Design and Synthesis of Novel Raman Reporters for Bioorthogonal SERS Nanoprobes Engineering

Surface-enhanced Raman spectroscopy (SERS) exploiting Raman reporter-labeled nanoparticles (RR@NPs) represents a powerful tool for the improvement of optical bio-assays due to RRs’ narrow peaks, SERS high sensitivity, and potential for multiplexing. In the present work, starting from low-cost and highly available raw materials such as cysteamine and substituted benzoic acids, novel bioorthogonal RRs, characterized by strong signal (103 counts with FWHM < 15 cm−1) in the biological Raman-silent region (>2000 cm−1), RRs are synthesized by implementing a versatile, modular, and straightforward method with high yields and requiring three steps lasting 18 h, thus overcoming the limitations of current reported procedures. The resulting RRs’ chemical structure has SH-pendant groups exploited for covalent conjugation to high anisotropic gold-NPs. RR@NPs constructs work as SERS nanoprobes demonstrating high colloidal stability while retaining NPs’ physical and vibrational properties, with a limit of detection down to 60 pM. RR@NPs constructs expose carboxylic moieties for further self-assembling of biomolecules (such as antibodies), conferring tagging capabilities to the SERS nanoprobes even in heterogeneous samples, as demonstrated with in vitro experiments by transmembrane proteins tagging in cell cultures. Finally, thanks to their non-overlapping spectra, we envision and preliminary prove the possibility of exploiting RR@NPs constructs simultaneously, aiming at improving current SERS-based multiplexing bioassays.


Introduction
Nowadays, the research in the field of diagnostics and biosensors is more and more focused on the development of novel methods for the molecular screening of biological fluids. As a matter of fact, liquid biopsies have emerged as powerful diagnostic tools not only because of their advantages related to the reduction in invasiveness and risks for the patients but also because it is possible to discriminate between physiological and pathological conditions by precociously identifying analytes (chemicals, biomolecules, cells, etc.) at concentration down to picomolar (pM) [1]. To address this issue, analytical devices must be characterized by high specificity, high sensitivity, quickness, and accuracy. In addition, there is a strong demand for high-throughput analysis enabling to qualitatively or quantitatively reveal multiple analytes within the same reduced volume of the sample [2]. The simultaneous profiling of multiple markers could improve the performances of these analytical devices, and moreover, it is fundamental for those diseases whose early-stage diagnosis relies on the ratio of at least two pathological biomarkers [3]. In this scenario,

Rational Design and Synthesis of Novel RRs Molecules
New molecular entities acting as bioorthogonal RRs were rationally designed to show (i) high Raman cross section, (ii) strong single narrow vibrational peaks in the biological silent region (1800-2800 cm −1 ), and (iii) proper pending functional groups to be attached to AuNPs avoiding alteration of colloidal stability. To this end, as depicted in the scheme in Figure 1, the RRs are conceived as constituted by two building blocks covalently coupled together: the "NPs anchoring block" and the "bioorthogonal Raman block". The "NPs anchoring block" should contain at least a thiol end-chain group for NPs covalent conjugation by Au-S bond formation and a second pendant functional group for covalently coupling with the "bioorthogonal Raman block" (Figure 1). With the purpose of developing a very modular, efficient, and versatile synthesis, we selected the amide bond formation as a coupling reaction. To this end, raw cysteamine 1 was identified as "NPs anchoring block", as the amine functional group was exploited to couple selected benzoic acid derivatives through standard peptide bonds formation. For the bioorthogonal Raman block, benzoic acids were preferred to aliphatic carboxylic acids since the π-electrons of the aromatic rings are characterized by high polarizability leading to a higher Raman scattering cross section [25]. Indeed, our approach enables us to customize a library of bioorthogonal RRs by simply exploring varied benzoic acid derivatives in the synthesis.
sandwiches and in vitro cell cultures [6]. Finally, the proposed method potentially enables the synthesis of a larger set of tags, hence providing a straightforward method for custom multicolor palette for application in SERS-based multiplexing analyses.

