Introduction

The growing application and environmental discharge of engineered nanomaterials (ENMs) today are posing high risks to the ecological system and human health. One major concern is that the ENMs can be transferred from low to high trophic levels through food chain. For example, polyvinylpyrrolidone-coated silver nanoparticles were found to be transferred and biomagnified through the Escherichia coli (E. coli) - C. elegans food chain1, or from green algae to grazers2, causing severe toxicity to the higher trophic level. The transfer of quantum dots (QDs) from ciliated protozoans to rotifers through dietary uptake3 and from bacteria (Pseudomonas aeruginosa) to ciliate (Tetrahymena thermophila) has also been demonstrated4, 5. Fluorescent QDs were favorably used in these studies because of their widespread use6,7,8 and easy traceability in-vivo by optical or transmission electron microscopy. The distribution, accumulation and toxicity of nanoparticles in specific living organisms have been intensively studied by using bacteria, mammalian cells and Caenorchabditis elegans as the models9,10,11. However, the possible changes of these ENMs during their food-chain transport were largely unknown, making their specific environmental behaviors, fate and toxicity poorly understood.

Characterization of EMNs in living organisms is challenging, due to the interference of soluble elements and complicated biological environment. The transformation between nanoparticles and their ionic forms may also occur, impeding an accurate quantification of EMNs in vivo 9. Especially, for fluorescent QDs, fluorescence quenching, emission spectrum red shift or blue shift frequently occur in conventional fluorescence microscopic observation9, making their quantification and localization difficult. Recently, X-ray absorption fine structure (XAFS) and micro-X-ray fluorescence (μ-XRF) spectroscopy was used to reveal detailed information about the chemical composition and distribution of elements or compounds in vivo at high spatial resolution12, 13. Raman mapping offers another powerful tool to visualize the molecular composition of compounds in vivo at subcellular scale without labeling14. While these methodologies have been proven useful for characterization of metallic nanoparticles15, here we for the first time used them to track the distribution and biotransformation of QDs during food-chain transport.

In this work, we used Shewanella onedensis as a representative of environmental bacteria and C. elegans as a representative of metazoan to constitute a simplified freshwater food chain, through which the transport and dynamic changes of CdSe QDs were investigated. C. elegans is broadly distributed in soil and water ecosystems and has been widely employed as a toxicology test model16. Since the uptake of nanoparticles by bacteria have been intensively reported and well known17, 18, here we focus on the downstream step, i.e., how the QDs are transported with the S. onedensis cells into C. elegans, To this end, we used in-vivo synthesized QDs by S. onedensis for simplification, In-situ fluorescence microscopy and Raman mapping were used to monitor the QDs distribution dynamics based on the fluorescence and chemical bond properties of QDs, respectively. The μ-XRF spectroscopy was used to localize the Se element distribution. The food footprints were also recorded by XAFS analysis. Together, the results reveal valuable information about the biotransformation and distribution of CdSe QDs during their food-chain transport.

Results and Discussion

Characterization of CdSe QDs in S. onedensis MR-1 cells

The fluorescent CdSe QDs in S. ondensis MR-1 were clearly observed by fluorescence microscopy (Fig. 1a). All the cells were lightened with bright yellow fluorescence ascribed to the in-vivo formed CdSe QDs. The in-situ Raman spectra of the QDs-containing S. ondensis cells showed several distinct peaks at 202 cm−1, 406 cm−1, 750 cm−1, 1128 cm−1, 1313 cm−1 and 1585 cm−1 (Fig. 1b). Among these, the peaks at 202 cm−1 and 406 cm−1 could be ascribed to the longitudinal-optical (LO) and 2 longitudinal-optical (2LO) phonons of CdSe, respectively19. The Raman characteristics of CdSe QDs were size-dependent20, 21, Thus, based on the Raman spectra, the average size of QDs was estimated as slightly larger than 3 nm. The Raman spectra also showed the characteristic peaks of ferrous and ferric state cytochrome C (750 cm−1, 1128 cm−1, 1313 cm−1 and 1585 cm−1) from S. onedensis 22. A direct observation of the purified QDs from S. ondensis cells by High-angle annular dark-field scanning transmission electron microscope (HAADF-STEM) confirmed the presence of the discrete nanoparticles with an average diameter of 3.3 nm. High-resolution TEM (HRTEM) analysis further revealed that the QDs had continuous lattice fringes with an interplanar lattice distance of 0.25 nm, corresponding to the (102) interplanar spacing of wurtzite structured CdSe.

Figure 1
figure 1

Characteristics of CdSe QDs in Shewanella onedensis cells. (a) Fluorescence image. (b) In-situ Raman spectrum (excited at 532 nm) shows characteristic peaks of QDs and cytochrome c. (c) HAADF-STEM and (d) HRTEM images of the purified CdSe QDs.

