Comprehensive multiscale analysis of lactate metabolic dynamics in vitro and in vivo using highly responsive biosensors

Abstract

As a key glycolytic metabolite, lactate has a central role in diverse physiological and pathological processes. However, comprehensive multiscale analysis of lactate metabolic dynamics in vitro and in vivo has remained an unsolved problem until now owing to the lack of a high-performance tool. We recently developed a series of genetically encoded fluorescent sensors for lactate, named FiLa, which illuminate lactate metabolism in cells, subcellular organelles, animals, and human serum and urine. In this protocol, we first describe the FiLa sensor-based strategies for real-time subcellular bioenergetic flux analysis by profiling the lactate metabolic response to different nutritional and pharmacological conditions, which provides a systematic-level view of cellular metabolic function at the subcellular scale for the first time. We also report detailed procedures for imaging lactate dynamics in live mice through a cell microcapsule system or recombinant adeno-associated virus and for the rapid and simple assay of lactate in human body fluids. This comprehensive multiscale metabolic analysis strategy may also be applied to other metabolite biosensors using various analytic platforms, further expanding its usability. The protocol is suited for users with expertise in biochemistry, molecular biology and cell biology. Typically, the preparation of FiLa-expressing cells or mice takes 2 days to 4 weeks, and live-cell and in vivo imaging can be performed within 1–2 hours. For the FiLa-based assay of body fluids, the whole measuring procedure generally takes ~1 min for one sample in a manual assay or ~3 min for 96 samples in an automatic microplate assay.

Introduction

Although lactate has traditionally been viewed as a metabolic waste product of glycolysis^1–4, the interest in lactate is increasing owing to its vital role in diverse biological functions. Lactate is one of the main respiratory fuels and is considered as important as glucose^1,3–5. It can be utilized as a building block for the biosynthesis of glucose, fatty acids and some amino acids^1,3–5. Lactate also plays a role in cellular signaling, as a natural suppressor of antiviral immunity⁶, and in the regulation of Mg²⁺ transport⁷ and gene expression by the posttranslational modification of histones⁸. Furthermore, lactate is actively shuttled between subcellular compartments, cells and organs⁹. Through these biochemical mechanisms, lactate metabolism is involved in diverse physiological conditions, such as spermatogenesis¹⁰, embryonic development¹¹, myocardial regeneration¹², autophagy^13,14, hypoxia response¹⁵, nutrient stress¹⁶, physical exercise^17,18, immune response^19,20, neural activity^21,22, the metabolism and function of the gut microbiome^17,23 and aging^24,25. Dysregulated lactate levels are associated with numerous life-threatening diseases, including obesity^26,27, diabetes mellitus²⁸, cancer^29–31, sepsis³², ischemic injury^33,34, heart failure³⁵ and Alzheimer’s disease³⁶. Therefore, a sensitive and simple assay to evaluate lactate metabolism is of interest for biologists and medical researchers.

Lactate is the end product of glycolysis. In mammalian cells, it is reversibly converted into pyruvate, catalyzed by lactate dehydrogenase (LDH) or released into the extracellular space via monocarboxylate transporters³. After conversion to pyruvate, lactate can produce a large amount of energy through entry into the tricarboxylic acid cycle and oxidative phosphorylation, generate glucose through the gluconeogenesis pathway or biosynthesize various metabolites such as amino acids, fatty acids and ketone bodies as building blocks⁹. Although the transport of lactate into mitochondria followed by its catabolism has been proposed and studied for a number of years, it remains a controversial topic^9,37. Moreover, lactate can be utilized as a substrate for protein modification via lactylation^8,38. Recently, it was found that lactylation ubiquitously occurs in mitochondrial proteins and histones^8,38, and we provided visual evidence of highly abundant mitochondrial lactate, suggesting an alternative metabolic fate of the metabolite³⁹. Extracellular lactate can either be taken up by cells in the microenvironment or circulate in the blood and supply distant organs as a fuel^3,5. In particular, lactate shuttles between muscle and liver, known as the Cori cycle, and has an important role in the systematic usage of fuels⁹. Physiological lactate levels have been estimated to be 0.6–1 mM within mammalian cells^6,9 or 0.8–2 mM in blood⁹; however, lactate abundance may fluctuate dramatically during physical exercise, metabolic stress and disease^9,40.

Development of FiLa biosensors and comparison with other methods

Traditional in vitro lactate assay methods

Despite the strong demand, tools for tracking lactate in live cells and in vivo are limited. Traditional biochemical methods, such as enzymatic cycling assays and chromatography and mass spectrometry, have been routinely used for many years. Feeding cells with isotopically labeled lactate and quantifying the metabolome by mass spectrometry enables the tracking of lactate metabolic changes. However, all of these methods require cell lysis during sample preparation, thus losing information about the spatial distribution and temporal dynamics of lactate. In addition, these methods are time intensive and, therefore, not suitable for quantitative, real-time high-throughput screening in living cells and in vivo. Recently, the Seahorse Analyzer has emerged as a noninvasive tool and has become increasingly commonly used in metabolic research⁴¹. However, the assay depends on the determination of the extracellular acidification rate and does not directly measure lactate levels. Although extracellular lactate levels are often balanced with intracellular levels, this equilibration does not always hold true. Moreover, the Seahorse Analyzer measures the average level of lactate in the cell population rather than in single cells, and cannot discern the lactate abundance in a single subcellular compartment or in vivo.

Genetically encoded fluorescent sensor-based lactate assay

Genetically encoded fluorescent sensors can overcome the technological challenge associated with determining the spatiotemporal characteristics of metabolites in living cells and organisms^42–45. These biosensors are generally classified into two categories, namely, single fluorescent protein (FP)-based sensors and Förster resonance energy transfer (FRET)-based sensors. Compared with single FP-based sensors, FRET-based sensors usually have a larger size but a smaller dynamic range (typically 10–150%)^44,46. For single FP-based sensors, a circular permuted FP (cpFP) is usually utilized and inserted into a loop region of a substrate-binding protein or between tandem units of proteins⁴⁴. In these cpFPs, the original N and C termini are joined with a short peptide linker, and new N and C termini are created around the fluorophore, making their fluorescence highly sensitive to metabolite changes^47–50. These sensors have been further optimized by screening the positions of cpFP insertion in the substrate-binding protein, screening the linker between the substrate-binding protein and cpFP, mutation of amino acid residues around the binding pocket, and truncation of the substrate-binding protein or linkers between protein domains to improve (and optimize) the fluorescent properties of sensors, such as the responsiveness, selectivity, affinity, signal intensity, and so on⁴⁴.

Since the first lactate biosensor Laconic was documented⁵¹, continuous efforts have been made to improve the sensors’ performance in the past decade, but the specificity, affinity and responsiveness, among other properties, of these biosensors remain to be optimized^52–56 (Fig. 1a,b and Table 1). A common disadvantage for these sensors is low responsiveness; for example, Laconic⁵¹ and GEM-IL⁵⁴ only have ~19% and ~88% fluorescence responses in vitro, respectively (Table 1). The fluorescence response of two subsequent lactate sensors, Green Lindoblum⁵² and eLACCO1.1 (ref. 53), increased to ~400%; however, their dissociation constants are ~30 μM and ~3.9 mM, respectively (Table 1), and in most cases they are either saturated or free in living cells, which seriously impairs their performance. Moreover, some sensors respond to metabolites other than lactate, including citrate for Laconic⁵¹ and calcium for eLACCO1.1 (ref. 53) and CanlonicSF⁵⁶ (Table 1). In addition, except for the FRET sensor Laconic and the fluorescence lifetime sensor LiLac⁵⁵, the readout of other sensors is intensity based, making them susceptible to the expression level of the reporters (Table 1).

Fig. 1 |. Genetically encoded lactate sensors.

a, General design of Laconic, Green Lindoblum, eLACCO1.1, CanlonicSF, GEM-IL, LiLac and FiLa. b, Timeline of the development of genetically encoded lactate sensors. c, Fluorescence intensities of FiLa, FiLa-H and FiLa-L with excitation at 485 nm or 420 nm in the presence of lactate and emission at 528 nm. Data are normalized to the fluorescence in the absence of lactate (n = 3). d, Lactate titration curves of the FiLa, FiLa-H, FiLa-L and FiLa-C sensors. Data are normalized to the initial value (n = 3). e, Lactate titration curves of FiLa, FiLa-H and FiLa-L. Data are converted to the 0–1 scale (n = 3). Data in d and e are derived from c. Data are presented as the mean ± s.e.m. (c–e).

Table 1 |.

Genetically encoded fluorescent sensors for lactate

Sensor	Binding protein	Species sensed	Affinity (Kd)	Dynamic range	Molecular size (kDa)	FP	Sensor type	Validated in subcellular compartments	Validated in vivo	Validated in clinical samples
FiLa	E. coli LLdR protein	Lactate	~130 μM	~1,500%	47.3	cpYFP	Ratiometric	Cytosol/nucleus/mitochondria	Yes	Human plasma/serum
FiLa-H	E. coli LLdR protein	Lactate	~20 μM	~2,600%	47.3	cpYFP	Ratiometric	Cytosol/nucleus	No	Human plasma/serum/urine
FiLa-La	E. coli LLdR protein	Lactate	~800 μM	~700%	47.3	cpYFP	Ratiometric	Cytosol/nucleus/mitochondria	No	No
Laconic51	E. coli LLdR protein	Lactate; citrate	~8 μM; ~830 μM	~19%	83.9	mTFP/Venus	Ratiometric	Cytosol/nucleus	Yes	No
Green Lindoblum52	E. coli LLdR protein	Lactate	~30 μM	~420%	48.5	GFP	Intensiometric	Cytosol/plasma membrane	No	Mouse pLasma
GEM-IL54	E. coli LLdR protein	Lactate	~661 μM	~87.5%	48.7	sfCFP(1-10)/sfGFP11	Intensiometric	Cytosol	No	No
eLACCO1.1 (ref. 53)	T. thermophilus TTHA0766 protein	Lactate; Ca2+	~3.9 μM	~400%	66.3	cpGFP	Intensiometric	Cytosol/cell surface	No	No
CanlonicSF56	T. thermophilus TTHA0766 protein	Lactate; Ca2+	~300 μM	~190%	65.8	cpGFP	Intensiometric	Cytosol/endoplasmic reticulum	No	No
LiLac55	H. pylori TlpC protein	Lactate	~0.62 mM; ~2.68 mMb	~1 ns (Ax)	57.3	mTurquoise2	Fluorescence lifetime	Cytosol	No	No

Bold indicates desired properties of the sensor.

^aFiLa-L is mutant of FiLa of Low affinity.

^bThe apparent K_d of LiLac is ~0.62 μM at 24 °C and ~2.68 μM at 34 °C.

Design and development of highly responsive FiLa biosensors

In previous studies, Escherichia coli LldR was demonstrated to show notable conformational changes in response to lactate; therefore, it served as a promising candidate for sensor design⁵⁷. To develop highly responsive lactate sensors, we first designed 90 chimeric proteins by inserting circularly permuted yellow FP (cpYFP) between two complete or truncated subunits of LldR or between amino acid residues located on the random coiled loops of LldR with or without the DNA-binding domain³⁹. Interestingly, the Y186/P189 chimera with cpYFP insertion in truncated LldR showed a ~200% fluorescence increase after lactate addition³⁹. We then created a series of truncated variants of this chimera in the 182–189 residue regions of LldR and found that the M185/P189 variant exhibited a ~420% fluorescence increase in response to lactate. To expand further the dynamic range of the sensors, we created a library of random mutants (~700 variants) targeting residues P189 and P190 based on the M185/P189 chimera. Among them, P189M/P190D, P189H/P190D and P189F/P190D show a dynamic range of ~740%, ~1,400% and ~2,600%, respectively, and their affinities for lactate are ~20–30 μM, which are far below the physiological concentrations of lactate (0.8–2 mM in blood, 0.6–1 mM in mammalian cells⁹). Thus, to tune the affinity of these lactate sensors, we next created a library of site-saturation mutants targeting residue M185, based on the P189M/P190D, P189H/P190D and P189F/P190D chimeras, and their affinities for lactate ranged from 19 to 1,000 μM. Considering the responsiveness, the selectivity and the affinity, we finally chose the P189F/P190D, M185L/P189H/P190D and M185A/P189M/P190D variants, denoted FiLa-H, FiLa and FiLa-L, respectively, for further research.

FiLa sensors exhibit several clear advantages over existing biosensors^39,58 and constitute the next generation of lactate sensors (Fig. 1c–e and Table 1). The FiLa sensors have opposing responses to lactate binding when excited at 420 and 485 nm (Fig. 1c and Supplementary Figs. 1a–c and 2a–c), leading to 700–2,600% fluorescence responses that were essentially unaffected by temperature fluctuations between 20 °C and 40 °C (Fig. 1d and Supplementary Figs. 1d and 2d). These fluorescence responses were almost 2- to 130-fold greater than those of the first-generation lactate sensors (Fig. 1d and Table 1). FiLa sensors are highly selective for lactate not responsive to any mono-, di- or tricarboxylic acids, intermediate metabolites of glycolysis, amino acids, or Ca²⁺/Mg²⁺ ions³⁹. As shown in Supplementary Figs. 1e,f and 2e,f, both FiLa-H and FiLa-L sensors also have good specificity for lactate over other relevant metabolites in the range of 0–10 mM, except that FiLa-H has a small response to 10 mM malate (Supplementary Fig. 1e,f), and FiLa-L has a small response to 10 mM pyruvate or citrate (Supplementary Fig. 2e,f). Physiological concentrations of malate, pyruvate and citrate are lower than 100 μM (ref. 59) in human plasma and are 20–800 μM (ref. 60), 10–200 μM (ref. 60) and 10–1,000 μM (ref. 60), respectively, in mammalian cells, and, therefore, these endogenous metabolites are not expected to interfere with lactate sensing under physiological conditions. FiLa, FiLa-H and FiLa-L sensors have pH-resistant apparent dissociation constants ($K_{\mathrm{d}}$ values) of ~130, ~20 and ~800 μM in the physiological pH range (pH 6.0–8.0) (Fig. 1d,e and Supplementary Figs. 1g and 2g), covering most physiological concentrations of lactate. When the lactate level is unusually abundant or scarce, the FiLa-L variant with low affinity or the FiLa-H variant with high affinity can be used. In addition, the FiLa family sensors with two excitation wavelengths are intrinsically ratiometric (Supplementary Figs. 1a–c and 2a–c) and provide the advantages of being quantitative with reduced susceptibility to imaging artifacts or the degree of expression of sensors⁶¹. Compared with other lactate biosensors, FiLa sensors also have a notably lower molecular size (~47 kDa), which is generally beneficial for sensor expression in cells and in vivo, as well as for transgenic studies. Thus, FiLa sensors are superior and fulfill most of the requirements of a near-ideal sensor^39,44,58, similar to the NADH sensor SoNar and the NADPH sensor iNap that we previously reported^48,49,62.

