Glycoconjugates Based on Supramolecular Tröger's Base Scaffold: Synthesis, Photophysics, Docking, and BSA Association Study.

This study presents new Tröger's bases bearing glycosyl moieties obtained from a copper-catalyzed azide-alkyne cycloaddition reaction. The Tröger's bases present absorption maxima close to 275 nm related to fully spin and symmetry-allowed electronic transitions. The main fluorescence emission located at 350 nm was observed with no influence on the glycosyl moieties. Furthermore, protein detection studies have been performed using bovine serum albumin (BSA) as a model protein, and results have shown a strong interaction between some of the compounds through a static fluorescence suppression mechanism related to the formation of a glycoconjugate-BSA complex favored by the glycosyl subunit. Moreover, docking was also studied for better understanding the suppression mechanism and indicated that the glycosyl and triazole moieties increase the affinity with BSA.

Introduction

Tröger’s bases are C2-symmetric chiral tetracyclic molecules, bearing a methano-1,5-diazocine ring system, disposed in a unique V-shaped twisted structure. In this conformation, the two aryl rings are almost at a 90o dihedral angle, as presented in Figure for 2,8-bis(methyl)-6H,12H-(5,11)-methanodibenzo[b,f][1,5]diazocine (1).

Figure 1

Structure of the Tröger’s base 1.

These molecules have been known for more than a century, since the first report from Tröger.[1] Thereafter, a number of applications in supramolecular chemistry have been proposed. In this sense, the pioneering work from Wilcox[2] reported the use of Tröger’s bases in molecular recognition studies as a torsional molecular balance to quantify weak interactions such as head–tail and CH−π, which have important implications for the folding of proteins.[3] Since these initial reports, different designs have been presented.[4−11] For example, these molecules have been described as synthetic receptors for carboxylic acids,[12] terpenes,[13] adenine and biotin,[14] monovalent cations,[15] optical sensing,[16−18] and dye-sensitized solar cell (DSSC) applications.[19] Similarly, polyaromatic arylated Tröger’s bases[20] and derivatives bearing heterocyclic ring systems such as phenanthroline,[21] naphthalimide,[22] and acridine[23] derivatives have been studied in DNA interacalation. Additional applications include chiral solvating agents,[24] catalysts,[25] ligands in metal complexes,[26] new materials,[27] and compounds with biological activity.[28] On the other hand, Tröger’s bases bearing carbohydrates in their structure have not received much attention in the literature, despite the interesting properties that might arise from the introduction of a sugar moiety into the supramolecular structure.[29,30] Carbohydrates are naturally available and therefore usually abundant and inexpensive.[31] The polyoxygenated scaffold of carbohydrates is a valuable source of functionality that has been used to introduce a complexation or recognition site in a supramolecular structure for diverse applications. In this sense, different chemosensors for the detection of metals such as nickel(II),[32,33] copper(II),[34,35] zinc(II),[36] silver(I),[37] and protein-responsive electrochemical sensors[38] are found in the literature.

In this way, herein are described new Tröger’s base scaffolds presenting different glycosyl moieties linked by the 1,2,3-triazole core through copper-catalyzed azide–alkyne cycloaddition (Scheme ).[39] Additionally, their photophysical properties were also studied using UV–vis and fluorescence spectroscopy. Furthermore, fluorescence titration and docking have been performed using bovine serum albumin (BSA) for association studies.

Scheme 1

Glycoconjugates Prepared in This Work from Tröger’s Base Scaffold

Results and Discussion

The first step of this investigation was the functionalization of the Tröger’s base scaffold as presented in Scheme .[40] All details can be found in the Supporting Information. The 2,8-bis-iodo derivative 2, obtained by the reaction of 4-iodoaniline with paraformaldehyde under acidic conditions, was reacted with trimethylsilylacetylene in a Pd-catalyzed Sonogashira coupling to prepare the required alkyne needed for the copper-catalyzed azide–alkyne cycloaddition (CuAAC) reaction. The corresponding 2,8-bis-alkyne 3 was isolated in 95% yield. In a second step, the copper-catalyzed azide–alkyne cycloaddition using azides derived from d-xylose, d-galactose, and d-mannose was performed.[41] It was found that the efficiency of the CuAAC reaction was particularly dependent on the solvent probably due to solubility issues. Glycoconjugate 4a was obtained by reaction of 3 with a d-xylose-derived azide, resulting in 66% yield when the reaction was performed in a 2:1 mixture of t-butanol/water as the solvent. For the d-galactose-derived azide, the best result was achieved in a binary mixture of dichloromethane/water, and the corresponding product 4b was isolated in 61% yield. On the other hand, when the azide derived from d-mannose was used, the best result was obtained in dichloromethane/water to produce the derivative 4c in 51% yield and the monofunctionalized derivative (16% yield), which were readily separated by column chromatography.[40]

Scheme 2

Functionalization of the Tröger’s Base Scaffold

(i) PdCl2(PPh3)2 (10 mol %), (ii) Et3N, trimethylsilylacetylene, 90 °C, 4 h, (iii) CuSO4 (20 mol %), sodium ascorbate, (iv) TBAF, t-BuOH/H2O (2:1), rt, 48 h (for 4a), TBAF, dichloromethane/H2O (1:1), rt, 29 h (for 4b) and TBAF, dichloromethane/H2O (1:1), rt, 29 h (for 4c).

