Mechanistic Insight into the Thermal “Blueing” of Cyanine Dyes

Abstract

In recent work to develop cyanine dyes with especially large Stokes shifts, we encountered a “blueing” reaction, in which the heptamethine cyanine dye Cy7 (IUPAC: 1,3,3-trimethyl-2-((1E,3E,5E)-7-((E)-1,3,3-trimethylindolin-2-ylidene)hepta-1,3,5-trien-1-yl)-3H-indol-1-ium) undergoes shortening in two-carbon steps to form the pentamethine (Cy5) and trimethine (Cy3) analogs. Each step blue-shifts the resulting absorbance wavelength by ca. 100 nm. Though photochemical and oxidative chain-shortening reactions had been noted previously, it is simple heating alone or with amine bases that effects this unexpected net C₂H₂ excision. Explicit acetylene loss would be too endothermic to merit consideration. Our mechanistic studies using ²H labeling, mass spectrometric and NMR spectroscopic analyses, and quantum chemical modeling point instead to electrocyclic closure and aromatization of the heptamethine chain in Cy7 forming Fischer’s base FB (1,3,3-trimethyl-2-methyleneindoline), a reactive carbon nucleophile that initiates chain shortening of the cyanine dyes by attack on their polymethine backbones. The byproduct is the cationic indolium species TMP (IUPAC: 1,3,3 trimethyl-2-phenyl indolium).

Introduction

Organic fluorescent dyes have been instrumental in areas ranging from cellular biology¹⁻¹² and drug design^13,14 to advanced materials engineering^7,15−27 including energy capture and conversion.^19,28−30 Key to continued progress in this field is the ability to tune the chromophores’ structural and photophysical properties to suit their target applications. Among the most widely utilized dye families, cyanines boast high brightness and easy synthetic tunability.³¹ Cyanine dyes are typically categorized based on the (odd) number of methine units in their electronic push–pull structure between the terminating alkyl-indolenine/indolium groups. Monomethine (Cy1) and trimethine (Cy3) absorb and emit light in the UV–vis range,^3,32 while the longer pentamethine (Cy5) and heptamethine (Cy7) cyanines span from visible to near-infrared (NIR) wavelengths,^{9,29,31,33−36} where the light’s deep tissue penetration and the dyes’ minimal autofluorescence enable their use for in vivo imaging and photodynamic therapy.^{4,5,10,11,37−41} Recently, because their absorption range covers a substantial fraction of the solar spectrum, Cy5 and Cy7 dyes have also been developed for energy harvesting.^23,28

As polyene iminium chromophores, the cyanine dyes show substantial chemical reactivity.^{5,13,20,42−45} Their degradation involving singlet oxygen has been investigated in recent decades, leading to novel insights and methods. For instance, Schnermann et al. found that in situ photooxidation of Cy7 systems yields truncated cyanines via net C₂H₂ excision from the parent polymethine chromophore.⁵ The resulting hypsochromic shift, dubbed “photoblueing”, has potential applications in super-resolution imaging and single-particle tracking.^40,43,46 As first discovered, this reaction gave <2% yield of the phototruncated product Cy5, but by raising the pH to 9.5 in 1 M CAPSO buffer, the Schnermann group was able to increase the observed truncation yield to 17.2% (Figure 1a).⁴

Figure 1.

(a) Prior work by Schnermann et al. on oxygen-mediated cyanine phototruncation; (b) nucleophilic chloride displacement from IR-786, a centrally constrained Cy7, by 2° amines to yield 4’-aminated IR-786 dyes with large Stokes shifts; (c) attempted Buchwald amination of 3’-Br-Cy7 and the isolated products 3’-Br-Cy5, 3’-Br-AsCy6(NEt₂) along with detection of FB (as FB-H⁺ via ESI-MS).

Our own work⁴⁷⁻⁴⁹ has recently extended the previously reported strategy^50,51 of attaching donor substituents midchromophore to tune the heptamethine cyanines’ Stokes shift, as shown in Figure 1b. Specifically, the synthesis entailed displacement of a halide by a secondary amine at C4’ of a Cy7 framework. Noting the large spectral changes brought on by this substitution, we sought to compare the effects of installing similar donors in the C3’ position, which lacks the conjugation pathways of the C4’ location. The required 3’-Br-Cy7 was readily accessed via Klan’s practical route to backbone functionalized heptamethine cyanines via pyridinium salt (Zincke) ring opening.^31,34 However, upon attempting Buchwald amination with diethylamine (Figure 1c), instead of the target 3’-aminated Cy7, we isolated the symmetric polymethine 3’-Br-Cy5, an apparent result of C₂H₂ loss from the parent cyanine. The other significant product observed was Fischer’s Base (1,3,3-trimethyl-2-methyleneindoline, designated herein as FB). Further investigation found that these transformations occurred with the amine components alone, omitting the Buchwald catalyst and excluding both molecular oxygen and light. They thus follow a path different from that reported by Schnermann and others.^4,5

