Quantitative and Label-Free Detection of Protein Kinase A Activity

Apr 20, 2018 - Laurent Bekale,. †. Malvin Wei Cherng Kang,. †. Susana Soo-Yeon Kim,. §. Ivan Tan,. §. Kong-Peng Lam,. § and James Chen Yong Kah...
0 downloads 0 Views 3MB Size
Download PDF
Article Cite This: Anal. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/ac

Quantitative and Label-Free Detection of Protein Kinase A Activity Based on Surface-Enhanced Raman Spectroscopy with Gold Nanostars Shuai He,† Yi Mon Ei Kyaw,† Eddie Khay Ming Tan,‡ Laurent Bekale,† Malvin Wei Cherng Kang,† Susana Soo-Yeon Kim,§ Ivan Tan,§ Kong-Peng Lam,§ and James Chen Yong Kah*,†,∥ †

Department of Biomedical Engineering, National University of Singapore, Singapore 117583 TechnoSpex Private Limited, Singapore 169203 § Bioprocessing Technology Institute, Agency for Science, Technology and Research, Singapore 138668 ∥ NUS Graduate School for Integrative Sciences and Engineering, National University of Singapore, Singapore 117456 ‡

S Supporting Information *

ABSTRACT: The activity of extracellular protein kinase A (PKA) is known to be a biomarker for cancer. However, conventional PKA assays based on colorimetric, radioactive, and fluorometric techniques suffer from intensive labeling-related preparations, background interference, photobleaching, and safety concerns. While surface-enhanced Raman spectroscopy (SERS)-based assays have been developed for various enzymes to address these limitations, their use in probing PKA activity is limited due to subtle changes in the Raman spectrum with phosphorylation. Here, we developed a robust colloidal SERS-based scheme for label-free quantitative measurement of PKA activity using gold nanostars (AuNS) as a SERS substrate functionalized with bovine serum albumin (BSA)−kemptide (Kem) bioconjugate (AuNS−BSA−Kem), where BSA conferred colloidal stability and Kem is a high-affinity peptide substrate for PKA. By performing principle component analysis (PCA) on the SERS spectrum, we identified two Raman peaks at 725 and 1395 cm−1, whose ratiometric intensity change provided a quantitative measure of Kem phosphorylation by PKA in vitro and allowed us to distinguish MDA-MB-231 and MCF-7 breast cancer cells known to overexpress extracellular PKA catalytic subunits from noncancerous human umbilical vein endothelial cells (HUVEC) based on their PKA activity in cell culture supernatant. The outcome demonstrated potential application of AuNS−BSA−Kem as a SERS probe for cancer screening based on PKA activity.

P

PKA activity is closely related to the development of cancer2,3 and has been shown to be a potential prognostic marker and predictor of clinical outcome.4−6 Extracellular secretion of PKA catalytic subunits by various tumor cells could also be detected in the serum of cancer patients.7−9 Reporter systems based on radioactive, colorimetric, and fluorometric techniques are conventionally used to probe PK activity.10 While radioactive assay raises safety and environmental concerns, colorimetric and fluorometric assays coupled

hosphorylation of proteins regulates almost all aspects of cell life, and abnormal phosphorylation is both a cause and consequence of diseases.1 Phosphorylation by cAMP-dependent protein kinase A (PKA) occurs through the transfer of a phosphate group from adenosine triphosphate (ATP) to the serine and threonine residues of substrate proteins or peptides. PKA comprises a complex of two regulatory subunits which contain a cAMP binding site and two catalytic subunits which are responsible for phosphorylation upon activation. The binding of cAMP to regulatory subunits changes their conformation and dissociates them from the catalytic subunits. The released catalytic subunits are then activated to phosphorylate specific targets. © XXXX American Chemical Society

Received: December 26, 2017 Accepted: April 20, 2018

A

DOI: 10.1021/acs.analchem.7b05417 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

cancer from noncancer cells based on their different levels of extracellular PKA activity probed in complex culture supernatant, demonstrating potential of this probe for cancer screening in biological fluids.

