Single- or few-layered BP NSs were synthesized by a modified mechanical exfoliation method from BP crystal powders in organic solvent [30], [31], [32]. As shown in Figure 1A, the transmission electron microscopic (TEM) image reveals an apparent flake structure of BP NSs with an average lateral size of about 300 nm, which reflects an appropriate size for efficient endocytosis [33]. The TEM image shows that the surface of BP NSs is coated with a great deal of globular Au nanosphere whose average size is about 30 nm (Figure 1B). These data are consistent with the result measured by dynamic light scattering analysis, which indicates that the average diameters of AuNPs and BP-Au NSs are 37 nm and 342 nm, respectively (Figure S1, Supporting Information). In the enlarged view (Figure 1C), a distinct lamellar structure can be observed at the edge of nanohybrids, which confirms the successful connection of Au nanoparticles with BP NSs. A lot of AuNPs are confined to the NS, forming nanoaggregates, which have been proven to be an excellent SERS substrate due to the abundant “hot spots” existing in the nanostructure [34]. The energy-dispersive X-ray (EDX) spectral mapping shows three elemental mapping images (Au, P and O) with good overlap, further verifying the successful fabrication of BP-Au NSs (Figure 1D). Compared to bare BP, BP-Au NSs display significantly improved absorption in the ultraviolet (UV)-visible region and show a characteristic absorption peak at 540 nm, which could attest the presence of Au nanoparticles (Figure 1E). The Raman characterization of the nanohybrids (Figure 1F) exhibits three primary Raman bands, one out-of-plane phonon mode (A¹_g) and two in-plane modes (B_2g and A²_g), which are located at about 358, 431 and 459 cm⁻¹, respectively [35], [36], [37]. Compared with the Raman spectrum of BP NSs, no significant change is observed from that of BP-Au NSs.

Figure 1:

(A) TEM image of BP nanosheets. (B) Overall and (C) enlarged TEM images of BP-Au NSs. (D) EDX elemental mapping of BP-Au nanocomposites. (E) UV-Vis absorption spectra of the BP and BP-Au NSs at the same concentration. (F) Raman spectra of BP NSs and BP-Au NSs.

To illustrate the SERS activity of BP-Au NSs, the SERS experiments were carried out using human hepatocellular carcinoma (HepG2) cells as the model. After treatment with the nanomaterials and fresh medium for 4 h, Raman signals of cells from different regions were detected. As shown in Figure 2A and B, cells display faint signals whether in Raman single-point measurement or overall imaging after incubation with fresh medium or BP NSs. On the contrary, ubiquitous intensive SERS signals from 620 to 1720 cm⁻¹ can be observed from almost all of the intracellular regions when the cells are incubated with BP-Au NSs (Figure 2C).

Figure 2:

Raman mapping and corresponding spectra of HepG2 cells incubated with (A) fresh medium, (B) BP NSs and (C) BP-Au NSs. Scale bar: 15 μm.

The discovery of the excellent SERS activity of BP-Au NSs endows them with the capability for ultrasensitive detection of various cellular components. The stretching vibration peaks of DNA mainly display at 835 cm⁻¹ (asymmetric O-P-O stretching vibration), 1316 cm⁻¹ (guanine) and 1375 cm⁻¹ (thymine and adenine). The characteristic bands of proteins basically appear at 655 cm⁻¹ (C-C twisting of tyrosine), 820 cm⁻¹ (tyrosine ring breath), 1004 cm⁻¹ (phenylalanine ring symmetrical breathing) and 1597 cm⁻¹ (C-C bending vibration of phenylalanine and tyrosine). Raman peaks at 1130 cm⁻¹ (C-O/C-C stretching vibration in disaccharide) and 1271 cm⁻¹ (lipid CH₂ deformation mode) represent the existence of carbohydrates and lipids in cells. These data indicate that the BP-Au NSs can be used as a SERS-active substrate to detect the intracellular complex biological components. Detailed assignments are summarized in Table S1 (Supporting Information).

We further investigated the process of cellular uptake of BP-Au NSs by the confocal Raman spectrometer. The SERS signals of HepG2 cells were acquired after the cells were incubated with BP-Au NSs for 1, 2, 4, 6, 8 and 12 h. As shown in Figures 3 and S2 (Supporting Information), visible Raman peaks originating from intracellular components confirm that the nanohybrids can be rapidly internalized into the cells within an hour. As time progresses, a moderate increase in the intensity of Raman signals is found over an incubation period of 6 h, meaning that the nanohybrids gradually accumulate inside the cells. Afterward, the intensity of Raman signals emerges a downward trend with elongation of incubation time. During the entire endocytosis process, the strongest Raman signals of cancer cells induced by BP-Au NSs appear at 6 h, which may be chosen as the best time point for further SERS investigation. There is a nonlinear relation between the incubation time of nanohybrids and the intensity of Raman signals. The main reason for the decrease in SERS intensity is probably related to the excessive irreversible aggregation of nanohybrids, which could not maintain appropriate inter-nanoparticle spacing on plasmon resonance, resulting in a reduction in the interparticle hot spots [38], [39].

