Determinants of Cyanuric Acid and Melamine Assembly in Water

ARTICLE pubs.acs.org/Langmuir

Determinants of Cyanuric Acid and Melamine Assembly in Water Mingming Ma and Dennis Bong* Department of Chemistry, The Ohio State University, Columbus, Ohio 43210, United States

bS Supporting Information ABSTRACT: While the recognition of cyanuric acid (CA) by melamine (M) and their derivatives has been known to occur in both water and organic solvents for some time, analysis of CA/M assembly in water has not been reported (Ranganathan, A.; Pedireddi, V. R.; Rao, C. N. R. J. Am. Chem. Soc. 1999, 121, 1752 1753; Mathias, J. P.; Simanek, E. E.; Seto, C. T.; Whitesides, G. M. Macromol. Symp. 1994, 77, 157 166; Zerkowski, J. A.; MacDonald, J. C.; Seto, C. T.; Wierda, D. A.; Whitesides, G. M. J. Am. Chem. Soc. 1994, 116, 2382 2391; Mathias, J. P.; Seto, C. T.; Whitesides, G. M. Polym. Prepr. 1993, 34, 92 93; Seto, C. T.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 905 916; Zerkowski, J. A.; Seto, C. T.; Whitesides, G. M. J. Am. Chem. Soc. 1992, 114, 5473 5475; Seto, C. T.; Whitesides, G. M. J. Am. Chem. Soc. 1990, 112, 6409 6411; Wang, Y.; Wei, B.; Wang, Q. J. Chem. Cryst. 1990, 20, 79 84; ten Cate, M. G. J.; Huskens, J.; Crego-Calama, M.; Reinhoudt, D. N. Chem.—Eur. J. 2004, 10, 3632 3639). We have examined assembly of CA/M, as well as assembly of soluble trivalent CA and M derivatives (TCA/TM), in aqueous solvent, using a combination of solution phase NMR, isothermal titration and differential scanning calorimetry (ITC/DSC), cryo-transmission electron microscopy (cryo-TEM), and synthetic chemistry. While the parent heterocycles coprecipitate in water, the trivalent system displays more controlled and cooperative assembly that occurs at lower concentrations than the parent and yields a stable nanoparticle suspension. The assembly of both parent and trivalent systems is rigorously 1:1 and proceeds as an exothermic, proton-transfer coupled process in neutral pH water. Though CA and M are considered canonical hydrogen-bonding motifs in organic solvents, we find that their assembly in water is driven in large part by enthalpically favorable surface-area burial, similar to what is observed with nucleic acid recognition. There are currently few synthetic systems capable of robust molecular recognition in water that do not rely on native recognition motifs, possibly due to an incomplete understanding of recognition processes in water. This study establishes a detailed conceptual framework for considering CA/M heterocycle recognition in water which enables the future design of molecular recognition systems that function in water.

’ INTRODUCTION Designed molecular recognition that functions in organic solvents rarely transitions successfully into aqueous milieu. Polar/polar interactions, such as hydrogen bonding, that typically direct recognition in low dielectric solvents are less effective at mediating recognition in high dielectric solvents that feature competitive hydrogen bonding groups, such as water. Designed assembly in water remains a challenge for chemists that is well worth investigating, given the ubiquitous importance of molecular recognition in biology. Thus, aqueous-phase assembly is a solved problem in Nature, but there are few examples of designed recognition motifs in water.10 14 An elegant example from Fenniri and co-workers features synthetic heterocycles that combine the base-pairing properties of guanine and cytosine in single module that assembles into “helical rosette” topologies in water.15,16 Interestingly, they also found that similar diamino triazine and barbiturate derivatives did not assemble in water;17 these compounds are analogues of melamine (M) and cyanuric acid (CA) in which a nitrogen has been replaced by carbon.16 It was not clear why the former system assembled while the latter system did not, r 2011 American Chemical Society

underscoring the need for more detailed understanding of assembly processes in water. We present herein a careful examination of aqueous-phase assembly of the canonical recognition pair, cyanuric acid and melamine, which has largely been ignored. To date, there are no reports that probe the origins of aqueous-phase selective heteromeric assembly of cyanuric acid and melamine. Chemists have long been interested in the assembly properties of cyanuric acid and melamine derivatives,1 9,18 though recognition properties have been mostly limited to organic solvents and the solid state. Hydrogen bond driven assembly of CA/M derivatives demonstrates how substitution can drive the formation of “rosette” and tape structures in organic solvents (Figure 1),3,5,6,19 22 a motif that has been employed in exquisitely designed molecular receptors.1,9,18,23 Though the parent compounds CA and M rapidly coprecipitate in water, monoderivatized CA/M heterocycles do not engage in specific molecular recognition in water, Received: April 18, 2011 Revised: May 27, 2011 Published: June 21, 2011 8841

dx.doi.org/10.1021/la201415d | Langmuir 2011, 27, 8841–8853

Langmuir consistent with the reports of Fenniri, Lehn, and co-workers.16,17 However, studies at the amphiphile water interface by Kunitake and Meijer and their co-workers24 30 revealed that melamine or barbituric/cyanuric acid recognition can direct noncovalent polymerization when the hydrogen bonding groups are buried. There are few studies that provide clear understanding of the determinants of aqueous-phase recognition of synthetic hydrogen bonding groups13,31 and whether these designed interactions are distinct from base-stacking.32 We recently reported that “trivalent” scaffolds (Figure 2) displaying three cyanuric acid or melamine heterocycles (TCA and TM, respectively) bind strongly to one another when anchored at the lipid water interface;33 35 we report herein that the freely soluble TCA/TM headgroups selectively recognize each other even in the absence of a lipid water interface. We were intrigued by the robust aqueous-phase assembly properties of this non-native system, especially in light of its minimal scaffold, the absence of a hydrophobic tail, and the inability of other water-soluble CA/M derivatives to assemble.16 We conducted a detailed investigation of the physical underpinnings of cyanuric acid/melamine recognition in water in the context of both the parent heterocycles and our trivalent derivatives. To the best of our knowledge, this is the first such study of this canonical molecular recognition system in water. The aqueous phase assembly of cyanuric acid/melamine (CA/M) systems has not been previously examined in part because these

ARTICLE

systems do not assemble to discrete oligomers in aqueous solvent. Precipitates are formed from the parent system while large aggregates are formed from amphiphilic systems; our trivalent derivatives assemble to form a uniform nanoparticle suspension. Display of CA/M recognition groups on a synthetic scaffold not only retains the recognition properties of the parent heterocycles, but also limits assembly to nanoparticle formation, which facilitates closer inspection of the process. Assembly enthalpy provides a useful probe of sensitivity to heat, pH, and scaffold structure and synthetic scaffold modification. Despite the obvious similarities between CA/M recognition and nucleobase-pairing, key differences emerge that are a direct result of CA/M molecular symmetry. Whereas nucleobase pairing is generally observed at a single interface at a time, the higher symmetry of CA/M and derivatives permits simultaneous molecular recognition on symmetry-related hydrogen-bonding faces of the molecule (Figure 1). This symmetry equates recognition with noncovalent polymerization and, further, places cyanuric acid pKa (∼6.5) near physiological pH (unlike native nucleobases). Thus, assembly at pH ∼ 7.0 requires base neutralization. Though these molecular characteristics put CA/M heterocycles at a disadvantage relative to native nucleobases in mediating discrete molecular recognition at physiological pH, the retention of selective and robust recognition properties when displayed on simple synthetic scaffolds indicates possible applications of derivatives in aqueous-phase adhesive materials and tools for biotechnology.