with enzyme-linked immunosorbent assay (ELISA) require intensive labeling-related preparations and the readout often suffers from background interference and photobleaching.11,12 Recent developments in PK assays include the use of gold nanoparticles (AuNPs) that exploits their unique optical properties based on surface plasmon resonance (SPR) to elicit an absorbance spectral change due to phosphorylation-induced aggregation of AuNPs.13−15 While sample preparation is straightforward, such assays suffer from absorbance interference by other molecules in the sample. The specific molecular fingerprint from surface-enhanced Raman spectroscopy (SERS) could overcome interferencerelated issues associated with absorbance measurement. SERS is specific, sensitive, rapid, and nondestructive and requires minimal sample preparation. SERS-based assays have been used to probe the activity of various enzymes including phosphatase, peroxidase, lipase, and protease.16−18 However, SERS-based PKA assays are limited due to weak changes in Raman spectra upon serine/threonine phosphorylation by PKA.19 We addressed this gap by developing a label-free SERS probe comprising a PKA-specific recognition peptide substrate (LRRASLG) called kemptide (Kem)20,21 conjugated to gold nanostars (AuNS) through a bovine serum albumin (BSA) spacer which maintained colloidal stability of the AuNS−BSA− Kem probe. AuNS demonstrated stronger SERS enhancement than spherical AuNPs due to their multiple sharp spikes serving as intrinsic SERS “hot spots”.22−24 We detected kemptide phosphorylation by PKA in vitro using AuNS−BSA−Kem by monitoring changes in intensity of two Raman markers at 725 and 1395 cm−1 identified from principal component analysis (PCA) (Figure 1). The probe achieved a detection limit of 5 mU/mL, which was about 3 orders lower than commercial immunoassays.25 With intensity ratio I725/I1395 defined as a PKA activity index, we distinguished

■

MATERIALS AND METHODS All reagents were purchased from Sigma-Aldrich unless otherwise specified. Milli-Q water with a resistivity of 18.2 MΩ cm was used for all experiments. Synthesis and Characterization of AuNS. AuNS were synthesized using a modified one-pot seedless protocol.23 Briefly, 360 μL of 10 mM gold(III) chloride (HAuCl4) and 20 μL of 10 mM silver nitrate (AgNO3) were mixed in 10 mL of H2O under vortex for 30 s in a 15 mL centrifuge tube, before adding 60 μL of 100 mM L-ascorbic acid (C6H8O6) as the reducing agent and vortexing for another 20 s. The solution turned from faint yellow to greenish blue almost immediately. The synthesized AuNS colloid was centrifuged at 300 rcf for 1 h, resuspended in 2 mL of H2O, and then stored at 4 °C until further use. The AuNS were characterized for their optical properties using UV−vis absorption spectroscopy (MultiSkan GO, Thermo Fisher Scientific Inc., U.S.A.), and their hydrodynamic diameter (DH) and ζ-potential (ζ) were characterized using a Zetasizer (Nano ZS, Malvern, U.K.). The morphology of AuNS was characterized using transmission electron microscopy (TEM) (JEM-1220, JEOL Ltd., Japan). The final AuNS concentration was estimated by mass calculation to be ≈50 pM, assuming Au3+ ions were completely reduced to Au0 in AuNS, and the concentration was determined by dividing the total mass of available Au by the mass of one AuNS, whose size is measured from the DH and density is given by 19.3 g/cm3. Bioconjugation of Kem to BSA. Kem powder with 95% purity [GenScript (Hong Kong) Limited] was conjugated to BSA by carbodiimide cross-linker chemistry.26 Briefly, BSA was dissolved in phosphate buffer at pH 5 to facilitate efficient cross-linking. 1-Ethyl-3-(3-(dimethylamino)propyl)-carbodiimide hydrochloride (EDC) powder was then added to BSA at twice the molar ratio of BSA and stirred at 4 °C for 5 min to activate the carboxylic acid groups on BSA. Kem powder dissolved in phosphate buffer at pH 5 was then mixed with BSA/EDC solution at twice the molar ratio of BSA and stirred overnight at 4 °C to allow cross-linking between the carboxylic acid groups on BSA and primary amine groups on Kem. The mixture was stirred for another 4 h in pH 7 at 4 °C, followed by dialysis against phosphate buffer to remove intermediate product and excess EDC. The concentration of BSA in purified BSA−Kem was estimated from fluorescence emission of the two tryptophan residues (Ex = 280 nm, Em = 332 nm) in BSA27 and compared against a linear standard curve of fluorescence intensity against BSA concentration from 0 to 500 μg/mL. Although both BSA and kemptide contained carboxylates and amines, random polymerization of polypeptides caused by EDC-mediated cross-linking could be minimized by a Kem-toBSA molar ratio of less than 3,26 where multiple Kem could be conjugated on one BSA without polymerization. The molar concentration of kemptide in purified BSA−Kem was estimated to be twice that of BSA since the initial Kem-to-BSA ratio was 2, with the assumption that all Kem were conjugated to BSA. Conjugation of BSA−Kem on AuNS. BSA−Kem was conjugated to AuNS by nonspecific adsorption. Here, 500 μL of BSA−Kem was first titrated at varying concentrations (0, 0.04,

