Substitutions of sulfonatocalix[4]arenes that lead to applications in biomolecular recognition and give rise to novel self-association phenomena

(1)

recognition and give rise to novel self-association phenomena

by

Graham Garnett

B.Sc., University of Victoria, 2011

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE

in the Department of Chemistry

 Graham Garnett, 2014 University of Victoria

(2)

Supervisory Committee

Substitutions of sulfonatocalix[4]arenes that lead to applications in biomolecular recognition and give rise to novel self-association phenomena

by

Graham Garnett

B.Sc., University of Victoria, 2011

Supervisory Committee

Dr. Fraser Hof, Department of Chemistry Supervisor

Dr. David Berg, Department of Chemistry Departmental Member

Dr. Cornelia Bohne, Department of Chemistry Departmental Member

(3)

Abstract

Supervisory Committee

Dr. Fraser Hof, Department of Chemistry Supervisor

Dr. David Berg, Department of Chemistry Departmental Member

Dr. Cornelia Bohne, Department of Chemistry Departmental Member

The epigenetic post-translational modifications (PTMs) of histone proteins are numerous and have an important role in regulating cellular development. Molecular recognition elements that can bind directly to epigenetic PTMs have previously been developed. The most synthetically accessible of these are a family of molecules called monoaryl-trisulfonate calix[4]arenes. The initial goal of this thesis was to develop research tools consisting of these asymmetrically-derivatized calixarenes immobilized onto a solid resin for the purpose of enrichment of PTM-bearing species. Seven novel resin-immobilized calixarene reagents were created and employed in batch-wise pulldown experiments, as well as chromatographic separations. These experiments produced mixed results: poor efficacy was demonstrated in batch-binding experiments but total separation of certain PTM bearing peptides was achieved in a chromatographic approach. During these studies, a subset of these calixarenes were found to undergo self-association in water in a fashion not previously observed for calixarenes. Secondary goals of the thesis were to create new examples of this self-associating motif, and to

characterize and develop structure-function relationships for their assemblies. Eight new self-associating calixarenes were developed and characterized extensively by 1H NMR, isothermal titration calorimetry (ITC), and X-ray crystallography. Self-association was shown to be enthalpically driven with Kd values ranging from 1-20 mM. The dimeric

assembly behaviours were remarkably consistent across many different family members, and were shown to persist even in highly competitive media like mock blood and urine. This system represents a novel class of ordered calixarenes assemblies that operate in biological media.

(4)

List of Tables

Table 2.1 Solution phase affinities of PSC for KMe3 and K bearing peptides as determined by ITC titration. Little discrimination is observed between KMe3 bearing peptides of different sequence. Reprinted with permission from Daze et al.46_{H3K4 =} +_H

3N-ARTKQTAY-C(O)NH2, H3K9 = Ac-TARKSTGY-C(O)NH2, H3K27 =

Ac-AARKSAPY-C(O)NH2, H3K36 = Ac-GGCKKPHY-C(O)NH2. ... 39

Table 2.2 Enrichment provided by resin AG2.6 is plotted against simulatneously against pH and organic co-solvents. IpOH = isopropanol, DMSO = dimethylsulfoxide, MeOH = methanol, TFE = 2,2,2-trifluoroethanol. Standard conditions buffered with phosphate (pH 6.0, 9.0, 11.8) or acetate (pH 3.7). ... 57 Table 2.3 Total Fraction Free data for the same experiments as Table 2.2 above. IpOH = isopropanol, DMSO = dimethylsulfoxide, MeOH = methanol, TFE =

2,2,2-trifluoroethanol. Standard conditions buffered with phosphate (pH 6.0, 9.0, 11.8) or acetate (pH 3.7). ... 57 Table 2.4 A more fine grained investigation of AG2.6 enrichment at different pH and TFE concentration. Standard conditions buffered with 10 mM phosphate at all pH. ... 57 Table 2.5 Enrichment values of resin AG2.6 with respect to changes in pH and

Hofmeister salt concentration. Standard conditions buffered with 5 mM phosphate (pH 4.5, 7.0, 9.0, 11.0) or CAPS buffer (pH 10.6)... 58 Table 2.6 Enrichment of resin AG2.6 with respect to changes in pH as well as phosphate buffer and TFE concentration. Standard conditions buffered with 5 mM phosphate (pH 7.0, 9.0, 11.0) or CAPS buffer (pH 10.6)... 58 Table 2.7 Enrichment data from resin AG2.9, investigating some of the better conditions observed for AG2.6. Standard conditions buffered with phosphate (pH 6.0, 9.0, 11.8) or acetate (pH 3.7). ... 59 Table 2.8 Total fraction free data from the same experiment with AG2.9 as Table 2.7 above. Standard conditions buffered with phosphate (pH 6.0, 9.0, 11.8) or acetate (pH 3.7). ... 59 Table 2.9 Enrichment for resin AG2.9 at different pH and buffer concentrations. Standard conditions buffered with 5 mM phosphate at all pH. ... 60 Table 2.10 Total fraction free data for resin AG2.9 at different pH and buffer

concentrations. Standard conditions buffered with 5 mM phosphate at all pH. ... 60 Table 2.11 Enrichment of a side-by-side comparison of identical experiments performed using resin AG2.6 produced in two different batches. Standard conditions buffered with phosphate (pH 6.0, 9.0, 11.8) or acetate (pH 3.7). ... 63 Table 2.12 Total fraction free of a side-by-side comparison of identical experiments performed using resin AG2.6 produced in two different batches (C83 vs. C90). Standard conditions buffered with phosphate (pH 6.0, 9.0, 11.8) or acetate (pH 3.7). ... 64 Table 2.13 Control experiments looking at the effect of TFE and the presence of

unfunctionalized AffiGel-102 resin on enrichment. Standard conditions buffered with phosphate (pH 6.0, 9.0, 11.8) or acetate (pH 3.7). ... 64 Table 2.14 Control experiments looking at the effect of TFE and the presence of

unfunctionalized AffiGel-102 resin on total fraction free. For the “no resin” experiments, 25 μL of diH2O was added in place of resin. Standard conditions buffered with phosphate

(8)

Table 2.15 Control experiments looking at the effect of increased concentrations of unfunctionalized AffiGel-102 resin on enrichment and total fraction free. Standard conditions buffered with phosphate (pH 6.0, 9.0, 11.8) or acetate (pH 3.7). ... 65 Table 3.1 Calculated hydrodynamic radii from the DOSY determined diffusion constant as compared to the radii as determined by X-ray crystal structure data ... 108 Table 3.2 ITC data for hosts in 100 mM phosphate buffer, pH 7.4. Data is the average of triplicate titrations with the standard deviations shown on the right. ... 112 Table 3.3 ITC data for hosts in 100 mM PBS, pH 7.4. Data is the average of triplicate titrations with the standard deviations shown on the right. ... 116

(9)

