Hard nut to crack: Solving the disulfide linkage pattern of the Neosartorya (Aspergillus) fischeri antifungal protein 2.

As a consequence of the fast resistance spreading, a limited number of drugs are available to treat fungal infections. Therefore, there is an urgent need to develop new antifungal treatment strategies. The features of a disulfide bond-stabilized antifungal protein, NFAP2 secreted by the mold Neosartorya (Aspergillus) fischeri render it to be a promising template for future protein-based antifungal drug design, which requires knowledge about the native disulfide linkage pattern as it is one of the prerequisites for biological activity. However, in the lack of tryptic and chymotryptic proteolytic sites in the ACNCPNNCK sequence, the determination of the disulfide linkage pattern of NFAP2 is not easy with traditional mass spectrometry-based methods. According to in silico predictions working with a preliminary nuclear magnetic resonance (NMR) solution structure, two disulfide isomers of NFAP2 (abbacc and abbcac) were possible. Both were chemically synthesized; and comparative reversed-phase high-performance liquid chromatography, electronic circular dichroism and NMR spectroscopy analyses, and antifungal susceptibility and efficacy tests indicated that the abbcac is the native pattern. This knowledge allowed rational modification of NAFP2 to improve the antifungal efficacy and spectrum through the modulation of the evolutionarily conserved γ-core region, which is responsible for the activity of several antimicrobial peptides. Disruption of the steric structure of NFAP2 upon γ-core modification led to the conclusions that this motif may affect the formation of the biologically active three-dimensional structure, and that the γ-core modulation is not an efficient tool to improve the antifungal efficacy or to change the antifungal spectrum of NFAP2.

Confirmation of the Disulfide Connectivity and Strategies for Chemical Synthesis of the Four-Disulfide-Bond-Stabilized Aspergillus giganteus Antifungal Protein, AFP.

Emerging fungal infections require new, more efficient antifungal agents and therapies. AFP, a protein from Aspergillus giganteus with four disulfide bonds, is a promising candidate because it selectively inhibits the growth of filamentous fungi. In this work, the reduced form of AFP was prepared using native chemical ligation. The native protein was synthesized via oxidative folding with uniform protection for cysteine thiols. AFP's biological activity depends heavily on the pattern of natural disulfide bonds. Enzymatic digestion and MS analysis provide proof for interlocking disulfide topology (abcdabcd) that was previously assumed. With this knowledge, a semi-orthogonal thiol protection method was designed. By following this strategy, out of a possible 105, only 6 disulfide isomers formed and 1 of them proved to be identical with the native protein. This approach allows the synthesis of analogs for examining structure-activity relationships and, thus, preparing AFP variants with higher antifungal activity.

DMSO-Induced Unfolding of the Antifungal Disulfide Protein PAF and Its Inactive Variant: A Combined NMR and DSC Study.

PAF and related antifungal proteins are promising antimicrobial agents. They have highly stable folds around room temperature due to the presence of 3-4 disulfide bonds. However, unfolded states persist and contribute to the thermal equilibrium in aqueous solution, and low-populated states might influence their biological impact. To explore such equilibria during dimethyl sulfoxide (DMSO)-induced chemical unfolding, we studied PAF and its inactive variant PAFD19S using nuclear magnetic resonance (NMR) and differential scanning calorimetry (DSC). According to the NMR monitoring at 310 K, the folded structures disappear above 80 v/v% DMSO concentration, while the unfolding is completely reversible. Evaluation of a few resolved peaks from viscosity-compensated 15N-1H HSQC spectra of PAF yielded ∆G = 23 ± 7 kJ/M as the average value for NMR unfolding enthalpy. The NMR-based structures of PAF and the mutant in 50 v/v% DMSO/H2O mixtures were more similar in the mixed solvents then they were in water. The 15N NMR relaxation dynamics in the same mixtures verified the rigid backbones of the NMR-visible fractions of the proteins; still, enhanced dynamics around the termini and some loops were observed. DSC monitoring of the Tm melting point showed parabolic dependence on the DMSO molar fraction and suggested that PAF is more stable than the inactive PAFD19S. The DSC experiments were irreversible due to the applied broad temperature range, but still suggestive of the endothermic unfolding of PAF.

Resonance assignment of the Shank1 PDZ domain.