Herein, we report a mechanistic analysis of this base-promoted, anaerobic, thermal shortening of cyanine dyes. We use the “Cy” names to denote the unfunctionalized N-methyl indoline/indolium polymethine structures as shown in Figure 1a; these share the core chromophoric framework found in the commercial “Cy7” dye (sulfonated on the indole moieties and bearing N-alkyl sulfonate side chains). Furthermore, the counteranion to all of the cationic structures in this manuscript are iodide (omitted in the structures), unless specified otherwise. The conclusion is that Cy7 is unique; even in the absence of other reagents, it can undergo facile electrocyclization and aromatization. This process forms a new species, TMP (IUPAC: 1,3,3 trimethyl-2-phenyl indolium) and liberates free FB, which can then nucleophilically attack Cy7 and related polymethine cyanines, shortening their polyene backbones. Like FB, secondary amines can react with cyanines, generating asymmetric analogs of varying length. A byproduct of this mechanistic analysis is a practical method for the “blueing” of cyanine dyes and their respective symmetric and asymmetric derivatives. We note the beautiful work of Klan and co-workers that describes findings related to those described in this report. Professor Klan and his team were gracious enough to wait to submit findings at the same time as ours; the reader is referred to the accompanying manuscript in this journal.⁵²

Results and Discussion

Cyanine Reactions with Amines

As noted above, we initially observed the Cy7 to Cy5 two-carbon truncation in the context of attempted Buchwald amination of 3’-Br-Cy7 (Figure 1c). With or without the Buchwald catalyst, treatment with diethylamine (10 equiv) and Hunig’s base, (diisopropylethylamine, DIPEA, 2 equiv) at 70 °C in acetonitrile took a surprising direction. In addition to free FB, ESI-MS revealed two new polyene products with molecular ions m/z = 387/389 and m/z = 461/463, corresponding to 3’-Br-AsCy6(NEt₂) and 3’-Br-Cy5, respectively (Figure 1c). Note a point of usage here: “AsCy6(NEt₂)” designates the asymmetric product with the secondary amine as one end group of a hexamethine chain. These new derivatives were isolated, purified, and structurally analyzed by NMR, HRMS, and ESI-MS. Interestingly, small amounts of dehalogenated products Cy7, Cy5, and Cy3 were also noted (see Figure S2 for full details of the products).

Omitting the halogen to simplify the system, we then examined Cy7 itself (m/z = 409). Treatment with diethylamine and DIPEA (Figure 2a) yielded the analogous AsCy6(NEt₂) polymethine in addition to chain shortened species AsCy4(NEt₂) and AsCy2(NEt₂). As with the previously isolated 3’-Br-Cy5, pentamethine Cy5 was also found in this reaction. We also noted the formation of two monomeric indolines, cationic 1,1,3-trimethyl-2-phenyl indolium (TMP, m/z = 236) and Fischer base itself, FB (observed by ESI-MS as FB-H). To track the polymethine components of the parent cyanine, pentadeutero Cy7-D₅ was synthesized as depicted in Figure 2b. Its treatment with diethylamine and DIPEA led to isolation of centrally trideuterated Cy5-D₃, along with the corresponding AsCy[2,4, and 6](NEt₂) species, mono-, tri-, and pentadeuterated, respectively (see Supporting Information for characterization). Notably, here, the TMP formed retained all five deuterium atoms, while the FB showed none.

Figure 2.

(a) Reaction of Cy7 with DIPEA and diethylamine (70 °C, 60 h) to yield chain-shortened dyes, both symmetric and asymmetric. Yields are given for isolated products, along with the amount of Cy7 recovered after 60 h. (b) Synthesis of Cy7-(D₅) and its truncation products corresponding to the penta, tri, and mono labeled AsCy products seen with the unlabeled Cy7. (c) Proposed mechanism for amine attack on Cy7 liberating FB, and AsCy6(NEt₂).

The above treatment of Cy7 with a mixture of DIPEA and diethylamine gave the isolated products AsCy6(NEt₂) (m/z = 309) and Cy5 (m/z = 383) in a ca. 2:1 ratio (Table 1, entry 1) and 35% overall yield. Further simplifying the reaction conditions to probe the role of each component, Cy7 was treated with diethylamine alone (10 equiv; Table 1, entry 2) giving a higher selectivity for AsCy6(NEt₂) and higher yield (45% after 60 h at 70 °C). The analogous room temperature reaction was slower, but after 160 h the yield was higher (73%), with similar selectivity (Table 1, entry 3). At even longer times, the AsCy6(NEt₂) was converted to AsCy4(NEt₂) and on to AsCy2(NEt₂). With 30 equiv of diethylamine, the preference for AsCy6(NEt₂) over Cy5 increased, as did the chain-shortening processes (Figure S3). At 70 °C, reaction of Cy7 with 10 equiv of another secondary amine, morpholine, was faster, giving a 51% yield in 20 h with an AsCy6(Morph):Cy5 selectivity of 3.6:1 (Table 1, entry 4). All the above reaction mixtures also showed substantial amounts of FB, the end group displaced by the secondary amine. Importantly, the polyene chain shortening such as AsCy6(NEt₂) → AsCy4(NEt₂) seen in these reaction mixtures also suggested the thermochemically reasonable (vide infra) release of the C₂H₂ fragments in the form of vinyldialkylamines.