In order to improve the synthesis of 4c, it was decided to separately perform the deprotection reaction and the CuAAC cycloaddition. Thus, instead of adding TBAF to the reaction mixture in the CuAAC, we first removed both TMS groups from 3 using potassium carbonate in methanol/THF to afford the bis-terminal alkyne 5 in 92% yield. This compound was then subjected to the CuAAC reaction, and the desired glycoproduct 4c was isolated in 75% yield (Scheme ).[40]

Scheme 3

Two-step Synthesis of 4c

(i) K2CO3, methanol/THF (1:1), (ii) CuSO4 (20 mol %), sodium ascorbate, dichloromethane/H2O (1:1).

The photophysical investigation of the synthesized Tröger’s bases was performed by UV–vis absorption and steady-state and time-resolved fluorescence emission spectroscopies, respectively. The organic solvents with different dielectric constants (dichloromethane, 1,4-dioxane, ethyl acetate, ethanol, and acetonitrile) were used at a concentration of 10–5 M. The absorption spectra of 4a–c are presented in Figure . The experimental data from UV–vis absorption spectroscopy are presented in Table .

Figure 2

Normalized UV–vis absorption spectra in solution of the Tröger’s bases (a) 4a, (b) 4b, and (c) 4c at a concentration of ∼10–5 M.

Table 1

Photophysical Data of the UV–vis Spectra of the Tröger’s Bases 4a–ca

Tröger’s base	solvent	conc.	λabs	ε	fe	ke0	τ0	λem	ΔλST	ΦFL
4a	1,4-dioxane	2.56	276	2.65	0.95	1.25	0.80	350	74/7660	0.21
	ethyl acetate	5.03	275	1.51	0.91	1.20	0.84	346	71/7462	0.52
	dichloromethane	3.14	276	2.38	0.90	1.19	0.84	358	82/8299	0.21
	ethanol	3.14	268	1.89	0.76	1.06	0.95	360	92/9536	0.36
	acetonitrile	3.43	275	1.73	0.72	0.95	1.05	362	87/8739	0.16
4b	1,4-dioxane	1.64	276	2.48	0.98	1.29	0.78	347	71/7413	0.12
	ethyl acetate	2.10	276	2.63	0.93	1.22	0.82	346	70/7330	0.24
	dichloromethane	3.47	276	2.83	0.92	1.21	0.83	352	76/7823	0.14
	ethanol	3.56	268	2.43	0.96	1.34	0.75	359	91/9458	0.22
	acetonitrile	3.01	274	2.18	0.94	1.25	0.80	357	83/8485	0.16
4c	1,4-dioxane	2.39	276	2.43	0.99	1.30	0.77	350	74/7660	0.16
	ethyl acetate	2.34	273	3.45	0.92	1.23	0.81	348	75/7894	0.23
	dichloromethane	4.37	276	2.07	0.89	1.16	0.86	355	79/8063	0.14
	ethanol	2.36	268	3.37	0.96	1.33	0.75	357	91/9583	0.19
	acetonitrile	2.71	274	3.04	0.92	1.22	0.82	358	83/8431	0.16

The concentration is presented in 10–5 M; λabs and λem are the absorption and emission maxima, respectively (nm); ε is the molar absorptivity (104 M–1·cm–1), fe is the calculated oscillator strength; ke0 is the calculated radiative rate constant (109 s–1); τ0 is the calculated pure radiative lifetime (ns); ΔλST is the Stokes shift (nm/cm–1); and ΦFL is the relative fluorescence quantum yield.

The Tröger’s bases 4a–c present absorption maxima located around 275 nm with an almost absent solvatochromic effect, as already observed in the literature.[42] In addition, the maxima location also indicates that the triazole moiety does not present effective conjugation with the aromatic rings present in the Tröger’s bases, since it is found that values are very close to the Tröger’s bases framework without any substituent.[29] A particular behavior in ethanol was observed, where the absorption maxima blueshift is ∼10 nm due to specific solute-solvent interactions afforded by this polar protic media.

The UV–vis absorption results allowed obtaining ground-state parameters, such as the rate constant for emission (ke0), the oscillator strength (fe), and the pure radiative lifetime τ0 applying eqs – 3.[43,44] The respective values are summarized in Table .In these equations, ε (M–1·cm–1) is the extinction coefficient and ∫εdv̅ is the area under the absorption curve from a plot of ε (M–1·cm–1) versus the wavenumber v̅ (cm–1), which corresponds to a single electron oscillator. The rate constant (ke0) for emission can be related to ε using eq , where the absorption maxima v̅02 is presented in cm–1. In addition, the calculated pure radiative lifetime τ0 can be obtained from the rate constant from eq . The molar absorptivity coefficient ε values (104 L·mol–1·cm–1) and the calculated radiative rate constant (ke0) (109 s–1) for all Tröger’s bases indicate spin and symmetry-allowed electronic transitions, which could be related to 1π → π*. Moreover, the radiative lifetime show quite similar values (τ0 ≈ 1 ns), suggesting that after excitation, these structures populate the same excited state, as already observed for different Tröger’s bases.[29]

The normalized fluorescence emission spectra of Tröger’s bases 4a–c are presented in Figure , obtained by exciting the compounds at the absorption maxima. The relevant data from this characterization are also presented in Table .