The FiLa-based assays outperformed the current mainstream methods for lactate analysis. For example, chromatography and mass spectrometry require cell or tissue lysis and, therefore, lose spatiotemporal information, and the Seahorse Analyzer infers glycolytic activity based on the extracellular acidification rate measurement. In contrast, FiLa sensors monitor lactate dynamics in real time at subcellular resolution. Moreover, they allow in vivo imaging, which is impossible by the previously existing methods. Importantly, this powerful biosensor assay is much faster and easier to use than the mass spectrometry technology and Seahorse Analyzer. Thus, FiLa sensors provide powerful, broadly applicable tools for defining the spatiotemporal landscape of lactate metabolism in health or disease, and will pave the way for the precise investigation of lactate metabolism in fundamental, translational and clinical research. We anticipate that these sensors will be favored by most researchers in the field of cell metabolism and will be widely used. This approach can also be adapted to different genetically encoded fluorescent sensors using various analytic platforms, such as microplate readers, fluorescence microscopy, high-content imaging systems and in vivo small-animal imaging systems. Owing to the central importance of lactate metabolism, FiLa assays provided by this protocol hold promise to become a favored tool among biologists and translational and clinical researchers.

Overview of the procedure

Here, we provide a detailed protocol for the comprehensive multiscale analysis of lactate metabolic dynamics in vitro and in vivo using our sensitive and highly responsive biosensors. The comprehensive multiscale analysis covers the most commonly used scenarios. The protocol can be divided into four sections. The first section of the protocol describes the preparation and characterization of the FiLa sensor protein (Procedure 1). According to the affinities of FiLa, FiLa-H and FiLa-L, researchers can choose the appropriate sensor for their applications in samples (i.e., organisms, cells or organelles) of interest. The second section of the protocol describes the steps for FiLa sensor-based, real-time subcellular bioenergetic flux analysis by profiling the lactate metabolic response to different nutritional and pharmacological conditions (Procedure 2). Compared with the Seahorse Analyzer by determining extracellular acidification rate (Box 1), FiLa-based measurements provide a systematic-level view of cellular metabolic function at the subcellular scale for the first time. The third section of the protocol describes the steps for imaging lactate dynamics in live mice through a cell microcapsule system (Procedure 3) and recombinant adeno-associated virus (rAAV; Procedure 4). The fourth section of the protocol describes the steps for establishing a rapid and simple assay for evaluating lactate metabolism in human body fluids (i.e., plasma, serum or urine) to manifest the potential use of the FiLa sensor in clinical screening (Procedure 5).

BOX 1. Determination of extracellular acidification rate by Seahorse Analyzer.

TIMING 14–16 h

Procedure

Preparation before the ECAR assay

TIMING 10–12 h

• 1Determine the ECAR by using a Seahorse XFe96 Analyzer. Turn on the equipment at least one day before the assay.

  CRITICAL STEP The Seahorse XFe96 Analyzer should be prewarmed sufficiently to maintain a constant temperature during the assessment.

• 2Seed HEK293 cells in Seahorse XF96 V3 PS cell culture microplate (included in Seahorse XFe96 FluxPak) at a density of 1.2 × 10⁴ cells/80 μL per well 1 d before assay. The background correction wells (A1, A12, H1, H12) should contain only culture medium (no cells).
• 3Hydrate the Seahorse XFe96 sensor cartridge 1 d before the assay. Remove the sensor cartridge and utility plate from the Seahorse XFe96 extracellular flux assay kit (included in Seahorse XFe96 FluxPak) and place the sensor cartridge it upside down. Add 200 μL of double-distilled water to each well of the utility plate and put the sensor cartridge and the lid back onto the plate. To prevent bubbles, the sensor cartridge should first be immersed and raised out of the double-distilled water ten times. Add 25 mL of Seahorse XF Calibrant into a sterile 50 mL tube. Place the utility plate with the sensor cartridge and the tube with calibrant at 37 °C in a humidified, non-CO₂ incubator overnight.

  CRITICAL STEP Ensure the bottom of the sensor cartridge is kept away from any surface to protect the probes inside.

  CRITICAL STEP It is necessary for the incubator to be non-CO₂ and to be humidified properly to prevent evaporation.

  CRITICAL STEP It is necessary to hydrate all of the sensor cartridge wells for the sensor to function correctly and avoiding damage of the sensor caused by drying out.

• 4Prepare 80 mL of glycolysis stress test assay medium, which is sufficient volume for a single test.

Running the ECAR assay

TIMING 2–3 h

• 5Remove the utility plate with the Seahorse XFe96 sensor cartridge and the tube with calibrant from the incubator. Take off the lid and lift the sensor cartridge with the bottom upside down and replace the double-distilled water with 200 μL of calibrant in each well. Place the sensor cartridge and the lid back, eliminate bubbles as described above (step 3 above) and place the sensor cartridge in the 37 °C, humified, non-CO₂ incubator for 45–60 min.
• 6Prepare reagents for the glycolysis stress test. Prewarm the prepared glycolysis stress test assay medium at 37 °C. Dilute glucose, oligomycin and 2-DG to 8 mM, 9 μM and 500 mM, respectively, in the assay medium. Dispense 25 μL of the diluted glucose, oligomycin, 2-DG and blank assay medium into Ports A, B, C and D with a multichannel pipette.

  CRITICAL STEP Ensure the XF assay cartridge is in the correct orientation with the row labels to the left.

  CRITICAL STEP It is necessary to load reagents with tips oriented at the proper angle and to gently dispense compounds and prevent compound leakage through the bottom of the ports.

  CRITICAL STEP Make sure that all of the ports are filled with assay medium.

• 7Open Seahorse Wave Desktop software and modify the assay template files (e.g., group definitions, plate map and protocol of the software files). Start the assay by clicking on the run assay button on the screen, take off the lid and carefully transfer the utility plate with the Seahorse sensor cartridge loaded with drugs to the instrument tray. Allow time calibration, which takes 15–30 min after starting the assay.

  CRITICAL STEP The utility plate with the loaded sensor cartridge must be moved gently to the Seahorse Analyzer to prevent compound leakage through the bottom of the ports.

  CRITICAL STEP Ensure no plate lids are on the cartridge before it enters the Seahorse Analyzer.

  CRITICAL STEP Ensure that the cartridge is loaded appropriately and in the right orientation with the triangular notch in the lower left corner.

• 8After calibrating the Seahorse, remove the cell culture plate from the 37 °C CO₂ incubator. Place the cell culture plate on a clean bench, aspirate 60 μL of culture medium from each well of the plate and rinse the cells with 200 μL of assay medium twice. Add assay medium to a final volume of 175 μL per well. Verify that the cells are adherent to the plate under a microscope, and incubate the plate in a 37 °C, humidified, non-CO₂ incubator for 45–60 min.

  CRITICAL STEP Make sure that the incubation time does not exceed 1 h to maintain a high level of cell viability and avoid cell death.

  CRITICAL STEP Gently rinse the cells to avoid washing them away.

  

• 9Load the cell culture microplate after the calibration is completed and run the Seahorse with a three-injection scheme by sequential injection of three reagents (e.g., glucose, oligomycin and 2-DG). For the glycolysis stress test assay, we perform ECAR measurements for the baseline and after subsequent compound addition. Each well plate is measured three times.

  CRITICAL STEP Ensure there are no plate lids on the microplate before it enters the Seahorse Analyzer.

  CRITICAL STEP Ensure the microplate is loaded appropriately and in the right orientation with the triangular notch in the lower left corner.

  

• 10Load the cell culture microplate after the calibration is completed and run the Seahorse with a three-injection scheme by sequential injection of three reagents (e.g., glucose, oligomycin and 2-DG). For the glycolysis stress test assay, we perform ECAR measurements for the baseline and after subsequent compound addition. Each well plate is measured three times.

  CRITICAL STEP Ensure there are no plate lids on the microplate before it enters the Seahorse Analyzer.

  CRITICAL STEP Ensure the microplate is loaded appropriately and in the right orientation with the triangular notch in the lower left corner.

  

Data analysis

TIMING 0.5–1 h

• 11Perform data analysis as follows:

  • Exclude the values of abnormal background correction wells. An abnormal value, or outlier, can easily be seen if it is clearly different from the other values. For the background correction, run following commands: select the background correction wells and then start the assay by ticking the background correction box on the screen. The software can automatically subtract the ECAR values of the background correction wells from the wells of the samples

  • Calculate the glycolysis parameters, including glycolysis, glycolytic capacity and glycolytic reserve. These can be automatically calculated by the report generator or manually after exporting the data to Excel. In detail, the glycolysis is calculated by subtracting value of the last rate measurement before glucose injection from the value of the maximum rate measurement before oligomycin injection. The glycolytic capacity is calculated by subtracting the last rate measurement before glucose injection from the maximum rate measurement after oligomycin injection. The glycolytic reserve is the difference between glycolytic capacity and glycolysis

  • Normalize the data per well to the number of cells, which is estimated based on seeding density

    PAUSE POINT The data can be stored for analysis at a later timepoint.

    

Applications of the method

FiLa sensors have broad utility, including protein-based assays, live-cell imaging, subcellular organelle imaging, cell-based high-throughput screening, in vivo imaging of live mice and assays of human body fluids (i.e., serum and urine)^39,58. They are able not only to be expressed in living cells and organisms by delivery of the sensor gene for real-time tracking of lactate dynamics but also to measure the lactate levels in vitro (i.e., extracellular environment, lysate or body fluids) via recombinant sensor proteins. Here, we list four typical applications.

The most commonly used application is to monitor the lactate levels in living cells, which usually represent glycolytic activity and often reflect mitochondrial metabolism due to the close link between glycolysis and oxidative phosphorylation. Importantly, the sensor allows measurement at a single-cell resolution and, thus, enables analysis of metabolic heterogeneity between cell subpopulations. For example, immune cells²⁰, stem cells⁶³ and cancer cells² often manifest the association of metabolic features of lactate accumulation with their proliferative and functional states; the FiLa sensor facilitates the isolation and analysis of specific cell subtypes according to their metabolic states. In addition, researchers can use the FiLa sensor as a sensitive metabolic reporter to carry out high-throughput screening for either chemical compounds or genes modulating energy metabolism. Utilizing FiLa, we have developed high-throughput, live cell-based chemical genetic assays and compiled an atlas of subcellular lactate metabolism that reveals lactate as a key hub sensing various metabolic activities³⁹.

The second important application is that FiLa can report the lactate abundance in specific subcellular compartments. Studies on lactate metabolism in specific subcellular compartments may be of biological importance since the transport and metabolism of lactate in mitochondria have long been proposed, but a clear consensus has yet to be reached^9,37. Very recently, the discovery of the lactylation of many proteins in mitochondria and of some in the cytosol and nucleus^8,38 highlights the importance of monitoring lactate metabolism at subcellular resolution. Utilizing FiLa, we demonstrated, for the first time, that lactate is highly enriched in mammalian mitochondria³⁹, and the lactate level in mitochondria is much higher than that in the cytosol and nucleus, which resolved a long-term controversy in the field⁵⁸. Further studies are warranted to identify the key mediators of lactate influx, efflux and metabolic conversion in mitochondria.

The third application involves in vivo monitoring. Lactate transport is highly active, such as shuttling between tumor cells and stromal cells in the tumor microenvironment^3,5, between neurons and astrocytes^20,21, and between the liver and muscle via the Cori cycle⁹. As a result of this active movement, in vivo monitoring of lactate is particularly physiologically relevant. Monitoring this movement has previously been challenging owing to the limitations of available sensors described above. The FiLa sensor is useful for in vivo imaging because of its bright fluorescence and superior sensitivity and, hence, appears to be a suitable choice to overcome this technical challenge. In addition to the procedures introduced in this protocol (a cell microcapsule system and rAAVs), transgenic organisms expressing the FiLa sensor are particularly valuable for investigating the function and mechanisms of lactate metabolism in vivo. Utilizing FiLa, we directly imaged and found elevated lactate levels in live mice with type 1 diabetes mellitus (T1DM) or type 2 diabetes mellitus (T2DM). Furthermore, after insulin injection, mice with T1DM showed an apparent decrease in muscle lactate levels, but mice with T2DM did not, possibly due to the insulin resistance of mice with T2DM³⁹.