List of Figures

Figure 1.1 The nucleosome. a) crystal structure showing the individual histone subunits colour coded as the cartoon representation in b). H3 = green, H4 = blue, H2A = red, H2B = magenta. DNA (orange) wraps the structure twice. Histone N-terminal tails are shown to extend out form the core structure. (PDB ID: 1AOI)26... 3 Figure 1.2 Epigenetic acetylation and methylation of histone lysine residues is catalyzed by their respective enzymatic “writer” proteins. K = lysine, HAT = histone

acetyltransferase, HDAC = histone deacetylase, KMT = lysine methyltransferase, KDM = lysine demethylase. ... 4 Figure 1.3 Crystal structure of the chromodomain module of CBX7 bound to a

peptidomimetic inhibitor bearing a KMe3 residue. The surface present model shows the KMe3 binding pocked (left). The surface removed model clearly shows the interactions between the KMe3 residue and the Trp-Trp-Phe aromatic cage (right). (PDB: 4MN3).38 6 Figure 1.4 The structure of the calix[4]arene. The base calix[4]arene is shown left. Center is para-sulfoantocalix[4]arene (PSC). Right is the larger para-sulfonatocalix[6]arene (PSC-6). ... 8 Figure 1.5 Some exemplary guests for PSC with Kass values as determined by 1H NMR or

ITC. (NH4+)66 (Li+, Na+, K+ Cs+)67 (Mg2+, Sm3+,La3+)68 (toluene, iodobenzene)69 (Phe =

phenyalanine)70 (pyridine, 4-methylpyridine)71 (methylviologen)72 (acetycholine)73 (tetramethylammonium)74 ... 10 Figure 1.6 Kass for the interaction of some the cationic amino acids lysine and arginine,

and their PTM methyl variants, with PSC as determined by 1H NMR dilution titration.44 R = arginine, MMA = monomethyl arginine, aDMA = asymmetric dimethyl arginine ... 11 Figure 1.7 The cation-pi interaction. Cartoon representation of showing the preferred geometry of the interaction. The gas-phase interaction of K+ with benzene (left) is

roughly energetically equivalent to the interaction of K+ with water (right). ... 13 Figure 1.8 Applications for sulfonated calixarenes in protein binding. a) PSC-6 (cartoon bowl) immobilized onto the surface of a plate binds both prion proteins (red) and normal proteins (blue). The normal proteins are digested by the protease (yellow) while the prion and PSC-6 are both protease resistant. Immunodetection occurs in a subsequent step not show in this cartoon. b) the binding interaction between two proteins (blue and grey) is disrupted through the interaction of a sulfonated calixarene and the surface of one of the proteins. PSC-(4-8) have all been examined for use in protein binding. ... 17 Figure 1.9 Dye displacement assay using PSC. The fluorescence of a fluorophore

(orange) is quenched when bound to PSC, but is displaced by a suitable guest (KMe3), restoring fluorescence. ... 18

(10)

Figure 2.1 Chromatin immunoprecipitation sequencing (ChIP-Seq) procedure. Full length chromosomes containing both DNA and noncovalently associated epigenetic proteins are isolated. These proteins are covalently crosslinked the local DNA (red stars) and the DNA is sheared into small fragments (green arrows). The resulting protein-DNA complexes are then enriched for an epigenetic motif of interest using

immunoprecipitation. The DNA pulled down along with the protein can then be isolated and sequenced (yellow lightning bolt) to give a “hit”, or correlation between that

epigenetic protein and that particular genetic locus. ... 24 Figure 2.2 Direct output from the UCSC ENCODE Genome Browser110 using data collected by the Broad Institute.112 H3K9Ac, H3K27Ac are overlaid over the gene CD86 (top), a surface protein used in antigen presentation and expressed only in certain blood cells. Epigenomic data from 3 cell types is reported, human embryonic stem cells (hESC, green), a leukemia (K562, blue) and a lymphoblast (GM12878, red). As can be inferred from the histone overlays, only the GM12878 displays histone PTMs that correlate with gene expression at this locus. ... 25 Figure 2.3 Cartoon of an indirect and direct, aka “sandwich” ELISA immunodetection experiments. In the indirect ELISA (left), the target (orange ball) is recognized by the anti-target primary antibody (purple), which is in turn recognized by the reporter secondary antibody (blue) which is covalently linked to a reporter enzyme, in this case alkaline phosphatase (black circle with AP) which hydrolyzes 4-nitrophenylphosphonate to yield the chromogenic 4-nitrophenol. In the sandwich ELISA (right), a primary

antibody is first immobilized to a plate to improve the selectivity in immobilization of the target. ... 27 Figure 2.4 Proteomics overview cartoon. Cells are first lysed, the proteins are then isolated and digested with a protease such as trypsin. The sample can then be prepared prior to analysis through immunoprecipitation or some other type of affinity purification. The resulting purified sample is then subjected to separation by RP-HPLC and analysis by MS/MS. ... 30 Figure 2.5 Cartoon summary of affinity chromatography chemistries. Each of the resin bound chemistries are used alternately for PTM specific retention of analyte. a) a lectin (blue) immobilized onto a resin binds and retards the transit of glycoproteins (grey). b) chemistry of the covalent capture of a citrulline residue which reacts to form the resin immobilized phenyl glyoxal. The protein can then be released from the resin through hydrolysis of a base labile linker. c) the interaction of a phosphopeptides (black) with a Ni2+ IMAC resin retards the transit of phosphopeptides. ... 37 Figure 2.6 Polymer chemistry of resin matrix a) agarose polymer is the polysaccharide of the repeat unit D-galactose and 3,6-anhydro-L-galactopyranose aka agarobiose. b)

TentaGel is a graft copolymer of polyethylene glycol chains bonded to a polystyrene matrix. ... 47

(11)

Figure 2.7 Coupling of 2.6 to AffiGel-102. Overlay of traces before and 1 hour after the addition of the coupling agent EDC. Depletion of calixarene from supernatant is observed while the levels of the caffeine standard remain constant ... 51 Figure 2.8 Structure of calixarene immobilized to resins. AffiGel-102 agarose based resin was functionalized with calixarenes 2.5, 2.6, 2.9 to create resins AG2.5, AG2.6 and

AG2.9, respectively. TentaGel S NH2 resin was functionalized with calixarenes 2.5, 2.6, 2.9 to create resins TG2.5, TG2.6 and TG2.9. Profinity resin was functionalized with

calixarene 2.7 to create resin PR2.7. ... 52 Figure 2.9 Typical LCMS results from a batch-based enrichment screen using AG2.6 (a, b and c) compared to the results of a similar experiment using AG2.9 (d). a) pre-resin levels of peptide. b) 18 h after addition of AG2.6. c) blown-up overlay of the

AARKSAPY K and KMe3 variant peptides in a) and b). The black trace shows pre-resin peptide levels and the blue trace shows peptide levels 18 h after addition of AG2.6. d) blown-up overlay of a similar experiment using resin AG2.9. The black trace shows pre-resin peptide levels and the blue trace shows peptide levels 18 h after addition of AG2.9. Note the high-affinity, low-selectivity binding exhibited by AG2.9 in d) compared to the lower-affinity, higher-selectivity binding of AG2.6 in c). ... 55 Figure 2.10 Column summary. A photograph of one of the prototype columns (left). a) the resin used to create column cAG2.6. b) the resin used to create column cAG2.9 ... 67 Figure 2.11 Results from the commercial SPXL ion exchange column using a 50 mM phosphate running buffer (pH 7.5) and 1 M NaCl + RB elution buffer under the standard gradient described. This column shows retention but not separation of K vs KMe3 variant peptides. a) 20 μL injection of AARKSAPY K peptide with no column present. b) 40 μL injections AARKSAPY K and KMe3 variants, chromatograms overlaid. c) 40 μL

injections of ARTKQTARKSTGY K, K4Me3, K9Me3 variants, chromatograms overlaid. d) 20 μL injection of solution of ARTKQTARKSTGY K and K9Me3 variants, both 1.0 mM. Baseline was subtracted from a blank injection. ... 68 Figure 2.12 Results from cAG2.9 using a 50 mM phosphate running buffer (pH 7.5) and 2 M NH4Cl + RB elution buffer under the standard gradient described. Under these

conditions total separation of AARKSAPY K and KMe3 variants and partial separation of ARTKQTARKSTGY K, K4Me3 and K9Me3 variants is observed. a) overlay of 20 μL and 40 μL injections of a solution of AARKSAPY K. b) overlay of 20 μL and 40 μL injections of a solution of AARKSAPY KMe3. c) overlay of 40 μL injections from of AARKSAPY K and KMe3 variants. d) 40 μL injection of solution consisting of

AARKSAPY K and KMe3 variants, both 1 mM. e) overlay of separate 20 μL injections of ARTKQTARKSTGY K, K4Me3 and K9Me3 variants. f) 20 μL injection of a solution consisting of ARTKQTARKSTGY K and K9Me3 variants, both 1.0 mM. Baseline was subtracted from a blank injection. ... 72 Figure 2.13 The effect of increased amounts of NH4Cl in the running buffer on cAG2.9.