Table 1. Summary of Reactions with Isolated yields^a.

entry	cyanine	nucleophile	3° amine	solvent	temp (°C)	time (h)	isolated products (yield)b
1c	Cy7	HNEt2	DIPEA	ACN	70	60	Cy5 (11%), AsCy6(NEt2) (24%)
2	Cy7	HNEt2		ACN	70	60	Cy5 (12%), AsCy6(NEt2) (33%)
3	Cy7	HNEt2		ACN	RT	160	Cy5 (17%), AsCy6(NEt2) 56%)
4	Cy7	morpholine		ACN	RT	20	Cy5 (11%), AsCy6(Morph) (30%)
5	Cy7		DIPEA	ACN	70	60	Cy5 (18%), Cy3 (2%)
6	Cy7		DIPEA	ACN	100	48	Cy5 (51%), Cy3 (3%)
7	Cy7		DIPEA	ACN	180	3	Cy5 (38%), Cy3 (8%)
8	Cy7		DIPEA	EtOH	70	60	Cy5 (2%)
9	Cy7		DIPEA	CH2Cl2	70	60	Cy5 (26%), Cy3 (2%)
10	Cy7		quinuclidine	ACN	70	60	Cy5 (5%), Cy3 (15%)
11	Cy7		quinuclidine	ACN	100	36	Cy5 (10%), Cy3 (24%)
12	Cy7.5		DIPEA	ACN	100	40	Cy5.5 (25%), Cy3.5 (17%)
13	IR-786	FB		ACN	100	48	Cy3 (30%)
14d	Cy7			ACN	100	60	Cy5, Cy3, TMP, FB

^aReactions were run in sealed vessels on a 0.04 mmol scale in degassed solvents [0.03 M cyanine dye] with 10 equiv of additive (amine or FB) used.

^bNote that FB and Cy7 were observed for all reactions with the exception of Cy7 for entries 10 and 11 (with quinuclidine). Recovered Cy7 yields for entries 1–4 are 17%, 15%, 10%, and 32%, respectively.

^cDiethyl amine (10 equiv) and DIPEA (2 equiv). See the SI for other products identified by ESI-MS.

^dProducts were identified by MS and crude NMR.

To further probe these end-group exchange events, we reacted pure AsCy6(NEt₂) with morpholine (10 equiv), the secondary amine depicted in Figure 3a. Almost instantaneously, exchange of AsCy6(NEt₂) (m/z = 309) to AsCy6(Morph) (m/z = 323) was observed, along with small amounts of Cy3, Cy5, and Cy7 (Figure 3b). We envision formation of these cyanines via liberation of FB from AsCy6(NR₂) by free morpholine forming conjugated diamines (denoted Morph₂CyX, X = 3, 5, and 7; see Figure 3c). The liberated FB in turn forms Cy7 by displacing HNR₂ from AsCy6(NR₂) at C6’ (see numbering in Figure 3b). Attack at C4’ leads to Cy5 and vinylamine (R₂NCH=CH₂) and analogously at C2’, giving Cy3 and the corresponding dienylamine. Particularly noteworthy is the observation of Morph₂Cy7, which implies displacement of an end group by N-vinylmorpholine (Figure 3c). These ready exchanges of secondary amines, FB, and vinylamines produce a diverse mixture of polymethines with increasing proportions of the chain-shortened products over time.

Figure 3.

(a) Attack pathways and (b) products from morpholine attack on AsCy6(NEt₂) at 70 °C, which opens up pathways of FB attacking AsCy6(Morph) leading to formation of Cy3, Cy5, and Cy7. (c) Dimorpholine (Morph)₂-polyene fragments identified by ESI-MS. Of particular note here is the presence of (Morph)₂Cy7, implying addition of vinylmorpholine. (d) Reaction of Cy7 in the presence of different 3° amines and FB-H iodide salt.

Having noted a greater preference for the truncated product Cy5 over AsCy6(NEt₂) when DIPEA was present, we next studied this tertiary amine acting alone on cyanines. Treated with only DIPEA (10 equiv) at 70 °C for 60 h, Cy7 reacted slowly to form 18% Cy5, along with 2% of the further truncated trimethine Cy3 (Table 1, entry 5). Higher temperatures accelerated polyene shortening (Table 1, entries 5–7). As in the previous reactions, we noted the early formation of indolines TMP and FB via ESI-MS. Replacement of the initially studied acetonitrile with the protic solvent ethanol dramatically decreased the yield of Cy5; only about 2% Cy5 was isolated after 60 h at 70 °C with 10 equiv of DIPEA. On the other hand, the Cy5 yield improved to 26% when the same reaction was run in dichloromethane (Table 1, entries 8 and 9). Lacking further evidence, we have added the speculation that both the DIPEA and any FB formed are inhibited by hydrogen bonding interactions with hydroxylic solvents, whereas in DCM, their full reactivity is available for attack on the less solvent-stabilized Cy7.

