Mutated Human P-Selectin Glycoprotein Ligand-1 and Viral Protein-1 of Enterovirus 71 Interactions on Au Nanoplasmonic Substrate for Specific Recognition by Surface-Enhanced Raman Spectroscopy

Protein tyrosine sulfation is a common post-translational modification that stimulates intercellular or extracellular protein-protein interactions and is responsible for various important biological processes, including coagulation, inflammation, and virus infections. Recently, human P-selectin glycoprotein ligand-1 (PSGL-1) has been shown to serve as a functional receptor for enterovirus 71 (EV71). It has been proposed that the capsid viral protein VP1 of EV71 is directly involved in this specific interaction with sulfated or mutated PSGL-1. Surface-enhanced Raman spectroscopy (SERS) is used to distinguish PSGL-1 and VP1 interactions on an Au nanoporous substrate and identify specific VP1 interaction positions of tyrosine residue sites (46, 48, and 51). The three tyrosine sites in PSGL-1 were replaced by phenylalanine (F), as determined using SERS. A strong phenylalanine SERS signal was obtained in three regions of the mutated protein on the nanoporous substrate. The mutated protein positions at (51F) and (48F, 51F) produced a strong SERS peak at 1599–1666 cm−1, which could be related to a binding with the mutated protein and anti-sulfotyrosine interactions on the nanoporous substrate. A strong SERS effect of the mutated protein and VP1 interactions appeared at (48F), (51F), and (46F, 48F). In these positions, there was less interaction with VP1, as indicated by a strong phenylalanine signal from the mutated protein.


Fabrication of the Au Nanoporous SERS Substrate
Various kinds of nanoporous surface have been fabricated in previous studies [33][34][35][36][37]. The high density and uniformity of Au nanoporous substrates make them suitable as SERS substrates. In this work, Au nanoporous substrates were fabricated using an Ag-Au alloy (Ag80Au20 at.%) target. Briefly, the experiment started with the deposition of 400-nm-thick Au layers via an electron beam (e-beam) evaporator (VT1-10CE, ULVAC, Tainan, Taiwan) on polished single-crystal silicon (100) wafers primed with a 5-nm-thick adhesion layer of Ti. Then, a 150-nm-thick Ag/Au alloy layer was deposited via e-beam at two deposition rates (0.3 and 1 Å/s, respectively). Finally, the Au nanoporous surface was formed by dealloying Ag in concentrated nitric acid at various durations (30-210 s). A bottom thick adhesion layer of Au was deposited to increase the adhesion strength of Ag/Au alloy to the Si substrate to avoid film breakage during dealloying. The sample was washed with deionized (DI) water to remove the dealloying reaction products and excess nitric acid. The morphology, chemical composition, and optical properties of as-prepared samples were investigated using field-emission scanning electron microscopy (FE-SEM, JSM-7001, JEOL, Tokyo, Japan), HR-TEM-EDS (JEM-2100F, JEOL), and ultraviolet-visible spectrophotometry. Water solution contact angles on as-fabricated substrates were measured at 25 °C with an angle meter. A 5-µL droplet of DI water was placed on the sample surface at three sites to assess the surface wettability for measuring the average contact angle.

Fabrication of the Au Nanoporous SERS Substrate
Various kinds of nanoporous surface have been fabricated in previous studies [33][34][35][36][37]. The high density and uniformity of Au nanoporous substrates make them suitable as SERS substrates. In this work, Au nanoporous substrates were fabricated using an Ag-Au alloy (Ag 80 Au 20 at.%) target. Briefly, the experiment started with the deposition of 400-nm-thick Au layers via an electron beam (e-beam) evaporator (VT1-10CE, ULVAC, Tainan, Taiwan) on polished single-crystal silicon (100) wafers primed with a 5-nm-thick adhesion layer of Ti. Then, a 150-nm-thick Ag/Au alloy layer was deposited via e-beam at two deposition rates (0.3 and 1 Å/s, respectively). Finally, the Au nanoporous surface was formed by dealloying Ag in concentrated nitric acid at various durations (30-210 s). A bottom thick adhesion layer of Au was deposited to increase the adhesion strength of Ag/Au alloy to the Si substrate to avoid film breakage during dealloying. The sample was washed with deionized (DI) water to remove the dealloying reaction products and excess nitric acid. The morphology, chemical composition, and optical properties of as-prepared samples were investigated using field-emission scanning electron microscopy (FE-SEM, JSM-7001, JEOL, Tokyo, Japan), HR-TEM-EDS (JEM-2100F, JEOL), and ultraviolet-visible spectrophotometry. Water solution contact angles on as-fabricated substrates were measured at 25 • C with an angle meter. A 5-µL droplet of DI water was placed on the sample surface at three sites to assess the surface wettability for measuring the average contact angle.