Experiments were using a 50 mM phosphate running buffer (pH 7.5) and eluting with a 2 M NH4Cl + RB elution buffer under the standard gradient described. A third buffer

(12)

system consisting of 150 mM NH4Cl + RB was applied to the running buffer in different

amounts to examine the effects to add NH4Cl to the running buffer. a) 20 μL injections of

ARTKQTARKSTGY K with increasing amounts of NH4Cl in the running buffer. b) 20

μL injections of ARTKQTARKSTGY K9Me3 with increasing amounts of NH4Cl in the

running buffer. c) overlay of separate 20 μL injections of solutions of

ARTKQTARKSTGY K and ARTKQTARKSTGY K9Me3 with 7.5 mM of NH4Cl in the

running buffer shows the K9Me3 variant is less susceptible to the effects of the NH4Cl in

the running buffer. d) 20 μL injections of a solution of ARKSAPY KMe3 with increasing amounts of NH4Cl in the running buffer shows the shorter KMe3 variant is eluted from

the resin more easily than the longer ARTKQTARKSTGY K9Me3. Baseline was

subtracted from a blank injection. ... 73 Figure 2.14 Results from the separation of propionylated samples – synthetic

ARTKQTARKSTGY K and K9Me3 variants and histones digested by trypsin after treatment with propionic anhydride – using the cAG2.9 column with a 50 mM phosphate running buffer (pH 7.5) and eluting with 2 M NH4Cl+ RB. a) overlay of propionylated,

trypsin digested histones (~3 mg/mL). The blown up window shows a small portion of the digest is retained. Similar traces were observed for the non-propionylated trypsin digest. b) overlay of 40 μL injections of propionic anhydride-treated

ARTKQTARKSTGY K and K9Me3. For clarity, these chromatograms have not been baseline corrected; the green trace is from a blank injection of dH2O and shows the UV

absorbance resulting from the mobile phase gradient alone. ... 77 Figure 2.15 Unpublished data printed with permission from R. van Nulen and O. Gozani. Peptide pulldown using a biotinylated calixarene in combination with fluorescently labelled H3 1-20aa K, K4Me3, K9Me3, and K4/9/14Ac peptide variants. Calixarene and peptides were first incubated in solution before being pulled down onto streptavidin beads, separated by gel electrophoresis and fluorescently imaged. The darkness of the band is directly proportional to the amount of fluorescent peptide pulled down by the calixarene. a) this image shows the selectivity for KMe3 vs K bearing peptides exhibited by the calixarene under increasingly higher concentrations of NaCl. b) in this image, the calixarene is simultaneously presented with 4 variant peptides in a single pulldown. c) structure of the biotinylated calixarene used in these experiments... 78 Figure 3.1 Self-associating systems that form discrete homodimers in organic solvents. a) ureidopyrimidones by Meijer.153_{b) glycouril clips by Rebek.}154_{c) urea functionalized}

calixarenes by Rebek.52 ... 83 Figure 3.2 Calixarenes that self-associate to form aggregate structures in water. a) PSC derivatives with long-chain alkyl substituents on the lower rim that form micelles and other aggregate structures.164 b) N,N-dimethyl-N-hydroxyethylammoniumethylene functionalized calix[4]arene self-associates to form colloidal aggregates.165 ... 86 Figure 3.3 Examples of charge driven calixarene heterodimers. a) the heterodimeric system involving oppositely charged sulfonato- (3.6) and tetramidinium- (3.7) functionalized calixarenes.175 b) the heterodimeric system involving PSC and a

(13)

Figure 3.4 This resorcinarene-based dimer system87 associates due to hydrophobic

attraction between the aromatic flaps lining the upper rim. Omitted for clarity is the guest required to template this 1:1 dimer that is bound in the interior of the assembled capsule. ... 88 Figure 3.5 1H NMR spectra of calixarene 3.10 in D2O and CD3OD shows the solvent

dependence on self-association. a) 1_{H NMR spectra of calixarene 3.10 in D}

2O (25.4 mM)

shows upfield shifting of several resonances. b) 1H NMR spectra of the same calixarene

3.10, this time in CD3OD (21.3 mM). The upfield shifting observed in D2O, is completely

absent from the CD3OD spectra. ... 96

Figure 3.6 The proposed dimer structure. a) the equilibrium between monomeric and dimeric state of host 3.10. b) a simple SPARTAN model of the assumed structure of the dimer state of calixarene 3.11; the calixarenes associate through complementary

inclusion of the binding partner’s pendant arm. The monomer structures are energy minimized, but the dimer model is not. ... 97 Figure 3.7 1H NMR spectra host 3.10 and the para-alkyl series of hosts 3.11-3.14 show assembly-induced upfield shifts D2O. a) 3.10 (25.4 mM). b) 3.11 (20.3 mM). c) 3.12

(26.0 mM). d) 3.13 (24.8 mM). e) 3.14 (24.0 mM). ... 99 Figure 3.8 1H NMR of hosts 2.7, 3.15 and 3.16 in D2O. All hosts show upfield shifting of

proton resonances on the pendant arm that are consistent with the dimer model. a) 2.7 (20.4 mM). b) 3.15 (22.3 mM). c) 3.16 (20.2 mM). ... 100 Figure 3.9 1H NMR spectrum of host 3.11 shows the ability to assemble in different competitive solvents. a) CD3OD (22.1 mM). b) D2O (20.3 mM). c) 100 mM phosphate

buffered D2O (18.4 mM). d) 100 mM PBS buffer D2O (20.0 mM). e) artificial urine in

D2O (20.0 mM). ... 101

Figure 3.10 NOE spectra of host 3.14 give evidence for assembly in D2O and not in

CD3OD. a) 1D 1H NMR and corresponding NOE spectrum of a solution of host 3.14 in

D2O shows clear NOE correlations between all aromatic protons and the t-butyl

functionality of the tip of the pendant arm. Orange arrows represent correlations that might be the result of intramolecular NOE, whereas green arrow correlations can only be explained by intermolecular self-association with the t-butyl group in contact with the upper rim of another monomer. b) 1D 1H NMR and corresponding NOE spectrum of host 3.14 in CD3OD shows intramolecular NOE contacts to Ha and Hc but not the upper

rim protons He, Hf and Hg. No correlations were observed to the methylene protons in either spectrum. X-axes are coloured to highlight that they are scaled differently. ... 102 Figure 3.11 Dilution titration of host 3.11 in 100 mM phosphate buffered D2O, pD = 7.0.

