Disaggregation is a Mechanism for Emission Turn ... - ACS Publications

Mar 30, 2017 - Joseph D. Larkin,*,‡,∥ and Eric V. Anslyn*,†. †. Department of Chemistry, University of Texas at Austin, Austin, Texas 78712, U...
0 downloads 0 Views 1MB Size
Subscriber access provided by University of Newcastle, Australia

Article

Disaggregation is a Mechanism for Emission Turn-On of orthoAminomethylphenylboronic Acid-Based Saccharide Sensors Brette M. Chapin, Pedro Metola, Sai Lakshmana Vankayala, H. Lee Woodcock, Tiddo J Mooibroek, Vincent M. Lynch, Joseph D. Larkin, and Eric Van Anslyn J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 30 Mar 2017 Downloaded from http://pubs.acs.org on March 30, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of the American Chemical Society is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Disaggregation is a Mechanism for Emission Turn-On of orthoAminomethylphenylboronic Acid-Based Saccharide Sensors Brette M. Chapin,‡ Pedro Metola,‡ Sai Lakshmana Vankayala,§ H. Lee Woodcock,§ Tiddo J. Mooibroek,† Vincent M. Lynch,‡ Joseph D. Larkin,§#* Eric V. Anslyn‡* ‡

Department of Chemistry, University of Texas at Austin, Austin, Texas 78712, United States

§

Department of Chemistry, University of South Florida, Tampa, Florida 33620, United States



Van’t Hoff Institute for Molecular Sciences, University of Amsterdam, 1098 XH Amsterdam, Netherlands

#

Department of Chemistry, Eckerd College, St. Petersburg, Florida 33711, United States

ABSTRACT: ortho-Aminomethylphenylboronic acid-based receptors with appended fluorophores are commonly used as molecular sensors for saccharides in aqueous media. The mechanism for fluorescence modulation in these sensors has been attributed to some form of photoinduced electron transfer (PET) quenching, which is diminished in the presence of saccharides. Using a well-known boronic acid-based saccharide sensor (3), this work reveals a new mechanism for fluorescence turn-on in these types of sensors. Compound 3 exhibits an excimer, and the associated ground-state aggregation is responsible for fluorescence modulation under certain conditions. When fructose was titrated into a solution of 3 in 2:1 water/methanol with NaCl, the fluorescence intensity increased. Yet, when the same titration was repeated in pure methanol, a solvent in which the sensor does not aggregate, no fluorescence response to fructose was observed. This reveals that the fluorescence increase is not fully associated with fructose binding, but instead disaggregation of the sensor in the presence of fructose. Further, an analogue of the sensor that does not contain a boronic acid (4) responded nearly identically to 3 in the presence of fructose, despite having no functional group with which to bind the saccharide. This further supports the claim that fluorescence modulation is not primarily a result of binding, but of disaggregation. Using an indicator displacement assay and isothermal titration calorimetry, it was confirmed that fructose does indeed bind to the sensor. Thus, our evidence reveals that while binding occurs with fructose in the aqueous solvent system used, it is not related to the majority of the fluorescence modulation. Instead, disaggregation dominates the signal turn-on, and is thus a mechanism that should be investigated in other ortho-aminomethylphenylboronic acid-based sensors.

INTRODUCTION Synthetic receptors can be exploited as molecular sensors using a variety of signal transduction mechanisms.1,2 One particular signaling mechanism was put on solid ground in the 1980’s and 1990’s in studies reported by the de Silva,3 Czarnik,4 and Shinkai5 laboratories. The unifying principle of their studies was the exploitation of a photoinduced electron transfer (PET) mechanism for optical signaling.6 PET can occur from an amine nitrogen lone pair to a neighboring fluorophore, and the electron transfer is curtailed upon binding of the guest. In the case of a sodium ion binding to compound 1, as well as a pyrophosphate acid binding to 2, it is clear that nitrogen lone pairs are increasingly tied up upon coordination to sodium or protonation by a phosphoric acid, respectively, thereby arresting their ability to quench a nearby fluorophore. Hence, a fluorescence turn-on is observed upon guest binding.

that is weak in the boronic acid becomes stronger in the boronate ester, and thus upon binding a saccharide, the nitrogen is more involved in the dative bond and the lone pair cannot participate in PET quenching (Scheme 1A).7–9 This was quite a reasonable postulate, because it is similar to the mechanisms of optical modulations shown for other sensors developed during the same time period. Importantly, irrespective of the specific mechanism of fluorescence modulation, based upon these studies from Shinkai,5–11 and later James,12–15 boronic acids are now the dominant functionality incorporated into synthetic receptors/sensors for saccharides.16

In the case of compound 3, however, the argument is more complex. The postulate was that an N-B interaction

ACS Paragon Plus Environment

Journal of the American Chemical Society

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The PET mechanism for 3 was widely accepted, yet Wang proposed an alternative that has come to be known as the “pKa switch” mechanism.17,18 He postulated that the boronic acid engages in an N-B bond, but upon saccharide binding to generate a boronate ester, solvent insertion occurs and leads to full protonation of the amine, thereby arresting the PET quenching and giving rise to an emission turn-on (Scheme 1B). Wang pointed to several experimental observations that were inconsistent with NB bonding in both the boronic acid and boronate esters,17 and thus his postulate was worthy of investigation. Scheme 1.

A) Change in the strength of an N-B bonding interaction upon sugar binding. B) Breaking of an N-B bond and insertion of solvent upon sugar binding. C) Solvent insertion dominates and does not change upon sugar binding.

In structural studies designed to test the pKa switch mechanism, our group found that there is no significant extent of N-B bonding in either the boronic acid or the boronate ester.19,20 Instead, both are predominantly solvent-inserted in protic media (Scheme 1C). Thus, neither the original postulate of changes in strengths of N-B bonds, nor the pKa switch alternative, could be the entire explanation for how the fluorescence intensity of orthoaminomethylphenylboronic acid-based saccharide sensors, such as 3, respond to saccharides. Scheme 2.