Figure 3:

Average SERS spectra of HepG2 cells incubated with BP-Au NSs for 1, 2, 4, 6, 8 and 12 h, respectively.

SERS spectroscopy has become a viable alternative to fluorescence-based techniques due to its narrow spectral bandwidth, label-free feature and resistance to autofluorescence and photobleaching. For cellular mechanism study on the internalization of nanohybrids, the cells were coincubated with BP-Au NSs and various endocytosis inhibitors, such as amiloride (Ami, macropinocytosis inhibitor), methyl-β-cyclodextrin (MβCD; caveolae inhibitor), chlorpromazine (CPZ; clathrin inhibitor) and sodium azide (NaN₃; adenosine triphosphate-depleting agent). These endocytosis inhibitors were administered at concentrations that could not affect cell viability (Figure S3, Supporting Information). SERS images of HepG2 cells were acquired by integrating the signals from 620 to 1720 cm⁻¹, which indirectly reflected the distribution of BP-Au NSs within the cross-section of the cells. As shown in Figure 4A, intense SERS signals can be observed in the cells incubated with BP-Au NSs in the absence of any endocytosis inhibitor. In contrast, the Raman signals of HepG2 cells are hardly measured when the cells were pretreated with NaN₃, showing that the endocytosis of BP-Au NSs is energy dependent. Strong Raman signals also appear in the cells coincubated with CPZ, indicating that the uptake of BP-Au NSs can essentially not be interrupted by inhibition of the clathrin-dependent pathway. In the case of cells exposed to Ami or MβCD, relatively hypointense SERS images are obtained, which implies that the nanohybrids are mainly ingested into the cells by macropinocytosis and caveolae-dependent endocytosis.

Figure 4:

(A) Raman mapping of HepG2 cells incubated with BP-Au NSs after treatment with different endocytic inhibitors (scale bar: 10 μm). (B) Fluorescence images HepG2 cells incubated with BP-Au/Rh B in the presence of various endocytosis inhibitors. Scale bar: 200 μm.

For quantitative estimation of the uptake of nanohybrids into cells, 150 Raman spectral lines per group were acquired from the cells coincubated with different endocytosis inhibitors. The spectral line that could clearly report the fingerprint information of intracellular components was judged to be the positive signal. As shown in Table 1, CPZ has little impact on the enhanced Raman signals of HepG2 cells containing BP-Au NSs, with about 84% positive spectral lines being detected, whereas the ratios of valid SERS spectral lines obtained from the cells pretreated with Ami, MβCD and NaN₃ are 44.7%, 48% and 38.7%, respectively. Moreover, the rate of positive signals decreases dramatically when multiple pathways of endocytosis are inhibited. The endocytosis mechanism of BP-Au NSs was also verified using traditional fluorescence microscopy. The fluorescent (FL) signals of BP-Au NSs were acquired by grafting an exogenous fluorescence dye, Rh B, onto the NSs. As shown in Figures 4B and S4 (Supporting Information), the most distinct FL signals of BP-Au NSs are observed in the cells coincubated with or without CPZ, while cells treated with nanohybrids in the presence of Ami, MβCD or NaN₃ only emit obscure fluorescence, which are consistent with the data assessed by the SERS technique, further demonstrating that energy-dependent macropinocytosis and caveolae-mediated endocytosis are the main endocytosis pathways of BP-Au NSs. A similar mechanism of endocytosis of BP NSs has been demonstrated [13], implying that the introduction of Au NPs has no significant impact on the process of cellular uptake.

