1. Introduction

Fluorophores have played crucial roles in fluorescence imaging techniques, including fluorescent tags [1-3], molecular probes [4-7], and super-resolution microscopy techniques [8-11]. The proper labeling of bright fluorophores, combined with fluorescence microscopy, provides a valuable tool for investigating cellular structures and processes with high spatio-temporal resolution [12]. To date, numerous fluorophores for biological imaging have been developed, including organic small molecules [1, 13-15], fluorescent proteins [16], and nanomaterials [17]. Among these fluorophores, organic small-molecule dyes offer several advantages, including high photostability, low molecular weight, and excellent biocompatibility. Conventional small-molecule fluorophores, such as cyanine [18], rhodamine [13, 19], fluorescein [20], rhodol [21], coumarin [22, 23], boron dipyrromethene (BODIPY) [24, 25], 4-dialkylamino-7-nitro-benzoxadiazole (NBD) [26], oxazine [27], and naphthalimide [28], have been widely utilized in fluorescence imaging techniques to investigate biological processes in living systems [1]. However, most of these dyes are limited in high-quality super-resolution fluorescence imaging due to their low brightness in biological media. Therefore, the rational design and synthesis of small-molecule fluorophores with high brightness across the visible and near-infrared (NIR) spectrum facilitate imaging performance and broaden their range of applications [29-31].

This review highlights the development of high-brightness small-molecule dyes for imaging, systematically presenting recent advances in enhancing their brightness through molecular structure modifications. Moreover, current limitations in chemical modifications aimed at enhancing the brightness of organic small-molecule dyes are discussed. We believe this review will assist researchers understand the chemical mechanisms involved in enhancing brightness of fluorophores for biomedical applications.

2. The process of fluorescence

A simplified Jablonski diagram illustrates the process of fluorescence and related phenomenon (Fig. 1). Upon absorption of a photon, electrons in the singlet ground state (S₀) are excited to a higher energy state (S_n) and then quickly undergo vibrational relaxation to the first singlet excited state (S₁). Subsequently, the S₁ state can return to the S₀ through various pathways. Fluorescence occurs when a photon is emitted during this transition. The S₁ state may also return to the ground state through nonradiative thermal deactivation transitions, inducing the generation of heat energy. Additionally, the excited molecule may undergo intersystem crossing (ISC) to the triplet state (T₁), which may decay back to the ground state either through heat release or phosphorescence.

The extinction coefficient (ɛ, expressed in L mol⁻¹ cm⁻¹) is a critical descriptor of the above processes. It indicates a molecule's ability to absorb light and varies with the excitation wavelength. Another key descriptor is the fluorescence quantum yield (ΦF), which indicates fluorescence efficiency. Brightness is a useful metric for comparing fluorophores and is calculated by multiplying the extinction coefficient and quantum yield (ɛ×ΦF). High brightness is preferable for achieving a high signal-to-noise ratio and a lower detection limit. Furthermore, high-brightness fluorophores facilitate molecular imaging at low excitation intensities or short exposure times, thereby minimizing tissue damage from excitation light. Importantly, the brightness of a fluorophore, which is directly related to its molecular structure, can be improved through structural modifications that increase extinction coefficients and quantum yields.

3. Improving the brightness of fluorescent dyes

Fluorophores with insufficient brightness have limited applicability. The brightness of a fluorophore is largely determined by its quantum yield. Theoretically, to improve quantum yield, non-radiative decay pathways must be eliminated. Consequently, significant research efforts have focused on enhancing quantum yield through chemical modifications.