Rational Design and Synthesis of Novel RRs Molecules
New molecular entities acting as bioorthogonal RRs were rationally designed to show (i) high Raman cross section, (ii) strong single narrow vibrational peaks in the biological silent region (1800-2800 cm −1 ), and (iii) proper pending functional groups to be attached to AuNPs avoiding alteration of colloidal stability. To this end, as depicted in the scheme in Figure 1, the RRs are conceived as constituted by two building blocks covalently coupled together: the "NPs anchoring block" and the "bioorthogonal Raman block". The "NPs anchoring block" should contain at least a thiol end-chain group for NPs covalent conjugation by Au-S bond formation and a second pendant functional group for covalently coupling with the "bioorthogonal Raman block" (Figure 1). With the purpose of developing a very modular, efficient, and versatile synthesis, we selected the amide bond formation as a coupling reaction. To this end, raw cysteamine 1 was identified as "NPs anchoring block", as the amine functional group was exploited to couple selected benzoic acid derivatives through standard peptide bonds formation. For the bioorthogonal Raman block, benzoic acids were preferred to aliphatic carboxylic acids since the π-electrons of the aromatic rings are characterized by high polarizability leading to a higher Raman scattering cross section [25]. Indeed, our approach enables us to customize a library of bioorthogonal RRs by simply exploring varied benzoic acid derivatives in the synthesis. Compared to state-of-the-art studies from the literature focusing on the synthesis of novel bioorthogonal RRs for NPs conjugation [16,26], our synthetic strategy consists of only three low-cost and high-yield steps: (i) protection of the thiol functionality of the NPs anchoring block with triphenylmethanol, (ii) peptide coupling reaction with differently substituted commercially available benzoic acids, (iii) thiol deprotection with trifluoroacetic acid and triisopropylsilane (Figure 2a). It is important to highlight that both compounds composing the backbone of the bioorthogonal RRs are low-cost and commercially available molecules, contrary to commercial RRs generally belonging to the class of bifunctional thiophenols where the simultaneous direct conjugation of the thiol group and the Raman active group to the benzene ring can alter the reactivity as well as Compared to state-of-the-art studies from the literature focusing on the synthesis of novel bioorthogonal RRs for NPs conjugation [16,26], our synthetic strategy consists of only three low-cost and high-yield steps: (i) protection of the thiol functionality of the NPs anchoring block with triphenylmethanol, (ii) peptide coupling reaction with differently substituted commercially available benzoic acids, (iii) thiol deprotection with trifluoroacetic acid and triisopropylsilane (Figure 2a). It is important to highlight that both compounds composing the backbone of the bioorthogonal RRs are low-cost and commercially available molecules, contrary to commercial RRs generally belonging to the class of bifunctional thiophenols where the simultaneous direct conjugation of the thiol group and the Raman active group to the benzene ring can alter the reactivity as well as the stability of both moieties. Here, as a proof of concept, four probes were successfully obtained by exploiting p-ethynyl-, p-phenylethynyl-, p-cyano-and m-fluoro-, and p-cyano-benzoic acids, all of them characterized by alkyne stretching oscillation with vibrational frequencies above 2000 cm −1 . As reported in Figure 2a, the thiol functionality of 1 was protected with triphenylmethanol in the presence of trifluoroacetic acid, thus achieving compound 2 in 89% yield, without the need for preventive protection of the amino group, due to the chemoselectivity of this reaction. obtained by exploiting p-ethynyl-, p-phenylethynyl-, p-cyano-and m-fluoro-, and pcyano-benzoic acids, all of them characterized by alkyne stretching oscillation with vibrational frequencies above 2000 cm −1 . As reported in Figure 2a, the thiol functionality of 1 was protected with triphenylmethanol in the presence of trifluoroacetic acid, thus achieving compound 2 in 89% yield, without the need for preventive protection of the amino group, due to the chemoselectivity of this reaction. Then, benzoic acids were attached to the amine functionality of 2 by using HATU as a selective coupling reagent in the presence of DIPEA and DMF. In this way, different N-(2-mercaptoethyl)benzamide 3a-d were obtained in a range of yields spanning from 54% to 92%, depending on the steric hindrance and the electrophilicity of the carboxylic acid functionality. p-bromo-benzamide derivative 3i was transformed into the desired pphenylethynyl-benzamide 3b by using phenylacetylene in the Sonogashira coupling reaction condition. Finally, the thiol protecting group of 3a-d was removed using a cocktail mixture of trifluoroacetic acid and triisopropylsilane, giving compounds 4a-d as highly stable solids. At the end of the synthesis, solid compounds 4a-d, from here on named RR1-RR4, were obtained with higher yields compared to reported cases from the literature, soluble at least in ethanol that is crucial for the step of conjugation to NPs.
