A Novel Ruthenium-based Molecular Sensor to Detect Endothelial Nitric Oxide

Nitric oxide (NO) is a key regulator of endothelial cell and vascular function. The direct measurement of NO is challenging due to its short half-life, and as such surrogate measurements are typically used to approximate its relative concentrations. Here we demonstrate that ruthenium-based [Ru(bpy)2(dabpy)]2+ is a potent sensor for NO in its irreversible, NO-bound active form, [Ru(bpy)2(T-bpy)]2+. Using spectrophotometry we established the sensor’s ability to detect and measure soluble NO in a concentration-dependent manner in cell-free media. Endothelial cells cultured with acetylcholine or hydrogen peroxide to induce endogenous NO production showed modest increases of 7.3 ± 7.1% and 36.3 ± 25.0% respectively in fluorescence signal from baseline state, while addition of exogenous NO increased their fluorescence by 5.2-fold. The changes in fluorescence signal were proportionate and comparable against conventional NO assays. Rabbit blood samples immediately exposed to [Ru(bpy)2(dabpy)]2+ displayed 8-fold higher mean fluorescence, relative to blood without sensor. Approximately 14% of the observed signal was NO/NO adduct-specific. Optimal readings were obtained when sensor was added to freshly collected blood, remaining stable during subsequent freeze-thaw cycles. Clinical studies are now required to test the utility of [Ru(bpy)2(dabpy)]2+ as a sensor to detect changes in NO from human blood samples in cardiovascular health and disease.


Concentration dependent changes in fluorescence counts with [Ru(bpy)2(dabpy)] 2+ over time with an NO donor, NOC5.
(a) Fluorescence counts under λex=450 nm and λem=615 nm on the Glomax Discover System with 2-100 µM NOC5 and 10 µM [Ru(bpy)2(dabpy)] 2+ in PBS with readings recorded over 90 minutes. (b) Approximate concentration of NO in the PBS solution after 30 and 90 minutes of addition of NOC5, calculated based on the half-life of 93 minutes at a room temperature of 22˚C. Concentration dependent changes in the fluorescence at (c) 30 minutes and (d) 90 minutes after the addition of NOC5. The discontinuous lines represent the best fit used for the regression analysis and to calculate the coefficient of determination (R 2 ) for each concentration dependent response. All data are represented as mean ± s.d. from 3 cell-free replicates.
Background controls with acetylcholine.
(a, b) Fluorescence count readings under λex=450 nm and λem=615 nm on the Glomax Discover System 30 minutes after the addition of 10 µM acetylcholine (Ach) to cell-free PBS or culture media with 10 µM and 50 µM [Ru(bpy)2(dabpy)] 2+ . The p-values were derived from one-way ANOVA followed by Tukey's multiple comparisons test and the differences in the fluorescence count after the addition of Ach were not statistically significant. (c, d) Fluorescence count readings after the addition of NOC 13 to different sensor concentrations with and without 10 µM Ach in PBS or cell culture media, followed up for 60 minutes. All data are represented as mean ± s.d. from 3 cell-free replicates.

Supplementary Figure S5
Background controls with hydrogen peroxide.
(a, b) Fluorescence count readings under λex=450 nm and λem=615 nm on the Glomax Discover System 30 minutes after the addition of 150 µM hydrogen peroxide (H2O2) to cell-free PBS or culture media with 10 µM and 50 µM [Ru(bpy)2(dabpy)] 2+ . The p-values were derived from one-way ANOVA followed by Tukey's multiple comparisons test and only the p<0.05 values are reported. (c, d) Fluorescence count readings after the addition of NOC 13 to different sensor concentrations with and without 150 µM H2O2 in (c) PBS or (d) cell culture media followed up for 60 minutes. All data are represented as mean ± s.d. from 3 cell-free replicates.
In the last condition, 200 µM of cPTIO was added to 150 µM H2O2, to demonstrate the reduction of the fluorescent count in the presence of a nitric oxide scavenger, despite the presence of H2O2.

