2.1.

Materials

The aptamer (H8DAE) was synthesized, characterized using MST, and purchased from Base Pair Biotechnologies (Pearland, Texas). The NC membrane (pore size of $0.45\mu \mathrm{m}$) and blotting absorbent filter paper (# 1703960) were purchased from BioRad (Hercules, California), and the pure cellulose chromatography paper (Cat. No. 05714-1) was acquired from Fisher Scientific (Waltham, Massachusetts). For the synthesis of the citrate-capped AuNPs, gold (III) chloride (${\mathrm{HAuCl}}_4$) and sodium citrate were obtained from Sigma-Aldrich (St. Louis, Missouri). Before conjugation of the thiolated aptamer, the disulfide bonds were reduced using tris(2-carboxyethyl)phosphine hydrochloride (TCEP) purchased from Sigma-Aldrich (St. Louis, Missouri). The immobilization of the AuNPs on the nitrocellulose membrane was performed using streptavidin, which was purchased from VWR (Radnor, Pennsylvania) and a PEGylated-biotin linker was obtained from ThermoFisher Scientific (Waltham, Massachusetts) and modified accordingly. All experiments were performed in $1\times$ phosphate-buffered saline ($1\times$ PBS) at pH 7.4 and Raman measurements were taken using an ID Raman mini 2.0 (Metrohm—formally purchased from Ocean Optics, Switzerland).

2.2.

Aptamer-Conjugated Gold

The 34-nm citric acid-capped AuNPs were prepared using a well-known seed-mediated method adopted from Bastús et.al.¹⁹ Briefly, a solution of gold seed was prepared by pipetting $335\mu \mathrm{L}$ of 25 mM ${\mathrm{HAuCl}}_4$ solution into 50 ml of sodium citrate solution (2.2 mM) being heated at 100°C. After $\sim 15\min$ of reaction, the solution was cooled to 90°C for kinetically controlled shell growth. To grow the nanoshell further to the desired size, $335\mu \mathrm{L}$ of 60 mM sodium citrate solution was mixed with the gold solution (90°C) for 2 min followed by an additional pipetting of 25 mM ${\mathrm{HAuCl}}_4$ solution ($335\mu \mathrm{l}$) to the solution. Afterward, the reaction was allowed to occur for 30 min. The size of the particles was confirmed via UV–Vis spectroscopy and the table provided by Haiss et al.²⁰ Additional cycles of $335\mu \mathrm{L}$ of 60 mM sodium citrate solution followed by 25 mM ${\mathrm{HAuCl}}_4$ solution ($335\mu \mathrm{l}$) after 2 min were repeated until the desired size was obtained.

Following AuNP synthesis, the thiolated, fluorescently labeled aptamer was conjugated on its surface in a $9200:1$ molar ratio. Briefly, the aptamers’ disulfide bonds were reduced to generate functional thiol groups by incubating the aptamers ($100\mu \mathrm{M}$) in equal volume of 20 mM TCEP solution for an hour. Following this incubation, the TCEP was removed via 3 kDa Nanosep centrifuge filters and resuspended in $1\times$ PBS. The final concentration of the aptamers was determined via UV–Vis spectroscopy. The aptamers were then folded into their tertiary state by heating the solution to 93°C for 5 min and allowing this solution to slowly cool to room temperature. After folding, the aptamers were added to the gold colloidal solution at a $9200:1$ ratio. This was allowed to shake for an hour and then incubated overnight at room temperature to ensure proper time was given for the thiolated aptamers to adhere to the nanoparticle surface. Following this reaction, 4.5 mM citrate buffer (pH 3) was added to the solution and allowed to shake for 10 min before centrifugation. Thiol-pegylated biotin ($200\mu \mathrm{M}$) was added at $1000:1$ molar ratio and allowed to shake for 10 min before centrifuging (1.2 krcf) for 10 min to remove any unconjugated chemistry. The newly conjugated particles were resuspended to a final AuNP probe concentration of 0.5 nM in distilled water for later characterization and use.

2.3.

Vertical Flow Paper Fluidics

After determining the functionality of the sensor via fluorescence in free solution, a paper-fluidic platform was developed for POC testing of the citrulline biomarker. The well-established vertical flow paper (VFP) fluidics platform was adopted for this design. Specifically, the paper-fluidic device was designed as shown in Fig. 1 where three different paper membranes of varying pore sizes were used to (1) immobilize the probes, (2) wick the fluid as it flows vertically through to the final paper used to (3) absorb the waste solution (in a blotting absorbent pad). First, NC paper of 450-nm pore size was chosen for probe immobilization. The membrane was cut into a 5-mm-diameter circle using a biopsy punch. Streptavidin ($2\mathrm{m}\mathrm{g}/\mathrm{ml}$ in DI ${\mathrm{H}}_2\mathrm{O}$) was then incubated on its surface by pipetting $2\mu \mathrm{l}$ of this solution to the membrane, immediately followed by $1\mu \mathrm{l}$ of PBS and $5\mu \mathrm{l}$ of the biotinylated AuNPs probes solution. Enough time was given for the membrane to dry $(\sim 1\mathrm{h})$. Next, a hydrophobic barrier was created on the wicking paper (i.e., pure cellulose membrane) by curing polydimethylsiloxane in the designated dark blue areas seen in Fig. 1. The three papers were then assembled together as depicted in Fig. 1 using 3M double-sided adhesive tape (#444).