A) Acid/base chemistry of an ortho-ammoniummethyl phenylboronic acid. The first pKa is hydroxylation of the boronic acid. Sugar binding (shown as a diol) keeps the boron hydroxylated. B) Mechanistic role of the orthoammoniummethyl group as an intramolecular general acid catalyst in the rate-determining step of boronate ester formation.

There is no doubt that an ortho-aminomethyl group on a phenylboronic acid increases both the favorable thermodynamics and kinetics of saccharide binding at neutral

Page 2 of 12

pH.21–23 A proximal cationic ammonium group lowers the pKa of the boronic acid, resulting in an anionic boronate conjugate base and an anionic boronate ester (Scheme 2A, diol represents a sugar) at neutral pH. Thus, at this pH boronic acids are stronger Lewis acids and therefore both water binding (solvent insertion) and sugar binding are expected to be stronger relative to the strength of binding in the absence of the ortho-aminomethyl group (we return to this point below). Further, our group has shown that the corresponding ortho-ammoniummethyl group acts as an intramolecular general acid catalyst to increase the speed of the rate-determining loss of water from the anionic boronate conjugate base (Scheme 2B).24 Thus, the only remaining question pertains to the ammonium group’s role in modulating the emission from a pendant fluorophore in boronic-acid based sensors incorporating this group. We were intrigued to understand how the amine plays a role in modulating fluorescence if the boronic acid and the boronate ester are both solventinserted. We chose compound 3 as a paradigmatic example of this class of sensors, and because fructose gave the largest fluorescence response with 3, this was our choice of saccharide. As described herein, our experiments reveal that the ammonium group plays little, if any, role in modulating the fluorescence in 2:1 water/methanol with 50 mM NaCl, as used in studies for pH titrations.5 In other studies, phosphate buffer was used for titrations with sugars, also showing fluorescence turn-on.8,9 But, with the condition of 2:1 water/methanol with 50 mM NaCl, we herein report that the majority of the fluorescence modulation arises from an aggregation/disaggregation phenomenon that is modulated by the addition of fructose to a solution of 3 without binding to the boronic acid, as well as to an additional small extent via binding to the boronic acid. In other words, binding and fluorescence turn-on both occur in the presence of saccharides, but while these two phenomena are correlated, there is no clear causal link between the two.

RESULTS AND DISCUSSION Structural Analysis N-B bonding versus solvent insertion in orthoaminomethylphenylboronic acids or boronate esters is readily distinguished via 11B NMR spectroscopy.19 While a series of 11B NMR spectroscopy studies have confirmed that N-B bonding does not occur to a significant extent in ortho-aminomethyl systems in protic media,19,20 this had not yet been confirmed with compound 3. The original work that studied the fluorescence response to sugars using 3 with pH titrations was carried out in 2:1 water/methanol with 50 mM NaCl,5 and in the studies discussed herein, we used these exact conditions whenever possible. However, for the 11B NMR spectroscopy studies that were aimed at revealing N-B bonding versus solvent insertion, the solubility of 3 was not high enough in this solvent mixture for a significant signal. Instead, we used pure methanol, in which 3 is readily soluble. Because lower dielectric media tend to increase the extent of N-B bonding,19,25 the original solvent system, with significant

ACS Paragon Plus Environment

Page 3 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

amounts of water, would lead to even lower extents of NB bonding. Compound 3 was examined in the presence and absence of fructose using this technique in methanold4 (Figure 1A).

A crystal structure grown from acetonitrile shows that 3 lacks an N-B bond in the solid state, even when the crystal is grown from this solvent (Figure 1B). This was surprising, because aprotic solvents, including acetonitrile, typically induce N-B bonds.18,24 This result is corroborated by the 11B spectrum of 3 in acetonitrile-d3 solution; the single resonance at 29.6 ppm is characteristic of a trigonal boron atom with no N-B bond.

Spectroscopic Studies Having confirmed that there is no detectable N-B bonding in either the boronic acid or the fructose-bound boronate ester, we turned our attention to an analysis of the role of the amine/ammonium group of 3 in the emission response to fructose. We first reproduced a pH titration reported for 3 in the presence and absence of fructose, confirming that we found identical emission turn-on behavior in response to fructose as originally reported (Figure S3). In the pH range between 6 and 10, the fluorescence intensity is considerably greater in the presence of fructose. Thus, the original results were completely reproducible.

Figure 1. A) 11B NMR of 10 mM 3 in CD3OD alone (spectrum a), with 10 mM fructose (spectrum b), and with 100 mM fructose (spectrum c). B) Crystal structure of 3 from acetonitrile. In both the presence and absence of fructose, the 11B NMR chemical shift is in the range of 7-8 ppm, which is characteristic of solvent insertion in both the acid and ester.19 Enlarging the region of 14-15 ppm in search of a peak that would signify N-B bonding did not reveal any evidence of such a resonance. Upon addition of fructose at concentrations that would induce nearly complete binding if an affinity near 103 M-1 is operable (10 mM 3; 10 and 100 mM fructose),26 the boron signal shifts slightly downfield (7.9 ppm) from its chemical shift in the absence of fructose (7.2 ppm), and still no resonance appears in the range of 15 ppm. This is unsurprising, because fructose is likely binding in a tridentate form to the boron atom,27 and this arrangement would be electronically similar to being bound to three solvent molecules. Further, because the 11B NMR chemical shifts undergo little change, one can conclude that the geometry about the boron atom and the strengths of the O-B bonds are undergoing little perturbation upon saccharide binding (a point to which we will return below).

Figure 2. Fluorescence spectra of a saturated solution of 3 in 2:1 water/methanol with 50 mM NaCl. 1-cm path length cuvette, slit widths 4 nm, integration time 0.5 s. Emission scan with λex = 368 nm (blue). Excitation scan with λem = 417 nm (red). Excitation scan with λem = 520 nm (green). Emission scan with λex = 408 nm (purple).

