Superiority of SpiroZin2 Versus FluoZin-3 for monitoring vesicular Zn2+ allows tracking of lysosomal Zn2+ pools

Small-molecule fluorescent probes are powerful and ubiquitous tools for measuring the concentration and distribution of analytes in living cells. However, accurate characterization of these analytes requires rigorous evaluation of cell-to-cell heterogeneity in fluorescence intensities and intracellular distribution of probes. In this study, we perform a parallel and systematic comparison of two small-molecule fluorescent vesicular Zn2+ probes, FluoZin-3 AM and SpiroZin2, to evaluate each probe for measurement of vesicular Zn2+ pools. Our results reveal that SpiroZin2 is a specific lysosomal vesicular Zn2+ probe and affords uniform measurement of resting Zn2+ levels at the single cell level with proper calibration. In contrast, FluoZin-3 AM produces highly variable fluorescence intensities and non-specifically localizes in the cytosol and multiple vesicular compartments. We further applied SpiroZin2 to lactating mouse mammary epithelial cells and detected a transient increase of lysosomal free Zn2+ at 24-hour after lactation hormone treatment, which implies that lysosomes play a role in the regulation of Zn2+ homeostasis during lactation. This study demonstrates the need for critical characterization of small-molecule fluorescent probes to define the concentration and localization of analytes in different cell populations, and reveals SpiroZin2 to be capable of reporting diverse perturbations to lysosomal Zn2+.


Results
Because FluoZin-3 and SpiroZin2 employ different fluorescence sensing mechanisms, we performed a side-by-side comparison of their response to Zn 2+ perturbations in HC11 cells. Figure 1 depicts the chemical structures of FluoZin-3 and SpiroZin2, along with representative fluorescence images and Zn 2+ response curves. In these experiments, we used the cell-permeable acetoxymethyl (AM) ester derivative of FluoZin-3, FluoZin-3 AM, in which the negatively charged carboxylate groups are masked as neutral esters. After the functionalized dye crosses the cell membrane, intracellular esterases hydrolyze the esters, restoring the carboxylic acids and trapping the negatively charged probe inside the cell. In the absence of Zn 2+ , FluoZin-3 fluorescence is quenched by photoinduced electron transfer (PET) from the chelating group. Upon Zn 2+ binding, PET becomes unfavorable, which causes an increase in the fluorescence emission 21 . In contrast, SpiroZin2 operates via a reaction-based turn-on mechanism that occurs when Zn 2+ binding induces a ring-opening reaction that converts a non-fluorescent spirobenzopyran to a fluorescent cyanine dye 23 . Figure 1(c,d) show representative fluorescence images of HC11 mouse mammary epithelial cells stained with SpiroZin2 and FluoZin-3, respectively. Cells treated with SpiroZin2 routinely exhibited fluorescence in a vesicular pattern. FluoZin-3 fluorescence signals were more variable, appearing to reside in vesicles, large puncta, and the cytosol.
In order to measure the response of each probe to Zn 2+ perturbation, cells stained with either SpiroZin2 or FluoZin-3 were treated with reagents to determine the minimum and maximum fluorescence signal of the apo and Zn 2+ -bound states. To obtain the minimum fluorescence signal for the apo form of the sensor (F min ), the cell-permeable Zn 2+ chelator tris(2-pyridylmethyl)amine (TPA) was added to cells. Subsequently, the maximal fluorescence signal for the bound form of the sensor (F max ) was obtained by addition of excess Zn 2+ with the ionophore pyrithione. The titration curves in Figure 1(c,d) demonstrate that both SpiroZin2 and FluoZin-3 respond to Zn 2+ chelating and saturating reagents within a few minutes in HC11 cells.
Single cell analysis of live cells treated with either FluoZin-3 or SpiroZin2 revealed significant heterogeneity in the fluorescence intensities and staining patterns of each dye. Although the fluorescence intensity of FluoZin-3 has been used widely to estimate intracellular resting labile Zn 2+ levels under resting conditions [26][27][28][29][30] , evidence suggests that the high fluorescence signal of FluoZin-3 in HeLa cells is at least partially due to accumulation of the dye in the Golgi complex rather than high concentrations of labile Zn 2+ 25 . Moreover, variability in the amount of dye internalized by cells causes inconsistencies in resting fluorescence signals (F), and, therefore, estimated labile Zn 2+ levels. This behavior is a notorious issue with all intensity-based dyes: the fluorescence signal (F) depends not only on the analyte of interest (in this case Zn 2+ ), but also on the amount of dye present in a cell or in an organelle. We proposed that labile Zn 2+ levels could be assessed more precisely by normalizing resting state fluorescence to apo state fluorescence after treatment with TPA (F/F min ), where F min provides a Zn 2+ -independent indicator of the amount of dye in each cell. In this case, greater F/F min ratios would indicate higher concentrations of labile Zn 2+ .
