Patrick Groves

NMR and Modeling Studies of Protein–Carbohydrate Interactions: Synthesis, Three-Dimensional Structure, and Recognition Properties of a Minimum Hevein Domain with Binding Affinity for Chitooligosaccharides

HEV32, a 32‐residue, truncated hevein lacking eleven C‐terminal amino acids, was synthesized by solid‐phase methodology and correctly folded with three cysteine bridge pairs. The affinities of HEV32 for small chitin fragments—in the forms of N,N′,N′′‐triacetylchitotriose ((GlcNAc)3) (millimolar) and N,N′,N′′,N′′′,N′′′′,N′′′′′‐hexaacetylchitohexaose ((GlcNAc)6) (micromolar)—as measured by NMR and fluorescence methods, are comparable with those of native hevein. The HEV32 ligand‐binding process is enthalpy driven, while entropy opposes binding. The NMR structure of ligand‐bound HEV32 in aqueous solution was determined to be highly similar to the NMR structure of ligand‐bound hevein. Solvated molecular‐dynamics simulations were performed in order to monitor the changes in side‐chain conformation of the binding site of HEV32 and hevein upon interaction with ligands. The calculations suggest that the Trp21 side‐chain orientation of HEV32 in the free form differs from that in the bound state; this agrees with fluorescence and thermodynamic data. HEV32 provides a simple molecular model for studying protein–carbohydrate interactions and for understanding the physiological relevance of small native hevein domains lacking C‐terminal residues.

Hevein Domains: An Attractive Model to Study Carbohydrate–Protein Interactions at Atomic Resolution

Publisher Summary This chapter focuses on hevein domains and presents an attractive model to study carbohydrate–protein interactions at atomic resolution. Among the various biological processes in which carbohydrates are involved as biochemical signals, it is noteworthy that many plants harbor defense proteins (lectins) against pathogenic attack. These proteins are able to bind to chitin. This natural biopolymer is a key structural component of the cell wall of fungi and of the exoskeleton of invertebrates such as insects, nematodes, and arthropods. Direct binding to the saccharide can occur for the respective lectin, while a particular domain can also be instrumental for chitin-degrading enzymes. The antifungal activity of plant chitinases is largely restricted to those chitinases that contain a noncatalytic, plant-specific, chitin-binding domain (ChBD), also termed as “hevein domain.” This domain displays a common structural motif of 30–43 residues, rich in glycine and cysteine residues in highly conserved positions and organized around a four-disulfide core. The chapter explains the concepts related to protein–carbohydrate interactions and elaborates the basic techniques for analyzing sugar–hevein interactions. It also discusses the structure of the Hevein–Saccharide complexes.

Calretinin and calbindin D28k have different domain organizations.

The domain organization of calretinin (CR) was predicted to involve all six EF‐hand motifs (labeled I to VI) condensed into a single domain, as characterized for calbindin D28k (Calb), the closest homolog of calretinin. Unperturbed 1H,15N HSQC NMR spectra of a 15N‐labeled calretinin fragment (CR III–VI, residues 100–271) in the presence of the unlabeled complimentary fragment (CR I–II, residues 1–100) show that these fragments do not interact. Size exclusion chromatography and affinity chromatography data support this conclusion. The HSQC spectrum of 15N‐labeled CR is similar to the overlaid spectra of individual 15N‐labeled CR fragments (CR I–II and CR III–VI), also suggesting that these regions do not interact within intact CR. In contrast to these observations, but in accordance with the Calb studies, we observed interactions between other CR fragments: CR I (1–60) with CR II–VI (61–271), and CR I–III (1–142) with CR IV–VI (145–271). We conclude that CR is formed from at least two independent domains consisting of CR I–II and CR III–VI. The differences in domain organization of Calb and CR may explain the specific target interaction of Calb with caspase‐3. Most importantly, the comparison of CR and Calb domain organizations questions the value of homologous modeling of EF‐hand proteins, and perhaps of other protein families.

A MODEL FOR TARGET PROTEIN BINDING TO CALCIUM-ACTIVATED S100 DIMERS

S100 proteins are a family of dimeric calcium‐binding proteins implicated in several cancers and neurological diseases. Calbindin D9k is an unusual monomeric member of the S100 family. A calbindin D9k mutant containing a novel calcium‐induced helix is characterized. Based on sequence comparison, this helix could be a component of other S100 proteins and a factor in target protein binding. The origin of structural differences between three reported apo S100 dimer structures is verified. We conclude that the differences are a result of modeling rather than a function of different target binding properties. A mechanism for target protein binding is suggested.

Engineering a Structure Switching Mechanism into a Steroid-Binding Aptamer and Hydrodynamic Analysis of the Ligand Binding Mechanism

The steroid binding mechanism of a DNA aptamer was studied using isothermal titration calorimetry (ITC), NMR spectroscopy, quasi-elastic light scattering (QELS), and small-angle X-ray spectroscopy (SAXS). Binding affinity determination of a series of steroid-binding aptamers derived from a parent cocaine-binding aptamer demonstrates that substituting a GA base pair with a GC base pair governs the switch in binding specificity from cocaine to the steroid deoxycholic acid (DCA). Binding of DCA to all aptamers is an enthalpically driven process with an unfavorable binding entropy. We engineered into the steroid-binding aptamer a ligand-induced folding mechanism by shortening the terminal stem by two base pairs. NMR methods were used to demonstrate that there is a transition from a state where base pairs are formed in one stem of the free aptamer, to where three stems are formed in the DCA-bound aptamer. The ability to generate a ligand-induced folding mechanism into a DNA aptamer architecture based on the three-way junction of the cocaine-binding aptamer opens the door to obtaining a series of aptamers all with ligand-induced folding mechanisms but triggered by different ligands. Hydrodynamic data from diffusion NMR spectroscopy, QELS, and SAXS show that for the aptamer with the full-length terminal stem there is a small amount of structure compaction with DCA binding. For ligand binding by the short terminal stem aptamer, we propose a binding mechanism where secondary structure forms upon DCA binding starting from a free structure where the aptamer exists in a compact form.