3.1. Scaffold rigidification

Fluorophores with significant conformational flexibility exhibit non-radiative energy loss during excitation, resulting in poor quantum yields. Scaffold rigidification effectively reduces non-radiative decay and consequently increases quantum yield, as demonstrated by cyanine scaffolds, which are prone to photoinduced excited-state trans-to-cis polyene rotations (Fig. 2). For instance, the rigidification of cyanine 1 (Cy3) results in a 10-fold enhancement in quantum yield for 2 [32]. The quantum yield increased from 0.08 with 1 to 0.8 with 2 in MeOH. In 2017, Schnermann et al. developed a pentamethine indocyanine 4 with structural rigidity [33]. In MeOH, compound 4 displayed a quantum yield of 0.69, which is 4.6 times higher than that of the conventional pentamethine cyanine 3 (Cy3, ΦF = 0.15). Furthermore, the fluorescence quantum yield of 4 is largely unaffected by solvent viscosity. They further expanded the strategy to the heptamethine cyanine scaffold [34]. The modified compound 6 showed a slight increase in quantum yield (from 0.24 to 0.29) compared to the conventional fluorophore 5 (Cy7) in MeOH. Recently, Guo et al. developed a series of conformationally restricted hemicyanines by incorporating a bicyclic ring system into the skeleton of conventional coumarin hemicyanines [35]. Among these dyes, 8 exhibited a much higher fluorescence quantum yield (ΦF = 0.21) in phosphate buffered saline (PBS) compared to the unmodified dye 7 (ΦF = 0.02). These results suggest that conformational restriction is an effective strategy for inhibiting non-radiative energy loss through photoisomerization.

Furthermore, energy is dissipated from the excited state through the rotation of aromatic substituents. Anthocyanidin is a natural flower pigment with a low quantum yield. To enhance its quantum yield, our group fixed the rotational aryl rings and developed novel AC-Fluor dyes [36]. As illustrated in Fig. 3, anthocyanidin 9 exhibited a quantum yield of 0.04 in EtOH, whereas AC-Fluor 10 demonstrated a 7-fold increase in quantum yield (ΦF = 0.28). In another study [37], compound 11 with flexible structures exhibited a quantum yields of 0.02 in PBS. In contrast, 12 possessed rigid and planar molecular structures, demonstrating a higher fluorescence quantum yield (ΦF = 0.37). 2, 4, 6-Triphenylpyrylium salt 13 has unique photophysical characteristics but is unsuitable for bioimaging due to its low fluorescence (ΦF = 0.04 in EtOH). By restricting the rotation of aryl rings [38], compound 14 exhibited a more than 12-fold increase in quantum yield (ΦF = 0.51) compared to compound 13.

An azepane-substituted boron dye 15 exhibited low quantum yields (ΦF = 0.02) due to free intramolecular bond rotation (Fig. 4) [39]. After introducing a methylene bridge in the structure and restricting the C-C bond rotation, 16 displayed significantly higher quantum yields (ΦF = 0.83). Solntsev's group discovered that the synthesized chromophores of green fluorescent protein (GFP) had dramatically different fluorescence characteristics compared to wild-type GFP (wtGFP). Due to its flexible structure, compound 17 exhibits a low fluorescence quantum yield (ΦF < 10⁻⁴) [40]. Therefore, they designed a locked molecule 18, which displayed a quantum yield of 0.73 in acetonitrile, similar to that of wtGFP (ΦF = 0.79). These findings also demonstrate that controlling intramolecular bond rotation could improve dye quantum yields.

In addition, increasing the rigidity of the plane of BODIPY dyes can improve their brightness. For example, the 3, 5-diaryl-substituted BODIPY dye 19 exhibited modest fluorescence intensities (ΦF = 0.20 in CHCl₃) due to rotation of the aryl rings [41]. The constrained 20 showed about a two-fold enhancement in quantum yield (ΦF = 0.38) [42]. Another example is BODIPY dye 21, which shows low fluorescence intensities (ΦF = 0.07) in CHCl₃ (Fig. 4). In contrast, dye 22 has relatively rigid conformations and exhibits higher fluorescence quantum yields (ΦF = 0.41) [43].

