Lipid droplets (LDs) are ubiquitous and highly dynamic intracellular organelles that function as central hubs for lipid synthesis [
1], storage [
2–
4], and mobilization. Beyond serving as inert lipid depots, LDs actively participate in cellular energy homeostasis, stress adaptation, signaling regulation, and organelle communication [
5–
8]. The functional state of LDs is fundamentally dictated by compositional remodeling within their neutral lipid core, primarily composed of triacylglycerols (TAGs) and cholesteryl esters (CEs) [
9,
10]. Variations in lipid species, degrees of unsaturation, and oxidative modification accompany metabolic reprogramming, lipotoxic stress, and cell fate transitions [
11]. Notably, the CE/TAG ratio has emerged as an important indicator for disease staging and prognosis, further highlighting the pathological relevance of LD compositional dynamics [
12,
13]. These compositional changes also directly reflect shifts in lipid metabolic flux and organelle function, underscoring the importance of resolving LD composition
in situ and in real time [
14–
16].
Traditionally, LD compositional analysis relies on liquid chromatography–tandem mass spectrometry (LC–MS/MS)-based lipidomics [
17–
19], which enables highly accurate quantification of purified lipid extracts at cellular or tissue levels and precise identification of individual molecular species. However, such approaches are destructive and endpoint measurements, making it difficult to capture dynamic metabolic remodeling in living cells. More importantly, lipidomics provides population-averaged information, whereas a single cell may contain tens to hundreds of LDs with potential heterogeneity [
20]. Consequently, compositional and physicochemical variations within individual LDs remain largely inaccessible. Fluorescence imaging offers a noninvasive strategy to visualize LD structure and dynamics under physiological conditions [
21–
24]. In particular, super-resolution techniques allow nanoscale visualization of LD fusion, lipophagy, and inter-organelle interactions. Classical LD dyes, including boron-dipyrromethene (BODIPY) 493 and Nile Red [
25–
27], have realized quantification of LD number, size, and morphology [
28–
32]. However, these probes primarily report structural features and provide limited insight into real-time compositional remodeling within individual droplets, thereby restricting functional interrogation of LD metabolic heterogeneity.
Microenvironmental polarity is intrinsically governed by molecular composition. Therefore, LD polarity serves as a physicochemical surrogate for internal lipid remodeling. Lipid exchange, oxidation, and saturation changes during metabolic perturbation or mitochondria-LD interactions generate subtle dielectric variations within the LD core. Thus, spatiotemporal mapping of LD polarity offers an indirect yet powerful strategy to decode compositional dynamics and associated biological functions [
33–
38]. However, the LD interior presents a unique challenge that it is uniformly weakly polar and lacks hydrogen-bonding interactions, resulting in a compressed polarity window. Even substantial compositional remodeling produces only minor dielectric changes. Conventional solvatochromic fluorophores, which typically rely on strong intramolecular charge transfer (ICT) responses triggered by large polarity gradients or hydrogen bonding (
Fig. 1a), exhibit attenuated or indistinguishable emission shifts under these conditions [
39–
42]. Therefore, probes capable of resolving small polarity differences with high sensitivity in weakly polar environments are urgently needed.
Here, we introduce a phenylethynyl-bridged naphthalimide probe (TS-N) that achieves enhanced solvatochromic responsiveness (
Fig. 1b). By inserting a conjugated phenylethynyl spacer between the electron-donating amino group and naphthalimide acceptor core, the charge-transfer distance (
dCT = 3.561 Å) is increased to approximately 2.5-fold relative to the non-bridged analogue, thereby amplifying polarity sensitivity. TS-N exhibits an emission shift spanning over 220 nm across solvent polarity gradients, with a relative polarity-dependent emission shift rate of 14.5 nm, enabling high-resolution decoding of subtle dielectric variations. Importantly, TS-N can precisely target LDs and enable super-resolution imaging of LD dynamic processes, including fusion events and mitochondria-LD interactions. Quantitative
in situ analysis demonstrates that TS-N can resolve metabolic perturbation-induced polarity fluctuations at the single-droplet level while correlating dielectric changes with droplet size remodeling. This probe holds strong potential to integrate structural imaging with compositional decoding, thereby enabling functional interrogation of LD metabolic heterogeneity in living cells.
