Programming chemical kinetics: engineering dynamic reaction networks with DNA strand displacement

Over the last century, the silicon revolution has enabled us to build faster, smaller and more sophisticated computers. Today, these computers control phones, cars, satellites, assembly lines, and other electromechanical devices. Just as electrical wiring controls electromechanical devices, living organisms employ "chemical wiring" to make decisions about their environment and control physical processes. Currently, the big difference between these two substrates is that while we have the abstractions, design principles, verification and fabrication techniques in place for programming with silicon, we have no comparable understanding or expertise for programming chemistry.

In this thesis we take a small step towards the goal of learning how to systematically engineer prescribed non-equilibrium dynamical behaviors in chemical systems. We use the formalism of chemical reaction networks (CRNs), combined with mass-action kinetics, as our programming language for specifying dynamical behaviors. Leveraging the tools of nucleic acid nanotechnology (introduced in Chapter 1), we employ synthetic DNA molecules as our molecular architecture and toehold-mediated DNA strand displacement as our reaction primitive.

Abstraction, modular design and systematic fabrication can work only with well-understood and quantitatively characterized tools. Therefore, we embark on a detailed study of the "device physics" of DNA strand displacement (Chapter 2). We present a unified view of strand displacement biophysics and kinetics by studying the process at multiple levels of detail, using an intuitive model of a random walk on a 1-dimensional energy landscape, a secondary structure kinetics model with single base-pair steps, and a coarse-grained molecular model that incorporates three-dimensional geometric and steric effects. Further, we experimentally investigate the thermodynamics of three-way branch migration. Our findings are consistent with previously measured or inferred rates for hybridization, fraying, and branch migration, and provide a biophysical explanation of strand displacement kinetics. Our work paves the way for accurate modeling of strand displacement cascades, which would facilitate the simulation and construction of more complex molecular systems.

In Chapters 3 and 4, we identify and overcome the crucial experimental challenges involved in using our general DNA-based technology for engineering dynamical behaviors in the test tube. In this process, we identify important design rules that inform our choice of molecular motifs and our algorithms for designing and verifying DNA sequences for our molecular implementation. We also develop flexible molecular strategies for "tuning" our reaction rates and stoichiometries in order to compensate for unavoidable non-idealities in the molecular implementation, such as imperfectly synthesized molecules and spurious "leak" pathways that compete with desired pathways.

We successfully implement three distinct autocatalytic reactions, which we then combine into a de novo chemical oscillator. Unlike biological networks, which use sophisticated evolved molecules (like proteins) to realize such behavior, our test tube realization is the first to demonstrate that Watson-Crick base pairing interactions alone suffice for oscillatory dynamics. Since our design pipeline is general and applicable to any CRN, our experimental demonstration of a de novo chemical oscillator could enable the systematic construction of CRNs with other dynamic behaviors.

Recognition of double helical DNA by purine oligonucleotides via triple helix formation

Oligonucleotide-directed triple helix formation is one of the most versatile methods for the sequence specific recognition of double helical DNA. Chapter 2 describes affinity cleaving experiments carried out to assess the recognition potential for purine-rich oligonucleotides via the formation of triple helices. Purine-rich oligodeoxyribonucleotides were shown to bind specifically to purine tracts of double helical DNA in the major groove antiparallel to the purine strand of the duplex. Specificity was derived from the formation of reverse Hoogsteen G•GC, A•AT and T•AT triplets and binding was limited to mostly purine tracts. This triple helical structure was stabilized by multivalent cations, destabilized by high concentrations of monovalent cations and was insensitive to pH. A single mismatched base triplet was shown to destabilize a 15 mer triple helix by 1.0 kcal/mole at 25°C. In addition, stability appeared to be correlated to the number of G•GC triplets formed in the triple helix. This structure provides an additional framework as a basis for the design of new sequence specific DNA binding molecules.