Finally, the sensors are capable of quantifying lactate levels in body fluids, such as plasma, serum and urine. Since the lactate levels fluctuate in response to different physiological and disease conditions, quantification appears to be a promising diagnostic tool⁶⁴. For in vitro measurement, body fluid samples are simply diluted and mixed with the sensor protein in 96-well black-bottom plate. Fluorescence intensity is measured immediately by a multimode microplate reader. Lactate levels can then be quantified after calibration of sensor fluorescence in body fluid samples with that of a lactate standard. Compared with the multicomponent enzymatic cycling assay^6,65, the single-component FiLa-based assay requires just one recombinant sensor protein, and is more resistant to cross-reaction with endogenous compounds and variations in ionic strength and temperature. Notably, the FiLa sensor-based assay of body fluid samples also does not require time-consuming sample preparation (i.e., pretreatment or purification) and, therefore, is quite rapid and convenient. Typically, the whole measuring procedure generally takes ~1 min for one sample in a manual assay or ~3 min for 96 samples in an automatic microplate assay. The measurement requires only 0.5 μL of plasma or serum or 2.5 μL of urine for the 96-well plate format with a total volume of 100 μL per well. In addition, the FiLa-H sensor can detect lactate as low as 0.1–0.5 μM, which is at least as sensitive as the best commercially available enzyme-based kit. These advantages make it a promising technology for metabolic diagnostics and screening. Based on the FiLa sensor, we have established a simple, rapid and sensitive lactate assay for point-of-care clinical diagnostics³⁹. Furthermore, patients with maternally inherited diabetes and deafness exhibit elevated lactate levels in urine compared with those of patients with T1DM or T2DM. However, we found that they are frequently misdiagnosed with either T1DM or T2DM, so our sensors allow the rapid determination of this subset of diabetes³⁹.

Limitations of the protocol

Two concerns should be emphasized when applying the FiLa sensors to metabolic research. First, as with other cpYFP-based sensors, FiLa fluorescence, when excited at 485 nm, is sensitive to pH. This pH effect is a limitation for all cpYFP-based sensors^{47–50,66,67}. Fortunately, its fluorescence when excited at 420 nm is pH resistant. In the cytosol, the effects of pH on FiLa and other cpYFP-based sensors are not evident in many circumstances; however, in the mitochondrial matrix, pH often varies depending on the energy metabolism state and the generation of reactive oxygen species. When intracellular pH fluctuations do occur, FiLa fluorescence can be measured with 420 nm excitation only at different lactate ranges³⁹. Alternatively, FiLa could also be fused with mCherry, a pH-insensitive red FP in the physiological pH range, to develop sensors with pH-resistant and ratiometric readouts⁴⁹. It is also possible to develop a dual-function biosensor by fusing FiLa and mKeima, a pH-sensitive red FP⁶⁸, allowing the simultaneous monitoring of lactate and pH in the same sample as a future sensor. In addition, the pH effects on FiLa fluorescence excited at 485 nm may be corrected by measuring the fluorescence of FiLa’s and FiLa-C (a nonresponsive control sensor) in parallel because of their very similar pH sensitivity³⁹. To correct for pH effects, we suggest that users always measure the fluorescence ratio of FiLa alone in cells in parallel with measurements in FiLa-C-expressing cells, and then calculate the pH-corrected ratio $\left(R\right)$ of cells as follows: $R=R_{\text{FiLa}}/R_{\text{FiLa-C}}(R_{\text{FiLa}}$ and $R_{\text{FiLa-C}}$ represent the excitation ratio 485/420 nm for FiLa and FiLa-C, respectively). Notably, FiLa and FiLa-C cannot be deconvolved in the same sample because of their similar fluorescence spectra. Second, although the ratiometric readouts of FiLa provide advantages such as expanding the dynamic range, eliminating the interference from the expression level of the sensor proteins and allowing the quantification of free lactate levels in live cells or in vivo, there are several disadvantages, such as the multicolour imaging application, necessitating filter changes and ratio calculation. When the multicolour microscope with automatic filter change is available, ratiometric imaging, which provides quantitative data, remains the best choice as its benefits outweigh its disadvantages.

Experimental design

Analysis of real-time subcellular lactate flux in single living cells

Cell metabolism is highly compartmentalized in subcellular organelles where metabolic pathways occur. The FiLa gene can be easily delivered into a wide range of cell types by plasmid transfection or recombinant lentivirus infection, and FiLa sensors can also be targeted to different subcellular compartments (i.e., cytosol, nucleus or mitochondria) by tagging with organelle-specific signal peptides. Cells stably expressing sensors are particularly suitable for long-term live-cell imaging. The fluorescence of FiLa sensors is inheritable and readily observable at any time with subcellular resolution. For live-cell fluorescence imaging, we generally seed cells of interest into a 35 mm glass-bottom dish or 96-well glass-bottom plate with culture medium 1 d before imaging. Dual-excitation ratios are acquired using a Leica TCS SP8 SMD confocal laser-scanning microscope system with supersensitive HyD hybrid detectors. Lactate metabolism is highly sensitive to various cellular processes. For long-term imaging, a high-content automated microscope that provides both a highly stable cell culture environment and an efficient imaging system is needed. The BioTek Lionheart FX is equipped with an environment control system, providing a highly stable cell culture environment and enabling the capture of cell state transitions during long-term imaging.

Imaging lactate dynamics in live mice by the cell microcapsule system

For the preparation of cell microcapsules, cells expressing sensors were encapsulated into coherent alginate-poly-L-lysine-alginate beads (400 μm; ~200 cells per capsule) using a BUCHI B-395 Pro encapsulator with parameters that included a 200 μm nozzle with a vibration frequency of 1,300 Hz, a 20 mL syringe and 1,100 V for bead dispersion^69,70 to ensure the quality of the cell microcapsules. Stable cell lines with high sensor expression are preferred for in vivo imaging. The microencapsulated cells expressing FiLa or FiLa-C (~1.2 × 10⁷) should be subcutaneously implanted into the dorsum of mice with a 10 mL syringe to avoid crushing the microcapsules. More microencapsulated cells can be subcutaneously implanted to improve the fluorescence signal of sensors, and the dorsum of mice usually needs to be depilated by a shaver to reduce the fluorescence interference of hairs. Notably, cell microcapsules are not imaged immediately after implantation. Typically, 36–48 h of incubation promotes the body’s absorption of the solution in the cell microcapsule and facilitates higher-quality imaging.

For the in vivo monitoring of the transient lactate response triggered by pharmacological modulations, mice should be treated with compounds of interest via tail vein or intraperitoneal injection. We acquired high-quality images with an IVIS SpectrumCT in vivo imaging system by adjusting the proper exposure time, excitation filters and emission filters.

Imaging lactate dynamics in live mice by rAAV

rAAVs expressing genetically encoded sensors in serotype 9 with the cytomegalovirus (CMV) promoter are the appropriate choice for the gastrocnemius muscle of mice. The final transduction efficiency and kinetics of transgene expression depend on different serotypes, promoters and tissues⁷¹. In addition, robust and reproducible protocols are required to obtain rAAVs with sufficient purity levels and titers high enough for in vivo experiments⁷². In our studies, 8-week-old mice were infected with an adeno-associated virus (AAV) expressing FiLa or FiLa-C via bilateral subcutaneous gastrocnemius injection to evaluate lactate metabolism in vivo. One key to successful experiments is that the virus should be injected at the shallow surface of the gastrocnemius muscle for imaging. To ensure the effective expression of sensors, mouse models were constructed ~4 weeks after AAV injection.

For the construction of the T1DM mouse model, we used mice that were fasted for 16 h and received intraperitoneal injection of 50 mg/kg streptozotocin (STZ) daily for 5 d. Notably, fasted mice with hyperglycemia (>300 mg/dL blood glucose levels) were considered a T1DM model. For the imaging of living mice, the skin over the gastrocnemius muscle should be depilated by a shaver to reduce the fluorescence interference of hair. The gastrocnemius muscle was exposed and placed on glass cover slips for imaging. We acquired high-quality images using a Leica TCS SP8 SMD confocal laser-scanning microscope system with supersensitive HyD hybrid detectors and an HC Plan APO CS2 20×/0.75 NA (numerical aperture) dry objective. For insulin therapy, mice with T1DM that had fasted for ~4 h were given intraperitoneally administration of insulin (1.25 units per kg body weight). To enable kinetic imaging of the gastrocnemius muscle, mice must be sufficiently anesthetized, and insulin must be carefully injected without moving the muscle.

Rapid and simple assay to determine lactate levels in human body fluids

FiLa and FiLa-H sensors have been successfully used in quantifying lactate levels in human body fluids such as serum and urine³⁹. In this protocol, we determined the lactate levels in plasma as an example. To obtain human plasma samples, we recruited ten male tai chi athletes and ten male rugby athletes, 15–22 years of age, from the School of Exercise and Health, Shanghai University of Sport. Written informed consent was obtained from all human subjects, and all procedures related to human research subjects received approval from the School of Exercise and Health, Shanghai University of Sport, and were conducted in strict accordance with institutional guidelines. We collected their venous blood in resting states after fasting overnight and in states after aerobic exercise (tai chi) or anaerobic exercise (rugby). A vacuum blood collection tube with heparin was used, and the blood samples were immediately mixed by gentle inversion and subjected to centrifugation at a low temperature to obtain plasma samples.

Special care should be taken when collecting plasma or serum samples as they are easily contaminated by red blood cell lysis. Blood samples should be contained in sterile tubes, and oscillation of samples should be strictly avoided before centrifugation. To avoid degradation of the tested metabolites, all samples were divided into aliquots, labeled and stored in an −80 °C ultra-low temperature freezer as soon as possible after acquisition. To prevent the evaporation of water and changes in substance concentration in the samples after long-term storage, the tubes were thoroughly sealed with Parafilm. Plasma samples and sensor protein were thawed on ice just before the beginning of the experiment. FiLa family sensors with different affinities could be used to determine lactate levels in various body fluids. For sensor protein-based lactate analysis, four important items are required in the experimental design: (i) a lactate sensor with proper affinity, (ii) a suitable dilution factor of the samples (the unwanted fluorescent substances in body fluids would increase the background value), (iii) a high-quality sensor protein and (iv) a black 96-well (or 384-well) flat-bottom plate. Fluorescence intensity was immediately measured by a Synergy Neo2 multimode microplate reader using the appropriate excitation and emission filters.

Materials

Biological materials

CRITICAL The indicated materials and suppliers listed below can be substituted with appropriate alternatives if necessary.

• Cells of interest: we used NCI-H1299 non-small cell lung cancer cells (Cell Bank of Chinese Academy of Science, cat. no. TCHu160), HEK-293T human embryonic kidney epithelial cells (Cell Bank of Chinese Academy of Science, cat. no. GNHu17) and HEK-293 human embryonic kidney epithelial cells (Cell Bank of Chinese Academy of Science, cat. no. GNHu43)

  CAUTION All cell lines should be regularly checked to ensure that they are authentic and not infected with mycoplasma.

• Mice: we have used 6- to 8-week-old male C57BL/6JGpt mice (GemPharmatech)

  CAUTION Any experiments involving mice must conform to relevant institutional and national regulations. The procedures for the care and use of animals were approved by the East China University of Science and Technology Animal Studies Committee.

• BL21(DE3) chemically competent cells (TansGen Biotech, cat. no. CD601–02)

• Human plasma samples: plasma from athletes before and after tai chi or rugby training was provided by the School of Exercise and Health, Shanghai University of Sport

  CAUTION To avoid cross-contamination, clinical samples should be stored and processed in a dedicated refrigerator, laboratory bench and medical waste bin. Operators should wear hats, masks and gloves during the experiment. Medical wastes should be subjected to uniform autoclaving treatment after the experiment. Written informed consent was notified and obtained from all human subjects, and all procedures related to human research subjects have received approval from the School of Exercise and Health, Shanghai University of Sport, and were conducted in strict accordance with institutional guidelines.

Reagents

CRITICAL The indicated reagents and suppliers listed below can be substituted with appropriate alternatives if necessary.

• pCDFDuet-1 vector (Novagen, cat. no.71340–3)

• pLVX–IRES–Puro vector (Clontech, cat. no.632186)

• pcDNA3.1–Fila (FR Biotechnology, cat. no. FLA1001)

• pcDNA3.1–Fila-L (FR Biotechnology, cat. no. FLL1001)

• pcDNA3.1–Fila-H (FR Biotechnology, cat. no. FLH1001)

• pcDNA3.1–Fila-C (FR Biotechnology, cat. no. FLC1001)

• Pfu polymerase (TIANGEN, cat. no. KP201–02)

• Dulbecco’s Modified Eagle’s Medium (DMEM; Corning, cat. no. 10–013-CV)

• RPMI 1640 medium (Corning, cat. no. 10–040-CV)

• FBS (Biological Industries, cat. no. 04–001-1ACS)

• Penicillin–streptomycin solution (Cytiva, cat. no. SV30010)

• 0.25% (wt/vol) Trypsin–ethylenediaminetetraacetic acid (Gibco, cat. no. 25200072)

• PBS solution (Biosharp, cat. no. BL302A)

• Dimethyl sulfoxide (DMSO; Sigma, cat. no. D2650)

• Hieff trans liposomal transfection reagent (Yeasen, cat. no. 40802ES03)

• Opti-MEM I reduced-serum medium (Gibco, cat. no. 31985070)

• Polybrene (Sigma, cat.no. H9268)

• Puromycin (Yeasen, cat.no. 60210ES25)

• Human fibronectin (Yeasen, cat. no. 40105ES08)

• Phenol red and glucose-free DMEM medium (Bio-Channel, cat. no. BC-M-035–500 mL)

• Seahorse XF (extracellular flux) DMEM medium pH 7.4 (Agilent, cat. no. 103575–100)

• Seahorse XF calibrant (Agilent, cat. no. 100840–000)

• D-Glucose (Glc; Macklin, cat. no. D810594)

• L-Glutamine (Macklin, cat. no. L810391)

• Sodium pyruvate (Aladdin, cat. no. S104174)

• Coomassie brilliant blue G250 (Macklin, cat. no. C6232)

• Ethanol (Shanghai Titan Scientific, cat. no. 01226776)

• Phosphoric acid (Macklin, cat. no. P816337)