Supplementary Figure S7
Indirect detection of NO using the Griess assay in HUVECs.
Representative results (mean ± s.d. of replicate readings) demonstrating the absorbance readings in the Griess assay with 10 µM acetylcholine (Ach, 15 min) and 150 µM hydrogen peroxide (H2O2, 150 min) as endogenous stimuli and with NOC13 as a source of exogenous NO in HUVECs at 37˚C and 5% CO2. The absorbance readings for (a) nitrites only and (b) total nitrites+nitrates are reported. The absorbance at 40 µM NO was calculated using the standard curves derived using the standard solutions provided in the kit by the manufacturer.

WST-1 assay
Water Soluble Tetrazolium 1 (WST-1) reagent (Roche, NSW, Australia) based colorimetric assay was used to assess the effect of up to 72 hours of exposure to [Ru(bpy)2(dabpy)] 2+ on cell viability before using it for NO sensing in HUVECs. The optimum cell density and incubation period with WST-1 was determined by preliminary experiments. HUVECs in passage 3 were seeded at 4000 cells/well using Fluorescence Activated Cell Sorting into 96 well plates pre-coated with gelatine. HUVECs were incubated overnight in a cell incubator at 37˚C and 5% CO2 with cell culture media and 0.001-200 µM [Ru(bpy)2(dabpy)] 2+ was added. All exposures were done in quadruplicate and cell-free, gelatine, and sensor only controls were matched for all concentrations. Cell culture media was replaced with 100 µl of phenol red-free RPMI cell culture media and 10 µl of WST-1 reagent was added to each well at the end of each time point. HUVECs were incubated for 4 hours at 37˚C and 5% CO2 and read using an iMark microplate reader (BIORAD) at 440 nm.

Assessment of angiogenic (vascular tube formation) capacity of the HUVECs following exposure [Ru(bpy)2(dabpy)] 2+
HUVECs in passage 2 were seeded in six well plates (1.2x10 5 cells/well) and were used for the experiments when 90% confluent. HUVECs were incubated with either PBS, 10 or 50 µM [Ru(bpy)2(dabpy)] 2+ for 24 hours in Meso-endo media at 37 °C, 5% CO2. The Meso-Endo media was removed from the plated cells and washed briefly with filtered 1X PBS. The cells were trypsinised with 700 μL trypsin-EDTA to each well, incubated for 5 min at 37°C followed by inactivation with equal volume of Meso-Endo media. The cells were centrifuged at 2500 rpm for 5 min, re-suspended in 1 mL of Meso-Endo media and counted for each condition. A cell suspension of 1.2 x 10 5 cells/mL with total volume 2.6 mL in Meso-Endo media was made for each condition. Matrigel (Corning Matrigel Growth Factor Reduced (GFR) Basement Membrane Matrix) was added to each experimental well (40 μL) in quadruplicate in a 96-well plate under sterile conditions and kept on ice.
The suspension of cells of 200 μL (1.2 x 10 4 cells/well) was added per well in Meso-Endo. The plates were incubated at 37°C, 5% CO2 6 and 24 hours and imaged using an AX10 Zeiss microscope. Two images (central and peripheral) were taken from each quadrant in a well. Counting of the number of branches and branch points was performed using Image J software (National Institutes of Health, USA).

Representative images from the Matrigel assay
Representative images of the cell culture wells of the Matrigel assay under normoxic conditions following 24 hours pre-exposure to [Ru(bpy)2(dabpy)] 2+ Magnification X 2.5

Supplementary Information -Section 2 Localisation of [Ru(bpy)2(dabpy)] 2+ in HUVECs Spectrophotometry
Spectrophotometry (SynergyMx Microplate Reader, BioTek) was used to analyse fluorescence signal in the supernatant of washed and unwashed HUVECs in passage 3 that were treated with 1 mM NOC13 or 10 µM acetylcholine (15 min) and 50 µM [Ru(bpy)2(dabpy)] 2+ . The readings were taken in two groups: 1) cell culture media changed before reading the plate 2) no cell culture media change before the initial reading and subjected to two media changes with a reading after each change All readings were in quadruplicate and repeated with [Ru(bpy)2(dabpy)] 2+ incubation at 37 °C, 5% CO2 for 4, 6 and 24 hours, read at λex=450 nm and λem=590, 605, 615 and 630 nm. In addition, the supernatant was isolated during the first media change and the fluorescence count was measured to determine the location of the fluorescence signal as cell surface or supernatant based.