The bottom spectra is host 3.11 in CD3OD... 104

Figure 3.12 Dilution titration of 3.14 in D2O, expanded over the t-butyl resonance. In this

(14)

Figure 3.13 DOSY spectra of compound 3.15. The X-axis is the 1HNMR spectra while Y-axis is the diffusion constant determined for each resonance. a) a high concentration of

3.15 (27.2 mM) gives a diffusion constant of 1.7 x 10-10_m2_{/s. b) the lower concentration}

(0.4 mM) gives a more rapid diffusion constant of 2.3 x 10-10 m2/s. ... 107 Figure 3.14 X-ray diffraction crystal structure of compound 3.14. a) The monomer asymmetric unit with associated Na+ _{(purple) and H}

2O (red). b) the same monomer

structure with Na+ and H2O hidden. c) the dimeric interaction adjacent monomers in the

crystal, with Na+ (purple) and H2O (red). d) the dimeric interaction adjacent monomers in

the crystal, space filling model with Na+ and H2O hidden. e) the rosette structure of six

symmetry-related monomers, with each of the 6 calixarene units colour coded differently. f) hydrogen bonding network between the two sulfonates that approach each other the closest in the solid state dimer, bridged through Na+ (purple) and H2O (red) ... 109

Figure 3.15 ITC dilution titrations provide thermodynamics of dimerization. a) raw endotherms for phenyl host 3.10 b) curve fitting for the change in molar enthalpy per amount of monomeric equivalent of 3.10 injected. This fit is good with a relatively low χ² (6.8) and visibly acceptable fit to the data. c) raw endotherms for the biphenyl host 3.16. d) curve fitting for 3.16 is not as accurate as 3.10 and the χ² value is higher (54.7). This trace represents one of the worst fits observed for these experiments. ... 114

(15)

List of Schemes

Scheme 2.1 Synthesis of key precursors: asymmetric monobromo and mononitro

calixarenes... 49 Scheme 2.2 Synthesis of biaryl-functionalized calixarenes from 2.2 ... 49 Scheme 2.3 Synthesis of sulfonamide-derived calixarenes from host 2.3 ... 49 Scheme 2.4 EDC driven coupling of a carboxylic acid bearing calixarene to an amine functionalized resin. ... 50 Scheme 3.1 Synthetic scheme continued from Chapter 2 for the preparation of dimerizing calixarenes... 90

(16)

Abbreviations

aa amino acid

Ac acetyl

ACN acetonitrile

aDMA asymmetric dimethylarginine

ADP adenosine diphosphate

AG AffiGel

AP alkaline phosphatase

Arg arginine

AUC area-under-the-curve

BSA bovine serum albumin

BzCl benzoyl chloride

CB7 cucurbit[7]uril

CBX chromobox homolog

CBX7 chromobox homolog 7

CD86 cluster of differentiation 86

ChIP-Seq chromatin immunoprecipitation sequencing

CnBR cyanogen bromide

CID collision-induced decay

Cit citrulline

diH2O deionized H2O

DLS dynamic light scattering

DMF dimethylformamide

DMSO dimethyl sulfoxide

DNA deoxyribonucleic acid

DOSY Diffusion Ordered NMR SpectroscopY

EDC 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide ELISA enzyme-linked immunosorbent assay

FPLC fast protein liquid chromatography

GST glutathione S-transferase

GTP guanosine triphosphate

h hour

HAT histone acetyltransferase

HDAC histone deacetylase

H2A histone 2A

H2B histone 2B

H3 histone 3

H3K27 lysine 27 on the histone 3 tail

H3K27Me3 trimethylated lysine 27 on the histone 3 tail H3K36 lysine 36 on the histone 3 tail

(17)

H3k36Me3 trimethylated lysine 36 on the histone 3 tail H3K4 lysine 4 on the histone 3 tail

H3K4Ac acetylated lysine 4 on the histone 3 tail H3K4Me3 trimethylated lysine 4 on the histone 3 tail H3K9 lysine 9 on the histone 3 tail

H3k9Ac acetylated lysine 9 on the histone 3 tail

H4 histone 4

His histidine

HPLC high-performance liquid chromatography

HR-ESI-MS high-resolution electrospray ionization mass spectrometry

IgG immunoglobulin gamma

IpOH isopropanol

IP immunoprecipitation

IR infrared

ITC isothermal titration calorimetry

K lysine KAc acetyllysine KDC lysine demethylase KLH keyhole-limpet hemocyanin KMe1 monomethyllysine KMe2 dimethyllysine KMe3 trimethyllysine KMT lysinemethyltransferase LC liquid chromatography Me methyl MeOH methanol min minute MMA monomethylarginine mp melting point MS mass spectrometry

MS/MS tandem mass spectrometry

μW microwave

NAD nicotinamide adenine dinucleotide

NMR nuclear magnetic resonance

N-terminal amino terminal

PDB Protein Databank

Phe phenylalanine

ppm parts per million

PSC para-sulfonatocalix[4]arene

(18)

PRMT1 protein arginine methyltransferase 1 QC quality control RP-HPLC reverse phase HPLC R arginine s second SAM s-adenosylmethionine

SDS sodium dodecyl sulfate

SEM scanning electron microscopy

TG TentaGel

TFA trifluoroacetic acid

TFE 2,2,2-trifluoroethanol

THF tetrahydrofuran

Trp tryptophan

UV ultra violet

VT-1_HNMR _{variable-temperature proton nuclear magnetic resonance}

(19)

Acknowledgments

I would like to the entire Department of Chemistry at UVic, for being my home away from home on and off for the last 8 years. I have had an amazing experience. I would especially like to thank Fraser for being an awesome supervisor and someone to really look up to. I would also like to thank all the 20+ Hof group members who have cycled through during my time, especially Kevin Daze, Amanda Whiting, Tom Pinter and Cindy Wang, all of whom helped me get my bearings in the lab as an undergraduate researcher.

I would also like to thank everyone who contributed, in one way or another, to the research presented in this thesis: Kevin Daze, Manuel Ma, Jorge Peña and Noah Fagen. If I have seen farther, it was only by standing on Kevin’s lanky shoulders. Thank you to Al Boraston and the Boraston lab for letting me use their ITC, especially Craig Robb and Melissa Cid. I would also like to thank the technical staff at UVic. Thank you Ori Granot and Andrew Macdonald for keeping the mass spec and HPLCs running and leak-free. Also a big thank you to Chris Barr for all the help using the NMR and for buying an autosampling cassette for the 500MHz NMR; thereby liberating me from the drudgery of performing low-concentration dilution titrations manually.

Thank you to everyone who helped support me as I was writing this thesis. Thank you to Rebecca Courtemanche for your help editing this thesis. I would also like to thank my girlfriend, Oriana, for always being in my corner and making sure I was well fed. I couldn’t have done it without your support.

Finally a special thank you to my parents, Susan and Peter, for inspiring me and teaching me about what is really valuable in life. Thank you for getting me interested in science at a young age, while never trying to plan my future for me.