’ EXPERIMENTAL SECTION General Sample Preparation and Concentration Determination. All aqueous solvents or buffers were filtered through a 0.22 μm

Figure 1. Assembly of monoderivatized cyanuric acid and melamine derivatives into a “rosette” pattern, as observed in organic solvents and solid state.

PVDF syringe filter. Stock solutions (10 mM) of cyanuric acid (CA) and melamine (M) were prepared gravimetrically in H2O or D2O with or without PBS buffer (50 mM sodium phosphate, 150 mM NaCl) and pH/pD was adjusted with 1 M HCl/H2O, DCl/D2O, 1 M NaOH/H2O or NaOD/D2O as appropriate. Stock solutions were diluted as necessary. HPLC-purified and lyophilized (Supporting Information) trivalent CA and M derivatives (TCA and TM) were dissolved in H2O or D2O to make a stock solution of 5 10 mg/mL and pH/pD adjusted. Concentration was determined by lyophilization of a known volume of stock (usually 100 μL) in an NMR tube followed by dissolution in 0.5 mL of 1 mM isopropanol in D2O and 1H NMR integration of sample peaks relative to isopropanol.

Figure 2. Structures of the trivalent scaffold CA/M system based on Tris. Tris-based recognition pairs were solubilized with negatively charged carboxylate (TCA/TM) and sulfonate (TCA-SO3/TM-SO3) groups. 8842

dx.doi.org/10.1021/la201415d |Langmuir 2011, 27, 8841–8853

Langmuir NMR of TCA/TM. NMR measurements were performed on the 1:1 TCA/TM complex (5 mM total concentration) in PBS, pH 6.8 in H2O with a D2O insert tube for locking signal. Spectra were acquired on a Bruker Avance DPX 400 ultrashield instrument, using a water-suppression sequence. Samples were heated from 24 to 80 °C with equilibration at each temperature stage prior to acquisition. Isopropanol (0.5 mM) was used as an internal standard for titration experiments (10 mM TCA was titrated into 1 mM TM) at 24 °C. In the 1H NMR titration experiments, the central CH2 unit of the Tris scaffold (O CH2 CH2 CH2 NH ) was integrated. This methylene appears at 1.66 and 1.72 ppm in TM and TCA in PBS (H2O) at pH 6.8. Measurement of M, CA, and TCA pH, pKa, and Fractional Ionization. Cyanuric acid and melamine pKa’s in H2O have been

reported to be 6.5 and 5.1, respectively.36,37 We redetermined these pKa’s in salt solutions in both protic and D2O buffer (5 mM sodium phosphate, 150 mM NaCl). The cyanurate anion concentration was followed by its unique absorption peak at 213 nm, while protonation of melamine was indicated by the disappearance of absorbance at 236 nm. Curves were fitted to obtain absorbance at infinite pH values and normalized to obtain fractional ionization. All measurements were performed at 28 °C. Experiments with unbuffered samples (0.6 mM CA, M, TCA, TM, TCA/TM, or CA/M, 100 mM NaCl) were performed at pH 6.8 or pD 7.2. Samples of mixture TCA/TM or CA/M were prepared at 1:1 ratio, 0.3 mM each. The samples were stirred while heated and cooled in a water bath with an Accumet semimicro pH electrode inserted into the solution. Measurements were made after two heating and cooling cycles to degas the solution, resulting in highly reproducible pH temperature curves. The pH and temperature were measured using a Beckman 720 pH meter with auto pH temperature compensation. All pD measurements were made by correcting the pH reading by addition of 0.4 units.38,39 Calorimetry. Stock solutions of TCA/TM derivatives or CA/M were diluted as necessary, and all samples were filtered through 0.22 μm PVDF syringe filter and degassed before loading. Isothermal titration calorimetry (ITC) was performed on a MicroCal VP-ITC microcalorimeter. Typically, for TCA/TM derivatives, TM was diluted to 0.05 mM and placed in the cell while TCA (0.5 mM) was injected into the cell. Titration experiments with the parent CA/M systems were performed at 10-fold higher concentrations: M was placed in the cell at 0.5 mM, while a 5 mM CA solution was injected. Reference ITC runs were performed by omission of either titrant or analyte; this heat was subtracted from the final ITC data. Differential scanning calorimetry (DSC) was performed on a MicroCal VP-DSC microcalorimeter. For the TCA/TM mixture or CA/M mixture, the diluted single compound solution was filtered and then two solutions were mixed. Typically, for TCA/TM derivatives, a total concentration of 0.6 mM was used, with 0.3 mM of each compound for mixed samples. For CA/M DSC, a total concentration of 2 mM (1 mM each component) was used though no defined features were observed in the DSC. A scan rate of 60 °C/h was used to scan between 90 and 10 °C. Reference DSC measurements were performed on water and buffer; these data were subtracted from the final trace. Dynamic Light Scattering (DLS). Dynamic light scattering (DLS) was performed on a Malvern Instruments Zeta Sizer Nano-ZS apparatus. Samples were filtered through a 0.22 μm PVDF syringe filter. For the TCA/TM mixture or CA/M mixtures, the diluted single component solution was filtered and then two filtered solutions were mixed. Typically, for TCA/TM derivatives, total concentration of 0.2 mM was used (0.1 mM each compound) while CA/M measurements were made at 2 mM (1 mM each compound). Typically, the DLS pH experiments were started at a pH value above pH 9, where there was no aggregation. pH was adjusted down with 1 M filtered HCl, and samples were measured at 25 °C. Samples were sonicated for 20 s and equilibrated for 2 min at each temperature or pH.

ARTICLE

Cryo-Transmission Electron Microscopy. Filtered samples were prepared at 0.6 mM concentration. Mixtures of TCA/TM were sonicated for 1 min at room temperature to give a slightly milky solution, and about 20 μL of this solution was applied to standard Formvar carbon support film EM grids from Ted Pella. The specimen was blotted by filter paper from behind and then plunged into an ethane slurry maintained at liquid nitrogen temperature. The frozen hydrated specimens were examined in Tecnai G2 Spirit (FEI company) electron microscope operating at 120 keV. Images were recorded using a Gatan CCD camera and pixel size was determined by imaging a standard sample grid (gold shadowed latex spheres with average size of 204 nm from Ted Pella) at each magnification used.