To quantify the cellular heterogeneity of fluorescence signals of the two probes, we measured the fluorescence intensities of single HC11 cells stained with either SpiroZin2 or FluoZin-3 AM, with or without pluronic F-127, before (F) and after treatment with TPA (F min ). Pluronic F-127 is a non-ionic detergent that facilitates the dispersion of hydrophobic dyes in aqueous media and is often applied in standard FluoZin-3 staining protocols. Figure 2(a,c) shows that both SpiroZin2 and FluoZin-3 staining resulted in highly variable resting fluorescence (F). In contrast, the normalized fluorescence (F/F min ) of SpiroZin2 gave consistent results across multiple trials. For example, in two independent experiments in the absence of pluronic, we found F/F min = 1.4 ± 0.17 and 1.3 ± 0.16. In two separate experiments, we found that co-incubation with pluronic F-127 gave similar results: F/F min = 1.2 ± 0.21 and 1.2 ± 0.16. These data suggest that the ratio F/F min is a better indicator of resting Zn 2+ levels than F. FluoZin-3 staining experiments gave a large range of F/F min (ratios = 6.8 ± 4.5 and 6.1 ± 4.5 for Scientific RepoRts | (2018) 8:15034 | DOI:10.1038/s41598-018-33102-w two independent trials in the absence of pluronic; 8.4 ± 4.7 and 6.3 ± 4.4 in the presence of pluronic; Fig. 2(c,d)). These results indicate that complete calibration measurements of F, F min , and F max are required to estimate relative Zn 2+ levels using FluoZin-3. We also determined that the addition of pluronic had only minor effects on the ratio F/F min for both SpiroZin2 (ratio = 1.3 ± 0.17, − pluronic; 1.2 ± 0.17, + pluronic, p < 0.0001, Student's t-test) and FluoZin-3 (ratio = 6.5 ± 4.6, − pluronic; 7.2 ± 4.6, + pluronic, p < 0.05, Student's t-test). These data suggest that, although pluronic F-127 may affect the absolute amount of sensor taken up by a given cell, it does not affect sensor performance inside the cell.
When estimating relative labile Zn 2+ concentrations in different cell populations, our experiments demonstrate the need for using normalized signals (F/F min ), which give uniform and consistent results from cell to cell, rather than resting fluorescence intensities (F), which we found to be highly variable. A quantitative measure of variability is the coefficient of variation (CV). The CV of F for SpiroZin2 was 20.3-36.8%, whereas the CV of F/ F min was 12.4-17.4%. On the other hand, FluoZin-3 produced highly variable F (CV = 58.8-86.3%) and F/F min (CV = 55.9-74.6%) results. Thus, when FluoZin-3 is used to estimate or compare intracellular labile Zn 2+ levels under different conditions or in different samples, a large number of cells must be investigated to compensate for the signal variability at the single cell level. Both SpiroZin2 and FluoZin-3 co-localize with lysosomal markers in a variety of eukaryotic cells including HeLa cells, cultured hippocampal neurons, breast cancer cells, and C. elegans intestinal cells 9,23,31,32 . FluoZin-3 was also reported to accumulate in the Golgi apparatus of HeLa cells and DT40 chicken B cells 25,33 . The mechanism by which dyes are sequestered in acidic vesicular compartments is not yet fully understood, and it is possible that a given dye can be distributed unequally across different types of vesicles (e.g. lysosomal, secretory, and Golgi apparatus vesicles). To study the distribution of SpiroZin2 and FluoZin-3 in different vesicular compartments, we performed a colocalization analysis of each dye with three vesicular markers, LAMP1-EBFP, GalT-CFP, and VAMP8-CFP. Lysosome associated membrane protein 1 (LAMP1) is a highly glycosylated integral membrane protein that is found in late endosomes and lysosomes, and it is frequently used as a marker for these organelles 34 . GalT-CFP comprises the first sixty amino acids of human galactosyltransferase (GalT), a trans-Golgi targeting sequence, fused to CFP. Vesicle-associated membrane protein 8 (VAMP8) is a marker for secretory vesicles that is associated with exocytosis in different types of mammalian cells including pancreatic acinar cells 35 , mast cells 36 , cytotoxic T lymphocytes 37 , platelets 38 , and kidney collecting duct epithelia 39 .
