A. Validation of microparticle–aptamer–protein complexes

We performed SEM, XPS, and SDS-PAGE analysis to validate the protein–aptamer–microparticle complexes of two kinds before the acoustofluidic particle-based liquid biopsy. Figure 3 shows the SEM images of (a) 5.12 μm-diameter PS microparticles (PS5), (b) 6.96 μm-diameter PS microparticles (PS7), (c) AptC-functionalized PS5 (PS5-AptC), (d) AptT-functionalized PS7 (PS7-AptT), (e) CRP-conjugated PS5-AptC (PS5-AptC-CRP), and (f) Thr-conjugated PS7-AptT (PS7-AptT-Thr). The streptavidin-coated PS microparticles in Figs. 3(a) and 3(b) seem to be smooth microspheres with approximate diameters of 5 and 7 μm, respectively. The two ssDNA aptamers (AptC and AptT designed for selective conjugation with CRP and Thr, respectively) coated onto the bare microspheres via streptavidin–biotin interactions can be visually confirmed by the roughened surfaces with white dots on the microspheres in Figs. 3(c) and 3(d). The two target biomarker proteins, CRP and Thr, conjugated with their corresponding aptamer-functionalized PS microparticles can be verified by the protein–aptamer–microparticle complexes in Figs. 3(e) and 3(f).Figures 4(a) and 4(b) show the XPS results of the CRP and Thr complexes, respectively, where streptavidin-coated PS microparticles, aptamer-functionalized PS microparticles, and target protein-conjugated aptamer-functionalized PS microparticle are comparatively analyzed. The inset figures represent the high-magnified views of the P2p peaks. In the streptavidin-coated PS microparticles (black lines), a distinct C1s peak at approximately 280 eV can be observed due to the chemical composition of PS, as well as O1s and N1s peaks at approximately 540 and 400 eV, respectively, but without the P2p peak. It should be noted that the O1s and N1s peaks are present due to the amino acid in streptavidin and that bare PS microparticles should only have a C1s peak. The XPS analysis results for PS5-AptC and PS5-AptT (blue lines) show an additional P2p peak at approximately 130 eV because of the phosphoric acid in the ssDNA aptamers (AptC and AptT). The high-affinity cross-linkage between the DNA aptamers and streptavidin-coated PS microparticles is verified by the distinct P2p peaks even after several medium exchange in the fabrication protocol. The target protein-conjugated aptamer-coated PS microparticles (red lines) show slightly decreased P2p peaks with others remaining almost the same confirms the fact that most CDRs of the DNA aptamers were occupied by the conjugated target proteins (CRP and Thr).Moreover, we conducted SDS-PAGE analysis on the protein-conjugated aptamer-functionalized microparticle complexes to validate that the conjugated proteins were the target CVD biomarker proteins (CRP and Thr), as in Fig. 4(c). The proteins conjugated onto the particles via ssDNA aptamers were isolated by introducing DNase into the protein-conjugated particle solutions. DNase was used to nonspecifically catalyze the degradation of ssDNA aptamers and thus to produce various lengths of oligonucleotides. After the DNase treatment, the protein-conjugated particle solutions were centrifuged at 1775 g for 10 min for the selective retrieval of the proteins from the PS microparticles. The protein extracts were then electrophoresed in a 12% Tris-glycine gel. As negative controls, pure CRP and Thr were also used for SDS-PAGE for comparison. The SDS-PAGE bands numbered 1 and 3 indicate CRP and Thr as negative controls while those numbered 2 and 4 represent CRP and Thr retrieved from the protein–aptamer–microparticle complexes. The CRP and Thr extracted from the complexes were found to have molecular weights of 244848. C. C. Yue, J. Muller-Greven, P. Dailey, G. Lozanski, V. Anderson, and S. Macintyre, J. Biol. Chem. 271(36), 22245–22250 (1996). https://doi.org/10.1074/jbc.271.36.22245 and 35 kDa,4949. J. F. Keijzer, J. Firet, and B. Albada, Chem. Commun. 57(96), 12960–12963 (2021). https://doi.org/10.1039/D1CC05446E respectively, which are in good agreement with the negative controls.

