Mechanisms of action of FdUMP[10]: metabolite activation and thymidylate synthase inhibition.

FdUMP[10] is a multimer of FdUMP, a suicide inhibitor of thymidylate synthase (TS), and was designed to bypass resistance to 5-fluorouracil (5FU). The aim of the study was to compare the effect of FdUMP[10] with 5FU and 5-fluoro-2-deoxyuridine (FUdR) in their efficacy to inhibit their target TS in resistant cells. Therefore cell lines FM3A/0, FM3A/TK- (deficient in thymidine kinase) and FM3A/TS- (deficient in thymidylate synthase) were used to determine TK dependency and specificity for TS inhibition. FdUMP[10] inhibited cell growth with greater potency than 5FU and FdUMP. Direct folate-based inhibitors Raltitrexed, GW1843U89 and Pemetrexed were also evaluated using these cell lines. In TK-deficient cells these folate-based inhibitors had greater potency than the fluoropyrimidines (FPs). Surprisingly, Pemetrexed even inhibited cell growth in TS-deficient cells. Incubation with nucleotidase and phosphatase inhibitors resulted in a reduction of cytotoxicity of FdUMP[10], indicating that the drug can be degraded outside the cells. In the TS in situ inhibition assay (TSIA) 24 h exposure of FM3A cells to 0.5 microM FdUMP and 0.05 microM FdUMP[10] decreased TSIA to 7 and 1% of control. Inhibition of nucleotidase and phosphatase activities reduced the effect of FdUMP[10], while the inhibitory effect was lower in cells lacking TK. FdUMP[10] can enter the cells intact, but also to some extent after dephosphorylation. In conclusion, FdUMP[10] can bypass resistance to FUdR by direct inhibition of TS.

Novel chemical strategies for thymidylate synthase inhibition.

Thymidylate synthase (TS) is a well-validated target for cancer chemotherapy. TS was established as the principal target of the widely used anticancer drug 5-fluorouracil (5FU). The 5FU metabolite FdUMP forms a covalent complex with TS that is stabilized by 5-formyl tetrahydrofolate (leucovorin; LV). Numerous chemical strategies have been employed to develop novel TS inhibitors that are superior to 5FU/LV. 5FU is non-ideal as a TS-inhibitory drug because it is only inefficiently converted to FdUMP, while the remainder of the administered dose is converted to toxic metabolites. My laboratory has explored the utility of FdUMP[N] compounds (oligodeoxynucleotides comprised of FdUMP nucleotides) as FdUMP pro-drugs. FdUMP[N] compounds result in potent TS-inhibition, and display many advantages relative to 5FU/LV. A number of other chemical strategies have also been employed to develop pro-drugs, or metabolic precursors of FdUMP, and several of these strategies will be reviewed. In addition to chemical strategies to develop FdUMP pro-drugs, a number of chemical strategies have been devised to develop molecules that resemble the reduced folate co-factor required for TS catalysis. The synthesis of antifolates that have TS-inhibitory activity, such as Raltitrexed, has resulted in compounds that are effective and specific TS-inhibitors and, in some cases, have clinical potential. Chemical strategies that target TS mRNA for destruction are also being explored as potential chemotherapeutics. These diverse chemical approaches to control TS activity in tumor cells for the treatment of cancer will be reviewed.

Comparison of SH3 and SH2 domain dynamics when expressed alone or in an SH(3+2) construct: the role of protein dynamics in functional regulation.

Protein dynamics play an important role in protein function and regulation of enzymatic activity. To determine how additional interactions with surrounding structure affects local protein dynamics, we have used hydrogen exchange and mass spectrometry to investigate the SH2 and SH3 domains of the protein tyrosine kinase Hck. Exchange rates of isolated Hck SH3 and SH2 domains were compared with rates for the same domains when part of a larger SH(3+2) construct. Increased deuterium incorporation was observed for the SH3 domain in the joint construct, particularly near the SH2 interface and the short sequence that connects SH3 to SH2, implying greater flexibility of SH3 when it is part of SH(3+2). Slow cooperative unfolding of the SH3 domain occurred at the same rate in isolated SH3 as in the SH(3+2) construct, suggesting a functional significance for this unfolding. The SH2 domain displayed relatively smaller changes in flexibility when part of the SH(3+2) construct. These results suggest that the domains influence each other. Further, our results imply a link between functional regulation and structural dynamics of SH3 and SH2 domains.