The brightness of dyes can also be improved by increasing the molar extinction coefficient. As reported in the literatures, improving the planarity of BODIPY fluorophores can increase the extinction coefficient. For instance, Zhao et al. reported conformationally constrained aza-BODIPY dyes with intense absorption and strong fluorescence [44, 45]. Among these dyes, 24 exhibited a moderate quantum yield of 0.28 and demonstrated a high extinction coefficient (ε = 159, 000 L mol⁻¹ cm⁻¹) in CHCl₃ (Fig. 4). Compared to dye 23 (ΦF = 0.36, ε = 78, 500 L mol⁻¹ cm⁻¹), dye 24 showed a slight reduction in fluorescence quantum yield but a significant improvement in the absorption coefficient, resulting in a notable increase in brightness. Furthermore, Chan and coworkers synthesized a series of conformationally restricted aza-BODIPY dyes [46]. Among these dyes, dye 26 (ε = 172, 000 L mol⁻¹ cm⁻¹) exhibited superior extinction coefficients than parent dye 25 (ε = 54, 000 L mol⁻¹ cm⁻¹). In other studies, BODIPY fluorophores possessing a rigid plane also exhibited high molar extinction coefficients [47-49]. These results demonstrate that controlling intramolecular bond rotation can effectively enhance dye brightness.

3.2. Suppressing the TICT to improve the quantum yield of dyes 3.2.1. The TICT mechanism

Donor-acceptor (D-A) type fluorophores, including rhodamine, rhodol, coumarin, oxazine, naphthalimide, acedan, NBD, and others, have been widely used in molecular probes and fluorescent labels. The dialkylamino group is commonly utilized as an electron donor in D-A dyes because of its exceptional electron-donating capabilities, which promote intramolecular charge transfer (ICT) in the electronic excited state. Typically, D-A dyes exhibit high fluorescence in nonpolar solvents, but their fluorescence intensity significantly diminishes in polar solvents. Twisted intramolecular charge transfer (TICT) is one possible explanation for this phenomenon (Fig. 5) [30, 50]. It suggests the formation of a charge transfer (CT) state characterized by a perpendicular conformation resulting from intramolecular rotation between a donor and acceptor. Due to the complete charge separation between the donor and acceptor groups, the TICT state exhibits high polarity and experiences significant stabilization in polar solvents. In D-A fluorophores, radiative decays are primarily caused by the locally excited (LE)/ICT state. In nonpolar dyes, the quasi-planar emissive state is commonly termed the LE state, while in dipolar dyes, it is referred to as the ICT state. In contrast, nonradiative decays are predominantly caused by the TICT state [30]. Thus, inhibiting TICT formation in D-A fluorophores is crucial for producing high-performance fluorescent labels.

3.2.2. Alkyl cyclization of the flexible amine group

To prevent the formation of the TICT state, cyclic substituents on the nitrogen atom are commonly employed. Chemists inhibited the twisting of the amino group by cyclizing it into a rigid six-membered ring structure. This structural modification impedes the formation of the TICT state, thereby enhancing the quantum yield of the dyes in polar solvents. Bergmark et al. fixed the amino group of coumarin 28 through amino-cyclization, inhibiting its torsion and resulting in an increased quantum yield of 0.66 in water (Fig. 6) [51]. In contrast, the freely twistable N, N-diethylamino-substituted compound 27 displayed a quantum yield of only 0.055 in water. Prendergast et al. employed the same modification method for rhodol dyes [52]. The quantum yield of cyclized rhodol dye 30 in PBS was 0.43, significantly higher than that of conventional compound 29 (ΦF = 0.10). To mitigate the deactivation effects induced by amino group twisting, Yampolsky et al. developed GFP analogue 32 with a rigid cycloalkylamino group [53]. Compared to the freely twistable compound 31 (ΦF = 0.03), compound 32 exhibits a higher quantum yield of 0.48 in water.

The strategy of inhibiting TICT formation through amino-cyclization is also applicable to rhodamine dyes (Fig. 6). For instance, Fletcher et al. reported that the quantum yield of amino-cyclized rhodamine 101 (34) in ethanol was 0.96 at room temperature, which was significantly higher than that of uncyclized rhodamine B (33, ΦF = 0.65) [54]. Similar results were observed for rhodamine derivative 35, where the donor amine incorporates a phenyl group. The strong electron-donating capability of the phenylamine group in compound 35 facilitates efficient TICT formation, leading to a very low quantum yield (ΦF < 0.01) in methanol. Urano et al. developed locked derivatives 36 and 37 to inhibit the twisting motion of the C-N bond [55]. Compound 36 exhibited a low quantum yield (ΦF = 0.01), whereas compound 37 showed a significantly higher quantum yield (ΦF = 0.43). The significant variation in the photophysical properties of compounds 36 and 37 might be attributed to the differing flexibilities of the six- and five- membered rings. In another study [56], compound 38 with a cyclic structure also displayed high fluorescence quantum yields (ΦF = 0.26). However, while the amino cyclization strategy significantly enhances the brightness of fluorescent dyes, the addition of the alkane portion may reduce their solubility in aqueous environments.