Figure 3

Normalized steady-state fluorescence emission spectra in solution of the Tröger’s bases (a) 4a, (b) 4b, and (c) 4c at a concentration of ∼10–5 M.

The glycoconjugates show a main fluorescence emission located around 350 nm, indicating that the glycosyl moieties do not present any role on their excited-state photophysics. As already observed in the ground state, a small positive solvatochromic effect was observed from 1,4-dioxane to acetonitrile, showing that their dipole moments are larger in the excited state than in the ground state (i.e., μe → μg). Concerning the fluorescence quantum yields, surprisingly higher values could be obtained for the glycoconjugates 4a–c (ΦFL ≈ 0.2) if compared to the values from the simpler Tröger’s base framework (ΦFL ≈ 0.06).[29] This result indicates that the glycosyl moieties are not efficient excited-state deactivation channels.

The excited-state dynamics of the Tröger’s bases was investigated by time-resolved fluorescence spectroscopy. The respective curves are depicted in Figure . The residuals are presented in the Supporting Information, and the relevant data are summarized in Table . Moreover, the radiative and nonradiative decay rate constants, kf and knr, respectively, were also estimated from eqs and 5.[45,46]andwhere τf is the experimental fluorescence lifetime and ΦFL is the fluorescence quantum yield. The fluorescence decay profiles indicated a monoexponential fluorescence lifetime with values between 1.4 and 2.0 ns with good χ2 values. The solvent seems do not significantly affect the fluorescence lifetime of these compounds. In acetonitrile and 1,4-dioxane, it could be observed that the Tröger’s base 4c has a decrease in the fluorescence rate constant (kf) attributable to more effective nonradiative processes due to this moiety in the Tröger’s base scaffold. The nonradiative process showed to be higher for all compounds in the studied solvents, as expected.

Figure 4

Fluorescence decay curves of Tröger’s bases 4a–c in (a) 1,4-dioxane and (b) acetonitrile (ca. 10–5 M). IRF = instrument response factor.

Table 2

Time-Resolved Fluorescence Data of the Tröger’s Bases 4a–ca

Tröger’s base	solvent	τ	A	χ2	kf	knr
4a	1,4-dioxane	1.914	1.089	1.190	10.97	4.13
	acetonitrile	1.937	1.070	1.162	8.26	4.34
4b	1,4-dioxane	1.431	1.251	1.137	8.39	6.15
	acetonitrile	1.471	1.248	1.111	10.88	5.71
4c	1,4-dioxane	1.827	0.780	1.083	8.76	4.60
	acetonitrile	1.954	1.056	1.178	8.19	4.30

τ is the experimental fluorescence lifetime (in ns), A is the pre-exponential factor, χ2 indicates the quality of the exponential fit, kf (× 107) is the calculated radiative decay constant (in s–1), and knf (× 108) is the calculated nonradiative decay constant (in s–1).

BSA presents well-known photophysics characterized by an absorption located around 280 nm and a fluorescence emission around 340 nm related to the tryptophan residues.[47] In this way, to investigate the interaction between the Tröger’s bases and BSA, its fluorescence quenching in the presence of 4a–c was studied. It is understood that this quenching can be associated to different methods, such as molecular rearrangements, energy transfer, excited-state reactions, ground-state complex formation, and collisional deactivation.[48] In this way, the variation of fluorescence intensity of BSA in the presence of the compounds was measured using 280 nm as the excitation wavelength. Figure depicts the suppression experiment using compound 4b as the BSA quencher. It is worth mentioning that the interaction studies with BSA were also performed for comparison with the Tröger’s base 1 in order to evaluate the glycosyl moitey role on the interaction between the Tröger’s bases and BSA.

Figure 5

Fluorescence emission spectra (λexc = 280 nm) of BSA (11 μM in PBS, pH 7.2) in the presence of different amounts of (a) the Tröger’s base 1 and (b) 4b. The blank sample (0 μM) is BSA in the absence of the Tröger’s base. The inset shows the fluorescence intensity versus Tröger’s base concentration relation.

It could be observed that as the concentration of the Tröger’s bases increases, the fluorescence intensity of BSA decreases, which already qualitatively indicates a significant interaction of these compounds with BSA. However, it can also be observed that the methylated Tröger’s base 1 (Figure a) does not present a linear relationship (R2 = 0.941) over the entire range of studied concentration. On the other hand, compound 4b shows a linear relation between the fluorophore concentration and the BSA fluorescence intensity (R2 = 0.989). In addition, the Tröger’s base 1 (Figure a) presented a lower reduction in the fluorescence intensity of BSA (14%) than the glycoconjugate 4b (67%). We would like to highlight that the additional glycoconjugates 4a and 4c also presented linear correlations with similar values in the fluorescence suppression of BSA (67 to 71%) (data not shown; see the Supporting Information), indicating that the carbohydrate moitey plays an important role in the interaction with BSA.