With the goal of multiplexing, their Raman spectroscopic properties were measured, and as expected, RR1-RR4 Raman profiles (Figure 2c) were still characterized by overelaborated spectra in the biological fingerprint region (from 1000 to 1800 cm −1 ), with multiple peaks stating for single C-H, C-C, and C-S bonds, carboxylic groups, amide as well as phenyl rings stretching and bending, but also characterized by a single sharp peak in the 2050 cm −1 to 2245 cm −1 silent region range and highly resolved from each other. Lower frequencies are associated with C≡C stretching, whose smaller reduced mass (μ = 6) produced shorter vibrations. On the other hand, as the reduced mass increases with the C≡N bond (μ = 6.5), peaks are located at higher frequencies. Moreover, the presence of electron-donating or electron-withdrawing substituting groups could influence and shift the vibrational resonance [21]. Then, benzoic acids were attached to the amine functionality of 2 by using HATU as a selective coupling reagent in the presence of DIPEA and DMF. In this way, different N-(2-mercaptoethyl)benzamide 3a-d were obtained in a range of yields spanning from 54% to 92%, depending on the steric hindrance and the electrophilicity of the carboxylic acid functionality. p-bromo-benzamide derivative 3i was transformed into the desired p-phenylethynyl-benzamide 3b by using phenylacetylene in the Sonogashira coupling reaction condition. Finally, the thiol protecting group of 3a-d was removed using a cocktail mixture of trifluoroacetic acid and triisopropylsilane, giving compounds 4a-d as highly stable solids. At the end of the synthesis, solid compounds 4a-d, from here on named RR 1 -RR 4 , were obtained with higher yields compared to reported cases from the literature, soluble at least in ethanol that is crucial for the step of conjugation to NPs.
With the goal of multiplexing, their Raman spectroscopic properties were measured, and as expected, RR 1 -RR 4 Raman profiles (Figure 2c) were still characterized by overelaborated spectra in the biological fingerprint region (from 1000 to 1800 cm −1 ), with multiple peaks stating for single C-H, C-C, and C-S bonds, carboxylic groups, amide as well as phenyl rings stretching and bending, but also characterized by a single sharp peak in the 2050 cm −1 to 2245 cm −1 silent region range and highly resolved from each other. Lower frequencies are associated with C≡C stretching, whose smaller reduced mass (µ = 6) produced shorter vibrations. On the other hand, as the reduced mass increases with the C≡N bond (µ = 6.5), peaks are located at higher frequencies. Moreover, the presence of electron-donating or electron-withdrawing substituting groups could influence and shift the vibrational resonance [21].

SERS Bioorthogonal Nanoprobes Engineering and Characterization
SERS bioorthogonal nanoprobes RR@AuNSts were engineered by conjugating RR 1 -RR 4 to the surfaces of AuNPs. Based on previous results from our groups [27,28], highly branched star-shaped particles (AuNSts) were considered as SERS enhancers due to the convenient overlap between AuNSts plasmonic properties matching the range of excitation wavelength used for Raman analysis of liquid systems as well as for AuNSts morphology with intrinsic "hot spots" [29]. Then, a "one-pot" competitive method was implemented to simultaneously intercalate thiolated polyethylene glycol (PEG) and RRs molecules, both co-working as NPs-capping agents ( Figure 3).
RR4 to the surfaces of AuNPs. Based on previous results from our groups [27,28], highly branched star-shaped particles (AuNSts) were considered as SERS enhancers due to the convenient overlap between AuNSts plasmonic properties matching the range of excitation wavelength used for Raman analysis of liquid systems as well as for AuNSts morphology with intrinsic "hot spots" [29]. Then, a "one-pot" competitive method was implemented to simultaneously intercalate thiolated polyethylene glycol (PEG) and RRs molecules, both co-working as NPs-capping agents ( Figure 3). To this end, while PEG/NP ratio was kept constant at 500 k to achieve the 100% particle coverage as previously demonstrated [6], different RR/NP ratios were investigated to study the effect of RRs at 5 k, 10 k, 25 k, and 50 k ratios. Optimal PEG/NPs/RRs incubation conditions were those giving brighter bioorthogonal Raman signals while retaining SERS nanotags morphology and avoiding particle aggregation. From this perspective, Raman and UV-Vis absorption spectra were acquired over time for different batches ( Figure 