To perform the co-localization assay, HC11 cells transiently expressing FP-tagged vesicular markers were stained with SpiroZin2 or FluoZin-3 and imaged by fluorescence microscopy. Dual color images were analyzed with an intensity correlation assay based on the Pearson correlation coefficient (PC), which ranges from −1 to 1, with −1 indicating negative correlation, 0 no correlation, and 1 complete positive correlation. PC analysis requires the intensities of the two channels to be comparable because uneven intensities of two channels can lead to artificially diminished PC values 40 . Under resting conditions in HC11 cells, the fluorescence intensities of the sensors were too low compared to those of the FP-tagged marker proteins, presumably due to the low levels of intracellular Zn 2+ . In order to achieve comparable intensities in the dye and marker-FP channels, we added a Zn 2+ /pyrithione solution to dye-stained cells, which increased the sensor fluorescence intensities after a few minutes. PC-based co-localization analysis revealed that SpiroZin2 was strongly co-localized with LAMP1 (PC = 0.68 ± 0.11), but poorly co-localized with GalT (PC = −0.07 ± 0.10) and VAMP8 (PC = 0.08 ± 0.13), indicating that it preferentially resides in late endosomes and lysosomes (Fig. 3). Because FluoZin-3 is commonly delivered to cells with pluronic, we tested whether this procedure affects subcellular localization. Figure 4 shows that FluoZin-3 co-localizes with LAMP1 reasonably well (PC = 0.54 ± 0.12 without pluronic, 0.64 ± 0.1 with pluronic). However, we observed that some portion of FluoZin-3 also populates the Golgi (PC = 0.22 ± 0.22 without pluronic, 0.23 ± 0.21 with pluronic) and secretory vesicles (PC = 0.41 ± 0.25 with pluronic). Because vesicular compartments may have different amounts of labile Zn 2+ , the non-specific localization of FluoZin-3 may contribute to the highly variable fluorescence intensities of FluoZin-3 staining in Fig. 2.
Similar to reports from other researchers 9,25 , we observed cytosolic signals in some cells treated with FluoZin-3 ( Fig. 1(d)). To rigorously quantify the cytosolic distribution of SpiroZin2 and FluoZin-3, we stained HC11 cells with each dye in the presence and absence of pluronic and then determined the percentage of cells exhibiting cytosolic signals. As shown in Fig. 5, none of the SpiroZin2-stained cells showed cytosolic signals under any condition, whereas 0.38% and 13.7% of FluoZin-3-treated cells exhibited cytosolic staining in the absence and presence of pluronic, respectively. In the presence of pluronic, the increase in cytosolic FluoZin-3 signals may be attributable to the increased aqueous solubility of FluoZin-3 AM. Alternatively, it is possible that pluronic facilitates the general uptake of FluoZin-3 into cells and the dye accumulates in cytosol after saturating vesicular organelles. Therefore, when using pluronic it is advisable to optimize the concentration of the FluoZin-3 to avoid accumulation of dye in the cytoplasm. In contrast, SpiroZin2 exhibited unambiguous and consistent vesicular localization.
To further characterize the ability of SpiroZin2 to detect lysosomal Zn 2+ , we investigated the dynamic range of the sensor in lysosomes and monitored the lysosomal Zn 2+ pool as a function of a series of perturbations. To define the full dynamic range, we carried out in situ calibrations using TPA to determine F min and Zn 2+ with pyrithione to determine F max . As shown in Fig. 6(a), the dynamic range of SpiroZin2 in lysosomes is 5.3 ± 1.2. An important feature of a lysosomal probe is insensitivity to pH. Whereas SpiroZin2 is insensitive to pH changes between 4.0 and 6.0 in vitro, this property was not previously tested in cells. To assess the pH dependence of SpiroZin2 in cells, we first validated that LysoSensor TM Yellow/Blue Dextran, a lysosomal pH indicator, could detect changes in pH from 4 to 6 (Supplementary Fig. 1(a)). We demonstrated that treatment of cells with NH 4 + lead to a lysosomal pH shift within 2.5 min (Supplementary Fig. 1(b)) and showed that this shift in pH did not alter SpiroZin2 fluorescence intensity (Supplementary Fig. 1(c)), confirming that SpiroZin2 fluorescence is not affected by changes in pH within this range in cells.
