2.1.

Materials

Heterobifunctional polyethylene glycol (PEG) linker thiol-PEG-n-hydroxysuccinimide with a 1-kDa molecular weight was purchased from Nanocs (New York). About 200 proof ethanol (EtOH) was purchased from VWR (Radnor, Pennsylvania). 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), 4-mercaptobenzoic acid (MBA), and l-citrulline were purchased from Fisher Scientific (Hampton, New Hampshire). Citrulline aptamers were discovered by Base Pair Biotechnologies (Pearland, Texas; Base Pair). After discovery, aptamers were synthesized with a 3′-dithiol group and a 5′-Cy5^™ fluorophore.

2.2.

Particle Characterization

A Zetasizer nano ZS90 (Malvern, United Kingdom) was used to measure the $\zeta$-potential of the functionalized nanoparticles. Functionalized nanoparticle samples were put into a Greiner 96 flat bottom transparent well plate and scans were recorded between 400 and 900 nm. A Tecan ultraviolet–visible (UV/VIS) spectrometer was used to measure the extinction spectra of the silver nanoparticles. Transmission electron microscope (TEM; JEOL JEM-2010 TEM) images of the 100-nm silver nanocubes were collected to show the nanoparticles’ size and morphology. A nanosight particle tracking system (Nanosight, LM10, Salisbury, United Kingdom) was used to characterize the silver nanoparticles’ size distribution. SERS spectra were recorded using a Thermo Scientific, DXR Raman microscope (Waltham, Massachusetts) to show the enhancement capabilities of the silver nanoparticles and find the limit of detection in the microwells. The functionalized nanoparticles were excited using a 24 mW, $780\pm 0.2\mathrm{nm}$, diode laser through a $10\times$ Olympus microscope objective configured with an $1800\text{grooves}/\mathrm{mm}$ grating. The 780-nm laser was chosen with the future biosensor in mind. Blood, a typical medium for biosensing, has a strong background fluorescence in the VIS region confounding the SERS signal. Also, due primarily to water, the absorption of blood is very high beyond the mid-infrared wavelength region. Therefore, it is useful to look outside these wavelength regions in the near-infrared wavelength band in which the background signal relative to SERS is low. The nanoparticles’ integration time was 5 s and they were exposed 30 times. The reported Raman/SERS signal is the average of all 30 scans. To verify the amide bond between citrulline and PEG, Fourier transform infrared spectrometry (FTIR) (Moore Analytical, Houston, Texas) was used. Microscale thermophoresis (MST) data were collected from two bind molecular interactions. Each test was repeated in triplicate and read twice on a Monolith NT.115 (Pico) MST instrument (NanoTemper Technologies GmbH, Germany) via thermophoresis + temperature jump at 80% laser power, 10% LED power at 25°C.

2.3.

Nanoparticle Synthesis

The 100-nm edge length silver nanocubes were obtained from Dr. Hung-Jen Wu’s group within chemical engineering at Texas A&M University. Their synthesis process is described in detail in Ref. 56. Briefly, pentanediol was heated in an oil bath with a magnetic stir bar. After reaching $\sim 130°\mathrm{C}$, a silver nitrate solution followed by a polyvinylpyrrolidone (PVP) solution was repeatedly injected into the pentanediol solution until the reactant was consumed. To finish the synthesis, the nanocubes were washed with ethanol and concentrated to the desired final concentration.

2.4.