4). Experimental data were analyzed as a function of RRs loading in terms of the Raman Intensity Value (RIV) parameter, which quantifies the intensity of the specific bioorthogonal peaks, and DeltA parameter, defined as the percentage of variation of the NPs resonant peak within 7 days. The highest RIVs combined with the lowest DeltAs identify the RR@NPs working window conditions. As shown in the graphs in Figure 4a,b, despite the decrease in the Raman signal when passing from 5 k to 10 k, RIVs for RR1 and RR3 @AuNSts 10 k did not significantly vary with respect to 5 k (NS), indicating a slight variability between 5 k and 10 k conjugation process. Conversely, in the same range from 5 k to 10 k, RIV for RR2@AuNSts significantly decreased (p < 0.0001). For RR/NP ratios higher than 10 k RIVs for RR1, RR2, and RR3 dramatically decreased (p < 0.0001). Finally, RR4@AuNSts showed high signal stability at all RR/NP ratios. With respect to the DeltA parameter, the graphs in Figure 4c,d clearly display that for RR1, RR2, and RR3, colloidal stability was immediately lost after the conjugation processes, as related UV-Vis spectra were characterized by flat, broad, and red-shifted curves, stating for uncontrolled aggregation phenomena. This supports the hypothesis that at higher ratios, there are too many RR molecules adhering to gold and competing with PEG stabilizing agent in decorating NPs surfaces, thus affecting PEG steric-stabilization layer efficacy and causing NPs clustering through irreversible interactions. Finally, in the 5 k-10 k range, improved stability at 7 days was observed for RR1, RR2, and RR3 @AuNSts, while again, To this end, while PEG/NP ratio was kept constant at 500 k to achieve the 100% particle coverage as previously demonstrated [6], different RR/NP ratios were investigated to study the effect of RRs at 5 k, 10 k, 25 k, and 50 k ratios. Optimal PEG/NPs/RRs incubation conditions were those giving brighter bioorthogonal Raman signals while retaining SERS nanotags morphology and avoiding particle aggregation. From this perspective, Raman and UV-Vis absorption spectra were acquired over time for different batches ( Figure 4). Experimental data were analyzed as a function of RRs loading in terms of the Raman Intensity Value (RIV) parameter, which quantifies the intensity of the specific bioorthogonal peaks, and DeltA parameter, defined as the percentage of variation of the NPs resonant peak within 7 days. The highest RIVs combined with the lowest DeltAs identify the RR@NPs working window conditions. As shown in the graphs in Figure 4a,b, despite the decrease in the Raman signal when passing from 5 k to 10 k, RIVs for RR 1 and RR 3 @AuNSts 10 k did not significantly vary with respect to 5 k (NS), indicating a slight variability between 5 k and 10 k conjugation process. Conversely, in the same range from 5 k to 10 k, RIV for RR 2 @AuNSts significantly decreased (p < 0.0001). For RR/NP ratios higher than 10 k RIVs for RR 1 , RR 2 , and RR 3 dramatically decreased (p < 0.0001). Finally, RR 4 @AuNSts showed high signal stability at all RR/NP ratios. With respect to the DeltA parameter, the graphs in Figure 4c,d clearly display that for RR 1 , RR 2 , and RR 3 , colloidal stability was immediately lost after the conjugation processes, as related UV-Vis spectra were characterized by flat, broad, and red-shifted curves, stating for uncontrolled aggregation phenomena. This supports the hypothesis that at higher ratios, there are too many RR molecules adhering to gold and competing with PEG stabilizing agent in decorating NPs surfaces, thus affecting PEG steric-stabilization layer efficacy and causing NPs clustering through irreversible interactions. Finally, in the 5 k-10 k range, improved stability at 7 days was observed for RR 1 , RR 2 , and RR 3 @AuNSts, while again, negligible changes were measured for RR 4 @AuNSts at all ratios. Finally, according to experimental data, RR/NP optimal ratios were 10 k for RR 1 and 5 k for RR 2 , RR 3 , and RR 4 . negligible changes were measured for RR4 @AuNSts at all ratios. Finally, according to experimental data, RR/NP optimal ratios were 10 k for RR1 and 5 k for RR2, RR3, and RR4. TEM micrographs showed that NPs morphology was perfectly retained when implementing the optimized RRs conjugation process, thus guaranteeing that also the resonant peak is retained (Figure 5a). DLS and zeta potential (ζ) measurements further attested that a successful ligand exchange process was achieved since the hydrodynamic diameter (Dh) increased by about 50-60 