’ RESULTS AND DISCUSSION Initial Findings on Aqueous-Phase Recognition of TCA/ TM. Assembly studies began with the previously reported TCA/

TM carboxylate used to functionalize lipid and peptide membrane recognition elements (Figure 2). Design was driven primarily by considerations of synthetic simplicity rather than a defined hypothesis of the requirements for aqueous phase recognition. We did draw upon elegant prior efforts to design and synthesize hydrogen bonding pairs that function in water; attention was focused on rigid scaffolds to minimize entropy loss on binding, but generally systems were not as efficient as those in organic solvents.9,13,40 In water, surface area burial upon recognition is of particular importance. Interactions between heteroaromatic rings (including nucleobase pairing) in water41,42 are not well-described by the “classical” hydrophobic effect,43,44 which is defined by poor solvation in water; many heteroaromatic rings that interact favorably in water, including cyanuric acid and melamine, are also very water-soluble. Interactions between heteroaromatic systems in water are thought to be driven by a number of factors in addition to the classical hydrophobic effect, such as dispersive interactions, dipole dipole interactions, and donor acceptor interactions, though the relative contributions of each of these effects remains unclear.41,45 47 A flexible scaffold is expected to be more tolerant of imperfect recognition register, thus maximizing binding interactions. With these considerations in mind, a flexible scaffold derived from tris-(hydroxymethyl)aminomethane (Tris) was functionalized with three hydrogen bonding heterocycles (CA or M) using thioether linkages at the end of each hydroxymethyl arm. We had previously found that these trivalent cyanuric acid (TCA) and melamine (TM) modules could mediate membrane membrane docking and fusion when appropriately installed on lipids and peptides. The amino group on Tris was formerly used to connect the trivalent module to lipid or peptide membrane anchors;33 here, we have simply acylated with succinic anhydride to install a negative charge on the scaffold to maintain water solubility. Immediately upon mixing the fully soluble TCA and TM carboxylates, a suspension formed. When examined by solution 1 H NMR, aggregate size prevented observation of signals from 1:1 mixtures; this aggregation was found to be thermally reversible when monitored by 1H NMR between 25 and 80 °C, with a melting transition around 60 °C (Figure 3). Interestingly, titration experiments revealed a rigorously maintained 1:1 stoichiometry of assembly by 1H NMR, even though the aggregate is of undetermined oligomerization state. TCA was titrated into a TM solution while monitoring 1H NMR peak integration, revealing a steady decrease in TM signal until a mole ratio of 1 was reached; following this equivalence point, TCA signals 8843


Langmuir

ARTICLE

Figure 3. (A) NMR peak integration of TCA and TM below (b) and above (O) 1:1 TCA/TM stoichiometry. (B) A 1:1 mixture of TCA/TM forms an aggregate invisible by 1H NMR in PBS 6.8 but may be melted into soluble components. (C) Raw titration calorimetry trace of TCA titrated into TM in PBS 6.8, 10 °C.

Figure 4. Structures of trivalent CA and M synthetic derivatives based on the (A) Tris and (B) Tren scaffolds, with sites of variation indicated (R1 4).

appeared and increased. This indicated that the TCA/TM aggregate consumed all TCA until a 1:1 ratio was reached, then additional TCA was excluded from the aggregate. The 1H NMR analysis was performed at millimolar concentrations of TCA/ TM; with 20-fold dilution (50 μM), aggregation was undetectable by DLS, though this does not preclude the existence of assembled structures. Recognition at this lower concentration regime was probed by ITC, which revealed strikingly robust exothermic mixing ( 31.5 kcal/mol) of TCA and TM at 10 °C. Though the aggregation state is unknown, the titration curve fit well to a 1:1 binding model. We were unfortunately unable to correlate the ITC binding curve at low concentration to direct structural information due to aggregation in the millimolar concentration regime required for NMR. We turned to synthetic chemistry to search for derivative structures that could assemble into defined oligomer states at millimolar concentrations compatible with NMR analysis. Synthetic Efforts to Optimize Trivalent CA/M Recognition Scaffolds. We prepared Tris-based TCA/TM derivatives with decreased number of hydrogen bonding sites, limited scaffold conformation, and increased charge, in an effort to inhibit noncovalent polymerization (Figure 4, Table 1). Scaffold charges, positive or negative, were essential for maintaining the solubility of the cyanuric acid and melamine derivatives. As melamine functionalization is much more facile than CA modification, additional alkylation of the exocyclic melamine amino groups was examined as a means of

controlling assembly by reducing the number of hydrogen bonding sites. Both methyl and aminoethyl substituents were prepared, with the aminoethyl group providing an additional cationic charge as well as blocking a recognition site (Table 1). Melamine aminoethylation did indeed inhibit aggregation as judged by DLS, but unfortunately also blocked all recognition of TCA as well. It was somewhat surprising that the relatively subtle perturbation of melamine monomethylation also abolished all molecular recognition of cyanuric acid derivatives, especially as partial melamine alkylation serves as a means to control oligomerization state in organic solvents. N-Methyl amide linkages were hypothesized to generate cis amide conformers that would favor a closed or “concave” conformation amenable to discrete oligomerization.48 However, amide methylation was insufficient to favor one particular conformation and instead resulted in a number of conformational states, as judged by a much more complicated NMR spectrum relative to that of the secondary amide scaffold; recognition properties were abolished. We also explored an alternative trivalent scaffold for CA/M presentation based on trisaminoethylamine (Tren) which permitted a longer linker arm length of 14 atoms instead of 11 (Figure 4B, Table 1). Tren derivatives bear a positive charge on the tertiary amine; this charge is not sufficient to completely solubilize the Tren-melamine derivative. However, the Tren-CA3 derivative is fully soluble and, despite differences in scaffold chemistry and structure, assembles with Tris-M3 (TM), yielding a similarly exothermic signature in ITC ( 29.5 kcal/mol) with 1:1 stoichiometry (Table 1). 8844


Langmuir

ARTICLE

Table 1. Summary of Derivatives and Recognition Propertiesa scaffold Tris Tris Tris

a

R1 OH NH(CH2)2SO3Na OH

R2

R3

R4

ring(s)

assembly?

recognition Pair

ΔH (kcal/mol)

H

H

CA/M

yes

TCA/TM

H

H

CA/M

yes

TCA-SO3/TM-SO3

31.5

Me

H

CA/M

no

n/a

nd

31.0

Tris

OH

H

Me

M

no

n/a

nd

Tris

OH

H

-Et-NH2

M

no

n/a

nd

Tren

H

M

yes

Tren-CA3/TM

Tren

CO2H

M

no

n/a

29.5 nd

Assembly and enthalpy determined by DLS and ITC at 10 °C in PBS, pH 6.8. nd = not determined.

Figure 5. (A) DLS counts of the TCA/TM assembly on cooling (b) and heating (O). (B) Particle size observed by DLS as a function of pH for TCA/ TM (O) and TCA-SO3/TM-SO3 (b). The gray region indicates data points with high polydispersity and unreliable size data. (C) DLS counts as a function of pH for TCA/TM (O), TCA-SO3/TM-SO3 (b), and CA/M (4) assemblies. Cryo-TEM of (D) TCA/TM and (E) TCA-SO3/TM-SO3 with 200 and 100 nm (inset) scale bars as indicated.