Molecular Probe Detection
Rhodamine 6G was diluted in aqueous solution in the range of 10 −6 to 10 −12 M. To investigate the SERS effect, the molecular solution was placed on the as-fabricated samples (created at rates of 0.3 and 1 Å/s), and then subjected to Raman spectroscopy. Raman spectra were obtained using a confocal microscopy Raman spectrometer (inVia Raman microscope, Renishaw, London, UK) using a He-Ne laser with an excitation wavelength of 633 nm, a magnification of 50×, and an integration time of 10 s over an area of 1 µm × 1 µm (the size of the laser spot was~1 µm). An air-cooled charge-coupled device (CCD) was used as the detector, and the incident power was 3 mW. Before each measurement, the Raman shift was calibrated using the signal at 520 cm −1 with the absolute intensity from a standard silicon wafer.
The enhancement factor (EF) was calculated according to the standard equation [25][26][27]: where I sers and I bulk are the SERS and normal Raman scattering intensities, respectively, and N sers and N bulk are the numbers of molecules contributing to the inelastic scattering intensity, respectively, as evaluated by SERS and normal Raman scattering measurements. The SERS effect was averaged from twenty consecutive measurements. The commercial software COMSOL (version Multiphysics ® 2019) was applied to calculate the EM field distribution on the Au nanoporous substrates [25,27].

Self-Assembled Monolayers of Proteins on Au Nanoporous Substrate
To assemble protein on Au nanoporous substrates, the optimized Au nanoporous substrates were cleaned with acetone for 10 min and then treated with oxygen plasma for 3 min. The treated Au nanoporous surface was soaked in a 2% 3 APTES/ethanol solution for 30 min and then carefully washed with ethanol and heated at 120 • C for 10 min to produce amine groups on the surface and remove surplus ethanol. The APTES-modified Au nanoporous surface was then incubated in phosphate-buffered saline (PBS; pH 7) with 12.5% glutaraldehyde for 1 h. The self-assembled monolayer (SAM) that formed on the surface was washed with double-distilled water (ddH 2 O) and dried with nitrogen gas. Then, the surface-modified Au nanoporous substrate was immersed in 1 µM PBS with GST-PSGL-1 or mutant GST-PSGL-1 peptides for 12 h, respectively. Next, the protein-immobilized Au nanoporous surface was washed with PBS solution and subjected to a PTS reaction for 4 h at 37 • C, as detailed in our earlier publication [21]. After the PTS reaction, the substrate was washed with PBS solution three times. Finally, the protein-modified Au nanoporous surface was soaked in anti-sulfotyrosine (1:1000 dilution) or 1 µM VP-1 in PBS for 12 h, and then unreacted probes were removed with PBS solution and characterized using Raman spectroscopy. The surface compositions of the pristine (Au) and protein-immobilized nanoporous substrates were analyzed using XPS with an Al Kα (1486.6 eV) X-ray source. For the pristine and protein-immobilized nanoporous surfaces, wide-scan C 1s, S 2p, N 1s, O 1s, and Au 4f spectra were collected. The morphology of the SAM surface was investigated using field-emission scanning electron microscopy (FE-SEM) and AFM (Dimension Icon, Bruker, Billerica, MA, USA).

Characterization of the Plasmonic Au Nanoporous Substrate
A continuous plasmonic Au nanoporous substrate was fabricated using both physical and chemical methods, as illustrated in Figure S1 in Supplementary Materials. Both methods can be used to precisely tune the plasmonic properties of nanoporous materials. The FE-SEM images in Figure 2a,b show the morphology of a plasmonic Au nanoporous surface. The pore size varied with dealloying time (30-210 s, with samples denoted as Aup_30s, Aup_60s, Aup_90s, Aup_120s, Aup_150s, Aup_180s, and Aup_210s) and thus with the deposition rate (0.3 and 1 Å/s). The Au pore sizes and contact angles are given in Table S1. After dealloying, the composition of the Ag-Au alloy film changed from 80:20 to~2:20 at.%, as determined from the energy-dispersive X-ray spectroscopy (EDS) spectra. The Au nanoporous surface had a different distribution of plasmonic porous cavity sizes, which significantly affects the local EM field effect. The deposition rate also influences the mechanical and physical properties of nanoporous materials. In addition, the reduction in contact angle is related to the deposition rate. Figure S2 shows a high-resolution transmission electron microscopy (HR-TEM) image of an optimized Au nanoporous substrate (Aup_120, 1 Å/s). The pore size was 50 ± 2 nm, which creates a strong local EM environment, enabling single-molecule detection [24]. The EDS mappings show the formation of a porous structure. Moreover, a continuous plasmonic Au nanoporous substrate has enhanced optical properties from the visible to near-infrared regions under irradiation, which may provide a high density of SERS hot spots.