nm with respect to citrate-capped gold nanostars (whose characteristic Dh was 81.3 ± 1.8 nm, data not shown). At the same time, ζ values decreased by at least 3 mV from −24.2 ± 1.5 mV upon PEG-RRs intercalating for RR@AuNSts conjugates (Figure 5b). TEM micrographs showed that NPs morphology was perfectly retained when implementing the optimized RRs conjugation process, thus guaranteeing that also the resonant peak is retained (Figure 5a). DLS and zeta potential (ζ) measurements further attested that a successful ligand exchange process was achieved since the hydrodynamic diameter (Dh) increased by about 50-60 nm with respect to citrate-capped gold nanostars (whose characteristic Dh was 81.3 ± 1.8 nm, data not shown). At the same time, ζ values decreased by at least 3 mV from −24.2 ± 1.5 mV upon PEG-RRs intercalating for RR@AuNSts conjugates (Figure 5b). Finally, to determine the limit of detection (LOD) of the SERS nanoprobes, a concentration series of RR1-4@AuNSts aqueous solutions was prepared in a triplicate manner within the 10-0.1 nM range that is considered relevant in clinical practice, also considering that no sample drying was due. As illustrated in Figure 6, the SERS intensity of the RR1-4 specific bioorthogonal peaks was plotted against the probe concentration to build the calibration curves. For all the RR@AuNSts, the linear trend was observed almost in the 0.1-1.5 nM range. Saturation was reached at higher values ( Figure S1). In the linear region, the calibration curves were fitted, and the error bars are standard deviations from three measurements of each sample. Then, LOD was determined as three times the mean of the standard deviation of blank divided by the slope of the calibration curves, while the standard deviation was calculated considering the error propagation (see Supporting Information Figures S1-S21). It resulted that all SERS bioorthogonal complexes are all characterized by LOD values below 1 nM, with slight differences ranging from 60 pM to 1 nM, a key aspect for their parallel exploitability in a bioassay for multiplexing analysis.  Finally, to determine the limit of detection (LOD) of the SERS nanoprobes, a concentration series of RR 1-4 @AuNSts aqueous solutions was prepared in a triplicate manner within the 10-0.1 nM range that is considered relevant in clinical practice, also considering that no sample drying was due. As illustrated in Figure 6, the SERS intensity of the RR 1-4 specific bioorthogonal peaks was plotted against the probe concentration to build the calibration curves. For all the RR@AuNSts, the linear trend was observed almost in the 0.1-1.5 nM range. Saturation was reached at higher values ( Figure S1). In the linear region, the calibration curves were fitted, and the error bars are standard deviations from three measurements of each sample. Then, LOD was determined as three times the mean of the standard deviation of blank divided by the slope of the calibration curves, while the standard deviation was calculated considering the error propagation (see Supporting Information Figures S1-S21). It resulted that all SERS bioorthogonal complexes are all characterized by LOD values below 1 nM, with slight differences ranging from 60 pM to 1 nM, a key aspect for their parallel exploitability in a bioassay for multiplexing analysis. Finally, to determine the limit of detection (LOD) of the SERS nanoprobes, a concentration series of RR1-4@AuNSts aqueous solutions was prepared in a triplicate manner within the 10-0.1 nM range that is considered relevant in clinical practice, also considering that no sample drying was due. As illustrated in Figure 6, the SERS intensity of the RR1-4 specific bioorthogonal peaks was plotted against the probe concentration to build the calibration curves. For all the RR@AuNSts, the linear trend was observed almost in the 0.1-1.5 nM range. Saturation was reached at higher values ( Figure S1). In the linear region, the calibration curves were fitted, and the error bars are standard deviations from three measurements of each sample. Then, LOD was determined as three times the mean of the standard deviation of blank divided by the slope of the calibration curves, while the standard deviation was calculated considering the error propagation (see Supporting Information Figures S1-S21). It resulted that all SERS bioorthogonal complexes are all characterized by LOD values below 1 nM, with slight differences ranging from 60 pM to 1 nM, a key aspect for their parallel exploitability in a bioassay for multiplexing analysis.