During these fluorescence studies, an emission scan with λex = 368 nm was carried out on a saturated solution (prepared to be 100 µM, but slightly cloudy due to incomplete dissolution) of 3 in 2:1 water/methanol with 50 mM NaCl in the absence of fructose. The expected monomer fluorescence was observed, with its maximum at 417 nm. However, unexpectedly, in some cases there was also a broad emission peak with its maximum at approximately 530 nm. In Figure 2, the blue curve shows an example of one of the largest occurrences of an emission peak found in this range. However, repeated analysis showed that the size of this peak was both condition-dependent and sample-dependent. When this secondary emission peak is small, it is easy to imagine that it could have been missed without prior knowledge of its existence. It could also be missed during a titration with fructose, because the overall emission of the monomer grows greatly at the expense of this far smaller peak.

ACS Paragon Plus Environment

Journal of the American Chemical Society

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Clearly, we were observing an excimer of anthracene. While this excimer emission, previously unknown for compound 3, was initially surprising, perusal of the literature revealed that anthracene has been known to exhibit an excimer.28–30 While anthracene can also form a photodimer, no evidence for this was ever observed under these conditions. As with other excimers, this emission peak is broad and structureless, and is observed at a longer wavelength than its corresponding monomer peak. These features are due to self-quenching and internal conversion.31 Additionally, except with pyrene, excimers generally have far lower quantum yields than their monomeric counterparts.31 In fact, various sensors using anthracene as the fluorescent reporter have exploited modulations in such excimers.32 Further, pyrene excimer and exciplex formation has been used in boronic acid-based sensors, albeit PET mechanisms were still the postulated mechanisms for turn-on.33,34 In order to provide further evidence for the formation of an excimer with 3, additional emission and excitation scans were carried out. An excitation scan was carried out while monitoring at λem = 417 nm, the emission maximum of the monomer (Figure 2, red curve). The excitation profile is the mirror image of the emission, as expected. A second excitation scan was carried out while monitoring λem = 520 nm, emission of the putative excimer. This excitation profile, the green curve, has a completely different shape from the monomer excitation profile, and resembles previously reported anthracene excimer excitation spectra.35 Furthermore, this difference in shape suggests that the excimer is a result of an interaction facilitated by a ground-state aggregate of 3. The difference in excitation profiles can also be exploited, as there are wavelengths where the excimer alone can be excited and the fluorescence of the excimer can be observed in isolation. To achieve this, an emission scan was carried out at λex = 408 nm (purple curve). At this wavelength, excitation of the monomer (red curve) is at nearly zero, but the excitation that leads to the broad emission (green curve) is approximately half of its maximum fluorescence intensity. As expected, the purple curve shows only broad fluorescence with no trace of the monomer maxima, and definitively confirms that the broad emission peak represents an excimer. To test if binding fructose modulated the monomer emission and the extent of excimer fluorescence, a series of titrations was performed. Compound 3 (12 µM, homogeneous and no longer cloudy, at the same concentration used in the original report5) was titrated with fructose in 2:1 water/methanol with 50 mM NaCl, allowing ten minutes for equilibration after each addition of fructose. This titration showed a lack of saturation (Figure S4), even though it was carried out over a concentration range of 0-200 mM fructose, where a binding constant on the order of 103 M-1 should have shown saturation of fluorescence intensity. Given that the literature suggests that fructose should bind 3 with an affinity near 103 M-1,26 as when a similar titration is performed using phosphate buffer,8,9 we postulated that there was some slow kinetic

Page 4 of 12

step involved in modulating the emission. Thus, the fluorescence titration was repeated, but with many samples, each with a different concentration of fructose, prepared individually from stock solutions and stored at room temperature in the dark overnight to allow equilibration of the emitting species. The emission spectrum for each individual sample was then obtained, and the resulting isotherm suggested that 3 had saturated with fructose. The data was fit to a one-to-one binding isotherm with an association constant of K = 1.3 × 102 ± 2.6 × 101 M-1 (Figure 3). Compared to literature precedent,5,8,9 the emission response did not accurately report the thermodynamics of fructose binding.

Figure 3. A) Fluorescence spectra of a 12 µM solution of 3 in 2:1 water/methanol with 50 mM NaCl, 0-32 mM fructose, λex = 368 nm. Semi-micro cuvette, slit widths 2 nm, integration time 1 s. B) Change in fluorescence at 417 nm plotted against concentration of fructose and fit to a one-to-one binding curve.

In our experience, boronic acids and boronate esters exchange and reach equilibrium within five to ten minutes, if not faster.36 The time dependence of the fructose titration in Figure 3 was thus puzzling, and suggested that perhaps the fluorescence of 3 changed over time on its own. In fact, this was found to be the case (Figure 4). Five repeat scans at λex = 368 and λex = 408 nm were carried out with 12 µM 3 alone in 2:1 water/methanol with 50 mM NaCl. With each pair of scans, monomer fluorescence increased and excimer fluorescence decreased. Solutions were then stored at room temperature in the dark

ACS Paragon Plus Environment

Page 5 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

overnight and scanned four additional times the following day. In all cases, fluorescence intensity of the monomer increased in the absence of fructose while the fluorescence of the excimer decreased. Furthermore, irradiating the sample repeatedly led to greater fluorescence intensity increases than time alone could produce. This supports the notion that excimer formation is facilitated by a ground-state aggregate which breaks up over time, and that the breakup is accelerated by excitation. In contrast, if the excimer formed during a collision between free monomers in solution, the extent of excimer fluorescence would not display this time dependence.

founded with one another, such that the binding curve in Figure 3B is not truly representative of binding alone. Similar behavior has been found in an entirely different sugar-binding study, in which Davis showed that the UVvisible absorption of a porphyrin-based host changes in the presence of glucose, but that these changes are best explained by a change in aggregation state of the porphyrin, and do not truly signify glucose binding.37 In order to better understand the time dependence of the monomer fluorescence of 3 alone, a solution of 12 µM 3 was monitored with repeated emission scans for approximately 72 hours. Even over a span of three days, the monomer fluorescence continued to increase and did not reach a plateau (Figure S5). Additionally, a sample of 3 in 2:1 water/methanol with 50 mM NaCl that was left in the dark became more fluorescent (at the monomer wavelength) over time, but an aliquot of the same solution that was repeatedly subjected to excitation over the same time period increased faster. In other words, as shown in Figure 4, irradiation increased the rate of disaggregation of 3. As with any excited compound, 3 can release energy and return to its ground state via fluorescence, internal conversion, or intersystem crossing followed by phosphorescence. In the particular case of 3, internal conversion can occur from an excited monomer or from the excimer, meaning that it can occur within the aggregate. Internal conversion is typically conceptualized as a simple relaxation, a radiationless transition that releases energy as heat. In this case, the released energy is localized and it assists the monomers to break away from the rest of the aggregate. Similarly, other methods of introducing thermal energy into the system, such as heating or sonication, should also result in accelerated aggregate breakup and a greater increase in monomer fluorescence without any addition of fructose.