Beyond solvent and temperature, the next parameter explored was the choice of tertiary amine. Reaction of Cy7 with triethylamine (Figure 3d) gave results like those from DIPEA treatment, albeit slightly faster. Other tertiary amines gave similar, but not identical, behavior. For instance, after 48 h at 100 °C with pyridine or DBU, (1,8-diazabicyclo[5.4.0]undec-7-ene) Cy7 was largely unconsumed; only trace chain shortened polymethines Cy5 and Cy3 were detected, along with the now familiar TMP and FB. Conversion of Cy7 using quinuclidine gave more rapid chain truncation; after 60 h at 70 °C, yields of Cy5 and Cy3 were 5% and 15%, respectively, with 13% recovery of Cy7. At 100 °C, Cy7 was nearly consumed in 36 h, yielding 10% and 24% of Cy5 and Cy3, together with TMP and FB. Here, the majority product was Cy3, not Cy5 (compare Table 1, entries 10, 11 vs entries 5–7). In this case, the observation of a large m/z peak at 435 (see Figure S4), corresponding to nonamethine Cy9, was of particular importance. As FB is the nucleophile that effects end group exchange, the byproduct implied by chain shortening is the vinylated Fischer base, FB-CH=CH₂. Like FB itself, this dienamine species may attack and displace FB, leading to polyene chain lengthening. This finding of chain extension is analogous to the observation of (Morph)₂Cy7 above (Figure 3c), implying the transfer of vinylated end groups.

The above Cy7 reactions offer selectivity to a range of symmetric/asymmetric polymethines with varying degrees of conjugation, and thus absorption/emission wavelengths ranging from the NIR to UV (Figure S1, Figure 2a). This conversion has potential for useful applications as highlighted by Schnermann et al. in the context of their related photoactivated truncations.^4,5 However, mechanisms by which cyanines could undergo the above anaerobic, thermal C₂H₂ excisions are not obvious. Thus, the activation and fate of the C₂H₂ fragment must be considered as part of any proposed chain-shortening mechanism.

Experimentally no C₂H₂ was detected in solution or headspace. On simple thermochemical grounds, it is difficult to imagine direct release of C₂H₂; such a process should be at least as endothermic (roughly uphill by 40 kcal/mol) as the simple polyene shortening in eq (1). What, then, becomes of this two-carbon fragment? To “pay” the energetic cost of net C₂H₂ excision, four candidate energy-compensating reactions may be envisioned (eq 2–5, Figure 4a) (a) net N–H addition across the C₂H₂ fragment by the secondary amine to form an N,N-dialkylvinylamine (e.g., eqs 2 and 3); (b) net addition of adventitious water across the acetylene to form acetaldehyde (eq 4); (c) net C–H addition of FB across C₂H₂ to form FB-CH=CH₂ (eq 5); or (d) net formal trimerization of acetylene to form benzene (exothermic by 143 kcal/mol or ca. 48 kcal/mol per C₂H₂ unit) or a phenyl moiety.

Figure 4.

(a) Except as otherwise specified, ΔH_rxn values shown above are computed from ΔH_f data in the NIST webbook database.⁵³ For diethylamine, the ΔH_f was computed using the G3(MP2) method;⁵⁴ (b) comparison of calculated relative acidities of FB and amines in acetonitrile; (c) NMR spectrum of acetaldehyde in the reaction mixture, and the calculated hydrolysis energy.

The simple mechanism depicted in Figure 2c rationalizes the conversion of Cy7 to the corresponding aminated (asymmetric) polymethines by secondary amines. Here, the secondary amine attacks the C2’ position of the methine backbone, neutralizing the indolium nitrogen. Proton transfer from the now cationic amine nitrogen to the C1’ carbon enables cleavage of the C1’-C2’ bond, releasing FB and the AsCy6 product. Consistent with this mechanism, as in Figure 2b, diethylamine treatment of C2’-C6’ deuterium-labeled Cy7 showed full label retention in the corresponding asymmetric product, AsCy6(NEt₂)-D₅, along with unlabeled FB. Evidently, secondary amine attack like that shown in Figure 3c removed the elements of C₂D₂, rather than degrading from the other end of the chain. Based on the order of product appearance, growth, and decay as detected by ESI-MS over the course of the reaction, we concluded that AsCy4(NEt₂) and AsCy2(NEt₂) form downstream from AsCy6(NEt₂) via sequential chain-shortening steps, rather than directly forming from Cy7. This was confirmed via the explicit reaction of AsCy6(NEt₂) with morpholine, as depicted in Figure 3. Notably, these processes lead to enamine byproducts, both the methylene indoline FB and (presumed) N,N-diethylvinylamine. This scenario bypasses high energy species such as free acetylene, removing the lost C₂H₂ fragment instead as the vinyl group in N,N-diethylvinylamine. Although the FB is detected, the latter vinylamine is a reactive, low molecular weight species that we were not able to directly detect via ESI-MS. In the presence of even traces of water, this species would be expected to readily hydrate and hydrolyze to liberate acetaldehyde and diethylamine. Indeed, acetaldehyde was detected by ¹H NMR in these reaction mixtures (Figure 4c).