SERS Bioorthogonal Nanoprobes against Biological Relevant Environments
In vitro experiments were performed to demonstrate that bioorthogonal SERS nanoprobes can be further modified to tag specific analytes, even in heterogeneous and molecular complex systems, such as cell cultures. As a matter of fact, we report the first example of bioorthogonal RRs SERS nanoprobes showing simultaneously multiple functionalities as cases from literature are either based on RRs molecules not available for direct covalent absorption on SERS NPs or do not provide pendant groups for further conjugation with molecules for biological targeting. For this experimental study, brighter RR 2 @AuNSts was used for demonstration, and -COOH NPs-pendant moieties from assembled PEG were exploited to conjugate Alexa550-WGA, a fluorescent probe selectively tagging cellular membrane proteins. Incidentally, the implemented bio-functionalization protocol is based on standard peptide bond formation, thus exploitable to conjugate all capturing agents exposing amino functionalities. HEK cell cultures were incubated for 1 h with Alexa550-WGA-RR 2 @AuNSts and then washed thoroughly with PBS buffer to remove unbound SERS probes. Cells incubation with bare RR 2 @AuNSts and free Alexa550-WGA solution were performed as controls. As shown in the confocal images reported in Figure 7a, the fluorescence signal from Alexa550-WGA was observed for cells incubated in the presence of WGA, while no signal was observed when incubating with bare RR 2 @AuNSts. Furthermore, the fluorescence intensity from cells incubated with free Alexa550-WGA solution was comparable with that from cells incubated with Alexa550-WGA-RR 2 @AuNsts (Figure 7b). This attested to the high selectivity and specificity of the SERS constructs and, more importantly, demonstrates that the conjugation to RRs@AuNSts does not interfere in the WGA tagging process of the protein's membrane. Then, the Alexa550-WGA fluorescence signal was exploited to localize SERS constructs and to acquire Raman spectra from cells incubated with Alexa550-WGA-RR 2 @AuNSts and fixed with paraformaldehyde solution since this would not interfere with the measurements in the biological Raman-silent region. Figure 7c displays representative SERS spectra from different regions in the cell cultures, marked with colored squares in Figure 7a (C states for control, S for sample). As shown, spectra obtained from controls C1, C2, and C3 revealed no peak characteristic for RR 2 . Contrariwise, the SERS spectra clearly revealed the strong narrow peak characteristic of the RR 2 when focusing on the cell membranes (marks S1-S3). Successful colocalization of fluorescence and Raman signal paves the way for the exploitation of these RRs@AuNSts for immuno-SERS microscopy.
Finally, proof-of-concept experiments were implemented to investigate the possibility of exploiting SERS nanoprobes for multiplexing analysis. To this end, we first verified that characteristic signals from the engineered probes were still distinguishable even upon their mixing. As shown in Figure 8a, the bioorthogonal peaks of each tag could be clearly identified when acquiring SERS spectra of RR 1 @AuNSts, RR 2 @AuNSts, and RR 4 @AuNSts solution. The 3:1:3 mixing volume ratio was selected to have a starting RRs mixture with comparable Raman intensity from each tag. RR 3 @AuNSts was not considered since it was completely hindered by the RR 4 @AuNSts signal ( Figure S2). This is due to the broadening of the peaks resulting when moving from raw RRs powder to their corresponding nanoconstructs. Then, the potential multiplexing properties of the RR@AuNSts mixed solution were verified in biologically relevant tests, such as biotin-avidin immunoassays used in standard clinical practices. To this end, RRs@AuNSts exposing biotin moieties (b-RRs@AuNSts) were obtained without affecting SERS nanoprobes properties ( Figure S3). Then, biotinmodified polymeric substrates were prepared and further incubated with streptavidin, whose tetrameric structure was exploited for further self-assembling of b-RRs@AuNSts through incubation (Figure 8b). Streptavidin concentration was decreased from 10 µg/mL to 0.1 µg/mL, and Raman-SERS spectra were acquired in random single points of the PDMS assay. For all the concentrations tested, raw spectra presented the RR@AuNSts peaks at 2050 cm −1 , 2218 cm −1 , and 2235 cm −1 , thus attesting that multiplexing behavior, previously tested in solution, was preserved even when implemented on immunoassay. No spurious signal from the background could interfere (Figure 8c), apart from the peak at 9 of 14 2150 cm −1 ascribable to a stretching vibration frequency coming from the polymeric device ( Figure S4). By plotting the Raman intensity of each bioorthogonal peak, in terms of S/N ratio, against streptavidin concentrations (Figure 8d), the same trend of decreasing Raman intensity with decreasing concentration was successfully observed for all the three probes. Furthermore, RRs