B. Acoustofluidic tri-separation of CVD biomarker proteins

Figure 5 shows the acoustofluidic separation of CVD biomarker proteins of two kinds from platelets in human plasma in a continuous manner. The target proteins were conjugated to the corresponding aptamer-functionalized microparticles, and the particle complexes were manipulated using the proposed acoustofluidic method in a label-free manner. With no acoustic field applied to the device in Fig. 5(a), all the particulate samples (PS5-AptC-CRP, PS7-AptT-Thr, and PRP) suspended in plasma were observed to travel without any disturbance along the flow. For the experiments, the flow rates of the sample, sheath 1, and sheath 2 solutions were 20, 150, 30 μl/h, respectively. The sample solution was pinched by the two sheath fluid flows so that all the suspensions in plasma were focused near the one microchannel sidewall close to the IDT, as in Fig. 1(c). As the sheath fluid, we have used 0.1% (w/v) PEO solution to prevent the adhesion to the microchannel wall and aggregation of the protein-conjugated microparticles suspended in human plasma. The microchannel width and height were 300 and 40 μm, respectively, where the particulate samples were exposed to the acoustic field [Figs. 5(a) and 5(b)].Unlike the conventional cross-type acoustofluidic devices including our previous studies28,5028. M. Afzal, J. Park, J. S. Jeon, M. Akmal, T. S. Yoon, and H. J. Sung, Anal. Chem. 93(23), 8309–8317 (2021). https://doi.org/10.1021/acs.analchem.1c0119850. R. Ahmad, G. Destgeer, M. Afzal, J. Park, H. Ahmed, J. H. Jung, K. Park, T. S. Yoon, and H. J. Sung, Anal. Chem. 89(24), 13313–13319 (2017). https://doi.org/10.1021/acs.analchem.7b03474 where SAWs were produced from an IDT and propagated perpendicular to the microparticles, in this study, an IDT was placed at an inclined angle of approximately 10° such that the suspended objects were subjected to the tilted-angle acoustic waves. In this configuration, we could use a closed-type PDMS microfluidic chip, composed of a 50-μm-thick PDMS membrane and a microchannel-patterned PDMS stamp, and place it right on top of the IDT. The microchannel can be replaceable, preventing potential cross-contamination of the samples used in the device from multiple patients; the IDT deposited on a piezoelectric substrate can be reused, reducing the fabrication and operation cost of the proposed method. The wave attenuation in PDMS was reported as 7 dB/cm⋅MHz,34,5134. J. Park, J. H. Jung, K. Park, G. Destgeer, H. Ahmed, R. Ahmad, and H. J. Sung, Lab Chip 18(3), 422–432 (2018). https://doi.org/10.1039/C7LC01083D51. D. Rabaud, P. Thibault, J.-P. Raven, O. Hugon, E. Lacot, and P. Marmottant, Phys. Fluids 23(4), 042003 (2011). https://doi.org/10.1063/1.3579263 and the transmitted wave energy through the membrane was imparted on the microparticle complexes inside the microchannel. The tilted-angle SAW configuration allowed utilization of the vertical component of the ARF in addition to the horizontal component of the ARF,5252. H. Ahmed, G. Destgeer, J. Park, J. H. Jung, and H. J. Sung, Adv. Sci. 5(2), 1700285 (2018). https://doi.org/10.1002/advs.201700285 resulting in the increased ARF magnitude acting on the microparticles in the acoustic field despite the wave attenuation in the PDMS membrane.Figure 5(b) demonstrates the tri-separation of two kinds of the proteins-conjugated microparticle complexes and platelets in plasma by 130 MHz SAW at 10.5 mW. As discussed earlier with Fig. 2, the PS5 and PS7 experienced the traveling SAW-induced ARF with different magnitudes at the working frequency of 130 MHz. The lateral migration of the PS5 particles with κ = 1.39 was greater than that of the PS7 particles with κ = 1.90 while the platelets, whose κ was smaller than 1, exhibited negligible deflection by the traveling acoustic field. Therefore, CRP conjugated onto the AptC-coated PS5 was translocated to the microchannel sidewall opposite to the IDT, while Thr conjugated onto the AptT-functionalized PS7 was deflected to the center of the microchannel. It