Though the formation of asymmetric polymethines (AsCys) from end group replacement/chain shortening of Cy7 by secondary amines appears mechanistically straightforward, as discussed above and shown in Figure 3, the conundrum of truncation promoted by tertiary amines must be examined more closely. To begin the tertiary amine-promoted conversion of Cy7 to symmetric Cy5, amine attack at C2’ or C4’ of the methine chain was considered. However, with no reasonably acidic protons available, further reaction, analogous to that shown in Figure 3c, seemed unlikely. Quantum chemical modeling, including the effects of the CH₃CN as a solvent, found attack at the C2’ or C4’ positions by trimethylamine (as a model tertiary amine) to be more than 10 kcal/mol endergonic. In all the above tertiary amine-promoted reactions, TMP and free FB were detected, and as further discussed below, these appear to be the essential players in the tertiary amine-promoted chain shortening reactions. We argue that the tertiary amines mainly serve as bases that work by promoting the proton transfers that isomerize adducts of FB with Cy7 and the shorter cyanines leading to net C₂H₂ excision. A key finding (vide infra) or this ultimate mechanistic interpretation was the fact that purified Cy5 showed no reaction (specifically, no chain shortening, release of FB, or presence of TMP) on treatment (100 °C for 48 h) with the tertiary amines quinuclidine or DIPEA (see Figure 7b).

Figure 7.

(a) Reaction and comparison of ESI-MS sequences of Cy7 and Cy7-(D₅) (48 and 20 h) with FB in acetonitrile at 100 °C. (b) I ESI-MS sequence after 12 h of Cy5 and DIPEA before addition of morpholine. II ESI-MS sequence after 22 h after adding morpholine to the mixture. III ESI-MS sequence after 36 h reaction time. (c) Reaction of Cy5 + FB-0.5 (right) and expansion of the scrambled FB cyanines (left) observed by ESI-MS. (d) Independent reactions of Cy7 + Cy7.5 (with DIPEA, right) and Cy7 + Cy5.5 (with quinuclidine, left) lead to the same crossover cyanine products, and truncated FB-0.5 and FB cyanines.

Mechanistic Experiments

If the tertiary amines are not directly acting on the Cy7, how do the chain-shortening reactions occur? Tertiary amines differ from the reactive secondary amines in lacking an exchangeable proton. A new set of hypotheses was sparked by our observation of free FB formed from Cy7 in the presence of tertiary amines. As the source of the FB, an electrocyclization path was suggested by the concomitant appearance of 1,3,3-trimethyl-2-phenylindolinium ion (TMP) (Figure 5a) and analogous TMP-D₅ noted above. The exchange of secondary amines and FB seen above as cyanine end groups also suggests a direct role for FB in chain shortening. Insights supporting this path came from two minimal reactions: (a) Cy7 heated alone at 100 °C in acetonitrile (Figure 6); (b) Cy5 treated similarly; these were supplemented by four simple binary reactions: (c) DIPEA + Cy7 (Figure 3d); (d) DIPEA + Cy5 (Figure 7b); (e) FB + Cy7 (Figure 7a); and (f) FB + Cy5 (Figure 7c, see Figure S8 for graphical summary).

Figure 5.

Proposed intramolecular electrocyclization pathway of Cy7, resulting in formation of TMP and FB. Computed TSs are shown for cis and trans cyclizations from coiled Int 1 and for the cleavage of Int 4 to release FB and TMP. At right is shown the strongly exothermic oxidative aromatization path of Int 2, illustrating the strong hydride accepting capacity of TMP. The prototropic reaction energetics are computed using FB as the proton trafficking base, but as seen in Figure 4b, the relative basicities of FB and aliphatic amines are nearly equal. Notably, like tautomerization, proton equilibration in the absence of added base is slow, so that heating of Cy7 without base or with a weak base (e.g., pyridine) leads to slow emergence of TMP and FB, which then autocatalytically accelerates the reaction. (b) Free energy diagram of FB attack pathways on Cy7, leading to Cy5 (via C4’ attack, right) and Cy3 (via C2’ attack, left), along with their corresponding FB-polyene species. Computed TSs are shown for attack and cleavage steps.

Figure 6.

Heating of Cy7 alone in acetonitrile at 100 °C, showing formation of species Cy5, Cy3, and Cy1, in addition to TMP, FB, and other indole products previously mentioned at 20 h (top) and 65 h (bottom).

In reaction (a), heating of Cy7 alone (Figure 6) forms substantial amounts of TMP, FB, and chain-shortening products Cy5 and Cy3. In contrast, in reaction (b), heating of the shorter Cy5 resulted in no change. Even in cases where Cy5 was subjected to reaction temperatures exceeding 175 °C, no chain shortening was observed. This comparison most starkly illustrates the unique additional reactivity of Cy7, whose heptamethine chain enables cyclization, TMP formation, and FB release, initiating chain shortening processes.

In reaction (c) as recounted previously, DIPEA + Cy7 forms TMP and FB (Figure 3d). As in the reagent-free reaction of Cy7, the chain shortening to form Cy5 and Cy3 is then proposed to occur via attack by the released FB on Cy7 at the C4’ and C2’ positions as shown in Figure 5b. Proton transfers, assisted by the DIPEA (when present) and/or by the FB formed when Cy7 is heated alone, then enable the release of Cy5 along with FB-CH=CH₂, the vinylated analog of FB. Like the vinylation of a secondary amine, this represents a thermochemically reasonable fate for the C₂H₂ fragment removed in the net chain shortening.

In case (d), treatment of Cy5 with DIPEA (10 equiv) at 100 °C (see Figure 7b, top panel) showed essentially no reaction. However, subsequent addition of the secondary amine morpholine (10 equiv) initiated an immediate reaction as expected, forming AsCy4(Morph) and (Morph)₂Cy3, and liberating FB. Small amounts of Cy3 were also observed.