Solution structure of the human Hck SH3 domain and identification of its ligand binding site.

SH3 domains are protein binding domains that occur widely among signal transduction proteins. Here, we present the NMR-determined solution structure of the SH3 domain from the cytoplasmic protein tyrosine kinase, Hck. Hck is involved in a number of cell signal transduction pathways, frequently in pathways associated with immune response. SH3 domains bind proteins via a left-handed polyproline type II helix on the target protein. We have assessed the structural impact of binding to a ligand through addition of a peptide corresponding to a proline-rich region of a Hck target, the GTPase activating protein of the Ras pathway. Ligand binding effects small structural changes and stabilizes the SH3 domain structure. Also, we have compared the solution structure of the Hck SH3 domain to the crystal structure of Hck, in which the SH3 domain exhibits an intramolecular binding to an interdomain linker region. These structures are interpreted as the apo- and holo- forms of the Hck SH3 domain.

Dynamics of the Hck-SH3 domain: comparison of experiment with multiple molecular dynamics simulations.

Molecular dynamics calculations provide a method by which the dynamic properties of molecules can be explored over timescales and at a level of detail that cannot be obtained experimentally from NMR or X-ray analyses. Recent work (Philippopoulos M, Mandel AM, Palmer AG III, Lim C, 1997, Proteins 28:481-493) has indicated that the accuracy of these simulations is high, as measured by the correspondence of parameters extracted from these calculations to those determined through experimental means. Here, we investigate the dynamic behavior of the Src homology 3 (SH3) domain of hematopoietic cell kinase (Hck) via 5N backbone relaxation NMR studies and a set of four independent 4 ns solvated molecular dynamics calculations. We also find that molecular dynamics simulations accurately reproduce fast motion dynamics as estimated from generalized order parameter (S2) analysis for regions of the protein that have experimentally well-defined coordinates (i.e., stable secondary structural elements). However, for regions where the coordinates are not well defined, as indicated by high local root-mean-square deviations among NMR-determined structural family members or high B-factors/low electron density in X-ray crystallography determined structures, the parameters calculated from a short to moderate length (less than 5-10 ns) molecular dynamics trajectory are dependent on the particular coordinates chosen as a starting point for the simulation.

Sequential assignment and secondary structure determination for the Src homology 2 domain of hematopoietic cellular kinase.

The hematopoietic cellular kinase (Hck) is a member of the Src family of non-receptor protein-tyrosine kinases and participates in signal transduction events regulating the growth, differentiation and function of phagocytes. The secondary structure of the SH2 domain for Hck was determined for a 13C/15N-enriched sample using multi-dimensional NMR spectroscopy. The secondary structure for the domain was determined from chemical shift indices [1H alpha, 13C alpha and 13C'], sequential NOEs [d(alphaN)(i, i+1) and d(NN)(i, i+1)], and 3J(alphaN) scalar coupling constants. The Hck SH2 domain consists of two alpha-helices and seven beta-strands. Complementary strands of beta-sheets were identified from long-range NOEs using a novel 3D, 13C/15N-edited HMQC-NOESY-(HCACO)NH experiment that correlated 1H alpha resonances between beta-strands. The secondary structure for Hck SH2 is similar to that predicted from the sequence alignment of the Src-family protein tyrosine kinases.

Binding of ethidium to DNA measured using a 2D diffusion-modulated gradient COSY NMR experiment.