Nanoparticle Functionalization

Functionalizing the nanoparticles and removing the capping agent PVP were based on a protocol of Moran and Xia.⁵⁷ To functionalize the nanoparticles, the capping agent PVP was removed by centrifuging the nanocubes for 30 min in water. Following centrifugation, the supernatant was removed, and the nanocubes were suspended in a 1:1 ratio of ethanol and $22.3\mu \mathrm{M}$ MBA. The pH of the sample was then adjusted and the nanocubes were allowed to shake 1 h on a shake plate. The sample was then centrifuged 30 min, and the supernatant was removed and replaced with a 1:1 ratio of deionized water (DI) water and $22.3\mu \mathrm{M}$ MBA. Ethanol was initially added to help dissolve the PVP capping agent, but ethanol has been shown to weaken thiol bonds; therefore, the second wash step was performed without ethanol.⁵⁸ The sample was then placed on a shake plate and left for 24 h. Near the end of the 24-h period, a sample of citrulline bonded to PEG was made in another centrifuge tube. To make the citrulline PEG sample, 5 mg of PEG was first dissolved in 1 ml of 50 mM HEPES resulting in a 5-mM PEG HEPES solution. Citrulline was then gently dissolved in HEPES making a 250 mM final concentration of citrulline HEPES solution. Citrulline was then added dropwise to PEG at a 5:1 ratio at pH 7.3. Citrulline forms an amide bond to PEG by displacing the NHS group on PEG. The reaction occurred over 2 h on a shake plate. Following the conjugation of citrulline to PEG, the citrulline PEG solution was slowly added dropwise at a 1:1 ratio to the nanocubes with MBA solution on a stir plate. The solution was then allowed to sit 24 h on a shake plate. Last, the sample was carefully washed one last time via centrifugation at slow speeds, 6000 rpm, to remove any excess MBA, PEG, or citrulline not bonded to the surface. The process is depicted schematically in Fig. 1. When making the protocol specific for the silver nanocubes, there are a few important parameters to consider. One of the most important parameters was having the correct pH for the various reactions to occur. Regarding the thiol group on MBA and PEG, hydrogen was displaced from sulfur and binds to the silver surface in basic solutions above pH 7. When MBA was dissolved in DI, the pH was $\sim 3.4$ and was raised using NaOH. NHS, the functional group on the heterobifunctional PEG, begins to hydrolyze after 4 h at pH 7. Hydrolysis occurs within 10 min at pH 8.6, meaning the citrulline would not be able to bind to the PEG after this process. Choosing the correct pH and performing the reaction before the occurrence of hydrolyzation is vital.

Fig. 1

3-D diagram of process to modify the surface of Ag nanocubes with MBA and citrulline.

2.5.

Citrulline Aptamer Discovery

A key advantage of in vitro aptamer selection is that selection and counter-selection steps can be easily alternated to obtain highly specific affinity agents. In order to select DNA aptamers to citrulline, a modified version of “structure-switching” SELEX described by Martin et al.,⁵⁹ was utilized. Briefly, structure-switching SELEX involves immobilization of the nucleic acid library rather than the target. The DNA library is immobilized by hybridization to a short capture oligonucleotide, and upon exposure to the target (or off-targets), candidate aptamers are eluted due to conformational change and higher affinity to the target over the capture probe. This approach is especially advantageous for selection of aptamers to small molecule targets as it does not require the chemical immobilization of the target, which can perturb the structure. Fourteen rounds of SELEX were performed. In even-numbered rounds, citrulline was used to elute the library captured on magnetic beads. In odd-numbered rounds, the library was first exposed to a cocktail of the 20 essential amino acids (Promega, catalog #L4461) followed by citrulline elution in phosphate-buffered saline (PBS) buffer, pH 7.4, $1\mathrm{mM}{\mathrm{MgCl}}_2$, 0.05% Tween 20, with 5% human serum albumin added. Stringency was increased by increasing the concentration of the 20 essential amino acids as well as decreasing the elution concentration of the citrulline in successive rounds. After the final round, the eluted library was sequenced using an Ion Torrent Personal Genome Machine^™ (Life Technologies). Approximately 100,000 sequences were then ranked by bioinformatic analysis for frequency, subsequence motifs, and secondary structure. By this process, individual aptamer candidates were then chosen for synthesis and further testing.

3.1.