(20)

Dedication

(21)

Chapter 1: Synthetic macrocycles with the potential for

discriminating epigenetic modifications

1.1 Epigenetics

1.1.1 What is epigenetics?

Epigenetics refers to heritable changes in gene expression that are not the result of modification to the underlying genome itself.1 These changes play a pivotal role in

determining the fate of cells and are heavily involved in the transition from pluripotent stem cells to differentiated somatic cells.2 To illustrate this point, an example: a human being contains many different cell types: blood cells, brain cells, muscle cells, etc. all of which arise from a single common genome. How is such diversity created from a single genome? The answer is that in any of these cell types, only a fraction of the genes present in the genome are expressed.3_{These genes can be grouped into two classes: the first class}

are basic genes of which expression is observed in all cell types, such as those for critical cellular functions such as metabolizing sugars or producing proteins, and the second class are genes for more specialized function in specific cell types only, such as producing antibodies in B-lymphocytes or rhodopsin in retinal cells. These differentiated cell types all arise from common stem cell progenitor cells through epigenetic alterations to gene expression.2, 4, 5 Unlike conventional genetic change (DNA mutation), epigenetic change refers to heritable changes in the expression patterns of the genome that are not the result of direct modifications to the underlying DNA code itself.1 Epigenetics is nonetheless a powerful operator of cellular change, being able to convert one cell type to another in a single generation. Even more dramatic is the phenomenon of phenotypic plasticity exhibited by certain animals, such as observed in the respective transitions of grasshoppers and caterpillars to locusts and butterflies; organisms which bear little outward similarity to their original form yet share identical genomes.6,7 Furthermore, epigenetics has direct implications for human health, especially cancer.8,9,10

Pivotal to the role of epigenetics in controlling gene expression are the proteins that directly act on DNA to either increase or decrease the expression of certain genes. Proteins deposit and recognize static epigenetic marks or signatures. These chemical

(22)

modifications to chromatin can dictate the local chromatin structure directly11 and can recruit other proteins to exert further downstream effects.12 Therefore the epigenome can be displayed as a set of data overlaid on the genome that represents the frequency with which a given epigenetic mark occurs at that genetic locus. The ability to profile the epigenome over the genome is a powerful tool as it allows for recognition of the

underlying epigenomic differences between cell types that lead to their differentiation.5 1.1.2 Epigenetic control can be exerted through chemical modification of histone

proteins

The direct methylation of guanosine nucleobases of DNA is the most well-known epigenetic modification. However, the epigenetic marks that occur on DNA-associated proteins are equally – if not more – important in controlling gene regulation13,14,15_{and in}

some cases these marks have shown an upstream regulatory role in DNA methylation.16 I restrict my discussion here to epigenetic modifications of proteins, as they are the most relevant to this thesis.

Modifications to proteins that are installed after translation (synthesis) of the protein are referred to as post-translational modifications (PTMs). Epigenetic protein modifications are created by upstream “writer” enzymes that catalyze their deposition.17

Once laid, these PTMs can exert their effects either directly by affecting the protein’s function or by recruiting other epigenetic “reader” proteins that recognize epigenetic PTMs to further perpetuate downstream signalling. Eventually the expression of the target gene is either silenced or encouraged through alteration of the local chromatin structure and the recruitment of proteins that either restrict or promote transcription. These PTMs involve the direct modification of the side chains of certain amino acids via methylation, phosphorylation, acetylation, sumoylation, malonylation, and more, leading to PTM-derivative amino acid residues that serve as the epigenetic marks dictating gene expression.18,19,20,21,22, 23_{These modifications are context sensitive to specific amino}

acid residues; lysine methylation may result in a different downstream outcome

depending on which lysine is methylated.24_{Multiple modifications to the same protein or}

interacting proteins can lead to combinatorial effects,25_{adding complexity to an already}

(23)

Figure 1.1The nucleosome. a) crystal structure showing the individual histone subunits colour coded as the cartoon representation in b). H3 = green, H4 = blue, H2A = red, H2B = magenta. DNA (orange) wraps the structure twice. Histone N-terminal tails are shown to extend out from the core structure. (PDB ID: 1AOI)26

One set of proteins most frequently implicated in epigenetics are the core protein components of the nucleosome called histones.23 The nucleosome is the base repeat unit of the genome consisting of the DNA double stranded helix wrapped nearly twice around a globular protein assembly of histone proteins (Figure 1.1).27 These proteins bind and wrap DNA around themselves so as to compact the genome in 3D space while at the same time positioning themselves to regulate gene expression. Nucleosomes are an octamer formed of two copies each of four distinct histone proteins and span the entire length of the genome, repeating every 180 BP or so. Widespread deposition of certain epigenetic PTMs on histones have been implicated in altering the chromatin morphology between transcriptionally active “euchromatin” and transcriptionally repressed

“heterochromatin” states.5,28, 29_{The heterochromatin state is tightly coiled so as to}

restrict the expression of the genes contained within and is a method of direct structural control of gene expression exerted by histone PTMs.11 However as mentioned the modification of histone proteins are multiple, complex and context sensitive. Often epigenetic PTMs are implicated in altering gene expression in areas of transcriptionally

(24)

accessible chromatin. In these cases, the epigenetic PTMs typically exert their effects through the recruitment of reader proteins, which then affect gene expression.23Histones H3 and H4 are especially important in epigenetic regulation and are heavily decorated with PTMs.Modifications occur most commonly on the N-terminal tails of histones, which are unstructured and extend out from the protein core by about 40 amino acid residues.30 Of especial interest in the library of histone modifications are the PTMs of lysine residues.

1.1.3 Histone lysine modifications are important epigenetic regulatory PTMs

Figure 1.2Epigenetic acetylation and methylation of histone lysine residues is catalyzed by their respective enzymatic “writer” proteins. K = lysine, HAT = histone acetyltransferase, HDAC = histone deacetylase, KMT = lysine methyltransferase, KDM = lysine demethylase.

Histone lysine residues on the N-terminal tails of H3 and H4 are frequently acetylated or methylated. These modifications are involved in the control of gene expression and their misregulation has been implicated in certain cancers.31,32 Histone lysine methylation is catalyzed by writer proteins called histone lysine methyltransferases (KMTs). Using the methyl donating cofactor S-adenosylmethionine, these enzymes can methylate the side chain amine of a lysine residue either once, twice, or three times, producing one of three distinct PTMs (Figure 1.2). Each resulting modification can be interpreted differently by the cell.25 In the shorthand of epigenetics, this turns a lysine or K residue into a KMe1, KMe2, or KMe3.Acetylation of lysine is catalyzed by histone acetyltransferases (HATs), enzymes which utilize the acetyl-CoA cofactor to deposit a single acetyl group onto the side chain amine. As with methylation, acetylation can regulate gene expression either through direct effects on the local chromatin structure or by being recognized by reader proteins. Hyperacetylation reduces the cationic nature of the histone proteins and loosens the histone’s grip on the negatively charged DNA strand.

(25)

Therefore these hyperacetylated areas are found in regions of transcriptionally active chromatin. Hypermethylation on the other hand can lead to tightening of the DNA-histone interaction and is often found in areas of transcriptionally repressed

heterochromatin.11 Histone lysine methylation and acetylation are important PTMs and also happen to be some of the most difficult for molecular recognition systems to discriminate. Antibodies, the premier molecular recognition unit of nature, are not well suited to target these motifs alone as they are too small and the differences too slight between modified and unmodified residue. A detailed discussion of the limitations of antibodies with regards to KMe3 discrimination will be provided in Chapter 2.