Confocal microscopy
HUVECs (20000 cells/well) were grown on gelatine coated 12 well glass bottom plates (Cell E&G, San Diego, CA, USA) and were incubated with 50 µM [Ru(bpy)2(dabpy)] 2+ for 24 hours and treated with excess (1 mM) NOC13 (1-Hydroxy-2-oxo-3-(3-amino-propyl)-3-methyl-1-3\triazene, T1/2 = 13.7 min) for 1 hour at 37˚C in an incubator or live imaging chamber attached to microscope set-up. A nucleic acid stain, 2 µM Hoechst 33342 (ThermoFisher) was added for 30 minutes before imaging to demarcate the nucleus. A Leica TCS SP8X/MP Confocal Microscope under λex=473 nm and λem=565-645 nm was used for imaging, with analysis performed using LAS-X imaging software (Leica Microsystems Pty Ltd, NSW, Australia). Cells were imaged with and without changing the media to confirm the cellular localisation of the fluorescence. Fluorescence intensity of different channels were quantified in each image, from three different sections for each condition before and after media change.
In the second experiment, the HUVECs were incubated with 50 µM [Ru(bpy)2(dabpy)] 2+ for 24 hours followed by 50 µM or 100 µM NOC13 for 1 hour at 37˚C in an incubator or live imaging chamber attached to microscope set-up. A nucleic acid stain, 2 µM Hoechst 33342 was added for 30 minutes before imaging. Cells were imaged with and without changing the media to confirm the cellular localisation of the fluorescence. Fluorescence intensity of different channels were quantified in each image, from three different sections for each condition before and after media change.
Mean intensity from the red channel ([Ru(bpy)2(dabpy)] 2+ ) was normalized to the mean intensity of the blue channel (Hoechst), to determine the mean fluorescence relative to the number of cells.

Localisation of fluorescence from NO bound [Ru(bpy)2(T-bpy)] 2+ in HUVECs using confocal microscopy
[Ru(bpy)2(T-bpy)] 2+ emits an extracellular signal with HUVECs as represented in the merged confocal images of HUVECs stained with Hoechst 33342 (a nuclear stain -Blue) and [Ru(bpy)2(T-bpy)] 2+ (NO bound sensor -Red). HUVECs in the presence of NOC13 (an exogenous NO donor) with (a-c) no media change or (d-f) with a media change prior to imaging. Mean intensities from the red channels in three stacked images are reported ( ) alongside the normalised intensities to the nuclear stain ( ) in cells subjected to (g) no media change and (h) with the media changed.

Confocal microscopic imaging of HUVECs for exogenous nitric oxide (NO) with NOC13
Representative confocal microscopic images of HUVECs in the presence of the NO sensor [Ru(bpy)2(dabpy)] 2+ and (a and c) 50 µM or (b and d) 100 µM NOC13 as an exogenous NO donor, in cells before media change.
Red -active/NO bound [Ru(bpy)2(T-bpy)] 2+ Blue -Hoechst 33342 (a nuclear stain) (e) Mean intensities from the red channels in three stacked images are reported ( ) alongside the normalised intensities to the nuclear stain ( ) in cells subjected to no media change.