We also examined how lysosomal Zn 2+ pools respond to an increase in extracellular Zn 2+ . Previously we demonstrated that elevation of extracellular Zn 2+ leads to a slow increase in cytosolic Zn 2+ in HeLa cells followed by sequestration into intracellular organelles such as the ER and mitochondria 14 . Here, we show that HC11 cells also exhibit a steady increase in cytosolic Zn 2+ detected via the genetically encoded fluorescent Zn 2+ sensor ZapCV2 41 , when treated with 40 μM extracellular Zn 2+ ( Fig. 6(a). Surprisingly, lysosomal Zn 2+ increased more quickly than cytosolic Zn 2+ , reaching a peak within 5 min, while cytosolic Zn 2+ continued to rise after 45 min of treatment. This result suggests that the extracellular Zn 2+ is rapidly internalized into lysosomes via endocytosis, which operates on a time scale of seconds to minutes 42 . Our data also revealed that this increase in lysosomal Zn 2+ was transient; the Zn 2+ level dropped quickly and stabilized within 20 min. The possible routes of lysosomal Zn 2+  43 , and/or export to the extracellular environment by lysosomal exocytosis 44 . To further validate that SpiroZin2 specifically monitors lysosomal Zn 2+ , we treated cells with ML-SA1, a TRPML1 agonist 45 that activates Ca 2+ release from lysosomes 46,47 . We discovered that ML-SA1 treatment led to a decrease in normalized SpiroZin2 fluorescence compared to vehicle (DMSO) treatment ( Fig. 6(b)), consistent with reported studies that TRPML1 mediates a Zn 2+ leak from lysosomes in human and mouse cells 10,43 . Together, these data demonstrate that SpiroZin2 is capable of monitoring changes in lysosomal Zn 2+ level upon direct modulation of lysosome Zn 2+ transport.
Finally, we used bafilomycin A1, a vacuolar H + ATPase inhibitor, 48 to dissipate the H + gradient across the lysosomal membrane (Fig. 6(c)), and show that this procedure decreases the lysosomal Zn 2+ pool, as detected by a decrease in normalized SpiroZin2 fluorescence intensity ( Fig. 6(d-f)). These results suggest that lysosomal Zn 2+ accumulation is facilitated by the proton gradient via Zn 2+ /H + exchange. The Cation Diffusion Facilitator (CDF) protein family transports divalent cations and is present in diverse organisms from bacteria to humans 49 . It is well established that the E.coli CDF protein YiiP acts as a Zn 2+ /H + antiporter 50,51 . YiiP is a homologue of mammalian ZnT transporters, and Zn 2+ transport by ZnT1 and ZnT5 has been suggested to be facilitated by proton gradients 52,53 . Given that ZnT2 and ZnT4 transport Zn 2+ into lysosome 43,54,55 , it is possible that ZnT2 and ZnT4 act as Zn 2+ /H + antiporters, whereby dissipation of the proton gradient disrupts lysosomal Zn 2+ accumulation.
Given its superior performance, we used SpiroZin2 to track vesicular Zn 2+ concentrations in mouse mammary epithelial cells (HC11), which differentiate in response to lactation hormones. Because accumulation of Zn 2+ in intracellular vesicles in secreting HC11 cells has been hypothesized to be a hormone-dependent process 13 , we chose to study this phenomenon. Previous reports indicate that lysosome-related organelles in C. elegans intestinal cells serve as major Zn 2+ storage sites that play important roles in maintaining Zn 2+ homeostasis 32 . To examine whether late endosomes/lysosomes serve a similar function in secreting HC11 cells, we induced  Fig. 7(a). Cells were stained simultaneously with a nuclear dye and SpiroZin2. Nucleus channel images were analyzed with ImageJ to identify the nuclear area, segment cells, and define regions of interest (ROIs), with each ROI corresponding to a single cell. The mean fluorescence intensity in the corresponding SpiroZin2 channel was quantified for each cell in the entire field of view. Normalized fluorescence intensities were measured by treating the cells with TPA and calculating F/F min , as described above. As shown in Fig. 7(b), we detected a substantial increase in labile Zn 2+ concentrations 24 hours after hormone treatment relative to proliferating resting cells (p < 0.0001), an observation that suggests late endosomes or lysosomes may serve as temporary storage sites for Zn 2+ during HC11 cell differentiation.

Discussion
Live-cell imaging of individual cells using small-molecule fluorescent probes is a powerful and widely used approach for estimating the concentration and distribution of analytes in cells. However, few studies rigorously evaluate the heterogeneity in fluorescence signal and localization from cell to cell and explicitly compare the functionality of different probes. In this study, we performed a parallel and systematic comparison of two small-molecule vesicular Zn 2+ probes, FluoZin-3 AM and SpiroZin2, to evaluate each probe for measurement of vesicular Zn 2+ pools. Our work reveals that SpiroZin2 specifically reports on lysosomal Zn 2+ pools and can be used to monitor changes in lysosomal Zn 2+ in response to a variety of perturbations (changes in extracellular Zn 2+ , activation of TRPML1, dissipation of the lysosomal H + gradient, and cell differentiation).
