Project 1: Chemistry and Biology of Non-Duplex DNA
Watson and Crick's model of fully hydrated (B-form) duplex DNA inspired our mechanistic understanding of the replication and flow of genetic information known as the central dogma of biology.[1] Since then, DNA has been regarded by most as a uniform double helix and a passive library of genetic information. DNA structure, however, is highly dynamic and its functions are potentially diverse.[2] Recent evidence suggests that non-duplex DNA structures may play a direct role in regulating gene expression, chromosome stability, cellular replication, and programmed cell death.
From the early days of structural biology, DNA was recognized as having the ability to adopt alternate folds.[3] In addition to A-, B-, and Z-form double helices, single-stranded DNA can form various hairpin, G-quadruplex and i-motif structures.[2] While the folding and function of these DNA structures are not well understood in vivo, the identification of proteins that selectively bind to either G-quadruplex[4] or i-motif DNA[5] in vitro and in vivo[6] provides strong evidence that these DNA structures are biologically relevant.
Research in our group is focused on the design, synthesis, and evaluation of chemical probes for diverse secondary and tertiary structures of DNA such as hairpins, G-quadruplexes, and i-motifs. Structure-specific probes capable of reporting and/or stabilizing these folds will provide important new tools for expanding our current understanding of DNA biology, and may provide important new leads for anti-cancer agents.
G-quadruplex DNA:
According to a recent survey, over 30,000 repetitive G-rich sequences with the motif (G3+N1-7)4 are dispersed throughout the human genome.[7] While the exact sequence requirements for G-quadruplex formation are poorly understood,[8] highly diverse single-stranded sequences containing this motif can fold into various types of G-quadruplex structures (some examples shown below).[9] These types of DNA structures have been implicated in regulating gene transcription,[10] recombination,[11] viral integration,[12] chromosome stability,[13] and cellular senescence.[14]
Molecules that bind to G-quadruplex and/or i-motif structures may inhibit cancer growth by at least two distinct mechanisms. First, the over-expression of oncogenes like c-Myc, c-Kit, and KRAS might be down-regulated by small-molecule stabilization of G-quadruplex/i-motif structures in the promoter regions of these genes.[15],[16] The second, more extensively studied mechanism, is the inhibition of telomerase, a ribonucleoprotein complex that catalyzes the 3' extension of telomeric DNA, which is essential for the continued division of most immortalized cells including cancer tissues.[17]
Telomeric DNA:
Human telomeres are nucleoprotein complexes containing the repeated DNA sequence (5'GGGTTA3')n (n = 100 - 3,000) that is paired with its complemeent strand, except for a short (24 - 400 base) single-stranded 3' overhang.[17] Single-stranded DNA containing this sequence can fold into intramolecular G-quadruplexes in vitro.[18] Telomeric DNA gets shortened by each successive round of cellular replication, possibly resulting in a molecular clock that limits the total number of replicative cycles a normal cell can undergo.[19] To overcome this "end replication problem," most cancer tissues maintain genomic stability by telomerase-mediated extension of telomeric DNA.[20] Since telomerase is expressed in the majority of malignant tumor cells and in relatively few somatic cells, it was recognized as a potential cancer-specific target.[20] With this in mind, some researchers have developed small-molecule inhibitors of telomerase by targeting its DNA substrate, the 3' overhangs of telomeres.[21,22,23,24] These studies provided evidence that the stabilization of G-quadruplex structures via small-molecule binding can effectively inhibit telomerase activity; and when applied to cells, these compounds can initiate apoptosis or replicative senescence. The G-quadruplex ligands used in these studies, however, have "frustratingly low quadruplex affinity"[23] and bind to other nucleic acids with similar affinity,[25] making alternate mechanisms for their biological effects possible.
Deciphering the potential relationships between telomerase inhibition, telomere shortening, genomic instability, senescence, and apoptosis is complicated by the very gradual loss of telomeric DNA that typically results from a continuous application of a telomerase antagonist.[24] In addition, some cell types (including 10-20% of cancer tissues) have one or more non-telomerase-based mechanisms for telomere extension.[26] The biological effects of even an ideal telomerase inhibitor can therefore be slow to appear or totally absent.[27] Along with the relatively low affinity and/or specificity of available G-quadruplex ligands,[23] the validation of telomeric DNA as a viable drug target has remained elusive.[21,22]
G-quadruplex ligands:
Three of the best characterized G-quadruplex ligands are N-methyl mesoporphyrin (NMM),[25] tetra(N-methyl-4-pyridyl)porphine (TMPyP4),[25,28] and Telomestatin.[29] The structures of these compounds are shown below. The cationic porphyrin TMPyP4 is the most extensively studied G-quadruplex ligand to date. Telomerase inhibition by TMPyP4 is relatively insensitive to metal coordination by the porphyrin,[30] but is highly dependent on the identities of the groups at the meso positions.[31] Extensive substitution of the meso positions with groups other than pyridinium failed to generate compounds with improved telomerase inhibition activity, but the resulting SARs illustrated that base-stacking and charge-charge interactions are important for porphyrin-DNA binding.[32] Like many porphyrins, TMPyP4 exhibits some anticancer activities in vivo,[33] but it exhibits very poor DNA specificity,[25] causes anaphase bridges in sea urchin embryos,[34] and (like other photosensitizers) it has problems with long-lasting skin toxicity.[35] New macrocylic ligands have been designed to circumvent these problems. The synthesis and evaluation of two new families of G-quadruplex ligands containing cationic phthalocyanines is being persued. These molecules have either a variable charge density, or the ability to make groove contacts with G-quadruplex DNA. The G-quadruplex affinity and specificity of these compounds will be evaluated using surface plasmon resonance, fluorescence spectroscopy, and equilibrium dialysis. Compounds with promising DNA specificity will be applied to cells, and changes in telomere length, chromosome stability, and viability will be determined using established methods.