Figure 4. Fluorescence spectra of a 12 µM solution of 3 in 2:1 water/methanol with 50 mM NaCl, λex = 368 nm (A) and λex = 408 nm (B). 1-cm path length cuvette, slit widths 4 nm, integration time 1 s.

Therefore, we propose that the fluorescence change over time is a result of long-lived aggregates that remain intact for several hours but gradually dissociate. These aggregates are likely present because compound 3, while not saturated at 12 µM, is close to its solubility limit. The aggregates were too small to detect by dynamic light scattering, and too insoluble to observe via DOSY NMR spectroscopy. Similarly, no detectable boroxines could be found in mass spectrometry. While the structures of the aggregates have not been characterized, we expect some combination of π-interactions and solvophobic effects are responsible for aggregation. While definitively present, the surprising feature of these aggregates is that they have a significant kinetic barrier to dissociation. Irrespective of their structure, their presence and time dependence lead to aggregate dissociation and fructose binding being con-

Following this hypothesis, and in an attempt to separate the variables of disaggregation and sugar-binding and their effect on fluorescence intensity, a 12 µM 2:1 water/methanol with 50 mM NaCl solution of 3 was sonicated in order to speed up the process of disaggregation (Figure 5). An emission scan with λex = 368 nm was carried out (first blue point) and the solution was sonicated for 90 minutes, and scanned again (red point). The solution was sonicated for a second 90-minute period, and then scanned (green point). Then scans were repeated until the fluorescence intensity remained constant, and finally fructose was added (50 mM, purple point). The 50 mM concentration of fructose was chosen to correspond to that used in the original studies.5 Repeated scans resulted in very small increases after the addition of fructose. The total turn-on of fluorescence is along the same magnitude as that shown in Figure 3. As Figure 5 shows, fluorescence intensity increased by more than fifteen-fold due to disaggregation, and then increased by nearly two-fold upon addition of fructose, for a total of approximately a thirty-fold turn-on. However, as can be seen in the titration, the magnitude of the increases along the y-axes are approximately the same.

ACS Paragon Plus Environment

Journal of the American Chemical Society

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

When similar studies were performed in phosphate buffer, a four-fold turn-on of fluorescence was found,8,9 far less than the total thirty-fold found herein. Important for this study, a large fraction of the fluorescence turn-on occurs without any sugar binding. However, fructose did reliably bring about this additional two-fold turn-on over multiple repeats of this experiment. Thus, it is clear that fructose does make a difference in the emission, but it could not yet be determined what role fructose plays in increasing emission of 3.

Page 6 of 12

the 2:1 water/methanol with 50 mM NaCl solvent system (Figure 5) is due to additional disaggregation upon addition of fructose. It also indicates that fructose binding gives no emission response when all aggregates are broken up from the start. De Silva has reported a similar phenomenon in which a PET-based sensor was fully turned on in lower dielectric media.39 Further, Fahrni reported that when a pH sensor was introduced to cells and adsorbed into the cytoskeleton, emission was fully on, irrespective of pH.40 Potentially, disaggregation plays a role in these systems also. The measurement of fluorescence lifetimes is another technique commonly used to characterize excimers.41 The existence of an excimer was also corroborated by lifetime studies that uncovered a complex decay composed of two distinct lifetimes (Figure S8A). Yet, upon addition of fructose (Figure S8B), the decay curve fit a single exponential function, revealing that the presence of fructose does lead exclusively to monomeric species in solution.

Figure 5. Fluorescence of a 12 µM solution of 3 in 2:1 water/methanol with 50 mM NaCl. λex = 368 nm, λem = 417 nm. 1-cm path length cuvette, slit widths 2 nm, integration time 1 s. Red: scan after 90 minutes of sonication. Green: scan after another 90 minutes of sonication. Purple: scan after adding fructose (50 mM).

The 2:1 water/methanol with 50 mM NaCl mixture is expected to favor aggregation of hydrophobic entities relative to lower dielectric media, as is akin to “salting out”.38 Thus, similar studies to those described above were carried out in pure methanol, where 3 is freely soluble. First, the fluorescence intensity of a 12 µM solution of 3 was monitored over many scans (Figure S6). Unlike in Figure S5, the fluorescence intensity did not increase over time, but instead remained approximately constant. We propose that this is due to the much greater solubility of 3 in pure methanol, resulting in little or no aggregation at the outset. Given that 3 did not appear to be aggregated in methanol, we then turned to determine whether the addition of fructose would lead to increased fluorescence intensity. A titration of fructose into 12 µM 3 was carried out in pure methanol. As in the titration depicted in Figure 3A, each sample was prepared individually from stock solutions, stored at room temperature in the dark overnight to equilibrate any kinetic issues, and emission spectra were obtained the next day (Figure S7). The result was a lack of modulation of the fluorescence of 3 upon addition of fructose. This is due to the fact that the fluorescence intensity of 3 alone in methanol is already high, and in fact, at its maximum; it cannot further increase over time or with the addition of fructose (and fructose binding). The fact that we did not find any emission enhancement upon fructose binding in methanol indicates, but is not necessarily definitive, that the nearly two-fold enhancement in

No Boronic Acid in the Receptor The discussion has, in part, relied on contrasting the behavior of 3 with the known thermodynamics and kinetics of sugar binding to other boronic acids, assuming that this was relevant to compound 3. To remove the possibility of binding of sugars altogether and follow only the aggregation state behavior, the boronic acid moiety would need to be absent. To this end, compound 4 was synthesized and treated to the same titration conditions as discussed above with 3.