The binding of ethidium bromide to a DNA hairpin (dU(5)-hairpin) was investigated using a novel 2D diffusion-modulated gradient correlation spectroscopy (DMG-COSY) experiment to evaluate the applicability of this technique for studying the binding of drugs to DNA. The DMG-COSY experiment includes a preparation period during which coherent magnetization is attenuated due to molecular self-diffusion. Magnetization then evolves due to scalar coupling during an evolution delay, and is detected using gradient pulses for coherence selection. The time-domain data are processed in an analogous manner as for gradient-selected COSY experiments. The diffusion coefficient for uridine in DMSO solution was determined from the H5-H6 crosspeak intensities for a series of 2D DMG-COSY experiments that differed in the magnitude of the gradient pulses applied during the preparation period of the DMG-COSY experiment. The diffusion coefficient for uridine calculated from the DMG-COSY experiments was identical (within experimental error) to that determined from 1D diffusion experiments (5.24x10(-6) cm(2)/s at 26 degrees C). The diffusion coefficients for ethidium bromide and for the dU(5)-hairpin were first measured separately using the DMG-COSY experiment, and then measured in the putative complex. The diffusion coefficient for free ethidium bromide (4.15x10(-6) cm(2)/s at 26 degrees C) was considerably larger than for the dU(5)-hairpin (1. 60x10(-6) cm(2)/s at 26 degrees C), as expected for the smaller molecule. The diffusion coefficient for ethidium was markedly decreased upon addition of the dU(5)-hairpin, consistent with complex formation (1.22x10(-6) cm(2)/s at 26 degrees C). Complex formation of 1:1 stoichiometry between ethidium and the stem of the dU(5)-hairpin was verified independently by fluorescence spectroscopy. These results demonstrate the utility of the DMG-COSY experiment for investigating the binding of drugs to DNA in aqueous solution.

NMR spectroscopy as a tool to investigate the structural basis of anticancer drugs.

NMR spectroscopy has been shown to be useful in determining the structures of nucleic acid fragments in solution. Over the last several years NMR spectroscopy, in conjunction with restrained molecular dynamics, has been employed to understand the 3D structures of a number of anticancer drugs and to rationalize their DNA binding behavior. In this review we address the methodologies used most frequently to determine nucleic acid structures in solution. In subsequent sections, we examine how these methods have been applied to rationalize the activities of a number of anticancer agents that target duplex DNA such as cisplatin, bleomycin and calicheamicin. Non-duplex DNA and RNA also represent interesting nucleic acid targets for anticancer drug design and applications of solution NMR spectroscopy to understanding the structures of these types of molecules (e.g. Okazaki fragments, DNA tetraplexes) are also reviewed. In the final sections, advances in NMR methodologies (e.g. linear prediction, superconducting probes) that are likely to impact the research conducted in this area are reviewed. The success of NMR spectroscopy in understanding the structural basis for clinically useful anticancer drugs bodes well for future applications of this methodology not only in rationalization of existing biological activity, but in the design of novel agents that will be useful in treating neoplastic disease.

Prospective utility of cerebral proton magnetic resonance spectroscopy in monitoring HIV infection and its associated neurological impairment.

Neurological manifestations of HIV disease occur in most adults and children with AIDS. Many of those affected will inevitably suffer clinical neurological deficits involving mental function, movement, and sensation. Surprisingly, there are not as yet adequate monitoring systems to predict the onset and/or progression of HIV infection of the CNS. Neurological, neuropsychological, CSF, and magnetic resonance imaging (MRI) analyses cannot accurately detect mental deterioration during advancing HIV disease. Reports suggest that in vivo proton MR spectroscopy (1H MRS) of the brain could be a predictor of virus-induced neurological deterioration. H MRS can measure N-acetylaspartate (NAA), a metabolite present only in neurons. Decreased NAA reflects neuronal loss seen during HIV infection of brain. To uncover possible associations between NAA levels and HIV-induced neurological disease we performed serial 1H MRS brain tests in HIV-infected patients with or at risk for encephalopathy. Serial testing, for 1 year, of 10 patients showed that brain NAA levels decreased in all HIV-infected subjects. The most severe NAA reductions were associated with progressive neurological impairment. These findings suggest that NAA can be used as a noninvasive measure of neuronal loss in patients with HIV disease. Most important, the results suggest that 1H MRS could be used to monitor therapeutics directed against HIV infection within the CNS.

Implications of SH3 domain structure and dynamics for protein regulation and drug design.

SH3 Domains provide interesting targets for investigations of protein structure and dynamics because of their compact size and importance for signal transduction. The present review summarizes recent research investigating SH3 domain structure and dynamics, the discovery of novel SH3 domains, the role of SH3 domains in disease, and progress in targeting SH3 domains for the development of novel therapeutics. Particular emphasis is placed on the unfolding/refolding characteristics of SH3 domains and the potential importance of these processes for regulation of signal transduction.

