms. We also use the model to derive the asymptotic power law for the dependence of the maximum ladder distance on chain length.

The extracellular architecture of adherens junctions revealed by crystal structures of type I cadherins.

Adherens junctions, which play a central role in intercellular adhesion, comprise clusters of type I classical cadherins that bind via extracellular domains extended from opposing cell surfaces. We show that a molecular layer seen in crystal structures of E- and N-cadherin ectodomains reported here and in a previous C-cadherin structure corresponds to the extracellular architecture of adherens junctions. In all three ectodomain crystals, cadherins dimerize through a trans adhesive interface and are connected by a second, cis, interface. Assemblies formed by E-cadherin ectodomains coated on liposomes also appear to adopt this structure. Fluorescent imaging of junctions formed from wild-type and mutant E-cadherins in cultured cells confirm conclusions derived from structural evidence. Mutations that interfere with the trans interface ablate adhesion, whereas cis interface mutations disrupt stable junction formation. Our observations are consistent with a model for junction assembly involving strong trans and weak cis interactions localized in the ectodomain.

The ends of a large RNA molecule are necessarily close.

We show on general theoretical grounds that the two ends of single-stranded (ss) RNA molecules (consisting of roughly equal proportions of A, C, G and U) are necessarily close together, largely independent of their length and sequence. This is demonstrated to be a direct consequence of two generic properties of the equilibrium secondary structures, namely that the average proportion of bases in pairs is ∼60% and that the average duplex length is ∼4. Based on mfold and Vienna computations on large numbers of ssRNAs of various lengths (1000-10 000 nt) and sequences (both random and biological), we find that the 5'-3' distance-defined as the sum of H-bond and covalent (ss) links separating the ends of the RNA chain-is small, averaging 15-20 for each set of viral sequences tested. For random sequences this distance is ∼12, consistent with the theory. We discuss the relevance of these results to evolved sequence complementarity and specific protein binding effects that are known to be important for keeping the two ends of viral and messenger RNAs in close proximity. Finally we speculate on how our conclusions imply indistinguishability in size and shape of equilibrated forms of linear and covalently circularized ssRNA molecules.

Cooperativity between trans and cis interactions in cadherin-mediated junction formation.

Intercellullar junctions formed by cadherins, including desmosomes and adherens junctions, comprise two dimensional arrays of "trans" dimers formed between monomers emanating from opposing cell surfaces. Lateral "cis" interfaces between cadherins from the same cell surface have been proposed to play a role in cadherin clustering. Although the molecular details of cis interactions remain uncertain, they must define an anisotropic arrangement where binding is favorable only in certain orientations. Here we report Monte Carlo simulations performed on a 2D lattice constructed to account for the anisotropy in cadherin cis interactions. A crucial finding is that the "phase transition" between freely diffusing cadherin monomers and dimers and a condensed ordered 2D junction formed by dimers alone is a cooperative process involving both trans and cis interactions. Moreover, cis interactions, despite being too weak to be measured in solution, are critical to the formation of an ordered junction structure. We discuss these results in light of available experimental information on cadherin binding free energies that are transformed from their bulk solution values to interaction energies on a 2D lattice.

From branched networks of actin filaments to bundles.

Cross-linking proteins can mediate the emergence of rigid bundles from a dense branched network of actin filaments. To enable their binding, the filaments must first bend towards each other. We derive an explicit criterion for the onset of bundling, in terms of the initial length of filaments L, their spacing b, and cross-linker concentration f, reflecting the balance between bending and binding energies. Our model system contains actin, the branching complex Arp2/3 and the bundling protein fascin. In the first distinct stage, during which only actin and Arp2/3 are active, an entangled aster-like mesh of actin filaments is formed. Tens of seconds later, when filaments at the aster periphery are long and barely branched, a sharp transition takes place into a star-like structure, marking the onset of bundling. Now fascin and actin govern bundle growth; Arp2/3 plays no role. Using kinetic Monte Carlo simulations we calculate the temporal evolution of b and L, and predict the onset of bundling as a function of f. Our predictions are in good qualitative agreement with several new experiments that are reported herein and demonstrate how f controls the aster-star transition and bundle length. We also present two models for aster growth corresponding to different experimental realizations. The first treats filament and bundle association as an irreversible sequence of elongation-association steps. The second, applicable for low f, treats bundling as a reversible self-assembly process, where the optimal bundle size is dictated by the balance between surface and bending energies. Finally, we discuss the relevance of our conclusions for the lamellipodium to filopodia transition in living cells, noting that bundles are more likely nucleated by "tip complex" cross-linkers (e.g. mDia2 or Ena/VASP), whereas fascin is mainly involved in bundle maintenance.

Predicting the sizes of large RNA molecules.