Matrix vesicles isolated from mineralization-competent Saos-2 cells are selectively enriched with annexins and S100 proteins.

Matrix vesicles (MVs) are cell-derived membranous entities crucial for mineral formation in the extracellular matrix. One of the dominant groups of constitutive proteins present in MVs, recognised as regulators of mineralization in norm and pathology, are annexins. In this report, besides the annexins already described (AnxA2 and AnxA6), we identified AnxA1 and AnxA7, but not AnxA4, to become selectively enriched in MVs of Saos-2 cells upon stimulation for mineralization. Among them, AnxA6 was found to be almost EGTA-non extractable from matrix vesicles. Moreover, our report provides the first evidence of annexin-binding S100 proteins to be present in MVs of mineralizing cells. We observed that S100A10 and S100A6, but not S100A11, were selectively translocated to the MVs of Saos-2 cells upon mineralization. This observation provides the rationale for more detailed studies on the role of annexin-S100 interactions in MV-mediated mineralization.

The relative orientation of the lipid and carbohydrate moieties of lipochitooligosaccharides related to nodulation factors depends on lipid chain saturation

Lipochitooligosaccharides (LCOs) signal the symbiosis of rhizobia with legumes and the formation of nitrogen-fixing root nodules. LCOs 1 and 2 share identical tetrasaccharide scaffolds but different lipid moieties (1, LCO-IV(C16:1[9Z], SNa) and , LCO-IV(C16:2[2E,9Z], SNa)). The conformational behaviors of both LCOs were studied by molecular modeling and NMR. Modeling predicts that a small lipid modification would result in a different relative orientation of the lipid and tetrasaccharide moieties. Diffusion ordered spectroscopy reports that both LCOs form small aggregates above 1 mM. Nuclear Overhauser spectroscopy (NOESY) data, collected under monomeric conditions, reveals lipid-carbohydrate contacts only for 1, in agreement with the modeling data. The distinct molecular structures of 1 and 2 have the potential to contribute to their selective binding by legume proteins.

Enthalpic (electrostatic) contribution to the chelate effect: a correlation between ligand binding constant and a specific hydrogen bond strength in complexes of glycopeptide antibiotics with cell wall analogues

The 1H NMR chemical shift of amide protons in the binding pocket of glycopeptide antibiotics has been used to monitor the interaction of these amide protons with the carboxylate group of cell wall analogues and related ligands. A good correlation is observed between overall ligand binding energy (ΔG°) and amide NH chemical shift. We conclude that the strength of the electrostatic interaction of the carboxylate group, which is crucial to recognition and binding by the antibiotics, is cooperatively enhanced by adjacent functional groups on the same ligand template. Hydrogen bonding and burial of hydrocarbon in adjacent sites produce an enhancement of electrostatic binding of the carboxylate group. The data provide experimental evidence for an enthalpic contribution to the chelate effect that is distinct from, and works in addition to, the classic entropic chelate effect. The correlation between amide NH chemical shift and overall binding energy has been used to show binding affinity for eremomycin and chloroeremomcin by di-N-Ac-Lys-D-Ala-D-Lac (Lac = lactate), which is a cell wall analogue of bacteria which exhibit vancomycin resistance. Binding constants for this ligand have also been determined by UV difference spectrophotometry (70 dm3 mol–1 and 240 dm3 mol–1 respectively).

Diffusion nuclear magnetic resonance spectroscopy detects substoichiometric concentrations of small molecules in protein samples

Small molecules are difficult to detect in protein solutions, whether they originate from elution during affinity chromatography (e.g., imidazole, lactose), buffer exchange (Tris), stabilizers (e.g., beta-mercaptoethanol, glycerol), or excess labeling reagents (fluorescent reagents). Mass spectrometry and high-pressure liquid chromatography (HPLC) often require substantial efforts in optimization and sample manipulation to provide sufficient sensitivity and reliability for their detection. One-dimensional (1D) (1)H nuclear magnetic resonance (NMR) could, in principle, detect residual amounts of small molecules in protein solutions down to equimolecular concentrations with the protein. However, at lower concentrations, the NMR signals of the contaminants can be hidden in the background spectrum of the protein. As an alternative, the 1D diffusion difference protocol used here is feasible. It even improves the detection level, picking up NMR signals from small-molecule contaminants at lower concentrations than the protein itself. We successfully observed 30 microM imidazole in the presence of four different proteins (1-1.5 mg/ml, 6-66 kDa, 25-250 microM) by 1D diffusion-ordered spectroscopy (DOSY) difference and 1-h total acquisition time. Of note, imidazole was not detected in the corresponding 1D (1)H NMR spectra. This protocol can be adapted to different sample preparation procedures and NMR acquisition methods with minimal manipulation in either deuterated or nondeuterated buffers.

Rapid Stoichiometric Analysis of G-Quadruplexes in Solution

Guanine-rich sequences of nucleic acids may fold into secondarystructural folds called quadruplex architectures.These architectures are a rapidly growing theme of interestwith promising repercussions in our understanding of biologyand practical applications in medical fields, materialsscience, and biotechnology. Currently, there is thus interestin determining both topology and their atomic detail 3Dstructures. The establishment of solution conditions suitablefor structural studies of G-quadruplex architectures requiresthe determination of the level of oligomerization (stoichiometry)of DNA strands. Various analytical techniques arecurrently applied for the routine assessment of the stoichiometrythat generally include conditions not representativeof the environment in which the structural studies are performed.