One of the primary applications of small molecule fluorescent Zn 2+ sensors is to compare relative amounts of Zn 2+ in different biological samples, for example between different cell types or between cells in different physiological states, such as differentiated versus undifferentiated. As is evident in published studies [26][27][28][29][30] , it is common practice to use resting fluorescence intensities of FluoZin-3 to estimate relative Zn 2+ levels in different cell populations. Our results indicate that the resting FluoZin-3 or SpiroZin2 fluorescence intensities (F) are highly variable among cells (Fig. 2), with an average coefficient of variation (CV) of 28% for SpiroZin2 and 73% for FluoZin-3, and should not be used to represent relative Zn 2+ levels. In contrast, for SpiroZin2 the ratio of fluorescence intensity of the resting state to that after TPA treatment (F/F min ) was uniform and consistent, with a substantially decreased average CV (13.7%, Fig. 2). Therefore, instead of F, F/F min is a more appropriate and precise parameter to assess relative labile Zn 2+ levels in cells when using SpiroZin2. Unfortunately for FluoZin-3, fluorescence signals normalized to F min (F/F min ) remained highly variable (average CV 66%). One possible explanation for this high variability in fluorescence signal under resting conditions is that individual cells internalize different amounts of dye. In this case, normalization of F to the F min signal in each individual cell provides an internal control for that cell and mitigates the variability. For FluoZin-3, the large degree of signal heterogeneity, even after normalization, most likely results from the widespread heterogeneity in dye localization in addition to affects from differential dye accumulation.
Unlike targeted genetically-encoded sensors that can be localized to a specific organelle by a tagging sequence, small molecule fluorescent probes tend to accumulate adventitiously in different compartments 25 . Because previous reports suggest that FluoZin-3 AM accumulates in multiple organelles in mammalian cells 58 , we rigorously determined the intracellular distribution of the dyes. Our data reveal that SpiroZin2 exclusively localizes in lysosomal vesicles. Whereas previous work 23 showed SpiroZin2 accumulated in lysosomes, it did not establish that SpiroZin2 was exclusively in lysosomes, as we do here. FluoZin-3 AM non-specifically accumulates in the cytosol, as well as across three populations of vesicles: lysosomal vesicles, trans-Golgi vesicles and secretory vesicles. Owing to possible different labile Zn 2+ levels in different vesicular compartments, the heterogeneous intracellular localization of FluoZin-3 may contribute to the variable fluorescence intensities of FluoZin-3 staining observed in Fig. 2, even after normalization to F min . This high degree of variation confounds interpretation of Zn 2+ measurements made by FluoZin-3 in different biological samples or under different conditions. In order to compare relative Zn 2+ levels in a particular subcellular compartment across samples or conditions using FluoZin-3, researchers should simultaneously measure FluoZin-3 signals and organelle markers, carry out full probe calibrations by recording F, F min , and F max in individual cells, and restrict analysis to those cells in which FluoZin-3 co-localizes with a genetically-encoded marker for the compartment of interest.
In the present study we used F/F min to provide a relative measure of the labile Zn 2+ pool in lysosomes. It should be noted that, although sensors can be used to quantify Zn 2+ , doing so requires a full sensor calibration, estimation of the dissociation constant (K d ) for Zn 2+ ideally in the compartment of interest, and demonstration that the probe is in equilibrium with the labile pool within a particular compartment. Small compartments, such as lysosomes, pose an additional challenge because of their extremely small volumes.