should be noted that the functionalized PS microparticle complexes were manipulated in a label-free manner in the proposed acoustofluidic approach. The surface modification and conjugated proteins on the surface of the microparticles were assumed to have negligible effects on the acoustic radiation since the acoustic wavelength (∼μm) far exceeded the surface modification layer (∼nm). The platelets with an average diameter of 2 μm at 130 MHz are in the Rayleigh scattering regime, and they were too small to experience the 130 MHz SAW-induced ARF. As a consequence, the PS5-AptC-CRP, PS7-AptT-Thr, and platelets located in the three different locations in the microchannel were separated at downstream tri-furcation and thus collected at the outlets 1 (top), 2 (middle), and 3 (bottom), respectively, as in Fig. 4(c) (see also supplementary material Video 1).A disposable PDMS microfluidic chip with a thin PDMS membrane has been used for detachable acoustofluidic devices. Our previous studies33,3433. J. Park, J. H. Jung, G. Destgeer, H. Ahmed, K. Park, and H. J. Sung, Lab Chip 17(6), 1031–1040 (2017). https://doi.org/10.1039/C6LC01405D34. J. Park, J. H. Jung, K. Park, G. Destgeer, H. Ahmed, R. Ahmad, and H. J. Sung, Lab Chip 18(3), 422–432 (2018). https://doi.org/10.1039/C7LC01083D used the disposable chips for parallel-type acoustofluidic devices to utilize the traveling SAWs propagating parallel to the flow direction for droplet manipulation. Similarly, Ma et al.3030. Z. Ma, D. J. Collins, and Y. Ai, Anal. Chem. 88(10), 5316–5323 (2016). https://doi.org/10.1021/acs.analchem.6b00605 and Zhao et al.3131. S. Zhao, M. Wu, S. Yang, Y. Wu, Y. Gu, C. Chen, J. Ye, Z. Xie, Z. Tian, H. Bachman, P. H. Huang, J. Xia, P. Zhang, H. Zhang, and T. J. Huang, Lab Chip 20(7), 1298–1308 (2020). https://doi.org/10.1039/D0LC00106F also proposed detachable cross-type acoustofluidic devices for traveling SAW and tilted-angle standing SAWs propagating perpendicular to the flow direction, respectively. The present work used the thin membrane-based disposable PDMS microfluidic chip to utilize the tilted-angle traveling SAWs in the detachable acoustofluidic device. In the proposed device, the lateral migration has been improved up to approximately 236 μm for PS5 and 118 μm for PS7, which can be further increased with wider microchannels in the proposed acoustofluidic device. Also, the inter-distance between the two target microparticles was also increased up to approximately 118 μm, resulting in the stable and efficient tri-separation of the two target biomarker proteins from platelets in human blood plasma.Figure 6 shows the influence of the total flow rate (Qtotal) on the separation efficiency. Figures 6(a)–6(c) correspond to the stacked microscopic images of the cases of Qtotal = 0.2, 0.8, and 1.6 ml/h with the flow rate ratio remained the same as sheath 1 (top): sample (middle): sheath 2 (bottom) = 15 : 2 : 3 (see also supplementary material Video 2). The experimental observation clearly demonstrated that the PS5-AptC-CRP, PS7-AptT-Thr, and platelets were separated by the SAW-induced ARF acting on the microparticle complexes even at the increased total flow rates compared to those in Fig. 5. The total flow rate with successful acoustofluidic tri-separation was 0.2 ml/h in the cross-type arrangement of the IDT with respect to the flow.2828. M. Afzal, J. Park, J. S. Jeon, M. Akmal, T. S. Yoon, and H. J. Sung, Anal. Chem. 93(23), 8309–8317 (2021). https://doi.org/10.1021/acs.analchem.1c01198 The increased sample flow rate can be attributed to the additional utilization of the upward migration induced by the vertical component of the ARF,5252. H. Ahmed, G. Destgeer, J. Park, J. H. Jung, and H. J. Sung, Adv. Sci. 5(2), 1700285 (2018). https://doi.org/10.1002/advs.201700285 and it allows improved throughput of the proposed method. For quantification on the separation efficiency, we conducted hemocytometry analysis on the samples collected at outlets 1 and 2. The separation efficiency was calculated as the number of the target