Real-time tracking of ER turnover during ERLAD by a rhenium complex via lifetime imaging

Abstract Endoplasmic reticulum (ER) degradation by autophagy (ER-phagy) is a recently revealed selective autophagy pathway that plays important roles in organelle turnover and protein degradation, but the biological functions of ER-phagy are largely unknown. Here, we present an ER-targeting Re(I) tricarbonyl complex (Re-ERLAD) that can accumulate in the ER, induce ER-to-lysosome-associated degradation (ERLAD) upon visible light irradiation, and label ER buds and track their morphological alterations during ER-phagy. The emission of Re-ERLAD is sensitive to viscosity, which is a key parameter reflecting the amount of unfolded protein in the ER. Quantitative detection using two-photon fluorescence lifetime imaging microscopy shows that ER viscosity initially increases and then decreases during ERLAD, which reveals that ERLAD is a pathway for alleviating ER stress caused by unfolded proteins. In conclusion, our work presents the first specific photoinducer and tracker of ERLAD, which can be used in studying the regulatory mechanism and function of this process.


INTRODUCTION
The endoplasmic reticulum (ER) plays important roles in protein and lipid synthesis, ion homeostasis and organelle communication [1]. To protect cells from both endogenous and exogenous threats, e.g. unfolded proteins [2] or reactive oxygen species (ROS) [3], the ER has developed complicated selfmodulation mechanisms [4][5][6]. Depending on the receptors involved and substrates degraded, at least three subtypes of ER degradation by autophagy (ER-phagy) have been reported in mammalian cells: macro ER-phagy, ER-to-lysosome-associated degradation (ERLAD) by micro ER-phagy and ER-phagy-related ERLAD [7]. Compared with ER-associated degradation (ERAD) [8], ERLAD provides a proteasome-free degradation pathway for protein quality control, which is crucial for ER turnover and cellular homeostasis [9][10][11]. However, very limited tools, including specific inducers and imaging agents, are currently available for the investigation of the regulatory mechanisms and functions of ER-phagy.
Viscosity is an important micro-environmental factor reflecting the functionalities of organelles, including mitochondria, lysosomes and the ER. Viscosity is closely related to intracellular diffusion, molecular interactions and membrane fluidity, which are crucial for biological interactions and biochemical reactions. Intracellular viscosity can be reflected by molecular rotors, whose emission is quenched by intramolecular rotation and recovers upon an increase in environmental viscosity [12][13][14]. The lifetime of a fluorophore does not depend on its concentration; however, it is sensitive to its environment [14]. Using fluorescence lifetime imaging microscopy (FLIM), environmental parameters, e.g. viscosity and pH, can be measured quantitatively with high accuracy [15][16][17]. Via exploitation of their intrinsic subcellular localization properties or modification with targeting groups, these molecular rotors can be used to measure the viscosity of subcellular organelles, e.g. mitochondria [16], lysosomes [15] and the ER [12].
To date, there have been few reports on the monitoring of dynamic turnover behavior at the subcellular level, which is particularly important for the maintenance of cell homeostasis. In this work, we attached the molecular rotor BODIPY [46,47] to the organometallic rhenium tricarbonyl moiety to produce an ER-targeting phototheranostic agent (Re-ERLAD) with a viscosity-sensitive emission property and lifetime (Fig. 1A). The attachment of the rhenium moiety to BODIPY alters its cellular localization and significantly increases its capability to generate singlet oxygen ( 1 O 2 ) upon irradiation due to the heavy atom effect. Re-ERLAD can induce ERLAD upon photoinitiation, specifically imaging ER buds (a special microstructure formed during ERLAD) and quantitatively tracking the viscosity parameters in ER buds during ER-phagy via twophoton fluorescence lifetime imaging microscopy (TPFLIM). The light-activated property makes Re-ERLAD more controllable and widely applicable than previously reported theranostic agents [18,19], since it can be used as a simple probe to measure the changes in ER viscosity initiated by other stimuli. Overall, we present here the first report on a specific photoinducer and viscosity tracker of ER-LAD, which shows the dynamic changes in microenvironments during ER turnover.
In luminescence with a maximum at ca. 520 nm, and Re-ERLAD shows a broader emission band centered at ca. 530 nm (Fig. S8). The quantum yield of Re-ERLAD is lower than that of py-BODIPY, which may be attributed to the energy transfer from the BODIPY ligand to the metal center [48]. The photostability of Re-ERLAD under physiological conditions was also verified by HPLC (Fig. S9). Re-ERLAD remains stable after light irradiation, which indicates that Re-ERLAD is suitable for long-term imaging in living cells.