Our results reveal that SpiroZin2 is a specific lysosomal Zn 2+ probe that provides reliable measurements of resting lysosomal Zn 2+ pools in single cell analyses, when properly normalized, and how these pools change with different cellular perturbations. We demonstrate that lysosomal Zn 2+ pools respond rapidly to perturbation of extracellular Zn 2+ . We further show that the lysosomal Zn 2+ pool decreases upon activation of the Zn 2+ -permeable channel TRPML1 and relies on activity of the V-ATPase of lysosomes. We also applied SpiroZin2 to lactating mouse mammary epithelial cells and detected a transient increase of lysosomal free Zn 2+ 24 hrs after treatment with a lactation hormone. This finding is significant because a growing number of studies suggest that lysosomes play a key role in Zn 2+ homeostasis. Zn 2+ can be delivered into lysosomes from the cytosol by two Zn 2+ transporters, Slc30a2 and Slc30a4 43,54,55 , or through other cellular processes such as autophagy 59,60 . Lysosomes serve as Zn 2+ storage reserves and protect cells from Zn 2+ overload in C. elegans intestinal cells and HeLa cells 32,44 . Even so, lysosomal accumulation of Zn 2+ can also be toxic, inducing lysosomal membrane permeabilization (LMP) and cell death under certain pathological conditions 9,24,61,62 . Lysosomal Zn 2+ can be transported into the cytosol by TRPML1 43 and/or exported to the extracellular environment by lysosomal exocytosis 44 . Although the mechanism of the transient increase of lysosomal Zn 2+ lactating mouse mammary epithelial cells is unknown, we propose that lysosomes serve as temporary Zn 2+ storage sites and play key roles in the regulation of Zn 2+ homeostasis during lactation. Previous work has implicated lysosomes as playing an important role in redistribution of Zn 2+ pools during involution of mammary tissue upon weaning [63][64][65] . Here, we show they also play a role during differentiation to a lactation phenotype, suggesting that lysosomes may be a major player during remodeling of Zn 2+ homeostasis and redistribution of labile Zn 2+ pools.

Methods
Chemicals. SpiroZin2 was synthesized as previously described and a 4 mM stock solution was prepared in DMSO 23  An insulin stock solution (4 mg/mL) was purchased from Life Technologies (catalog number 12585014). Human recombinant epidermal growth factor (EGF, VWR 47743-566) was prepared as a 10 μg/ mL stock solution in water. Prolactin (Sigma-Aldrich L6520) was prepared as a 5 mg/mL stock solution in water. A stock solution (138 μM) of hydrocortisone (Sigma-Aldrich H0888) in absolute ethanol was also prepared. ML-SA1 (Abcam ab144662) was prepared as a 50 mM stock solution in DMSO. Bafilomycin A1 (Sigma-Aldrich B1793) was prepared as a 100 μM stock solution in DMSO.
Cell Loading with Zinc Probes and Calibration. FluoZin-3 AM and SpiroZin2 were loaded as previously described according to manufacturer's instructions 23 . Briefly, a 2 mM stock solution of FluoZin-3 AM DMSO stock solution was diluted to a final concentration of 3 μM in phosphate-free HHBSS buffer. Phosphate-free HHBSS buffer was used to avoid precipitation of Zn 2+ by phosphate. HC11 cells were washed with phosphate-free HHBSS and then incubated with FluoZin-3 AM/phosphate-free HHBSS for 1 h at 37 °C with 5% CO 2 , followed by a 30 min wash in phosphate-free HHBSS buffer at 37 °C with 5% CO 2 . Imaging was performed immediately after washing. For SpiroZin2, a 4 mM DMSO stock solution was diluted to a final concentration of 5 μM in phosphate-free HHBSS buffer. HC11 cells were washed with phosphate-free HHBSS buffer and then incubated with 2 mL of SpiroZin2 solution for 1 h at 37 °C with 5% CO 2 prior to imaging. To apply pluronic during the staining procedure, the stock solution of zinc probe was mixed with an equal volume of Pluronic F-127 20% (w/v) solution and then the mixture was diluted in phosphate-free HHBSS buffer, making the final pluronic concentration about 0.025%. For Zn 2+ calibrations, F min measurements were acquired by treating cells with 150 μM TPA in phosphate-free HHBSS buffer. To obtain values for F max , a solution containing 5 μM pyrithione/20 μM ZnCl 2 in phosphate-free HHBSS buffer was added to HC11 cells. Fluorescence images for colocalization of fluorescent probes with VAMP8-CFP in Fig. 3 and all fluorescence images in Fig. 4 were acquired on a Nikon Ti-E spinning disc confocal microscope fitted with a Yokogawa CSU-X1 spinning disc head, equipped with the Nikon Elements software platform and Ti-E Perfect Focus system. Cells were imaged using a 60x oil objective (NA 1.40) or a 100x oil objective (NA 1.45)

Responses of Probes to Zn 2+ Perturbation in HC11 Cells.