Intramolecular binding of a proximal PPII helix to an SH3 domain in the fusion protein SH3Hck : PPIIhGAP.

SH3 domains are a conserved feature of many nonreceptor protein tyrosine kinases, such as Hck, and often function in substrate recruitment and regulation of kinase activity. SH3 domains modulate kinase activity by binding to polyproline helices (PPII helix) either intramolecularly or in target proteins. The preponderance of bimolecular and distal interactions between SH3 domains and PPII helices led us to investigate whether proximal placement of a PPII helix relative to an SH3 domain would result in tight, intramolecular binding. We have fused the PPII helix region of human GAP to the C-terminus of Hck SH3 and expressed the recombinant fusion protein in Escherichia coli. The fusion protein, SH3Hck : PPIIhGAP, folded spontaneously into a structure in which the PPII helix was bound intramolecularly to the hydrophobic crevice of the SH3 domain. The SH3Hck : PPIIhGAP fusion protein is useful for investigating SH3: PPII helix interactions, for studying concepts in protein folding and design, and may represent a protein structural motif that is widely distributed in nature.

The affinity of an oligodeoxynucleotide-peptide conjugate for an RNA hairpin loop depends on stereochemistry at the DNA-peptide junction.

Molecular models of an oligodeoxynucleotide-peptide conjugate complexed to an RNA hairpin loop were constructed to assess the effect of stereoisomerism at the point of attachment of the peptide to the oligodeoxynucleotide on the affinity of the conjugate for an RNA target. The peptide portion of the oligodeoxynucleotide-peptide conjugate, (L-lysine)8, was covalently attached to the N-allyl group of (D)- or (L)-aspartic alcohol that was incorporated into the interior of an antisense oligodeoxynucleotide. The stereocenter in the oligodeoxynucleotide interior originates from either (D)- or (L)-aspartic alcohol. The oligodeoxynucleotide portion of the oligodeoxynucleotide-peptide conjugate forms Watson-Crick base pairs with the single-stranded RNA that flanks the RNA hairpin loop. The positively charged peptide makes specific electrostatic contacts with the negatively charged phosphate backbone of the RNA hairpin loop when attached to the N-allyl of (D)-aspartic alcohol but does not have the proper orientation to make these electrostatic contacts when attached to the N-allyl of (L)-aspartic alcohol. This modelling study emphasizes the importance of stereocontrol at the point of branching in synthesizing oligodeoxynucleotide-peptide conjugates for binding of RNA hairpin loops.

NMR structure of the thrombin-binding DNA aptamer stabilized by Sr2+.

The structure of thrombin-binding DNA aptamer complexed with a single Sr2+ ion (Sr2+:TBA complex) has been determined using NMR spectroscopy and restrained molecular dynamics simulations. The quadruplex structure for the Sr2+:TBA complex is similar in topology, but distinct in structure, from that previously reported for the K+:TBA complex. The inter-tetrad distance of the Sr2+:TBA complex is 3.8 angstroms, or 0.7 angstroms larger than in the K+:TBA complex. This substantial difference can be attributed to a different binding site for Sr2+ in the Sr2+:TBA complex than for K+ in the K+:TBA complex. The Sr2+:TBA complex assumes a 1:1 stoichiometry, and it is very likely that the Sr2+ ion simultaneously interacts with the eight O6 atoms of the two G-tetrads. The results indicate that quadruplex DNA structures are highly sensitive to the presence of specific metal ions. The binding of specific metal ions may modulate the biological activity of quadruplex DNA structures in vivo.

Structure of the Sm binding site from human U4 snRNA derived from a 3 ns PME molecular dynamics simulation.