Coatings 2020, 10, x FOR PEER REVIEW 5 of 14 A continuous plasmonic Au nanoporous substrate was fabricated using both physical and chemical methods, as illustrated in Figure S1 in Supplementary Materials. Both methods can be used to precisely tune the plasmonic properties of nanoporous materials. The FE-SEM images in Figure  2a,b show the morphology of a plasmonic Au nanoporous surface. The pore size varied with dealloying time (30-210 s, with samples denoted as Aup_30s, Aup_60s, Aup_90s, Aup_120s, Aup_150s, Aup_180s, and Aup_210s) and thus with the deposition rate (0.3 and 1 Å/s). The Au pore sizes and contact angles are given in Table S1. After dealloying, the composition of the Ag-Au alloy film changed from ~80:20 to ~2:20 at.%, as determined from the energy-dispersive X-ray spectroscopy (EDS) spectra. The Au nanoporous surface had a different distribution of plasmonic porous cavity sizes, which significantly affects the local EM field effect. The deposition rate also influences the mechanical and physical properties of nanoporous materials. In addition, the reduction in contact angle is related to the deposition rate. Figure S2 shows a high-resolution transmission electron microscopy (HR-TEM) image of an optimized Au nanoporous substrate (Aup_120, 1 Å/s). The pore size was 50 ± 2 nm, which creates a strong local EM environment, enabling single-molecule detection [24]. The EDS mappings show the formation of a porous structure. Moreover, a continuous plasmonic Au nanoporous substrate has enhanced optical properties from the visible to near-infrared regions under irradiation, which may provide a high density of SERS hot spots. The surface plasmon resonance (SPR) spectra of Aup_120s (0.3 and 1 Å/s) and Aup_150s (0.3 and 1 Å/s) substrates are shown in Figure S3. The Ag/Au alloy film SPR peak at 485 nm significantly shifted after the Ag/Au alloy film was dealloyed. Two strong SPR shifts appeared at 681 nm for Aup_150s (1 Å/s) and 735 nm for Aup_120s (1 Å/s) compared to those for a deposition rate of 0.3 Å/s (649 nm for Aup_150s and 688 nm for Aup_120s); these peaks are attributed to the strong multi-polarization electric field coupling in small porous cavities [24][25][26][27]33]. COMSOL was used to calculate the optical properties of the optimized Au nanoporous substrate as detailed in SI. We believe that the higher EM field was generated by continuous nanoporous cavities, which allow the detection of lower concentration target species via SERS.
The as-fabricated Au nanoporous samples (30 to 210 s, 0.3 and 1 Å/s) created strong Raman vibrational modes when rhodamine 6G (R6G) was used as the molecular probe. The SERS spectra of 10 −6 M R6G on as-fabricated Au nanoporous samples were obtained at a laser wavelength of 633 nm, as shown in Figure 3a,b. A strong R6G SERS characteristic peak (ISERS) was obtained at 1363 cm −1 for all samples and corresponds to the ν(C=C) aromatics mode [26]. The large local EM field enhancement was found in the nanoporous cavity zone, which may allow for multipolar excitations and enhanced Raman scattering. In addition, the matched Raman laser energy with the nanoporous SPR band can enhance the maximum SERS response. Sample Aup_120s (1 Å/s) had an increased SERS intensity due to its optimal spacing between porous cavities, as shown in Figure 3c. The enhancement factor (EF) was calculated using Equation (1); the optimized Au nanoporous substrate Aup_120s (1 Å/s) obtained enhancement in the order of 10 7 , using the relatively strong band at 1363 The surface plasmon resonance (SPR) spectra of Aup_120s (0.3 and 1 Å/s) and Aup_150s (0.3 and 1 Å/s) substrates are shown in Figure S3. The Ag/Au alloy film SPR peak at 485 nm significantly shifted after the Ag/Au alloy film was dealloyed. Two strong SPR shifts appeared at 681 nm for Aup_150s (1 Å/s) and 735 nm for Aup_120s (1 Å/s) compared to those for a deposition rate of 0.3 Å/s (649 nm for Aup_150s and 688 nm for Aup_120s); these peaks are attributed to the strong multi-polarization electric field coupling in small porous cavities [24][25][26][27]33]. COMSOL was used to calculate the optical properties of the optimized Au nanoporous substrate as detailed in SI. We believe that the higher EM field was generated by continuous nanoporous cavities, which allow the detection of lower concentration target species via SERS.