In work described in Chapter 3, the triplet specificities and required strand orientations of two classes of DNA triple helices were combined to target double helical sequences containing all four base pairs by alternate strand triple helix formation. This allowed for the use of oligonucleotides containing only natural 3'-5' phosphodiester linkages to simultaneously bind both strands of double helical DNA in the major groove. The stabilities and structures of these alternate strand triple helices depended on whether the binding site sequence was 5'-(purine)_m (pyrimidine)_n-3' or 5'- (pyrimidine)_m (purine)_n-3'.

In Chapter 4, the ability of oligonucleotide-cerium(III) chelates to direct the transesterfication of RNA was investigated. Procedures were developed for the modification of DNA and RNA oligonucleotides with a hexadentate Schiff-base macrocyclic cerium(III) complex. In addition, oligoribonucleotides modified by covalent attachment of the metal complex through two different linker structures were prepared. The ability of these structures to direct transesterification to specific RNA phosphodiesters was assessed by gel electrophoresis. No reproducible cleavage of the RNA strand consistent with transesterification could be detected in any of these experiments.

Distinct intron DNA structures in simian virus 40 T-antigen and adenovirus 2 E1A genes

Distinct structures delineating the introns of Simian Virus 40 T-antigen and Adenovirus 2 E1A genes have been discovered. The structures, which are centered around the branch points of the genes inserted in supercoiled double-stranded plasmids, are specifically targeted through photoactivated strand cleavage by the metal complex tris(4,7-diphenyl-1,10-phenanthroline)rhodium(III). The DNA sites that are recognized lack sequence homology but are similar in demarcating functionally important sites on the RNA level. The single-stranded DNA fragments corresponding to the coding strands of the genes were also found to fold into a structure apparently identical to that in the supercoiled genes based on the recognition by the metal complex. Further investigation of different single-stranded DNA fragments with other structural probes, such as another metal complex bis(1,10-phenanthroline)(phenanthrenequinone diimine)rhodium(III), AMT (4'aminomethyl-4,5',8 trimethylpsoralen), restriction enzyme Mse I, and mung bean nuclease, showed that the structures require the sequ ences at both ends of the intron plus the flanking sequences but not the middle of the intron. The two ends form independent helices which interact with each other to form the global tertiary structures. Both of the intron structures share similarities to the structure of the Holliday junction, which is also known to be specifically targeted by the former metal complex. These structures may have arisen from early RNA intron structures and may have been used to facilitate the evolution of genes through exon shuffling by acting as target sites for recombinase enzymes.

Cleavage of DNA by polyamide-seco-CBI conjugates

Small molecules that bind to any predetermined DNA sequence in the human genome are potentially useful tools for molecular biology and human medicine. Polyamides containing N-methylimidazole (Im) N-methylpyrrole (Py) are cell permeable small molecules that bind DNA according to a set of "pairing rules" with affinities and specificities similar to many naturally occurring DNA binding proteins. Py-Im polyamides offer a general approach to the chemical regulation of gene expression. We demonstrate here that polyamide containing a DNA alkylating moiety seco-CBI can specifically direct sequence specific DNA alkylation. We can also control the strand of DNA that is alkylated, depending on the enantiomer of seco-CBI used and the orientation of the polyamide relative to the alkylation site (Chapter 2). This class of molecules has been applied to a gene repair system in collaboration with the Baltimore group at Caltech (Chapter 3). Also reported are additional seco-CBI polyamide conjugates synthesized to study other systems (HIV-1 and COX-2) (Appendix 1).

Structure-, stereochemistry-, and metal-regulated DNA binding/cleaving molecules

In the cell, the binding of proteins to specific sequences of double helical DNA is essential for controlling the processes of protein synthesis (at the level of DNA transcription) and cell proliferation (at the level of DNA replication). In the laboratory, the sequence-specific DNA binding/cleaving properties of restriction endonuclease enzymes (secreted by microorganisms to protect them from foreign DNA molecules) have helped to fuel a revolution in molecular biology. The strength and specificity of a protein:DNA interaction depend upon structural features inherent to the protein and DNA sequences, but it is now appreciated that these features (and therefore protein:DNA complexation) may be altered (regulated) by other protein:DNA complexes, or by environmental factors such as temperature or the presence of specific organic molecules or inorganic ions. It is also now appreciated that molecules much smaller than proteins (including antibiotics of molecular weight less than 2000 and oligonucleotides) can bind to double-helical DNA in sequence-specific fashion. Elucidation of structural motifs and microscopic interactions responsible for the specific molecular recognition of DNA leads to greater understanding of natural processes and provides a basis for the design of novel sequence-specific DNA binding molecules. This thesis describes the synthesis and DNA binding/cleaving characteristics of molecules designed to probe structural, stereochemical, and environmental factors that regulate sequence-specific DNA recognition.