A molecular dynamics simulation of the Sm binding site from human U4 snRNA was undertaken to determine the conformational flexibility of this region and to identify RNA conformations that were important for binding of the Sm proteins. The RNA was fully-solvated (>9,000 water molecules) and charge neutralized by inclusion of potassium ions. A three nanosecond MD simulation was conducted using AMBER with long-range electrostatic forces considered using the particle mesh Ewald summation method. The initial model of the Sm binding site region had the central and 3' stem-loops that flanked the Sm site co-axial with one another, and with the single-stranded Sm binding site region ([I] conformation). During the course of the trajectory, the axes of the 3' stem-loop, and later the central stem-loop, became roughly orthogonal from their original anti-parallel orientation. As these conformational changes occurred, the snRNA adopted first an [L] conformation, and finally a [U] conformation. The [U] conformation was more stable than either the [I] or [L] conformations, and persisted for the final 1 ns of the trajectory. Analysis of the structure resulting from the MD simulations revealed the bulged nucleotide, U114, and the mismatched Ag91-G110 base pair provided distinctive structural features that may enhance Sm protein binding. Based on the results of the MD simulation and the available experimental data, we proposed a mechanism for the binding of the Sm protein sub-complexes to the snRNA. In this model, the D1/D2 and E/F/G Sm protein sub-complexes first bind the snRNA in the [U] conformation, followed by conformational re-arrangement to the [I] conformation and binding of the D3/B Sm protein sub-complex.

The 5' stem-loop from human U4 snRNA adopts a novel conformation stabilized by G-C and G-U base pairs.

1H NMR and molecular modeling studies of the 5' stem-loop from human U4 snRNA were undertaken to determine the conformation of this stem-loop that is essential for spliceosome formation and pre-mRNA splicing. Sixteen of the 35 nucleotides of this stem-loop are in the loop region and inspection of the loop sequence revealed no decomposition into elements of secondary structure commonly found in other RNA stem-loops. An analysis of possible base pairing interactions for this stem-loop using the methods of Zuker revealed the lowest energy secondary structure for the 16 nucleotide loop consisted of four base pairs at the base of a non-canonical tetraloop (UUUA). This shorter stem-loop was joined to the nine base pair stem by two A residues on the 5' side and a single bulged A on the 3' side. Both stems also had bulged A residues. 1H NMR experiments performed on solutions of the 35 mer stem-loop, the stem region, and the loop region confirmed the 35 mer adopted this secondary structure in solution. A 3D molecular model of this structure consistent with the NMR data was generated to assist in visualization of this novel structure.

A linear 23-residue peptide reveals a propensity to form an unusual native-like conformation.

To gain insight into the earliest events of protein folding, a 23-residue peptide with a sequence corresponding to the 38-60 fragment of the beta-subunit of human chorionic gonadotropin (hCG beta) was studied by NMR. In aqueous solution the majority of the peptide residues adopted an extended polyproline II (PII) conformation similar to those in mature, fully folded hCG beta. The finding that the isolated protein fragment may acquire native-like structural motifs, even without alpha-helices or beta-structures, extends the possibility of using free peptides as model systems to better understand the protein folding mechanisms. It was shown that the PII-rich structural motif can be determined efficiently by NMR spectroscopy. The observation that in the absence of extensive medium- and long-range interactions the majority of amino acid residues may adopt the PII conformation suggests that the PII-rich structural motifs may play an important role in early events of protein folding.

Shape-selective recognition of a model Okazaki fragment by geometrically-constrained bis-distamycins.

Okazaki fragments represent interesting targets for the design of anticancer drugs because of their selective occurrence during DNA replication, a process often elevated in aggressive malignancies. Structural studies have indicated a bend occurs in the helical axis at the junction region (JR) that joins the DNA duplex region (DDR) and the RNA-DNA hybrid duplex region (HDR) of model Okazaki fragments. To identify a structural motif that provides a shape complementary to the Okazaki fragment minor groove, we have investigated the binding of geometrically-constrained bis-distamycins to a model Okazaki fragment, [OKA], with a sequence derived from the genome of simian virus 40 (SV40). Both the JR and the DDR of [OKA] contain consecutive A/T base pairs that could accommodate distamycin binding. Of the six bis-distamycins selected for analysis, the two with a para configuration of the distamycins on the benzene or pyridine scaffold bound [OKA] tightly (Kd approximately 10(-6) M from gel-shift assays; Kd approximately 10(-8) M from deltaT(M)) while the four with a meta orientation did not bind. The two mono-distamycins studied also did not bind [OKA]. Molecular modeling of the complex between the para bis-distamycin MT-9 and [OKA] revealed MT-9 adopted an S- shape complementary to the minor groove of the model Okazaki fragment.