A titration of fructose into 4 (12 µM) was carried out in 2:1 water/methanol with 50 mM NaCl. Each sample was prepared individually from stock solutions, stored at room temperature in the dark overnight, and emission spectra were obtained the next day (Figure 6A), just as was done with compound 3. The result is striking, because although 4 does not contain a boronic acid with which to bind fructose, the titration still produces a curve that fits to a one-to-one binding isotherm (Figure 6B) with a binding constant of 87 ± 18 M-1. It is, of course, unreasonable to call this one-to-one binding because no binding is taking place. Instead, this curve was simply fit in order to compare to the titration with compound 3 (Figure 3B). However, the fact that it fits, with an apparent “affinity” basically the same (within error) as that for compound 3, further brings into question the validity of the one-to-one binding curve for compound 3. We postulate that both 3 and 4 are undergoing disaggregation due to the addition of fructose (while 3 is also binding fructose), but that the binding of fructose is not being reported by the majority of the emission response.

ACS Paragon Plus Environment

Page 7 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society another 90 minutes and scanned again (green point). Scans were repeated until fluorescence intensity remained constant, and then fructose was added to be 50 mM (purple point). Adding fructose resulted in little, if any, increase in fluorescence intensity, and repeat scans had no effect. This indicates that the two-fold change in fluorescence of 3 upon addition of fructose must be associated with fructose in solution, but the manner in which fructose influences fluorescence was still an open question.

Confirming Fructose Binding While Figures 5 and 7 present evidence that fluorescence modulation of 3 in the presence of fructose may not be due to binding fructose, we do not mean to suggest that binding does not occur. As such, we now lay out evidence that 3 does indeed bind fructose, and measure the strength of that binding.

Figure 6. A) Fluorescence spectra of a 12 µM solution of 4 in 2:1 water/methanol with 50 mM NaCl, 0-120 mM fructose, λex = 368 nm. Semi-micro cuvette, slit widths 2 nm, integration time 1 s. B) Change in fluorescence at 417 nm plotted against concentration of fructose and fit to a one-to-one binding curve.

Figure 7. Fluorescence of a 12 µM solution of 4 in 2:1 water/methanol with 50 mM NaCl. λex = 368 nm, λem = 417 nm. 1-cm path length cuvette, slit widths 2 nm, integration time 1 s. Red: scan after 90 minutes of sonication. Green: scan after another 90 minutes of sonication. Purple: scan after adding fructose.

Investigation of compound 4 continued with a study (Figure 7) similar to that shown in Figure 5. A 12 µM solution of 4 in 2:1 water/methanol with 50 mM NaCl was scanned (initial blue point), sonicated for 90 minutes, and scanned again (red point). The solution was sonicated for

Figure 8. A) UV-Vis absorption spectra of a 80 µM solution of PV in methanol, 0-833 µM 3. 1-cm path length cuvette, 10min. equilibration time between additions. B) Change in absorbance at 525 nm plotted against concentration of 3 and fit to a one-to-one binding curve. The lack of a clean isosbestic point is due to a small pH change during the titration because the methanol was not buffered.

The first method used to measure a binding constant was an indicator displacement assay (IDA). Pure methanol had to be used rather than the water/methanol mixture due to limited solubility of 3 in the mixed solvent. First, a one-to-one binding curve had to be established for 3 and an indicator, Pyrocatechol Violet (PV). Figure 8

ACS Paragon Plus Environment

Journal of the American Chemical Society

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

shows the absorbance curves (A) and one-to-one binding curve (B). Fitting the data to a one-to-one binding curve produced a binding constant of 1.5 × 104 ± 3.0 × 103 M-1. Next, the IDA was performed. The concentrations of PV (80 µM) and 3 (170 µM) were kept constant while fructose was titrated in (Figure 9). This curve was fit to an indicator displacement isotherm.42 The fit gave a binding constant of 3.8 × 103 ± 7.9 × 102 M-1 for the binding of 3 and fructose. This value is in the range of reported fructose affinities to boronic acids,26 and that found in a phosphate buffer solution of 3, rather than the 1.3 × 102 M-1 we found using fluorescence spectroscopy.

Page 8 of 12

be observed in isolation. This binding curve is plotted in Figure 10B. The curve was fit to give a binding constant of 2.9 × 103 ± 5.4 × 102 M-1, which is in reasonable agreement with the binding constant found via indicator displacement. Since ITC measures both ∆Ho and Ka, ∆Go and ∆So can also be calculated. The results shown in Figure 10 give ∆Ho = -0.2 kcal/mol and ∆So = 15.8 cal/(mol·K). A titration of fructose into compound 4 gave no ITC response beyond heat of dilution, as may be expected because it is not aggregated in methanol and does not bind fructose. The ITC with 3 and fructose lends insight with regard to boronic acid-sugar recognition. The results indicate that ∆Ho barely contributes to the strength of binding, and that the binding is almost entirely entropy-driven. This is explained by the fact that one molecule of fructose binds in a tridentate manner and releases three molecules of solvent upon binding. Because enthalpy is dictated by bond strengths, the small ∆Ho indicates that the trimethoxy boronate ester (or boronic acid) and the fructose boronate ester are very similar in the strength of their OB bonds, as may be expected. Moreover, it is in agreement with the fact that the 11B NMR signals of 3 shifted only slightly in the presence of fructose, indicating little difference in the boron environment upon binding a sugar (Figure 1A).

Figure 9. A) UV-Vis absorption spectra of a solution of 80 µM PV and 170 µM 3 in methanol, 0-5.2 mM fructose. 1-cm path length cuvette, 10-minute equilibration time between additions. B) Change in absorbance at 610 nm plotted against concentration of fructose and fit to an indicator displacement curve. The clean isosbestic point here reveals no pH change during this titration.