Figure 1. AuNS-based SERS detection scheme for PKA activity. AuNS were conjugated to bovine serum albumin (BSA)−kemptide (Kem) conjugates to form AuNS−BSA−Kem before introduction to biological samples. The change in acquired SERS spectra upon phosphorylation of Kem by PKA was processed by PCA to identify two Raman peaks at 725 and 1395 cm−1 with ratiometric intensity change as potential Raman markers of PKA activity that allowed us to differentiate cancer from noncancer cells in extracellular culture supernatant. B

DOI: 10.1021/acs.analchem.7b05417 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry 0.4, 4, 40 mg/mL) to 500 μL of 50 pM AuNS to determine the minimum concentration of BSA−Kem for conjugation to AuNS. During the 1 h incubation at room temperature under continuous shaking, the BSA−Kem formed a corona around the AuNS through nonspecific adsorption of BSA to the AuNS via its free thiol (−SH) or disulfide (−S−S−) groups,28 to form a Au−thiol coordinate bond. Excess unbound proteins were then removed by repeated centrifugation at 1700 rpm for 20 min, and the resulting AuNS−BSA−Kem was resuspended in 500 μL of 1× phosphate-buffered saline (PBS) to determine the minimum concentration of BSA−Kem needed for conjugation based on colloidal stability of AuNS−BSA−Kem as measured from UV−vis spectroscopy (MultiSkan GO, Thermo Fisher Scientific Inc., U.S.A.). The DH and ζ of AuNS−BSA−Kem were determined using a Zetasizer (Nano ZS, Malvern, U.K.). From their absorption spectra, the concentration of AuNS−BSA−Kem was estimated to be 20 pM. Phosphorylation of AuNS−BSA−Kem in Buffer. PKA catalytic subunits (New England BioLabs Inc., U.S.A.) were used to phosphorylate BSA−Kem. Here, the term PKA will subsequently be used to refer only to the catalytic subunits of PKA. AuNS−BSA−Kem was phosphorylated by adding 10 μL of 250 kU/mL PKA, 10 μL of 10 mM ATP, and 10 μL of 10 mM MgCl2 into two volumes of AuNS−BSA−Kem (70 and 470 μL) to give two final concentrations of 100 and 20 pM, respectively, followed by 37 °C incubation for 1 h. The two volumes would allow us to determine the effect of AuNS− BSA−Kem concentration on its phosphorylation efficiency. PKA(−) and ATP(−) samples in the absence of PKA and ATP, respectively, were also included as negative controls. AuNS−BSA conjugate without Kem was used as the Kem(−) negative control. Commercially available pre-phosphorylated Kem (p-Kem) powder with 95% purity [GenScript (Hong Kong) Limited] was used as our positive control. Phosphorylation was repeated with varying PKA concentrations (5 mU/ mL, 0.5 U/mL, 5 U/mL, 0.5 kU/mL, 5 kU/mL) in the same setup to produce a standard calibration curve for PKA activity. The specificity of PKA assay was examined with protein kinase C (PKC) of different isozymes, i.e., PKC-α (bovine)/ε (human)/η (human)/ζ (rat), which is also a serine/threonine kinase, by incubating them with AuNS−BSA−Kem in the presence of MgCl2 and ATP at 37 °C for 1 h. PKC-α/ε/η/ζ isozymes were prepared by cloning the genes into pXJ40-FLAG vector and overexpressed in HEK293T cells before being immunoprecipitated from cell lysates using FLAG-beads overnight and eluted with FLAG-peptide in PBS. PKC expression was verified using Western blot with anti-FLAG and anti-PKC (Santa Cruz) antibodies (Figure S1, Supporting Information). Phosphorylation of BSA−Kem in the absence of AuNS was also performed to serve as a protein marker for subsequent SDS−PAGE experiments. Briefly, 1 μL of 250 kU/mL PKA catalytic subunits, 1 μL of 10 mM ATP, and 1 μL of 10 mM MgCl2 were mixed with 7 μL of 4 mg/mL BSA−Kem and incubated at 37 °C for 30 min to allow phosphorylation. SDS−PAGE Gel Electrophoresis. Sodium dodecyl sulfate−polyacrylamide gel electrophoresis (SDS−PAGE) analysis for phosphoprotein staining and total protein identification was performed with dual-color prestained protein standard, 4−15% Tris−glycine precast gels, and 10× Tris/glycine running buffer (Bio-Rad, U.S.A.). A 5× sample loading buffer was prepared with 10% SDS, 10 mM dithiothreitol (DTT), 20% glycerol, 0.2

