RICERCA ITALIANA     

Research program

Supramolecular assemblies. The Dps (DNA-binding proteins from starved cells)- DNA and sorcin - calcium channels systems and their biological function

University Co-ordinator

Università degli Studi di ROMA "La Sapienza" - SCIENZE BIOCHIMICHE - ROMA(RM)

Research Unit Leader

Emilia CHIANCONE

Description

1. Dps proteins - DNA

A wealth of experimental data points to the involvement of the Dps N-terminal region in the binding of DNA, but only recently the proponent's group has proven that the lysine residues it contains have a specific role not only in determining the mode of interaction between the two macromolecules, but also in promoting self-aggregation of the protein in the absence of DNA (2). This conclusion was reached by comparing the behavior of wt E. coli Dps and two N-terminus deletion mutants, Dps-delta8 e Dps-delta18, which lack two or all three N-terminal lysine residues, respectively. Gel electrophoresis and atomic force microscopy experiments carried out as a function of pH have demonstrated that the self-aggregation capacity parallels the ability to condense plasmid DNA, i.e. to form large Dps-DNA complexes that contain many Dps moelcules and one or few plasmids. Self-aggregation and DNA-condensation both require that at least one N-terminal lysine residue be protonated. Thus, Dps-delta8 self-aggregates and condenses DNA at pH values lower than or euql to 7.3, but not at pH 7.7, a condition where DNA binding without condensation takes place. The fact that Dps-delta18 is able to bind DNA only at pH values around 6.3 indicates that residues on the surface of the molecule other than lysines, which are ionized at these pH values, e. g. histidines 112 and 117, can participate in DNA binding.
The proposed research activity is based on these results and on other unpublished observations. The latter show that N-terminal lysine residues of E. coli Dps are methylated in part when the protein is expressed during the stationary growth phase, but are unmethylated when the protein is expressed during the exponential phase. They also suggest that methylation has a profound effect on the interaction with DNA, since it appears to favour its condensation in agreement with the results discussed above. Methylation of the N-terminal lysine residues therefore may constitute an as yet unknown regulatory factor, of obvious evolutionary interest. The project will investigate the structural aspects of the methylation reaction. The mode of DNA binding and self-aggregation of the protein in different methylation states will be assessed as a function of pH by means of gel eletrophoresis, atomic forse microscopy (in collaboration with C. Rivetti and G. L. Rossi, Univerity of Parma), and analytical ultracentrifugation.
In order to unveil the functional ramifications of the methylation reaction, its estent will be evaluated under different growth conditions. Based on the knowledge acquired to date on other members of the Dps family, one may hypothesize that oxidative stress should not influence methylation since this reaction does not affect the protein's ferroxidase activity. Therefore, mainly other stress conditions will be tested (e.g. thermal stress, effect of alkylating agents). The data obtained will be discussed in the framework of the reported variations in chromatin structure that take place in E. coli during the different growth phases and have been attributed to the formation of ordered Dps-DNA complexes of different nature (19).
The study of Thermosynechococcus elongatus Dps, the first Dps protein from a thermophilic bacterium, is relevant in this connection. The N-terminal sequence contains a single solvent-exposed lysine residue (Lys 8) and thus resembles the E. coli Dps-delta8 mutant discussed previously. Complex formation with DNA will be characterized by means of the methodological approach described above not only as a function of pH, as for the Dps-delta8 protein, but also as a function of temperature. Thereby the importance of the entropic contribution to the formation of Dps-DNA complexes "in vitro" will be assessed and in turn its biological relevance under the extreme temperature conditions that characterize the habitat of the bacterium.
During the research activity carried out within PRIN 2003, co-crystallization experiments of E. coli Dps and plasmid DNA have been initiated; X-ray quality, albeit extremely fragile co-crystals have been obtained. After improving the stability of the wt E. coli Dps-DNA co-crystals, these experiments will be carried out on the methylated forms of Dps. The results obtained will allow the description of the interaction between the two macromolecules at a resolution never achieved before and in turn enable one to identify the correct model among those described in the literature (1, 19, 20).
In the Dps dodecamer, as alluded to above, the N-terminal regions of the individual subunits are located in the vicinity of the channels that permit access of Fe(II) to the ferroxidase center. However, their effect on the kinetics of iron incorporation has never been studied, despite the possible biological importance of this feature. Rapid mixing stopped-flow experiments will be performed which entail mixing solutions of the protein under study with anaerobic Fe(II) solutions. Since it is known that Fe(II) binding to the ferroxidase center results in a decrease of the intrinsic fluorescence intensity, this signal will be monitored. These results, taken together with those on Dps-DNA complex formation, will furnish an overall picture of how bacteria may integrate the physical and the chemical protection mechanims exterted by Dps proteins on DNA.