• BSA (Yeasen, cat. no. 36101ES60)

• Isopropyl-β-D-thiogalactoside (IPTG; Macklin, cat. no. I811719)

• Tryptone (Oxoid, cat. no. LP0042)

• Agarose (Yeasen, cat. no. 10208ES60)

• Yeast extract (Oxoid, cat. no. LP0021)

• Streptomycin (Yeasen, cat. no. 60211ES25)

• Imidazole (Shanghai D&B Biological Science and Technology, cat. no. K392001)

• Na₃PO₄ (Shanghai Titan Scientific, cat. no. G81528A)

• NaCl (Shanghai Titan Scientific, cat. no. G81793F)

• KCl (Shanghai Titan Scientific, cat. no. G80636B)

• KH₂PO₄ (Shanghai Titan Scientific, cat. no. G82821B)

• Na₂HPO₄ (Shanghai Titan Scientific, cat. no. G81291A)

• MgSO₄ · 7H₂O (Yonghua Chemical Technology, cat. no. 213602129)

• CaCl₂ (Vetec, cat. no. V900266)

• NaHCO₃ (Macklin, cat. no. S818080)

• HEPES (Macklin, cat. no. H811129)

• 4-Morpholinepropanesulfonic acid (MOPS; Macklin, cat. no. M813154)

• NaOH (Sinopharm Chemical Reagent, cat. no. 10019718)

• Sodium L-lactate (Sigma, cat. no. 71718)

• β-Nicotinamide adenine dinucleotide, oxidized form (NAD⁺; Yeasen, cat. no. 60323ES08)

• β-Nicotinamide adenine dinucleotide, reduced form (NADH; Yeasen, cat. no. 60301ES03)

• β-Nicotinamide adenine dinucleotide phosphate, oxidized form (NADP⁺; Yeasen, cat. no. 60324ES03)

• β-Nicotinamide adenine dinucleotide phosphate, reduced form (NADPH; Yeasen, cat. no. 60302ES01)

• Adenosine 5′-diphosphate (ADP; Yeasen, cat. no. 60604ES03)

• Adenosine 5′-triphosphate (ATP; Sigma, cat. no. A26209)

• Phosphoenolpyruvic acid monopotassium salt (PEP; Alfa Aesar, cat. no. B20358)

• Sodium citrate (Aladdin, cat. no. S189183)

• dl-Isocitric acid trisodium salt hydrate (Sigma, cat. no. I1252)

• α-Ketoglutaric acid disodium salt dihydrate (Sigma, cat. no. 75892)

• Succinic acid (Aladdin, cat. no. S108855)

• Fumaric acid (Aladdin, cat. no. F118459)

• β-Hydroxybutyric acid (BHB; MedChemExpress, cat. no. HY-113378)

• Oxaloacetic acid (Sigma, cat. no. O4126)

• Sodium acetate (Shanghai Titan Scientific, cat. no. G18900B)

• L-Malic acid (Sigma, cat. no. 112577)

• Sodium oxamate (Sigma, cat. no. O2751)

  CAUTION Sodium oxamate is an inhibitor of glycolysis. Wear gloves, a mask, safety glasses and a laboratory coat when handling sodium oxamate.

• Rotenone (Sigma, cat. no. R8875)

  CAUTION Rotenone is a potent inhibitor of mitochondrial complex I. Wear gloves, a mask, safety glasses and a laboratory coat when handling rotenone.

• Oligomycin (Oligo; MedChemExpress, cat. no. HY-16589)

  CAUTION Oligomycin is a potent inhibitor of mitochondrial complex V. Wear gloves, a mask, safety glasses and a laboratory coat when handling oligomycin.

• 2-Deoxy-D-glucose (2-DG; MedChemExpress, cat. no. HY-13966)

  CAUTION 2-DG is an inhibitor of glycolysis. Wear gloves, a mask, safety glasses and a laboratory coat when handling 2-DG.

• Ionomycin (Aladdin, cat. no. I139530)

  CAUTION Ionomycin is a calcium ionophore. Wear gloves, a mask, safety glasses and a laboratory coat when handling ionomycin.

• AR-C155858 (MedChemExpress, cat. no. HY-13248)

  CAUTION AR-C155858 is a monocarboxylate transporter inhibitor. Wear gloves, a mask, safety glasses and a laboratory coat when handling AR-C155858.

• AZD-3965 (MedChemExpress, cat.no. HY-12750)

  CAUTION AZD-3965 is a monocarboxylate transporter inhibitor. Wear gloves, a mask, safety glasses and a laboratory coat when handling AZD- 3965.

• SB-204990 (MedChemExpress, cat.no. HY-16450)

  CAUTION SB-204990 is an ATP citrate–lyase inhibitor. Wear gloves, a mask, safety glasses and a laboratory coat when handling SB-204990.

• Poly-L-lysine (Sigma, cat. no. P7890)

• Sodium alginate (BUCHI, cat. no. 11059994)

• Insulin solution (Sigma, cat. no. I9278)

• STZ (Sigma, cat. no. S0130)

  CAUTION STZ is a potent DNA-methylating antibiotic. Wear gloves, a mask, safety glasses and a laboratory coat when handling STZ.

• Sodium pentobarbital (Sigma, cat. no. P3761)

  CAUTION Sodium pentobarbital is a general anesthetic agent, and it may impair human thinking or reactions. Wear gloves, a mask, safety glasses and a laboratory coat when handling sodium pentobarbital.

• Normal saline (Meilunbio, cat. no. MA0083)

• Depilatory cream (Mayllie, cat. no. TOMO-01)

Equipment

• Automated microscope system (Agilent BioTek, model no. Lionheart FX)

• Leica confocal laser scanning microscope system (Leica, model no. TCS SP8 SMD)

• Stage top incubator (PeCon, model no. Incubator PM 2000 RBT)

• SpectrumCT In Vivo Imaging System (PerkinElmer, model no. IVIS)

• Automated cell counter (Isogen Life Science, model no. Countstar IC1000)

• Multimode microplate reader (Agilent BioTek, model no. Synergy Neo2)

• Universal mounting frame (PeCon, model no. Universal mounting frame AK)

• Encapsulator (BUCHI, model no. B-395 Pro)

• Blood glucose monitor (Microtech Medical, model no. Beiwen IIS)

• Fluorescent spectrometer (PerkinElmer, model no. FL6500)

• pH meter (Mettler Toledo, cat. no. FE28)

• Seahorse XFe96 analyzer (Agilent, cat. no. S7800B)

• Seahorse XFe96 FluxPak (Agilent, cat. no. 102601–100)

• CO₂ incubator (Thermo, cat. no. 4111)

• Thermostatic chamber (SAST, cat. no. PD-12)

• Ultrasonic homogenizer (SCIENTZ, model. no. SCIENTZ-IID)

• Prepacked column with Sephadex G-25 medium (GE Healthcare, cat. no. 17003301)

• NTA (nitrilotriacetate) column (Beyotime, cat. no. FCL30)

• Ni Sepharose 6 Fast Flow resin (GE Healthcare, cat. no. 17531804)

• 4-Chamber 35 mm glass-bottom dish (Cellvis, cat. no. D35C4–20-1-N)

• 96-well glass-bottom plate (Cellvis, cat. no. P96–1-N)

• Clear 96-well flat-bottom plate (Leiloer, cat. no. 4209601)

• Black 96-well flat-bottom plate (WHB, cat. no. 96–02)

• 0.22 μm filter membrane (Merck, cat. no. SLGP033RB)

• 0.45 μm filter membrane (Merck, cat. no. SLGP033RA)

• Picus electronic pipette, 8 channel, 10–300 μL (Sartorius, cat. no. 735361)

• Picus electronic pipette, 8 channel, 50–1,200 μL (Sartorius, cat. no. 735391)

• Proline plus mechanical pipette, 8 channel (Sartorius, cat. no. 728140)

• 10, 50, 200 and 1,000 μL manual pipettes (Sartorius, cat. nos. 720015, 720025, 720070 and 720060)

• Optifit flexibulk nonsterile pipette tips, 350 μL (Sartorius, cat. no. LH-B790354)

• Optifit flexibulk nonsterile pipette tips, 1,200 μL (Sartorius, cat. no. LH-B791204)

• 1, 10, 20 and 50 mL disposable syringe with a needle (Zhonghe, cat. nos. HM-0541, HM-0544, HM-0545 and HM-0546)

• Centrifuge tube, 1.5 mL (Sangon Biotech, cat. no. F601620–9001)

• Vacuum blood collection tube with heparin (BD, cat. no. 367878)

• Electric shaver (Codos, cat. no. CP-6800)

• ImageJ software (ImageJ 1.50d, https://imagej.nih.gov/ij/)

• Leica Application Suite X 1.8.1.13759 (https://www.leica-microsystems.com/products/microscope-software/p/leica-las-x-ls/downloads/)

• GraphPad Prism v8.4.0 (https://www.graphpad.com/scientific-software/prism/)

• Gen5 3.03 (https://www.biotek.com/products/software-robotics/)

• Wave Desktop (https://www.agilent.com.cn/zh-cn/products/cell-analysis/software-download-for-wave-desktop)

• SnapGene (https://download.cnet.com/s/snapgene/)

Reagent setup

CRITICAL All buffers should be sterilized using a 0.22 μm filter membrane.

FiLa sensors

Amplify the coding sequences of FiLa sensors from pcDNA3.1–FiLa plasmids (FR Biotechnology) using Pfu polymerase; users should design their own primers for FiLa sequences using SnapGene software with consideration of general rules, such as primer length (typically 18–48 nt), GC% (typically 40–60%), melting temperature (typically 45–65 °C), etc. Subclone into the pCDFDuet-1 vector (Novagen) using restriction sites BamHI and HindIII for prokaryotic expression or into the pLVX–IRES–Puro (Clontech) vector behind a Kozak sequence (GCCACC) using restriction sites EcoRI and NotI for mammalian expression. For nuclear targeting, fuse the threefold nuclear localization sequence, (DPKKKRKV)₃, at the C terminus of the sensor. For mitochondrial targeting, fuse a signal sequence from Neurospora crassa ATP synthase at the N terminus of sensors⁷³. All constructs, including different versions of FiLa, FiLa-L, FiLa-H, FiLa-C, are available commercially (FR Biotechnology, https://www.fr-biotechnology.com/).

E. coli BL21(DE3) cells expressing FiLa sensors

E. coli BL21(DE3) cells expressing FiLa sensors were obtained by transforming FiLa sensor plasmid of pCDFDuet-1 vector into bacteria cells. Pipette 2 μL of FiLa sensor plasmid and 10 μL of E. coli BL21(DE3) cells thawed on an ice bath to the 1.5 mL sterilized centrifuge tube, gently mix and place in an ice bath for 30 min. Then, heat the tube for 45 s in a water bath at 42 °C and place it in the ice bath for 2 min. Add 500 μL of LB medium to the centrifuge tube subsequently. Mix and culture the tube at 37 °C for 1 h. Expand the transformed bacterial cells by further culture in LB medium containing 50 μg/mL streptomycin.

Mammalian cells expressing FiLa sensors

To generate the stable mammalian cell lines expressing FiLa sensors, construct the the pLVX lentiviral plasmids encoding FiLa sensor. Produce lentivirus by cotransfecting two lentiviral packaging vectors (pMD2.G and psPAX2) in HEK-293T cells. Collect lentiviral supernatants 48 and 72 h after transfection. Infect HEK-293 or H1299 cells in 6-well tissue culture plates in media containing 8 μg/mL polybrene and centrifuged at 1,000g for 1 h. After infection, remove the virus, and select cells with 0.2–1 μg/mL puromycin for 1 week.

rAAVs

AAVs expressing FiLa or FiLa-C in serotype 9 with CMV promoter were generated and purified by OBiO Technology. For users, acquire the pcDNA3.1–FiLa plasmid from FR Biotechnology, and provide it to OBiO Technology or similar suppliers to construct the pAAV–CMV–MCS plasmids encoding the FiLa sensor and generate the rAAV. Aliquot and store the virus at −80 °C for 6 months.

CRITICAL Avoid repeated freeze—thaw cycles, as this can reduce viral titers.

H1299 cell culture medium

H1299 cell culture medium is RPMI 1640 medium supplemented with 10% (vol/vol) FBS and 1% (vol/vol) penicillin–streptomycin. Store at 4 °C for up to 1 month and preheat to 37 °C before use.

HEK-293 and HEK-293T cell culture medium

To make HEK-293 and HEK-293T cell culture medium, supplement DMEM with 10% (vol/vol) FBS and 1% (vol/vol) penicillin–streptomycin. Store at 4 °C for up to 1 month and preheat to 37 °C before use.

Fibronectin

Prepare 1 mg/mL fibronectin stock solution in PBS. Store in small aliquots at −80 °C for up to 6 months. Dilute with PBS to the final concentration of 10 μg/mL before use. Dilutions should be freshly prepared.

To coat the imaging dish or plate, add 200 μL or 50 μL of 10 μg/mL fibronectin onto a four-chamber 35 mm glass-bottom dish or 96-well glass-bottom plate per well, respectively. Incubate the dish or plate at 37 °C for ~2 h. Remove the solution before use.

HBSS buffer

Prepare HBSS buffer (100 mM HEPES, 5.4 mM KCl, 136.7 mM NaCl, 0.44 mM KH₂PO₄, 0.35 mM Na₂HPO₄, 0.81 mM MgSO₄, 1.26 mM CaCl₂, 4.2 mM NaHCO₃) in ultrapure water, adjust the pH to 7.4 with 4 M NaOH and filter sterilize by 0.22 μm filter membrane. Store at 4 °C for 1 month and preheat to 37 °C before use.

HBSS buffer with 25 mM D-glucose (HBSS-G buffer)

Supplement HBSS buffer (pH 7.4) with 25 mM D-glucose. Prepare fresh and preheat to 37 °C before use.

