Real Time and Spatiotemporal Quantification of pH and H2O2 Imbalances with a Multiplex Surface-Enhanced Raman Spectroscopy Nanosensor

Oxidative stress is involved in many aging-related pathological disorders and is the result of defective cellular management of redox reactions. Particularly, hydrogen peroxide (H2O2), is a major byproduct and a common oxidative stress biomarker. Monitoring its dynamics and a direct correlation to diseases remains a challenge due to the complexity of redox reactions. Sensitivity and specificity are major drawbacks for H2O2 sensors regardless of their readout. Luminiscent boronate-based probes such as 3-mercaptophenylboronic acid (3-MPBA) are emerging as the most effective quantitation tool due to their specificity and sensitivity. Problems associated with these probes are limited intracellular sensing, water solubility, selectivity, and quenching. We have synthesized a boronate-based nanosensor with a surface-enhanced Raman spectroscopy (SERS) readout to solve these challenges. Furthermore, we found out that environmental pH gradients, as found in biological samples, affect the sensitivity of boronate-based sensors. When the sensor is in an alkaline environment, the oxidation of 3-MPBA by H2O2 is more favored than in an acidic environment. This leads to different H2O2 measurements depending on pH. To solve this issue, we synthesized a multiplex nanosensor capable of concomitantly quantifying pH and H2O2. Our nanosensor first measures the local pH and based on this value, provides the amount of H2O2. It seems that this pH-dependent sensitivity effect applies to all boronic acid based probes. We tested the multiplexing ability by quantitatively measuring intra- and extracellular pH and H2O2 dynamics under physiological and pathological conditions on healthy cells and cells in which H+ and/or H2O2 homeostasis has been altered.