The proteins writing and reading PTMs are of interest as potential therapeutic targets for chemotherapy for certain cancers.33,34 Since the epigenome must be laid down onto the new genome of the daughter cell each time the cell divides, the inhibition of writer enzymes associated with cancer aggression seems like a viable therapeutic target. There are currently two small molecule inhibitors35, 36 of histone deacetylases (HDAC) available for the treatment of cutaneous T-cell lymphoma and there are several more inhibitors of epigenetic proteins in clinical trials for the treatment of other cancers.37 However, it is not feasible to generate small molecule inhibitors that are able to recognize the presence of KMe3 vs K on a dynamic protein tail, because these marks do not occur within the kinds of structured binding pockets needed for small molecule binding. For ideas on how to bind these marks directly, we looked first at the biological systems that operate on them. How do the epigenetic reader proteins involved in these epigenetic pathways recognize their respective PTM marks?

(26)

1.1.4 Biological recognition of KMe3 through specially evolved binding modules

Figure 1.3Crystal structure of the chromodomain module of CBX7 bound to a

peptidomimetic inhibitor bearing a KMe3 residue. The surface present model shows the KMe3 binding pocked (left). The surface removed model clearly shows the interactions between the KMe3 residue and the Trp-Trp-Phe aromatic cage (right). (PDB: 4MN3).38

The trimethylammonium group of the KMe3 PTM is not drastically different than the primary amine of an unmodified K residue. Both hold a positive charge under cellular conditions and are not that different in size when considered in the context of whole proteins. However, KMe3 marks can be recognized by certain reader proteins. One exemplary family of such proteins are those that include a Chromatin Organization Modification domain (chromodomain), as depicted in Figure 1.3.38, 39,40 This domain contains several electron rich aromatic amino acid residues that are aligned in such a fashion as to line the surface of the KMe3 binding pocket with the faces of the aromatic residues, allowing it to benefit from hydrophobic and cation-pi interactions.

Chromodomains are able to discriminate ≥10 fold between KMe3 and K40,41 with Kass for

KMe3 bearing peptides in the range of ~100,000 M-1. This discrimination is largely accomplished through recognition of the more diffuse and hydrophobic nature of the cationic charge of KMe3 residue vs K, as well as important differences in solvation between the molecules. These molecular recognition features are in fact common to almost all known proteins that bind methyllysine selectively over unmethylated lysine.42

(27)

KMe3 binding domains have all evolved to specifically recognize the KMe3 residue within the context of a specific surrounding sequence. For example, the

chromodomains of CBX proteins typically bind KMe3 at histone 3, lysine 9 (H3K9) or histone 3, lysine 27 (H3K27), but do not bind methylated lysine at H3K4 or H4K20.40, 41 There are many applications in which a sequence-independent binding of a certain PTM would be highly desirable. We wondered if we could use synthetic chemistry to design a receptor for KMe3 residues displaying similar discrimination and affinities to their biological counterparts, but without strong sequence selectivity. As we shall see, this was a task well-suited to one of the most well-known macrocycles in supramolecular

chemistry, the calixarene.

1.2 Introduction to calixarenes and host-guest encapsulation

The field of supramolecular chemistry is focused on the interactions between molecules.43 Supramolecular chemistry is then an understandably broad term, the

concepts of which are central to the fields of molecular recognition, materials science, as well as many (if not all) biological processes. Host-guest chemistry involves the

supramolecular complexation of one binding partner by another binding partner. Host-guest chemistry is distinguished from other supramolecular interactions in that one host is usually a macrocycle with a concave binding pocket that is well suited to form

interactions with a smaller, simpler guest molecule. One macrocycle,

para-sulfonatocalix[4]arene (PSC), is particularly well suited for the binding of the KMe3 PTM.44,45,46 Below I will explore the structure and binding preferences of PSC. I will then investigate some of the structural and functional similarities between PSC and the binding pockets of methyllysine reader proteins.

1.2.1 Calixarene, background and structure

From the Latin “calyx” for “cup” or “chalice,” the cup-like macrocycle calixarene was first reported in 1944 by German polymer chemists.47 One of the most well-known macrocyclic hosts (along with cyclodextrins), calixarenes have enjoyed a long and diverse history in supramolecular chemistry. This is partly due to their relative ease in synthesis and derivatization as compared to other macrocycles. Calix[n]arenes are characterized structurally by “n” phenol rings joined through methylene bridges at the

(28)

meta position. The macrocycles are stable and most commonly formed in sizes ranging

from 4 – 8 phenol rings. This meta bridging restricts free rotation of the phenol rings and leads to the availability of several structural conformations, depending on ring size.48 As the size of the calixarene increases, so does the conformational flexibility of the

molecule, making the calix[4]arene the most conformationally rigid. Four low-energy conformations are available for the calix[4]arene in solution and of these, the cup-shaped “cone” conformation—which is the iconic origin of the calixarene’s name—is most desirable from host-guest binding standpoint.48_{The other conformations (partial cone,}

1,2-alternate and 1,3-alternate) involve the inversion of one or more rings of the

calixarene and are not well-suited for guest inclusion. The two opening of the calixarene are referred to as the upper rim (wider) and the lower rim (narrower) (Figure 1.4).

Figure 1.4The structure of the calix[4]arene. The base calix[4]arene is shown left. Center is para-sulfoantocalix[4]arene (PSC). Right is the larger para-sulfonatocalix[6]arene (PSC-6). The lower rim hydroxyls of calix[4]arenes are in close proximity and interact with one another, forming a cyclic hydrogen bonding ring around the lower rim that stabilizes the cone conformation.48Modification of the lower rim phenols can have marked effects on conformation and binding properties48,49,50,51 and is synthetically limited to

alkylations and acylations of the hydroxyls. Conversely, modification at the upper rim does not strongly disrupt the preference for the cone conformation and larger functional groups are tolerated while still allowing for guest binding.50Electrophilic aromatic substitution para to the phenol is a common method of derivatization and there many examples of calix[4]arenes with different chemistry on the upper rim.50,52,53

1.2.2 Sulfonation provides a water soluble calixarene

For a macrocycle to have relevance in biomolecular recognition applications it must be able to operate in water, the medium of biology. The addition of strongly polar

(29)

water-solubilizing groups onto the upper rim of calixarenes were first reported in 1984.54,

55_{Of the several water soluble calixarenes that have been created,}

para-sulfonatocalix[4]arene (PSC) is the most widely reported and has enjoyed a storied history as one of the premier water soluble macrocycles.56,57

Characterized by symmetrical sulfonation at the four para positions (Figure 1.4), PSC is readily soluble in water up to mM concentrations. The lower rim is typically left unmodified in PSC; both to improve solubility and to benefit from the intramolecular hydrogen bond network around the lower rim phenols that stabilizes the cone

conformation. In water, this hydrogen bond network increases the acidity of the first phenol deprotonation to near pKa = 3, which along with the four upper rim sulfonates

(pKa < 0) brings the formal charge to -5 in water at neutral pH.58

1.2.3 Guest preferences of PSC

In the cone formation, para-sulfonatocalix[4]arene has a somewhat flexible internal cavity well suited for the inclusion of small molecular guests 3-4 Å wide. PSC will bind a large variety of hydrophobic and cationic guest molecules in aqueous solution. Much research has been devoted to studying the inclusion preferences of PSC for

different biomolecules such as amino acids,59_peptides60_{and proteins}61,62_{as well as small}

molecules such as steroids63,64_{and neurotransmitters like acetylcholine}65_{in water. These}

studies typically involve some combination of 1H NMR titration, isothermal titration calorimetry (ITC) and/or fluorescence to measure the strength of the binding interaction. As we will see, the aromatic rings lining the cavity directly influence guest preference through hydrophobic and pi-interactions. A complete review is beyond the scope of this thesis, but some illustrative examples are discussed here.