To investigate the intracellular fate of nanohybrids after endocytosis, fluorescence colocalization imaging experiment was performed. The subcellular structures of lysosomes, mitochondria and the endoplasmic reticulum (ER) were stained with FL organelle-specific trackers to facilitate colocation analysis [40], [41], [42]. Nanoparticles and other extracellular substances taken up by macropinocytosis and caveolae-dependent endocytosis could be enclosed in multiple organelles [43], [44], [45]. As shown in Figure 5, the fluorescence ascribable to BP-Au NS nanohybrids does not match that of mitochondria tracker but coincided with that of lysosomes and ER stains during the entire incubation period. These results indicate that BP-Au NS nanohybrids are mainly distributed in the ER and lysosomes. To accurately evaluate the subcellular distribution of the internalized nanohybrids, we calculated the colocalization coefficient of BP-Au NSs with ER tracker, lysotracker and mitochondria tracker (Figure 5D–F). With increasing incubation time, the amount of nanohybrids internalized by macropinocytosis within lysosomes gradually decreased (Figure 5F), indicating the possibility of either intracellular degradation of BP-Au NSs or transportation of the hybrids to extracellular and intracellular regions. The colocalization coefficient value for BP-Au NSs with the ER exhibits an increase at 2 h, followed by a slight decline at 4 h (Figure 5D). These results demonstrate that prolonged retention of BP-Au NSs inside living cells can be achieved when the nanohybrids are localized in the lysosome and ER via multi-endocytosis pathways.

Figure 5:

Fluorescence co-localization analysis of BP-Au NSs with the (A) ER, (B) mitochondria and (C) lysosome at different time points. Scale bar: 20 μm. (D–F) Colocalization coefficients between nanohybrids and different organelles.

The exact subcellular localization of BP-Au NSs in the lysosome and ER makes it possible to identify the fingerprints of these two organelles through Raman-fluorescence dual-mode mapping using Raman microspectroscopy. The fluorescence images of the two organelle-specific probes were acquired under 514.5 nm laser excitation in the 515–600 nm range (Figure S5, Supporting Information). The SERS images of HepG2 cells mediated by BP-Au NSs were obtained using 785 nm laser excitation. Figure 6A displays the Raman-fluorescence dual imaging of the cancer cells, where a good overlap between the Raman signals and the FL images of the two organelles (ER and lysosomes) is observed, which is in accordance with the data detected by fluorescence colocalization assay. The overlap analysis between the two imaging modes was further performed using Matlab software, with the best anastomotic pixels being separated (Figure 6B). The overlap rates in the lysosome and ER regions are 52% and 61%, respectively. Then, the SERS spectra of cellular components were extracted from these overlapped pixel points, which could reflect the molecular information on the ER and lysosome. As shown in Figure 6C, the mean SERS spectra of the ER and lysosome exhibit similar spectral patterns but notable differences in some Raman peaks, indicating the diversity of biochemical components in the two organelles. For example, Raman bands at 1160 cm⁻¹ (C-C/C-N stretching of protein), 1214 cm⁻¹ (amide III β-sheet), 1353 cm⁻¹ (CH₂/CH₃ wag of protein), 1399 cm⁻¹ (COO⁻ symmetrical stretching of protein), 1471 cm⁻¹ (C-H deformation/bending of protein) and 1570 cm⁻¹ (amide II) mainly appear in the mean Raman spectrum of the ER [46], [47]. However, the dominant Raman peaks in the spectral lines of lysosome are located at 1135 cm⁻¹ (C-C stretching mode of protein, chain C-C stretching of lipid or C-O/C-C in disaccharide), 1284 cm⁻¹ (Amide III, C-O-C vibrations of sugar rings), 1448 cm⁻¹ (CH₂/CH₃ deformations in proteins and lipids) and 1519 cm⁻¹ (Amide II) [47], [48], [49]. Not only the distributions of biochemical molecular species but also those of active lysosomes and ER in the cells were successfully visualized via the SERS technique. It may pave new insights for synthesizing functional targeted BP-based nanomaterials and the understanding of the interaction between cells and nanomaterials.

Figure 6:

(A) SERS-fluorescence dual-mode imaging analysis of the subcellular localization of BP-Au NSs. Scale bar: 20 μm. (B) Overlap analysis of the FL and SERS images. (C) SERS spectra extracted from the ER and lysosome of HepG2 cells.

The BP crystal powders were purchased from a commercial supplier (XFNANO, Nanjing, China) and stored in a dark Ar glovebox. Chloroauric acid (HAuCl₄·4H₂O) was bought from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Ami and methyl thiazolyl tetrazolium (MTT) were obtained from Sigma-Aldrich (St. Louis, MO, USA). N-methyl-2-pyrrolidone (NMP), CPZ, dimethyl sulfoxide (DMSO), MβCD, NaN₃ and Rh B were provided by Aladdin biological technology Co., Ltd. (Shanghai, China). Dulbecco’s Modified Eagle Medium (DMEM), fetal bovine serum (FBS) and trypsin were obtained from GIBCO (Grand Island, NY, USA). Lyso-Tracker Green, Mito-Tracker Green, ER-Tracker Green and ER-Tracker Blue-White DPX were obtained from Yesen Biotechnology Technology Co., Ltd. (Shanghai, China). Deionized distilled water (Milli-Q System, Millipore, Billerica, MA, USA) was used throughout the experiment.