The generally unfavorable recognition performance of derivatives closely related to unfunctionalized Tris CA/M reveals the sensitivities of the CA/M recognition process in water. Tris-M3 (TM) and Tren-CA3 bind to one another, as judged by DLS and ITC, indicating that recognition may occur despite obvious dissimilarities in structure, linker length, and charge. Therefore, it appears that recognition only requires flexible multivalent presentation. If the linker is conformationally limited (such as by amide alkylation) or if the recognition interface is modified (such as by melamine alkylation), all binding is lost. Structurally, CA/M recognition in water is at once highly tolerant of the nature of the scaffold and highly discriminating about the interface. We hypothesized that the extensive surface area burial achieved by aggregation is the actual driver of assembly rather than hydrogen bonding, consistent with the finding that molecules designed to limit aggregation did not assemble at all. We investigated the basis for assembly of cyanuric acid and melamine

in water using primarily TCA/TM, the Tris-based recognition pair we discovered initially (Figure 2). Effect of Ionization on Assembly. Analysis of the TCA/TM aggregation process by DLS indicated both temperature and pH dependent assembly (Figure 5). The hydrogen bonded network requires that each heterocycle is neutral; thus, disassembly was expected at pH regimes beyond the pKa's of melamine (5.0) and cyanuric acid (6.5). The 1:1 suspension exhibited a sharp heatinduced disassembly transition and a cooling-induced assembly transition, with marked hysteresis. Interestingly, TCA/TM assembly revealed a more pH-controlled process than the parent CA/M system. The aggregate formed by assembly of the parent CA/M system in water may also be melted reversibly and also exhibits a pH dependence. The total light scattering of the parent CA/M system indicated that assembly begins at pH e 8.5, which presumably reflects protonation of the cyanurate anion. As cyanuric acid pKa is ∼6.5, it appears that assembly induces a considerable pKa shift. Striking changes in pKa upon host guest 8845


Langmuir binding in synthetic systems have been reported by Nau and coworkers using a curbituril host49 51 as well as in the designed metal ligand host system reported by Raymond et al.52,53 In CA/M assembly, each cyanurate anion must be protonated to be incorporated into the assembly, suggesting that the surface surface of the nascent nanoparticle must bear anionic charge from partially bound cyanurates that is neutralized as additional layers of melamine are added. Dropping the pH below the pKa of melamine did not induce disassembly as would be expected from the protonation of melamine, possibly due to kinetic stability of the aggregate. Consistent with this observation, it has been previously reported that diffraction quality crystals of the 1:1 complex1 can form from HCl solutions, indicating that assembly forces prevail kinetically. In contrast, TCA/TM assembly begins at pH 8, slightly closer to the cyanuric acid pKa and begins to disassemble at pH < 5, consistent with disassembly being triggered by TM protonation. The sensitivity of TCA/TM assembly to acidic pH conflates melamine protonation with neutralization of the carboxylate tail of TCA and TM. Indeed, while aggregation of CA/M results in burial of neutralized heterocycles, assembly of TCA with TM requires that the negatively charged carboxylate must somehow be incorporated in the assembly, or itself be neutralized. To probe the effect of scaffold ionization, we replaced the carboxylate tail with a sulfonate (Figure 2), which should remain ionized over a wider pH range than the carboxylate (CH3SO3H pKa = 2.6). If scaffold neutralization is required for assembly, assembly of the sulfonate system (TCA-SO3 /TM-SO3 , Table 1) should be inhibited relative to the carboxylate system (TCA/ TM) given the more significant energetic cost of sulfonate protonation relative to carboxylate protonation. To our surprise, assembly of TCA-SO3 /TM-SO3 at near neutral pH was very similar to TCA/TM assembly, as judged by DLS. Both sulfonate and carboxylate systems assemble into 100 200 nm diameter particles with low polydispersity between pH 5.5 7.5. The two systems diverge in behavior at pH 5.5, where carboxylate neutralization drives a rapid increase in particle heterogeneity; the TCA/TM sulfonate nanoparticle assembly remains stable until pH ∼ 4.2 where melamine protonation is expected (Figure 5B). It appears that the sulfonate tail remains ionized at pH 4.2, while melamine protonation (pKa = 5.0) results in an increase in polydispersity and decrease in scattering counts, consistent with disassembly upon loss of hydrogen bonding sites on melamine. Thus, DLS reveals key differences in CA/M driven assembly that result from changes in the pKa of the solubilizing scaffold charge; both sulfonate and carboxylate systems exhibit greater pH sensitivity than the parent CA/M system. It is apparent that the solubilizing charge plays a critical role in limiting aggregate size, as the parent CA/M system, which has no solubilizing charges, aggregates into a wide range of particle sizes, from micrometer-sized to visible aggregates. The charged carboxylate or sulfonate tails provide some degree of control over the aggregation process, and the similar assembly behavior at neutral pH suggests that the charges are mostly surface-exposed. Unfortunately, this was difficult to probe using standard methods such as zeta potential measurement, as the particles were not stable to the measurement conditions, which caused a yellowing of the solution and precipitation, possibly a result of electrochemical oxidation. Cryo-TEM analysis of TCA/TM carboxylate aggregates at pH 6.8 confirmed a particle population mostly in the 100 200 nm diameter range, with some larger aggregates as well. Similarly, cryo-TEM of the nanoparticle suspension formed

ARTICLE

by TCA-SO3/TM-SO3 indicated particles in the ∼100 nm range, with similar uniformity (Figure 5). In contrast to the stable particle suspensions formed by CA/M assembly on the Tris scaffold, assembly of unfunctionalized CA/M heterocycles leads to large visible precipitates that quickly sediment and are not readily analyzed by cryo-TEM. Assembly pH Dependence and Proton Transfer. Spontaneous CA/M assembly observed by DLS at pH 8.5 was remarkable, as this is fully 2 pK units higher than the reported pKa value of cyanuric acid of 6.5. This indicated that favorable assembly energetics can drive cyanurate protonation at a pH greater than cyanuric acid pKa. To confirm the requisite proton transfer from solvent to cyanurate anion upon assembly, unbuffered aqueous solutions of CA and M adjusted to pH 6.8 were mixed at room temperature and the pH monitored. Upon mixing of the recognition components of either the parent CA/M or TCA/ TM system at pH = 6.8, a suspension was formed immediately and the pH rose to >9. The CA/M assemblies were thermally cycled while pH was monitored. The pH returned to 6.8 upon melting (disassembly) and returned to 9.4 upon cooling (assembly) in a highly reversible manner (Figure 6). This behavior was not found in the separate CA, TCA, M, or TM unbuffered solutions. It appears that CA/M and TCA/TM assembly enables deprotonation of water by cyanurate; melting/disassembly then returns the proton to hydroxide (Figure 7). This particularly noteworthy, since water (pKa = 15.7) is a weaker acid than cyanuric acid (pKa = 6.5) by ∼9 orders of magnitude. Cyanuric acid generally is nondissociated in organic solvents, and thus, proton transfer-coupled assembly is exclusively an aqueous-phase phenomenon that we find in both the unfunctionalized parent CA/M system as well in the Trisscaffold TCA/TM system. Therefore, although there are significant differences in pH sensitivity and assembly morphology, the recognition processes of trivalent CA/M are similar mechanistically to parent CA/M assembly. Calorimetry and Proton Transfer in CA/M Assembly. Our examination of the CA/M aqueous phase assembly process thus far allowed us to construct a model (Figure 7) which features proton-transfer coupled assembly, unique to the aqueous environment. The protonation of cyanurate anion by water against the pKa gradient was intriguing to us, and we turned to ITC to obtain a more quantitative view of assembly. As it was not possible to confirm a specific aggregation state, we did not fit the ITC curves to extract equilibrium constants and free energy, but rather just determined the heat flow upon CA/M mixing. As previously mentioned, isothermal titration calorimetry of TCA/TM mixing revealed a robust exothermic signal, indicative of 1:1 TCA/TM binding stoichiometry. The mixing of CA with M is also exothermic, but requires millimolar concentration for similar heat flow, while the trivalent system exhibits a more defined and cooperative calorimetric trace at micromolar concentration, suggesting enhanced assembly from multivalent presentation (Figure 8). In ITC traces of both systems, there is a strong dependence on temperature and pH, confirming our observations with DLS. At pH 6.8 in PBS, diminishing heats of mixing of 31.5, 28.8, and 22.5 kcal/mol were observed at 10, 25, and 40 °C, respectively, overall more exothermic heats than observed with CA/M titrations (Figure 8, Table 2). Similarly, titration enthalpy became more exothermic as pH was changed from 8.0 ( 27 kcal/mol) to 6.0 ( 36 kcal/mol). The system was not investigated below pH 6 as protonation of the carboxylate would 8846