The as-fabricated Au nanoporous samples (30 to 210 s, 0.3 and 1 Å/s) created strong Raman vibrational modes when rhodamine 6G (R6G) was used as the molecular probe. The SERS spectra of 10 −6 M R6G on as-fabricated Au nanoporous samples were obtained at a laser wavelength of 633 nm, as shown in Figure 3a,b. A strong R6G SERS characteristic peak (I SERS ) was obtained at 1363 cm −1 for all samples and corresponds to the ν(C=C) aromatics mode [26]. The large local EM field enhancement was found in the nanoporous cavity zone, which may allow for multipolar excitations and enhanced Raman scattering. In addition, the matched Raman laser energy with the nanoporous SPR band can enhance the maximum SERS response. Sample Aup_120s (1 Å/s) had an increased SERS intensity due to its optimal spacing between porous cavities, as shown in Figure 3c. The enhancement factor (EF) was calculated using Equation (1); the optimized Au nanoporous substrate Aup_120s (1 Å/s) obtained enhancement in the order of 10 7 , using the relatively strong band at 1363 cm −1 , and was thus used for further studies. The SERS peak intensity at 1363 cm −1 for an R6G concentration of 10 −7 to 10 −12 M was examined on the optimized Au nanoporous sample, as shown in Figure 3d. The optimized Au nanoporous substrate had a strong SERS intensity due to multipolar enhancement, making it suitable for detection at the single-molecule level. The SERS spectra of R6G were enhanced greatly and equally at each 20 point sampling position on the Au nanoporous substrate, indicating good activity and reproducibility of the sample.
Coatings 2020, 10, x FOR PEER REVIEW 6 of 14 cm −1 , and was thus used for further studies. The SERS peak intensity at 1363 cm −1 for an R6G concentration of 10 −7 to 10 −12 M was examined on the optimized Au nanoporous sample, as shown in Figure 3d. The optimized Au nanoporous substrate had a strong SERS intensity due to multipolar enhancement, making it suitable for detection at the single-molecule level. The SERS spectra of R6G were enhanced greatly and equally at each 20 point sampling position on the Au nanoporous substrate, indicating good activity and reproducibility of the sample.

Chemisorption of Protein-Immobilized Plasmonic Au Nanoporous Substrate
The active peptide (ATEYEYLDYDFL)-containing sulfation site of GST-PSGL-1 was immobilized on the functionalized Au nanoporous surface by 3-amino-propyltriethoxysilane (APTES) and glutaraldehyde as linker molecules. The protein immobilization process was confirmed using X-ray photoelectron spectroscopy (XPS) and SERS characterization. The XPS spectra of C 1s, O 1s, N 1s, S 2p, and Au 4f are shown in Figure 4a-e.