Viscosity response of Re-ERLAD
The fluorescence response of Re-ERLAD to environmental viscosity was measured in a mixed methanol-glycerol system representing different viscosities. Re-ERLAD exhibits ca. 8-fold emission enhancement in high-viscosity media compared with low-viscosity media (Fig. 1C). The emission recovery that can be recognized by the naked eye is ascribed to the restricted rotation of py-BODIPY.
In addition, the fluorescence lifetime of Re-ERLAD increases from 1.5 ns to 7.9 ns and shows a linear correlation with the viscosity parameters (Fig. 1D). Moreover, the emission intensity and lifetime of Re-ERLAD show negligible responses to polarity, common cations, glutathione and human serum albumin, which indicates the specificity of the fluorescence response of Re-ERLAD to viscosity (Figs S10 and S11). Disrupted protein folding and accumulation of unfolded proteins can cause oxidative damage and ER stress [49,50]. The impact of unfolded proteins on the solution viscosity was investigated by using β-lactoglobulin as a model protein and urea as the unfolding agent. As expected, the mobility of the protein solution significantly decreased after treatment with urea for 24 h, and the lifetime of Re-ERLAD was obviously enhanced in the solution containing unfolded proteins (Fig. 1E). The results indicate that the changes in environmental viscosity are correlated with the degree of protein unfolding, which can be reflected by the emission intensity and lifetime of Re-ERLAD.

Cytotoxicity and cellular localization of Re-ERLAD
The in vitro cytotoxicities of py-BODIPY and Re-ERLAD were evaluated in A549 (human lung adenocarcinoma) cells after 48 h of incubation (Fig. S12). Re-ERLAD shows low cytotoxicity in the dark, while it exhibits obviously increased cytotoxicity after irradiation with a 450 nm laser for 15 min. In contrast, py-BODIPY is non-cytotoxic both in the dark and under irradiation. Under 450 nm light irradiation, the quantum yields of 1 O 2 photosensitization for py-BODIPY and Re-ERLAD are 0.07 and 0.28, respectively. The higher photosensitization capability of Re-ERLAD can be ascribed to the heavy atom effect of the metal center [51]. Compared to the control group, the emission of the ROS probe significantly increased by ∼17-fold in A549 cells treated with Re-ERLAD and irradiated with a 450 nm laser for 15 min (Fig. S13).
Many organometallic rhenium tricarbonyl complexes have been reported to accumulate in mitochondria [33]; however, colocalization studies have revealed that the positively charged Re-ERLAD localizes in the ER. The colocalization coefficient of Re-ERLAD with the ER (0.97; Fig. S14A) was much higher than that of Re-ERLAD with mitochondria (0.37; Fig. S14B). The neutral hydrophobic molecule py-BODIPY shows a high colocalization coefficient (0.89) with the lipid droplet-specific fluorescent dye (Fig. S14C). These results suggest that the increased photocytotoxicity of Re-ERLAD compared with py-BODIPY is attributed to its ability to produce large amounts of cellular ROS and may also be associated with its subcellular localization properties.

Re-ERLAD initiates ERLAD
Re-ERLAD colocalizes with ER-Tracker and shows a weaker emission intensity in the dark. After irradiation with a 450 nm laser for 15 min, however, highly emissive punctate areas that colocalized with ER-Tracker were observed (Fig. 2A). These structures show minimal colocalization with fluorescent probes that stain punctate subcellular organelles, including lipid droplets and lysosomes (Fig. S15). We speculate that these punctate areas are likely to be ER buds grown from the reticular structure of the ER [11], and the strong emission may be attributed to the accumulation of unfolded proteins caused by light-induced oxidative damage.
To confirm what these highly emissive structures are, transmission electron microscopy (TEM) was used to observe the morphological alterations in the ultrastructure of A549 cells. Compared with the control group, ER fragments and vesicles with monolayer structures and contents inside were found in cells treated with Re-ERLAD in the presence of light (Fig. 2B). In addition, ER buds with a monolayer membrane budding from reticular structures were also observed. These structural changes are consistent with the phenomenon described for ER-LAD, which forms monolayer vesicles rather than bilayer autophagosomes to fuse with lysosomes [7]. Time-dependent upregulation of the key ER-phagy receptor FAM134B, which is involved in ERLAD by generating high membrane curvature [52], was further confirmed by western blotting (Fig. 2C). Similar results were also obtained for the positive controls rapamycin and tunicamycin, which are reported to induce ER stress and ER-phagy in the literature (Fig. S16) [9]. In addition, the interaction between FAM134B and LC3B, two key proteins regulating the process of ERLAD, was confirmed with immunofluorescence double staining (Fig. 2D), and the overlap coefficient increased from 0.01 (dark) to 0.31 (light). The interactions of FAM134B and LC3B were also detected by Airyscan super-resolution imaging techniques (Fig. 2E). Collectively, these results indicate that Re-ERLAD can specifically initiate the ERLAD process upon irradiation.
Then, we used Re-ERLAD to track ER-associated morphological events during ERLAD. Before irradiation, the emission of Re-ERLAD was very weak, and there was no obvious colocalization between  Re-ERLAD and LC3B. After irradiation, the punctate emission of Re-ERLAD was greatly enhanced, and colocalization of Re-ERLAD with punctate LC3B was detected, which showed that ER buds were formed and began to recruit LC3B (Fig. 3A). At the same time, with prolonged incubation time, the punctate emission of Re-ERLAD gradually fuses with Lyso-Tracker Deep Red (LTDR) and disappears, indicating that ER buds are gradually engulfed by lysosomes (Fig. 3B). The Pearson correlation coefficients increase gradually from 60 min to 120 min but decrease at 150 min. The reason might be that the signal intensity of Re-ERLAD is weakened along with the degradation of ER buds in lysosomes at 150 min. These results show that Re-ERLAD can track the morphological alterations in ER buds during ERLAD [53].