We first investigated the photophysical properties of the phenylethynyl-bridged derivatives. Both TS-O and TS-N exhibited pronounced solvatochromic behavior in different solvents (
Figs. 2a and b, Figs. S1-S3 in Supporting information). Notably, TS-N displayed an exceptionally large emission shift spanning over 220 nm across the solvent series, with maximum λ
em located at 513 nm in hexane and red-shifted to 736 nm in methanol (
Fig. 2c). In contrast, its absorption spectra remained nearly unchanged in different solvents, indicating that the polarity response predominantly arises from excited-state modulation rather than ground-state perturbation. By comparison, the electron-deficient derivative TS-CN showed minimal polarity responsiveness, with almost invariant fluorescence emission across solvents, confirming the essential role of the donor–acceptor framework in enabling solvatochromism (Fig. S4 in Supporting information). We also evaluated Nile Red and a reference naphthalimide derivative lacking the phenylethynyl bridge (4-dimethylamino-1,8-naphthalimide (DMN)) under identical solvent conditions (Figs. S5 and S6 in Supporting information). For quantitative comparison, we calculated the average emission shift per solvent polarity increment. TS-N and TS-O exhibited sensitivity values of 14.5 nm and 7.0 nm, respectively, whereas Nile Red displayed only 4.4 nm and DMN merely 3.5 nm (
Fig. 2d and Fig. S7 in Supporting information). These results clearly demonstrate that incorporation of the phenylethynyl bridge dramatically enhances polarity responsiveness, endowing TS-N and TS-O with superior capability for detecting subtle dielectric variations.
To elucidate the origin of the enhanced polarity sensitivity, we calculated the charge redistribution between the ground and excited states using DMN as the reference structure (
Fig. 2e). The degree of ICT was quantified using the charge transfer distance (
dCT), defined as the spatial separation between the barycenters of positive (Δρ
+) and negative (Δρ
−) electron density increments upon excitation [
43]. The calculated
dCT value for TS-N (3.561 Å) was approximately 2.5 times larger than that of DMN (1.438 Å), indicating a significantly enhanced ICT character in the S
1 state of TS-N. This value was also considerably larger than those reported for Nile Red (0.946 Å) and BODIPY dyes (0.445 Å) [
44]. Furthermore, TS-N exhibited a substantially higher oscillator strength (
f) than the reference compound, corresponding to enhanced absorption capability. This theoretical prediction agreed well with experimental data that TS-N showed a molar extinction coefficient of ~25,000 L mol
−1 cm
−1, nearly double that of DMN (~12,000 L mol
−1 cm
−1). Compared with TS-O and TS-CN, TS-N displays larger
dCT and
f values, further confirming that the phenylethynyl unit functioned as an efficient charge-transfer bridge that increases the donor–acceptor separation and amplifies excited-state polarization. The resulting enhanced charge-transfer character increased the dielectric sensitivity of the excited state, thereby leading to a more pronounced emission response to polarity variations.
To further quantify polarity-related electronic redistribution, we evaluated the molecular polarity index (MPI) [
45], derived from the surface electrostatic potential distribution (
Fig. 2f). Although TS-N showed a relatively lower MPI in the ground state (13.2 kcal/mol) than DMN, its excited-state MPI markedly increased to 27.4 kcal/mol, surpassing that of DMN (23.0 kcal/mol). This substantial excited-state polarity amplification further explained the pronounced solvatochromic behavior of TS-N and confirms that phenylethynyl-mediated conjugation engineering enhanced sensitivity to subtle environmental polarity changes. Such characteristics made TS-N particularly suitable for probing weakly polar microenvironments with narrow dielectric variation ranges.
We then assessed the LD targeting capability of TS-N in live cells. Colocalization experiments were performed using our previously reported probe LD 688P (
Figs. 3a and b) [
46]. TS-N displayed specific accumulation within LDs, exhibiting a Pearson’s correlation coefficient exceeding 0.88 with LD 688P, thereby confirming its excellent LD selectivity (Fig. S8 in Supporting information). For long-term visualization of LD dynamics, high photostability is essential. Environment-sensitive fluorophores frequently suffer from rapid photobleaching, as extended and flexible
π-conjugated systems are prone to singlet oxygen-mediated oxidative degradation under continuous irradiation. Given that TS-N incorporates a phenylethynyl spacer between the amino donor and the naphthalimide acceptor, we anticipated that the increased structural rigidity and extended conjugation would enhance photostability by reducing conformational flexibility. To evaluate this hypothesis, we examined the photostability of TS-N under confocal imaging by repeatedly acquiring fluorescence images from the identical cellular region under continuous 488 nm excitation (
Fig. 3c). For comparison, the commercial LD dye BODIPY 493 was tested under identical conditions (
Fig. 3d). HeLa cells were incubated with TS-N (2 µmol/L) for 0.5 h prior to imaging. Upon continuous irradiation, TS-N retained over 83% of its initial fluorescence intensity, whereas BODIPY 493 exhibited faster photobleaching, with intensity decreasing to below 70%. Such enhanced photostability provided a reliable foundation for capturing LD fusion, morphological remodeling, and organelle interactions in living cells. Notably, TS-O also demonstrated precise LD localization in live cells (Fig. S9 in Supporting information), highlighting the general applicability of this phenylethynyl-engineered design strategy for LD targeting.