HEPES buffer

Prepare HEPES buffer (100 mM HEPES and 100 mM NaCl) in ultrapure water, adjust pH to 7.4 using 4 M NaOH and filter sterilize by 0.22 μm filter membrane. Store at 4 °C for up to 1 month.

Citrate buffer

Prepare liquid A by dissolving 2.1 g of citric acid powder in 100 mL ultrapure water. Prepare liquid B by dissolving 2.94 g of trisodium citrate powder in 100 mL ultrapure water. Mix liquid A and liquid B in the volume ratio 1:1.32 before use, and adjust the pH to 4.2–4.5. Prepare fresh and precool on ice before use.

D-Glucose stock solution

Prepare 1 M D-glucose stock solution in phenol red and glucose-free DMEM medium or Seahorse XF DMEM medium and filter sterilize by 0.22 μm filter membrane. Store at 4 °C for up to 1 month.

L-Glutamine stock solution

Prepare 200 mM L-glutamine stock solution in Seahorse XF DMEM medium and filter sterilize by 0.22 μm filter membrane. Store at 4 °C for up to 1 month.

Seahorse glycolysis stress test assay medium

Take 80 mL of Seahorse XF DMEM medium and add 800 μL of 200 mM L-glutamine stock solution resulting in assay medium containing 2 mM L-glutamine. Prepare Seahorse media fresh on day of experiment or the day before the assay and keep the medium at 4 °C. Preheat to 37 °C just before use.

Buffer A

Prepare buffer A (20 mM Na₃PO₄, 500 mM NaCl and 10 mM imidazole) in ultrapure water, adjust pH to 7.4 using 1 M NaOH. Keep at 4 °C for up to 1 month.

Buffer B

Prepare buffer B (20 mM Na₃PO₄, 500 mM NaCl and 500 mM imidazole) in ultrapure water, adjust pH to 7.4 using 1 M NaOH. Keep at 4 °C for up to 1 month.

Wash buffer (34 mM imidazole)

Mix buffer A and buffer B in the volume ratio 19:1 before use. Prepare fresh.

Eluent buffer (300 mM imidazole)

Mix buffer A and buffer B in the volume ratio 2:3 before use. Prepare fresh.

LB

Dissolve 10 g of NaCl, 10 g of tryptone, and 5 g of yeast extract in 1 L of ultrapure water. Immediately after preparation sterilize by autoclaving. Store at 4 °C for up to 1 month.

Streptomycin (100 mg/mL)

Dissolve 5 g of streptomycin in 50 mL of ultrapure water. Store in small aliquots at −20 °C for up to 6 months.

Bradford solution

Dissolve 50 mg of Coomassie Brilliant Blue G250, 50 mL of ethanol and 60 mL of phosphoric acid in 500 mL of ultrapure water, store at room temperature (20–25 °C) and keep away from light for up to 1 month.

BSA standard solutions

Dissolve 14.4 mg of BSA powder in 10 mL of ultrapure water. Dilute the BSA solution of 1.44 mg/mL to serial concentrations of (0, 0.288, 0.576, 0.864 and 1.152 mg/mL) in the ultrapure water to prepare protein standards. Store at −20 °C for up to 6 months.

Oxamate

Freshly prepare 200 mM oxamate stock solution in HBSS-G. Immediately dilute to the final concentration of 10 mM before use.

Oligomycin

Prepare 10 mM oligomycin stock solution in DMSO. Store in small aliquots at −80 °C for up to 6 months. Dilute with HBSS to the final concentration of 1 μM before use. Prepare fresh.

Ionomycin

Prepare 10 mM ionomycin stock solution in DMSO. Store in small aliquots at −80 °C for up to 6 months. Dilute with HBSS-G to the final concentration of 2 μM before use. Prepare fresh.

Rotenone

Prepare 10 mM rotenone stock solution in DMSO. Store in small aliquots at −80 °C for up to 6 months. Dilute with HBSS-G to the final concentration of 10 μM or 0.5 μM before use. Prepare fresh.

2-DG

Freshly prepare 500 mM 2-DG stock solution in HBSS for FiLa-based glycolysis stress assay, or in Seahorse glycolysis stress test assay medium for the extracellular acidification rate (ECAR) assay, respectively. Immediately dilute to a final concentration of 50 mM before use.

AR-C155858

Prepare 10 mM AR-C155858 stock solution in DMSO. Store in small aliquots at −80 °C for up to 6 months. Dilute with HBSS-G to the final concentration of 2 μM before use. Prepare fresh.

SB-204990

Prepare 30 mM SB-204990 stock solution in DMSO. Store in small aliquots at −80 °C for up to 6 months. Dilute with normal saline to 5 mM and adjust the pH to 7.4 with 4 M NaOH before use. Prepare fresh.

AZD-3965

Prepare 10 mM AZD-3965 stock solution in DMSO. Store in small aliquots at −80 °C for up to 6 months. Dilute with normal saline to 1 mM and adjust the pH to 7.4 with 4 M NaOH before use. Prepare fresh.

L-Lactate

Freshly prepare 500 mM L-lactate stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–50 mM (0, 0.85, 2.54, 7.62, 22.86, 68.59, 205.76, 617.28, 1,851.85, 5,555.56, 16,666.67 and 50,000 μM; or 0, 0.05, 0.19, 0.76, 3.05, 12.21, 48.83, 195.31, 781.25, 3,125, 12,500 and 50,000 μM) before use.

NAD⁺

Freshly prepare 20 mM NAD⁺ stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

NADH

Freshly prepare 20 mM NADH stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

NADP⁺

Freshly prepare 20 mM NADP⁺ stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

NADPH

Freshly prepare 20 mM NADPH stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

ADP

Freshly prepare 20 mM ADP stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

ATP

Freshly prepare 20 mM ATP stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

PEP

Freshly prepare 20 mM PEP stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

Citrate

Freshly prepare 20 mM citrate stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

Isocitrate

Freshly prepare 20 mM isocitrate stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

α-Ketoglutarate

Freshly prepare 20 mM α-ketoglutarate stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

Succinate

Freshly prepare 20 mM succinate stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

Fumarate

Freshly prepare 20 mM fumarate stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

BHB

Freshly prepare 20 mM BHB stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

Oxaloacetate

Freshly prepare 20 mM oxaloacetate stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

Acetate

Freshly prepare 20 mM acetate stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

Malate

Freshly prepare 20 mM malate stock solution in HEPES buffer (pH 7.4). Immediately dilute with HEPES buffer (pH 7.4) to a final concentration of 0–10 mM (0, 0.1, 0.3, 1, 3 and 10 mM) before use.

MOPS solution

Prepare MOPS solution (10 mM MOPS, 145.3 mM NaCl) in ultrapure water. Adjust pH to 7.2 with 4 M NaOH. Store at 4 °C for 1 month.

Polymerization solution

Prepare polymerization solution (10 mM MOPS, 100 mM CaCl₂) in ultrapure water. Adjust pH to 7.2 with 4 M NaOH. Store at 4 °C for 1 month.

Depolymerization solution

Prepare depolymerization solution (10 mM MOPS, 76.9 mM NaCl, 50 mM sodium citrate) in ultrapure water. Adjust the pH to 7.2 with 4 M NaOH. Store at 4 °C for 1 month.

1.5% (wt/vol) Sodium alginate solution

Prepare 1.5% sodium alginate solution by dissolving 1.5 g of sodium alginate powder into 100 mL MOPS solution. Store at 4 °C for 1 month.

CRITICAL Stir overnight and sterilize with a 0.45 μm filter membrane.

0.03% (wt/vol) Sodium alginate solution

Dilute 5 mL 1.5% (wt/vol) sodium alginate solution in 250 mL MOPS solution. Store at 4 °C for 1 month.

0.05% (wt/vol) Poly-L-lysine solution

Dissolve 73 mg poly-L-lysine powder into 150 mL MOPS solution. Prepare fresh.

Insulin

Dilute the insulin solution with PBS solution to a final concentration of 1.25 units/mL. Prepare fresh.

Sodium pentobarbital

Dissolve 0.2 g of sodium pentobarbital powder in 10 mL normal saline to a final concentration of 20 mg/mL. Store at 4 °C for up to 1 week.

STZ

Dissolve 5 mg STZ powder in 1 mL of precooled citrate buffer. Prepare fresh, keep on ice, avoid illumination and use within 30 min.

Equipment setup

Lionheart FX automated microscope system

The Agilent BioTek Lionheart FX automated microscope system is equipped with an environment control cover; a CO₂/O₂ gas control; a humidity chamber; Olympus Plan Fluorite 4× 0.13 NA, 10× 0.3 NA, 20× 0.45 NA, 40× 0.6 NA objectives; a laser autofocus; four user-replaceable light-emitting diodes (LEDs; available wavelengths: 365, 390, 405, 465, 505, 523, 590, 623, 655 and 740 nm); four user-replaceable fluorescence cubes; a Sony CMOS 16-bit grayscale camera and Gen5 software control. For dual-excitation ratio imaging, use 405 nm and 465 nm LED light sources, 400 BP 40 nm or 469 BP 35 nm bandpass (BP) excitation filters, and 550 BP 49 nm or 525 BP 39 nm emission filters.

Leica TCS SP8 confocal laser scanning microscope system

The Leica confocal microscope system should be set up to include the following components: a Leica DMI6000 inverted microscope equipped with supersensitive HyD hybrid detectors, universal mounting frame AK, stage top incubator, an HC Plan Fluotar 5×/0.15 NA objective, an HC Plan APO CS2 20×/0.75 NA objective, an HC Plan Apo CS 40×/0.85 NA objective and an HC Plan Apo CS2 63×/1.40 NA oil objective, an anti-vibration table, image-based autofocus for transmission light combined with adaptive focus control (AFC), lasers (diode, 50 mW: 405 nm; Ar, 65 mW: 488 nm; diode-pumped solid-state laser, 20 mW: 561 nm), photomultiplier tube detection for imaging, an EL6000 fluorescence lamp and the software platform Leica Application Suite X.

Seahorse XFe96 analyzer

The Seahorse XFe96 analyzer, consisting of a detecting instrument and a controller, is suitable for assays in 96-well plate formats. The detecting instrument is equipped with a four-dosing port system with an automatic mixing function and a precise temperature control heating tray that can be maintained at 16–42 °C: turn on at least 1 d before the assay to prewarm the equipment. The recommended loading volume in the ports is 25 μL, and the recommended assay volume is 150–275 μL per well. The controller is combined with a Windows 10 computer, a touch screen monitor and Wave software, which can be used to control the instrument and to acquire and analyze data. Wave software contains several default assay template files to design the experiment and customize when necessary. In combination with the Seahorse XFe96 FluxPaks, the oxygen consumption rate values and ECAR values of living cells under different conditions can be monitored in real time.

Encapsulator B-395 pro system

The BUCHI Encapsulator B-395 pro system consists of a control unit and a reaction unit (Supplementary Fig. 6). The control unit is equipped with a syringe pump, a magnetic stirrer, a vibration setting, electrical and pneumatic systems, and a control panel. The reaction unit forms a closed, autoclavable unit in which the beads are formed under sterile conditions, which is used for the sterile production and collection of microcapsules, and includes a reaction vessel, a bead producing module, a bead collecting flask (glass bottle with lid) and a liquid waste flask, with valves, fittings, tubes and filters. For in vivo imaging, use the 0.20 mm nozzle of the syringe pump to keep the bead size in range of ~0.15–2.0 mm. The integrated magnetic stirrer bead and the closed reaction unit should be sterilized as a whole by autoclaving before use.

IVIS SpectrumCT in vivo imaging system

The PerkinElmer IVIS SpectrumCT in vivo imaging system integrates optical and micro-computed tomography modalities with three-dimensional tomography for fluorescence and bioluminescence, enhanced spectral unmixing for multispectral imaging and dynamic contrast enhancement, including a back-illuminated grade 1 charge-coupled device camera (−90 °C), X-ray detector, low-dose micro-computed tomography, heated chamber, gas anesthesia ports, injector ports, constant horizontal gantry motion, flat panel detector, stable revolving animal platform table and 10 excitation fluorescence filters and 18 emission fluorescence filters. For dual-excitation ratio imaging, use 430 BP 30 nm and 500 BP 30 nm excitation filters and 540 BP 20 nm emission filters.

Procedure

Procedure 1: preparation and characterization of FiLa sensors

TIMING 2–3 d

1. Culture E. coli BL21(DE3) cells carrying the pCDFDuet–FiLa expression plasmid in 100 mL of LB medium containing 50 μg/mL streptomycin at 37 °C until the OD₆₀₀ value reaches ~0.4–0.6. Induce protein expression in the recombinant bacteria with 1 mM IPTG at 18 °C for 36 h.

   CRITICAL STEP The OD600 value must be between 0.4 and 0.6. Otherwise, it will be difficult to express the sensor protein.

2. Collect bacteria by centrifugation at 4,000g for 30 min at 4 °C, and suspend the cell pellets in 25 mL of buffer A.

3. Lyse the bacteria via ultrasonication. We use a total sonication time of 600 s with a pattern of 1 s sonication and a 3 s rest period. Set the number of cycles to four and set the amplitude percentage to 45%. Keep cells on ice during the sonication process.

   CRITICAL STEP Cell pellets should be thoroughly sonicated.

   CRITICAL STEP The lysate should be kept at low temperature while sonicating.

   

4. Purify proteins using a column prepacked with Ni Sepharose 6 fast flow resin (2 mL). First, load the lysate onto the column. Then, wash the column with five column volumes (10 mL) of wash buffer, and discard the wash buffer. Elute the proteins with two column volumes (4 mL) of eluent buffer, and then collect the eluent in 1.5 mL centrifuge tubes, with a final volume in each tube of 1 mL.