Langmuir

ARTICLE

Figure 6. Solution pH as a function of temperature for unbuffered aqueous solutions of (A) heating (9) and cooling (0) of cyanuric acid alone, heating (2) and cooling (4) of melamine alone, and heating (b) and cooling (O) of the CA/M complex; (B) heating (9) and cooling (0) of TCA alone, heating (2) and cooling (4) of TM alone, and heating (b) and cooling (O) of the TCA/TM complex. Solutions of CA/M and TCA/TM prepared by mixing unbuffered aqueous pH 6.8 solutions of the individual compounds at room temperature.

Figure 7. Schematic illustration of proton transfer coupled assembly of CA/M and TCA/TM in water, based on the assumption that assemblies adopt hydrogen bonding patterns similar to those observed in organic solvents and the solid state.

begin and the assembly components would precipitate, complicating analysis. The strong enthalpy dependence on pH again points toward the importance of ionization state; in the pH range 6 8, only cyanuric acid (CA or on TCA) is significantly changing in ionization. As the pKa of TCA is ∼6.5, it should be close to 50% deprotonated at pH 6.8. We reasoned that lowering pH would generate a higher concentration of neutral TCA, thus bypassing protonation and associated energetic penalties. In phosphate buffer, the proton donor is presumably NaH2PO4 (pKa2 = 6.82), which has an ionization enthalpy at 10 °C of +1.8 kcal, while in unbuffered water the proton donor is water (pKa = 15.7), which carries an ionization enthalpy penalty of +14.2 kcal/ mol.54,55 Indeed, when ITC was performed in unbuffered water at pH 6.8 instead of PBS, a dramatic decrease in the heat released upon mixing was observed by ITC, consistent with a larger endothermic penalty for proton transfer from water to cyanurate relative to proton transfer from NaH2PO4 to cyanurate. In PBS 6.8, assembly enthalpy was found to be ∼ 32 kcal/mol by both ITC and DSC, while in unbuffered water at pH 6.8 ITC indicated a heat of mixing of about 10 kcal/mol. A full accounting of enthalpic contributions is complicated by the large increase in pH during the course of the unbuffered reaction between TCA and TM (Figure 6). As pH increases, TCA ionization increases and enthalpy of mixing decreases, resulting in further suppression of

heat flow. Despite these uncertainties, it is apparent from the early part of the titration experiment that proton ionization enthalpies have a direct effect on the heat of TCA/TM assembly. Thus, protonation of CA serves as an assembly trigger in both the parent and trivalent systems. We confirmed the assembly calorimetric profile using DSC (Figure 9). Despite the presence of an assembly melting transition by DLS and exothermic ITC trace, DSC of the parent CA/M complex did not reveal a well-defined transition peak on heating or cooling scans. In contrast, the TCA/TM aggregate suspension exhibited a broad endothermic heating transition, and a very sharp exothermic peak was seen on cooling. The exothermic peak observed on cooling by DSC correlates well with the assembly transition observed by DLS (Figures 5 and 9). Integration of the DSC cooling transition peak at pH 6.8 revealed a transition enthalpy of 31 kcal/mol, similar to our observations by ITC ( 32 kcal/mol). Cycling resulted in loss of the endothermic melting features, but the exothermic peak on cooling was very reproducible. This heating cooling asymmetry in the DSC trace is not uncommon: proteins often exhibit an endothermic melting peak and no exothermic peak on cooling, while lipid assemblies exhibit both endothermic melting transitions and exothermic cooling transitions, coinciding with disassembly and reassembly, respectively. The presence of a defined peak on cooling but not heating DSC scans of the TCA/TM complex may be explained if one considers that aggregates of heterogeneous size are melting, which could lead to broadened transition heats, while on cooling soluble molecules are condensed into uniform nanoparticles at a defined temperature, leading to a coordinated release of heat. This is consistent with our observations by DLS that prolonged “aging” of TCA/TM suspensions at lowered temperatures leads to larger aggregates, while freshly prepared suspensions are relatively uniform. Solvent Isotope Effects in CA/TM Assembly. In the course of NMR experiments on the TCA/TM system, it was observed that the use of deuterated water as solvent would induce aggregation at lower concentrations than in protic water. We examined this phenomenon more precisely using DSC and ITC and found that the enthalpy of mixing was much more exothermic ( 38.6 kcal/mol) in D2O PBS (pD 6.8) than observed with 8847


Langmuir

ARTICLE

Figure 8. ITC of buffered solutions of TCA titrated into TM at (A) 10 (0), 25 (O), and 40 °C (4) and (B) at pHs of 6.0 (0), 6.8 (O), and 8.0 (4) measured at 10 °C. (C) CA titrated into M at 10 (0), 25 (O), and 40 °C (4) in PBS, pH 6.8. (D) ITC of TCA titrated into TM at 10 °C in unbuffered water at pH 6.8 (O) and in PBS pH 6.8 (b).

Table 2. Assembly Enthalpies of CA/M and TCA/TMa temperature (°C)

pH 6.0

pD

10

25

40

5.0

6.8

8.0

6.8

TCA/TM

31.5

28.8

22.5

nd

36.5

31.5 ( 30.8)

27.5

38.6 ( 40.3)

CA/M

25.5

8.7

0

20.6

27.0

25.4

16.9

28.7

a

Enthalpy reported in kcal/mol and obtained by ITC except for values in parentheses, obtained by DSC. Temperature variation was carried out at pH 6.8 in PBS, while pH/pD experiments were performed at 10 °C in PBS. nd = not determined.

protic PBS at pH 6.8 ( 31.5 kcal/mol) (Figure 9, Table 2). Analogous DSC experiments revealed assembly transition enthalpies at pH/pD = 6.8 ( 30.8 and 40.3 kcal/mol, respectively) that confirmed the more strongly exothermic heats in deuterated buffer observed by ITC. Furthermore, DSC indicated a higher assembly temperature (65 °C) relative to protic buffer (51 °C), underscoring facilitated assembly in deuterated solvents. This trend was also observed in ITC analysis of the parent CA/M system; titration experiments in deuterated