In order to better investigate the suppression mechamism between the Tröger’s bases and BSA, additional experiments were performed at different temperatures (25, 30, 35, and 40 °C) by applying the Stern–Volmer relation presented in eq :where F0 and F are the fluorescence intensities of BSA in the absence and presence of certain amounts of the Tröger’s bases, respectively; Kq and KSV are the bimolecular suppression and Stern–Volmer constants, respectively; and [Q] is the quencher concentration. In this equation, Kq is related to the suppression efficiency, and τ0 is the fluorophore lifetime in the absence of the suppressor, with values around 10–8 s to biomolecules.[49] The relevant data are summarized in Table .

Table 3

Results from the Stern–Volmer Relation Form the Tröger’s Bases at Different Temperaturesa

compound	T (°C)	linear fit	R2	KSV	Kq
1b	25	F0/F = 0.96 + 4.05 × 103[Q]	0.988	4.05	4.05
	30	F0/F = 0.98 + 4.03 × 103[Q]	0.934	4.03	4.03
	35	F0/F = 0.91 + 4.43 × 103[Q]	0.983	4.43	4.43
	40	F0/F = 0.94 + 3.72 × 103[Q]	0.946	3.72	3.72
4a	25	F0/F = 0.60 + 5.25 × 104[Q]	0.987	52.5	52.5
	30	F0/F = 0.64 + 5.24 × 104[Q]	0.990	52.4	52.4
	35	F0/F = 0.60 + 5.05 × 104[Q]	0.985	50.5	50.4
	40	F0/F = 0.58 + 5.21 × 104[Q]	0.985	52.1	52.1
4b	25	F0/F = 0.66 + 4.55 × 104[Q]	0.991	45.5	45.5
	30	F0/F = 0.67 + 4.46 × 104[Q]	0.993	44.6	44.6
	35	F0/F = 0.67 + 4.14 × 104[Q]	0.989	41.4	41.4
	40	F0/F = 0.60 + 4.35 × 104[Q]	0.989	43.5	43.5
4c	25	F0/F = 0.84 + 3.50 × 104[Q]	0.931	35.0	35.0
	30	F0/F = 0.68 + 4.32 × 104[Q]	0.933	43.2	43.2
	35	F0/F = 0.72 + 3.63 × 104[Q]	0.941	36.3	36.3
	40	F0/F = 0.72 + 3.74 × 104[Q]	0.940	37.4	37.4

Kq (× 1011 L·mol–1·s–1) is the bimolecular suppression constant, and KSV (× 103 L·mol–1) is the suppression constant.

2,8-Bis(methyl)-6H,12H-(5,11)-methanodibenzo[b,f][1,5]diazocine (Figure ).

It can be observed that the constants KSV and Kq independent on the temperature present values higher than 1011 L·mol–1·s–1, which exceed the maximum value of 2.0 × 1010 L·mol–1·s–1 for the diffusion-controlled mechanism (dynamic quenching). Thus, it can be considered in this study that the suppression can be related to a static mechanism, where the formation of a glycoconjugate–BSA complex takes place in the ground state. Despite the observed mechanism for all studied compounds, the Tröger’s bases 1 presented lower values for the KSV and Kq in comparison with its glycoconjugates 4a–c, which can be related to a weaker interaction with BSA.

It is worth mentioning that UV–vis absorption spectroscopy can also be a powerful tool to discuss the nature of the interaction between BSA and the Tröger’s base (static or dynamic). In this sense, since the dynamic mechanism involves only the excited state, it is expected that no changes will take place in the absorption spectrum (ground state). On the other hand, in the static mechanism, changes in the absorption spectrum are expected due to the formation of a new species (suppressor–BSA complex). Thus, equimolar amounts of the suppressors were added to a solution of BSA and evaluated by UV–vis absorption spectroscopy.[50] In addition, in order to prove that the spectrum of BSA in the presence of the suppressor is not only the sum of the absorbances of BSA and the suppressor, the difference between the absorbance of the complex and the suppressor was also obtained (Figure ).

Figure 6

UV–vis absorption spectra of BSA in the absence and presence of the Tröger’s bases (a) 1, (b) 4a, (c) 4b, and (d) 4c.