3.2.3. Reducing steric hindrance of amino group

The non-planarity between the electron-donating group of the amino group and the fluorophore scaffold may lead to an increase in the spatial repulsion, leading the molecule to enter the TICT state. Consequently, reducing spatial repulsion is a common practice aimed at diminishing the molecule's propensity to enter the TICT state and enhancing the quantum yield. For example, spatial repulsion with the fluorophore scaffold can be minimized by reducing the spatial volume of the amino electron-donating group (Fig. 7). Foley et al. replaced the traditional dimethylamino group in sulforhodamine dyes with a 7-azabicyclo[2.2.1]heptane [57]. Theoretical calculations indicated that 7-azabicyclo[2.2.1]heptane exhibits lower steric hindrance than N, N-dimethyl. In methanol, compound 40 achieved a quantum yield of 0.95, significantly higher than the typical tetramethyl analog 39 (ΦF = 0.65). While this modification effectively enhanced the brightness of fluorophores, it also resulted in reduced water solubility.

In 2015, Lavis and co-workers demonstrated a general structural modification approach aimed at enhancing the brightness of conventional fluorophores [58]. A pivotal discovery was that replacing an N, N-dimethylamino group with an azetidine moiety in a classical dye significantly increased its fluorescence quantum yield by suppressing TICT. The azetidine-substituted rhodamine 42 (JF₅₄₉) exhibits a quantum yield of 0.88 in aqueous solution, which is significantly higher than that of compound 41 (tetramethylrhodamine (TMR), ΦF = 0.41). In addition, this minor structural modification maintained the cell membrane permeability of the dyes. Lavis et al. compared the super-resolution imaging performance of HaloTag ligands 55 and 56 (Fig. 8). They used 55 and 56 to label the nucleus in live HeLa cells expressing HaloTag-histone 2B (Fig. 8a), and the results showed that 55 exhibited improved brightness and track length compared to 56 (Fig. 8b). Subsequently, they successfully applied this method to a variety of fluorophores, including coumarin, naphthylimine, acridine, rhodol, carbon rhodamine, silicon rhodamine, and oxazine derivatives, resulting in a significant increase in brightness. For example, dimethylamino coumarin 43 exhibits a quantum yield of 0.19, whereas azetidine ring substitution in compound 44 increases the quantum yield to 0.96. Compared with the parent naphthalimide 45 (ε = 9500 L mol⁻¹ cm⁻¹, ΦF < 0.01), the azetidine derivative 46 showed a larger extinction coefficient (ε = 18, 000 L mol⁻¹ cm⁻¹) and a high quantum yield (ΦF = 0.28) in water. Replacing the N, N-dimethylamino group in rhodol 47 (ΦF = 0.21) with azetidine resulted in a 4-fold increase in quantum yield (48, ΦF = 0.85). The azetidine-containing analog 50 (ΦF = 0.24) exhibited a notable increase in quantum yield compared to the classic dye oxazine 49 (ΦF = 0.07).

Inspired by this study, Baranov et al. designed GFP analogues 52 with an azetidinyl ring substituent, which proved effective at enhancing the quantum yield of 51 in water (from 0.05 with 51 to 0.53 with 52) [59]. Miller and Citterio's groups extended this strategy to firefly luciferin analogues, which demonstrated enhanced bioluminescence due to azetidine substitution [60, 61]. Compared to the N, N-dimethylamino compound 53 (ΦF = 0.28 in water), azetidine-substituted compound 54 demonstrated a 2.3-fold increase in fluorescence quantum yield (ΦF = 0.65).