characteristic signals were detectable even down to 0.1 ug/mL, thus in accordance with the LODs of the systems. Preliminary results support the potentiality of SERS bioorthogonal tags for background-free multiplex and quantitative detection of liquid samples. Finally, proof-of-concept experiments were implemented to investigate possibility of exploiting SERS nanoprobes for multiplexing analysis. To this end, w verified that characteristic signals from the engineered probes were still distinguis even upon their mixing. As shown in Figure 8a, the bioorthogonal peaks of each tag be clearly identified when acquiring SERS spectra of RR1@AuNSts, RR2@AuNSts RR4@AuNSts solution. The 3:1:3 mixing volume ratio was selected to have a startin mixture with comparable Raman intensity from each tag. RR3@AuNSts wa considered since it was completely hindered by the RR4@AuNSts signal ( Figure S2) is due to the broadening of the peaks resulting when moving from raw RRs powd their corresponding nanoconstructs. Then, the potential multiplexing properties RR@AuNSts mixed solution were verified in biologically relevant tests, such as b avidin immunoassays used in standard clinical practices. To this end, RRs@Au exposing biotin moieties (b-RRs@AuNSts) were obtained without affecting nanoprobes properties ( Figure S3). Then, biotin-modified polymeric substrates prepared and further incubated with streptavidin, whose tetrameric structure exploited for further self-assembling of b-RRs@AuNSts through incubation (Figur Streptavidin concentration was decreased from 10 μg/mL to 0.1 μg/mL, and Ramanspectra were acquired in random single points of the PDMS assay. For a concentrations tested, raw spectra presented the RR@AuNSts peaks at 2050 cm −1 cm −1 , and 2235 cm −1 , thus attesting that multiplexing behavior, previously test solution, was preserved even when implemented on immunoassay. No spurious intensity with decreasing concentration was successfully observed for all the three probes. Furthermore, RRs characteristic signals were detectable even down to 0.1 ug/mL, thus in accordance with the LODs of the systems. Preliminary results support the potentiality of SERS bioorthogonal tags for background-free multiplex and quantitative detection of liquid samples.

Synthesis of Raman Reporters (RR) Molecules
Experimental procedures and characterization data for N-(2-mercaptoethyl)benzamide 3a-d and Raman Reporters 4a-d, as well as copies of 1 H and 13 C NMR spectra of all new compounds, are reported in detail in the Supporting information file.

Gold-Nanoparticles Synthesis
Gold nanoparticles (Au-NPs) were synthesized by the seeded-growth process described by Yuan et al. [30]. The seed solution of 15 nm nanospheres (NSps) was prepared by adding 1.5 mL of 30 mM HAuCl 4 ·3H 2 O (1%) to 48.5 mL of boiling and stirring MilliQ (stirring 7 position, 250 • C). After 10 s, 4.5 mL of 38.8 mM sodium citrate solution was added to the solution. The solution was stirred under heating for 15 min and then stirred without heating for 30 min. For nanostars (NSts) synthesis, 0.083 mL of 30 mM HAuCl 4 ·3H 2 O (1% solution) was added to 9.917 mL of MilliQ; 60 µL of 1 M HCl and 100 µL of the NSps solution were added to the solution. Then, 100 µL of 2 mM AgNO 3 and 50 µL of 0.1 M ascorbic acid were added simultaneously. The solution was stirred for 30-60 s, while its color turned from light red to dark grey (or blue). Immediately afterward, NSts were centrifuged for 20 min at 2500× g rpm in 1.5 mL Eppendorf tubes and redispersed in 100 µL of distilled water.

Preparation of RR@AuNSts Complexes
SH-PEG-COOH and RR molecules were exploited in a ligand-exchange process to replace the citrate layer with PEG in citrate-capped nanostars (Cit-NSts). AuNPs solution was diluted to 2 nM, and the same volume of 1 mM SH-PEG-COOH aqueous solution was added together with 2 µL of different concentrations of RR molecules in ethanol. Conditions are reported in the table in Figure S5. The solution was left stirred at RT for at least 2 h, then centrifuged at 25 • C, 2500× g rpm for 40 min, and redispersed in milliQ water or PBS for further bioconjugation. The concentration of particles was determined according to a formula reported by the method of De Puig et al. [31].

Preparation of RR@AuNSts Bioconjugates
RR@AuNSts complexes were prepared with 1 nM solution of the previously prepared solution. Incubation was in 0.1 M phosphate saline buffer (PBS) at a pH of 7.4 for 2 h at room temperature after 15 min activation with 0.4 mM EDC and 0.1 mM NHS (final concentration). Then, 0.5 µL of 1 mg/mL stock solution of antibody (Biotinylated-Ab and Alexa550-WGA) was added to 500 µL of 1 nM RR@AuNSts, resulting in a 100-times dilution (1 µg/mL antibodies solution). Samples were then centrifuged for 30 min at 4 • C and 2500 rpm, and the pellet was redispersed in 100 µL of PBS buffer with BSA 0.5% w/v. For cell incubation, pellets were dispersed in 100 µL of Leibovitz buffer with 3% FBS and 0.5% BSA.