microparticles out of the total number of the microparticles at each outlet, i.e., purity, using the hemocytometry images obtained from five independent experiments. The target microparticle complexes were PS5- AptC-CRP and PS7-AptT-Thr in outlets 1 and 2, respectively, as shown in Fig. 1(c). In Fig. 6(d), the green bar indicates the separation efficiency calculated from the number of the PS5-AptC-CRP complexes out of all, while the red bar indicates the separation efficiency calculated from the number of the PS7-AptT-Thr complexes out of all. In all the cases of Qtotal = 0.2–1.6 ml/h, the purity was measured to exceed 95%, which verifies that almost all the target proteins were separated at the designated outlets. Obviously, the lower the total flow rate was, the higher the purity was measured with 99% purity in the case of Qtotal = 0.2 ml/h. In the proposed acoustofluidic device, we utilized the tilted-angle traveling SAWs to improve the processing throughput, in addition to the lateral migration, for the stable and efficient separation of the two biomarker proteins from platelets in human blood plasma. The proposed tilted-angle configuration for the traveling SAWs allows the utilization of the vertical component of the acoustic radiation force (ARF) in addition to the horizontal component of the ARF.5252. H. Ahmed, G. Destgeer, J. Park, J. H. Jung, and H. J. Sung, Adv. Sci. 5(2), 1700285 (2018). https://doi.org/10.1002/advs.201700285 Compared to the conventional cross-type SAW-based acoustofluidic devices,53,5453. R. W. Rambach, V. Skowronek, and T. Franke, RSC Adv. 4(105), 60534–60542 (2014). https://doi.org/10.1039/C4RA13002B54. J. Duan, M. Ji, and B. Zhang, Biosensors 12(8), 611 (2022). https://doi.org/10.3390/bios12080611 the magnitude of the ARF that the microparticles experienced was enhanced. This can be verified in the improvement in the sample flow rate, the flow velocity, and the lateral migration. The maximum total volumetric flow rate with the tri-separation efficiency of >95% was 1.6 ml/h in the microchannel with a cross section of 300 μm (width) × 40 μm (height). The corresponding maximum average flow velocity was calculated as 37.04 mm/s, which was much higher than that in the previous acoustofluidic devices.30,31,53–5930. Z. Ma, D. J. Collins, and Y. Ai, Anal. Chem. 88(10), 5316–5323 (2016). https://doi.org/10.1021/acs.analchem.6b0060531. S. Zhao, M. Wu, S. Yang, Y. Wu, Y. Gu, C. Chen, J. Ye, Z. Xie, Z. Tian, H. Bachman, P. H. Huang, J. Xia, P. Zhang, H. Zhang, and T. J. Huang, Lab Chip 20(7), 1298–1308 (2020). https://doi.org/10.1039/D0LC00106F53. R. W. Rambach, V. Skowronek, and T. Franke, RSC Adv. 4(105), 60534–60542 (2014). https://doi.org/10.1039/C4RA13002B54. J. Duan, M. Ji, and B. Zhang, Biosensors 12(8), 611 (2022). https://doi.org/10.3390/bios1208061155. J. Nam, Y. Lee, and S. Shin, Microfluid. Nanofluid. 11(3), 317–326 (2011). https://doi.org/10.1007/s10404-011-0798-156. L. Schmid, D. A. Weitz, and T. Franke, Lab Chip 14(19), 3710–3718 (2014). https://doi.org/10.1039/C4LC00588K57. J. Behrens, S. Langelier, A. R. Rezk, G. Lindner, L. Y. Yeo, and J. R. Friend, Lab Chip 15(1), 43–46 (2015). https://doi.org/10.1039/C4LC00704B58. G. Simon, M. A. B. Andrade, J. Reboud, J. Marques-Hueso, M. P. Y. Desmulliez, J. M. Cooper, M. O. Riehle, and A. L. Bernassau, Biomicrofluidics 11(5), 054115 (2017). https://doi.org/10.1063/1.500199859. M. Wu, Y. Ouyang, Z. Wang, R. Zhang, P. H. Huang, C. Chen, H. Li, P. Li, D. Quinn, M. Dao, S. Suresh, Y. Sadovsky, and T. J. Huang, Proc. Natl. Acad. Sci. U.S.A. 114(40), 10584–10589 (2017). https://doi.org/10.1073/pnas.1709210114It should be noted that, unlike the previous studies where protein-conjugated carrier microparticles were spiked in buffer solution,17,50,6017. J. Hwang, Y. Seo, Y. Jo, J. Son, and J. Choi, Sci. Rep. 6, 34778 (2016). https://doi.org/10.1038/srep3477850. R. Ahmad, G. Destgeer, M. Afzal, J. Park, H. Ahmed, J. H. Jung, K. Park, T. S. Yoon, and H. J. Sung, Anal. Chem. 89(24), 13313–13319 (2017). https://doi.org/10.1021/acs.analchem.7b0347460. R. Huang, J. Quan, B. Su, C. Cai, S. Cai, Y. Chen, Z. Mou, P. Zhou, D. Ma, and X. Cui, Sens. Actuators B 359, 131583 (2022). https://doi.org/10.1016/j.snb.2022.131583 human blood plasma with platelets suspended was used as a sample solution in the present study. As in Fig. 7(a), we found in the experiments that the protein–microparticle complexes aggregated and adhered to the PDMS microchannel wall. The irreversible protein adsorption to the PDMS surface is attributed to the conformational changes and denaturation of the protein molecules induced by the hydrophobic interactions between the PDMS surface and the proteins.44,6144. H. Chen, M. A. Brook, and H. Sheardown, Biomaterials 25(12), 2273–2282 (2004). https://doi.org/10.1016/j.biomaterials.2003.09.02361. H. Zhang and M. Chiao, J. Med. Biol. Eng. 35(2), 143–155 (2015). https://doi.org/10.1007/s40846-015-0029-4 This undesirable side-effect was not observed when the bare PS microparticles in buffer solution were exposed to the SAW-induced acoustic field. All the experimental conditions in Fig. 7 were the same as those of Fig. 6(a). As the practical application of the proposed technique involves human plasma, the aggregation and adhesion of the microparticle complex to the wall should be addressed. We addressed this issue by introducing PEO solution as a sheath fluid flow and PEO coating on the PDMS microchannel inner-wall. PEO is a typical biocompatible polymer known to prevent protein damage and adhesion.6262. H. Chen, L. Yuan, W. Song, Z. Wu, and D. Li, Prog. Polym. Sci. 33(11), 1059–1087 (2008). https://doi.org/10.1016/j.progpolymsci.2008.07.006 Therefore, Fig. 7(b) shows that the PEO sheath flow significantly restrained the protein-carrying PS microparticles from being attached to the PDMS wall and accumulated, while the PEO coating was observed to have a similar effect, as found in Fig. 7(c). The protein adsorption into to the hydrophobic PDMS was observed to be more significant especially when the human plasma was used as the liquid medium. We assume that the plasma further accelerated the interactions between the protein molecules and the hydrophobic PDMS surface. Hydrophobicity in the PDMS microchannel surface caused the platelets suspended in plasma to be activated and, as a consequence, the proteins conjugated on the particles to denature and adhere to the wall. The PEO molecules inhibited the platelet activation and thus prevent the protein adsorption and aggregation.63,6463. K. S. Siow, S. Kumar, and H. J. Griesser, Plasma Process. Polym. 12(1), 8–24 (2015). https://doi.org/10.1002/ppap.20140011664. N. P. Desai and J. A. Hubbe, Biomaterials 12, 144–153 (1990). https://doi.org/10.1016/0142-9612(91)90193-E The PEO used as the coating agent and sheath fluid served as protein repellant molecules to prevent the undesirable protein aggregation and adsorption to the PDMS wall.4444. H. Chen, M. A. Brook, and H. Sheardown, Biomaterials 25(12), 2273–2282 (2004). https://doi.org/10.1016/j.biomaterials.2003.09.023 Nevertheless, either of the two PEO usage approaches did not completely address the unwanted phenomenon of the protein–particle complex in human plasma in Fig. 7(a). We discovered that the adhesion and accumulation of the protein-carrying particle complex in plasma can be resolved by applying both PEO coating on the PDMS wall and PEO sheath flow, as in Fig. 7(d). The biofouling of proteins suspended in biofluids has been one of major challenges in biomedical devices, and non-specific protein adsorption and aggregation should be prevented for practical applications.6161. H. Zhang and M. Chiao, J. Med. Biol. Eng. 35(2), 143–155 (2015). https://doi.org/10.1007/s40846-015-0029-4

C. Verification of target proteins after acoustofluidic separation