It is observed that the BSA absorption bands in the presence and in the absence of the Tröger’s bases 4a–c differ from each other, which can be related to a change in the aromatic residues of BSA due to the formation of complexes BSA–Tröger’s bases, agreeing with the assumption of the static mechanism. It is worth mentioning that this experiment also indicates that the Tröger’s base 1, presented for comparison, shows close absorption spectra (intensity and maxima location), which can be related to a lower interaction with BSA, as already discussed in this study. Thus, considering the static mechanism, the binding constant (KA) and the number of binding sites (n) between BSA and suppressor can be calculated from eq :[51]

Table presents the results from the double logarithmic plot relating the fluorescence intensities from BSA and the quencher concentration. According to these results, it was observed that glycoconjugates 4a–c present a strong interaction with BSA due to the high values of binding constant KA higher than 105 L·mol–1. Further, the binding constants (KA) for Tröger’s 4a and 4b bases increase with thetemperature, and there are two binding sites with BSA. On the other hand, for glycoconjugate 4c, there is only one binding site with BSA, and there is no tendency for KA with the temperature. The observed differences between glycoconjugates such as the magnitude of KA, number of binding sites, and linearity allow us to conclude that the glycosyl portion plays a fundamental role on the interaction with the protein. For the Tröger’s base containing the methyl group, this analysis was not conclusive, since the results of KA and the number of binding sites present a great variation in a very random way.

Table 4

Results from Eq from Tröger’s Bases 1 and 4a–ca

compound	T (°C)	linear fit	KA (L·mol–1)	n
1	25	log(F0 – F/F) = 10.12 + 2.24 log[Q]	1.41 × 1010	2.24
	30	log(F0 – F/F) = 4.21 + 1.16 log[Q]	1.62 × 104	1.16
	35	log(F0 – F/F) = 1.60 + 0.59 log[Q]	3.95 × 101	0.59
	40	log(F0 – F/F) = 7.92 + 2.01 log[Q]	8.37 × 107	2.01
4a	25	log(F0 – F/F) = 7.31 + 1.61 log[Q]	2.02 × 107	1.61
	30	log(F0 – F/F) = 7.65 + 1.68 log[Q]	4.45 × 107	1.68
	35	log(F0 – F/F) = 7.92 + 1.75 log[Q]	8.36 × 107	1.75
	40	log(F0 – F/F) = 10.13 + 2.24 log[Q]	1.30 × 1010	2.24
4b	25	log(F0 – F/F) = 7.40 + 1.64 log[Q]	2.52 × 107	1.64
	30	log(F0 – F/F) = 8.36 + 1.86 log[Q]	2.27 × 108	1.86
	35	log(F0 – F/F) = 8.98 + 2.00 log[Q]	9.58 × 108	2.00
	40	log(F0 – F/F) = 9.75 + 2.18 log[Q]	5.64 × 109	2.18
4c	25	log(F0 – F/F) = 5.34 + 1.19 log[Q]	2.20 × 105	1.19
	30	log(F0 – F/F) = 5.74 + 1.28 log[Q]	5.55 × 105	1.28
	35	log(F0 – F/F) = 5.64 + 1.28 log[Q]	4.36 × 105	1.28
	40	log(F0 – F/F) = 5.49 + 1.25 log[Q]	3.11 × 105	1.25

KA (L·mol–1) is the binding constant, and n is the binding sites.

In order to investigate the interactions between the Tröger’s base derivatives 4a–c and 1 with BSA, docking simulations were performed. It could be observed from the predicted binding energy (ΔGbind) summarized in Table that glycoconjugates 4a, 4b, and 4c present a more spontaneous interaction with BSA than the Tröger’s base model 1. These results are correlated with the experimental data, which indicates that the glycoconjugates present the major reduction in the BSA fluorescence. In relation to the binding pose, the molecular docking show that 4a and 4b interact in the IB subdomain of BSA closely to Trp134 (∼10 Å) (Figure a,b), while compounds 4c and 1 binds in the II A region, interacting with Trp213 (3.4–6.4 Å) (Figure c,d). In general, the compounds demonstrated H-bonds, hydrophobic, and electrostatic interactions with the BSA protein. It is worth mentioning that the BSA domains I A (1–112), I B (113–195), II A (196–303), II B (304–383), III A (384–500), and III B (501–583) presented in Figure were based on previous studies.[52,53]

Table 5

Amino Acids Residues Involved in the Interactions between BSA and the Tröger’s Base Derivatives 1 and 4a–c and the Predicted Binding Energy Obtained from the Docking

	Interactions
Tröger’s base	hydrophobic	electrostatic	H-bonds	ΔGbind (kcal·mol–1)	Kda
1	Leu237, Leu259, Ile289, and Ala290	Arg198 and Arg256		–8.1	1.16 × 10–6
4a	Leu115, Lys116, Tyr160, and Arg185	Arg185	Lys136, Lys114, and Arg185	–11.8	2.24 × 10–9
4b	Leu115, Pro117, Leu122, and His145		Arg144	–13.3	1.78 × 10–10
4c	Tyr149, Lys187, and Pro440	Arg194, Arg217, and Asp450	Arg198 and Ser442	–11.6	3.14 × 10–9

Kd refers to the dissociation constant between BSA and ligands obtained from Kd = exp[(ΔG × 1000)/(R × T))], where R is 1.98719 cal and T is 298.15 K.