Preparation of the Proof Immunoassay
To fabricate polymeric devices, PDMS liquid precursor and its curing agent were mixed in a 10:1 w/w and poured into the glass mold. Crosslinking of the PDMS was achieved after 12 h at RT and 2 h at 70 • C. Once at RT, the PDMS replica was carefully peeled off from the molds and left again at 70 • C for 2 h to ensure that the PDMS was properly cured at the PDMS/air interface. Then, the surface of the device was functionalized with biotin by incubating with biotinylated-BSA solution (1 mg/mL for 3 min) and further incubating with streptavidin at decreasing concentrations (10 µg/mL, 1 µg/mL, and 0.1 µg/mL for 10 min). The tetrameric structure of streptavidin was exploited to further assemble antibody-conjugated RR@AuNSts (5 nM overnight incubation, 4 • C). After each incubation step, substrates were washed thrice with PBS.

Characterization Techniques
The plasmonic resonant properties of gold nanoparticles colloidal solution before and after functionalization processes were characterized in the range from 400 nm to 850 nm with a UV-Vis-NIR spectrophotometer (Lambda 950 instrument, Perkin Elmer). UV WinLab Software was used to acquire spectra, and data were processed with Origin software. The hydrodynamic dimensions and the polydispersity were acquired by dynamic light scattering (DLS) analysis performed with a Malvern Zetasizer Nano series ZS90. Measurements were performed with a fixed scattering angle of 90 • , at 25 • C. Each sample was measured three times, and each measurement consisted of about 30 acquisitions. Cumulating statistics were used to measure the hydrodynamic diameter and polydispersity. Data were then processed with Origin software. The structural features of the NPs were characterized by transmission electron microscopy (TEM, CM 12 PHILIPS).
Living HEK neuroblastoma cells were fixed with 4% paraformaldehyde solution for 10 min at 4 • C, then thoroughly rinsed with PBS three times. The analysis of Alexa550-WGA fluorescence was performed after excitation at 561 nm, using a Nikon C2 laser scanning confocal microscope and a Plan Fluor 100 × 1.49 NA oil immersion objective. A series of optical sections (1024 × 1024 pixels) at the cell median planes was taken and analyzed using ImageJ software. All settings, including pinhole diameter, detector gain, and laser power, were kept constant for each analysis.
Raman spectra were collected with a conventional micro-Raman setup (XploRA PLUS Confocal Raman Microscope, Horiba), consisting of a 785 nm laser (Coherent) and a spectrometer with a focal length of 500 mm, equipped with a 600 lines/mm grating. The incident laser power on the sample was about 20 mW. The scattered light was detected by a CCD camera operating at about 350 K. Raman-SERS spectra were recorded in the wavenumber range of 800-2500 cm −1 , the acquisition time was 3 s, and the measurement was repeated three times for spectral averaging. To avoid spurious signals, calcium fluoride Raman slides (CaF 2 , Crystan) were used as substrates. To extract the Raman signal of interest, fluorescence and background signals were subtracted from the acquired raw spectra through Vancouver Raman Algorithm, dedicated software for automatic autofluorescence background subtraction for Raman spectroscopy [32]. Data were further analyzed with Origin software.

Conclusions
Herein, we report a versatile, modular, and straightforward method for the high yield synthesis of a set of novel molecules with advanced bioorthogonal Raman properties potentially exploitable for SERS multiplexing bioassays. By combining low-cost and commercially available raw materials, such as cysteamine and differently substituted derivatives of benzoic acids, a series of novel bioorthogonal RRs are obtained and further conjugated to gold nanostars to engineer highly stable SERS nanoprobes retaining optical and physical properties and with LODs in the picomolar clinical range of interest. These background-free RRs@AuNSts can be modified to have bio-tagging specificity for selective targeting of analytes of interest expressed in highly heterogeneous environments such as biological samples. In addition to the stability and to the bio-tagging capability, SERS nanoprobes can be used simultaneously as they retain their bioorthogonal properties of RRs@AuNSts with non-overlapping peaks in the cellular Raman-silent region, thus representing a potential tool to be exploited for parallel detection of more analytes for multiplexed encoding and detection.
Author Contributions: C.D. and C.C. designed and performed the experiments, analyzed the data, and wrote the manuscript. R.I., E.L. and A.T. were involved in designing and performing experiments related to the organic synthesis of molecules. F.S.P. aided with technical advice and fruitful discussion