Fig. 1

Schematic of citrulline SERS bioassay.

2.4.

SERS Response to Citrulline

The SERS response of the assay to varying concentrations of citrulline was observed using a portable handheld Raman spectrometer at 638-nm excitation and 40-mW power. First, the SERS signal of the membrane in its dry state was measured as baseline. Then $10\mu \mathrm{l}$ of PBS was added to the membrane at the sample loading zone and Raman measurements were taken after 20 min. Then $10\mu \mathrm{l}$ of $10\mu \mathrm{M}$ of citrulline was added to the same zone and after equal amount of time another Raman spectrum was taken. This was repeated for 0, 10, 20, 40, and $50\mu \mathrm{M}$ concentrations. This experiment was performed in triplicate.

To determine the feasibility of these chips to be used as one-time disposable chips, the different concentrations of citrulline were tested on separate chips. Again, the SERS signal of the membrane in its dry state was measured as baseline. Then $10\mu \mathrm{l}$ of PBS (i.e., $0\mu \mathrm{M}$ citrulline) was added to the membrane at the sample loading zone and Raman measurements were taken after 20 min. Then $10\mu \mathrm{l}$ of $10\mu \mathrm{M}$ of citrulline was added to the same zone and after equal amount of time another Raman spectrum was taken. This was repeated for citrulline concentrations of 20, 40, and $50\mu \mathrm{M}$ all on separate chips. This experiment was repeated in triplicates for citrulline concentrations of 0, 10, 20, 40, and $50\mu \mathrm{M}$, where the $0\mu \mathrm{M}$ was used as the baseline normalization factor for each chip.

2.5.

Selectivity of the SERS-Based Paper-Fluidic Assay

The selectivity of the SERS-based paper-fluidic assay was investigated by comparing the percentage of the SERS intensity signal in the presence of $200\mu \mathrm{M}$ citrulline and three noncomplementary amino acids: valine (Val), histidine (His), and glycine (Gly), all at $200\mu \mathrm{M}$ concentration. Like before, the paper-fluidic chip was interrogated using the Ocean Optics portable handheld Raman spectrometer at 638-nm excitation and 40-mW power. First, the SERS signal of each membrane in its dry state was measured as baseline. Then $10\mu \mathrm{l}$ of PBS was added to the membrane at the sample loading zone and Raman measurements were taken after 20 min. Then $10\mu \mathrm{l}$ of the specific amino acid ($200\mu \mathrm{M}$) was added to the same zone and after an equal amount of time another Raman spectrum was taken. This was repeated for two more times for each amino acid.

3.1.

Characterization of Aptamer-Conjugated Gold

The Cy3-labeled aptamer was conjugated on the surface of AuNPs to cause a distance-dependent change in the signal as depicted in Fig. 1. The size of the gold was measured via absorbance and confirmed by TEM images to be $\sim 34\mathrm{nm}$ in diameter (Fig. 2). The conjugation of these particles was verified via absorbance measurements and resulted in a 2-nm redshift change between bare citrate-capped AuNPs and the aptamer-conjugated AuNPs (Fig. S1 in the Supplementary Material). This shift is a result in the change of the particles’ dielectric constant due to the presence of the thiolated aptamer covalently bound to the particles’ surface, thus resulting in a change in localized surface plasmon resonance and confirming successful conjugation. This conjugation was further confirmed by the change in zeta potential of the particle surface of citrate-capped AuNPs ($-42+0.9\mathrm{mV}$) and aptamer-AuNPs ($-36+0.67\mathrm{mV}$). Furthermore, the obtained surface charge of less than $-25\mathrm{mV}$ indicates the stability of both particles.²¹

Fig. 2

TEM images of bare citrate-capped AuNPs ($\sim 34\mathrm{nm}$).

3.2.

SERS in Response to Citrulline Concentration

To create a portable platform to collect and concentrate the citrulline analyte and to amplify the SERS signal, a VFP fluidic chip was created (Fig. 3).

Fig. 3

Schematic of VFP chips immobilized with aptamer-conjugated SERS-active probes for citrulline detection.