(30)

Figure 1.5Some exemplary guests for PSC with Kass values as determined by 1H NMR or

ITC. (NH4+)66 (Li+, Na+, K+ Cs+)67 (Mg2+, Sm3+,La3+)68 (toluene, iodobenzene)69 (Phe =

phenyalanine)70_{(pyridine, 4-methylpyridine)}71_{(methylviologen)}72_{(acetylcholine)}73

(tetramethylammonium)74

A set of exemplary binding constants between PSC and different guests at neutral pH are shown in Figure 1.5. In addition to biomolecules, PSC is known to form

complexes with various transition metals, and alkali metal cations with Kass in the range

of 3 to 170 000 M-1_{, with stronger binding observed for the more highly charged}

cations.66,67 When it comes to organic molecules, PSC is a fairly prolific hosts that will bind a large array of simple hydrophobes with low to moderate affinity. Examples

include toluene, iodobenzene and biphenyl, with binding constants ranging from 20 to 60 M–1. 69_{Similar affinities for planar hydrophobic biomolecules such as steroids have been}

documented, with Kass in the range of 10-100 M-1.63Other heteroaromatics with cationic

nature such as pyridine, 4-methylpyridineand methylviologen are more suitable guests for PSC, and associate with Kass of 302 M-1, 1350 M-1 and93 300 M-1 respectively.71 ,72

Of course the binding modes of all of these guests is not the same, some guests will prefer to interact with the upper rim sulfonate groups than the pi-cavity. For a thorough review of literature pertaining to host-guest inclusion of PSC in water see Guo.75

(31)

Figure 1.6. Kass for the interaction of some the cationic amino acids lysine and arginine, and

their PTM methyl variants, with PSC as determined by 1_{H NMR dilution titration.}44_{R =}

arginine, MMA = monomethyl arginine, aDMA = asymmetric dimethyl arginine

Investigation into the binding of various amines revealed the high affinity of PSC for quaternary ammonium cations. For instance, 1H NMR studies by Lehn76 put the binding by PSC of tetramethylammonium as Kass 79 000 M-1, compared to the Kass

ammonium which had previously been determined as 6.9 M-1_.66_{These high affinities for}

tetramethylammonium sparked interest in the use of calixarenes for binding to biological quaternary amines such as acetylcholine.65, 76_{Cationic amino acids lysine and arginine}

both bind PSC in water with low affinities (Figure 1.6).44_{Early work in the Hof group}

showed that PSC binds the free amino acid trimethyllysine much more strongly than other cationic amino acids, with a Kass = 37 000 M–1.44 Importantly, the unmethylated K

amino acid associates with a much weaker Kass = 520 M-1 (Figure 1.6). The affinities for

KMe3 become slightly stronger when the trimethyllysine is in the context of a peptide, for example Kass for the short peptide Ac-R(KMe3)ST-NH2 = 96 600 M–1, bringing the

affinities to values that are similar to those observed for binding of KMe3 residues by chromodomains as discussed above.

1.3 Forces driving guest binding and discrimination by PSC

As in most complexation events, guest inclusion by PSC results from the contribution of multiple weak interactions. Understanding these interactions is key to understanding the affinity and selectivity of PSC for one guest over another. To gain a more comprehensive understanding of the nature of the interaction, it is helpful to take a reductionist approach towards the forces governing it.

(32)

1.3.1 Electrostatic and van Der Waals effects

Electrostatics often play an important role in binding events that involve charged species and the host-guest interactions of PSC and KMe3 bearing peptides are no

exception. Complementary charge between molecules can be a strong driving force for association, even in water where ions are strongly solvated. As PSC has a formal charge of 5− at neutral pH, interactions with cations are favoured, while anion binding is less represented. Besides the upper rim sulfonate, the lower rim of PSC also possess a 1− charge shared amongst the four phenolic hydroxyls which no doubt provides some driving force for the inclusion of cationic guests into the cavity. Histone N-terminal tails contain many cationic residues and KMe3 itself possesses a positive charge on the sidechain. As such the attraction between calixarene and histone-tail peptides based on electrostatics alone is significant. In general, KMe3 and K, having the same charge at neutral pH, should behave similarly from an electrostatic standpoint.

Van der Waals interactions between guest and cavity also play a role in the association. These include dipole, induced dipole or dispersion forces experienced between the molecules. These interactions should be strongest for guest molecules large enough to form tight interactions with the walls of the cavity. However the difference in strength of these interactions among slightly different guest molecules is relatively small compared to the other interactions involved. The interaction geometries of unmethylated lysine and trimethyllysine are distinct, and each maximize van der Waals contact surface area between the host and guest in different ways.61_{In general, the difference in energy}

between the van der Waals interactions of a primary or quaternary ammonium ion binding in the cavity of PSC would be unlikely to contribute to the large binding preference observed for KMe3 over K.

1.3.2 Hydrogen bonding

The IUPAC definition of a hydrogen bond is an attractive interaction between an electronegative atom and a hydrogen atom attached to a second, relatively electronegative atom.77_{Although they involve an element of electrostatic attraction between the partially}

positive H atom (on the H-bond donor) and a partially negative heteroatom (acceptor), hydrogen bonds differ from purely electrostatic interactions in that there is a covalent component shared between H bond donor and acceptor atoms through the proton.77

(33)

Although they are able to donate hydrogen bonds, the lower rim phenols are rarely observed to interact with substrate. Salt bridges are a type of hydrogen bond formed between two oppositely charged species and have an additional strong electrostatic component.78 In cocrystal structures of proteins and PSC, salt bridges and hydrogen bonds have been observed between the upper rim sulfonates of the calixarene and the side chain amines and backbone amide N-H’s of peripheral amino acid residues.61

1.3.3 Cation-pi

Figure 1.7 The cation-pi interaction. Cartoon representation of showing the preferred geometry of the interaction. The gas-phase interaction of K+ _{with benzene (left) is roughly}

energetically equivalent to the interaction of K+_{with water (right).}

Perhaps one of the more important factors when it comes to binding quaternary ammonium ions, both by synthetic systems such as calixarenes and biological systems, is the cation-pi effect. First discovered through computational simulations of benzene hydration in 1981,79 the cation-pi effect has since been recognized as being important to a number of biological phenomena. These include protein folding80 and the biological recognition of quaternary ammonium ions81 such the recognition of KMe3 residues by the chromodomain as discussed earlier.40, 82_{Although charge transfer effects can be}

involved, the cation-pi effect can be effectively modelled in purely electrostatic terms.83 The electrostatic potential over the plane of the pi-system is partially negative and can interact favourably with cations. Early gas phase experiments and simulations determined the attraction between a simple benzene ring and K+_{cation to be 19}_{kcal/mol (Figure}

1.7), which is comparable to the ion-dipole interaction energy between an K+ cation and a water molecule (18 kcal/mol).79_{Attractive force has been shown to correlate with cation}

charge density, with harder cations expected to make stronger interactions.57, 79, 84 While this holds true in the gas phase, solvent effects must be taken into consideration to understand the relative strengths of cation-pi interactions in solution. This is especially

(34)

true in water as the desolvation cost associated with harder ions is often greater than softer ions.78,85 The strength of the interaction also correlates well with the electronic character of the pi-system, and a linear decrease in binding has been observed for mono-di-tri-and tetra-fluorinated aromatics that are made less electron rich (and therefore

weaker participants in cation-pi interactions) with each subsequent fluorination.81_{As with}

electrostatic attraction in general, the strength of the cation-pi interaction diminishes with increasing solvent polarity.78

