Estudio por RMN del reconocimiento molecular carbohidrato/proteína con discriminación diferencial de anómeros en equilibrio

Abstract

Carbohydrates, together with proteins, nucleic acids and lipids, are essential molecules of biological systems involved in structural, metabolic and cell signaling functions. Carbohydrates are composed of polyhydroxylated monosaccharide units with an aldehyde or ketone functional group. Chemically monosaccharides exist in cyclic structures closed with an acetal or hemiacetal function. Depending on the stereochemistry and regiochemistry of the acetal function, the same monosaccharide may exist in various anomeric forms. Monosaccharides generally exist in aqueous solutions in the equilibrium form between two major anomeric forms, alpha and beta. The unique reactivity of the acetal function allows monosaccharides to build oligomeric structures by forming the glycosidic bond between the anomeric position of a monosaccharide and one of the hydroxyls of the following monosaccharide resulting oligo-and polysaccharide in both lineal and branched structures of great complexity and diversity. These oligosaccharide structures can also be conjugated to other biomolecules to render glycoproteins, glycolipids, proteoglycans, etc., and in many cases are the main component of the cell outer surface (glycocalyx olisacaridicas capsules etc) or even constitute the chemical backbone of the cell walls as peptidoglycan in bacteria or chitin in fungal microorganism. Glycoconjugates are under active research area because, in addition to its structural role, have been found participating in many fundamental biological processes, fertilization, development and differentiation, immune system, recognition of pathogens, cancer metastasis and so on. Many of these phenomena involve processes of molecular recognition between carbohydrates and glycoconjugates, expressed on cell surfaces and specific carbohydrate binding proteins including lectins. Lectins are defined by their ability to bind carbohydrates but are not antibody-like proteins of immune system neither have enzymatic activity. Carbohydrate-protein recognition processes implied transient noncovalent interactions with middle to low affinity in biochemical terms (millimolar of micro-order) but high selectivity when it is referring to distinguish different types of carbohydrates. Lectins are able to very selectively discriminate subtle differences as the stereochemistry of a single hydroxyl group that differentiates galactose from glucose. Significantly different stereochemistry of the anomeric position is what differentiates the alpha and beta anomers of monosaccharides however in this case the chemical properties of the acetal function causes the chemical equilibrium of both anomers in aqueous solution and therefore to analyze the selectivity towards each anomer individually is not an easy task and in most cases the affinity studies are conducted on mixtures of anomers that only allow to drawn average values for the equilibrium mixture but do not allow to establish preference towards anyone of the two anomers.

Methodologies based on nuclear magnetic resonance are very suitable to study carbohydrate-protein transient interactions. Both approaches are possible, observing the NMR signals of the protein spins and its perturbations by the presence of ligands or the other way around, observing the NMR signals of the ligand and its perturbations by the presence of lithe receptor. When applying strategies from the point of view of the protein then it is necessary to previously assign the protein´s signals to the corresponding atoms of the protein . The assignment can be performed by means of homonuclear or heteronuclear correlation NMR experiments. In the case of heteronuclear experiments, mandatory for proteins over 10kD, it is necessary to express and produce, by means of molecular biology methodologies, the protein isotopically labeled with 13C and/or 15N to surpass the very low sensitivity of NMR. Once the chemical shift are assigned to each protein ´s proton then it is possible to determine the tridimensional structure of the protein recording nuclear Overhauser effects (NOE) correlations among the protons of the protein. Those NOE corresponds to dipolar interactions between hydrogen pair that are inversely proportional to the six power of interprotein distances. The interproton NOE signals can be translated into distance restrictions that become into experimental inputs for the calculations of the tridimensional structure of the by means of molecular modeling protocols. Once the structure is calculated then the perturbations of the protein NMR signals because of the interactions with the ligands can be use to estimate the affinity constants performing titrations with increasing concentration of ligand and additionally, the identification of the most perturbed signals allows to map the carbohydrate binding site in the protein. Alternatively applying NMR experiments from the point of view of the ligand make use of the difference of in the ligand‘s NMR parameters on passing from the small molecule regime with fast tumbling correlated with fast rotational correlation time and slow spin relaxation to the large molecule regime when the ligand is bound to the protein with slow rotational correlation time corresponding now a fast spin relaxation. Making use of these differences it is possible determine affinity constants and map the binding epitope on the ligand with NMR experiments like the saturation transfer difference (STD). STD