Following the immobilization of the probes on the nitrocellulose membrane and the assembly of the VFP chip, the SERS signal was measured for varying concentrations of citrulline. First, the chips were tested by varying the concentration of citrulline from 0 to $50\mu \mathrm{M}$ [Fig. 4(a)] within the same chip to determine the conformational change and saturation limit of the assay. Figure 4(a) shows the SERS response decreasing with increasing concentration of the analyte, which indicated the movement of the dye away from the nanoparticle surface (i.e., outside of the electromagnetic field) as citrulline binds to the aptamer and the aptamer changes its conformation in response to this binding. The bands near 1197, 1390, 1468, and $1585{\mathrm{cm}}^{-1}$ are indicative of the bands for Cy3 and not of the Raman bands of the DNA aptamer or citrulline.^22,23 The limit of detection (LOD) for the VFP chip was determined to be $0.56\mu \mathrm{M}$ (Fig. S2 in the Supplementary Material). This detection limit is well below the physiological relevant concentrations needed to determine disease state of the gut. Furthermore, the small error bars indicate the ability to be able to reproduce this response across different chips ($n=3$) with a significant statistical difference among concentrations $\le 40\mu \mathrm{M}$ ($p=0.00015$), thus indicating the potential use of this paper-based assay for the detection of low levels of citrulline relating to diseases such as villous atrophy and short bowel syndrome.

Fig. 4

(a) SERS response on the VFP chip in the presence of varying concentrations of citrulline within the same chip. (b) Calibration curve for the average of three different chips at $1389{\mathrm{cm}}^{-1}$ peak.

The chips were tested for the ability to be used as a disposable, one-time use lab on a chip platform. To accomplish this, different chips were loaded with different concentrations of the target with different citrulline concentrations (0, 10, 20, 40, and $50\mu \mathrm{M}$). Due to fluctuating baseline from chip to chip, all chips were first measured with $0\mu \mathrm{M}$ solution (i.e., PBS) before the desired concentration was added and final measurements were taken at that concentration. The data for each chip were then normalized to its zero concentration to account for this variation. Similar to what was observed in Fig. 4, the SERS response indicated a decrease in signal intensity with higher concentrations of the citrulline target (Fig. 5).

Fig. 5

SERS response on the VFP chip in the presence of different concentrations of citrulline in PBS across different chips (all data normalized to its $0\mu \mathrm{M}$ concentration).

The LOD was calculated as 9.5 and $7.8\mu \mathrm{M}$ for 1386 and $1585{\mathrm{cm}}^{-1}$, respectively (Fig. S3 in the Supplementary Material). This LOD is again well below the physiological relevant concentration of $40\mu \mathrm{M}$ indicating that the VFP assay’s ability to detect low concentrations of citrulline necessary to make a diagnosis of disease of the gut. The upper limit of quantification, defined as the maximum detection that would result in a true positive, was calculated to be 79.6 and $58.8\mu \mathrm{M}$ for 1386 and $1585{\mathrm{cm}}^{-1}$, respectively. A calibration of three different chips exposed to the same concentration of citrulline (Fig. 6) shows the reproducibility of the assay (denoted by the error bars). To assess the precision of the chips, the mean absolute difference was calculated to be 3.3% and 2.3% for 1386 and $1585{\mathrm{cm}}^{-1}$, respectively. According to literature, citrulline concentrations $<30\mu \mathrm{M}$ indicate a certain level of disease state of the gut. This assay is able to detect changes in levels around that concentration; however, the performance can be improved by changing the detection platform. One limitation of paper fluidics is the inability to control the rate at which it dries, which in turn may cause baseline drift from chip to chip. One potential way to overcome this issue is immobilizing this assay in a microfluidic device or using a combination vertical and lateral flow paper with an optically transparent cover over the probe area of the paper, which can create a stable fluid environment for the assay and allow for optimum observation of the conformational change that the aptamer undergoes in the presence of citrulline.

Fig. 6

Calibration curve at 1386 and $1585{\mathrm{cm}}^{-1}$ peak for the average of three sets of disposable chips incubated with different concentrations of citrulline.

3.3.

Selectivity of the SERS in Response to Citrulline

To determine the aptamer selectivity to citrulline, three other nonspecific amino acids, histidine (His), valine (Val), and glycine (Gly), were tested on the paper-fluidic device at high concentrations ($200\mu \mathrm{M}$). This concentration was selected based on previously reported average blood plasma concentrations of these amino acids: His ($83\mu \mathrm{M}$), Val ($237\mu \mathrm{M}$), and Gly ($224\mu \mathrm{M}$).⁴ Figure 7 shows the change in the normalized SERS intensity at $1585{\mathrm{cm}}^{-1}$ of the assay in the presence of $200\mu \mathrm{M}$ ($I_{200}$). The data were normalized to the assay initial exposure to only PBS ($I_0$) prior to the addition of the amino acid. The percentage decrease for each amino acid was calculated as ($I_{200}/I_0\times 100$) and results indicated a percentage of change of 15%, 2%, 3%, and 75% for His, Val, Gly, and Cit, respectively. Although His showed a 15% change for the assay exposed to $200\mu \mathrm{M}$, it is typically found in blood plasma around $83\mu \mathrm{M}$; therefore, this percentage change is not likely to be evident and thus affecting the sensing of citrulline.

Fig. 7

Normalized SERS intensity at $1585{\mathrm{cm}}^{-1}$ of histidine (His), valine (Val), and glycine (Gly) compared to citrulline (Cit) at high concentrations demonstrating VFP assay selectivity.