Chemisorption of Protein-Immobilized Plasmonic Au Nanoporous Substrate
The active peptide (ATEYEYLDYDFL)-containing sulfation site of GST-PSGL-1 was immobilized on the functionalized Au nanoporous surface by 3-amino-propyltriethoxysilane (APTES) and glutaraldehyde as linker molecules. The protein immobilization process was confirmed using For the functionalized Au nanoporous surface with oxygen plasma treatment, the AuOx spectrum exhibited a weak signal peak at binding energies of 85.01 and 88.05 eV, which confirms the formation of O and OH on the Au nanoporous surface, and SAM formation [38,39]. The Au 4f5/2 and Au 4f7/2 spectra exhibited strong signal peaks at 83.50 and 87.23 eV, respectively. The peak intensity decreased after protein assembly on the surface due to multi-SAM formation. The O 1s spectrum of the functionalized Au nanoporous surface exhibited an emission peak at 532.24 eV, which indicates the formation of a nano-sized oxidation layer [38]. For the APTES-modified Au nanoporous surface, the C 1s spectrum exhibits strong emission peaks at 284.78 and 288.96 eV, which correspond to C-C and O-C=O, respectively [21,27,40]. The O 1s spectrum of the APTES-modified Au nanoporous surface exhibits a strong emission peak at 532.73 eV, which corresponds to O-C=O from the APTES layer. The N 1s spectrum of the APTES-modified Au nanoporous surface exhibits a strong emission peak at 398.45 eV, which corresponds to NH2 from the APTES layer. The glutaraldehyde-modified Au nanoporous surface C 1s spectrum exhibits an emission peak at 288.96 eV, which corresponds to H-C=O; these results confirm the chemisorption of glutaraldehyde [21]. The O 1s and N 1s spectra of the glutaraldehyde-modified Au nanoporous surface are similar to those obtained for the APTES-modified surface. The XPS peaks at 287.04 and 289.17 eV for C 1s are from the immobilized GST-PSGL-1 on the Au nanoporous surface, which correspond to C-N, H-C=O, and C=O-N, respectively [21,40]. The disappearing glutaraldehyde H-C=O peak at 288.96 eV confirms the immobilization of the proteins. The XPS peaks in the O 1s spectrum from the immobilized GST-PSGL-1 on the Au nanoporous surface appear at 530.24 and 531.84 eV, which can be assigned to N=O and C=O-OH, respectively. The shifted N 1s spectrum shows peaks corresponding to C-N, N=O, and the oxide NH2/NH3 + at 397.70, 399.22, and 402.57 eV, respectively, which strongly indicate the presence of proteins. It shows that greatly influential interactions between protein molecules via amino groups can be found. The S 2p spectrum peak at 167.90 eV can be assigned to SO4 2− , which confirms the PTS reaction on the GST-PSGL-1-assembled surface. These peaks are consistent with SAM chemical states. Therefore, there is strong chemisorption between proteins and the Au nanoporous surface. The fabricated protein-assembled substrate surface was used for further SERS studies. Figure 5 shows the SERS spectra of GST-PSGL-1 and S-GST-PSGL-1 immobilized Au nanoporous substrates obtained at a Raman laser wavelength of 633 nm. The SERS spectra of the APTES-modified functionalized Au nanoporous substrate show weak C-O-C, Si-O-Si, or Si-O-C Raman bands at 813 and 1007 cm −1 , respectively, confirming that APTES had self-assembled on the substrate [41]. Notably, the O-C=O stretching modes had significant shifts at 1590-1700 cm −1 after the protein was immobilized on the Au nanoporous substrate via the SAM procedure, as marked in Figure 5a. For the functionalized Au nanoporous surface with oxygen plasma treatment, the AuO x spectrum exhibited a weak signal peak at binding energies of 85.01 and 88.05 eV, which confirms the formation of O and OH on the Au nanoporous surface, and SAM formation [38,39]. The Au 4f 5/2 and Au 4f 7/2 spectra exhibited strong signal peaks at 83.50 and 87.23 eV, respectively. The peak intensity decreased after protein assembly on the surface due to multi-SAM formation. The O 1s spectrum of the functionalized Au nanoporous surface exhibited an emission peak at 532.24 eV, which indicates the formation of a nano-sized oxidation layer [38]. For the APTES-modified Au nanoporous surface, the C 1s spectrum exhibits strong emission peaks at 284.78 and 288.96 eV, which correspond to C-C and O-C=O, respectively [21,27,40]. The O 1s spectrum of the APTES-modified Au nanoporous surface exhibits a strong emission peak at 532.73 eV, which corresponds to O-C=O from the APTES layer. The N 1s spectrum of the APTES-modified Au nanoporous surface exhibits a strong emission peak at 398.45 eV, which corresponds to NH 2 from the APTES layer. The glutaraldehyde-modified Au nanoporous