After the acoustofluidic separation of the aptamer-functionalized PS microparticles on which the target proteins were conjugated, we conducted the SEM, XPS, and SDS-PAGE analysis to confirm the successful separation of the target proteins in plasma. Figures 8(a) and 8(b) show the SEM images of the PS5-AptC-CRP and PS7-AptT-Thr collected at outlets 1 and 2 after being separated by the SAW-induced ARF, respectively. The target CRP and Thr were found to remain attached to their respective carrier microparticles (PS5 and PS7), as similar to Figs. 3(e) and 3(f). These SEM results confirmed that the strong binding strength between the engineered DNA aptamers and the corresponding target proteins, as well as the streptavidin-biotin binding, stayed robust against the acoustic field formed in the microchannel. Figures 8(c) and 8(d) show the XPS analysis results of the aptamer-functionalized microparticles carrying CRP and Thr. The CRP-conjugated PS5-AptC and Thr-conjugated PS7-AptT exhibited the similar peaks for O1s at 540 eV, N1s at 400 eV, C1s at 280 eV, and P2p at 130 eV in the XPS results similar to those in Figs. 3(c) and 3(d), respectively. With the SEM and XPS analysis results, we confirmed that the target proteins on the functionalized microparticles remained attached to the particle surface after exposure to the acoustic field due to high affinity of the protein–aptamer conjugation. Figure 8(e) represents the SDS-PAGE results of the isolated proteins by applying DNase into the microparticle–aptamer–protein complexes. For the negative control bands, platelet-rich-plasma, and pure CRP and thrombin were used in each SDS-PAGE gel. Without SAW application in Fig. 5(a), CRP, Thr, and platelets were all focused by the two sheath fluid flows and thus collected in outlet 3 while no particulate sample was collected in the other outlets. The three bands of the outlet 3 sample in the “SAW OFF” column of Fig. 8(e) were observed to match with the negative control bands of each sample in the left-most column. As confirmed in Fig. 4(c), CRP and thrombin collected at the outlet 3 were measured to be 24 and 35 kDa after the acoustofluidic separation while the platelets were measured to be approximately 65 kDa.6565. P. R. Erickson and M. C. Herzberg, J. Biol. Chem. 265(24), 14080–14087 (1990). https://doi.org/10.1016/S0021-9258(18)77270-0 The PRP band may have included other non-target plasma proteins, such as albumin (65–70 kDa),6666. H. Kaab, M. M. Bain, K. Bartley, F. Turnbull, H. W. Wright, A. J. Nisbet, R. Birchmore, and P. D. Eckersall, Poult. Sci. 98(2), 679–687 (2019). https://doi.org/10.3382/ps/pey431 which were not separated by the acoustic field.Once the sample solution was subject to the acoustic field formed in the microchannel, the two elastic PS microspheres carrying their target proteins experienced the SAW-induced ARF and therefore deflected from their original pathlines, as in Fig. 5(b). At the downstream trifurcation, the three kinds of the particulate suspensions located in three different locations within the microchannel were separated into the three different outlets, as shown in Fig. 5(c). The collected sample at each outlet was qualitatively analyzed by SDS-PAGE, and the results are shown in the “SAW ON” column in Fig. 8(e). As observed in the microscopic observation in Fig. 5(c), the separated particulate samples were found to be CRP, Thr, and platelets in outlets 1, 2, and 3, respectively. These results verified that the 130 MHz SAW-induced acoustic field successfully separated PS5-AptC-CRP and PS7-AptT-Thr from the platelets in human plasma. We expect that the proposed acoustofluidic particle-based liquid biopsy for multiplexed protein assays using human blood samples can be utilized as a promising clinical tool toward early diagnosis of various diseases. As a follow-up study, on-chip acoustofluidic microparticle enrichment5050. R. Ahmad, G. Destgeer, M. Afzal, J. Park, H. Ahmed, J. H. Jung, K. Park, T. S. Yoon, and H. J. Sung, Anal. Chem. 89(24), 13313–13319 (2017). https://doi.org/10.1021/acs.analchem.7b03474 and fluorescent protein quantification6767. L. Wu and X. Qu, Chem. Soc. Rev. 44(10), 2963–2997 (2015). https://doi.org/10.1039/C4CS00370E usage can further advance the applicability of the proposed acoustofluidic approach for clinical particle-based liquid biopsy.