Ag Nanocube Functionalization and Characterization

Ag nanocubes were initially analyzed with a zetasizer to quantify the surface charge ($\zeta$-charge) before and after functionalization. About $10\mu \mathrm{l}$ of Ag nanocubes were diluted in $990\mu \mathrm{l}$ of HEPES buffer to machine specifications and $10\mu \mathrm{l}$ of Ag functionalized nanocubes were also diluted in the same manner as the nanocubes. The $\zeta$-charge of the nanoparticles decreased from $-24.2\mathrm{mV}$ before functionalization to $-27.9\mathrm{mV}$. The surface charge is similar due to the capping agents, initially PVP before functionalization and then PEG after functionalization, both of which prevent aggregation. UV/VIS spectrophotometry was used to verify the extinction spectra from the Ag nanocube nanoparticles and these nanocube extinction spectra are tuned primarily by the edge curvature and particles size, which is why the experiments were performed using 100 nm size nanocubes. As depicted in Fig. 2(a), the functionalized nanoparticles extinction spectra were at the edge of the resonance with the 780-nm excitation wavelength. It should be noted that while it is desirable to align the extinction spectra maximum with the excitation wavelength, enhancement can still be depicted as long as there is overlap and, in addition, the literature has stated that localized resonances at different locations on the nanoparticles due to dimers, etc., are actually more effective for creating SERS enhancements than the excitation wavelength relative to location of extinction spectra maximum peak.⁶⁰ After functionalizing the nanocubes, no major shift was noticed, indicating no major aggregation occurred during the functionalizing process that would have shifted the extinction spectra [Fig. 2(a)]. TEM images shown in Fig. 2(b) of the silver nanocubes were used to show their size, 100 nm and morphology, both of which are important for understanding the overall SERS intensity.⁶¹ After proving that the extinction spectra of the functionalized nanoparticles were in resonance with the excitation wavelength and did not shift after functionalization using the UV/VIS spectrometer, multiple devices were used to further characterize the surface chemistry.

Fig. 2

(a) UV/VIS spectrum of Ag nanocubes (black line) and functionalized Ag nanocubes (red line). (b) TEM image of Ag nanocubes.

FTIR spectroscopy (Moore Analytical, Houston, Texas) was used to verify that the amide bond was formed between citrulline and PEG. To insure that the NHS group on heterobifunctional PEG did not prevent the amine group on citrulline from bonding to the PEG, an important step was to confirm the citrulline did bond to PEG. A peak analysis was performed of the FTIR spectra in Fig. 3(a), which revealed that an amide bond did form, which was depicted in the spectrum by the peak between 1630 and $1670{\mathrm{cm}}^{-1}$.

Fig. 3

(a) FTIR spectrum of functionalized Ag nanocubes, highlighting the amide bond formed between PEG and citrulline. (b) Raman spectrum of MBA powder (top plot) and SERS spectrum of $22.3\mu \mathrm{M}$ MBA bonded to Ag nanocubes (middle plot). Raman spectrum of $22.3\mu \mathrm{M}$ MBA dissolved in PBS showing no detectable signal at this concentration as expected (bottom plot).

3.2.

Raman and Surface-Enhanced Raman Spectroscopy Analysis

Knowing that PEG and citrulline are present on the functionalized nanoparticles, Raman and SERS scans were performed to verify the Raman reporter MBA was present on the nanoparticles. As depicted in Fig. 3(b), to understand the SERS spectra, first, a Raman scan of MBA powder was taken. A peak analysis was done to show where the vibrational modes are for the two main Raman and SERS MBA peaks, 1099/1596 and $1075/1586{\mathrm{cm}}^{-1}$, respectively. The shift from 1099 to $1075{\mathrm{cm}}^{-1}$ and 1596 to $1586{\mathrm{cm}}^{-1}$ from Raman to SERS corresponds to MBA bonding to the metal surface.⁶² For the $\sim 1075{\mathrm{cm}}^{-1}$ peak, the vibrational modes correspond to the $\mathrm{C}─\mathrm{H}$ in-plane bending. The $\sim 1586{\mathrm{cm}}^{-1}$ peak corresponds to the $\mathrm{C}─\mathrm{C}$ stretching. Using the Raman spectra of MBA powder to know where the Raman active peaks of MBA are located, a SERS scan of the functionalized nanoparticles with $22.3\mu \mathrm{M}$ of MBA was taken to verify MBA did bond to the surface of the nanoparticles. A Raman scan of $22.3\mu \mathrm{M}$ of MBA dissolved in PBS was then taken to see if that concentration of MBA gives off a detectable Raman signal. Figure 3(b) shows no detectable signal from $22.3\mu \mathrm{M}$ of MBA dissolved in PBS, which verifies that the MBA peaks detectable when bonded to the Ag nanocubes can come only from the Ag surface enhancement. Also, the enhancement factor of nanoparticles was calculated to be ${10}^5$. Raman spectra of 100-mM MBA dissolved in ethanol were used to compare with the intensity from $22.3\mu \mathrm{M}$ of MBA adsorbed to the surfaces of the nanoparticles, to calculate the enhancement factor. The enhancement factor (EF) for the nanoparticles was calculated using Eq. (1)⁶⁰ to be

After characterization of the functionalized nanoparticles, a serial dilution was performed of the functionalized nanoparticles to find the limit of detection. The functionalized nanoparticles were first pipetted into the microwell and an initial SERS scan was performed. Following the initial scan, the nanoparticles were diluted by mixing them with PBS at a 1:1 ratio, the solution was then pipetted in the microwell and a SERS scan was performed. The surface modified nanoparticles were found to have a 24.50 pM limit of detection, as depicted in Fig. 4(a), which was defined as where the 1075 and $1586{\mathrm{cm}}^{-1}$ MBA peaks were no longer easily discernable after multiple dilution steps.