Shank proteins are among the most abundant and well-studied postsynaptic scaffold proteins. Their PDZ domain has unique characteristics as one of its loop regions flanking the ligand-binding site is uniquely long and has also been implicated in the formation of PDZ dimers. Here we report the initial characterization of the Shank1 PDZ domain by solution NMR spectroscopy. The assigned chemical shifts are largely consistent with the common features of PDZ domains in general and the available Shank PDZ crystal structures in particular. Our analysis suggests that under the conditions investigated, the domain is monomeric and the unique loop harbors a short helical segment, observed in only one of the known X-ray structures so far. Our work stresses the importance of solution-state investigations to fully decipher the functional relevance of the structural and dynamical features unique to Shank PDZ domains.

Solution Structure, Dynamics, and New Antifungal Aspects of the Cysteine-Rich Miniprotein PAFC.

The genome of Penicillium chrysogenum Q176 contains a gene coding for the 88-amino-acid (aa)-long glycine- and cysteine-rich P. chrysogenum antifungal protein C (PAFC). After maturation, the secreted antifungal miniprotein (MP) comprises 64 aa and shares 80% aa identity with the bubble protein (BP) from Penicillium brevicompactum, which has a published X-ray structure. Our team expressed isotope (15N, 13C)-labeled, recombinant PAFC in high yields, which allowed us to determine the solution structure and molecular dynamics by nuclear magnetic resonance (NMR) experiments. The primary structure of PAFC is dominated by 14 glycines, and therefore, whether the four disulfide bonds can stabilize the fold is challenging. Indeed, unlike the few published solution structures of other antifungal MPs from filamentous ascomycetes, the NMR data indicate that PAFC has shorter secondary structure elements and lacks the typical ß-barrel structure, though it has a positively charged cavity and a hydrophobic core around the disulfide bonds. Some parts within the two putative γ-core motifs exhibited enhanced dynamics according to a new disorder index presentation of 15N-NMR relaxation data. Furthermore, we also provided a more detailed insight into the antifungal spectrum of PAFC, with specific emphasis on fungal plant pathogens. Our results suggest that PAFC could be an effective candidate for the development of new antifungal strategies in agriculture.

Solution structure and novel insights into phylogeny and mode of action of the Neosartorya (Aspergillus) fischeri antifungal protein (NFAP).

Small, cysteine-rich and cationic antifungal proteins from natural sources are promising candidates for the development of novel treatment strategies to prevent and combat infections caused by drug-resistant fungi. However, limited information about their structure and antifungal mechanism hampers their future applications. In the present study, we determined the solution structure, dynamics and associated solvent areas of the Neosartorya (Aspergillus) fischeri antifungal protein NFAP. Genome mining within the genus revealed the presence of orthologous genes in N. fischeri and Neosartorya spathulata, and genes encoding closely related proteins can be found in Penicillium brasiliensis and Penicillium oxalicum. We show that the tertiary structure of these putative proteins can be resolved using the structure of NFAP as reliable template for in silico prediction. Localization studies with fluorescence-labelled protein pointed at an energy-dependent uptake mechanism of NFAP in the sensitive model fungus Neurospora crassa and subsequent cytoplasmic localization coincided with cell-death induction. The presented results contribute to a better understanding of the structure/function relationship of NFAP and related proteins and pave the way towards future antifungal drug development.

Calcium binding of the antifungal protein PAF: Structure, dynamics and function aspects by NMR and MD simulations.