M Tris−HCl (pH 6.8), and 0.05% bromophenol blue. In preparing BSA−Kem for SDS−PAGE analysis, 5 μL of 5× sample loading buffer was mixed with 10 μL of BSA−Kem before heating in boiling water for 10 min. The preparation of AuNS−BSA−Kem for SDS−PAGE analysis was similar except that AuNS−BSA−Kem was centrifuged after phosphorylation and 5 μL of 5× sample loading buffer was added to 10 μL of pellet for gel loading. The SDS−PAGE electrophoresis was performed at 20 mA for 40 min. Phosphoprotein Staining. Phosphorylation of BSA−Kem before and after conjugation to AuNS was verified by Pro-Q Diamond phosphoprotein gel stain (Invitrogen, U.S.A.) after the SDS−PAGE gel electrophoresis. The SDS−PAGE gel was immersed in 100 mL of fix solution (50% methanol and 10% acetic acid) and incubated at room temperature with gentle agitation for 30 min. This fixation step was repeated once and the gel was left overnight to ensure complete removal of SDS out of the gel. The fixed gel was then washed by immersing in 100 mL of ultrapure water with gentle agitation for 10 min and repeated twice to remove all the methanol and acetic acid from the gel. The fixed gel was then incubated in 50 mL of Pro-Q Diamond phosphoprotein gel stain with gentle agitation in the dark for 1.5 h, followed by destaining (20% acetonitrile and 50 mM sodium acetate at pH 4.0) under gentle agitation for 30 min at room temperature in the dark. The stained gel was washed twice with ultrapure water at room temperature for 5 min per wash in the dark and imaged under UV light to identify fluorescent phosphoprotein bands. After imaging, the gel was stained with 50 mL of Coomassie blue total protein stain solution (0.1% Coomassie brilliant blue R-250, 50% methanol, and 10% acetic acid) for total protein staining under gentle agitation for 15 min and destained overnight with H2O. The gel was imaged under white light to quantify total protein amount using ImageJ. Cell Culture. Human breast adenocarcinoma cells (MCF-7 and MDA-MB-231) were cultured in Roswell Park Memorial Institute medium (RPMI 1640) with 10% fetal bovine serum (FBS) at 37 °C and 5% CO2. Human umbilical vein endothelial cells (HUVEC) were cultured in endothelial cell growth medium (EGM-2, Lonza, U.S.A.) at 37 °C and 5% CO2. Both MCF-7 and MDA-MB-231 cells were reported to have elevated extracellular PKA activity,7,9 while HUVECs were used as our negative healthy control with no detectable PKA activity.29 The cells were monitored for confluency under bright-field microscopy. Phosphorylation of AuNS−BSA−Kem by Extracellular Secretion of PKA. To probe for phosphorylation of AuNS− BSA−Kem by secreted PKA in the extracellular environment of cancer cells, the culture supernatant from three cell lines (MDA-MB-231, MCF-7, HUVEC) was collected when the cells reached confluency and centrifuged at 180g for 5 min to remove any remnant cells in the collected supernatant. AuNS− BSA−Kem was then introduced to 500 μL of the purified culture supernatant to a final concentration of 20 pM, and the mixture was incubated at 37 °C for 1 h for phosphorylation of AuNS−BSA−Kem by PKA present in the culture supernatant to occur. Raman Spectra Acquisition and Processing. All Raman spectra were acquired with a uRaman-785-Ci Raman spectroscopy system (Technospex Pte. Ltd., Singapore), which excited the samples using a 50 mW frequency stabilized laser at 785 nm with line width of ∼100 MHz. The system was equipped with a C