1.1 Synthesis of the nanosensors based on silica coated plasmonic LbL (AuNPs@NCs).Briefly, the synthesis of the NCs is performed wrapping polyelectrolytes of opposite charge (PSS and PAH) layer by layer (LbL) onto a template (polystyrene, PS) being the last layer, a deposition of gold seeds.We grow a final silica (SiO2) shell to protect the plasmonic nanostructure and to preserve the SERS signal.Afterwards, we dissolve the PS template chemically and grow the gold seeds inside into a plasmonic nanostructure showing a characteristic LSPR.DLS measurements show a hydrodynamic size of 467.6 nm and PDI 0.044 (SI-1D) and a ζ-potential mean value -36.7 mV.The hollow plasmonic NCs are functionalized with the Raman probes for sensing.We selected 3-MPBA as a H2O2 sensor molecule because it can be oxidized into 3-hydroxyl thiophenol (3-MP) in the presence of H2O2 showing new SERS characteristic bands of 3-MP 1 (figure SI-1F).Figure SI-1G shows the Raman spectrum of 3-MPBA in powder form (black line) and SERS spectrum of 3-MPBA modified NCs (NCs@3-MPBA; red line).The spectral fingerprint differences between both spectra are the consequence of the surface selection rules and the surface enhancement due to the media interaction and the resonance coupling occurring when 3-MPBA is adsorbed onto the NCs' metallic surface.In this context, the disappearance of the peak at 910 cm -1 associated to the vibration mode of the thiol group (-SH) is a confirmation of the deprotonation of this group and a consequence of 3-MPBA bonding to the NCs 2 .Moreover, an intense SERS spectrum of 3-MPBA confirming the successful functionalization of the NCs appears 1 .Characteristic features include the bands at 783 cm -1 assigned to C-H out of plane bending mode, at 996 cm -1 resulting from C-C in plane bending mode, at 1020 cm -1 attributed to C-H in plane bending mode, at 1070 cm -1 issued from C-C in plane bending coupled with C-S stretching modes, at 1553 cm -1 referred to non-totally symmetric benzene ring stretching mode, and at 1570 cm -1 imputed to totally symmetric benzene ring stretching mode.
To confirm the H2O2 sensing capability of NCs@3-MPBA, we first elucidated the peaks sensitive to H2O2 (figure SI-2).To emphasize the spectra modification, a statistical analysis of the spectra variation based on standard deviation was conducted and shown in figure SI-2A (blue spectrum), allowing us to detect significant spectra variation.With variation higher than 25 %, they are sensible to H2O2 concentration.These spectra variations can be categorized into three different types: i) peak intensity reduction associated to bond relative concentration reduction (peaks at 1020 and 1110 cm -1 , orange shadow), ii) peak intensity increase attributed to bond relative H2O2 concentration increase (peaks at 882 and 1240 cm -1 and broad band at 1550-1600, yellow shadow areas), and iii) peaks shift associated to modification in the bond close environment (peaks at 996 and 1070 cm -1 ).The emerged peaks at 882 and 1589 cm -1 with the contribution of H2O2 concentration are assigned to the benzene ring stretching (ν12) and the totally symmetric ring stretching (ν89) of 3-MP, respectively.Figure SI-2B included a summary of the assignment of the peaks in the SERS spectra and the evolution of 3-MPBA before and after the oxidation by H2O2.If the H2O2 concentration increases, the SERS intensity of these bands also increases because of the oxidation of 3-MPBA to 3-MP and the conversion of boronate to hydroxyl functional group (figure SI-1G), in consistency with previous studies 1 .Moreover, the oxidation of 3-MPBA did not change significantly the molecular orientation of the mercaptobenzene group on the gold surface since we did not monitor a large set of different perturbations on the SERS spectra of the mercaptobenzoyl moiety 9 (figure SI-1F).We did observe that the intensity of the peak associated with the C-C in plane bending mode at 996 cm -1 was not affected by the oxidation of the molecule.This invariance of the intensity allows us to use this contribution as a reference band for ratiometric analysis.This independence is the consequence that C-C bonds are not created or destroyed during the oxidation process.There is a minimal blue shift (2 cm -1 ) of the C-C peak associated to the modification of the environment of the C-C bond.However, this effect does not compromise to use of the peak intensity as internal calibration to correct signal fluctuation and to minimize the impact of external parameters such as NCs batch-to-batch variability or different cellular loading.
Then we confirmed a dynamic responsiveness by analyzing the reaction time of NCs@3-MPBA interacting with H2O2 in saline buffer (pH 7) (figure SI-3A).We set up the timing for future experiments at 30 min.Then, we measured the SERS spectra of the NCs@3-MPBA dispersed in cells growth medium (pH 7.2-7.4,figure SI-3B), containing variable amounts of H2O2. Figure SI-3C shows the relations between intensity ratio of 882 cm -1 and 996 cm -1 (log [I882/I996]) and the concentration of H2O2 (log[H2O2]) of the NCs@3-MPBA dispersed in phosphate buffered saline (black line) and in cell growth media (red line).Both calibration curves were similar, proving the value of our sensor in biological environments.Each spectrum was the average of 5 spectra gotten from 5 different NCs@3-MPBA.(C) Calibration curves of NCs@3-MPBA determining H2O2 in phosphate buffer and in cells growth media followed the same trend, indicating that calibration curves gotten in buffer system can be easily used in growth media environment.
4-MBA has been previously used for pH sensing with SERS 10,11 since the ratiometric intensity signal of COO -stretching vibration mode can be calibrated for pH sensing.The peaks at around 1075 cm -1 and 1590 cm -1 correspond to aromatic ring vibrations.The peaks at around 1385 cm -1 and 1700 cm -1 are attributed to symmetric carboxyl stretching mode and C=O stretching vibrations of non-dissociated COOH groups, respectivel 10,12 .Following previously results from us and other groups, we conjugate the NCs with 4-MBA for pH sensing and we further proofed that the pH signaling was not altered by the presence of H2O2.As shown in figure SI-4, we collected SERS spectra of 4-MBA modified NCs (NCs@4-MBA) in phosphate buffered saline under pH 4 and pH 7 with and without H2O2.We did not observed changes in the vibrational modes, confirming that the commonly used pH sensitive peak at around 1385 cm -1 can be used for pH sensing even in the presence of H2O2.