experiments are base on the selective transfer of spin saturation from the protein when it is selectively irradiated (without affecting directly the signals of the ligand) and this saturation is efficiently transferred only to spins that are in close contact with the protein as it happens only when the ligand is bound to the protein. relaxation. This differences between the free and the bound state amplifies the signal perturbations on the ligand spectrum and the saturation of the spins of the protein will be esasily transfer to those protons of the ligand in close contact with the protein in the bound state. In this thesis two different model systems of carbohydrate-protein interactions have been studied. Lectins that recognizes N-acetylglucosamine (GlcNAc) related oligosacarides. N-AcetylGlucosamine (GlcNAc) derived oligo- and polysaccharides participate in many biological important structures. Chitin (GlcNAc polysaccharide) is the main structural component of insect and crustaceans exoskeleton and is present in the cell wall of fungus and other microorganism, as well as bacterial cell walls are made of peptidoglycan that has its glycan core chain built with dissacharide repeating units of GlcNAc and related N-AcetylMuramic acid (GlcNAc-N-Acetylmuramyl disaccharides) crosslinked with short polypeptide fragments. Quite diverse proteins recognize those chitooligosaccharide related structures accomplishing functions as, for example, defense proteins (hevein domains), symbiosis signal recognition (Lys-M domains), carbohydrate processing (chitinases, glucosaminidases) etc. In this thesis NMR methodologies have been applied for studying those interactions in model systems based on hevein, Lys-M and PASTA protein domains. -Hevein Domains: Many plants express chitin binding proteins that have been identified as plant defense proteins. A family of these plant defense proteins contains one or several chitin binding sites localized on the so call "hevein domains". Those domains have been named after hevein, a small 43 residues protein with 4 disulfide bonds that binds chitin and chitoligosaccharides. Wheat Germ agglutinin (WGA) is a lectin compose of four homologous hevein domains. It has been shown that WGA binds N-AcetylGlucosamine (GlcNAc) related oligosaccharides and in fact it is use as commercial reagent to detect glycans containing GlcNAc residues. It has been shown that WGA is able to bind bacterial cells, data that could correlate with its plant-defense capacities, but there is no information at molecular level how WGA binds to peptidoglycan. In this thesis the binding of short peptidoglycan fragments to WGA have been characterized by means of STD-NMR studies. The results obtained demostrate that disaccharide dipeptide peptidoglycan fragment (glucosaminyl-muramyl-dipeptide, GMDP) is recognized by WGA and that the GlcNAc residue establishes the mayor contacts with WGA followed by the NAcetylmuramic acid residue while the peptidic moiety is pointing outside of the binding site. Additionally, as the GMDP is present as the mixture of anomers of its reducing end (muramic residue), the STD experiments have shown that WGA preferentially binds the the beta anomer . -LysM-Domains