DOI: 10.1021/acs.analchem.7b05417 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 2. Characterization of synthesized AuNS, AuNS−BSA, and AuNS−BSA−Kem. (A) UV−vis absorption spectra showing a red shift of ∼20 nm in the SPR peak indicating the presence of BSA on AuNS. (B) Hydrodynamic diameter, DH, from dynamic light scattering (DLS) increased by ∼20 nm due to the binding of BSA and BSA−Kem on AuNS, while the ζ-potential (ζ) became less negative upon conjugation, especially for BSA− Kem since Kem carries a net positive charge. (C) Histogram size distribution showing collective increase in DH of AuNS after BSA and BSA−Kem conjugation. (D) TEM images of synthesized AuNS showing a typical spiky morphology.

Figure 3. Colloidal stability of AuNS−BSA−Kem prepared from different concentrations of BSA−Kem added to AuNS as determined from its UV− vis absorption spectrum (A) before and (B) after resuspension in 1× PBS. The aggregation induced by 1× PBS allowed us to determine the minimum BSA−Kem concentration of 0.4 mg/mL necessary to confer colloidal stability to 50 pM AuNS against PBS-induced aggregation of AuNS.

Principal Component Analysis (PCA). The spectra were processed by standard normal variate (SNV) normalization for principal component analysis (PCA). We consolidated the Raman spectra of phosphorylated and non-phosphorylated samples and applied the princomp function in MATLAB on the Raman spectra to obtain the loadings and scores of various principal components calculated for the data set. The first three components (PC1, PC2, and PC3) were plotted and analyzed since they accounted for the majority of the variance in the data set. After a clear separation of the phosphorylated group from

TEC-cooled 2048-pixel CCD detector which achieved a spectral resolution of ∼8.6 cm−1. The spectra range spanned between 150 and 2400 cm−1. All spectra were obtained from samples contained in a quartz cuvette under an integration time of 60 s. The raw Raman spectra were processed using MATLAB to remove the autofluorescence background signal by third-order polynomial curve fitting. The spectra were averaged from three measurements to ensure reliability. D

DOI: 10.1021/acs.analchem.7b05417 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 4. Detection of phosphorylation in BSA−Kem with and without AuNS conjugation through Pro-Q Diamond phosphoprotein stain (top). The bright fluorescent bands indicated successful phosphorylation of BSA−Kem in the presence of PKA. The brightness was comparable to a commercially available pre-phosphorylated p-Kem as a positive control, which was chemically phosphorylated before bioconjugation to AuNS. Total protein identification was completed by Coomassie blue stain (bottom).