Synthesis and characterization of the multiplex nanosensor (NCs@3-MPBA&4-MBA).
3-MPBA and 4-MBA modified NCs (NCs@3-MPBA&4-MBA) were synthesized following the same procedure as described before.The general approach was to incubate the synthesized NCs with diluted feedstock solution of the two Raman probes (3-MPBA and 4-MBA) in ethanolic solution.The surface composition of functionalized NCs was monitored by SERS.By simply comparing the strong bands intensities at 996 cm -1 and 1075 cm -1 , the relative proportions on NCs surface of each probe can be estimated.The peak around 996 cm -1 is attributed to C-C in plane bending mode of 3-MPBA, and the peak around 1075 cm -1 attributed to aromatic ring vibrations of 3-MPBA and 4-MBA.We merged the SERS spectra of 3-MPBA and 4-MBA with the ratio 1:1 and calculated the intensity ratio between 1075 cm -1 to 996 cm -1 (I1075/I996) as shown in Figure SI-5A.When I1075/I996 is around 1.55, the signal proportion of both Raman probes will be equivalent.The surface composition is the result of competition for surface site between 4-MBA and 3-MPBA.Thus, for the synthesis of the multiplex sensor, we adjust the concentration ratio between 4-MBA and 3-MPBA to achieve equivalent signals.We then measured the response of NCs@3-MPBA&4-MBA to different concentration of H2O2 and pH to confirm the preservation of the SERS signal for both probes (figure SI-6).The signal changes in NCs@3-MPBA&4-MBA agree with the phenomenon we observed with the NCs@3-MPBA i.e., a log(I882/I996) increase with both H2O2 and pH (figure SI-2-SI-3).Bands at around 996 and 1075 cm -1 are attributed to aromatic ring vibrations, which are not sensitive to pH neither H2O2.Bands at around 1385 cm -1 and 1700 cm -1 are related to carboxyl group vibrations and are only sensitive to pH (figure SI-4).Intensity at 1385 cm -1 increased along with pH increase, while intensity at 1700 cm -1 decreased when pH increased.We chose the intensity ratio between the well-known sensitive peak at around 1385 cm -1 10 and the insensitive peak at 996 cm -1 (I1385/I996) to obtain the calibration curves for pH measurement (figure SI-6A).NCs@3-MPBA&4-MBA were sensitive to changes in the pH ranging from 5 to 7, presenting typical Henderson-Hasselbalch plots 13 , which is in agreement with the SERS results published of 4-MBA covered gold nanoparticles 10 .The NCs@3-MPBA&4-MBA maintain the sensitivity of the band at 882 cm -1 , corresponding to the benzene ring stretching mode of 3-MPBA, to H2O2.Its intensity increases along with H2O2 (figure SI-6B).

Section 2:
Influence of physiological pH levels on boronate-based H2O2 sensors' sensitivity.

NCs@3-MPBA's response to H2O2 dynamics depends on environmental pH values.
Figure SI-7 shows how the SERS intensity ratios log(I882/I996) of NCs@3-MPBA varied at different pH and H2O2 concentrations, ranging from pH 4 to pH 9 and H2O2 concentrations from 10 -2 M to 10 -8 M. The SERS intensity ratio log(I882/I996) increases along with increasing H2O2 concentrations.Moreover, log(I882/I996) also increased when the pH became more alkaline under same H2O2 concentration.For example, I882/I996 at pH 9 and [H2O2] = 10 -6 M was approx.4 times higher than at pH 4. These results indicate an influence of pH on H2O2 quantification although exclude cross talks since the H2O2 sensitive bands are not affected in the absence of the analyte.

SERS response of NCs@3-MP (the oxidized form of 3-MPBA) does not depend on environmental pH.
H2O2 oxidized 3-MPBA into 3-MP.We functionalized AuNPs@NCs with 3-MP and measured the SERS spectra at different pH values in the absence of H2O2 to check the stability and specificity of the signal (figure SI-8).We observed no change (e.g., band shift, intensity ratios, among others) in its vibrational mode, thus concluding no cross talk and confirming signal specificity of 3-MPBA to H2O2.