The LysM (Lysin motiv) named after Eschericihia c. lytic transglycosidase is a protein domain of about 50 aminoacid residues being part of larger multidomain proteins extensively distributed in virus, bacteria, plants, fungi and animals. In bacteria and plants some of the proteins containing LysM domains are associated with recognition of peptidoglycan and quitoologosaccharides acting as sensor proteins of pathogens (recognition of patogen by binding fragmentes of the cell walls of the patoghens) or simbionts (for example the receptors of bacterial nodulation factors (chitolipooligosaccharides) contains three LysM domain sequentially organized). However there is not yet a generalized description of the binding mode between oligosaccharides and LyM domain. In this thesis a mutant variant of LysM domain from Eschericihia c. lytic transglycosidase has been used as model system to study LysMoligosaccharide interactions. First the domain was successfully expressed and produced from bacteria. The attempts to similarly express some LysM domains from plant proteins involved in detecting nodulation factors were unsuccessful. The correct folding of the model LysM domain was confirmed by assignment of the homonuclear NMR spectra of the domain. After assignment and structural calculations, binding studies were performed following possible perturbation of signals from the protein. Only with polymerized peptidoglican was possible to shown binding of LysM-domain to peptidoglican following NMR signal intensity decrease after several additions of peptidoglycan to the sample however no interaction was observed neither with GMDP nor with quitooligosaccharides. -PASTA Domain: Reactivation from dormancy, growth and division of bacteria require cleavage of the cell-wall peptidoglycan. Bacteria turn over their cell-wall material due to the actions of peptidoglycan hydrolases and amidases Muropeptide-driven exit from dormancy requires a member of the serine/threonine kinase (STPK) family. Proteins of this family are expressed in many prokaryotes, including a broad range of pathogens, and modulate a wide number of cellular processes, such as biofilm formation, cell wall biosynthesis and cell division, sporulation, stress response. This proteins presents in their domain organizations, an intracellular serine/threonine kinase domain, a transmembrane region and an extra-cellular portion that contains domains denoted as PASTA (Penicillin binding Associated and Serine/Threonine kinase Associated domains) It has been described that the PASTA domain is responsible of the binding to muropeptides. In this thesis it has been determined by NMR, applying heteronuclear NMR correlations experiments, the tridimensional structure of the second PASTA domain of Prkc protein from Bacillus subtilis expresed heterogously with isotopic enrichment in 13C and 15N. Interestingly, when the isolated domain was studied in the presence of muropetides isolated from Bacillus s. cells walls, no interaction was observed. Additionally no binding was observed with betalactams neither with GMDP muroppetide. The results obtained rise the hipothesis that more than one PASTA domain should be necessary to express the full functionality of this protein domains in order to recognize their natural ligands the tyrosine site of viscumin.

Lectins that binds Galactose. Anomers selection by viscumin Viscumin is plant toxin-type lectin extracted from latex of mistletoe and has been proposed to be useful therapeutic drug for treatment some types of cancer. It is a heterodimer with a toxin subunit A and a lectin subunit B. The lectin subunit has two carbohydrate binding sites, one located around tryptophan 38 y other around tyrosine 249. The tryptophan site has been characterized as the one with high aaffinity but its functionality is very much dependent of the quaternary structure of the lectin. Previous work have shoun some preference for alfa galactosides but there is no data if the viscumin discriminate the anomers of plain reducing galactose monosaccharide. Reducing carbohydrates can be characterized in two different structures depending of the configuration at the anomeric position of the reducing end. Anomeric selectivity has been observed and studied in enzymes involved in carbohydrate transformations by diverse kinetic approaches since long time ago, however those approaches are not possible in the case of just binding studies of carbohydrate in solution. Lectins and other carbohydrates receptors do not modified their ligands, thus do not generate time dependent responses feasible of monitoring. Taking into account that free anomers equilibrate in no more than a few hours, many of the available biophysical techniques in solution only provide a mean macroscopic view of the recognition event involving the mixture of anomers in equilibrium. The direct discrimination of the recognition of each anomer is not an easy task, and NMR can be technique of choice because it allows observing independent signals for each anomer in spite of they are in chemical equilibrium exchange. In fact, 13C NMR has been used to show the -anomer preference of bacterial chemotaxis sugar binding proteins and more exotic tritium 3H -NMR has been used to study anomeric preference in maltose binding protein. In this thesis, by means of STD NMR strategies combined with molecular docking, it has been possible to obtain quantitative information on the individual interaction of each anomer, in order to characterize the differential monosaccharide anomer affinity by viscumin lectin. the obtained data shows an slight preference for the alfa anomer over the beta anomer in both cases, the free reducing galactose as mixture of anomers and the alfa- and beta- methyl-galactosides with fix configuration at the anomeric carbon. On the other side the comparison of the experimental STD data with theoretical predictions of saturation for each proton of the ligand applying the CORCEMA protocol allows to define the binding epitope of each ligand and identify the binding site on the protein. Interestingly, under the conditions of viscumin concentration used in the NMR experiments that favors the dimer of heterodimers ABAB quaternary structure, the fitting of the experimental with the calculated STD data shows that the tryptophan site of viscumin is not full accessible to the ligand while the experimental data can be explained assuming that the ligand binds mostly to This result is in accordance with previous reports that shown the dependence of the functionality of the tryptophan site on the viscumin concentration.