Calcium ions (Ca2+) play an important role in the toxicity of the cysteine-rich and cationic antifungal protein PAF from Penicillium chrysogenum: high extracellular Ca2+ levels reduce the toxicity of PAF in the sensitive model fungus Neurospora crassa in a concentration dependent way. However, little is known about the mechanistic details of the Ca2+ ion impact and the Ca2+ binding capabilities of PAF outside the fungal cell, which might be the reason for the activity loss. Using nuclear magnetic resonance (NMR), isothermal titration calorimetry and molecular dynamics (MD) simulations we demonstrated that PAF weakly, but specifically binds Ca2+ ions. MD simulations of PAF predicted one major Ca2+ binding site at the C-terminus involving Asp53 and Asp55, while Asp19 was considered as putative Ca2+ binding site. The exchange of Asp19 to serine had little impact on the Ca2+ binding, however caused the loss of antifungal activity, as was shown in our recent study. Now we replaced the C-terminal aspartates and expressed the serine variant PAFD53S/D55S. The specific Ca2+ binding affinity of PAFD53S/D55S decreased significantly if compared to PAF, whereas the antifungal activity was retained. To understand more details of Ca2+ interactions, we investigated the NMR and MD structure/dynamics of the free and Ca2+-bound PAF and PAFD53S/D55S. Though we found some differences between these protein variants and the Ca2+ complexes, these effects cannot explain the observed Ca2+ influence. In conclusion, PAF binds Ca2+ ions selectively at the C-terminus; however, this Ca2+ binding does not seem to play a direct role in the previously documented modulation of the antifungal activity of PAF.

New Antimicrobial Potential and Structural Properties of PAFB: A Cationic, Cysteine-Rich Protein from Penicillium chrysogenum Q176.

Small, cysteine-rich and cationic proteins with antimicrobial activity are produced by diverse organisms of all kingdoms and represent promising molecules for drug development. The ancestor of all industrial penicillin producing strains, the ascomycete Penicillium chryosgenum Q176, secretes the extensively studied antifungal protein PAF. However, the genome of this strain harbours at least two more genes that code for other small, cysteine-rich and cationic proteins with potential antifungal activity. In this study, we characterized the pafB gene product that shows high similarity to PgAFP from P. chrysogenum R42C. Although abundant and timely regulated pafB gene transcripts were detected, we could not identify PAFB in the culture broth of P. chrysogenum Q176. Therefore, we applied a P. chrysogenum-based expression system to produce sufficient amounts of recombinant PAFB to address unanswered questions concerning the structure and antimicrobial function. Nuclear magnetic resonance (NMR)-based analyses revealed a compact ß-folded structure, comprising five ß-strands connected by four solvent exposed and flexible loops and an "abcabc" disulphide bond pattern. We identified PAFB as an inhibitor of growth of human pathogenic moulds and yeasts. Furthermore, we document for the first time an anti-viral activity for two members of the small, cysteine-rich and cationic protein group from ascomycetes.

Structural features of human DJ-1 in distinct Cys106 oxidative states and their relevance to its loss of function in disease.

DJ-1 (PARK7) is a multifunctional protein linked to the onset and progression of a number of diseases, most of which are associated with high oxidative stress. The Cys106 of DJ-1 is unusually reactive and thus sensitive to oxidation, and due to high oxidative stress it was observed to be in various oxidized states in disease condition. The oxidation state of Cys106 of DJ-1 is believed to determine the specific functions of the protein in normal and disease conditions. Here we report molecular dynamics simulation and biophysical experimental studies on DJ-1 in reduced (Cys106, S-), oxidized (Cys106, SO2-), and over-oxidized (Cys106, SO3-) states. To simulate the different oxidation states of Cys106 in DJ-1, AMBER related force field parameters were developed and reported for 3-sulfinoalanine and cysteine sulfonic acid. Our studies found that the overall structure of DJ-1 in different oxidation states was similar globally, while it differed locally significantly, which have implications on its stability, function and its link to disease on-set. Importantly, the results suggest that over-oxidation may trigger loss of functions due to local structural modification in the Cys106 containing pocket of DJ-1 and structurally destabilize the dimeric state of DJ-1, which is believed to be its bioactive conformation. Such loss of functions would result in reduced ability of DJ-1 to protect from oxidative stress insults and may lead to increased progression of disease.

Fine-tuning the extent and dynamics of binding cleft opening as a potential general regulatory mechanism in parvulin-type peptidyl prolyl isomerases.

Parvulins or rotamases form a distinct group within peptidyl prolyl cis-trans isomerases. Their exact mode of action as well as the role of conserved residues in the family are still not unambiguously resolved. Using backbone S2 order parameters and NOEs as restraints, we have generated dynamic structural ensembles of three distinct parvulins, SaPrsA, TbPin1 and CsPinA. The resulting ensembles are in good agreement with the experimental data but reveal important differences between the three enzymes. The largest difference can be attributed to the extent of the opening of the substrate binding cleft, along which motional mode the three molecules occupy distinct regions. Comparison with a wide range of other available parvulin structures highlights structural divergence along the bottom of the binding cleft acting as a hinge during the opening-closing motion. In the prototype WW-domain containing parvulin, Pin1, this region is also important in forming contacts with the WW domain known to modulate enzymatic activity of the catalytic domain. We hypothesize that modulation of the extent and dynamics of the identified 'breathing motion' might be one of the factors responsible for functional differences in the distinct parvulin subfamilies.