Antineoplastic activity of a novel multimeric gemcitabine-monophosphate prodrug against thyroid cancer cells in vitro.

Gemcitabine (Gem) is a deoxycytidine analog that is effective against pancreatic cancer and other malignancies following conversion to the 5'-O-mono-, di- and tri-phosphate forms. We evaluated the cytotoxicity of GemMP[10], a novel multimeric form of 2'-deoxy-2',2"-difluorocytidine-5'-O-monophosphate (gemcitabine monophosphate) against three thyroid carcinoma cell lines established from anaplastic (8505C), papillary (B-CPAP) and poorly-differentiated papillary (BHT-101) cancer. GemMP[10] decreased tumor cell growth at concentrations ranging from 1 to 50 nM. These concentrations were 5- to 10-fold lower than those required for inhibition of tumor cell growth by monomeric Gem. GemMP[10] cytotoxicity occurred via induction of apoptosis. Flow cytometric analysis of GemMP[10] treated cells revealed growth arrest in S-phase. Fas-antigen expression was increased in thyroid cancer cells treated with GemMP[10], whereas Fas-L and Bcl-2 expression were not significantly affected. These results demonstrated that GemMP[10] is a potent cytotoxic agent that serves to induce apoptosis in association with increased Fas expression in cultured thyroid cancer cell lines.

Shape-selective binding of geometrically-constrained bis-distamycins to a DNA duplex and a model Okazaki fragment of identical sequence.

The binding of ligands to nucleic acids is of great interest for the control of gene expression and other nucleic acid mediated processes. We have evaluated the binding of several geometrically-constrained bis-distamycins to a model Okazaki fragment [OKA], or a DNA duplex having identical base sequence [DD], using gel-shift assays, optical spectroscopy and differential scanning calorimetry. In the case of covalent attachment of two distamycins to a central benzene ring, a similar binding profile was observed for [DD] as was observed for [OKA] (para binds [K(app) > 10(6) M(-1)], meta binds only weakly). For a central pyridyl ring, however, clear distinction between the binding to [DD] and binding to [OKA] was observed. While none of the three meta isomers having a central pyridyl ring bound [OKA], two of them (MT-17 and MT-12) bound [DD] [K(app) > 10(6) M(-1)]. These results demonstrate subtle differences in lexitropsin shape and placement of electronegative atoms may result in selective binding to a nucleic acid duplex based both on base sequence and chemical composition. Selective binding to DNA duplexes may be useful for designing ligands that regulate transcription, but do not interfere in other nucleic acid mediated processes.

Quantum mechanical calculations and experimental measurement of N-terminal charge effects on 1HN and 1HC alpha chemical shifts in peptides.

Nuclear magnetic resonance spectroscopists are increasingly utilizing chemical shifts to characterize the secondary structure of proteins. The present study addresses the effects that the positively charged amino group at the N-terminus of a peptide has on 1HN and 1HC alpha chemical shifts along the chain. This information is necessary for interpreting chemical shift data for proteins and/or for peptides that are used as models for protein structure. The chemical shifts for the 1H resonances of four peptides that differ only in the location of their N-terminii are assigned using two-dimensional nmr spectroscopy. The peptides have sequences derived from the beta subunit of the glycoprotein hormone human chorionic gonadotropin (hCG-beta). Comparison of the 1HN and the 1HC alpha chemical shifts for residues common to all four peptides reveals downfield shifts for 1HN and the 1HC alpha resonances within three residues of the N-terminus compared with chemical shifts in the interior of the peptide. The magnitude of the downfield shift is larger for resonances nearer the N-terminus. Quantum mechanical calculations of the 1HN and 1HC alpha chemical shifts in peptides constructed with six alanine units also predict a significant terminus effect. The calculations agree both qualitatively and quantitatively with the experimental data. The inductive nature of the end effect is confirmed in the calculations by Mulliken population analysis. End effects should be taken into account in determining protein secondary structures from chemical shifts.