Leveraging its high photostability and enhanced environmental sensitivity, TS-N was employed for long-term super-resolution imaging of LD dynamics in living cells. In LD-enriched regions, multiple fusion events were observed within the first 10 min (
Fig. 4a, white arrows), with fusion occurring frequently during this initial period (
Fig. 4b). Interestingly, although small LDs-presumably characterized by higher surface tension-remained in close proximity thereafter, no additional fusion events were detected during the subsequent 30 min.
Notably, we captured a complete LD fusion and subsequent morphological relaxation process. Between 540 s and 640 s, two adjacent LDs underwent coalescence (
Fig. 4c), forming a newly generated LD with an irregular morphology. Merging super-resolution images at 640 s and 660 s revealed a contracted green region (dotted frame), demonstrating gradual morphological relaxation. This fusion-induced rounding process was consistent with surface tension-driven minimization of interfacial free energy, whereby the newly formed droplet might undergo shape relaxation to reach a thermodynamically favorable spherical configuration.
Long-term super-resolution imaging further enabled visualization of dynamic mitochondria-LD interactions (
Figs. 4d and e), revealing both sustained and transient contacts at the nanoscale. In one example, the LD indicated by the white arrow maintained tight contact with mitochondria for over 8 min, suggesting prolonged metabolic coupling. In another instance, a mitochondrion extended a protrusion toward a neighboring LD at 260 s (blue arrow), followed by gradual retraction at 360 s, representing typical transient contact. Additional observations revealed an LD, labelled with orange arrow, serving as a nodal point for three linear mitochondria, with simultaneous interactions occurring between 40 s and 80 s. In certain cases, LDs appeared partially wrapped by mitochondria (
Fig. 4e, white arrow), which was further confirmed by intensity profile analysis (
Figs. 4f and g). Furthermore, the orange-labeled LD gradually migrated toward the center of a ring-like mitochondrial network at 180 s, potentially increasing contact area and facilitating lipid transfer. These observations highlight the heterogeneous and dynamic nature of mitochondria-LD interactions and underscore the ability of TS-N to resolve nanoscale organelle crosstalk in real time.
To assess the capability of TS-N in reporting real-time LD polarity changes, cells were subjected to starvation, oleic acid (OA) supplementation, and deferoxamine (DFO) treatment (
Fig. 5a). Starvation significantly reduced LD numbers, whereas OA supplementation increased LD abundance. DFO-treated cells maintained LD numbers comparable to control group.
In situ fluorescence spectral analysis of >600 LDs per group revealed that starvation slightly increased LD polarity, with the average emission wavelength red-shifting from 550 nm (control) to 554 nm, accompanied by a modest size increase from 0.31 µm
2 to 0.33 µm
2 (
Figs. 5b and c). Here, LD size was quantified using projected 2D area. In contrast, OA treatment resulted in substantial LD enlargement (0.40 µm
2) with only a minor polarity decrease (549 nm), consistent with neutral lipid accumulation and increased hydrophobicity. Similarly, DFO treatment induced LD enlargement (0.39 µm
2) while maintaining polarity levels close to control group, suggesting lipid accumulation with reduced oxidative modification. Correlation analysis between LD size and polarity (
Figs. 5d and e) revealed pronounced heterogeneity in control cells (0.2–0.6 µm
2). OA-treated cells exhibited increased dispersion and reduced uniformity. In contrast, LDs in DFO-treated cells appeared more homogeneous and tightly distributed, possibly due to suppression of lipid peroxidation and reduced compositional variability within the droplets. Collectively, these results demonstrate that TS-N not only resolves subtle polarity fluctuations within individual LDs but also simultaneously enables super-resolution visualization of LD morphological remodeling and organelle interactions, providing a powerful tool for functional interrogation of lipid metabolic heterogeneity in living cells.
In conclusion, we had developed a phenylethynyl-bridged naphthalimide probe (TS-N) that enables sensitive visualization of LD polarity and morphological dynamics in living cells. Incorporation of a conjugated phenylethynyl spacer between the donor and acceptor units significantly extends the intramolecular charge-transfer distance, resulting in enhanced excited-state polarization and amplified solvatochromic responsiveness. Consequently, TS-N exhibits a large emission shift across solvent polarity gradients and high polarity sensitivity, enabling detection of subtle dielectric variations within the weakly polar interiors of LDs. In addition, TS-N displays excellent LD selectivity and high photostability, supporting long-term super-resolution imaging of LD dynamic processes, including droplet fusion, morphological relaxation, and mitochondria-LD interactions. In situ fluorescence spectra further demonstrate that TS-N can resolve metabolic perturbation-induced polarity fluctuations at the single-droplet level and uncover correlations between droplet polarity and size remodeling under different metabolic conditions. The phenylethynyl-bridged design strategy presented here provides a general framework for developing highly sensitive probes for weakly polar biological microenvironments, thereby facilitating mechanistic investigations of lipid metabolism, organelle communication, and metabolic heterogeneity in living cells.
PDF
Full
81
45
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