Reaction (e), FB + Cy7 combined alone without an amine base (see Figure 7a) led mainly to Cy3. The implied byproduct, FB-CH=CH–CH=CH₂, is a polyene even more susceptible to intermolecular reactions than FB-CH=CH₂.⁵⁵ As a control experiment, treatment of Cy7 with the iodide salt of FB-H in the absence of base was explored (Figure 3d); this reagent, lacking the reactive methylene group, led simply to ESI traces analogous to those of the Cy7 reaction alone, a process that begins by forming the cyclization product TMP and releasing FB.

In reaction (f), FB + Cy5 slowly produced Cy3 but interestingly, also generated small amounts of Cy7. This latter finding implies release and recapture of FB-CH=CH₂ by Cy5, confirming the reversibility of these attacks and exchanges. This is analogous to the N-vinylmorpholine addition shown in Figure 3c. In an extension of this reaction, treatment of Cy5 with FB(0.5) was explored (Figure 7c). Here, the “0.5” designation signifies the presence of benzo[e]indolium, IUPAC: 1,1,3-trimethyl-2-methylidenebenzo[e]indole. This experiment produced a mixture of crossover products Cy(0.5) and Cy(0.25) (“0.25” refers to a polymethine with FB and FB(0.5) end groups) pentamethine and trimethine cyanines (Figure 7c). This points to nucleophilic attack at C2’ of Cy5 by FB, efficiently exchanging end groups, while more slowly effecting chain shortening. These processes are directly analogous to the above-described reactions with secondary amines.

Though the formation of Cy7 from Cy5 in experiment (f) implies its intermediacy and reactivity, the vinylated FB byproduct was not directly observed. Previous attempts to synthesize and isolate it in monomeric form led to dimer- or oligomerization, as described by Hubschwerlen and co-workers.⁵⁵ Like vinylamines, this diene may be easily protonated and hydrolyzed by trace water, forming acetaldehyde and regenerating FB. Quantum chemical modeling finds this hydrolysis to be −7.9 kcal/mol exothermic in acetonitrile (Figure 4c, shaded box). Meanwhile, no byproducts that could be attributed to breakdown of DIPEA (or any of the tertiary amines studied) itself were observed.

Isotopic labeling studies provided further support for the scenario outlined in Figure 5a. As noted earlier, DIPEA + Cy7-D₅ yielded the centrally deuterated pentamethine Cy5-D₃, reflecting removal of the C2’-C3’ (CD)₂ fragment from the polymethine chain, while the TMP product retained all five deuteriums. In addition, the protons at the C1’ and C7’ backbone positions were revealed to be slowly exchangeable by stirring an acetonitrile solution of Cy7 that included 10 equiv of D₂O at ambient temperature for 48 h. The resulting C1’/C7’-deuterated-Cy7 was isolated and characterized by NMR and ESI-MS (Figure S6). This finding presumably reflects the proton exchange expected at the C1’/C7’ sites that occurs during reversible FB attack and end group exchange. Importantly, little or no exchange is seen at sites C2’-C6’ under these conditions.

To further explore the reversibility of attack and release of FB end groups, we heated a 1:1 mixture of Cy7 and Cy5.5 in acetonitrile with quinuclidine (Figure 7d). The resulting cyanine products, Cy5.25 and Cy3.25 identified by their m/z peaks, had exchanged end groups. Additionally, Cy7 and Cy7.5 were reacted with DIPEA as described above, and the crossover cyanine products including Cy7.25 were formed in the reaction. In both experiments, the corresponding TMP, TMP(0.5), and FB(0.5) fragments were also seen.

This observation of end group exchange confirmed that the methylene indoline units such as FB readily undergo release and reattachment, consistent with their participation in chain shortening. The chain shortening mechanism presented in Figure 5 above also implies a reasonable pathway for direct truncation of Cy7 to symmetric trimethine, Cy3. The initial part of the mechanism is the same as described for Cy7 undergoing end group exchange in which FB attacks the Cy7 backbone at the C2’ position. Proton transfer then may occur to the former C1’ position or the former C3’ position; the energetically degenerate protonation at C1’ activates end group exchange, replacing the original FB moiety with that of the attacking FB. The higher energy protonation at C3’ enables cleavage of the original C2’-C3’ bond to release Cy3 and (by implication) FB-CH=CH–CH=CH₂ (Figure 5b).

The fate of the excised C₂H₂ fragment remains somewhat elusive. ESI-MS peaks corresponding to the proposed vinyl and dienyl FB species FB-CH=CH₂ and FB-CH=CH–CH=CH₂ were not directly observed. However, as noted above, the appearance of small amounts of chain lengthening product Cy7, formed during FB treatment of Cy5, implies the presence and participation as a nucleophile of FB-CH=CH₂. Analogous C₂H₂ transfers were evidenced in the reactions of morpholine and AsCy6(NEt₂) (see Figure 3c). Such species are highly reactive, and may undergo further chemistry, either in the form of self-condensation or by reaction with adventitious water.⁵⁵ Despite our best efforts to ensure dryness, hydrolysis is clearly possible in light of the observation of other oxygen-containing fragments, most notably 1,3,3-trimethyl-2-indolone (depicted below as IN–OH, the protonated M+H adduct detected by ESI-MS), the lactam corresponding to FB.