respectively), which was typical of proteins adsorbed on the surface of gold nanoparticles.32 The surface ζ-potential became less negative (−22.7 mV for AuNS−BSA and −11.6 mV for AuNS−BSA−Kem) (Figure 2B). This was especially true for AuNS−BSA−Kem due to the net positive charge of Kem. The relatively weaker surface charge indicated that AuNS−BSA− Kem was less colloidally stable than AuNS−BSA due to less charge stabilization. The conjugation of BSA and BSA−Kem to AuNS also caused the average DH to increase from 106.37 to 123.70 nm for AuNS−BSA (ΔDH = 17.33 nm) and 125.48 nm for AuNS− BSA−Kem (ΔDH = 19.11 nm) (Figure 2B). Since the size of a BSA molecule was known to be 4 × 4 × 14 nm3 in aqueous solution,33 this was equivalent to a modeled sphere of a diameter of ∼8.5 nm, which agreed well with the observed ΔDH. Interestingly, the size histogram distribution of AuNS− BSA and AuNS−BSA−Kem shifted to larger DH collectively (Figure 2C) without significant change in the size distribution profile. This indicated that both BSA and BSA−Kem were conjugated on the AuNS homogeneously. Phosphorylation of AuNS−BSA−Kem. PKA recognizes the amino acid sequence motif of RRX(S/T)X and phosphorylates targets with this sequence surrounding the phosphorylation site. Kem had a sequence LRRASLG with kinetic constants comparable to those of native protein substrates.34 Apart from the enzyme PKA, ATP and Mg2+ were also essential for phosphorylation to occur as ATP was the source of phosphate to be transferred to Kem while Mg2+ was crucial to form the transition state of phosphoryl transfer reaction.35 The Mg2+ in the active sites also enhanced the binding affinities of ATP to PKA through electrostatic interactions.35 We verified the phosphorylation of BSA−Kem with Pro-Q Diamond phosphoprotein stain which combined a trivalent transition metal ion (Ga3+) and a fluorescent metal ion indicator36 to complex with the phosphate groups on Kem after phosphorylation by PKA and produce a bright fluorescent band on the gel. This band was observed in BSA−Kem in the presence of PKA, ATP, and MgCl2 to indicate successful phosphorylation of BSA−Kem bioconjugates in vitro [Figure 4, lane 1; without AuNS, PKA(+)], and its intensity was comparable to that of commercially available pre-phosphorylated Kem (p-Kem) at the same concentration (Figure 4, lane 3), indicating almost 100% phosphorylation of Kem by PKA based on our in vitro protocol.

the non-phosphorylated group was observed on the scores plot of a principal component, the loadings plot of that particular principal component was used to identify potential Raman markers for PKA phosphorylation.