Figure 7

Results from molecular docking between BSA and the Tröger’s base derivatives (A) 4a, (B) 4b, (C) 4c, and (D) 1. The hydrogen bonds, hydrophobic (π–alkyl, alkyl–alkyl, and π–π stacking), and electrostatic (π–anion and π–cation) interactions are shown in green, purple, and orange dashed lines, respectively. Red dashed lines show the distances (Å) between the Tröger’s bases and tryptophan (Trp) residues.

It could be observed that the H-bonds presented by the glycoconjugates 4a–c are essential for the better binding affinity to BSA. Likewise, the Tröger’s base 1 that does not show H-bonds presented low BSA fluorescence quenching. In other words, the Tröger’s base functionalization with glycosyl and triazole moieties increases the affinity with BSA. In fact, the H-bonds have an important role in the complex stabilization, as already reported in the literature.[54,55] The interaction of the Tröger’s base in I B[52,56] and II A[51,57,58] subdomains could induce conformational changes in BSA. As a result, the tryptophan fluorescence quenching can be affected.

Considering that the AutoDock Vina does not calculate the Ki values, in this study, it was performed the Ki prediction based on the ΔG from docking using the relation Ki = exp((ΔG × 1000)/(R × T)), where R is 1.98719 cal and T is 298.15 K, according to previous studies.[59,60] We would like to highlight that, in this investigation, it was decided to use the term Kd (dissociation constant) rather than Ki (inhibition constant) due to the protein nature of BSA despite the enzymatic one, where the term inhibition is better used. As shown in Table , the predicted Kd values are in accordance with the KA data, where compound 4b presents the preferential binding for BSA, while compound 1 shows the lowest one.

Conclusions

It was described, in this investigation, the modification of the Tröger’s base scaffold with carbohydrates. Both moieties have been linked together efficiently by a 1,2,3-triazole through copper-catalyzed azide–alkyne cycloaddition of a Tröger’s base-alkyne and carbohydrate-derived azides. The Tröger’s bases presented absorption maxima located in the UV region with an almost absent solvatochromic effect. The Tröger’s bases presented a main emission located around 350 nm. The unexpected higher fluorescence quantum yields if compared to the values from the Tröger’s base framework indicated that the glycosyl moieties are not efficient excited-state deactivation channels. The time-resolved fluorescence spectroscopy showed a monoexponential time decay with lifetime values between 1.4 and 2.0 ns with good χ2 values. BSA fluorescence suppression was observed in glycoconjugates 4a–c mainly due to the carbohydrate moiety, which presented a higher Stern–Volmer constant despite the simpler Tröger’s base 1.

Experimental Section

General Information

Air-sensitive reactions were carried out in oven-dried glassware equipped with tightly fitted rubber septa. In these reactions, a positive pressure of dry argon was applied. Thin-layer chromatography (TLC) was performed using Silica Gel 60 F254. Column chromatography was performed using Silica Gel 60 Å (70–230 mesh). Carbohydrate-derived azides were synthesized according to the literature.[61,62] NMR spectra were recorded in CDCl3 solution on a Varian VNMRS 300 or 400 MHz or a Bruker 400 MHz spectrometer. Assignment of chemical shifts is based on standard NMR experiments and reported in parts per million (ppm) related to tetramethylsilane (δ = 0.00 ppm in 1H NMR) or from the solvent peak of CDCl3 (δ = 77.23 ppm in 13C NMR). The data are presented as follows: chemical shift (δ), multiplicity, coupling constant (J) in Hz, and the integrated intensity. 13C NMR spectra were recorded at 75 or 100 MHz in CDCl3 solution. The multiplicities are given as s (singlet), d (doublet), t (triplet), dd (double doublet), m (multiplet), q (quartet), and br (broad singlet). High-resolution mass spectrometry with electrospray ionization (HRMS-ESI) measurements was performed on the positive mode. FTIR spectra were acquired with a Bruker spectrometer (Alpha) using an attenuated total reflectance (FTIR-ATR) device. Melting points were obtained on a Buchi M-565 and are uncorrected. In the photophysical studies, spectroscopic-grade solvents were used. All measurements were performed at 25 °C. The UV–vis absorption spectra were recorded on a Shimadzu UV-2450 spectrophotometer. The steady-state fluorescence spectra were recorded on a Shimadzu spectrofluorometer model RF-5301PC. Time-resolved fluorescence spectroscopy was performed using an EasyLife V spectrophotometer from Optical Building Blocks (OBB). The decay curves were analyzed using the software EasyLife V (OBB). Curve fitting was performed by a nonlinear least-squares method. The quality of the fit was evaluated by the χ2 values, the respective residuals, and the autocorrelation function. The quantum yield of the fluorescence (ΦFL) was calculated by applying the dilute optical method using spectroscopic-grade cyclohexane. Naphthalene (ΦFL = 0.23) was used as the quantum yield standard.[63] The Tröger’s base 2,8-bis(methyl)-6H,12H-(5,11)-methanodibenzo[b,f][1,5]diazocine presented in Figure and used in this study was prepared as described in the literature.[40]

BSA Interaction Study

Due to the low solubility in water, the Tröger’s bases were first solubilized in dimethylformamide (DMF) (∼2 × 10–4 M) to prepare a stock solution in phosphate buffer solution (PBS) (25 mL) to a final dye concentration of 14 μM. The BSA suppression study was performed using a constant BSA concentration (11 μM in phosphate buffer solution, pH 7.2). In this study, different amounts of the dye solutions (5–50 μM in PBS) were added. The fluorescence emission spectra were obtained at 25 °C and under an excitation wavelength located at 280 nm. For the experiments at different temperatures, the samples were kept in a water bath, with a thermometer-controlled temperature of one decimal place.