To test the responses of probes to Zn 2+ perturbation, 8 × 10 5 HC11 cells were plated in a 35-mm glass bottom dish one day prior to imaging. Cells were stained with FluoZin-3 AM (with pluronic) or SpiroZin2 as described above. The nuclear dye was added to the cell culture medium 30 min prior to imaging. Cells were washed with phosphate-free HHBSS buffer and then imaged on a Nikon A1R laser scanning confocal microscope as described in the "Live Cell Imaging" section. Z-series images were recorded every 3 min throughout the experiment. After adding Zn 2+ /pyrithione to cells, PMT gain was lowered from 50 to 15 for SpiroZin2 and from 45 to 25 for FluoZin-3 to avoid detector saturation. Image analysis was performed using ImageJ. Z-series images of each fluorescence channel per time point were flattened using the maximum intensity Z-projection. To perform single cell analyses, 15 cells were randomly chosen and each cell was defined as an individual region of interest (ROI). Background fluorescence of each channel was calculated by averaging the signal intensities of an area (mean intensities) without cells. The fluorescence intensities of the FluoZin-3 or SpiroZin2 channel of each ROI were calculated by subtracting background fluorescence from the raw mean fluorescence intensities for each channel. imaging. Cells were stained with FluoZin-3 or SpiroZin2 with or without pluronic as described above. Nuclear dye was added to the cell culture medium 30 min prior to imaging. Cells were washed with phosphate-free HHBSS buffer and then imaged on a Nikon A1R laser scanning confocal microscope as described in the "Live Cell Imaging" section. Z-series images were recorded in the resting state (F) and 15 min post TPA treatment (F min ). Image analysis was performed using ImageJ. Z-series images of each fluorescence channel per time point were flattened using the maximum intensity Z-projection. To perform single cell analyses, nucleus channel images were thresholded to identify the nucleus area. Touching objects were separated by a watershed algorithm. Next, segmented particles were identified by the "find maxima" command with "segmented particles" as the output type, and "light background" chosen when the background color was white. To create ROIs with each ROI corresponding to a single cell, segmented pictures were processed by the "analyze particles" command as follows: the range of the size of particles was adjusted so that each particle incorporated a single cell and then the option "add to manager" was chosen to generate a list of ROIs. Applying the ROIs to the FluoZin-3 or Spirozin2 channel images allowed the quantification of the mean raw intensity of the sensor fluorescence at the single-cell level. The background fluorescence of each channel was calculated by averaging the signal intensities of an area (mean intensities) without cells. The fluorescence intensities of FluoZin-3 or SpiroZin2 channel of each cell were calculated by subtracting the background fluorescence from the raw mean fluorescence intensities.

Colocalization Assay of Zinc Probes with Vesicular Markers.
To transiently express vesicular markers in HC11 cells, 500 ng of DNA encoding each fusion protein (LAMP1-EBFP, VAMP8-CFP or GalT-CFP) was electroporated into 1 × 10 6 cells with the Neon system (Life Technologies) 24-48 h prior to imaging using the following parameters: pulse voltage: 1400 V; pulse width: 20 ms; pulse number: 2 pulses. Electroporated cells were plated in 35-mm glass-bottom imaging dishes. Before imaging, cells were stained with FluoZin-3 or SpiroZin2 with or without pluronic as described above. After staining, cells were washed with phosphate-free HHBSS buffer. To enhance the fluorescence intensities of zinc probe channels, 5 μM pyrithione/20 μM ZnCl 2 was added to cells. Fluorescence images were recorded 5-10 min after Zn 2+ addition. Sensor co-localization experiments with LAMP1-EBFP and GalT-CFP were performed on a Nikon A1R laser scanning confocal microscope as described in the "Live Cell Imaging" section. Colocalization experiments with VAMP8-CFP were performed on a Nikon Ti-E spinning disc confocal microscope as described in the "Live Cell Imaging" section.
To perform the Pearson coefficient-based colocalization assay, cells with fluorescence signals in both the zinc sensor channel and the vesicular maker channel were selected for analysis with JACoP plugin for ImageJ (https:// imagej.nih.gov/ij/plugins/track/jacop.html). An important criterion for choosing cells is that the fluorescence intensities of the two channels must be comparable. A region of interest (ROI) was drawn within a cell and the Pearson coefficient was calculated by JACoP. Data were plotted with KaleidaGraph.

Cytosolic Distribution of Zinc Probes.
To determine the cytosolic distribution of FluoZin-3 and SpiroZin2, 8 × 10 5 HC11 cells were plated in a 35-mm glass bottom dish one day prior to imaging. Cells were stained with FluoZin-3 or SpiroZin2 with or without pluronic as described above. Nuclear dye was added to the cell culture medium 30 min prior to imaging. After staining, cells were washed with phosphate-free HHBSS buffer. To enhance the fluorescence intensities of the zinc sensor channels, 5 μM pyrithione/20 μM ZnCl 2 was added to the cells. Z-series fluorescence images were recorded 5-10 min after Zn 2+ addition on a Nikon Ti-E spinning disc confocal microscope as described in the "Live Cell Imaging" section. Image analysis was performed using ImageJ. Z-series images of each fluorescence channel were flattened using the maximum intensity Z-projection.