■

RESULTS AND DISCUSSION Synthesis and Characterization of AuNS−BSA−Kem. AuNS as synthesized by a SERS-optimized one-pot protocol using ascorbic acid as the reducing agent and silver nitrate as the shaping agent showed an SPR peak at ∼680 nm, with a DH of 106.37 ± 1.44 nm and ζ-potential of −29.3 ± 2.47 mV (Figure 2, parts A and B, red). The spiky morphology of AuNS was observed under TEM (Figure 2D). In our study, BSA was used as a capping agent to maintain colloidal stability of AuNS.30 From our ζ-measurements, BSA carried a net negative charge of −19.9 ± 3.28 mV at neutral pH, which agreed well with the −18 mV reported by others.31 Upon bioconjugation to Kem, the ζ of BSA−Kem became less negative from −19.9 ± 3.28 mV to −4.38 ± 0.64 mV due to the positive charge of Kem, since the sequence of Kem, i.e., LRRASLG, was calculated to carry a net charge of +2 at neutral pH based on calculations from each amino acid residue in its sequence. This indicated successful bioconjugation of Kem to BSA. On conjugating BSA−Kem to AuNS, the BSA−Kem corona around AuNS was formed through nonspecific adsorption of BSA to the AuNS. Since each BSA molecule contained one free thiol (−SH) group and 17 disulfide (−S−S−) bonds, direct covalent conjugation through a Au−thiol bond was possible. The colloidal stability of AuNS−BSA−Kem prepared from different BSA−Kem concentrations was examined from their SPR peak in the UV−vis absorption spectrum after constituting AuNS−BSA−Kem in 500 μL of 1× PBS. At a low BSA−Kem concentration of 0.04 mg/mL, the AuNS aggregated even in the absence of salt (1× PBS) (Figure 3A, yellow). This could be due to random electrostatic interactions between the positively charged Kem and negatively charged AuNS when the amount of BSA was insufficient to confer steric stabilization of AuNS. We found that a minimum of 0.4 mg/mL BSA−Kem was needed to stabilize 50 pM AuNS, and therefore, any amount below the minimum would cause AuNS−BSA− Kem to aggregate in 1× PBS (Figure 3B, green). The conjugation of BSA and BSA−Kem to AuNS at this minimum concentration caused an ∼20 nm red shift in the UV−vis peak absorbance of AuNS (Figure 2A, green and blue, E

DOI: 10.1021/acs.analchem.7b05417 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 5. (A) Normalized SERS spectra of AuNS−BSA−Kem that was not phosphorylated [PKA(−)], and phosphorylated by PKA [PKA(+)], in comparison to SERS spectra of AuNS−BSA−p-Kem using a pre-phosphorylated Kem [p-Kem(+)]. (B) PCA scores plot of various principal components, showing that only the PC2 component was able to distinguish non-phosphorylated samples [Phos(−), red] comprising PKA(−) from phosphorylated [Phos(+), blue] samples, which comprised PKA(+) and p-Kem(+), as demarcated by the yellow dotted line. (C) Loadings plot showing two distinct variations in SERS peak intensities at 725 and 1395 cm−1 for the PC2 component that were responsible for the clear separation between the phosphorylated and non-phosphorylated samples on the PCA scores plot along the PC2 axis.

protein load for samples without AuNS was ∼28 μg (7 μL × 4 mg/mL). Hence, the protein load for AuNS−BSA−Kem could be calculated from the ratio of band intensities to be ∼10.92 μg. With the assumption that one BSA molecule was bioconjugated to two Kem, the molar ratio of BSA−Kem to AuNS was estimated to be 3200 BSA per AuNS during phosphorylation. Raman Spectra Acquisition and Analysis. AuNS were known to be competent SERS substrates that enhanced the weak Raman spectra of molecules attached on their surface to produce much stronger SERS spectra. Here, we acquired the SERS spectra of AuNS−BSA−Kem in the presence [PKA(+)] and absence [PKA(−)] of PKA to probe for spectral changes following phosphorylation of Kem. The commercially available pre-phosphorylated Kem was conjugated to form AuNS−BSA− p-Kem as our positive control [p-Kem(+)] for phosphorylation. The SERS spectra between PKA(−), PKA(+), and pKem(+) showed indiscernible differences that were difficult to distinguish visually (Figure 5A). Hence, we applied PCA on the SERS spectra after the preprocessing steps of autofluorescence removal and SNV normalization. PCA is a nonparametric classification technique commonly used in analysis of Raman spectra to discriminate samples with seemingly similar spectroscopic fingerprints. Here, we used PCA to identify subtle variations in the SERS peaks attributed to phosphorylation. The subtle variations in the SERS spectra were identified by applying PCA on 12 samples, with each sample measured thrice (N = 36) to ensure reliability. We observed distinct clustering between the non-phosphorylated samples (Figure 5B, red) comprising PKA(−) and phosphorylated samples (Figure 5B,