A second method used to measure a binding constant was isothermal titration calorimetry (ITC). Again, compound 3 was not soluble enough in 2:1 water/methanol with 50 mM NaCl, so the titration was carried out in pure methanol. Fructose (10 mM) was titrated into 3 (1 mM), and a control in which fructose was titrated into pure methanol was carried out with the same fructose solution. Figure 10A shows the titration of fructose into 3 (black) and the titration of fructose into methanol (red). This control allows the heat of dilution of fructose from each addition to be subtracted, and thus the binding curve can

Figure 10. A) ITC of 10 mM fructose titrated into 1 mM 3 in methanol (black), and ITC of 10 mM fructose titrated into methanol (red). 25 ºC, reference power 5 µcal/s, stir speed

ACS Paragon Plus Environment

Page 9 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

300. Initial delay 60 s, all injections 12 µL with 600-s spacing. B) Plot of the difference of the two titrations in (A).

Combining the Structural and Thermodynamic Data We noted in the introduction that both boronic acids and boronate esters are solvent-inserted in protic solvent. Because both species are solvent-inserted with three O-B bonds, it is logical to question whether an enthalpic driving force for binding a sugar to the boronic acid would exist. In fact, the ITC reveals that the driving force is primarily an entropic one, and indeed any enthalpy changes (bond strengths and solvation changes) are minimal. This calls into question whether the ortho-aminomethyl group enhances boronate ester formation relative to boronic acid solvent insertion. The classic postulate has been that the boronic acid is induced to be pyramidal by the neighboring amine, which alleviates strain in a boronate ester derived from binding a sugar compared to trigonal planar boron. This is because the geometry around boron is better if it is tetrahedral in a five-membered ring when binding vicinal diols of sugars.42 While this argument may hold when comparing a trigonal boronic acid to a tetrahedral boronate ester, it does not hold at a pH where the boronic acid conjugate base and the boronate ester formed with the sugar are both tetrahedral. The increased electrophilicity of the boron, by virtue of the neighboring ammonium group, acts nearly equally to bind a third solvent or a sugar. Our ITC and 11B NMR studies show this to be the case; there is no relief of strain as would be evident in a ∆H term, and there is no significant change in 11B NMR chemical shift as would occur with a significant geometry change. Molecular Dynamics Recall that there is a nearly two-fold increase in the fluorescence intensity of 3 observed upon addition of fructose (Figure 5). There are several possibilities for how fructose leads to this small emission increase. The first is a PET-based mechanism such as originally postulated by Shinkai/James and Wang, but the fact that no significant fraction of an N-B bond exists seems to negate this possibility. Further, the fact that 3 shows no fluorescence response to fructose at all in methanol (when fructose was confirmed to be binding) supports the claim that when fructose binds to 3 in 2:1 water/methanol with 50 mM NaCl, there should also be no corresponding fluorescence increase due to binding. Second, the fructose could act as a catalyst for disaggregation – binding 3 within the aggregate, and pulling it into bulk solution and releasing it. Yet, we note that after sonication and irradiation of a solution of 3, an equilibrium distribution of 3 in bulk solution and some extent of disaggregation has been achieved (Figure 5). Thus, addition of fructose should not change an equilibrium distribution if it is solely a catalyst (catalysts do not change thermodynamics). A third possibility is that binding of 3 to fructose decreases aggregation because when the fructose is bound to 3 the complex is more soluble. Lastly, the mere fact that fructose is in an extremely large excess over 3 in the original studies (~ 4200 to 1),5 as well as in the studies shown in Figures 3

and 5, means that fructose could be significantly changing the solvent properties and thereby the solubility of 3 (or 4), thus breaking up aggregation. These latter two possibilities seem most likely, and thus we turned to computational modeling to characterize the aggregates and how the aggregates change both with fructose in solution and with fructose bound to 3. 180-Nanosecond simulations were performed in a cubic 2:1 water/methanol explicit solvent box using CHARMM.43 For computational simplicity, a 2:1 water/methanol ratio minus the NaCl was used. To analyze the simulations, radial distribution function (RDF) plots were used for their ability to capture aromatic interactions between small molecules. The RDF describes how closely aromatic carbons are radially packed around each other in 3dimensional space from any given aromatic carbon as a reference point, and serves as a useful tool to capture the average structure of disordered molecules such as liquids and mixed solvents.

Figure 11. g(r) vs. r of A) Compound 3 in 2:1 water/methanol (no NaCl present) explicit solvent, B) Compound 3 and fruc-

ACS Paragon Plus Environment

Journal of the American Chemical Society

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

tose uncomplexed in 2:1 water/methanol explicit solvent, and C) Compound 3 complexed with fructose, in 2:1 water/methanol explicit solvent.

The density of probability, g(r), of an aromatic carbon is evaluated to find neighboring aromatic carbons at a given distance r. In Figure 11 we plot the g(r) of compound 3 alone (A), fructose in solution with 3 (uncomplexed) (B), and fructose complexed with 3 (C). The RDF plots in Figure 11 compare well with experiment and show a broadening when fructose is included in the simulations. The occurrence of broad peaks at a longer distance r(Å) spread across the x-axis indicate minimal aromatic interactions, i.e. a lack of aromatic aggregates due to a high degree of distant ordering in the system. In contrast, sharper peaks at shorter distances are indicative of stronger aromatic stacking interactions leading to aggregation. Figures 11B and 11C demonstrate the effect of disaggregation upon inclusion of fructose in solution and when bound, as seen by the aggregate density having a smaller g(r). We further tested the aggregation/disaggregation hypothesis by performing analogous simulations on compound 4 alone and in the presence of fructose. The RDF plots in Figure 12 reveal a similar shift in the presence of fructose, indicating decreased aggregation. The radius of aggregate density increases when fructose is included in the simulation.