Inductively coupled plasma mass spectrometry (ICP-MS) to assess the cellular uptake of Ruthenium
HUVECs were grown in T75 tissue flasks until 80-90% confluent. Cells were then washed in PBS and trypsinised, harvested and seeded into 6 well plates at 1.2x10 5 cells/well in 2 mL of Meso Endo Cell Growth Medium (Cell Applications INC) and were allowed to adhere to the plates for 24 hours after seeding in the cell incubator (37°C, 5% CO2). HUVECs were then incubated in triplicate with either PBS, 10 µM or 50 µM [Ru(bpy)2(dabpy)] 2+ in Meso Endo media for 24 hours at 37°C, 5% CO2 in the cell incubator. After 24 hours of incubation with [Ru(bpy)2(dabpy)] 2+ , the supernatant from each well was removed and placed into a 15 mL tube and spun at 1400 rpm for 3 mins to remove cell debris. HUVECs were washed gently (3 times) with 1 x PBS (1mL), trypsinized and were spun at 7000 rpm for 3 minutes to pellet the cells. The supernatant was removed by suction. Cells were washed by resuspending in 1 x PBS (1mL) and spun at 7000 rpm for 3 minutes to pellet cells, and repeated 3 times. After the last wash, the cell pellet was desiccated in a dry block heater at 100 °C.
The dry cell samples were then re-suspended in 400 μL of 37% Hydrochloric Acid (HCl, Sigma Aldrich 258148-2.5L) and heated at 100°C until dry (for approximately 2 hours) to digest organic material and liberate Ruthenium from the sample, simplifying the matrix for ICP-MS analysis. The dry samples were then reconstituted in 2% HCl made up in MillQ H2O (4 mL total volume). All controls as well as the PBS supernatant were diluted 1:20 and the supernatant samples were diluted to fit the calibration series. The 10 µM [Ru(bpy)2(dabpy)] 2+ supernatant samples were diluted 1:21.4 and the 50 µM [Ru(bpy)2(dabpy)] 2+ supernatant were diluted 1:107.19 in 2% HCl up to a total volume of 4 mL). All samples were sonicated for 20 min to produce a clear solution. The samples were then filtered through a syringe filter cartridge, pore diameter of 0.22 μm (Millex-GV PVDF 0.22 µm, 33 mm) into a 5 mL tube.
The standards were prepared using a 100 mM stock solution of [Ru(bpy)2(dabpy)] 2+ . Based on the molecular weight of 1071.9, the following dilutions was used to make the standard solutions of 0, 1, 10, 20, 50, 100, 200, 250, 400 and 500 particles per billion (ppb).  1,10,20,50,100,200,250,400, 500 ppb were used for Ru quantification. At the end of the experiments the following standard curves were developed to confirm the accuracy of the standards solutions used to determine the ppb values from the cell lysates and supernatants.

Supplementary Figure S13
Standard curves used for the ICP-MS analysis  (2013). New Zealand White, male rabbits were bred in-house at the SAHMRI Preclinical Imaging and Research Laboratory animal facility (Gilles Plains, South Australia) or in the Animal Facility of the Flinders University. Nine rabbits were anaesthetised at two years of age with intramuscular ketamine (35 mg/kg)/xylazine (5 mg/kg) and 2-5% isoflurane inhalation. After thoracotomy, blood was directly drawn from the heart in to a 1 mL syringe containing EDTA/PBS solution and either 500 µL of 100 µM [Ru(bpy)2(dabpy)] 2+ in PBS or 500 µL PBS as the vehicle control. The samples were processed in two groups and all readings were done on the SynergyMx Microplate Reader at λex=450 nm and λem=615nm.
Group A (n=4): Thirty minutes after collection (on ice and protected from light), blood samples were centrifuged at 3000 rpm for 10 min, plasma snap-frozen, stored at -80˚C and later thawed prior to reading. Following the initial reading, 50 µL of 1 mM Ru(bpy)2(dabpy)] 2+ was added to plasma from the negative control (containing only PBS with plasma) and re-read to determine the changes of NO levels in plasma over time after a freeze-thaw cycle. NOC13 (1 mM) was then added to all samples containing [Ru(bpy)2(dabpy)] 2+ and re-read after 10 minutes to confirm the presence of active Ru(bpy)2(dabpy)] 2+ in plasma after sample processing.

Supplementary Figure S16
Scheme for sample processing in Group A rabbits. Group B (n=5): Both samples of blood were left on ice for 20 min covered from light. Then, 50 µL of 1 mM Ru(bpy)2(dabpy)] 2+ was added to the negative control containing PBS and left for 10 min. 50 µL of PBS was added to the other sample to control for the volume. All samples were centrifuged 3000 rpm 10 min, plasma separated, snap frozen, stored at -80˚C and thawed prior to reading. Following the initial read, the plate was left in -80˚C for 48 hours and re-read to determine the stability of NO-[Ru(bpy)2(dabpy)] 2+ complex over time and following freezing.

Supplementary Figure S17
Scheme for sample processing in Group B rabbits. In addition, blood from four rabbits (3-4 years old) were used for control experiments with and without a scavenger for NO.

Supplementary figure S18
Scheme for sample processing in Group C rabbits.