The natural product telomestatin, is one of the most potent small-molecule inhibitors of telomerase reported to date.[29] Telomestatin is selective for intra- over intermolecular G-quadruplexes, while TMPyP4 has the opposite selectivity.[36] Telomestatin binds to G-quadruplexes in vitro (Kd = ~20 nM),[37] and induces telomere shortening in treated cells more rapidly than is expected for a single mechanism involving telomerase inhibition.[38] Importantly, telomestatin induces apoptosis in a number of different tumor cell types in animal models while exhibiting little toxicity towards normal progenitor cells.[39] No structure-activity relationships have previously been reported for telomestatin due to the difficulties encountered with its synthesis.[40] A straightforward route for the total synthesis of telomestatin is being pursued that allows, in its most general form, any dipeptide combination of Ser, Thr, and Cys to be converted into highly variable macrocycles, thus providing a potential approach for making large combinatorial libraries of macrocyclic polyzoles.
Project 2: Fluorescent Detection of DNA Folding
Fluorescent techniques provide a powerful, yet readily accessible technology for deciphering biomolecule folding and activity.[41] The adenosine isostere 2-aminopurine (2-AP) is the most commonly used internal probe for nucleic acid folding and binding.[42] 2-AP exhibits a high quantum yield in water (~0.7) but is quenched upon its incorporation into duplex DNA,[43] folded RNA,[44] and G-quadruplex DNA.[45] Given the dramatic changes in guanosine hydration upon G-quadruplex folding, fluorescent analogs of guanosine might be ideal candidates for monitoring DNA folding and binding. Few examples of fluorescent G-analogs are currently found in the literature and all have disrupted Watson-Crick and/or Hoogstein base-pairing faces, both of which are needed for G-tetrad formation. A recent paper by Prof. Yitzhak Tor and Nicolas Greco summarize some of the fluorescent base analogs (shown below) that have been incorporated into DNA for subsequent photophysical studies.[46]
In collaboration with Prof. Andreas Vasella (ETH, Zurich) we are synthesizing a variety of fluorescent base analogs for subsequent incorporation into G-quadruplex DNA at positions known to favor either a syn- or anti-conformation. The quantum yields of each construct will be determined in its single-standed, duplex, and G-quadruplex forms to determine which guanosine derivatives can accurately report DNA folding and ligand binding. Following optimization, we will evaluate the suitability of this approach for generating fluorescently labeled G-quadruplex DNA in the telomeric regions of live cells, by utilizing available ALT+ cell lines that are known to extend telomeric regions through homologous recombination with foreign DNA substrates containing multiple copies of a potentially fluorescent, telomeric repeat sequence.[47]
Project 3: Synthesis of Defined DNA-DNA Interstrand Cross-Links
Preparing well-defined DNA-DNA interstrand cross-links (ICLs) is a challenging synthetic problem, and an important prerequisite for studies aimed at deciphering the mechanisms by which ICLs are identified and repaired. These repair pathways are critical aspects of normal cellular function, and have a profound impact on the anti-cancer and mutagenic activities of bifunctional alkylating agents. Currently, two new synthetic strategies for generating defined ICLs that result from chloroethyl nitrosourea (CENU) and nitrogen mustard treatment of DNA are being developed. These approaches utilize single-stranded DNAs containing stable ICL precursors prepared from modified phosphoramidites and solid-phase DNA synthesis. Following hybridization with a complimentary stand of DNA, the ICL precursor is photochemically activated and it reacts with a base in the opposing strand. Since the position of the ICL is determined by the placement of the of the ICL precursor, a single product can be isolated. In addition, since cross-link formation is the last step of this synthesis, chemically unstable ICLs (for example the acid-labile ICL’s resulting from formaldehyde treatment) can be generated and maintained under neutral conditions. This approach, once validated, should provide a flexible route for preparing milligram quantities of ICL DNA, and may provide the first approach for triggering formation of well-defined ICLs inside of living cells.