2.3
The pKa of 3-MPBA determines the sensitivity of boronic acid-based H2O2 sensors.
Let's have a closer look to the molecular H2O2 sensor, 3-MPBA.An important parameter of 3-MPBA is its pKa value, which is a measure of its Lewis acidity.This value determines the ratio between the trigonal boronic acid and the tetragonal boronate ion (negatively charged) at a specific pH value (figure SI-9A).In the case of NCs@3-MPBA, the inner gold surface will be fully negatively charged when dispersed in a strong basic solution because pKa of phenylboronic acid monolayers is reported to be 9.2 14 .These structural changes of the molecule can be observed using SERS spectroscopy.More specifically, the relative SERS intensity of the 3-MPBA band that is assigned to the non-totally symmetric ring stretching mode (1553 cm - 1 ) varies depending on the environmental pH values, which is a manifestation of charge transfer (CT) processes 15 .Figure SI-9B shows SERS spectra of NCs@3-MPBA dispersed in phosphate buffered saline with pH ranging from 4 to 9. The band at 1553 cm -1 decreases in intensity when pH increases.More importantly, pH had relatively low effect on the H2O2 sensitive peak at 882 cm -1 in the absence of H2O2, confirming no crosstalk between the signals and ensuring that changes in the 882 cm -1 peak were due to H2O2 variations and not to pH changes.3-MPBA oxidized into 3-MP converting boronate to hydroxyl functional group.This oxidative process produces the rupture of the B-C chemical bond.Comparing with trigonal boronic acid, the complexation with a third hydroxyl group in tetragonal boronate ion facilitates the oxidation to hydroxyl in the presence of H2O2, which enhanced this B-C bond cleavage sensitivity and enhances 3-MPBA oxidation by H2O2 base 16 .The equilibrium constants of this oxidation reaction vary at specific pH.This can be estimated using the Henderson-Hasselbach equation using the pKa of the 3-MPBA.Under basic pH, lower amount of H2O2 is needed than in acid pH to achieve equivalent SERS readout because the reaction is favored.With same amount of H2O2, since the equilibrium is different for different pH values, the SERS readout will be different, and thus the limit of detection (LODs) under different pH are going to be also different.The sensitivity of our sensor NCs@3-MPBA is pH-dependent, being maximum at high pH (7-9) and lowers with decreasing pH (6-4).In general, H2O2 measurements are based on direct or indirect oxidation of a probe by H2O2 5 , thus the pH effect on H2O2 measurements can be applied to all H2O2 sensors which are based on aromatic boronic acid coupled with fluorescence or SERS.(B) SERS spectra of NCs@3-MPBA dispersed in phosphate buffer with pH ranging from 4 to 9, showing a decreasing intensity at 1553 cm -1 (peak ascribed to charge transfer processes) with increasing pH.Each spectrum was the average of 5 spectra obtained from 5 different NCs@3-MPBA.
Our H2O2 sensor's (NCs@3-MPBA&4-MBA) LOD was calculated based on visual definition 17 .We used logarithmic scales for both the horizontal and vertical axes (i.e., log (I882/I996) and log [H2O2]), which broaden the linear range to four orders of magnitude (e.g., for pH7 [H2O2] from 10 -2 M to 10 -6 M), since the linear ranges previous reported were in two orders or less based on numerical scales 1,8,18 .Figure 2     One of the most common and simply method to understand intracellular H2O2 functions is to add H2O2 itself directly to the experimental system.To demonstrate the feasibility of mimicking cell stress upon exposure to H2O2, we used a fluorescent probe to monitor intracellular changes (figure SI-13A). Figure SI-13A shows cells transfected with orp1-GFP specifically labeling intracellular H2O2.After addition of 500 μM H2O2 to the cell's growth media, the fluorescence decreases, indicating the presence of intracellular H2O2.By calculating integrated density of CLSM images, we observed that the intracellular H2O2 level reached a plateau after 10 min treatment (figure SI-13B).We further measure possible cytotoxicity issues that could be derived from the exposure to H2O2 to discard erratic cell stress that could affect our sensing.
We measured toxicity at the level of mitochondrial activity (figure SI-13C) and cell membrane integrity (figure SI-13D) and confirmed that exposure to H2O2 was not cytotoxic and the cells exhibited a cell viability of higher than 90%.

Inducing intracellular alkalinization to alter pH homeostasis.
Bafilomycin A1 is one commonly used agent modifying lysosomal acidification 10 .The vacuolar ATPase (V-ATPase) is a proton pump hydrolysing ATP, controlling the acidification of endosomes and lysosomes.Bafilomycin A1 could inhibit the activity of the V-ATPase 19,20,21 .Thus, lysosomal pH increase upon the addition of Bafilomycin A1 20 .We also checked the ability in our cell line by using lysotracker as a fluorescence pH indicator shown in figure SI-14.With Bafilomycin A1 500 nM treated after 2 hours, there was no signals coming from lysotracker, which meant the lysosomal pH was not acidic.

4.5
High irradiation time causes photosublimation of 4-MBA which results in an altered spectrum.
SERS measurements in biological environment such as a cell is complex due to many interferences 22 this cause that the irradiation parameters set up for cells are more extreme than in a buffer.For example, irradiation time of the NCs is higher due to the scattering from the cellular structures and the difficulty to focus as much as light as possible to obtain enough SERS intensity.We observed that increasing the irradiation time affects the stability of the molecular sensor.We took NCs@3-MPBA&4-MBA functionalized at a ratio 5:1 and dispersed in a phosphate saline buffer at pH 9 without H2O2.Figure SI-17 shows the SERS response after their irradiation of the same NC at a 5 mW laser power during 5 s and 20 s.The intensities of peaks at 1385 cm -1 and 1590 cm -1 , corresponding to symmetric carboxyl stretching mode and aromatic ring vibrations of 4-MBA, decreased after 20 s.Possibly, high irradiation times induce longlasting increase in the local temperature of the plasmonic nanostructure (Au) where 4-MBA is conjugated.An increased and continuous heating of 4-MBA after laser irradiation can result in a photo-induced sublimation 23 of 4-MBA.We observed this effect on 4-MBA but not on 3-MPBA at the irradiation times described.Thus, the photosensitivity of the Raman probes is different most probably because of their different molecular structure.We can conclude that the irradiation time must be check when selecting a Raman probe for biosensing as it may interfere with the quantification.

