The full spectrum tuning of fluorescent molecules via a one-pot multicomponent reaction

Abstract

Fluorogenic probes for imaging enable visualization and analysis of difficult-to-reach cells and organelles. However, there are limited efficient examples of tuning these fluorescent molecules to higher wavelengths. This is vital since different tissues are sensitive to varying wavelength emissions. To address this need, we report the discovery, tuning, structure-photophysical property relationships (SPPR), and time-dependent DFT (TD-DFT) computations of 400–700+ nm fluorescent pyrido[2’,1’:2,3]imidazo[4,5-c]isoquinolines and substituted imidazo[1,2-a]pyridin-3-amines. The syntheses involve the trimethylsilylcyanide (TMSCN) modified Groebke-Blackburn-Bienaymé (GBB) multicomponent reaction as well as the TMSCN modified GBB combined with subsequent condensation of an aldehyde, and Aza-Friedel-Crafts-Intramolecular Cyclization-Oxidation all in one pot. The SPPR reveals that electron-withdrawing strength in the para-position of the aminopyridine starting material has direct control over the absorption and fluorescence emission wavelengths of these molecules. The TD-DFT computations show the changes in the natural transition orbitals (NTOs) with differing substitutions to the parent molecule that dictate the observed excitations, emissions, and fluorescence intensities. These findings give insights and directions for tuning the fluorescent properties of these motifs for various uses as probes and imaging agents.

Graphical Abstract

INTRODUCTION

Multicomponent reactions (MCRs) are commonly defined as chemical reactions that combine three or more reactants into a single product with every reactant providing a high proportion of its atoms to that product.¹ As such, MCR-based synthetic routes are often advantageous over conventional synthetic approaches because they inherently have high bond-forming efficiency, high atom-economy and deliver diverse scaffolds in one step. Within the multitude of MCRs lies the Groebke–Blackburn–Bienaymé (GBB) reaction first reported by three independent groups in 1998 (Scheme 1).^2,3 It is comprised of reaction of amidines 1, aldehydes 2 and isocyanides 3 facilitated by acid catalysis which affords aza-heterocycles 4.

Scheme 1.

Evolution of TMSCN MCR chemistry

Eight years after the publication of the original GBB reaction, this laboratory published the versatile TMSCN-modified version of the GBB reaction which negated the use of pungent isonitriles, affording products with an amino-handle primed for diversification.⁴ Nitrogen heterocycles derived from MCR methodology often exhibit fluorescence properties due to their high degrees of conjugation,⁵ a desirable property with high utility in diverse arenas such as sensing, clinical diagnostics and imaging.⁶ Of these, imaging is the most ubiquitous, relying on photoinduced electron transfer (PET) which involves fluorescence modulation.⁷ PET is critical as it enables highly sensitive fluorescent measurements where fluorometers are able to detect at 10⁻¹⁰ M.⁸ Whilst this technique is extremely powerful, there is often no facile route to tuning molecular fluorescence within a central chemotype, which is essential as living tissue is more transparent to higher wavelength radiation, and few molecules emit in that range.⁹ Despite the need, practical use, and efforts of many, the fluorescence properties of only a few fluorophores, such as BODIPY,⁹ Cyanine,¹⁰ and fluorescein¹¹, have been tuned over the entire visible light spectrum (400–700 nm+). The nitrogenous heterocycles presented herein are fluorescent across the full visible spectrum often with good quantum yields. Moreover, being derived from MCR methodology they can be prepared in one or two simple steps alleviating overall cost of goods.¹² As fused nitrogen-containing heterocycles make up 14% of all drugs,¹³ one would also envision utility in the pharmaceutical sector, exemplified by notable imidazopyridines that include Zolpidem (GABA_A receptor agonist 6).¹⁴ While little biological data has been gathered on pyrido[2’,1’:2,3]imidazo[4,5-c]isoquinolines (5) - a new two-step synthetic route is detailed herein - close analogs show antimitotic and cytotoxic effects respectively.^15,16

RESULTS AND DISCUSSION

Synthetic schemes describing the synthesis of pyrido[2’,1’:2,3]imidazo[4,5-c]isoquinolines 5 and imidazo[1,2-a]pyridin-3-amines 10 are predicated by previous work in this laboratory that employed TMSCN as a formal isonitrile replacement in the GBB reaction (conversion 7 – 10).¹⁷ Study of alternate reagent stoichiometry facilitated direct conversion to the imine 11 (Scheme 2).¹⁸ We subsequently employed a similar strategy (RNC to TMSCN switch) in a Knoevenagel/[4+1] cycloaddition-condensation-cyclization sequence to afford the indolizine tetracycles 14.¹⁹

Initial chemistry optimization toward pyrido-imidazo[4,5-c]isoquinolines 5 began with imine 11d, formed in one pot from the TMSCN-modified GBB reaction (Table 1). Modest increases in reaction time afforded better yields of 16d and increased catalyst loading further improved yields (Entries 1–7). Near quantitative yields were observed on switching from dichloromethane to dichloroethane which allowed for higher reaction temperatures (Entries 8–9).