5. Desalt the protein preparations by loading the eluent onto a prepacked column with Sephadex G-25 medium and exchange the buffer to HEPES buffer (pH 7.4) for in vitro characterization. Process 1 mL of protein preparation for a prepacked column. In detail, pretreat the column with 15 mL of ultrapure water and 15 mL of HEPES buffer, respectively. Discard the eluent, and then load 1 mL of protein preparation onto the column. Elute the proteins with 3 mL HEPES buffer (pH 7.4). Determine the concentration of purified protein, aliquot the protein in 1.5 mL centrifuge tubes.

   PAUSE POINT The purified protein can be stored at −80 °C for up to 6 months for analysis at a later timepoint.

6. Determine the concentration of purified protein by the Bradford assay. In detail, perform each assay by combining 5 μL of BSA standard solutions or purified protein and 195 μL of Bradford solution in a clear 96-well flat-bottom plate. Incubate the plate at room temperature for 3 min and measure the absorbance at 595 nm using a Synergy Neo2 multimode microplate reader. Calculate the protein concentration according to the standard curve of the BSA standard solutions, and transform the weight/volume percent concentration to molarity/volume percent concentration according to the molecular weight of the sensor (Table 1).

7. For the in vitro measurement, dilute the purified sensor protein to 0.4 μM in HEPES buffer just before the assay and perform each assay by combining 50 μL of the different concentrations of lactate and 50 μL of sensor protein (0.4 μM) in a black 96-well flat-bottom plate (final sensor protein concentration: 0.2 μM). Immediately measure the fluorescence intensity using a Synergy Neo2 multimode microplate reader with 420 BP 27 nm or 485 BP 20 nm excitation and 532 BP 40 nm emission bandpass filters.

   CRITICAL STEP For titration experiments, sensor concentrations must be as low as possible (i.e., 0.2 μM) to obtain accurate binding and dissociation constants. High sensor concentrations lead to sensory affinity overestimation in titration experiments.

8. Export fluorescence data. Correct the fluorescence by subtracting the background fluorescence values of samples without sensor protein. Calculate the ratio ($R_{485/420}$) of fluorescence excited at 485 nm to that excited at 420 nm.

Procedure 2: analysis of real-time subcellular lactate flux in single living cells

TIMING 1–2 d

1. Seed ~8,000 or 60,000 HEK-293 cells expressing FiLa, FiLa-Nuc and FiLa-L-Mit on fibronectin coated 96-well glass-bottom plates or four-chamber 35 mm glass-bottom dishes, respectively. Incubate the cells at 37 °C in a humidified atmosphere using a CO₂ incubator for 12–24 h. Cells should reach 30–50% confluency before imaging.

   CRITICAL STEP Typically, we suggest that researchers use FiLa-C, FiLa-C-Nuc and FiLa-C- Mit as the control sensors.

   CRITICAL STEP Coat the imaging plate or dish with fibronectin, collagen I, collagen IV, laminin or fibrinogen according to the cell type^71,72 to prevent cell detachment.

   CRITICAL STEP Transient transfection usually causes lower cell viability and is not ideal for long-term imaging. We recommend making stable cell lines using lentiviral infection.

   

2. For imaging subcellular lactate metabolism under different nutritional and pharmacological conditions, please follow option A and option B, respectively.

   1. Imaging subcellular lactate metabolism under different nutritional conditions
      CRITICAL The steps described here are for use with 96-well glass-bottom plates.

      1. Prepare assay medium with different nutritional conditions by combining various concentrations of glucose (0, 0.1, 1 and 25 mM) and FBS (1%, 5% and 10% (vol/vol)) in phenol red and glucose-free DMEM
      2. Before the experiment, remove the culture medium, rinse cells twice with 100 μL phenol red and glucose-free DMEM, and then add 100 μL of medium containing the various dilutions of glucose and FBS to each well for imaging
         CRITICAL STEP All operations must be carefully and gently performed to avoid dislodging the cells from the glass surface.
      3. Get ready to acquire dual-excitation fluorescence images of cells using a Lionheart FX automated microscope system with a Plan Fluorite 20× 0.45 NA objective, 405 nm and 465 nm LED sources with 400 BP 40 nm or 469 BP 35 nm bandpass excitation filters and 550 BP 49 nm or 525 BP 39 nm emission filters. First, set the temperature (37 °C) and CO₂ concentration (5%) 30 min before imaging
      4. Next, load cells into the Lionheart FX automated microscope system equipped with a 96-well plate adapter. Set optimal parameters for imaging, including light intensity, exposure time, gain and time intervals. Active laser autofocus and set a standard focal plane for each plate. Choose two to four fields with moderately fluorescent cells. Set the time interval to 30 min and perform imaging for 1–2 h
         CRITICAL STEP Measure cells stably expressing FiLa-C, FiLa-C-Nuc and FiLa-C-Mit treated with different nutritional conditions on the same plate in parallel experiments to account for the pH effects.
         CRITICAL STEP Make sure that temperature and CO₂ concentration reach standard requirements during the entire process. Disturbance of the temperature or CO₂ concentration would cause undesirable deviation.
         
   2. Imaging subcellular lactate metabolism under different pharmacological conditions
      CRITICAL The steps here describe use with both four-chamber 35 mm glass-bottom dishes and 96-well glass-bottom plates.

      1. Before the assay, prepare the experimental reagents (common agents for perturbing cell metabolism) listed in Table 2, as described in the ‘Reagent setup’ section
      2. Before the experiment, remove the culture medium, rinse cells twice with 500 μL if using four-chamber 35 mm glass-bottom dishes (or 100 μL per well for 96-well glass-bottom plate for high throughput) HBSS or HBSS- G, and then add 500 μL (or 100 μL per well for 96-well glass-bottom plates) HBSS or HBSS-G to each well for imaging. For high-throughput assays, proceed to step (vi)
         CRITICAL STEP All operations must be carefully and gently performed to avoid dislodging the cells from the glass surface.
         
      3. Maintain cells at 37 °C for 10–20 min in a stage top incubator. Then, prepare to image cells with a Leica TCS SP8 SMD confocal laser scanning microscope system with a 63× oil objective. First, place plates on the microscope stage and set the imaging parameters (i.e., laser power, pinhole size, photomultiplier tube sensitivity and time intervals)
         CRITICAL STEP High axial resolution imaging of mitochondria may be obtained with a smaller pinhole.
         CRITICAL STEP Shorter exposure times are generally helpful for reducing photobleaching and phototoxicity.
         
      4. Open the Adaptive Focus Control and choose two to four moderately fluorescent cells. For dual-excitation ratio imaging, use a 405 nm excitation laser and 488 nm excitation laser with an emission range of 500–550 nm. Monitor the FiLa, FiLa-L or FiLa-C sensors until the two-channel fluorescence intensities are stable for 2 min (typically, time intervals of 30 s, run time 10 min), and then pause the imaging program
         CRITICAL STEP It is important to obtain a stable baseline ratio, based on an average pixel intensity, before treatment.
         
      5. Add agents of interest that perturb cellular lactate metabolism at the desired timepoint, mix gently and continue the imaging. For example, we treated HEK-293 cells expressing FiLa sensors with sequential addition of D-glucose, oligomycin and 2-DG at the indicated time (see ‘Anticipated results’)
         CRITICAL STEP Pipette carefully to avoid touching the dish.
         
      6. (Optional) For high-throughput, cell-based chemical genetic assays, add agents of interest that perturb cellular lactate metabolism (for examples see step (v)) in a 96-well glass-bottom plate and, using the Lionheart FX automated microscope system equipped with a 96-well plate adapter, immediately measure the dual-excitation ratios of cells expressing FiLa sensors (Step 2A, Procedure 2)

3. Export raw data to ImageJ software as 12-bit or 16-bit TIFs for dual-excitation ratio image analysis. Determine the excitation ratio of 488/405 nm or 465/405 nm pixel by pixel and pseudocolor using the hue–saturation–brightness (HSB) color space by ImageJ software. Briefly, the HSB values (300, 100 and 100) represent the lowest ratio, and HSB values (0, 100 and 100) represent the highest ratio.

   PAUSE POINT The data can be stored for analysis at a later timepoint.

Table 2 |.

Effects of common cell metabolism-perturbing agents on subcellular lactate metabolism

Pathway	Agent	Mechanism(s) of action	Stock solution	Final solution	Cytosolic changes	Nuclear changes	Mitochondrial changes	Extracellular changes
Glycolysis	3-BrPA39	HK II	100 mM in HBSS-G	500 μM	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease
	2-DG	HK II	500 mM in HBSS-G	50 mM	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease
	Oxamate39	LDH	400 mM in HBSS-G	10 mM	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease
	IAA	GAPDH	10 mM in HBSS-G	500 μM	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease
	Oxalate	PK	5 mM in HBSS-G	500 μM	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease	Free [lactate] decrease
	UK5099	MPC	10 mM in DMSO	5–10 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
TCA	Devimistat	PDH/aKGDH	100 mM in DMSO	100 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] decrease	Free [lactate] increase
ETC	Rotenone39	Complex I	30 mM in DMSO	10 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
	Oligomycin A39	Complex V	5 mM in DMSO	5 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
pH Regulation	AZD-3965 (ref. 39)	MCT1	10 mM in DMSO	2 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] decrease
	AR-C155858 (ref. 39)	MCT1/2	10 mM in DMSO	2 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] decrease
Fatty acid metabolism	SB-204990 (ref. 39)	ACLY	30 mM in DMSO	50 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] decrease	Free [lactate] increase
	BMS-303141 (ref. 39)	ACLY	30 mM in DMSO	10 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] decrease	Free [lactate] increase
	Perhexiline maleate	CPT1	30 mM in DMSO	100 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
	Etomoxir	CPT1	30 mM in DMSO	100 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
	JZL195	MAGL	30 mM in DMSO	30 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
	KML29	MAGL	15 mM in DMSO	100 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
mTOR	Rapamycin39	mTOR	10 mM in DMSO	100 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
Metabolites	Octyl-(R)-2-HG39	ATP synthase/mTOR	100 mM in DMSO	500 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
	Octyl-(S)-2-HG39	ATP synthase/mTOR	100 mM in DMSO	500 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
Ca2+ Regulation	Ionomycin39	Calcium ionophore	10 mM in DMSO	2 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] decrease	Free [lactate] increase
Medical/health Drug	CoQ39	Essential Cofactor of ETC	10 mM in ethanol	100 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
	MitoQ39	Mitochondri-targeted antioxidant	10 mM in DMSO	10 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase	Free [lactate] increase
	Troglitazone39	Antidiabetes medicine	100 mM in DMSO	100 μM	Free [lactate] increase	Free [lactate] increase	Free [lactate] decrease	Free [lactate] decrease

Bold indicates an increase in the concentration of free Lactate ('free [Lactate]'). 3-BrPA, 3-bromopyruvic acid; HK II, hexokinase II; IAA, iodoacetamide; GAPDH, gLyceraLdehyde-3- phosphate dehydrogenase; PK, pyruvate kinase; MPC, mitochondrial pyruvate carrier; TCA, tricarboxylic acid; ETC, electron transport chain; PDH, pyruvate dehydrogenase; α-KGDH, α-ketoglutarate dehydrogenase; ATMA, antimycin A; MCT1/2, monocarboxylate transporter 1/2; ACLY, ATP citrate lyase; CPT1, carnitine palmitoyltransferase 1; MAGL, monoacylglycerol lipase; mTOR, mammalian target of rapamycin; 2-HG, 2-hydroxyglutarate; CoQ, Coenzyme Q10; MitoQ, Mitoquinone.

Procedure 3: imaging lactate dynamics in live mice by the cell microcapsule system

TIMING 40–48 h

CAUTION All handling procedures involving mice must conform to governmental and institutional guidelines and regulations.

Preparation of cell microcapsules

TIMING 3–4 h

• 1Prepare solutions, including MOPS solution, 1.5% (wt/vol) and 0.03% (wt/vol) sodium alginate solution, polymerization solution, depolymerization solution and poly-L-lysine solution, as described in the ‘Reagent setup’ section.
• 2Prepare the Encapsulator as described in the ‘Equipment setup’ section.
• 3Collect H1299 cells (~1.0 × 10⁸) stably expressing FiLa or FiLa-C and suspend the cells in culture medium. Centrifuge at 200g for 5 min and remove the supernatant. Wash the pellets with RPMI 1640 medium.
• 4Resuspend the cell pellets in 1 mL of RPMI 1640 medium and mix with 50 mL of 1.5% sodium alginate solution to produce the cell–sodium alginate suspension.

  CRITICAL STEP Pipette the cells gently to avoid producing air bubbles.

• 5Fill the autoclaved reaction vessel with 250–300 mL of polymerization solution. Aspirate the cell–sodium alginate suspension using a 20 mL syringe and link it to the reaction vessel in a laminar air hood. Fix the reaction vessel to the encapsulator control unit (follow the encapsulator setup as described in the ‘Equipment setup’ section and in Supplementary Fig. 6).
• 6Set parameters to start bead forming: frequency, 1,300 Hz; electrode, 1,100 V; stirrer, 55%; ‘pump 1’, 20.20 mL/min. Choose the ‘turbo’ button at the beginning to allow droplet injection and then revert to ‘Pump 1’.
• 7Start bead formation with previously established parameters. Allow the beads to harden in the polymerization solution for 15 min. Then, stop the stirrer and drain the polymerization solution to 100 mL.
• 8Add 150 mL of 0.05% (wt/vol) poly-L-lysine solution to the reaction vessel and react for 10 min. Drain the polymerization solution to 100 mL.
• 9Add 200 mL of MOPS solution, stir for 1 min and drain to 100 mL. Repeat this process.
• 10Add 250 mL of 0.03% (wt/vol) sodium alginate solution and stir for 10 min to form the external alginate septum. Drain the polymerization solution to 100 mL.
• 11Add 200 mL of MOPS solution, stir for 1 min and drain to 100 mL. Repeat this process.
• 12Add 250–300 mL of depolymerization solution and stir for up to 10 min to dissolve the alginate of the bead core. Drain the depolymerization buffer to 100 mL.