surface C 1s spectrum exhibits an emission peak at 288.96 eV, which corresponds to H-C=O; these results confirm the chemisorption of glutaraldehyde [21]. The O 1s and N 1s spectra of the glutaraldehyde-modified Au nanoporous surface are similar to those obtained for the APTES-modified surface. The XPS peaks at 287.04 and 289.17 eV for C 1s are from the immobilized GST-PSGL-1 on the Au nanoporous surface, which correspond to C-N, H-C=O, and C=O-N, respectively [21,40]. The disappearing glutaraldehyde H-C=O peak at 288.96 eV confirms the immobilization of the proteins. The XPS peaks in the O 1s spectrum from the immobilized GST-PSGL-1 on the Au nanoporous surface appear at 530.24 and 531.84 eV, which can be assigned to N=O and C=O-OH, respectively. The shifted N 1s spectrum shows peaks corresponding to C-N, N=O, and the oxide NH 2 /NH 3 + at 397.70, 399.22, and 402.57 eV, respectively, which strongly indicate the presence of proteins. It shows that greatly influential interactions between protein molecules via amino groups can be found. The S 2p spectrum peak at 167.90 eV can be assigned to SO 4 2− , which confirms the PTS reaction on the GST-PSGL-1-assembled surface. These peaks are consistent with SAM chemical states. Therefore, there is strong chemisorption between proteins and the Au nanoporous surface. The fabricated protein-assembled substrate surface was used for further SERS studies. Figure 5 shows the SERS spectra of GST-PSGL-1 and S-GST-PSGL-1 immobilized Au nanoporous substrates obtained at a Raman laser wavelength of 633 nm. The SERS spectra of the APTES-modified functionalized Au nanoporous substrate show weak C-O-C, Si-O-Si, or Si-O-C Raman bands at 813 and 1007 cm −1 , respectively, confirming that APTES had self-assembled on the substrate [41]. Notably, the O-C=O stretching modes had significant shifts at 1590-1700 cm −1 after the protein was immobilized on the Au nanoporous substrate via the SAM procedure, as marked in Figure 5a. The SERS spectrum of the APTES-modified Au nanoporous substrate shows the O-C=O stretching mode at 1601 cm −1 , which is correlated to impurities in the ATPES solution, as shown in Figure 5b. The glutaraldehyde-modified substrate SERS spectrum shows strongly shifted O-C=O stretching modes at 1594 and 1655 cm −1 , which correspond to glutaraldehyde immobilization, as shown in Figure 5c [42]. The stretching mode of the O-C=O region can greatly influence interactions between protein molecules via amino groups. The SERS spectra of the GST-PSGL-1 modified Au nanoporous substrate show a strong C=C Raman band at 1586 cm −1 , confirming the immobilization of GST-PSGL-1 [43][44][45][46]. In addition, the stretching mode of O-C=O at 1619 cm −1 shifted to a lower region compared to that of the glutaraldehyde-modified Au nanoporous substrates. S-GST-PSGL-1 formed on the Au nanoporous substrates via the PTS reaction. An anionic sulfate is assigned to the asymmetric SO4 2− and -SO3 stretching modes at 984, 1030, and 1192 cm −1 , confirming GST-PSGL-1 modification via the PTS reaction [27,47]. After the PTS reaction, the observed C=C and O-C=O peaks shifted to 1593 and 1627 cm −1 , as shown in Figure 5d,e, respectively, indicating the presence of The SERS spectrum of the APTES-modified Au nanoporous substrate shows the O-C=O stretching mode at 1601 cm −1 , which is correlated to impurities in the ATPES solution, as shown in Figure 5b. The glutaraldehyde-modified substrate SERS spectrum shows strongly shifted O-C=O stretching modes at 1594 and 1655 cm −1 , which correspond to glutaraldehyde immobilization, as shown in Figure 5c [42]. The stretching mode of the O-C=O region can greatly influence interactions between protein molecules via amino groups. The SERS spectra of the GST-PSGL-1 modified Au nanoporous substrate show a strong C=C Raman band at 1586 cm −1 , confirming the immobilization of GST-PSGL-1 [43][44][45][46]. In addition, the stretching mode of O-C=O at 1619 cm −1 shifted to a lower region compared to that of the glutaraldehyde-modified Au nanoporous substrates. S-GST-PSGL-1 formed on the Au nanoporous substrates via the PTS reaction. An anionic sulfate is assigned to the asymmetric SO 4 2− and -SO 3 stretching modes at 984, 1030, and 1192 cm −1 , confirming GST-PSGL-1 modification via the PTS Coatings 2020, 10, 403 9 of 14 reaction [27,47]. After the PTS reaction, the observed C=C and O-C=O peaks shifted to 1593 and 1627 cm −1 , as shown in Figure 5d,e, respectively, indicating the presence of a protein. The SERS shifts are listed in Table S2. These results are consistent with the XPS results, as detailed in Figure 4.