Real-time tracking of ERLAD with TPFLIM
As Re-ERLAD can induce ERLAD in the presence of light, specifically label ER buds and possess viscositydependent emission properties, we then used it as a theranostic probe to monitor the changes in the viscosity of ER buds during light-initiated ERLAD via TPFLIM. The lifetimes acquired from TPFLIM were plugged into the correlation curve in Fig. 1D to give the viscosity values. Considering the complicated cellular environment, the viscosity values may not be very accurate. Therefore, the trend of changes in viscosity is more meaningful and will give us more information than the specific values.
To study the dynamic change in the viscosity of ER buds during ERLAD, a long period of real-time monitoring was carried out. Interestingly, we found that the changes in viscosity reflecting the degree of unfolded protein aggregation during ERLAD can be divided into two distinct stages. At the earlier stage, a time-dependent increase in viscosity is observed in A549 cells treated with Re-ERLAD upon light irradiation (Fig. 4A), during which ER buds with higher viscosity are separated from the reticular structure (Fig. 4C). The growth is sustained for 2.5 h until the viscosity increases to ca. 240 cP from ca. 190 cP. The increase in viscosity indicates the gradual aggregation of unfolded proteins in ER buds, which may serve as a signal to activate ERLAD [10,53]. At the later stage (2.5∼4 h), the exorbitant viscosity gradually decreases to ca. 180 cP. According to the results from the colocalization experiment, the decrease in viscosity implies the lysosome-mediated degradation of ER buds, which serves as a cellular turnover process to relieve ER stress.
However, when 3-methyladenine (3-MA; a specific inhibitor of autophagy that blocks blocking autophagosome formation) is introduced to the process, the turnover process reflected by the viscosity is reversed (Fig. 4B and D). Instead, the increase in ER viscosity is prolonged, increasing to ca. 260 cP at 4 h, accompanied by obvious cell shrinkage and cellular vacuolation. Because 3-MA reverses the ER turnover process by inhibiting ER-phagy, the survival rates of Re-ERLAD-treated cells decrease in the presence of light (Fig. S17), which indicates that ERLAD is a cytoprotective mechanism against Re-ERLADinduced cell death. Calreticulin is correlated with immunogenic cell death, which could be induced by ER damage [54,55]. Consistent with the enhancing effects of 3-MA on the photocytotoxic effects of Re-ERLAD, a higher rate of calreticulin expression is observed in 3-MA-treated cells (Fig. S18). These results show that Re-ERLAD can induce and be used to monitor the dynamic changes in the viscosity of ER buds during photoinitiated ERLAD.

CONCLUSION
Overall, we developed the first small moleculebased specific ERLAD inducer, Re-ERLAD, which has viscosity-sensitive emission properties and can specifically image ER buds. Re-ERLAD can photoinitiate ERLAD, a specific form of ER-phagy. Re-ERLAD can also be used to monitor morphological alterations in the ER during ERLAD, including accumulation of unfolded proteins in ER buds, recruitment of LC3B proteins and subsequent fusion with lysosomes for degradation. Moreover, Re-ERLAD can be used to quantitatively monitor the dynamic changes in the viscosity of ER buds during ERLAD via TPFLIM, revealing that the microenvironmental turnover of ER is cytoprotective. In summary, we present the first example of the dynamic turnover process of subcellular organelles by quantitatively measuring their micro-environmental parameters, an approach that may be used in ER monitoring or screening of compounds/methods for intervention with ERLAD.

SUPPLEMENTARY DATA
Supplementary data are available at NSR online.

FUNDING
This work was supported by the National Natural Science Foundation of China (22022707, 21778078, 21837006, 91953117 and 22177142) and Fundamental Research Funds for the Central Universities.