Synthesis

2,8-Diiodo-6H,12H-5,11-methanodibenzo[b,f][1,5]diazocine (2)

In an open round-bottom flask containing 10 mL of TFA at −15 °C, p-iodoaniline (3 mmol) and p-formaldehyde (4.5 mmol) were added. After the addition, the mixture was stirred for 24 h at room temperature. After this time, the reaction mixture was poured into cold water, and ammonium hydroxide was added until pH = 8 was reached. The crude product was filtered, dried, and purified by column chromatography using dichloromethane as the eluent, and the pure product was obtained as a yellow solid in 37% yield (0.260 g). mp 177–180 °C. 1H NMR (300 MHz, CDCl3) δ (ppm): 7.45 (dd, J = 8.5, 2.0 Hz, 2H), 7.23 (d, J = 2.0 Hz, 2H), 6.87 (d, J = 8.5 Hz, 2H), 4.61 (d, J = 16.7 Hz, 2H), 4.23 (s, 2H), 4.08 (d, J = 16.7 Hz, 2H). 13C NMR (75.5 MHz, CDCl3) δ (ppm): 147.7, 136.6, 135.9, 130.3, 127.2, 87.8, 66.7, 58.3.

2,8-Bis[2-(trimethylsilyl)ethynyl]-6H,12H-5,11-methanodibenzo[b,f][1,5]diazocine (3)

In a dry tube, iodinated Tröger’s base 1 (1 mmol), CuI (8 mol %), PPh3 (9 mol %), PdCl2(PPh3)2 (10 mol %), Et3N (4 mL), and ethynyltrimethylsilane (4 mmol) were added under a N2 atmosphere. The tube was sealed and stirred at 90 °C for 4 h. Finally, the reaction was cooled at room temperature, ethyl acetate was added, and the mixture was filtered under silica. The solvent was removed, and product 2 was obtained as a yellow solid in 95% yield (0.475 g). mp 194–196 °C. 1H NMR (300 MHz, CDCl3) δ (ppm): 7.24 (dd, J = 8.2, 2.0 Hz, 2H), 7.03 (d, J = 2.0 Hz, 2H), 7.02 (d, J = 8.2 Hz, 2H), 4.61 (d, J = 16.7 Hz, 2H), 4.26 (s, 2H), 4.10 (d, J = 16.7 Hz, 2H), 0.21 (s, 18H). 13C NMR (75.5 MHz, CDCl3) δ (ppm): 148.3, 131.0, 130.8, 127.6, 124.8, 118.6, 104.7, 93.4, 66.8, 58.5.

General Procedure for the CuAAC

In an open flask were added the Tröger’s base 3 (1.0 equiv), the appropriate carbohydrate-derived azide (2.0 equiv), CuSO4·5H2O (0.2 equiv), sodium ascorbate (0.4 equiv), and the appropriate solvent (as indicated in Scheme ). After stirring for 5 min, TBAF (3 equiv, 1 M solution in THF) was added, and the reaction mixture was stirred for the time indicated in Scheme . The crude product was extracted with dichloromethane. The organic layer washed with 0.1 M EDTA solution, dried with Na2SO4, and filtered, and the solvent was removed under vacuum. The crude product was purified by column chromatography eluting with a gradient of dichloromethane and acetone.

Glycoconjugate 4a

Pale yellow solid, yield: 0.100 g, 66%. mp 152–155 °C. FTIR (KBr, cm–1): 3434, 3132, 2987, 2937, 2898, 1492, 1217, 1072. 1H NMR (300 MHz, DMSO-d6) δ (ppm): 8.42 (s, 2H), 7.62 (dd, J = 8.2 Hz and 1.5 Hz, 2H), 7.46 (d, J = 1.5 Hz, 2H), 7.21 (d, J = 8.2 Hz, 2H), 5.88 (d, J = 3.8 Hz, 2H), 5.67 (d, J = 5.0 Hz, 2H), 4.71 (d, J = 17.0 Hz, 2H), 4.47 (d, J = 3.8 Hz, 2H), 4.45–4.38 (m, 4H), 4.30 (s, 2H), 4.23 (d, J = 17.0 Hz, 2H), 4.08 (dd, J = 5.0 Hz and 2.6 Hz, 4H), 1.33 (s, 6H), 1.22 (s, 6H). 13C NMR (75.5 MHz, DMSO-d6) δ (ppm): 147.8, 146.2, 146.1, 128.5, 125.9, 125.3, 124.1, 123.6, 123.5, 121.4, 110.9, 104.5, 85.0, 79.1, 73.5, 66.2, 58.2, 49.0, 26.6, 26.0. HRMS (ESI-MS) calcd for C35H41N8O8 [M + H]+ 701.3047, found 701.3013.