Protein-Protein
Interactions on the Plasmonic Au Nanoporous Substrate Figure 6 shows a SERS spectrum of the GST-PSGL-1 mutant's immobilization on an Au nanoporous substrate obtained at a Raman laser wavelength of 633 nm. The SERS band positions were significantly changed due to the various GST-PSGL-1 mutants; these peaks are assigned to proteins, amides, and lipids, as listed in Table S3 [25,26,[47][48][49][50]. The tyrosine residue positions 46, 48, and 51 of GST-PSGL-1 were positions mutated to phenylalanine (F). Weak phenylalanine SERS bands were observed at 998-1005 cm −1 , confirming the presence of the mutant on the surface. Strong SERS peaks appeared at 1595-1650 cm -1 , which are related to the O-C=O stretching modes of phenylalanine [44,51]. In addition, a strong anionic sulfate SERS peak appeared at 986 cm −1 for the non-mutant GST-PSGL-1. In the case of the GST-PSGL-1 mutant's immobilization on an Au nanoporous substrate, the anionic sulfate SERS peak appeared at 1100-1300 cm  (Figure 6b). The maximum mutated protein SERS intensity (at O-C=O) was reduced compared to that of the control GST-PSGL-1 on the Au nanoporous substrate due to the structure being modified by phenylalanine. Nevertheless, the mutated GST-PSGL-1 positions at 48F, 51F, (46F, 48F), and (46F, 51F) on the Au nanoporous substrate showed greatly enhanced SERS intensity (at O-C=O). These positions could be used to understand interactions with antibodies or VP1 of EV71.
The S-GST-PSGL-1-immobilized Au nanoporous substrate was allowed to interact with anti-sulfotyrosine and examined at a Raman laser wavelength of 633 nm. Figure 6c shows a SERS spectrum of S-GST-PSGL-1 interaction with anti-sulfotyrosine. Strong anionic sulfate SERS bands appeared at 978 and 1282 cm −1 , corresponding to stretching modes of SO 4 2− , -SO 3 , and S=O, which may correspond to the strong interaction between the sulfated protein and the antibody. The major and minor Raman assignments of the protein and antibody were performed in the 600-1800 cm −1 range; they are listed in Table S4. The SERS peaks shifted to higher and lower regions due to the chemical configuration of protein interactions with the antibody, as shown in Figure S4  of the mutated protein. Three regions of the mutated protein (46F, 48F, 51F) on the Au nanoporous substrate showed highly enhanced SERS intensity at 1001 cm −1 due to a high quantity of phenylalanine. Furthermore, the SERS relative intensity of the peak positions at 1626-1649 cm −1 (O-C=O) were compared to detect the SERS activity of the mutated regions on the Au nanoporous substrate (Figure 6b). The maximum mutated protein SERS intensity (at O-C=O) was reduced compared to that of the control GST-PSGL-1 on the Au nanoporous substrate due to the structure being modified by phenylalanine. Nevertheless, the mutated GST-PSGL-1 positions at 48F, 51F, (46F, 48F), and (46F, 51F) on the Au nanoporous substrate showed greatly enhanced SERS intensity (at O-C=O). These positions could be used to understand interactions with antibodies or VP1 of EV71.  It is known that S-GST-PSGL-1 acts as a mediator between cells and an external virus. The VP1 (1 µm) of EV71 was diluted using PBS solution, with which about 2 mL was taken and immobilized onto the protein-modified samples for 12 h. The mutated GST-PSGL-1 and VP1 of EV71 interaction on the Au nanoporous substrate was examined at a Raman laser wavelength of 633 nm; the results are shown in Figure 7. The major and minor Raman assignments of PPIs were performed in the 600-1800 cm −1 range; they correspond to amides, amino acids, and lipids, as listed in Table S5 The strong SERS bands of PPIs (mutated GST-PSGL-1 and VP1) that appear at 1361 cm −1 for (46,48,51) are assigned to lipids, those at 1121 cm −1 for (46F) are assigned to C-C trans-conformation, those at 826 cm −1 for (48F) are assigned to NH 2 , those at 1601 cm −1 for (51F) are assigned to O-C=O, those at 1139 cm −1 for (46F, 48F) are assigned to C-C trans-conformation, those at 1567 cm −1 for (46F, 51F) are assigned to amides, those at 838 and 858 cm −1 for (48F, 51F) are assigned to NH 2 and lipids, and those at 1131 cm −1 for (46F, 48F, 51F) are assigned to C-C trans-conformation [25,26,44,52,53]. The results indicate that the SERS signals were significantly different for each mutated protein interaction with VP1 due to the different structural composition of protein. The SERS characteristic peaks are mostly from VP1, which indicated