Development of cyclic NGR peptides with thioether linkage: structure and dynamics determining deamidation and bioactivity.

NGR peptides that recognize CD13 receptors in tumor neovasculature are of high interest, in particular due to their potential applications in drug targeting. Here we report the synthesis and structural analysis of novel thioether bond-linked cyclic NGR peptides. Our results show that their chemostability (resistance against spontaneous decomposition forming isoAsp and Asp derivatives) strongly depends on both sample handling conditions and structural properties. A significant correlation was found between chemostability and structural measures, such as NH(Gly)-CO(Asn-sc) distances. The side-chain orientation of Asn is a key determining factor; if it is turned away from HN(Gly), the chemostability increases. Structure stabilizing factors (e.g., H-bonds) lower their internal dynamics, and thus biomolecules become even more resistant against spontaneous decomposition. The effect of cyclic NGR peptides on cell adhesion was examined in A2058 melanoma cell lines. It was found that some of the investigated peptides gradually increased cell adhesion with long-term characteristics, indicating time-dependent formation of integrin binding isoAsp derivatives that are responsible for the adhesion-inducing effect.

¹H, ¹5N, and ¹³C backbone chemical shift assignment of titin domains A59-A60 and A60 alone.

The giant protein titin is the third most abundant protein of vertebrate striated muscle. The titin molecule is >1 µm long and spans half the sarcomere, from the Z-disk to the M-line, and has important roles in sarcomere assembly, elasticity and intracellular signaling. In the A-band of the sarcomere titin is attached to the thick filaments and mainly consists immunoglobulin-like and fibronectin type III-like domains. These are mostly arranged in long-range patterns or 'super-repeats'. The large super-repeats each contain 11 domains and are repeated 11 times, thus forming nearly half the titin molecule. Through interactions with myosin and C-protein, they are involved in thick filament assembly. The importance of titin in muscle assembly is highlighted by the effect of mutations in the A-band portion, which are the commonest cause of dilated cardiomyopathy, affecting ~1 in 250 (Herman et al. in N Engl J Med 366:619-628, 2012). Here we report backbone (15)N, (13)C and (1)H chemical shift and (13)Cß assignments for the A59-A60 domain tandem from the titin A59-A69 large super-repeat, completed using triple resonance NMR. Since, some regions of the backbone remained unassigned in A60 domain of the complete A59-A60 tandem, a construct containing a single A60 domain, A60sd, was also studied using the same methods. Considerably improved assignment coverage was achieved using A60sd due to its lower mass and improved molecular tumbling rate; these assignments also allowed the analysis of inter-domain interactions using chemical shift mapping against A59-A60.

¹H, ¹5N and ¹³C backbone chemical shift assignment of the titin A67-A68 domain tandem.

Single molecules of the giant protein titin extend across half of the muscle sarcomere, from the Z-line to the M-line, and have roles in muscle assembly and elasticity. In the A-band titin is attached to thick filaments and here the domain arrangement occurs in regular patterns of eleven called the large super-repeat. The large super-repeat itself occurs eleven times and forms nearly half the titin molecule. Interactions of the large super-repeats with myosin are consistent with a role in thick filament assembly. Here we report backbone assignments of the titin A67-A68 domain tandem (Fn-Ig) from the third super-repeat (A65-A75) completed using triple resonance NMR experiments.

Local protein backbone folds determined by calculated NMR chemical shifts.