Fig. 4

(a) Serial dilution of functionalized nanocubes. (b) MST binding curve for citrulline aptamers. T1, T2, and T3 represent each test run on the MST. The 20 amino acids were separated into three different samples EA1, EA2, EA3 are tested to see how they bind to the citrulline aptamer.

3.3.

Microscale Thermophoresis Binding Studies

After finding the limit of detection for the nanoparticles, the aptamers were tested with a MST instrument. To perform the MST experiments, a Cy5-labeled version of the natural DNA citrulline aptamer ($100\mu \mathrm{M}$ stock in $1\times$ PBS, $1\mathrm{mM}{\mathrm{MgCl}}_2$ was folded by incubating at 95°C for 5 min, and then allowed to slowly cool to room temperature ($\sim 25°\mathrm{C}$). The aptamer was diluted in a solution of $1\times$ PBS, $1\mathrm{mM}{\mathrm{MgCl}}_2$, and 0.05% Tween 20. This was then added to an equal volume of a twofold dilution series of citrulline in a solution of $1\times$ PBS, $1\mathrm{mM}{\mathrm{MgCl}}_2$, and 0.05% Tween 20. The final concentration of Cy5-citrulline aptamer was 5 nM. The final concentration of citrulline ranged from $50\mu \mathrm{M}$ down to 1.526 nM, as depicted in Fig. 4(b). For the tests against the cocktail of 20 essential amino acids (EA), EA1, EA2, and EA3 are technical replicates of the same solution of 5 nM Cy5-labeled aptamer in a mixture of all 20 essential amino acids (arginine, lysine, aspartic acid, glutamic acid, glutamine, asparagine, histidine, serine, threonine, tyrosine, cysteine, methionine, tryptophan, alanine, isoleucine, leucine, phenylalanine, valine, proline, and glycine). There is a dilution series of the mixture in $1\times$ PBS pH 7.4, $1\mathrm{mM}{\mathrm{MgCl}}_2$, 0.05% Tween 20, such that at the highest concentration, the amino acids are each $500\mu \mathrm{M}$, with a twofold serial dilution down from there, to a final concentration of 15.26 nM. The aptamer was held constant at 5 nM. Last, tests (T)—T1, T2, and T3—are technical replicates of the amino acid citrulline with 5 nM Cy5-labeled aptamer. The highest final concentration of citrulline is $50\mu \mathrm{M}$, with a twofold serial dilution down from there in $1\times$ PBS pH 7.4, $1\mathrm{mM}{\mathrm{MgCl}}_2$, 0.05% Tween 20, to a final concentration of 1.526-nM citrulline. The aptamer was held constant at 5 nM. The dissociation constant, $K_{\mathrm{D}}$, for the citrulline aptamer binding to citrulline was calculated to be $4.96\mu \mathrm{M}$ [the vertical dashed line in Fig. 4(b)], with a 95% confidence interval of 3.325 to $7.378\mu \mathrm{M}$. To find the $K_{\mathrm{D}}$, three technical replicates were collected for each concentration. Technical replicates 2 and 3 were baseline corrected toward the technical replicate one. This was done by computing the average fluorescence for each replicate and then by subtracting that value from replicate 1’s average fluorescence. Based on the MST-based affinity data, the $K_{\mathrm{D}}$ for citrulline aptamer to citrulline was determined to be $\sim 5\mu \mathrm{M}$. The MST affinity data were also used for the citrulline aptamer to the 20 amino acids. The aptamer does appear to have some minor affinity for the other 20 amino acids, but the affinity is weak, $\sim 250\mu \mathrm{M}$. Thus, the aptamer is at least 50-fold more selective for citrulline than the other 20 amino acids.