Figure SI-17: SERS spectra of NCs@3-MPBA&4-MBA:
The NCs were functionalized at a 5:1 ratio (5x10 -3 M 3-MPBA : 10 -3 M 4-MBA) and dispersed in a phosphate buffered saline at pH 9. Individual NCs were irradiated for 5 s and 20 s at a fixed laser power (5 mW) and the spectra were collected.Red arrows indicate affected peaks at 1385 cm -1 and 1590 cm -1 corresponding to 4-MBA.

Figure SI- 1
Figure SI-1 shows the functional nanosensors' characterization during different steps of the synthesis with Transmission Electron Microscopy (TEM) (A-C), Dynamic Light Scattering (DLS) (D-E), and SERS (F).Briefly, the synthesis of the NCs is performed wrapping polyelectrolytes of opposite charge (PSS and PAH) layer by layer (LbL) onto a template (polystyrene, PS) being the last layer, a deposition of gold seeds.We grow a final silica (SiO2) shell to protect the plasmonic nanostructure and to preserve the SERS signal.Afterwards, we dissolve the PS template chemically and grow the gold seeds inside into a plasmonic nanostructure showing a characteristic LSPR.DLS measurements show a hydrodynamic size of 467.6 nm and PDI 0.044 (SI-1D) and a ζ-potential mean value -36.7 mV.

Figure SI- 3 :
Figure SI-3: SERS response characterization of the H2O2 nanosensors, NCs@3-MPBA, in different biological media.(A) Kinetic study based on the intensity ratio of 882 cm -1 to 996 cm -1 by measuring NCs solution mixed with 10 -3 M H2O2 in pH7 phosphate buffered saline.(B) NCs@3-MPBA dispersed in cells growth media with different H2O2 concentrations (10 -2 M, 10 -3 M, 10 -4 M, 10 -5 M, 10 -6 M, 10 -7 M, 10 -8 M) and without H2O2 (from red to green).Each spectrum was the average of 5 spectra gotten from 5 different NCs@3-MPBA.(C) Calibration curves of NCs@3-MPBA determining H2O2 in phosphate buffer and in cells growth media followed the same trend, indicating that calibration curves gotten in buffer system can be easily used in growth media environment.
Figure SI-5B shows the SERS spectra of a series of NCs prepared with different 3-MPBA and 4-MBA ratios, and the zoomed are of the spectra showing the variations of the selected bands.The relative intensities of the marker bands at 1075 cm -1 to 996 cm -1 changed dramatically.Only 4-MBA bands at 1075 cm -1 appeared in the spectra at lower concentration ratio.Figure SI-5C shows the relation of I1075/I996 against the ratios of the two Raman probes.When the 3-MPBA:4-MBA ratio increased from 5 to 20 times, I1075/I996 decreased from 3 to 1.4.We selected a ratio of 15:1 (3-MPBA:4-MBA) as working condition, because the I1075/I996 was round 1.55, thus indicating equivalent signals for both probes.

Figure SI- 6 .
Figure SI-6.NCs@3-MPBA&4-MBA's responsiveness to H2O2 and pH dynamics.(A) Calibration curves of NCs@3-MPBA&4-MBA for pH in phosphate buffer (from 4 to 9).The averages and standard deviations were calculated with the spectra in the presence and absence of H2O2 within the same pH.(B) SERS spectra of NCs@3-MPBA&4-MBA in phosphate buffer at different pH and H2O2 concentrations, with pH ranging from 4 to 9, and H2O2 from 10 -2 M to 10 -8 M and without H2O2 (from red to green, H2O2 concentration 10 -2 M, 10 -3 M, 10 -4 M, 10 -5 M, 10 -6 M, 10 -7 M, 10 -8 M, and without H2O2), showing how the intensity at 882 cm -1 decreased from red to green and the intensity at 1385 cm -1 varied along with pH.Each spectrum was the average of 5 spectra gotten from 5 different NCs@3-MPBA&4-MBA.