The single-layered BP NSs were prepared by a liquid exfoliation method according to a previously reported protocol. Typically, the bulk BP crystal (25 mg) powers were immersed in NMP (25 ml) and kept sonicated with a sonic tip for 6 h at a power of 600 W (on/off cycle: 4 s/2 s). Subsequently, the mixture solution was ultrasonicated for 12 h at 4°C using a power of 400 W. The dispersion was dealt with centrifugation (7000 rpm, 10 min) to collect the sediment containing BP NSs and then resuspended in NMP (5 ml). For the synthesis of BP-Au NSs, the BP (150 μl, 5 mg ml⁻¹) and HAuCl₄ (300 μl, 10 mm) solutions were successively dispersed in boiling water (20 ml) for 2 min until the color of the finished product was altered to purple red. Afterward, the solution was concentrated by centrifugation at 7000 rpm for 10 min. For fluorescence microscopy imaging, the BP-Au NSs and Rh B (20 mg) solutions were stirred overnight in a dark room and precipitation of BP-Au/Rh B was obtained after centrifugation (7000 rpm, 10 min).

TEM images and energy dispersive spectrometer mapping were characterized by a 120 kV JEM-2100HR transmission electron microscope (JEOL, Japan) equipped with an EDX spectrometer. UV-vis absorption spectra were taken on a spectrophotometer (UV-6100S, MAPADA, China). Raman spectra were measured by a high-resolution confocal Raman spectrometer (Derbyshire, England) equipped with excitation wavelengths of 514.5 nm and 785 nm. The hydrodynamic diameters of the Au nanoparticles and BP-Au NSs were measured by a Malvern Mastersizer 2000 (Zetasizer Nano ZS90, Malvern Instruments Ltd., UK).

For SERS bioanalysis, the HepG2 cells were cultured in high-glucose DMEM medium with 10% FBS under humidified 5% CO₂ at 37°C. Cells precultured on the quartz slide were incubated with fresh medium, BP NSs and BP-Au NSs for 4 h. Then, the samples were washed with phosphate buffer solution (PBS) several times for SERS measurement. To study the process of uptake of BP-Au NSs, HepG2 cells were incubated with 50 μm of BP-Au NSs for 1, 2, 4, 6, 8 and 12 h, respectively. After rinsing with PBS for three times, 30 cells were measured per time point by the Raman spectrometer under the nm laser excitation.

HepG2 cells were cultured in a 96-well plate at a density of 4×10⁴ cells/well overnight, then the medium was replaced by fresh medium containing Ami (80 μg ml⁻¹), CPZ (15 μg ml⁻¹), MβCD (8 mm) and NaN₃ (40 mm). After 24 h of incubation, 10 μl of MTT (5 mg ml⁻¹) was added and incubated in the dark for another 4 h, then MTT was replaced by DMSO to completely dissolve the purple formazan and detected the absorbance at 490 nm.

The HepG2 cells in the logarithm growth period were seeded in a culture dish at a density of 3×10⁵ cells per well for 10 h. After that, Ami (80 μg ml⁻¹), CPZ (15 μg ml⁻¹), MβCD (8 mm) and NaN₃ (40 mm) were pre-incubated cells for 2 h and then substituted with BP-Au/Rh B (50 μm) for another 6 h. Fluorescence imaging studies were performed with a fluorescence microscope. For SERS measurement, the cells were incubated with BP-Au NSs (50 μm) and detected under the Raman microspectrometer.

To study the subcellular localization of nanohybrids, HepG2 cells were seeded in 20-mm glass-bottom dishes and then treated with BP-Au NSs (50 μm) for 1, 2 and 4 h, respectively. After washing with PBS and fixing by 4% paraformaldehyde for 10 min, the cells were incubated with the primary FL probe: Lyso-Tracker Green, Mito-Tracker Green and ER-Tracker Blue-White DPX. Finally, the cells were washed to eliminate excess dyestuffs and observed under the confocal microscopy (LSM-710).

For fluorescence Raman imaging study, the HepG2 cells were incubated in fresh medium containing BP-Au (50 μm) for 2 h. Soon afterward, Lyso-Tracker Green and ER-Tracker Green were added for 15 min. Raman imaging and fluorescence imaging were performed under Raman microspectroscopy with excitations at 514.5 nm (FL) and 785 nm (Raman).