Glycoconjugate 4b

Pale yellow solid, yield: 0.180 g, 61%. mp 275–280 °C. FTIR (KBr, cm–1): 3143, 2987, 2937, 2906, 1383, 1213, 1066. 1H NMR (400 MHz, CDCl3) δ (ppm): 7.83 (s, 1H), 7.82 (s, 1H), 7.56 (d, J = 8.2 Hz, 2H), 7.43 (s, 2H), 7.17 (d, J = 8.2 Hz, 2H), 5.49 (d, J = 5.1 Hz, 2H), 4.73 (d, J = 16.8 Hz, 2H), 4.65–4.58 (m, 4H), 4.42 (dd, J = 13.7 Hz and 8.2 Hz, 2H) 4.34 (s, 2H), 4.30 (dd, J = 5.0 Hz and 2,4 Hz, 2H), 4.24 (d, J = 16.8 Hz, 2H), 4.17 (d, J = 7.8 Hz, 4H), 1.47 (s, 6H), 1.34 (s, 6H), 1.33 (s, 6H), 1.25 (s, 6H). 13C NMR (100.5 MHz, CDCl3) δ (ppm): 147.8, 147.1, 128.2, 128.2, 126.8, 126.7, 125.3, 125.0, 124.2, 120.5, 120.4, 109.8, 109.0, 96.2, 71.2, 70.7, 70.3, 67.0, 58.7, 50.5, 26.0, 25.9, 24.8, 24.4. HRMS (ESI-MS) calcd for C43H53N8O10 [M + H]+ 841.3885, found 841.3852.

Glycoconjugate 4c

Pale yellow solid, yield: 51%. mp 204–208 °C. FTIR (KBr, cm–1): 3168, 2985, 2941, 2898, 1375, 1209, 1070. 1H NMR (400 MHz, CDCl3) δ (ppm): 7.95 (s, 1H), 7.95 (s, 1H), 7.58 (dd, J = 8.8 Hz and 1.6 Hz, 2H), 7.55 (dd, J = 8.8 Hz and 1.6 Hz, 2H), 7.42 (d, J = 1.6 Hz, 1H), 7.40 (d, J = 1.6 Hz, 1H), 7.18 (d, J = 8.2 Hz, 1H), 7.16 (d, J = 8.2 Hz, 1H), 6.06 (d, J = 3.5 Hz, 2H), 4.91 (dd, J = 5.8 Hz and 3.5 Hz, 2H), 4.84 (dd, J = 5.2 Hz and 3.5 Hz, 2H), 4.73 (d, J = 16.8 Hz, 2H), 4.47 (ddd, J = 7.6 Hz, 6.6 Hz and 4.7 Hz, 2H), 4.34 (s, 2H), 4.24 (d, J = 16.8 Hz, 2H), 4.09 (dd, J = 8.9 Hz and 6.3 Hz, 2H), 4.03 (dd, J = 8.9 Hz and 4.3 Hz, 2H), 3.75 (dd, J = 7.6 Hz and 3.5 Hz, 2H), 1.54 (s, 3H), 1.52 (s, 3H), 1.42 (s, 6H), 1.36 (s, 6H), 1.29 (s, 3H), 1.28 (s, 3H). 13C NMR (100.5 MHz, CDCl3) δ (ppm): 148.1, 147.9, 128.2, 126.4, 126.3, 125.4, 125.3, 125.0, 124.3, 119.9, 113.7, 113.6, 109.5, 88.8, 79.6, 79.2, 72.6, 66.9, 66.7, 58.8, 58.7. 26.9, 25.6, 25.5, 25.1, 24.0. HRMS (ESI-MS) calcd for C43H53N8O10 [M + H]+ 841.3885, found 841.3820.

Docking

The structure of bovine serum albumin (BSA) was obtained from the Protein Data Bank.[64,65] The addition of chain B, water, and other small molecules, as well as hydrogen atoms was performed by the Chimera 1.8 software.[66] The software Avogadro 1.1.1 was used to build the chemical structure of the Tröger’s base derivatives[67] following the semiempirical PM6 geometry optimization (Program MOPAC2012).[68,69] Here, as a model of interaction, only the R isomers were tested. The Tröger’s bases and BSA were generated in the pdbqt format by AutoDockTools. In this methodology, the Tröger’s bases were considered flexible (with PM6 charges), and BSA was considered rigid (with Gasteiger charges).[70] The blind docking was performed by the AutoDock Vina 1.1.1 software[71] with a gridbox of 96 × 64 × 88 and the coordinates x = 9.46, y = 23.36, and z = 98.15 (exhaustiveness of 150). For the binding pose, the respective conformation with the lowest binding free energy (ΔG) was selected. The results from docking were analyzed using the Discovery Studio Visualizer 17.2.0 (DSV) software.[72]