Chapter One introduces the DNA minor groove binding antibiotics Netropsin and Distamycin A, which are di- and tri(N-methylpyrrolecarboxamide) peptides, respectively. The method of DNA affinity cleaving, which has been employed to determine DNA binding properties of designed synthetic molecules is described. The design and synthesis of a series of Netropsin dimers linked in tail-to-tail fashion (by oxalic, malonic, succinic, or fumaric acid), or in head-to-tail fashion (by glycine, β-alanine, and γ-aminobutanoic acid (Gaba)) are presented. These Bis(Netropsin)s were appended with the iron-chelating functionality EDTA in order to make use of the technique of DNA affinity cleaving. Bis(Netropsin)-EDTA compounds are analogs of penta(N-methylpyrrolecarboxamide)-EDTA (P5E), which may be considered a head-to-tail Netropsin dimer linked by Nmethylpyrrolecarboxamide. Low- and high-resolution analysis of pBR322 DNA affinity cleaving by the iron complexes of these molecules indicated that small changes in the length and nature of the linker had significant effects on DNA binding/cleaving efficiency (a measure of DNA binding affinity). DNA binding/cleaving efficiency was found to decrease with changes in the linker in the order β-alanine > succinamide > fumaramide > N-methylpyrrolecarboxamide > malonamide >glycine, γ-aminobutanamide > oxalamide. In general, the Bis(Netropsin)-EDTA:Fe compounds retained the specificity for seven contiguous A:T base pairs characteristic of P5E:Fe binding. However, Bis(Netropsin)Oxalamide- EDTA:Fe exhibited decreased specificity for A:T base pairs, and Bis(Netropsin)-Gaba-EDT A:Fe exhibited some DNA binding sites of less than seven base pairs. Bis(Netropsin)s linked with diacids have C₂-symmmetrical DNA binding subunits and exhibited little DNA binding orientation preference. Bis(Netropsin)s linked with amino acids lack C₂-symmetrical DNA binding subunits and exhibited higher orientation preferences. A model for the high DNA binding orientation preferences observed with head-to-tail DNA minor groove binding molecules is presented.

Chapter Two describes the design, synthesis, and DNA binding properties of a series of chiral molecules: Bis(Netropsin)-EDTA compounds with linkers derived from (R,R)-, (S,S)-, and (RS,SR)-tartaric acids, (R,R)-, (S,S)-, and (RS,SR)-tartaric acid acetonides, (R)- and (S)-malic acids, N ,N-dimethylaminoaspartic acid, and (R)- and (S)-alanine, as well as three constitutional isomers in which an N-methylpyrrolecarboxamide (P1) subunit and a tri(N-methylpyrrolecarboxamide)-EDTA (P3-EDTA) subunit were linked by succinic acid, (R ,R)-, and (S ,S)-tartaric acids. DNA binding/cleaving efficiencies among this series of molecules and the Bis(Netropsin)s described in Chapter One were found to decrease with changes in the linker in the order β-alanine > succinamide > P1-succinamide-P3 > fumaramide > (S)-malicamide > N-methylpyrrolecarboxamide > (R)-malicamide > malonamide > N ,N-dimethylaminoaspanamide > glycine = Gaba = (S,S)-tartaramide = P1-(S,S)-tanaramide-P3 > oxalamide > (RS,SR)-tartaramide = P1- (R,R)-tanaramide-P3 > (R,R)-tartaramide (no sequence-specific DNA binding was detected for Bis(Netropsin)s linked by (R)- or (S)-alanine or by tartaric acid acetonides). The chiral molecules retained DNA binding specificity for seven contiguous A:T base pairs. From the DNA affinity cleaving data it could be determined that: 1) Addition of one or two substituents to the linker of Bis(Netropsin)-Succinamide resulted in stepwise decreases in DNA binding affinity; 2) molecules with single hydroxyl substituents bound DNA more strongly than molecules with single dimethylamino substituents; 3) hydroxyl-substituted molecules of (S) configuration bound more strongly to DNA than molecules of (R) configuration. This stereochemical regulation of DNA binding is proposed to arise from the inherent right-handed twist of (S)-enantiomeric Bis(Netropsin)s versus the inherent lefthanded twist of (R)-enantiomeric Bis(Netropsin)s, which makes the (S)-enantiomers more complementary to the right-handed twist of B form DNA.