  CRITICAL STEP Make sure the stirring process is less than 10 min to prevent the layer of microcapsules from being too thin.

• 13Add 200 mL of MOPS solution, stir for 1 min and drain to 100 mL.
• 14Add 200 mL of MOPS solution and resuspend the microcapsules. Transfer the solution to a bead collecting flask and allow the microcapsules to settle for 5–10 min. Remove as much of the supernatant as possible by pipetting, leaving the microcapsules in a total volume of 50 mL, which is sufficient volume for ten implants.

Imaging lactate metabolism in live mice

TIMING 36–44 h

• 15Intraperitoneally inject sodium pentobarbital (100 mg per kg body weight) to anesthetize a C57 mice. After 5 min, monitor anesthetic depth by verifying the pedal withdrawal reflex. Remove dorsal hair of mice by electric shaver and depilatory cream. Implant the obtained 5 mL of cell microcapsules (~1.2 × 10⁷ microencapsulated cells) into the shaved dorsum of C57 mice via subcutaneous injection. After implantation, keep mice at room temperature and monitored until fully recovered. Continue feeding of mice for 36–44 h to eliminate edema.

  

• 16After 36–44 h, anesthetize the mice with sodium pentobarbital as described in step 15.
• 17Transfer a mouse to the IVIS SpectrumCT in vivo imaging system, keeping it under an artificially warmed environment by heated chamber. Place the mouse on its stomach.
• 18Place the mouse in the imaging chamber and image in a prone position, with excitation at 430 nm and emission at 540 nm for 6 s exposure time, excitation at 500 nm and emission at 540 nm for 4 s exposure time. Take images for 3 min and then inject a regulating agent (for example, 30 mg per kg body weight SB-204990 or 5 mg per kg body weight AZD-3965) via tail-vein injection to monitor the changes in lactate metabolism in mice in real time.

  CRITICAL STEP Start image acquisition 3 min before the addition of agents that perturb metabolism to obtain a baseline for the FiLa fluorescence ratio.

• 19Export the raw data to ImageJ software as 16 bit TIFs for dual-excitation ratio image analysis (Step 3, Procedure 2).

  PAUSE POINT The data can be stored for analysis at a later timepoint.

Procedure 4: imaging lactate dynamics in live mice by rAAV

TIMING 6–7 weeks

CAUTION All handling procedures involving mice must conform to governmental and institutional guidelines and regulations.

1. Generate and purify AAVs expressing FiLa or FiLa-C in serotype 9 with the CMV promoter (see ‘Reagent setup’).

   CRITICAL STEP Researchers should choose the appropriate serotype and promoter of AAVs for different tissues, which directly influence the final sensor transduction efficiency.

2. Bilaterally inject 50 μL of diluted AAV (1.2 × 10¹² vg/mL) in PBS in the gastrocnemius muscle.

   CRITICAL STEP The virus should be injected at the shallow surface of the gastrocnemius muscle, as deep injection is not suitable for imaging.

3. Allow the sensors to be expressed for 4 weeks in vivo, and monitor the sensor expression by visualizing fluorescence; too low or too high is not suitable for imaging. For different tissue types, the expression period is generally between 4 and 6 weeks.

   

4. To construct the T1DM mouse model, 4 weeks after injection of virus, intraperitoneally administer STZ (50 mg per kg body weight/day) to infected mice daily for 5 d. Raise the mice for 2 weeks after the STZ injection and monitor the body weight and blood glucose concentration every 3 d using a blood glucose monitor. Use mice with hyperglycemia (>300 mg/dL blood glucose levels) in the subsequent study.

   CRITICAL STEP The mice develop variable levels of hyperglycemia 2–3 weeks after STZ injection and ~90% mice should reach 300 mg/kg blood glucose level. Some diabetic mice with mild hyperglycemia (<300 mg/dL) spontaneously recover⁷³. As a result, only diabetic mice with medium/high hyperglycemia (>300 mg/dL) are used in this study.

5. Fast mice for ~4 h before imaging and then anesthetize mice with sodium pentobarbital (Step 15, Procedure 3) and depilate the skin over the gastrocnemius muscle of mice with an electric shaver.

   CRITICAL STEP Fasting for ~4 h contributes to physiological synchronization.

   CRITICAL STEP Depilating helps to reduce the fluorescence interference of hairs.

6. Using the universal mounting frame AK, hold a glass cover slips on the stage of a Leica TSC SP8 SMD confocal laser scanning microscope system with supersensitive HyD hybrid detectors and an HC Plan APO CS2 20×/0.75 NA dry objective. Place the mouse in a lateral position and place the leg of mouse on glass cover slips, its leg positioned opposite the objective. Keep mice under a warming blanket.

7. Using a 405 nm excitation laser and 488 nm excitation laser with an emission range of 500–550 nm for dual-excitation ratio imaging, take images for 5 min first and then administer insulin treatment (1.25 units per kg body weight) via intraperitoneal injection to monitor the changes in lactate metabolism in mice with T1DM in real time.

   CRITICAL STEP Insulin injection must be performed carefully and gently to avoid moving the image out of focus.

   CRITICAL STEP Start image acquisition 5 min before the addition of insulin to obtain a baseline for the FiLa fluorescence ratio.

   

8. Export the raw data to ImageJ software as 12 bit TIFs for dual-excitation ratio image analysis (Step 3, Procedure 2).

   PAUSE POINT The data can be stored for analysis at a later timepoint.

Procedure 5: rapid and simple assay to determine lactate levels in human body fluids

TIMING 24–26 h

Preparation of plasma samples

TIMING 24–26 h

• 1Collect human plasma samples. In our experiment, samples were collected from ten male tai chi athletes and ten male rugby athletes, 15–22 years of age, from the School of Exercise and Health, Shanghai University of Sport. Have the recruited athletes fast overnight and draw 10 mL of venous blood before they have breakfast.
• 2Collect venous blood samples in a vacuum blood collection tube with heparin and mix by inverting the vial gently several times.
• 3Process samples immediately after collecting. Centrifuge blood samples at 4 °C and 1,000g for 10 min, and then carefully aspirate supernatant plasma into 1.5 mL centrifuge tubes. Divide the samples into aliquots and label and store them at −80 °C for further examination.

  CRITICAL STEP Store 200–500 μL aliquots at −80 °C. Repeated freeze—thaw cycles should be avoided, as this may degrade metabolites.

  PAUSE POINT The obtained plasma samples can generally be stored at −80 °C for 2–3 years and used for metabolite detection at a later timepoint.

  

• 4After 30 min of rest after breakfast, let the athletes conduct tai chi training for 1 h on the same day or rugby training for 1 h on the following day followed by 1 h of rest. Then, draw 10 mL of venous blood, and collect and process samples as described above (Steps 2–3).

Lactate assay using human plasma

TIMING 3–5 min

• 5Thaw the plasma samples and sensor protein (FiLa or FiLa-H; prepared and purified as detailed in Steps 1–5, Procedure 1) on ice before assay. Dilute the plasma samples by a dilution factor of 40 (detected by FiLa) or 100 (detected by FiLa-H) in HEPES buffer (pH 7.4).
• 6Perform the assay in a black 96-well flat-bottom plate by combining 50 μL of diluted samples and 50 μL of FiLa sensor protein (final concentration, 0.2 μM). Alternatively, the assay can also be conducted by mixing a 0.5 μL or 2.5 μL plasma sample with 100 μL of sensor protein (final concentration, 0.2 μM). For background values, combine 50 μL of samples and 50 μL HEPES buffer (pH 7.4).

  

• 7Set up a lactate titration experiment on the same plate by combining 50 μL of the different concentrations of lactate and 50 μL of sensor protein.
• 8Immediately measure the fluorescence intensity by a Synergy Neo2 multimode microplate reader using 485 BP 20 nm or 420 BP 27 nm excitation and 532 BP 40 nm emission bandpass filters.
• 9
  Export fluorescence data. Correct the fluorescence by subtracting the background values. Calculate the ratio of fluorescence excited at 485 nm and 420 nm. From the lactate titration experiment, obtain fluorescence ratios in the absence of lactate ($R_{\text{min}}$) and saturated with lactate ($R_{\text{max}}$). Calculate the lactate concentration in the sample with the following equations:

  $${\mathrm{V}}_{\mathrm{L}\mathrm{a}\mathrm{c}}=\left[\mathrm{L}\mathrm{a}\mathrm{c}\right]/\left(K_{\mathrm{d}}+\left[\mathrm{L}\mathrm{a}\mathrm{c}\right]\right)$$ (1)

  $${\mathrm{V}}_{\mathrm{L}\mathrm{a}\mathrm{c}}=\left(R-R_{\mathrm{m}\mathrm{i}\mathrm{n}}\right)/\left(R_{\mathrm{m}\mathrm{a}\mathrm{x}}-R_{\mathrm{m}\mathrm{i}\mathrm{n}}\right)$$ (2)

  $$\left[\mathrm{L}\mathrm{a}\mathrm{c}\right]=K_{\mathrm{d}}\left(R-R_{\mathrm{m}\mathrm{i}\mathrm{n}}\right)/\left(R_{\mathrm{m}\mathrm{a}\mathrm{x}}-R\right)$$ (3)

  $$[\mathrm{L}\mathrm{a}\mathrm{c}{]}_{\text{Final}}=N\times [\mathrm{L}\mathrm{a}\mathrm{c}]=N\times K_{\mathrm{d}}\left(R-R_{\mathrm{m}\mathrm{i}\mathrm{n}}\right)/\left(R_{\mathrm{m}\mathrm{a}\mathrm{x}}-R\right)$$ (4)

  $V_{\mathrm{L}\mathrm{a}\mathrm{c}}$ is the fraction of FiLa or FiLa-H sensor bound to lactate, [Lac] is the free concentration of lactate, and $N$ is the dilution factor of the samples. $K_{\mathrm{d}}$ represents the affinity of the sensors for lactate (FiLa’s $K_{\mathrm{d}}$: ~130 μM; FiLa-H’s $K_{\mathrm{d}}$: ~20 μM; FiLa-L’s $K_{\mathrm{d}}$: ~800 μM). $R$ represents the R485/420 ratio of the samples.

  

Troubleshooting

Troubleshooting advice can be found in Table 3.

Table 3 |.

Troubleshooting table

Step	Problem	Possible reason	Solution
Procedure 1
3	Unclear bacterial fluid after sonication	Insufficient sonication	Dilute the bacterial fluid and increase sonication cycles
Procedure 2
1	Poor cell attachment	Glass-bottom surface is not ideal for cell attachment	Coat the glass-bottom plate or dish with poly-D-lysine, fibronectin or collagen
		Too high or too low cell confluency	Confirm cell viability and proper confluency
		Rough handling	Pipette carefully during washing or medium changing
2A(iv)	Failure in autofocus	Overexposed particles	Plate cells at proper density and confirm cell viability
			Avoid fields containing dead cells or cell debris, which could result in overexposed particles
		Vibrations	Place the whole equipment setup on a shockproof platform
	Cell death during long-term imaging	Excessive exposure time	Reduce exposure time, especially when a short wavelength excitation light source is utilized
		Unstable culture environment	Check temperate control, CO2 injection and water supplementation during the whole imaging process
2B(ii-iv)	Blurry images	Image out of focus	Pipette carefully
		AFC is not activated	Open AFC before imaging
		3D structure of different subcellular compartments	Choose independent focal plane and pinhole size for specific subcellular components
2B(v)	Low fluorescence responses in cells treated with chemicals	Reagent solutions are not fully mixed with cell culture medium	Mix fully after adding reagents
Procedure 3
15	Dim fluorescence of cell microcapsules	Low-level FiLa expression	Choose a stable monoclonal cell line with high expression of FiLa sensors
		Low cell viability	Control the preparation time of sodium alginate-cell suspension, and preheat buffer to room temperature in advance
		Inappropriate exposure time	Adjust the exposure time to improve image quality
		Low injection dose of microcapsule	Increase the injection dose appropriately
		or ruptured microcapsules	Reduce time of depolymerization or transplant the microcapsules with 10 mL syringe
Procedure 4
3	Low fluorescence signal in mouse muscle tissue	Inappropriate serotype of AAV	Choose the appropriate tissue or cell specific serotype of AAV
		Poor quality of AAV	Improve the titre and purity of AAV
		Low fluorescence intensity	Increase the injection dose of virus or increase expression time after virus injection
		Too deep an injection	Inject virus at the shallow surface of the gastrocnemius muscle
7	The field of view moves	Image out of focus	Inject insulin carefully and avoid touching the mouse or the microscope stage
		The mouse moves during the injection or imaging process	Anesthetize the mouse sufficiently by adjusting the dose of sodium pentobarbital
Procedure 5
3	Abnormal values	Plasma or serum samples are contaminated by red blood cell rupture	Contain blood samples in sterile tubes
			Avoid oscillating samples before centrifugation
		Degradation of the substrate to be tested	Divide samples into aliquots, store in ultra-low temperature freezer and avoid repeated freezing and thawing
		Evaporation of water in samples	Seal the tubes of samples with parafilm
6	The background value is too high	Unsuitable dilution factor	Increase the dilution factor or use the sensor with high affinity
9	Sample values are out of the sensor detection range	Unsuitable dilution factor	Test dilution factor in advance
		Use of sensor with inappropriate affinity	Use sensor with appropriate affinity
Box 1, step 8	Cells detach after washing with assay medium	Aspirating or adding the medium too hard	Remove or add medium carefully
	Low or high signal detection for measurements	The assay is not sensitive enough or the detection range is exceeded	Test cell density in advance and adjustment for different cell lines
Box 1, step 9	Did not yield the anticipated results	Drugs are not loaded properly into the injection ports, or not all injection ports are fulfilled with drugs or medium	Loaded drugs or medium should cover the injection ports properly to ensure enough air pressure. Make sure to pipette carefully, in one stream and avoid air bubbles
		Injection ports are clogged	Make sure that drugs are dissolved sufficiently without crystallization
		Incubation time is too short	Increase measurement points and wait for curve to flatten out
		Drug concentrations are not optimal	Test drug concentrations in advance and adjust for different cell lines
Box 1, step 10	Abnormal values	Data are disrupted by abnormal values of background correction wells	Check the values of the background correction wells and exclude abnormal values
		Cell density is not uniform in some of the wells	Trypsinize the cells into single-cell suspension and thoroughly mix before seeding
			Exclude the outliers