Quantum Chemical Modeling

To develop a more complete picture of the reaction paths involved in the thermal “blueing” of the Cy7 cyanine dye, quantum chemical simulations were run using the ωB97X-D density functional,⁵⁶ the 6-31G* basis set,⁵⁷⁻⁵⁹ and the conductor-like polarizable continuum model (CPCM) to account for the acetonitrile solvent;⁶⁰ we designate this overall model ωB97X-D/6-31G*/CPCM(MeCN). These calculations were performed with the Spartan program.⁶¹ Candidate structure lists for each species were initially generated using Spartan’s molecular mechanics conformation generator. Reoptimization and energy ordering of the lists (typically tens to hundreds of structures) were performed with the semiempirical PM6 method. After removal of duplicate structures, single-point ωB97X-D/6-31G*/CPCM(MeCN) energies were computed for structures within the lowest 5 kcal/mol energy window, and the lowest ten of these were then fully optimized at this level, including vibrational analysis. For transition structures (TSs), intrinsic reaction coordinate (IRC) calculations were used to verify their connections to the respective starting materials and products. The resulting free energy changes for the computed chain shortening events are summarized in Figure 5a,b. Additional relevant computed reaction energetics are shown in Figure 4.

Figure 5a depicts the ring closure of Cy7, whose ground state is the extended linear structure shown. Regarding the trans-cis isomerizations required to coil the polyene to enable the ring closure leading ultimately to TMP, multiple TSs and energy barriers were computed, with none exceeding 14 kcal/mol. Given the well-established experimental conformational mobility of Cy7, synthesized itself from a cyclic building block via the Zincke reaction,³⁴ we omit here a detailed exploration of the many minima en route from the linear to the coiled conformations of Cy7 that set the stage for the electrocyclization. Importantly, however, the requisite coiled forms are calculated to be close in energy to the extended global minimum indicating that such geometries are thermally accessible at 100 °C among the equilibrating conformations of Cy7.

Figure 5b shows the sequence of steps by which free FB is able to reversibly attack Cy7 and achieve chain shortening. Transition state calculations for these attacks at C2’/C6’ and at C4’ of Cy7 find free energy barriers of 21.6 and 20.6 kcal/mol, respectively, suggesting little selectivity between these two paths. Similarly, the respective adducts are close in energy. Nonetheless, the former path has twice the site degeneracy of the latter, suggesting that the C2’ attack should predominate. In fact, relatively rapid end group exchange without chain shortening is experimentally observed for both Cy7 and Cy5, as depicted in Figure 7. The uphill proton transfer to form the intermediate that liberates FB-CH=CHCH=CH₂ translates into easier exchange (degenerate) of FB end groups than release of the chain shortened Cy3. Similar logic applies to the attack at C4’ in that reversibility of the FB attack is easier than proton transfer and cleavage to release FB-CH=CH₂. Likewise, in Cy5, FB attack shows a barrier of 22.4 kcal/mol, but a relatively higher barrier for chain shortening than for simple end group release. Interestingly, the barrier to FB addition to Cy3 is even higher, at 26.2 kcal/mol.

That cyclization of Cy7 occurs to form TMP and FB is experimentally clear, as is the chain-shortening action of free FB on Cy7 and Cy5. Though the proposed vinylated and dienylated FB byproducts are not directly observed, their presence is implied by the appearance of chain-extended products, as noted above. Acetaldehyde was also detected by NMR in a Cy7 + DIPEA reaction mixture (Figures 4c and S7). Trace water is hard to completely eliminate, especially in the handling of the ionic cyanine dyes, the highly polar acetonitrile solvent, and the hygroscopic amines used. We therefore propose that the initially formed alkenylated FB derivatives undergo hydrolysis, either in situ or in postreaction handling, to release acetaldehyde (or potentially acrolein) which would not be easily seen in ESI-MS. Calculations indicate that, like vinylamine, this reaction is substantially exothermic, supporting the proposed hydrolytic modes of breakdown shown in Figure 4c.

Reactivity and Structure of TMP

To explore the possible chemical behaviors of the phenyl indolium species TMP, it was independently synthesized and its reactions with amines and FB were examined (Figure S5). The computational results indicate that TMP is a potent oxidant, capable of hydride abstraction from alkyl amines such as triethylamine (−7.8 kcal/mol). This reactivity is easily rationalized in terms of the phenyl ring’s steeply twisted orientation to the indoline, enforced by three flanking methyl groups (for the X-ray crystal structure of TMP see Supporting Information). This twist (46°) limits effective resonance stabilization of its attached cationic C=N⁺ moiety. The reduction product TMP-H (detected by ESI-MS as TMP-H₂⁺, m/z = 238), seen in reactions that generate TMP, implies formation of an oxidized byproduct. The calculated reaction free energy for hydride transfer from the deprotonated species (Int-3 in Figure 5a) formed after Cy7 cyclization is −38.2 kcal/mol, as expected for an aromatizing reaction. Similarly, hydride transfer from the FB-Cy3 adduct is calculated to release 20.4 kcal/mol. In reactions involving Cy7, FB, and TMP, the expected ions resulting from these processes, m/z = 407 and m/z = 528 respectively, do appear in conjunction with the TMP-H formation. Another reaction of TMP is demethylation of the N–CH₃ group by amines and by FB, as verified via study of the N-CD₃ and N–¹³CH₃ isotopomers (Figure S5). We denote the resulting 2-phenylindolenine (m/z = 222) TMP-[DeMe] (DeMethylated). The analogous TMP-0.5-[DeMe] was likewise noted in reactions of the parent heptamethine Cy7.5 via HRMS. Substantial further rearrangement and reaction chemistry follows from TMP formation but untangling the details of the resulting convoluted reaction manifold is deferred to a later publication.