Chapter Three describes the design and synthesis of molecules for the study of metalloregulated DNA binding phenomena. Among a series of Bis(Netropsin)-EDTA compounds linked by homologous tethers bearing four, five, or six oxygen atoms, the Bis(Netropsin) linked by a pentaether tether exhibited strongly enhanced DNA binding/cleaving in the presence of strontium or barium cations. The observed metallospecificity was consistent with the known affinities of metal cations for the cyclic hexaether 18-crown-6 in water. High-resolution DNA affinity cleaving analysis indicated that DNA binding by this molecule in the presence of strontium or barium was not only stronger but of different sequence-specificity than the (weak) binding observed in the absence of metal cations. The metalloregulated binding sites were consistent with A:T binding by the Netropsin subunits and G:C binding by a strontium or barium:pentaether complex. A model for the observed positive metalloregulation and novel sequence-specificity is presented. The effects of 44 different cations on DNA affinity cleaving by P5E:Fe were examined. A series of Bis(Netropsin)-EDTA compounds linked by tethers bearing two, three, four, or five amino groups was also synthesized. These molecules exhibited strong and specific binding to A:T rich regions of DNA. It was found that the iron complexes of these molecules bound and cleaved DNA most efficiently at pH 6.0-6.5, while P5E:Fe bound and cleaved most efficiently at pH 7.5-8.0. Incubating the Bis(Netropsin) Polyamine-EDTA:Fe molecules with K₂PdCl₄ abolished their DNA binding/cleaving activity. It is proposed that the observed negative metalloregulation arises from kinetically inert Bis(Netropsin) Polyamine:Pd(II) complexes or aggregates, which are sterically unsuitable for DNA complexation. Finally, attempts to produce a synthetic metalloregulated DNA binding protein are described. For this study, five derivatives of a synthetic 52 amino acid residue DNA binding/cleaving protein were produced. The synthetic mutant proteins carried a novel pentaether ionophoric amino acid residue at different positions within the primary sequence. The proteins did not exhibit significant DNA binding/cleaving activity, but they served to illustrate the potential for introducing novel amino acid residues within DNA binding protein sequences, and for the development of the tricyclohexyl ester of EDTA as a superior reagent for the introduction of EDT A into synthetic proteins.

Chapter Four describes the discovery and characterization of a new DNA binding/cleaving agent, [SalenMn(III)]OAc. This metal complex produces single- and double-strand cleavage of DNA, with specificity for A:T rich regions, in the presence of oxygen atom donors such as iodosyl benzene, hydrogen peroxide, or peracids. Maximal cleavage by [SalenMn(III)]OAc was produced at pH 6-7. A comparison of DNA singleand double-strand cleavage by [SalenMn(III)]⁺ and other small molecules (Methidiumpropyl-EDTA:Fe, Distamycin-EDTA:Fe, Neocarzinostatin, Bleomycin:Fe) is presented. It was found that DNA cleavage by [SalenMn(III)]⁺ did not require the presence of dioxygen, and that base treatment of DNA subsequent to cleavage by [SalenMn(III)]⁺ afforded greater cleavage and alterations in the cleavage patterns. Analysis of DNA products formed upon DNA cleavage by [SalenMn(III)] indicated that cleavage was due to oxidation of the sugar-phosphate backbone of DNA. Several mechanisms consistent with the observed products and reaction requirements are discussed.