Further functionalization of azetidine groups (e.g., alkyl, methoxy, carboxyl, cyano, fluorine) can be conducted to tune the emission wavelengths and the lactone-zwitterionic equilibrium of rhodamine dyes without affecting the fluorescence quantum yield (such as compounds 57–62, Fig. 9) [62]. Subsequently, Lavis et al. developed a series of fluorine-substituted rhodamine dyes covering the entire visible absorption spectrum [62-64], such as 62 (JF₅₂₅, λ_em = 549 nm, ΦF = 0.91), 63 (JF₅₀₃, λ_em = 529 nm, ΦF = 0.87), 64 (JF₅₈₅, λ_em = 609 nm, ΦF = 0.78), 65 (JF₆₃₅, λ_em = 652 nm, ΦF = 0.56), 66 (JF₄₇₉, λ_em = 517 nm, ΦF = 0.62), 67 (JF₅₅₉, λ_em = 579 nm, ΦF = 0.85) and 68 (JF₇₁₁, λ_em = 732 nm, ΦF = 0.17). These fluorine-substituted JF dyes maintain excellent brightness and demonstrate a high signal-to-noise ratio in "no wash" cellular imaging experiments. Moreover, they used JF₆₃₅-Halo to label living brain tissue from Drosophila larvae. As shown in Fig. 10, JF₆₃₅-Halo displayed consistent labeling throughout the living tissue with low background staining. Lavis and co-workers also applied the JF dyes (such as JF₅₂₅) for in vivo voltage imaging [65].

Tor et al. replaced the N, N-dimethylamino group with a 3, 3-difluoroazetidine, which dramatically increased the quantum yield of thieno purine analogues (Fig. 11) [66, 67]. Compound 71 exhibited a quantum yield of 0.64 in water, about tenfold greater than that of compound 70 (ΦF = 0.06) and 1280-fold higher than that of compound 69 (ΦF = 0.0005). Zhao et al. reported a method to enhance the performance of D-A fluorophores by integrating azetidine-containing heterospirocycles into their molecular scaffolds [68]. For instance, the dimethylamino derivative 72 exhibited minimal fluorescence quantum yield in aqueous conditions (ΦF < 0.01). In contrast, the azetidine-modified naphthalimide 73 displayed a moderate quantum yield of 0.19. The spirocyclic naphthalimide 74 demonstrated stronger emission in aqueous environments (ΦF = 0.42). They successfully extended this strategy to a diverse array of fluorophores, including rhodamine, rhodol, coumarin, NBD, and oxazone derivatives, resulting in a substantial enhancement in brightness. In 2016, Xu et al. investigated the impact of aziridine on TICT in the naphthimide system [69]. Aziridinyl dye 77 exhibited a quantum yield of 0.708 in EtOH, which was higher than that of azetidinyl dye 76 (ΦF = 0.631) and the dimethylamino derivative 75 (ΦF = 0.009). They also applied the method of incorporating aziridine into various fluorophores, such as coumarin, phthalimide, and NBD, thereby enhancing the quantum yields.

Therefore, in more polar solvents, reducing the ring size resulted in a significant increase in quantum yield. However, the degree to which the azetidine approach improves quantum yield varies depending on the fluorophore matrices. Some studies have indicated that replacing N, N-dimethylamino with azetidine does not enhance the brightness of fluorophores [70-72]. Furthermore, aziridines are susceptible to ring cleavage reactions due to nucleophilic attack, compromising their chemical stability [73].

3.2.4. Reducing the electron-donating capacity of auxochromes

Another factor that influences the formation of TICT states is electronic effects. This is primarily because the formation of TICT states relies on complete charge separation between the electron donor and acceptor. Therefore, reducing the electron-donating strength of the electron donor would destabilize the TICT state, increasing the energy barrier to TICT formation and thereby suppressing it.