PBS (pD 6.8) revealed a more exothermic heat of mixing ( 28.7 kcal/mol) than in protic PBS (pH 6.8) ( 25.4 kcal/mol). In addition, the ITC trace in deuterated PBS was more well-defined and indicative of cooperative assembly than the calorimetric trace in protic PBS, supportive of enhanced recognition at pD 6.8. Origins and Implications of Solvent Isotope Effect on Assembly. There are two possible origins of this solvent isotope effect that are not mutually exclusive. The first is the known suppressive effect on D2O on acid dissociation.56 A solvent isotope 8848


Langmuir

ARTICLE

Figure 9. Enthalpy of assembly exhibits a significant solvent isotope effect. (A) ITC of TCA into TM at 10 °C, pH/pD = 6.8. (B) DSC of TCA/TM complexes at pH/pD = 6.8. (C) ITC PBS of CA into M at at 10 °C, pH/pD = 6.8.

derived pKa shift of +0.5 is expected based on empirical methods of estimation in D2O relative to H2O. For a given pH/pD, one expects greater acid dissociation of CA or TCA in water versus D2O. Indeed, our measurements of pKa of CA and TCA in H2O and D2O reflect a half unit pKa shift, which leads to a significant difference in ionization, with 48% of CA ionized in H2O and only 26% ionized in D2O at pH/pD 6.8 (Figure 10). The difference is even greater for TCA, with 48% TCA ionized in H2O and 19% ionized in D2O at pH/pD 6.8. Thus, the energetic penalty for protonation in D2O should be 40 54% of that exacted in H2O when pH = pD, corresponding to the decreased cyanurate anion concentration in D2O; this would certainly contribute to the more robust assembly enthalpy observed by ITC. An additional explanation for the solvent isotope effect on enthalpy derives from the known enhancement of the hydrophobic effect in D2O relative to H2O.57 The transfer of hydrophobic amino acids from H2O to D2O results in a negative enthalpy change;58 this solvent isotope effect also manifests in lower critical micelle concentrations, higher protein stability, higher melting DNA duplexes,59 and more compactly folded proteins in D2O.57,60 Our observations of more enthalpically favorable assembly in D2O than in H2O are consistent with these prior reports. These two effects of suppressed ionization

and increased hydrophobic effect could both operate to improve assembly in deuterated water relative to protic water. Consideration of the TCA ionization curves (Figure 9) and the pH dependent ITC traces provides some indication of the solvent isotope effect on ionization and assembly. TCA/TM mixing at pD = 6.8 is significantly more favorable ( 38.6 kcal/mol) than mixing at pH = 6.8 ( 31.5 kcal/mol) but is closer to the heat of mixing found at pH 6.0 ( 36.5 kcal/mol) (Figure 10D, Table 2). Inspection of the TCA ionization curve (Figure 10B) reveals that pD 6.8 has a fractional TCA ionization identical to that of pH 5.9 (pH 6.0 has an fractional ionization of 22% vs 19% at pH 5.9). Thus, a large portion (∼5 kcal/mol) of the solvent isotope derived enthalpy change (ΔΔHH2OfD2O ∼ 7 kcal/mol) may be attributed to changes in fractional TCA ionization, with the remaining ∼2 kcal stabilization in D2O possibly the result of solvent isotope effects on hydrophobic burial. The identical effect is observed in analysis of parent CA/M assembly. There is a striking difference in calorimetric profiles derived from CA/M ITC experiments run at pH 6.8 and pD 6.8. Though numerically the difference is only ∼1.5 kcal/mol favoring assembly in deuterated buffer, the shape of titration curve is much more cooperative and defined in deuterated PBS relative to protic PBS (Figure 9C). Indeed, when the ITC measurement is carried 8849


Langmuir

ARTICLE

Figure 10. Room temperature measurement of pH/pD-dependent ionization derived from UV measurement of (A) CA and (B) TCA in protic buffer (b) and D2O buffer (O). ITC in PBS at 10 °C of (C) CA into M and (D) TCA into TM at pD 6.8 (O) and at pH 6.0 (b).

out at pH 6.0 where the fractional CA ionization (15%) is similar to that in pD 6.8 (26%), the shape of the ITC curve exhibits a similar cooperative transition and heat as found at pD 6.8 (Figure 10C, Table 2). Thus, the ionization state of the CA or TCA molecule is a critical determinant in the CA/M recognition process in water. That changes in ionization state may account for most of the enthalpy advantage observed in D2O relative to H2O is telling. When corrected for differences in ionization state, the solvent isotope effect favoring CA/M assembly in D2O is much smaller (∼2 kcal/mol), which may result from an increased classical hydrophobic effect. This relatively minor hydrophobic contribution suggests that the favorable interaction between neutralized cyanuric acid and melamine heterocycles may derive from other factors, such as dispersive and dipole interactions. The exothermic signature of CA/M recognition is more consistent with the binding of heteroaromatic rings41,42,61 in aqueous solvent, including nucleic acid recognition,62 rather than a classical hydrophobic effect. Also implicit in the observation of weakly stabilizing deuterium solvent isotope effects is the minor contribution of hydrogen bonding to binding energy. Certainly, there can be no assembly in the absence of hydrogen bond donor acceptor recognition, as evidenced by

melamine methylation experiments. However, hydrogen bonding is expected to diminish in strength when the shared proton is exchanged for a deuteron;56,57,63,64 thus, one expects a weaker assembly in deuterated buffer if hydrogen bonding is the major driving force. Increased stability in deuterated buffer, in contrast, suggests a minor role in stability but a defining role in selectivity. Influence of Hofmeister Ion Series on TCA/TM Assembly. Given the structural similarities between cyanuric acid and melamine heterocycles and native nucleobases, it seems reasonable that CA/M complexation should also be driven in large part by hydrophobic burial. Assembly should therefore be subject to the influence of Hofmeister salts.65 67 We measured the cooling transition temperature in DSC as a function of salt concentration for three anions that span the Hofmeister series. The salts Na2SO4, NaSCN, and NaCl were used, which define the “salting out” and “salting in” extremes and the middle of the Hofmeister ion series, respectively. A clear trend was observed in the relationship between the temperature at which assembly begins (Ta) and salt concentration, wherein Na2SO4 increased Ta up to 67 °C, NaSCN depressed Ta to 49 °C, and NaCl yielded Ta in between at 54 °C, supporting the notion that there is a significant 8850


Langmuir

ARTICLE

Figure 11. DSC assembly transition temperatures (on cooling) of TCA/TM mixtures as a function of NaSCN (4), NaCl (0), and Na2SO4 (O) salt concentration (A) and ionic strength (B).