Figure SI- 8 :
Figure SI-8: SERS spectra of NCs@3-MP at varying pH in the absence of H2O2.We took the oxidized form, 3-MP, of the molecular sensor, 3-MPBA, and functionalized the NCs with it.We dispersed the NCs in phosphate buffer at different pH ranging from pH 4 to 9 and measured the SERS response in the absence of H2O2.

Figure SI- 9 :
Figure SI-9: SERS spectra of NCs@3-MPBA at varying pH in the absence of H2O2.(A) Scheme of 3-MPBA format in acidic pH (phenylboronic acid) and alkaline pH (boronate acid).(B) SERS spectra of NCs@3-MPBA dispersed in phosphate buffer with pH ranging from 4 to 9, showing a decreasing intensity at 1553 cm -1 (peak ascribed to charge transfer processes) with increasing pH.Each spectrum was the average of 5 spectra obtained from 5 different NCs@3-MPBA.
shows the different H2O2 calibration curves for each pH and the pH-dependent LODs of the sensor while figure SI-10 shows the corresponding calibration equations.The LOD of H2O2 at acid and neutral pH (pH 4 to pH 7) was around 10 -6 M and close to 10 -8 M for pH 8 and pH 9. The behavior was like the individual nanosensor (NCs@3-MPBA) (figure SI-11).

Figure SI- 11 : 9 . 3 :
Figure SI-11: NCs@3-MPBA calibration curves and linear ranges for H2O2.(A) SERS measurements were performed under different pH (from 4 to 9).Red lines are linear fitting results.Green dots were masked (data points not included in the linear fitting).Error bars represented the standard deviations of five probes.(B) Calibration curves equations and LODs of NCs@3-MPBA for H2O2 sensing in phosphate buffer with pH ranging from 4 to 9.

Figure SI- 12 : 4 : 4 . 1
Figure SI-12: NCs cellular internalization and biocompatibility.(A-B) Cellular uptake of NCs by HT29 cells using confocal laser scanning microscopy (CLSM).(A) Z-scan of a cells area.(B) Different planes (X/Y; X/Z; Y/Z) of selected area (dashed square in A).NCs were shown with white arrow and white dash circle.Internalized NCs were localized within lysosomes as observed by the co-localization of the signal intensities of the different dyes and the reflected light of the NCs.Lysosomes: green; NCs: red; and cells membrane: magenta.(C) Cytotoxicity assay of NCs in HT29 cells.Cell viability was determined by Resazurin-Based Assay after the internalization of NCs.The concentrations of NCs were calculated by number.Optical images were collected with OLYMPUS CKX41 inverted microscope.NCs concentration increased from C1 to C8.Since C7 (NCs concentration higher than 1.15 pmol/L), the HT29 cells were fully covered by NCs.

Figure SI- 15
Figure SI-15 shows the intracellular and extracellular NCs@3-MPBA SERS spectra collected with untreated HT29 and H2O2 treated HT29 cells.By comparing intensity ratio log(I882/I996) of untreated cells and H2O2 treated cells (10 mM, 1 mM, and 0.5 mM), we observed that both the intracellular and extracellular H2O2 level increased when H2O2 concentration increased, indicating the sensing ability of our sensor.Meanwhile, the intracellular signal was much lower than the extracellular signal.

Figure 4
Figure 4 shows the distribution of I1385/I996 and log(I882/I996) obtained from individual NCs' spectra corresponding to the different pH and H2O2 values measured for all samples whereas figures 4C and 4D show the average value of different experiments and the standard deviation of I1385/I996 (pH) and log(I882/I996) (H2O2) separately.By knowing the intensity ratio of I1385/I996 and log(I882/I996) (figure 4B-4D) and relating them to the pH (cf., SI §1, figure SI-6A) and H2O2 calibration curves (figure 2 and cf., SI §2, figure SI-10), we can estimate the pH and H2O2 values for all samples.

Figure SI- 16
Figure SI-16 shows the complete SERS spectra from which the zoomed are shown in figure 4A has been taken.

Table SI -
1 shows a comparison of other systems different than 3-MPBA published for hydrogen peroxide detection.It includes commercial boronate based fluorescent probes, genetically encoded fluorescent probes, and other SERS system.In generally, SERS assays show better sensitivity than fluorescence assays, and great potential in multiplex analysis because of the easy functionalization of SERS probes.