In 2019, Xiao and coworkers developed a series of quaternary piperazine-substituted dyes with enhanced brightness (Fig. 12) [74]. Due to the positive charge of piperazine salts, the electron-donating ability of the amino group was significantly weakened, making it difficult for the dyes to enter the TICT state. The quantum yield of the piperazine-substituted rhodamine 79 (ΦF = 0.93 in water) was nearly 30 times higher than that of the parent dye 78 (ΦF = 0.032) and showed an approximate 2-fold increase compared to TMR (41). Subsequently, they developed a probe Mem-R to specifically label plasma membranes, allowing for effective use in super-resolution imaging experiments in live cells (Fig. 13a). They also utilized this strategy to improve the quantum yields of the naphthylimide and NBD derivatives. At the same time, Guo et al. designed a series of sulfone-substituted fluorophores employing a similar strategy [75]. The negative inductive effect of the sulfone group significantly reduced the electron-donating ability of the amino group and increased the ion potentials of the molecules, effectively inhibiting the TICT process. In comparison to piperidine derivatives 80 (ΦF = 0.06), sulfone-substituted rhodamine 81 exhibited a higher quantum yield of 0.99 in water. This strategy was also employed for carbon rhodamine, silicon rhodamine, phosphor rhodamine, oxazine, coumarin, naphthylamide, and BODIPY derivatives [76], resulting in enhanced brightness.

Lavis and coworkers proposed a comprehensive method to enhance the brightness and photostability of fluorophores by substituting hydrogen (H) atoms in N-alkyl groups with deuterium (D) [77]. Deuteration reduces the electron donor strength of the auxochrome, thereby reducing the effectiveness of the TICT process. For example, deuterated pyrrolidine dye 83 (ΦF = 0.8) exhibited a higher quantum yield and greater photostability compared to parent dye 82 (ΦF = 0.7). They successfully applied this approach to a wide range of fluorophores, including coumarin, phenoxazine, carbon rhodamine, and silicon rhodamine, resulting in enhanced brightness. Furthermore, Broichhagen et al. reported that a deuterium congener of tetramethyl(silicon)rhodamine 85 (ΦF = 0.46, ɛ×ΦF = 71, 760) exhibited a higher brightness compared to the parent dye 84 (ΦF = 0.35, ɛ×ΦF = 49, 350) in PBS [78]. Recently, Stacko and co-workers utilized this strategy to improve the brightness of heptamethine cyanine dyes [79]. In comparison to compound 86 (ΦF = 1.33% in CH₂Cl₂, ɛ×ΦF = 3, 730), deuterium-substituted cyanine 87 displayed higher brightness (ΦF = 2.39% in CH₂Cl₂, ɛ×ΦF = 7, 080). In 2018, our group designed and synthesized asymmetric rhodamines incorporating 1, 4-diethyl-decahydroquinoxaline. These novel dyes exhibited an enhanced Stokes shift (> 80 nm), but their quantum yields significantly decreased [80]. Recently, our group introduced a stronger electron-withdrawing ability of quinoxaline in these dyes, resulting in higher quantum yield [81]. For example, compound 89 (YL578) exhibited a 7.4-fold increase in quantum yield (ΦF = 0.74) in aqueous solution compared to its parent fluorophore 88 (ΦF = 0.1). The YL578-based probe YL578-Halo could be used for 3D STED imaging of whole mitochondria (Fig. 13b). This strategy could also be extended to other dye skeletons, such as coumarin, rhodol, and acridine derivatives, resulting in a significant increase in brightness. However, it is worth mentioning that the emission wavelengths of these dyes (such as 79, 81, 89) are blue-shifted due to the reduced electron-donating capacity of the auxochrome compared to their parent compounds.

3.2.5. Minimizing solvent–solute interactions

Chemists have also investigated the interaction between dyes and solvents to reduce the TICT effect and enhance brightness. Many D-A dyes possess a dipole structure, which facilitates dipole-dipole interactions with polar solvents that stabilize TICT states. Therefore, minimizing the interaction between dyes and solvents is essential to inhibit TICT state formation (Fig. 14). For example, Ahn et al. synthesized a series of acedan derivatives by modifying the auxochrome of acedan dyes [82]. The aminocyclohexanol-substituted molecule 91 exhibited a quantum yield of 0.40 in water, significantly higher than that of the N, N-dimethyl-substituted compound 90 (ΦF = 0.20). The authors attributed this finding to the cyclohexanol substituent, which prevents hydrogen bonding between the amine nitrogen and water molecules, effectively inhibiting the formation of TICT. Furthermore, they applied this approach to other popular dipolar dyes, including naphthylimide, coumarin, and NBD, successfully increasing their quantum yields. In 2019, Zhang's group reduced the solvent interactions of the amino group by introducing a β-carbonyl-based polar substituent [83]. Compared to N, N-dimethylamino NBD 92 (ΦF = 0.02), the β-carbonyl-substituted NBD 93 displayed a nine-fold increase in fluorescence quantum yield (ΦF = 0.18 in ethanol). The group found that other β-carbonyl-substituted D-A dyes, such as naphthalimide, coumarin, and PRODAN, also exhibited higher quantum yields and brightness.