Chapter Five describes progress on some additional studies. In one study, the DNA binding/cleaving specificities of Distamycin-EDTA derivatives bearing pyrrole N-isopropyl substituents were found to be the same as those of derivatives bearing pyrrole N-methyl substituents. In a second study, the design of and synthetic progress towards a series of nucleopeptide activators of transcription are presented. Five synthetic plasmids designed to test for activation of in vitro run-off transcription by DNA triple helix-forming oligonucleotides or nucleopeptides are described.

Chapter Six contains the experimental documentation of the thesis work.

The interactions of basic proteins and DNA

I. ELECTROPHORESIS OF THE NUCLEIC ACIDS

A zone electrophoresis apparatus using ultraviolet optics has been constructed to study nucleic acids at concentrations less than 0.004%. Native DNA has a mobility about 15% higher than denatured DNA over a range of conditions. Otherwise, the electrophoretic mobility is independent of molecular weight, base composition or source. DNA mobilities change in the expected way with pH but the fractional change in mobility is less than the calculated change in charge. A small decrease in mobility accompanies an increase in ionic strength. RNA’s from various sources have mobilities slightly lower than denatured DNA except for s-RNA which travels slightly faster. The important considerations governing the mobility of nucleic acids appear to be the nature of the hydrodynamic segment, and the binding of counterions. The differences between electrophoresis and sedimentation stem from the fact that all random coil polyelectrolytes are fundamentally free draining in electrophoresis.

II. THE CYTOCHROME C/DNA COMPLEX

The basic protein, cytochrome c, has been complexed to DNA. Up to a cytochrome:DNA mass ratio of 2, a single type of complex is formed. Dissociation of this complex occurs between 0.05F and 0.1F NaCl. The complexing of cytochrome to DNA causes a slight increase in the melting temperature of the DNA, and a reduction of the electrophoretic mobility proportional to the decrease in net charge. Above a cytochrome:DNA mass ratio of 2.5, a different type of complex is formed. The results suggest that complexes such as are formed in the Kleinschmidt technique of electron microscopy would not exist in bulk solution and are exclusively film phenomena.

III. STUDIES OF THE ELECTROPHORESIS AND MELTING BEHAVIOUR OF NUCLEOHISTONES

Electrophoresis studies on reconstituted nucleohistones indicate that the electrophoretic mobility for these complexes is a function of the net charge of the complex. The mobility is therefore dependent on the charge density of the histone complexing the DNA, as well as on the histone/DNA ratio. It is found that the different histones affect the transition from native to denatured DNA in different ways. It appears that histone I is exchanging quite rapidly between DNA molecules in 0.01 F salt, while histone II is irreversibly bound. Histone III-IV enhances the capacity of non-strand separated denatured DNA to reanneal. Studies on native nucleoproteins indicate that there are no gene-sized uncomplexed DNA regions in any preparations studied.

IV. THE DISSOCIATION OF HISTONE FROM CALF THYMUS CROMATIN

Calf thymus nucleoprotein was treated with varying concentrations of NaCl. The identity of the histones associated and dissociated from the DNA at each salt concentration was determined by gel electrophoresis. It was found that there is no appreciable histone dissociation below 0.4 F NaCl. The lysine rich histones dissociate between 0.4 and 0.5 F NaCl. Their dissociation is accompanies by a marked increase in the solubility of the chromatin. The moderately lysine rich histones dissociate mainly between 0.8 and 1.1 F NaCl. There are two arginine rich histone components: the first dissociates between 0.8 F and 1.1 F NaCl, but the second class is the very last to be dissociated from the DNA (dissociation beginning at 1.0 F NaCl). By 2.0 F NaCl, essentially all the histones are dissociated.

The properties of the extracted nucleoprotein were studied. The electrophoretic mobility increases and the melting temperature decreases as more histones are dissociated from the DNA. A comparison with the dissociation of histones from DNA in NaClO₄ shows that to dissociate the same class of histones, the concentration of NaCl required is twice that of NaClO₄.