Table 1.

Optimization of Aza-Friedel-Crafts-Intramolecular Cyclization-Oxidation.

Entrya	LA (eq.)	T (°C)	Solvent (M)	Time (min)	Yield (%)
1	Cu(OTf)2(0.5)	100	DCM(0.1)	5	20
2	Cu(OTf)2(0.5)	100	DCM(0.1)	10	28
3	Cu(OTf)2(0.5)	100	DCM(0.1)	20	28
4	Cu(OTf)2(1.0)	100	DCM(0.1)	5	68
5	Cu(OTf)2(1.0)	100	DCM(0.1)	10	83
6	Cu(OTf)2(1.0)	100	DCM(0.1)	20	91
7	Cu(OTf)2(2.0)	100	DCM(0.1)	20	93
8	Cu(OTf)2(1.0)	100	DCM(0.1)	20	99
9	Cu(OTf)2(2.0)	100	DCM(0.1)	20	99

^aScale: 1.0 mmol. Rxns performed under microwave irradiation.

We then explored feasibility of a one-pot procedure to 16d (Table 2). Early results with copper triflate (Cu(OTf)₂) failed to furnish the desired product 16d in different solvents (Entries 1–3). Interestingly, cyclized product 16d was observed upon reaction in acetic acid (Entry 4), a phenomenon likely related its ability to promote aza-Friedel-Crafts reactions.²⁰ Evaluating increased Bronsted acid strength, sulfuric acid (Entry 5–6) diminished the yields which we attributed to its ability to interfere with imine formation (11d). To circumvent this, sulfuric acid was added to the reaction post-imine formation, thus maintaining one-pot conditions and affording excellent yields (Entry 7–9).

Table 2.

One-Pot Optimization Toward Fluorescent Pyrido-imidazo[4,5-c]isoquinolines

Entrya	Acid (eq.)	Solvent (0.3 M)	Time (min)	Yield (%)
1	Cu(OTf)2(1.0)	MeOH	5	0
2	Cu(OTf)2(1.0)	DCMb	20	0
3	Cu(OTf)2(1.0)	ACN	20	0
4	-	AcOH	20	28
5	H2SO4(30 eq.)	AcOH	20	6
6	H2SO4(50 eq.)	MeOH	20	0
7	H2SO4(50 eq.)	MeOH	20, 5c	81
8	H2SO4(50 eq.)	MeOH	20, 10c	87
9	H2SO4(50 eq.)	MeOH	20, 20c	96

^aScale: 1.0 mmol

^b100 °C

^cAcid was added after 1^st indicated time and set to run for the 2^nd indicated time at 140 °C (microwave irradiation).

A set of 5 congeners (16a – e) and their corresponding intermediate amines (10a – 10e) were then prepared. During characterization, a color difference was observed between the two chemotypes (Fig. 2) with increasing wavelength and quantum yield favoring the TMSCN-modified GBB⁴ derived primary amines (10a-e). With a goal of tuning toward a high quantum yield red fluorophore, the free amines (10) were thus prioritized for further study.

Figure 2.

Fluorescent Perturbations

We next evaluated varying substituents on the aldehyde input 8 (Fig. 3) on amino-imidazopyridine fluorescence 10 (Scheme 1). The aryl 10f had a modest quantum yield of 0.30 and Em_max of 473nm and the introduction of a 4-methoxyl 10g increased quantum yield 30% to 0.39. However, this did not significantly affect fluorescence wavelength. Indeed, introduction of one or more methoxyl groups in the aryl ring consistently increased the quantum yield without affecting the fluorescence wavelength. The introduction of methyl groups 10l or the naphthalene 10e decreased quantum yield whilst Em_max was unchanged. Conversely, introducing an electron withdrawing group (EWG) (10h-k) quenched fluorescence. However, a para-N-ethyl group 10q significantly increased Em_max with lower quantum yield.

Figure 3.

Aldehyde Substituent Variation

Upon study of the amine input (7), products displayed diverse fluorescent character (Fig. 4). Notably a substantial red-shift with a decrease in quantum yield was observed on switching from pyridine 10d (Em_max 473 nm, 0.35 ϕ) to pyrimidine 10r (Em_max 553 nm, 0.02 ϕ). The pyrazine 10s had similar quantum yield to 10d and pyridazine 10t had similar properties to 10r. The amino-pyrimidine precursor to 10u failed to react.

Figure 4.

Aminopyridine Variation

Introduction of ester functionality proved mixed. Para-methoxy ester 10v increased both wavelength and quantum yield (Em_max 514 nm, 0.42 ϕ), whilst the meta-analog 10w exhibited poor quantum yield, and the ortho-analog 10x simply quenched fluorescence. We thus concluded the aldehyde diversity element of the high Em_max (546nm) pyridazine core 10t were worthy of SPPR evaluation and 8 analogs were prepared (Fig. 5). In summary, whilst Em_max was high for 10ag (Em_max 572 nm, 0.06 ϕ), quantum yields were consistently poor across the series. Thus, we reverted to the high quantum yield meta-methoxy-ester 10v and sought to optimize aldehyde-derived variants (Fig. 5). The Em_max varied little for the 10ah-am at ~510 nm, however as seen in Fig. 3, methoxyl groups increased quantum yields [10al, 0.51 ϕ]. More interestingly, the first fluorescent color change and red-shift was seen (green to yellow) promoted by the para-N-ethyl group 10ao (Em_max 540 nm, 0.25 ϕ) with good quantum yield.