Timing

Reagent setup: 1–2 h

Procedure 1: preparation and characterization of FiLa sensors

Steps 1–8, preparation and characterization of FiLa sensors: 2–3 d

Procedure 2: analysis of real-time subcellular lactate flux in single living cells

Steps 1–3, cell seeding, imaging and analysis of real-time subcellular lactate flux in single living cells: 1–2 d

Procedure 3: imaging lactate dynamics in live mice by the cell microcapsule system

Steps 1–14, preparation of FiLa-expressing cell microcapsules: 3–4 h

Steps 15–19, microcapsule injection, in vivo incubation and imaging lactate metabolism in live mice: 36–44 h

Procedure 4: imaging lactate dynamics in live mice by rAAV

Steps 1–8, AAV injection and expression, and construction of the T1DM mouse model, fasting and in vivo imaging of mice with T1DM: 6–7 weeks

Procedure 5: rapid and simple assay to determine lactate levels in human body fluids

Steps 1–4, preparation of plasma samples from athletes: 24–26 h

Steps 5–9, assay of lactate in human plasma: 3–5 min

Box 1: determination of extracellular acidification rate by Seahorse Analyzer

Steps 1–4, preparation before ECAR assay: 10–12 h

Steps 5–9, running the ECAR assay: 2–3 h

Step 10, data analysis: 0.5–1 h

Anticipated results

In this protocol, we describe strategies for a FiLa sensor-based comprehensive multiscale analysis to explore lactate metabolic dynamics in cells in exquisite detail, in subcellular organelles, and in living animals and body fluids in both healthy and diseased states. This comprehensive multiscale analysis should be substantially more beneficial than static or cell population analysis by providing dynamic phenotypes of lactate metabolism at different times, spaces and scales. Extracellular lactate accumulation during cell culture is universally known; however, intracellular lactate metabolism, especially subcellular lactate metabolism, is rarely studied. When cells were cultured with medium containing different concentrations of glucose or FBS, two widely used growth supplements, the cytosolic, nuclear and mitochondrial lactate levels notably increased in a dose-dependent way, as shown by the FiLa or FiLa-L fluorescence (Fig. 2a–d). There was also a synergistic effect of glucose and FBS on subcellular lactate levels (Fig. 2a–d). In contrast, only slight changes in fluorescence were observed in control FiLa-C-expressing cells with glucose or FBS (Supplementary Fig. 3a–c).

Fig. 2 |. Imaging subcellular lactate metabolism under various nutritional conditions.

a, General overview of the procedure for imaging subcellular lactate metabolism under various nutritional conditions. b, Ratiometric fluorescence images of HEK-293 cells expressing FiLa family sensors in the cytosol (top), nucleus (middle) and mitochondria (bottom). Cells were cultured under different nutritional conditions. Scale bars, 10 μm. c,d, Quantification of FiLa expressed in the cytosol/nucleus (c) and FiLa-L expressed in the mitochondria (d) of HEK-293 cells (n = 20). Data are presented as the mean ± s.d. (c and d).

To compare the methods for bioenergetic flux analysis by Seahorse Analyzer and FiLa sensors (Fig. 3a–d), we conducted a FiLa-based glycolysis stress test assay by successively treating HEK-293 cells with 1 mM glucose, 1 μM oligomycin and 50 mM 2-DG at the indicated time, which induced a rapid increase, further increase and subsequent decrease in subcellular lactate levels, respectively (Fig. 3e–j). As the control, the FiLa-C fluorescence did not notably change after treatment with glucose, oligomycin or 2-DG (Supplementary Fig. 4a–f). The resulting data showed that FiLa responded rapidly and was suitable for use in bioenergetic flux analysis, and FiLa sensors have higher space–time resolution than the Seahorse Analyzer.

Fig. 3 |. Comparison study of the glycolysis stress test in live cells by Seahorse Analyzer or FiLa.

a,b, General overview of the procedure for the glycolysis stress test in live cells by Seahorse Analyzer (a) or FiLa (b). c, A glycolysis stress test was carried out by measuring the ECAR with Seahorse XF96 analyzer following sequential injections of glucose (Glc, 1 mM), oligomycin (Oligo, 1 μM) and 2-DG (50 mM) in HEK-293 cells (n = 9). d, Glycolytic parameters, including glycolysis, glycolytic capacity and glycolytic reserve, were calculated (n = 9). e–j, Fluorescence images (e,g and i) and quantification (f,h and j) of FiLa expressed in the cytosol (e and f) and nucleus (g and h) or FiLa-L in mitochondria (i and j) of HEK-293 cells. Cells were successively treated with 1 mM glucose, 1 μM oligomycin and 50 mM 2-DG at the indicated times. Scale bars, 10 μm. Data are presented as the mean ± s.e.m. (c and d).

To study how subcellular lactate metabolism is regulated, we have been able to develop this system into a high-throughput, cell-based chemical genetic assay³⁹ (Fig. 3b) and investigate the impact of different chemical modulators on lactate levels in four subcellular compartments (i.e., the cytosol, nucleus, mitochondria and extracellular space). This strategy has allowed us to show that lactate metabolism shifts in the same direction in most cases; however, it can dynamically change depending on the subcellular compartments (Table 2). For example, as expected, the LDH inhibitor oxamate reduced the lactate level in all four compartments, and the complex I inhibitor rotenone increased the lactate level in all four compartments due to the compensatory upregulation of glycolysis (Fig. 4a–d and Table 2). However, the monocarboxylate transporter (MCT) inhibitor AR-C155858 decreased the extracellular lactate level and increased the lactate levels in the other three compartments (Fig. 4e,f and Table 2). Surprisingly, the Ca²⁺ ionophore ionomycin decreased mitochondrial lactate levels and increased the lactate levels in the other three compartments (Fig. 4g,h and Table 2). As the control, the fluorescence of FiLa-C did not notably change after treatment with oxamate, rotenone, AR-C155858 or ionomycin (Supplementary Fig. 5a–h).

Fig. 4 |. Effects of different metabolic modulators on subcellular lactate levels.

Fluorescence images (a,c,e and g) and quantification (b,d,f and h) of FiLa expressed in the cytosol (top), nucleus (middle) or FiLa-L in mitochondria (bottom) of HEK-293 cells. Cells were treated with 10 mM oxamate (a and b), 10 μM rotenone (c and d), 2 μM AR-C155858 (e and f) or 2 μM ionomycin (g and h). Scale bars, 10 μm.

Our system can also be used to visualize lactate dynamics in vivo, by microencapsulating transgenic H1299-FiLa or H1299-FiLa-C cells into coherent, semi-permeable and immunoprotective alginate-poly-(L-lysine)-alginate beads. These beads, which allow the free diffusion of substances with low molecular sizes (<72 kDa) across the membrane⁷⁰, are then subcutaneously transplanted into the mouse dorsum (Fig. 5a and Supplementary Fig. 6). We were able to show that after 36–44 h, treatment with the ATP citrate lyase inhibitor SB-204990 or the monocarboxylate transporter 1 inhibitor AZD-3965 rapidly induced a cytosolic lactate increase, as shown by FiLa sensors (Fig. 5b–e). As a control, minimal changes in fluorescence were observed in FiLa-C-expressing cell microcapsules (Fig. 5b–e).

Fig. 5 |. Imaging lactate metabolism in vivo by cell microcapsules.

a, Flowchart showing an overview of the procedure for cell microcapsule preparation and in vivo imaging. b–e, In vivo fluorescence imaging (b and d) and quantification (c and e) of H1299 cell microcapsules expressing FiLa or FiLa-C in response to intravenously administered SB-204990 (30 mg per kg body weight; b and c) or AZD-3965 (5 mg per kg body weight; d and e). Scale bars, 0.5 cm. All procedures involving animals were approved by the Institutional Animal Care and Use Committee of East China University of Science and Technology.

We also show a typical example of how to evaluate in situ lactate metabolism by using the FiLa sensor in live mice with T1DM. The gastrocnemius muscles of mice were infected with rAAVs encoding FiLa or FiLa-C, and the mice were then given multiple low doses of STZ to establish the T1DM model (Fig. 6a). Mice with T1DM showed notably elevated cytosolic lactate levels in the muscle tissue compared with those in the muscle tissue of wild-type mice ~6 weeks after AAV infection, while no notable changes in FiLa-C were observed in the same mice (Fig. 6b,c). After insulin injection, mice with T1DM exhibited a marked decrease in lactate and glucose levels, as monitored by the FiLa sensor and biochemical assay (Fig. 6d,e and Supplementary Fig. 7), and minimal changes were observed in FiLa-C-expressing muscle tissues (Fig. 6d,e).

Fig. 6 |. Imaging lactate metabolism in live mice with T1DM.

a, General overview of the procedure for imaging lactate metabolism in living mice. b, Images of a wild-type (WT) mouse and a mouse with STZ-induced T1DM. c, In vivo fluorescence imaging of muscle tissue expressing FiLa in living wild-type mice and mice with T1DM. d,e, In vivo fluorescence imaging (d) and quantification (e) of FiLa in muscle tissue of living mice with T1DM treated with insulin. Scale bars, 100 μm. Data are presented as the mean ± s.e.m. (e). Lactate levels can be measured after calibration of FiLa fluorescence in live mice with that of recombinant FiLa protein. The equation is as follows: $\left[\mathrm{L}\mathrm{a}\mathrm{c}\right]=K_{\mathrm{d}}\left(R-R_{\text{min}}\right)/\left(R_{\text{max}}-R\right)$. [Lac] is the free concentration of lactate. $R_{\text{min}}$ and $R_{\text{max}}$ represent the minimum $F_{488\mathrm{n}\mathrm{m}}/F_{405\mathrm{n}\mathrm{m}}$ ratio ($R_{\text{min}}$) and the maximum $F_{488\mathrm{n}\mathrm{m}}/F_{405\mathrm{n}\mathrm{m}}$ ratio ($R_{\text{max}}$) of the recombinant FiLa protein corrected by the corresponding FiLa-C protein values at saturating conditions. $R$ represents the calibrated $F_{488\mathrm{n}\mathrm{m}}/F_{405\mathrm{n}\mathrm{m}}$ ratio of FiLa in the live mice. All procedures involving animals were approved by the Institutional Animal Care and Use Committee of East China University of Science and Technology.

We have further explored the use of FiLa for the rapid and simple determination of lactate levels in body fluids and recruited 20 male athletes (10 tai chi athletes and 10 rugby athletes) 15–22 years of age (Fig. 7a), whose characteristics are shown in Supplementary Fig. 8. As detected by FiLa-H, the plasma lactate levels in the group after rugby training (2.01 ± 0.34 mM) were significantly higher than those in the group before rugby training (1.57 ± 0.26 mM); however, there were no significant differences between groups before and after tai chi training (before: 1.52 ± 0.32 mM; after: 1.69 ± 0.32 mM) (Fig. 7b). These results might reflect the different effects of aerobic exercise (tai chi) and anaerobic exercise (rugby) on lactate levels. To validate our assay, we analysed the samples in parallel using FiLa (Fig. 7c). The results showed a strong correlation between the two independent sensors (Fig. 7d,e).

Fig. 7 |. Rapid and simple assay to determine the plasma lactate levels of tai chi athletes and rugby athletes.

a, General overview of the procedure for the plasma lactate assay by FiLa-H or FiLa. b,c, Plasma lactate concentration of athletes before and after tai chi or rugby training determined by FiLa-H (b) or FiLa (c) (n = 20). d, Quantification of lactate levels in plasma samples. The test results obtained by FiLa are plotted against the results obtained by FiLa-H. r, correlation coefficient. Data are replotted from b and c. e, Bland–Altman analysis for plasma samples measured by FiLa-H and FiLa. Data are replotted from b and c. All P values were obtained using paired two-tailed Student’s t-test. ****P < 0.0001, N.S., not significant. All procedures related to human research subjects received approval from the School of Exercise and Health, Shanghai University of Sport and were conducted in strict accordance with institutional guidelines.

Supplementary Material

Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41596-023-00948-y.

Key points.

• This protocol describes the use of FiLa biosensors, fluorescence-based sensors for the analysis of lactate metabolism. FiLa biosensors can be used in both in vitro assays and in vivo assays, and under different nutritional and pharmacological conditions.

• Unlike traditional methods of lactate analysis, FiLa biosensors allow subcellular analysis of lactate in real time and show a larger fluorescence ratio response than existing biosensors.

Related links

Key references using this protocol

Li, X. et al. Cell Metab. 35, 200–211 (2023): https://doi.org/10.1016/j.cmet.2022.10.002

Jia, M. et al. Sci. Adv. 9, eadg4993 (2023): https://doi.org/10.1126/sciadv.adg4993

Dou, X. et al. Nat. Metab. 5, 1887–1910 (2023): https://doi.org/10.1038/s42255-023-00912-w

Data availability

The data used to generate the example results presented in Table 2 are available in the supporting primary research paper³⁹. All other data supporting the findings of this study are available for research purposes from the authors upon reasonable request. Source data are provided with this paper.