1.3.4 The hydrophobic effect

For full thermodynamic characterization of a binding event one must not only take into account energetic contributions from the interactions between the two binding solutes, but also the interactions of those solutes with solvent in both the free and bound states. This is especially true for systems that associate in water, as the interactions between water-water and water-solute are often very strong and can outweigh the

energetic contributions of the solute-solute interactions.86 Consequently the strengths and trends of binding interactions in water are more difficult to predict compared to those in organic solvents. The differential solvation of free and bound states in water can be a potent driving force for association.87_{This is because bulk water, due to its optimal}

hydrogen bonding and polar contacts with other water molecules, is in such an

energetically low state that the solvation of organic solutes leads to unfavourable change in the energy of the system. As a result, hydrophobes will tend to associate in such a fashion that minimizes the amount of wetted hydrophobic surface area, as water on the hydrophobic surface is energetically frustrated as “high enthalpy water”.78, 86_When

hydrophobic surfaces come together, water is expelled to the bulk phase, a result that can be both entropically and enthalpically favoured.88,89 This water specific driving force for the association of hydrophobes is termed the “the hydrophobic effect”. This effect tends to be much stronger than the van der Waals attraction experienced between the assembled molecules. The same intermolecular interactions giving rise to the hydrophobic effect are also responsible for the other interesting properties of water, such as high heat capacity and surface tension.78

The effect was first brought to the attention of modern physical chemistry through studies of oil-water partitioning.90 In these studies, separation of oil from water was

(35)

shown to be entropically favoured due to the increased disorder of released water molecules from interacting surface areas. This entropic driving force was initially considered a hallmark of the hydrophobic effect.90 Although it is common to observe favourable entropies in the association of hydrophobic systems in aqueous solvent, other associations also presumably driven in large part through the hydrophobic effect have been characterized by favourable enthalpies and unfavourable entropies. The enthalpy-driven association of hydrophobes can be explained by water molecules forming more hydrogen bonds in the bulk phase than when solvating a hydrophobic surface area. This has been termed the “non-classical hydrophobic effect”,91_{relegating the entropy-driven}

variant the title of “classical hydrophobic effect.” Ultimately, it is difficult to rationalize the enthalpic and entropic contributions to association in any given system that is driven by the hydrophobic effect. Other researchers have suggested that other measures, such as a negative change in heat capacity upon binding, as being better correlated to the

hydrophobic effect.91,92

It is important to realize that the interaction between hydrophobic molecules and water is not repulsive. In fact, computations have shown that water-alkane interactions are actually more favourable than alkane-alkane interactions, but are just not as strong as water-water interactions.93 Similarly, water inside the cavity of PSC – though frustrated94 relative to water in the bulk state – is making some energetically favourable contacts with the cavity. These include not only dipole-induced dipole interactions and van der Waals interactions, but also as was demonstrated for the first time in a crystal structure of PSC, OH-pi interactions.95 Therefore, even though these interactions may be relatively weak there is still an energetic cost of desolvation that must be paid before guest inclusion can occur.

1.3.5 Structural differences between K and KMe3

All of these factors contribute in one way or another to the binding of analytes like KMe3 by PSC. I have discussed briefly the structural characteristics of PSC and how the cavity is similar to the biological chromodomains used to recognize KMe3. What are the structural differences between KMe3 and K that allow for each of these binding pockets to discriminate between the two?

(36)

Size and shape complementarity is important to ensure that the aromatic pocket is able to make optimal interactions with the guest molecule. A guest too small to

adequately fill the cavity will not be able to make optimal interactions (of all of the above kinds) to the entire cavity and will be too small to fully occupy the space left by water. A guest too large will not be able to fit into the cavity without forcing energetically

unfavourable steric overlap with elements of the macrocycle. X-ray structures and molecular models show that trimethylammonium functionalities – such as is present on the KMe3 side chain – is perfectly sized for these two respective pockets, while

unmethylated lysine is too small to make contacts with all walls of the pockets at once.42,

61, 74_{Additionally, trimethylation alters the electrostatic and hydrophobic character of the}

lysine side chain and leads to large differences in solvation which can be exploited for the discrimination of KMe3 vs K.42,82 Although both KMe3 and K side chains are cationic at neutral pH, the charge on the trimethylammonium of KMe3 is spread more diffusely over the methyl groups as opposed to the protons of the primary ammonium K side chain. Most importantly, the trimethylammonium sidechain of KMe3 is no longer able to form hydrogen bonds with water. As a result of both of these factors the KMe3 sidechain is much more hydrophobic than the K (or KMe1 and KMe2) species. The result of this differential solvation of guests is a much stronger association of the calixarene cavity towards KMe3 than to K, driven largely by the hydrophobic effect. A consequence of this solvation can be seen in the crystal structure of PSC bound to surface lysine residues on a model protein (cytochrome C).61 The unmethylated lysine residues on the protein surface often bind in an orientation with their hydrophobic beta, gamma, delta, epsilon methylene groups folded up and sitting inside of the PSC cavity. The -NH3+ group of lysine sits

above the cavity, preferring to retain partial hydration and form salt bridges to the upper rim sulfonates in over becoming completely desolvated, entering the calixarene pocket, and forming cation-pi interactions with the aromatic walls. The authors suggest the appearance of multiple distinct states in the crystal structure indicate a relatively shallow energetic potential map for K binding to the calixarene in different orientations (but none of which have cation-pi contacts as a major player).61 Conversely, KMe3, which is not able to hydrogen bond to solvent or to the upper rim sulfonates, is always found bound

(37)

deeply within aromatic pockets by X-ray structures (in the case of KMe3 reader proteins40, 96 and NMR evidence (in the case of PSC binding).44

1.4 Previous applications of sulfonated calixarenes as biochemical tools

Targeted biological applications for sulfonated calixarenes typically involve the binding of amines or other cationic species, either to disrupt interactions or to act as sensors for those species. Although antibodies and other protein modules can often be used as affinity reagents in these experiments, one of the overarching themes behind the use of macrocycles such as calixarenes is that synthetic compounds offer potential benefits over protein-based reagents. These advantages will be discussed in detail in Chapter 2, but for now will be summarized by stating that calixarenes are cheaper, easier to produce, more robust, and have less batch-to-batch variability than protein alternatives.

Figure 1.8 Applications for sulfonated calixarenes in protein binding. a) PSC-6 (cartoon bowl) immobilized onto the surface of a plate binds both prion proteins (red) and normal proteins (blue). The normal proteins are digested by the protease (yellow) while the prion and PSC-6 are both protease resistant. Immunodetection occurs in a subsequent step not show in this cartoon. b) the binding interaction between two proteins (blue and grey) is disrupted through the interaction of a sulfonated calixarene and the surface of one of the proteins. PSC-(4-8) have all been examined for use in protein binding.

There have been a few recent advances in potential applications for sulfonated calixarenes as molecular recognition and protein surface binding elements. A larger analog of PSC, p-sulfonatocalix[6]arene (PSC-6), has been used to aid in the detection of

Substitutions of sulfonatocalix[4]arenes that lead to applications in biomolecular recognition and give rise to novel self-association phenomena

Supervisory Committee

Abstract

Table of Contents

List of Tables

List of Figures

List of Schemes

Abbreviations

Acknowledgments

Dedication

Chapter 1: Synthetic macrocycles with the potential for

discriminating epigenetic modifications