Conclusions & Outlook

This work has presented several key findings regarding the chemistry of cyanine dyes when heated in acetonitrile (see Figure 8 for a pictorial summary):

Figure 8.

Heating of Cy7 alone in acetonitrile at 100 °C forms TMP and its byproduct FB, which enables shortening of the Cy7 polymethine chain to form species Cy5, Cy3, and Cy1. Analogous reactions with secondary amines form asymmetric polymethines as well as the shortened Cy species.

• A new mechanism is described for initiation of thermal cyanine “blueing” in the absence of reagents, oxygen, or light. This process is unique to Cy7, in which the polyene chromophore undergoes 6-membered ring cyclization and aromatization, releasing Fischer’s base (FB) and the 1,3,3-trimethyl-2-phenylindolium (TMP) ion.

• The carbon nucleophile FB attacks Cy7, Cy5, and Cy3 at C2’ or C4’ positions. Degenerate proton shuffling in the resulting intermediates enables rapid replacement of their indoline end groups with the attacking FB group.

• Proton exchange in the above FB adducts may also form higher energy species, enabling chain-shortening to form Cy5, Cy3, and Cy1. These modes of reaction are verified by explicit treatment of Cy7 with FB, from which the same suite of products quickly arise. Release of vinylated and dienylated FB end groups account for the 2- or 4-carbon fragments excised from the polymethine chains.

• In contrast to Cy7, Cy5 is unreactive in isolation, but does react with added FB or secondary amines to undergo end group exchange and chain shortening analogous to that seen with Cy7.

• Secondary amines mimic the FB reactivity described above, nucleophilically attacking Cy7 at C2’ or C4’ sites, exchanging protons, and reversibly displacing the FB or amine end groups to form asymmetric cyanines of type AsCy6(NR₂) and further, to form (R₂N)₂Cy species. As with FB, proton shuffling within the amine adducts enables facile exchange of the amine end groups. Analogous chain shortening also occurs, implying release of R₂NCH=CH₂ enamines or dienamines R₂NCH=CHCH=CH₂. Like the parent cyanine dyes, the isolable AsCy products’ spectral properties span across the visible range, allowing for applications in multicolor imaging.

• Cyanines undergo ready end group exchange, equilibrating FB and secondary amines. Displacement of an aniline by FB is in fact intrinsic to the Zincke and classic related syntheses of the cyanines. Thus, in FB, itself an enamine, the exo-methylene site reacts much like an amine. This similarity is supported by theoretical calculations that find very similar acidities for the protonated forms of FB and aliphatic amine bases in acetonitrile.

• The alkenylated end groups such as FB-CH=CH₂ and R₂N–CH=CH₂, invoked as the carriers of the C₂H₂ fragments lost in chain shortening, have evaded direct observation. However, their presence is confirmed by the observation of small amounts of chain extended cyanines, for instance in reaction of FB with Cy5, and of morpholine with Cy7. Literature efforts to isolate and directly observe FB-CH=CH₂ noted failures due to dimerization and polymerization. In the presence of water, these enamines are subject to exothermic hydrolysis, which readily generates acetaldehyde, and indeed, small amounts of acetaldehyde can be detected in reaction mixtures.

• Secondary amines such as morpholine react with Cy7 and Cy5 to form AsCy chromophores even at room temperature. This reactivity offers potential for in situ adjustment of the chromophores’ spectral coverage, potentially via both chain length and amine π donor ability variations.

• The role of tertiary amines in the present studies appears to be limited to assisting in the proton shuffling involved in the cyclization of Cy7 and the end group exchange and chain shortening by FB or secondary amines. Their steric preferences in acid–base reactions may affect the selectivity in conversion of Cy7 to Cy5 and Cy3, as seen in comparing the reactivity of DIPEA (favoring Cy5) and quinuclidine (favoring Cy3). However, no products or intermediates incorporating tertiary amine or their fragments were noted, though as supported by the calculations, it is clear that they can reduce TMP to TMP-H₂.

• Lastly, we believe that the new understanding of the cyclization behavior of longer cyanine dyes such as Cy7, and the broader reactivity of this family of dyes with amine nucleophiles will help in interpreting and exploiting the rich chemistry involved in the in situ reactions of cyanine chromophores.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.4c02171.

• Additional experimental details, supplimentary figures, X-ray data, and coordinates and energies from quantum chemical studies (PDF)

Author Contributions

^† A.V. and M.M. contributed equally. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.