4. Enhancing the brightness of dyes in water

Organic fluorophores often possess highly conjugated structures, which render them hydrophobic. In aqueous phases, organic dyes are prone to aggregation due to π-π stacking interactions, leading to nonradiative decay of the excited state and subsequently quenched fluorescence. Typically, the addition of one or more hydrophilic groups to a dye through chemical modification can significantly increase its water solubility. Consequently, the brightness of these dyes can be enhanced in aqueous environments. Several comprehensive reviews have systematically summarized recent advances in improving the water solubility of dyes for biological imaging [84, 85]. Furthermore, in contrast to conventional dyes that showed quenched fluorescence in water, the novel aggregation-induced emission luminogens (AIEgens) exhibited remarkable brightness. Since 2001 [86], the application of AIEgens has been significantly advanced by Tang's group and has been comprehensively reviewed by numerous reviews [87-90]. Therefore, we will focus on additional strategies for enhancing dye brightness in water, including the use of hydrogen bond networks, the introduction of shielding groups, the formation of supramolecular host-guest complexes, and protein encapsulation.

4.1. Hydrogen-bond-induced enhanced emission (HIEE)

Utilizing hydrogen bond networks between dyes and polar solvents can enhance dye brightness in certain situations (Fig. 15) [91]. For example, dye 94 exhibited fluorescence quantum yields of 0.80 in PBS, 0.53 in EtOH, 0.12 in MeCN, and 0.02 in CH₂Cl₂, demonstrating 113-fold brighter emission in PBS compared to CH₂Cl₂. Furthermore, dye 94 showed significant fluorescence turn-on as the PBS ratio increased (Fig. 15). According to DFT studies, hydrogen bond networks can restrict intramolecular rotation, minimizing non-radiative decay and thereby increasing quantum yields in aqueous environments. This phenomenon is known as HIEE.

4.2. Introducing shielding groups

Heptamethine cyanine dyes have large hydrophobic surface areas, which promote aggregation in aqueous environments and thus significantly reduce brightness. For example, dye 95 (ɛ×ΦF = 9600 in PBS) exhibits moderate brightness due to its rigid hydrophobic core (Fig. 16). To improve the performance of dye 95 in aqueous environments, Smith et al. designed a sterically shielded dye 96 (ɛ×ΦF = 18, 000) with two shielding arms, which enhance both photostability and brightness [92]. Subsequently, they synthesized dye 98 with constrained flanking straps, which exhibited a quantum yield of 0.11 in PBS, higher than that of the unstrapped dye 97 (ΦF = 0.081) [93]. The results indicate that the restricted flanking straps in dye 98 are more effective at inhibiting nonradiative decay of the excited state.

4.3. Forming supramolecular host-guest complexes

Macrocyclic structures can serve as molecular containers for fluorescent dyes, facilitating the formation of host-guest complexes. The geometric confinement of fluorophores within the host can modify their photophysical characteristics by restricting intramolecular rotation and thus inhibiting nonradiative decay processes. Dyes enclosed in a rigid, complementary molecular container can reduce aggregation and protect against external quenchers, resulting in increased brightness in aqueous conditions [94]. In a recent study, Smith et al. employed dye 99 encapsulated in cucurbit[7]uril (CB7) to increase the quantum yield from 0.12% to 0.23% in water (Fig. 17) [95]. Ge and co-workers also reported host–guest complexation between an auxochrome and CB7 to enhance the fluorescence quantum yield of dyes in water [96]. For example, the quantum yields of complexes 100@CB7 (ΦF = 0.62), 101@CB7 (ΦF = 0.15), and 102@CB7 (ΦF = 0.149) were significantly higher than those of the free dyes 100 (ΦF = 0.22), 101 (ΦF = not detected (n.d.)), and 102 (ΦF = 0.057), respectively.