2. Sorcin- Ca2+ channels

The research project aims at the characterization of the interactions that sorcin establishes with calcium channels in molecular and physiological terms. In particular, the dependence of the kinetic and thermodynamic parameters of the individual binding reactions on Ca2+ concentration and the topology of the interactions will be studied. It is conceivable that during the excitation-contraction-decontraction cycle a network of interactions is established between sorcin and Ca2+ channels whose timing and extent allow sorcin to regulate intracellular calcium fluxes.
Special attention will be given to the ryanodine receptors (RyRs) and the Na+-Ca2+ exchanger (NCX); in addition to wt sorcin, its C-terminal calcium-binding domain (SCBD) and the sorcin N-terminal domain will be studied, together with site-specific mutants of the highest affinity EF3 hand (E124A) and of the D-helix in view of its critical role in the transmission of information of the binding of calcium to EF3 (e.g. the W99G and W105G mutants in which the only tryptophan residues of the polypeptide chain are substituted). A further mutant will be studied, F112L, which has been expressed in E.coli and purified both as a full-length protein and as calcium binding domain. The F112L mutation, located where the D-helix is connected to the EF3 loop, has been proposed to account for an inherited form of hypertrophic cardiomyopathy and hypertension in several families (21). The goal is to obtain X-ray quality crystals, to solve the three-dimensional structure in order to establish whether molecular basis of the pathology rests on the impairment of sorcin's interaction with its target proteins.
The interaction between sorcin and RyR1 or RyR2, that are highly expressed in the skeletal and cardiac muscle, respectively, is the most studied one also from a physiological viewpoint. Calcium release from the sarcoplasmic reticulum occurs via clusters of RyR units. Opening of individual RyR clusters generates localized calcium release events termed Ca2+ sparks, a form of elementary stochastic events from spatially localised calcium stores (22). In single permeabilized cardiomyocytes, sorcin (together with other proteins such as calmodulin and FKBP12) inhibits RyR activity (23), as shown by the decrease in Ca2+ spark frequency, amplitude and duration (17, 24, 25). The interaction of sorcin with RyR2 takes place in the millisecond timescale, at a rate that allows for sorcin modulation of the channel on a beat-to-beat basis (24).
Sorcin interacts with RyR via its C-terminal domain (26); the details of the interaction topology are unknown, although the hydrophobic region that comprises the D-helix, the EF-loop and the G-helix is likely to be involved. In collaboration with Prof. G. L. Smith (University of Glasgow), single rabbit cardiomyocytes will be permeabilized with b-escin and superfused with calcium, ATP and creatine phosphate, in the presence of wt sorcin, SCBD or the mutants mentioned above. Calcium sparks will be imaged using a confocal microscope with the fluorophore Fluo-3, that allows determination of frequency, duration, width and amplitude of the calcium release events. The effect of wt sorcin, SCBD and the various mutants on such parameters will provide information on the specific region of the C-terminal domain involved in the interaction and in addition validate the mechanism of sorcin activation proposed by Mella et al. (13) on the basis of in vitro experiments.
The study of the topology of interaction between sorcin and the Na+-Ca2+ exchanger will be based on immunoprecipitation experiments and Western blots, and in addition on the use of isolated cardiomyocytes. In rabbit ventricular myocytes NCX activity has been demonstrated to be increased upon transfection with sorcin (17). The Na+-Ca2+ exchanger, located in the cardiac plasma membrane and in other cell types, is electrogenic, since the ion exchange involves the movement of three Na+ ions for each Ca2+ ion, with a net transfer of one positive charge in the direction of Na+ movement; a further property is that NCX can operate in both directions depending on the ionic concentration and on membrane voltage (27). In different phases of the excitation-contraction cycle, NCX can generate either depolarizing or repolarizing currents. Under physiological conditions however, it works mainly in the Ca2+ extrusion mode that is driven mostly by calcium release from SR (28). In isolated rabbit cardiomyocytes, membrane currents will be measured (in collaboration with Prof. G. L. Smith) in response to a 3-second ramp from –120 to 80 mV in the presence of sorcin, SCBD and the various mutants. Background current will be obtained in the presence of 5 mM NiCl2, that inhibits NCX activity fully (17, 29). These experiments will allow us to map the sorcin region interacting with NCX.
The identification of the RyR and NCX regions interacting with sorcin is more complex, because of the dimensions of the cytoplasmic regions of the calcium channels. Therefore, a different strategy based on the phage display technique will be employed (30). Sorcin will be immobilized on a support (either a polystyrene dish or a Sepharose matrix) and phage-display libraries of filamentous phages belonging to the fd-tet family, exposing foreign peptides fused to regions of the pIII coat protein, will be subjected to several selection rounds that will allow the isolation of the peptides endowed with the highest calcium-dependent affinity for sorcin. Two types of phage display libraries will be used: a random 15-aminoacid peptide library, whose DNA inserts will be obtained by degenerate oligonucleotide synthesis, and a 30-400-residue library derived from random fragments of the target proteins. Such fragments will be obtained by random fragmentation of the RyR1, RyR2, NCX and RyR3 cDNA with Dnase I. RyR3 will be studied since it is characterized by a very high sequence homology relative to RyR1 and RyR2, and therefore is likely to interact with sorcin. Ryr3 is highly expressed in corpus striatum, thalamus, and hippocampus, and is important in synaptic plasticity, hippocampal long-term potentiation and spatial learning.
After identification of the RyR and NCX peptides that constitute the sites of interaction with sorcin, the kinetic and thermodynamic parameters of the individual interactions will be determined by surface plasmon resonance (SPR). Techniques based on SPR are powerful tools for the study of macromolecule interactions. The biosensor is a glass surface coated on one side with a thin gold film and on the other with a dextran matrix where the ligand is immobilized and reaction with the analyte occurs. When a protein sample is injected into the flow cell and binds to the ligand, a change in refractive index occurs that produces a shift in the angle at which the incident light on the other side of the metal film is absorbed due to the SPR phenomenon. This change is recorded by the detector and is displayed in the form of a sensorgram; the technique allows a real-time ligand-analyte interaction analysis that provides both kinetic information about complex formation (association and dissociation rates) and equilibrium constants.
The RyR and NCX peptides selected on the basis of the phage display experiments will be immobilized on a CM5 biosensor; sorcin will be injected into the flow cell at different protein and calcium concentrations and the association, steady state and dissociation phases of the binding reaction will be followed. The kinetic and equilibrium parameters thus obtained will provide insight into the structural features of the interacting partners that determine specificity and affinity of the individual binding reactions. In turn, these data will shed light into the complex network of interactions that sorcin establishes upon translocation to membranes, and will thereby contribute to the knowledge of sorcin's function as a regulator during the excitation-contraction cycle.