We present a theory of the dependence on sequence of the three-dimensional size of large single-stranded (ss) RNA molecules. The work is motivated by the fact that the genomes of many viruses are large ssRNA molecules-often several thousand nucleotides long-and that these RNAs are spontaneously packaged into small rigid protein shells. We argue that there has been evolutionary pressure for the genome to have overall spatial properties-including an appropriate radius of gyration, R(g)-that facilitate this assembly process. For an arbitrary RNA sequence, we introduce the (thermal) average maximum ladder distance (MLD) and use it as a measure of the "extendedness" of the RNA secondary structure. The MLD values of viral ssRNAs that package into capsids of fixed size are shown to be consistently smaller than those for randomly permuted sequences of the same length and base composition, and also smaller than those of natural ssRNAs that are not under evolutionary pressure to have a compact native form. By mapping these secondary structures onto a linear polymer model and by using MLD as a measure of effective contour length, we predict the R(g) values of viral ssRNAs are smaller than those of nonviral sequences. More generally, we predict the average MLD values of large nonviral ssRNAs scale as N(0.67+/-0.01), where N is the number of nucleotides, and that their R(g) values vary as MLD(0.5) in an ideal solvent, and hence as N(0.34). An alternative analysis, which explicitly includes all branches, is introduced and shown to yield consistent results.

Arp2/3 branched actin network mediates filopodia-like bundles formation in vitro.

During cellular migration, regulated actin assembly takes place at the cell leading edge, with continuous disassembly deeper in the cell interior. Actin polymerization at the plasma membrane results in the extension of cellular protrusions in the form of lamellipodia and filopodia. To understand how cells regulate the transformation of lamellipodia into filopodia, and to determine the major factors that control their transition, we studied actin self-assembly in the presence of Arp2/3 complex, WASp-VCA and fascin, the major proteins participating in the assembly of lamellipodia and filopodia. We show that in the early stages of actin polymerization fascin is passive while Arp2/3 mediates the formation of dense and highly branched aster-like networks of actin. Once filaments in the periphery of an aster get long enough, fascin becomes active, linking the filaments into bundles which emanate radially from the aster's surface, resulting in the formation of star-like structures. We show that the number of bundles nucleated per star, as well as their thickness and length, is controlled by the initial concentration of Arp2/3 complex ([Arp2/3]). Specifically, we tested several values of [Arp2/3] and found that for given initial concentrations of actin and fascin, the number of bundles per star, as well as their length and thickness are larger when [Arp2/3] is lower. Our experimental findings can be interpreted and explained using a theoretical scheme which combines Kinetic Monte Carlo simulations for aster growth, with a simple mechanistic model for bundles' formation and growth. According to this model, bundles emerge from the aster's (sparsely branched) surface layer. Bundles begin to form when the bending energy associated with bringing two filaments into contact is compensated by the energetic gain resulting from their fascin linking energy. As time evolves the initially thin and short bundles elongate, thus reducing their bending energy and allowing them to further associate and create thicker bundles, until all actin monomers are consumed. This process is essentially irreversible on the time scale of actin polymerization. Two structural parameters, L, which is proportional to the length of filament tips at the aster periphery and b, the spacing between their origins, dictate the onset of bundling; both depending on [Arp2/3]. Cells may use a similar mechanism to regulate filopodia formation along the cell leading edge. Such a mechanism may allow cells to have control over the localization of filopodia by recruiting specific proteins that regulate filaments length (e.g., Dia2) to specific sites along lamellipodia.

The "electrostatic-switch" mechanism: Monte Carlo study of MARCKS-membrane interaction.

The binding of the myristoylated alanine-rich C kinase substrate (MARCKS) to mixed, fluid, phospholipid membranes is modeled with a recently developed Monte Carlo simulation scheme. The central domain of MARCKS is both basic (zeta = +13) and hydrophobic (five Phe residues), and is flanked with two long chains, one ending with the myristoylated N-terminus. This natively unfolded protein is modeled as a flexible chain of "beads" representing the amino acid residues. The membranes contain neutral (zeta = 0), monovalent (zeta = -1), and tetravalent (zeta = -4) lipids, all of which are laterally mobile. MARCKS-membrane interaction is modeled by Debye-Hückel electrostatic potentials and semiempirical hydrophobic energies. In agreement with experiment, we find that membrane binding is mediated by electrostatic attraction of the basic domain to acidic lipids and membrane penetration of its hydrophobic moieties. The binding is opposed by configurational entropy losses and electrostatic membrane repulsion of the two long chains, and by lipid demixing upon adsorption. The simulations provide a physical model for how membrane-adsorbed MARCKS attracts several PIP(2) lipids (zeta = -4) to its vicinity, and how phosphorylation of the central domain (zeta = +13 to zeta = +7) triggers an "electrostatic switch", which weakens both the membrane interaction and PIP(2) sequestration. This scheme captures the essence of "discreteness of charge" at membrane surfaces and can examine the formation of membrane-mediated multicomponent macromolecular complexes that function in many cellular processes.