4.4. Encapsulating dyes within proteins

The protein also provided a confined environment for the dye, inhibiting intramolecular mobility and resulting in intense fluorescence (Fig. 18) [3]. GFP is a well-known example of a fluorophore enclosed within a protein that emits light when excited at specific wavelengths. Another example is stilbene derivative 103 (ΦF = 0.02), which showed very low fluorescence in solution due to efficient nonradiative decay through cis-trans isomerization [97]. In striking contrast, the antibody-stilbene complex 103@mAb display higher fluorescence quantum yield (ΦF = 0.28–0.78) in PBS. In 2008, Waggoner et al. developed fluorogen-activating proteins (FAPs), which enhance the fluorescence emission performance of fluorophores [98]. For instance, compounds 104 (TO1-2p) and 105 (MG-2p) exhibited no fluorescence in PBS. Upon binding to FAPs, the dye complexes 104@FAP (ΦF = 0.47) and 105@FAP (ΦF = 0.25) displayed strong fluorescence. Gautier et al. also reported fluorescence-activating and absorption-shifting tags (FASTs), which lock the fluorophore into a planar conformation, resulting in high fluorescence [99, 100]. The rhodanine derivatives 106 (4-hydroxybenzylidene-rhodanine (HBR), ΦF = 0.0002) and 107 (4-hydroxy-3-methylbenzylidene-rhodanine (HMBR), ΦF = 0.0004) showed low fluorescence in solution, while dye complexes 106@FAST (ΦF = 0.09) and 107@FAST (ΦF = 0.33) exhibited higher fluorescence. In 2017, Mishin et al. reported a protein capable of binding dye 108 (M739, ΦF = 0.035), which shows a remarkable enhancement in fluorescence upon binding to the protein (108@DiB1, ΦF = 0.32) [101].

NIR-Ⅱ fluorescence imaging reduces scattering coefficients in tissue at long wavelengths, making it ideal for deep tissue imaging [102-105]. However, the majority of NIR-Ⅱ cyanine fluorophores have poor quantum yields in aqueous environments. To address this issue, researchers employed proteins (such as fetal bovine serum (FBS) [106, 107], bovine serum albumin (BSA) [108], β-lactoglobulin (β-LG) [109], and human serum albumin (HSA) [110]) complexed with cyanine dyes to produce highly bright NIR-Ⅱ cyanine@albumin fluorophores [111]. These studies demonstrate that encapsulating dyes within proteins can effectively improve the dye brightness.

5. Conclusions and future outlook

Small molecule fluorophores play a crucial role in biology and biochemistry. For most biological applications, the ideal fluorophore would exhibit high brightness, emit in the red to NIR range, possess high photostability, and offer good biocompatibility. This review summarizes the recent efforts and significant progress made by chemists in enhancing the brightness of organic small-molecule fluorescent dyes by increasing extinction coefficients and quantum yields, including scaffold rigidification and inhibition of TICT (such as cyclizing the flexible amine group, modifying steric hindrance, utilizing inductive effects and minimizing solvent–solute interactions). These strategies effectively improve the brightness of various small-molecule fluorophores, such as rhodamine, rhodol, coumarin, BODIPY, naphthylimine, NBD, and oxazine. However, enhancing the brightness of fluorescent dyes with NIR emission (NIR-Ⅰ and NIR-Ⅱ) remains challenging. Additionally, optimizing brightness often leads to low solubility [57], poor photostability [112], and limited cellular permeability [74, 113]. Therefore, future studies should focus on enhancing the brightness of dyes emitting in the NIR region while maintaining the biocompatibility of fluorophores.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

CRediT authorship contribution statement

Junliang Zhou: Writing – original draft, Investigation, Conceptualization. Tian-Bing Ren: Writing – review & editing, Investigation, Conceptualization. Lin Yuan: Writing – review & editing, Supervision, Conceptualization.

Acknowledgments

This work was supported by the National Science Foundation for Distinguished Young Scholars (No. 22325401), the National Natural Science Foundation of China (No. 22404049), and the China Postdoctoral Science Foundation (No. 2024M750866).