Figure 12. g(r) vs. r of A) Compound 4 in 2:1 water/methanol explicit solvent, and B) Compound 4 and fructose uncomplexed in 2:1 water/methanol explicit solvent.

The molecular dynamics simulations validate that simply having fructose in solution with 3 leads to a larger

Page 10 of 12

radial distance between the individual receptor molecules, most likely due to a change in solvent properties. Further, the binding of fructose to 3 also leads to a lesser extent of aggregation, likely because the complex is more soluble than 3 alone. While the computational analysis cannot confirm which effect causes the nearly two-fold enhancement in the emission of 3 upon addition of fructose, it does support or claim that disaggregation accounts for the turn-on without any contribution of a PETbased mechanism. While the molecular dynamics simulations support that 3 is indeed aggregated in water/methanol mixtures, they do not lend insight into the unexpected kinetic barrier that leads to the persistence of the aggregates over several hours. Thus, the precise nature of the aggregates, the mechanism of disaggregation, and the exact manner in which fructose influences aggregation are still under investigation in our laboratories.

CONCLUSIONS Compound 3 forms ground-state aggregates in 2:1 water/methanol with 50 mM NaCl, likely due to the fact that the solution is nearly saturated in 3, and these aggregates lead to the formation of excimers upon irradiation. Disaggregation is observed via increase of the monomer fluorescence emission at λem = 417 nm and/or decrease of the excimer fluorescence emission at λem = 520 nm as a function of irradiation, time, sonication, and fructose addition. In pure methanol, the excimer is not observed and the fluorescence of the monomer does not increase over time or with the addition of fructose, indicating that 3 does not aggregate in methanol. While fructose certainly binds, as shown by an indicator displacement assay and isothermal titration calorimetry, we assert that the vast majority of fluorescence modulation is unaffected by this phenomenon, while a small portion (two-fold turn on in 2:1 water/methanol with NaCl, or four-fold turn on in phosphate buffer8,9) of the emission response is either simply a result of changes in solvent properties (due to a large excess of fructose), or due to fructose binding that causes an even further increase in disaggregation, or that phosphate buffer can play additional roles (see below). These claims are supported by the fact that compound 4 also appears to demonstrate binding when the fluorescence spectra in the presence of fructose are examined, despite the fact that it contains no boronic acid with which to bind fructose. The molecular dynamics simulations further support the claim that the presence of fructose alone leads to disaggregation of 3 or 4. We do not propose that all boronic acid fluorescence turn-on sensors operate by a disaggregation mechanism. For insteance, polymer-bound sensors44–49 are presumably spaced apart, and cannot operate in this manner. Further, as repeatedly noted herein, when phosphate buffer is used in methanol/water mixtures four-fold fluorescence turnon is observed in titrations with sugars, and these titrations reveal the correct affinities.8,9 We note that Fyles/James and Wang have found that phosphate binds to boronic acids,50,51 and thus may act to break up aggre-

ACS Paragon Plus Environment

Page 11 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

gates. The mechanism of turn-on in these cases still awaits further investigation that is on-going in our laboratories. Therefore, we have uncovered an unprecedented mechanism by which ortho-aminomethylphenylboronic acids can signal (sense) the presence of sugars. We suggest that the role of disaggregation should be examined when mechanistic postulates are considered for fluorescence sensing. This phenomenon could be even more widespread than just sugar sensing using boronic acids.

ASSOCIATED CONTENT

(12) (13) (14) (15) (16) (17) (18) (19) (20)

Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Synthetic preparations, experimental procedures, and control experiments with figures (PDF) Crystal structure (CIF)

AUTHOR INFORMATION

(21) (22) (23)

(24) (25)

Corresponding Authors

(26)

* [email protected] * [email protected]

(27)

ACKNOWLEDGMENTS Drs. Claudia Turro, Dario Bassani, Tony Davis, and Marco Bonizzoni are gratefully acknowledged for their suggestions and advice. Steve Sorey and Angela Spangenberg are thanked for their help with NMR spectroscopy and Dr. Raluca Gearba is recognized for her help with fluorescence lifetime studies. Further, we gratefully acknowledge insights from Binghe Wang and Seiji Shinkai, and extensive discussions with Tony James. EVA, BMC, and PM thank the National Science Foundation (grant CHE-1212971) and EVA additionally thanks the Welch Regents Chair (F-0046) for funding. SLV, HLW, and JDL would like to thank the National Heart, Lung, and Blood Institute of the National Institutes of Health (grant K22HL113045) and the National Science Foundation (grant CHE-1531590) for support.

(28) (29) (30) (31) (32) (33) (34) (35) (36) (37) (38) (39)

REFERENCES (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

You, L.; Zha, D.; Anslyn, E. V. Chem. Rev. 2015, 115, 7840. Wu, J.; Kwon, B.; Liu, W.; Anslyn, E. V.; Wang, P.; Kim, J. S. Chem. Rev. 2015, 115, 7893. de Silva, A. P.; de Silva, S. A. J. Chem. Soc., Chem. Commun. 1986, 23, 1709. Vance, D. H.; Czarnik, A. W. J. Am. Chem. Soc. 1994, 116, 9397. James, T. D.; Sandanayake, K. R. A. S.; Shinkai, S. J. Chem. Soc., Chem. Commun. 1994, 4, 477. James, T. D.; Linnane, P.; Shinkai, S. Chem. Commun. 1996, 3, 281. James, T. D.; Sandanayake, K. R. A. S.; Shinkai, S. Angew. Chem. 1994, 106, 2287. James, T. D.; Sandanayake, K. R. A. S.; Shinkai, S. Nature 1995, 374, 345. James, T. D.; Sandanayake, K. R. A. S.; Iguchi, R.; Shinkai, S. J. Am. Chem. Soc. 1995, 117, 8982. Sandanayake, K. R. A.; Shinkai, S. J. Am. Chem. Soc., Chem. Commun. 1994, 9, 1083. Kimura, T.; Takeuchi, M.; Nagasaki, T.; Shinkai, S.