Cells with cytosolic distribution of dyes were counted by eye. The total number of cells was calculated with ImageJ. Nucleus channel images were thresholded to identify the nuclear area. Touching objects were separated with a watershed algorithm. The number of nuclei (i.e. cells) was determined with the command "analyze particles. " The quantification of cytosolic distribution of dyes was determined by dividing the number of cells with cytosolic signal by the total number of cells per image.
Lysosomal pH Measurement. HC11 cells were grown in a 96-well plate in growth media supplemented with 0.5 mg/mL of LysoSensor TM Yellow/Blue Dextran (ThermoFisher Scientific, L22460) overnight. The standard curve was generated by incubating cells in 10 μM monensin and 10 μM nigericin in MES buffer (5 mM NaCl, 115 mM KCl, 1.3 mM MgSO 4 , 25 mM MES), with the pH adjusted to within the range of 4.0-7.0 for 10 min. After the incubation, fluorescence was quantified with a fluorescence microplate reader (BioTek Synergy H1 hybrid) at emission wavelengths of 535 and 440 nm with excitation at 340 nm. To generate a standard curve, the ratio of emission 535 nm/440 nm was plotted against the pH value of MES buffer. The lysosomal pH post Bafilomycin A1 or DMSO treatment was calculated using the standard curve. TRPML1 Activation by ML-SA1. To monitor Zn 2+ release from lysosomes, 8 × 10 5 HC11 cells were plated in a 35-mm glass bottom dish one day prior to imaging. Prior to imaging, cells were stained with SpiroZin2 and nuclear dye as described above, and then imaged on a Nikon A1R laser scanning confocal microscope as described in the "Live Cell Imaging" section. Z-series images were recorded every 3 min. ML-SA1 (final concentration 20 μM) or DMSO was added to cells for 30 min and then washed off, followed by 12 min treatment with 150 μM TPA. Image analysis was performed with ImageJ. Z-series images of each fluorescence channel per time point were flattened using the maximum intensity Z-projection. Single cell analysis was performed as described in the "Variability of Fluorescence Intensities of Zinc Probes". Fluorescence after ML-SA1/DMSO treatment was defined as F and fluorescence after TPA treatment was used as F min . The lysosomal Zn 2+ level after ML-SA1 or DMSO treatment is represented as F/F min .
Scientific RepoRts | (2018) 8:15034 | DOI:10.1038/s41598-018-33102-w Monitoring Cytosolic Zn 2+ Using a Genetically-encoded Zn 2+ FRET Sensor. HC11 cells stably expressing a Zn 2+ FRET Sensor (ZapCV2) were generated using the PiggyBac ™ Transposon Vector System (System Biosciences). Fluorescence images were acquired in the YFP FRET and CFP channels. To perform data analysis, a cytoplasmic region (ROI) and a region without cells (background) were selected and the FRET ratio was calculated by dividing the background corrected YFP FRET intensity by the background corrected CFP FRET intensity.
Measurement of Lysosomal Zn 2+ Level During Lactation. HC11 cells were induced to differentiation as described above. To study the change in lysosomal Zn 2+ over the progression of lactation, cells at different stages were stained with SpiroZin2 (without pluronic) and nuclear dye as described above, and then imaged on a Nikon A1R laser scanning confocal microscope as described in the "Live Cell Imaging" section. Z-series images were recorded in the resting state (F) and 15 min after TPA treatment (F min ). Image analysis was performed with ImageJ. Z-series images of each fluorescence channel per time point were flattened using the maximum intensity Z-projection. To perform single cell analyses, the same workflow as described in the above section "Variability of Fluorescence Intensities of Zinc Probes" was used: nucleus channel images were analyzed to identify the nucleus area, segment cells, and create ROIs, with each ROI identifying a single cell. Applying the ROIs to the Spirozin2 channel allowed us to quantify the mean intensity of SpiroZin2 at the single-cell level in an entire field of view. Background fluorescence of each channel was calculated by averaging the signal intensities of an area (mean intensities) without cells. The fluorescence intensities of the SpiroZin2 channel of each cell were calculated by subtracting background fluorescence from the raw mean fluorescence intensities. Data plots and statistical analyses (ANOVA with Tukey's HSD posthoc test) were generated with KaleidaGraph software.