NMR chemical shifts (CSs: δN(NH), δC(α), δC(ß), δC', δH(NH), and δH(α)) were computed for the amino acid backbone conformers (α(L), ß(L), γ(L), δ(L), ε(L), α(D), γ(D), δ(D), and ε(D) [Perczel et al., J Am Chem Soc 1991, 113, 6256]) modeled by oligoalanine structures. Topological differences of the extended fold were investigated on single ß-strands, hairpins with type I and II ß-turns, as well as double- and triple-stranded ß-sheet models. The so-called "capping effect" was analyzed: residues at the termini of a homoconformer sequence unit usually have different CSs than the central residues of an adequately long homoconformer model. In heteroconformer sequences capping effect ruins the direct applicability of several chemical shift types (δH(NH), δC', and δN(NH)) for backbone structure determination of the parent residue. Experimental δH(α), δC(α), and δC(ß) values retrieved from protein database are in good agreement with the relevant computed data in the case of the common backbone conformers (α(L), ß(L), γ(L), and ε(L)), even though neighboring residue effects were not accounted for. Experimental and computed ΔδH(α)-ΔδC(α), ΔδH(α)-ΔδC(ß), and ΔδC(α)-ΔδC(ß) maps give qualitatively the same picture, that is, the positions of the backbone conformers relative to each other are very similar. This indicates that the H(α), C(α), and C(ß) chemical shifts of alanine depend considerably on the backbone fold of the parent residue also in proteins. We provide tabulated CSs of the chiral amino acids that may predict the various structures of the residues.

Structure and stability of short beta-peptide nanotubes: a non-natural representative of collagen?

Since secondary structure elements are known to play a key role in stabilizing the 3D-fold of proteins for the design of non-natural proteins composed of beta-amino acid residues, the construction of suitable secondary structural elements is mandatory. Folding analogues of alpha-helices and beta-strands of beta-polypeptides were already described (Chem. Biodiversity 2004, 1, 1111 (1)). Here, we present several collagen-like folds composed exclusively of beta-Ala(s). Unlike their natural counterpart, these tubular nanostructures can be composed of more than three polypeptide chains aligned parallel and/or antiparallel. By using ab initio and DFT calculations we have optimized a large number of versatile collagen-like antiparallel nanostructures. In these tubular systems, oligopeptide strands are interconnected by i --> (i) type H-bonds, except for the "closing" set. This latter is called "the H-bond zipper" and is either (i) --> i, ( i + 1) --> i, or ( i + 2) --> i type. Antiparallel, tubular foldamers composed of l number of strands, each of k number of beta-amino acid residues (e.g., apbeta-T(l) i+l ) k , ap(beta-T(l) i+1 ) k , or ap(beta-T(l) i+2 ) k ), are unexpectedly stable supramolecular complexes. Independent of k and l, the local backbone fold of the amino acid residues is usually spiral, abbreviated as "S(P)" or "S*(P)". Nevertheless, in contrast to parallel, in antiparallel nanotubes the backbone fold can occasionally twist out from S(P) or S*(P) type into an alternative local structure. However, the more the local geometry of the strands resembles to S(P) or S*(P), the higher the stability is. Besides the backbone twisting, the overall stability is determined by the type and the geometrical properties of the constituent H-bonds. Interestingly, higher number of total H-bonds can provide a lower overall stability, when H-bond parameters are inferior. In general, the increase of both the number of strands and their length stabilize the supramolecular complex. Now that, for beta-peptides, collagen-like overall folds with their stability were determined, their POG- or PPG-like sequence specificity has to be revealed.

A theoretical comparison of self-assembling alpha- and beta-peptide nanostructures: toward design of beta-barrel frameworks.

Self-assembling peptide-based nanotubes are among the most investigated bioactive compounds as a result of their numerous potential applications as novel biomaterials. To support rational bottom-up design of such artificial nanosystems, here we investigate structural and energetic properties of various sheet-derived nanotubes. We carried out high level quantum chemical calculations on large models, composed of up to 32 amino acids, and characterized structures from extended beta-sheets to the molecular framework of beta-barrel proteins. Surprisingly, enzyme-resistant nonnatural beta-peptides have an affinity to form nanotubes that is remarkably higher than that of natural alpha-peptides. We analyzed the stability of both systems depending on (i) parallel or antiparallel orientation, (ii) the number of peptide strands, and (iii) the formed hydrogen bond pattern. Applicability is outlined by investigating guest molecules in the tubes. It is hoped that the structural and energetic data presented here will be effectively used in the design of novel peptide nanosystems.

Structure, enzymatic stability and antitumor activity of sea lamprey GnRH-III and its dimer derivatives.