hydrophobic burial component to assembly (Figure 11). We find that this trend is indeed a specific salt effect and not merely a manifestation of increasing ionic strength; transition temperature as a function of ionic strength reveals the same Hofmeister ion effect (Figure 11), indicating that the identity of the salt has an effect beyond nonspecific electrostatic screening. When corrected for ionic strength, there remains a 9 °C range in assembly temperature, approximately twice the range observed with ion specific effects on lower critical solution temperatures of poly-Nisopropylacrylamide;68 this heightened sensitivity to ions may be due to the greater number of assembly interfaces in an aggregate of many small molecules as opposed to an aggregate formed from a polymer. Observation of both ion-specific increases and decreases in assembly temperature distinguishes CA/M recognition from that of nucleic acids. It is known that protein folds may be both stabilized and destabilized by specific ions, whereas there are no known ions that can stabilize DNA duplexes. Record and coworkers recently correlated ion specific sensitivity to the nature of the buried surface; micelles and proteins bury surfaces that are >65% hydrocarbon from surfaces while DNA duplex formation buries a surface that is only ∼35% hydrocarbon, with the remainder being mostly polar heteroatoms (N,O).69 Ion specific effects on assembly in water are thought to derive from the degree to which ions partition to the buried surface, which is favored at polar interfaces (salting in) and disfavored at nonpolar interfaces (salting out).65 For largely polar buried surfaces, as found in DNA, ion partitioning to the interface is always favored; thus, one would not predict non-Coulombic stabilizing ion effects on DNA.69 Our experiments on TCA/TM assembly reveal both stabilizing and destabilizing ion effects, a breakdown in similarity between CA/M recognition and nucleobase pairing. Analysis using Chem3D (Supporting Information) indicated that approximately ∼56 59% of the total surface of TCA or TM is contributed by the hydrocarbon linkages, higher than the buried hydrocarbon fraction in DNA duplexes. Despite this significant hydrocarbon fraction, it would appear that favorable heterocycle binding interactions dominate over classical hydrophobic effects in assembly, based on our solvent isotope experiments and ionization state dependent ITC. Overall, there appears to be a reasonable origin for both stabilizing and destabilizing Hofmeister ion effects observed for

TCA/TM assembly, though it is likely that these effects are scaffold specific.

’ CONCLUSIONS Cyanuric acid/melamine self-assembly, considered a canonical example of hydrogen-bond driven assembly in organic solvents, appears to be driven in large part by exothermic surface burial or “base-stacking” effects in water. Given the similarities to nucleobase recognition, this is perhaps to be expected. Aqueous phase CA/M assembly is distinct from base-stacking and even CA/M assembly in organic solvents in that recognition is coupled to an endothermic cyanurate anion protonation. Otherwise, cyanuric acid/melamine assembly behaves in a manner generally consistent with other modes of aqueous-phase recognition: while hydrogen-bonding pattern recognition affords selectivity, assembly is driven significantly by exothermic surface area burial, as observed with other heterocycle systems in water. It is possible that the difficulties associated with the design of discrete molecular recognition based on cyanuric acid and melamine derivatives may derive from insufficient burial of the interacting heterocycles. Extensive surface burial may be required to offset the energetic penalties of protonation of the cyanurate anion. We found that engineering charge, conformation, or number of hydrogen bonding sites on the scaffold was generally detrimental to assembly; the simplest flexible trivalent presentation was optimal. We note that designed recognition in water remains a growth area of research and is in need of additional design guidelines to broaden the scope of controllable artificial interactions in water, including defined interactions with biomolecules.70 73 The lessons learned from assembly of the parent and trivalent CA/M systems will be useful in pursuit of new synthetic recognition modules that function in water. ’ ASSOCIATED CONTENT

bS

Supporting Information. Detailed synthetic procedures, compound characterization, and additional calorimetry data. This material is available free of charge via the Internet at http://pubs.acs.org.

8851


Langmuir

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT This work was supported in part by the NSF (NSF-0747194, NSF-927778) and the Institute of Materials Research at the Ohio State University. ’ REFERENCES (1) Ranganathan, A.; Pedireddi, V. R.; Rao, C. N. R. J. Am. Chem. Soc. 1999, 121, 1752–1753. (2) Mathias, J. P.; Simanek, E. E.; Seto, C. T.; Whitesides, G. M. Macromol. Symp. 1994, 77, 157–166. (3) Zerkowski, J. A.; MacDonald, J. C.; Seto, C. T.; Wierda, D. A.; Whitesides, G. M. J. Am. Chem. Soc. 1994, 116, 2382–2391. (4) Mathias, J. P.; Seto, C. T.; Whitesides, G. M. Polym. Prepr. 1993, 34, 92–93. (5) Seto, C. T.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 905–916. (6) Zerkowski, J. A.; Seto, C. T.; Whitesides, G. M. J. Am. Chem. Soc. 1992, 114, 5473–5475. (7) Seto, C. T.; Whitesides, G. M. J. Am. Chem. Soc. 1990, 112, 6409–6411. (8) Wang, Y.; Wei, B.; Wang, Q. J. Chem. Cryst. 1990, 20, 79–84. (9) ten Cate, M. G. J.; Huskens, J.; Crego-Calama, M.; Reinhoudt, D. N. Chem.—Eur. J. 2004, 10, 3632–3639. (10) Nowick, J. S. Acc. Chem. Res. 2008, 41, 1319–1330. (11) Nowick, J. S.; Cao, T.; Noronha, G. J. Am. Chem. Soc. 1994, 116, 3285–3289. (12) Nowick, J. S.; Chen, J. S. J. Am. Chem. Soc. 1992, 114, 1107– 1108. (13) Oshovsky, G. V.; Reinhoudt, D. N.; Verboom, W. Angew. Chem., Int. Ed. 2007, 46, 2366–2393. (14) Bong, D. T.; Clark, T. D.; Granja, J. R.; Ghadiri, M. R. Angew. Chem., Int. Ed. 2001, 40, 988–1011. (15) Borzsonyi, G.; Beingessner, R. L.; Yamazaki, T.; Cho, J. Y.; Myles, A. J.; Malac, M.; Egerton, R.; Kawasaki, M.; Ishizuka, K.; Kovalenko, A.; Fenniri, H. J. Am. Chem. Soc. 2010, 132, 15136–15139. (16) Fenniri, H.; Packiarajan, M.; Vidale, K. L.; Sherman, D. M.; Hallenga, K.; Wood, K. V.; Stowell, J. G. J. Am. Chem. Soc. 2001, 123, 3854–3855. (17) Marchi-Artzner, V.; Gulik-Krzywicki, T.; Guedeau-Boudeville, M.-A.; Gosse, C.; Sanderson, J. M.; Dedieu, J.-C.; Lehn, J.-M. ChemPhysChem 2001, 2, 367–376. (18) Damodaran, K.; Sanjayan, G. J.; Rajamohanan, P. R.; Ganapathy, S.; Ganesh, K. N. Org. Lett. 2001, 3, 1921–1924. (19) Tazawa, T.; Yagai, S.; Kikkawa, Y.; Karatsu, T.; Kitamura, A.; Ajayaghosh, A. Chem. Commun. 2010, 46, 1076–1078. (20) Yagai, S.; Aonuma, H.; Kikkawa, Y.; Kubota, S.; Karatsu, T.; Kitamura, A.; Mahesh, S.; Ajayaghosh, A. Chem.—Eur. J. 2010, 16, 8652–8661. (21) Yagai, S.; Mahesh, S.; Kikkawa, Y.; Unoike, K.; Karatsu, T.; Kitamura, A.; Ajayaghosh, A. Angew. Chem., Int. Ed. 2008, 47, 4691– 4694. (22) Mahesh, S.; Thirumalai, R.; Yagai, S.; Kitamura, A.; Ajayaghosh, A. Chem. Commun. 2009, 5984–5986. (23) Whitesides, G. M.; Simanek, E. E.; Mathias, J. P.; Seto, C. T.; Chin, D.; Mammen, M.; Gordon, D. M. Acc. Chem. Res. 1995, 28, 37–44. (24) Kawasaki, T.; Tokuhiro, M.; Kimizuka, N.; Kunitake, T. J. Am. Chem. Soc. 2001, 123, 6792–800. (25) Ariga, K.; Kunitake, T. Acc. Chem. Res. 1998, 31, 371–378. (26) Kimizuka, N.; Kawasaki, T.; Hirata, K.; Kunitake, T. J. Am. Chem. Soc. 1998, 120, 4094–4104.