(40) (41) (42) (43)

(44) (45) (46)

Tetrahedron Lett. 1995, 36, 559. Cooper, C. R.; James, T. D. Chem. Commun. 1997, 15, 1419. Cooper, C. R.; James, T. D. Chem. Lett. 1998, 9, 883. Ward, C. J.; Patel, P.; Ashton, P. R.; James, T. D. Chem. Commun. 2000, 3, 229. Arimori, S.; Bell, M. L.; Oh, C. S.; Frimat, K. A.; James, T. D. Chem. Commun. 2001, 18, 1836. James, T. D. Top. Curr. Chem. 2007, 277, 107. Franzen, S.; Ni, W.; Wang, B. J. Phys. Chem. B 2003, 107, 12942. Ni, W.; Kaur, G.; Springsteen, G.; Wang, B. Bioorg. Chem. 2004, 32, 571. Zhu, L.; Shabbir, S. H.; Gray, M.; Lynch, V.; Sorey, S.; Anslyn, E. V. J. Am. Chem. Soc. 2006, 128, 1222. Collins, B. E.; Sorey, S.; Hargrove, A. E.; Shabbir, S. H.; Lynch, V.; Anslyn, E. V. J. Org. Chem. 2009, 74, 4055. Lorand, J. P.; Edwards, J. O. J. Org. Chem. 1959, 24, 769. Wulff, G. Pure Appl. Chem. 1982, 54, 2093. Wulff, G.; Dederichs, W.; Grotstollen, R.; Jupe, C. In Analytical Chemistry Symposia Series, Affinity Chromatography and Related Techniques; Gribnau, T. C. J., Visser, J., Nivard, R. J. F., Eds.; Elsevier: Amsterdam, 1982; pp 207–216. Collins, B. E.; Metola, P.; Anslyn, E. V. Supramol. Chem. 2013, 25, 79. Chapin, B. M.; Metola, P.; Lynch, V. M.; Stanton, J. F.; James, T. D.; Anslyn, E. V. J. Org. Chem. 2016, 81, 8319. Nagai, Y.; Kobayashi, K.; Toi, H.; Aoyama, Y. Bull. Chem. Soc. Jpn. 1993, 66, 2965. Norrild, J. C.; Eggert, H. J. Chem. Soc., Perkin Trans. 2 1996, 12, 2583. Chandross, E. A. J. Chem. Phys. 1965, 43, 4175. Chandross, E. A.; Ferguson, J.; McRae, E. G. J. Chem. Phys. 1966, 45, 3546. McVey, J. K.; Shold, D. M.; Yang, N. C. J. Chem. Phys. 1976, 65, 3375. Birks, J. B. Rep. Prog. Phys. 1975, 38, 903. Van Arman, S. A.; Czarnik, A. W. J. Am. Chem. Soc. 1990, 112, 5376. Jin, T. Chem. Commun. 1999, 24, 2491. Arimori, S.; Bell, M. L.; Oh, C. S. Org. Lett. 2002, 4, 4249. Momiji, I.; Yoza, C.; Matsui, K. J. Phys. Chem. B 2000, 104, 1552. Seifert, H. M.; Trejo, K. R.; Anslyn, E. V. J. Am. Chem. Soc. 2016, 138, 10916. Renney, C. M.; Fukuhara, G.; Inoue, Y.; Davis, A. P. Chem. Commun. 2015, 51, 9551. Pal, M. K. Histochemie 1965, 5, 24. Bissell, R. A.; de Silva, A. P.; Thilak, W.; Fernando, M. L.; Patuwathavithana, S. T.; Shantha, T. K.; Samarasinghe, D. Tetrahedron Lett. 1991, 32, 425. Fahrni, C. J.; Yang, L.; VanDerveer, D. G. J. Am. Chem. Soc. 2003, 125, 3799. Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3rd ed.; Springer: New York, 2006. Hargrove, A. E.; Zhong, Z.; Sessler, J. L.; Anslyn, E. V. New J. Chem. 2010, 34, 348. Brooks, B. R.; Brooks, C. L. I.; MacKerell, A. D. J.; Nilsson, L.; Petrella, R. J.; Roux, B.; Won, Y.; Archontis, G.; Bartels, C.; Boresch, S.; Caflisch, A.; Caves, L.; Cui, Q.; Dinner, A. R.; Feig, M.; Fischer, S.; Gao, J.; Hodoscek, M.; Im, W.; Kuczera, K.; Lazaridis, T.; Ma, J.; Ovchinnikov, V.; Paci, E.; Pastor, R. W.; Post, C. B.; Pu, J. Z.; Schaefer, M.; Tidor, B.; Venable, R. M.; Woodcock, H. L.; Wu, X.; Yang, W.; York, D. M.; Karplus, M. J. Comput. Chem. 2009, 30, 1545. Wu, X.; Li, Z.; Chen, X.-X.; Fossey, J. S.; James, T. D. Chem. Soc. Rev. 2013, 42, 8032. Lacina, K.; Skladal, P.; James, T. D. Chem. Cent. J. 2014, 8, 60. Crane, B. C.; Barwell, N. P.; Gopal, P.; Gopichand, M.; Higgs, T.; James, T. D.; Jones, C. M.; MacKenzie, A.;

ACS Paragon Plus Environment

Journal of the American Chemical Society

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(47) (48) (49)

Mulavisala, K. P.; Paterson, W. J. Diabetes Sci. Technol. 2015, 9, 751. Zhai, W.; Sun, X.; James, T. D.; Fossey, J. S. Chem. Asian J. 2015, 10, 1836. Sun, X.; Zhai, W.; Fossey, J. S.; James, T. D. Chem. Commun. 2016, 52, 3456. Brooks, W. L. A.; Sumerlin, B. S. Chem. Rev. 2016, 116, 1375.

(50) (51)

Page 12 of 12

Bosch, L. I.; Fyles, T. M.; James, T. D. Tetrahedron 2004, 60, 11175. Springsteen, G.; Wang, B. Tetrahedron 2002, 58, 5291.

Insert Table of Contents artwork here

ACS Paragon Plus Environment