Direct antitumor activity of sea lamprey (Petromyzon marinus) gonadotropin-releasing hormone III (Glp-His-Trp-Ser-His-Asp-Trp-Lys-Pro-Gly-NH(2); lGnRH-III) was described on several tumor cells. To improve the selectivity of antitumor effects without increasing the hormone releasing activity and to enhance the enzymatic stability, lGnRH-III dimers were prepared via disulfide bond formation. Our results demonstrate that the lGnRH-III dimer derivatives exhibited higher antiproliferative effect and enzymatic stability in comparison with the native lGnRH-III, while lower LH-releasing potency was determined. In order to find a correlation between the biological and structural features of these compounds, the conformation of lGnRH-III and its dimer derivatives was determined by ECD, VCD, FT-IR and (1)H NMR.

Determining suitable lego-structures to estimate stability of larger peptide nanostructures using computational methods.

Nanofibers, nanofilms and nanotubes constructed of one to four strands of oligo-alpha- and oligo-beta-peptides were obtained by using carefully selected building units. Lego-type approaches based on thermoneutral isodesmic reactions can be used to reconstruct the total energies of both linear and tubular periodic nanostructures with acceptable accuracy. Total energies of several different nanostructures were accurately determined with errors typically falling in the subchemical range. Thus, attention will be focused on the description of suitable isodesmic reactions that have enabled the determination of the total energy of polypeptides and therefore offer a very fast, efficient and accurate method to obtain energetic information on large and even very large nanosystems.

Investigation of penetratin peptides. Part 2. In vitro uptake of penetratin and two of its derivatives.

As endocytic uptake of the Antennapedia homeodomain-derived penetratin peptide (RQIKIWFQNRRMKWKK) is finally being revealed, some of the early views about penetratin need to be reconsidered. Endocytic uptake seems to contradict the indispensability of tryptophans and also the minimum length of 16 amino acid residues for efficient internalization. To revise the membrane translocation of penetratin, two penetratin analogs were designed and synthesized: a peptide in which tryptophans were replaced by phenylalanines (Phe(6,14)-penetratin, RQIKIFFQNRRMKFKK) and a shortened analog (dodeca-penetratin, RQIKIWF-R-KWKK) made up of only 12 residues. The peptides were fluorescently labeled and applied to live, unfixed cells from various lines. Cellular uptake was analysed by confocal microscopy and flow cytometry. Low temperature or ATP-depletion blocked the intracellular entry of all three penetratin peptides. A decrease in membrane fluidity or cholesterol depletion with methyl-beta-cyclodextrin greatly inhibited peptide uptake, showing the involvement of cholesterol-rich lipid rafts in internalization. Exogenous heparan sulfate also diminished the internalization of penetratin and its derivatives, reflecting the paramount importance of electrostatic interactions with polyanionic cell-surface proteoglycans. The beneficial presence of tryptophans is supported by observations on the decreased cellular uptake of Phe(6, 14)-penetratin. The maintained translocational efficiency of dodeca-penetratin demonstrates that a thorough understanding of penetratin internalization can yield new penetratin analogs with unaltered translocational abilities. This study provides evidence on the energy-dependent and lipid raft-mediated endocytic uptake of penetratin and highlights the necessity of revealing those pathways that cationic cell-penetrating peptides employ to enter live cells.

Binding crevice for TT-232 in a homology model of type 1 somatostatin receptor.

Somatostatin receptor type 1 was modelled based on the atomic structure of bovine rhodopsin. Possible ways of binding interaction between somatostatin receptor type 1 and TT-232, a cycloheptapeptide analogue of somatostatin with broad therapeutic potential, were analysed by molecular docking. The twelve TT-232 conformations, obtained by NMR measurements in H(2)O-D(2)O mixture, were similar, disclosing a consensus backbone conformation. Several residues interacting with TT-232, such as Val133, Asp137 (helix 3), Arg197 (helix 4), Phe287, Gln291, Asn294 (helix 6), Ser305, and Tyr313 (helix 7), were found. In accordance, in vitro binding experiments indicated high-affinity binding of TT-232 to (125)I labelled somatostatin sites in brain membranes. The single binding crevice obtained by docking may allow the design and discovery of new peptidomimetics of TT-232 in the future.