the binding strength of PPIs. Figure S5 shows the SERS spectra of VP1 dropped on an Au nanoporous substrate. The peaks correspond to the amides, amino acids, and lipids of VP1. Moreover, the SERS relative intensity of the peak position at 1600 cm -1 was compared to the detected SERS intensity of PPIs (Figure 7b). The strong SERS intensity of the mutated protein at (48F), (51F), and (46F, 48F) indicates less interaction with VP1 due to the presence of a strong phenylalanine signal from the mutated protein. Therefore, the activity of mutated GST-PSGL-1 at positions (48F), (51F), and (46F, 48F) may control the path for virus entry, which is consistent with our previous work and detailed in SI. The AFM measurements of S-GST-PSGL-1 mutants and VP1 of EV71, as shown in Figure 7c. Generally, selectin-PSGL-1 interactions depend on the microenviroment of the interacting molecules (molecular structure). PTS-based GST-PSGL-1 may regulate the interaction of VP1 of EV71. The force measurement procedure described in a previous paper was applied in this study. AFM was used to measure the unbinding force between the outer sheath protein VP-1 of EV71 and GST-PSGL-1. Through APTES and glutaraldehyde (GA), the proteins were, respectively, covalently attached on the apex of the AFM tip and silicon substrate. Then, the unbinding forces between the proteins were analyzed. The unbinding force between sulfated-GST-PSGL-1 and VP-1 was obviously stronger than that of nonsulfated-GST-PSGL-1 since the specific bonding between the proteins affects the unbinding force.
Coatings 2020, 10, x FOR PEER REVIEW 11 of 14 paper was applied in this study. AFM was used to measure the unbinding force between the outer sheath protein VP-1 of EV71 and GST-PSGL-1. Through APTES and glutaraldehyde (GA), the proteins were, respectively, covalently attached on the apex of the AFM tip and silicon substrate. Then, the unbinding forces between the proteins were analyzed. The unbinding force between sulfated-GST-PSGL-1 and VP-1 was obviously stronger than that of nonsulfated-GST-PSGL-1 since the specific bonding between the proteins affects the unbinding force.

Conclusion
In this study, Au nanoporous SERS-active substrates were utilized for detecting PPIs. Protein immobilization on the Au nanoporous substrate was confirmed using XPS and SERS. The mutated protein SERS effect was investigated through the observation of a strong phenylalanine SERS signal. The mutated protein positons at (51F) and (48F, 51F) exhibited a high SERS effect at 1599-1666 cm −1 , which could be related to non-binding with the mutated protein and anti-sulfotyrosine interaction on the Au nanoporous substrate. The stretching mode of O-C=O (phenylalanine) at the 1600 cm −1 SERS band was compared to the mutated protein and VP1 interactions. The strong SERS effect of the mutated protein and VP1 interactions was obtained at (48F), (51F), and (46F, 48F); these positions indicated less interaction with VP1 due to the presence of a strong phenylalanine signal from the mutated protein. The proposed method will help in drug development and contribute to the understanding of physiological and pathological mechanisms.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Figure S1: Schematic illustration of fabrication of Au nanoporous substrate, Figure S2: HR-FETEM image of optimized Au nanoporous substrate and EDS mappings of areas marked in HR-FETEM image, Figure S3: SPR spectra of

Conclusions
In this study, Au nanoporous SERS-active substrates were utilized for detecting PPIs. Protein immobilization on the Au nanoporous substrate was confirmed using XPS and SERS. The mutated protein SERS effect was investigated through the observation of a strong phenylalanine SERS signal. The mutated protein positons at (51F) and (48F, 51F) exhibited a high SERS effect at 1599-1666 cm −1 , which could be related to non-binding with the mutated protein and anti-sulfotyrosine interaction on the Au nanoporous substrate. The stretching mode of O-C=O (phenylalanine) at the 1600 cm −1 SERS band was compared to the mutated protein and VP1 interactions. The strong SERS effect of the mutated protein and VP1 interactions was obtained at (48F), (51F), and (46F, 48F); these positions indicated less interaction with VP1 due to the presence of a strong phenylalanine signal from the mutated protein. The proposed method will help in drug development and contribute to the understanding of physiological and pathological mechanisms.