Figure 5.

Pyridazine and Para-Methoxy Ester Variation

Noting a para-EWG in 7 was optimal, alternate groups were studied (Fig 6.). Thus, a 4-pyridyl group in combination with trimethoxy-benzaldehyde 10ax had a similar fluorescent profile to 10ao. On incorporation of para-N-ethyl 10ay, a color change was noted (yellow to orange) (Em_max 581 nm, 0.06 ϕ), albeit with poor quantum yield.

Figure 6.

4-Pyridyl and 4-Nitro Variation

When EWG strength was further increased with the para-nitro analog 10az, red fluorescence was observed (Em_max 651 nm]. Indeed, on introduction of a para-N-ethyl group 10ba Em_max jumped from visible range to near-infrared (NIR) (720 nm) which may explain quenching seen with 10k.

Different pairings of EDGs: EWGs were then studied (Fig. 7). Double methoxyl-substitution was deleterious to fluorescence (10at, Em_max 500 nm, 0.10 ϕ). Other EDGs on the aminopyridine 10as-aw had poor outcomes.

Figure 7.

Exploration of EDG-EWG Relationships

Driven by prior SPPR, we returned to the pyrido-imidazo[4,5-c]isoquinolines and prepared products 16f-16i (Fig. 8). Fluorescence emission increased for 16f compared to the aminopyridine 16c, but it had a lower Em_max compared to its amine congener 10ad.

Figure 8.

Second Look at Tetracyclics

This confirmed our prior observation that tetracyclic pyrido-imidazo[4,5-c]isoquinolines had poorer fluorescent properties than the imidazopyridines 10. Finally, additional tetracyclics were evaluated with 2 distinct sources of diversity from aldehydes (Table 3). This required the development of a 2-step procedure with introduction of ‘aldehyde 1’ in 10d, followed by purification and reaction with ‘aldehyde 2’ in a condensation-cyclization step (Entry 2,3,6).

Table 3.

Asymmetric Aldehyde Variation

Entrya	Acid (eq.)	Solvent (0.1 M)	Time (min)	Yield (%)
1	Cu(OTf)2(1.0)	DCM	5	89
2	Cu(OTf)2(1.0)	DCM	10	95
3	Cu(OTf)2(1.0)	DCM	20	99
4	H2SO4(50 eq.)	MeOH	5	41
5	H2SO4(50 eq.)	MeOH	10	72
6	H2SO4(50 eq.)	MeOH	20	99

^aScale: 0.5 mmol, DCM ran at 100 °C and MeOH at 140 °C (microwave irradiation).

With the introduction of a naked aryl 18a, fluorescence wavelength decreased compared to 16c (Fig 9). Similar to the amine 10k, introduction of an EWG quenched the fluorescence fully in 18b. When cinnamaldehyde was introduced (18c), a similar pattern was followed with wavelength and quantum yield decreasing compared to the amine 10v.

Figure 9.

Electronic Effects of Aldehyde Variations

During these studies, we note the amino-imidazopyridines had two Abs_max values (Fig. 10). Fluorescence quantum yield was measured for both Abs_max values and consistently the higher Abs_max value provided higher quantum yields (10al, 10e). Fluorescence measurements throughout were conducted in methanol, which although afforded lowest wavelength emissions, garnered twice the quantum yield seen in DMSO and DMF with 10e.^{21, 22}

Figure 10.

Quantum Yield and Solvent Effects

TD-DFT computations.

Time-Dependent Density Functional Theory (TD-DFT) computations were used to study a representative sample of the molecules and a detailed description of calculated adsorption compared to experimentally observed is discussed with good alignment in the Supplementary Information.

CONCLUSION

This work provides a methodology for the preparation of fully tunable fluorophores with good quantum yield as well as a novel one-pot synthesis of cyclized pyrido[2’,1’:2,3]imidazo[4,5-c]isoquinolines. Structure-photophysical property relationships (SPPR), solvent effects, and time-dependent DFT (TD-DFT) computations are presented. The computations reveal the effects of various chemical modifications on the absorption and subsequent fluorescence properties. Modification of the para position of the imidazopyridine moiety proved to be the most effective for modulating the wavelengths of excitation and fluorescence due to the differing character of the NTO’s. Substitutions that result in excitations with hole-electron transfers to opposite sides of the molecule correspond to low fluorescence intensity. The ability to match the spatial overlap and minimize the hole-electron separation associated with the excitation should be considered in the design of future fluorogenic probes.

Figure 1.

Fused Nitrogen Heterocycles

Scheme A.

Groebke–Blackburn–Bienaymé Reaction