ARTICLE

(27) Marchi-Artzner, V.; Lehn, J. M.; Kunitake, T. Langmuir 1998, 14, 6470–6478. (28) Brunsveld, L.; Vekemans, J. A.; Hirschberg, J. H.; Sijbesma, R. P.; Meijer, E. W. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4977–4982. (29) Brunsveld, L.; Folmer, B. J. B.; Meijer, E. W.; Sijbesma, R. P. Chem. Rev. 2001, 101, 4071–4098. (30) Hirschberg, J. H.; Brunsveld, L.; Ramzi, A.; Vekemans, J. A.; Sijbesma, R. P.; Meijer, E. W. Nature 2000, 407, 167–170. (31) Mateos-Timoneda, M. A.; Crego-Calama, M.; Reinhoudt, D. N. Chem. Soc. Rev. 2004, 33, 363–372. (32) McKay, S. L.; Haptonstall, B.; Gellman, S. H. J. Am. Chem. Soc. 2001, 123, 1244–1245. (33) Ma, M.; Gong, Y.; Bong, D. J. Am. Chem. Soc. 2009, 131, 16919– 16926. (34) Ma, M. M.; Paredes, A.; Bong, D. J. Am. Chem. Soc. 2008, 130, 14456–. (35) Ma, M.; Bong, D. Langmuir 2011, 27, 1480–1486. (36) Hirt, R. C.; Schmitt, R. G. Spectrochim. Acta 1958, 12, 127–138. (37) Koebel, M.; Elsener, M. J. Chromatogr., A 1995, 689, 164–169. (38) Mohlin, K.; Karlsson, P.; Holmberg, K. Colloids Surf., A 2006, 274, 200–210. (39) Covington, A. K.; Paabo, M.; Robinson, R. A.; Bates, R. G. Anal. Chem. 1968, 40, 700–706. (40) Schmuck, C.; Wienand, W. J. Am. Chem. Soc. 2003, 125, 452–459. (41) McKay, S. L.; Haptonstall, B.; Gellman, S. H. J. Am. Chem. Soc. 2001, 123, 1244–1245. (42) Gabriel, G. J.; Iverson, B. L. J. Am. Chem. Soc. 2002, 124, 15174– 15175. (43) Dill, K. A. Biochemistry 1990, 29, 7133–7155. (44) Chandler, D. Nature 2005, 437, 640–647. (45) Yakovchuk, P.; Protozanova, E.; Frank-Kamenetskii, M. D. Nucleic Acids Res. 2006, 34, 564–574. (46) Kool, E. T. Annu. Rev. Biophys. Biomol. Struct. 2003, 30, 1–22. (47) Guckian, K. M.; Schweitzer, B. A.; Ren, R. X. F.; Sheils, C. J.; Tahmassebi, D. C.; Kool, E. T. J. Am. Chem. Soc. 2000, 122, 2213–2222. (48) Whitlock, B. J.; Whitlock, H. W. J. Am. Chem. Soc. 1990, 112, 3910–3915. (49) Klock, C.; Dsouza, R. N.; Nau, W. M. Org. Lett. 2009, 11, 2595–2598. (50) Bakirci, H.; Koner, A. L.; Schwarzlose, T.; Nau, W. M. Chem.— Eur. J. 2006, 12, 4799–4807. (51) Marquez, C.; Nau, W. M. Angew. Chem., Int. Ed. 2001, 40, 3155–. (52) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Acc. Chem. Res. 2009, 42, 1650–1659. (53) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Science 2007, 316, 85–88. (54) Christensen, J. J.; Hansen, L. D.; Izatt, R. M. Handbook of proton ionization heats and related thermodynamic quantities; Wiley: New York, 1976; p xiv, 269 pp (55) Fukada, H.; Takahashi, K. Proteins 1998, 33, 159–166. (56) Bunton, C. A.; Shiner, V. J. J. Am. Chem. Soc. 1961, 83, 42–47. (57) Efimova, Y. M.; Haemers, S.; Wierczinski, B.; Norde, W.; van Well, A. A. Biopolymers 2007, 85, 264–273. (58) Kresheck, G. C.; Schneider, H.; Scheraga, H. A. J. Phys. Chem. 1965, 69, 3132–3144. (59) Blake, R. D. Biochemistry 1981, 20, 5735–5743. (60) Cho, Y.; Sagle, L. B.; Iimura, S.; Zhang, Y.; Kherb, J.; Chilkoti, A.; Scholtz, J. M.; Cremer, P. S. J. Am. Chem. Soc. 2009, 131, 15188–15193. (61) Meyer, E. A.; Castellano, R. K.; Diederich, F. Angew. Chem., Int. Ed. 2003, 42, 1210–1250. (62) SantaLucia, J., Jr.; Hicks, D. Annu. Rev. Biophys. Biomol. Struct. 2004, 33, 415–440. (63) Makhatadze, G. I.; Clore, G. M.; Gronenborn, A. M. Nat. Struct. Biol. 1995, 2, 852–855. (64) Steiner, T. Angew. Chem., Int. Ed. 2002, 41, 48–76. (65) Zhang, Y.; Cremer, P. S. Annu. Rev. Phys. Chem. 2010, 61, 63–83. 8852


Langmuir

ARTICLE

(66) Zhang, Y.; Cremer, P. S. Curr. Opin. Chem. Biol. 2006, 10, 658–663. (67) Pegram, L. M.; Record, M. T. J. Phys. Chem. B 2008, 112, 9428–9436. (68) Zhang, Y.; Furyk, S.; Bergbreiter, D. E.; Cremer, P. S. J. Am. Chem. Soc. 2005, 127, 14505–14510. (69) Pegram, L. M.; Wendorff, T.; Erdmann, R.; Shkel, I.; Bellissimo, D.; Felitsky, D. J.; Record, M. T. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 7716–7721. (70) Mittapalli, G. K.; Osornio, Y. M.; Guerrero, M. A.; Reddy, K. R.; Krishnamurthy, R.; Eschenmoser, A. Angew. Chem., Int. Ed. 2007, 46, 2478–2484. (71) Mittapalli, G. K.; Reddy, K. R.; Xiong, H.; Munoz, O.; Han, B.; De Riccardis, F.; Krishnamurthy, R.; Eschenmoser, A. Angew. Chem., Int. Ed. 2007, 46, 2470–2477. (72) Vysabhattar, R.; Ganesh, K. N. Tetrahedron Lett. 2008, 49, 1314–1318. (73) Dervan, P. B.; Edelson, B. S. Curr. Opin. Struct. Biol. 2003, 13, 284–299.

8853


Determinants of Cyanuric Acid and Melamine Assembly in Water

Recommend Documents