We also noted a faint fluorescent band in the absence of PKA [Figure 4, lane 2; without AuNS, PKA(−)], which could be due to incomplete removal of background staining by Pro-Q Diamond dye on ATP in the gel. Also, the transition metal ion (Ga3+) in the Pro-Q Diamond could interact with negatively charged compounds such as BSA and contribute to the weak background staining. Phosphorylation of BSA−Kem by PKA was also possible after adsorption to form AuNS−BSA−Kem as similar bright bands were also observed in both volumes with the same amount of PKA, ATP, Mg2+, and AuNS−BSA−Kem [Figure 4, lanes 4 and 6; with AuNS, PKA(+) for 100 and 500 μL]. The fluorescence intensities in the two reaction volumes were also comparable to each other and that of commercially available pKem as our positive control (Figure 4, lane 8), indicating that concentration of phosphorylating reactants did not greatly affect the phosphorylation efficiency of AuNS−BSA−Kem. In the absence of PKA, we observed no fluorescent band from AuNS−BSA−Kem, indicating the absence of any phosphorylation activity and background staining (Figure 4, lanes 5 and 7). Here, the dark band stained by Coomassie blue proved the presence of BSA−Kem on AuNS, although the band was weaker compared to BSA−Kem without AuNS. The weaker Coomassie blue staining could mean less BSA−Kem present on AuNS−BSA−Kem compared to the amount of free BSA−Kem used in the gel staining, thus accounting for weaker background staining between samples with and without AuNS (lanes 5 and 7 vs lane 2) in the absence of PKA. We estimated the protein load on AuNS by analyzing the band intensities of the Coomassie blue stain using ImageJ. The F

DOI: 10.1021/acs.analchem.7b05417 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 6. PKA activity index (IPKA), defined as the ratio of SERS peak intensities at 725 cm−1 to those at 1395 cm−1 (I725/I1395) in the SERS spectra normalized to that of AuNS−BSA−Kem alone as reference, for various phosphorylated samples measured in vitro. These include pre-phosphorylated AuNS−BSA−p-Kem [p-Kem(+)] and AuNS−BSA−Kem in the presence of PKA and ATP [PKA(+)], as well as non-phosphorylated negative controls including AuNS−BSA−Kem in the presence of PKA but not ATP [ATP(−)], AuNS−BSA−Kem with ATP but not PKA [PKA(−)], and AuNS−BSA in the presence of PKA and ATP, but without Kem [Kem(−)]. The specificity of IPKA to PKA was also determined with various isozymes of PKC, i.e., PKC-α/ε/η/ζ isozymes from bovine, human, human, and rat, respectively. The normalized ratio for phosphorylated samples was significantly higher than non-phosphorylated ones including those incubated in PKC isozymes (p < 0.0001, Student t test).

1.03 ± 0.03], and AuNS−BSA in the presence of PKA and ATP, but without Kem [Kem(−), IPKA = 1.02 ± 0.05] (Figure 6). While the IPKA of PKA(+) was slightly lower than that of AuNS−BSA−p-Kem that was commercially pre-phosphorylated (IPKA = 2.09 ± 0.16), the values were not quite statistically different (p = 0.0549, Student t test). Our results thus showed that IPKA, based on I725/I1395, in the SERS spectra could be a quantifiable measure of PKA activity level to distinguish phosphorylated from non-phosphorylated samples. This index was also specific to phosphorylation since the presence of PKA or ATP alone did not result in a significantly higher value of IPKA. Moreover, the measure of PKA activity level using IPKA was PKA-specific since the presence of PKC, another serine/ threonine kinase, did not result in any false positive as we observed insignificant level of phosphorylation by all isozymes of PKA. Here, the IPKA values of AuNS−BSA−Kem in the presence of four PKC-α/ε/η/ζ isozymes were calculated to be 0.98 ± 0.03, 0.95 ± 0.07, 1.01 ± 0.02, and 1.03 ± 0.04, respectively, all of which were significantly less than phosphorylated AuNS−BSA−Kem by PKA, thus demonstrating the specificity of this assay using IPKA. The coefficient of variance (CV) of all IPKA values was calculated to be between 2.9% and 8.6%. The relatively low CV (