Full size image

Distribution and movement of CdSe QDs in C. elegans

Fluorescence microscopy was used to track the distribution and movement of QD in C. elegans over time. According to the fluorescence images in Fig. 2, the negative control emitted blue autofluorescence only, while the experimental group showed numerous dots of bright fluorescence on the background of faint blue autofluorescence. Apparently, these bright fluorescent dots should be ascribed to the ingested QDs together with the S. onedensis cells. Here, the fluorescent color of QDs was distinct from the typical yellow green fluorescence of CdSe QDs because the QDs were located inside the nematode body that gave blue autofluorescence to interfere the observation. A further examination on the fluorescence dynamics showed that the cell-embodied QDs were ingested from the mouth, through the pharynx lumen, gut and finally to the rectum of the C. elegans. In this process, the fluorescence color changed from yellow to light red, likely attributed to the aggregation of the QDs in the acidic microenvironment of pharynx lumen23. A closer observation showed that, after 6-h exposure to the QDs-containing S. onedensis cells, the aggregated QDs (in red color) were abundantly present in the gut lumen and some dispersed QDs (in blue-green color) also entered the adjacent intestinal cells. The color change indicates that the QDs shifted back to the original state when transported from the gut lumen to adjacent intestinal cells. After 12-h exposure, the QDs became more enriched in the gut lumen and a part of the QDs were excreted from the anus of nematode. A comparison of the intestinal fluorescence signals at different time points clearly showed an expansion of the lightened area from the intestinal tract to other subcutaneous tissue and at the same time the fading of fluorescence in the gut lumen, further confirming the movement of QDs from gut to subcutaneous tissue and excretion.

Figure 2
figure 2

Distribution dynamics of CdSe QDs in C. elegans over time recorded using fluorescence microscopy. The blue signals represent the autofluorescence of C. elegans. The bright blue-green and light red fluorescence signals represent the dispersed and aggregated CdSe QDs, respectively. All the scale bars are 10 μm.

Full size image

Co-localization of QDs and S. onedensis in C. elegans

To systematically evaluate the fate of QDs through food chain, the Raman signal of both the prey S. onedensis (750 cm−1, 1128 cm−1, 1313 cm−1 and 1585 cm−1) and the CdSe QDs (202 cm−1) were recorded by Raman mapping after exposuring the nematode to the QD-containing bacteria (Fig. 3). After 6-h, the C. elegans showed obvious CdSe accumulation in the digestive tracts of  the C. elegans was observed, consistent with the fluorescence microscopy results. The co-localization of QDs and cytochrome C, an important membrane-anchored protein in S. onedensis, in the gut lumen confirms that the QDs were ingested together with S. onedensis. Furthermore, the expansion of cytochrome C signal to the whole nematode body over time revealed a fast assimilation of S. onedensis cells by C. elegans into subcutaneous tissue. Notably, the distribution of QDs was not fully overlapped with that of cytochrome C signals, indicating that  the CdSe QDs might have been partially transformed to other chemical forms.

Figure 3
figure 3

Raman mapping of CdSe QDs (202 cm−1) and cytochrome C (750 cm−1, 1128 cm−1, 1313 cm−1 and 1585 cm−1) bond signals in C. elegans after 6-h exposure.

Full size image

Distribution of Se element in C. elegans

The distribution of Se element in C. elegans body after 48-h exposure to S. onedensis cells was examined by μ-XRF spectroscopy. This technique can reduce the possible interference of the biological background compared with other optical and chemical characterization approaches. Figure 4 shows strong Se signal at the digestive tracts, consistent with the fluorescence and Raman results. Notably, the μ-XRF spectral signal showed a broader distribution area than the fluorescence and Raman signals in the nematode body, which further supports the transformation of fluorescent CdSe QDs to the other Se species. In contrast, the negative control showed no signals of QDs confirming that the Se was originated from CdSe QDs.

Figure 4
figure 4

Micro-X-ray fluorescence (μ-XRF) spectroscopy of Se element in C. elegans after 48-h exposure.

Full size image

Biotransformation of CdSe QDs in C. elegans

To shed light on the biotransformation of CdSe QDs in C. elegans, Se K-edge XAFS was used to examine the chemical composition changes of the QDs during food-chain transfer (Fig. 5). Both X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) spectroscopy were applied for the analysis. The oxidation state of Se can be reflected by the peak positions in XANES (Fig. 5a)24. The peak position of QDs-containing S. onedensis matched that of CdSe−II references (12,662 eV). Interestingly, after being ingested by C. elegans, new peaks emerged at both lower and higher energy positions, matching those of Se0 and Na2SeO3 IV standards respectively.

Figure 5
figure 5

Changes of Se chemical species during food-chain transfer. (a) Se K-edge normalized XANES spectra of references Se (Se: elemental selenium, Na2SeO3: sodium selenite and CdSe: cadmium selenide) and biological samples. (b) Se K-edge EXAFS k 3 χ(k) functions and (c) Fourier transformed of Se K-edge EXAFS k 3 χ(k) functions of biological samples and standards.

Full size image

The partial transformation of CdSe QDs to Se0 and Na2SeO3 IV was further evidenced by the Fourier transforms of the Se K-edge EXAFS (Fig. 5c). Although the QDs-containing S. onedensis and C. elegans both showed Se-Se and Se-Cd bonds consistent with the Se0 and CdSe−II standards, the peak intensities were different. The QDs-containing S. onedensis had a higher intensity of Cd-Se bond, while the C. elegans showed high intensity of Se-Se bond. In addition, the Se-O bond signal observed in C. elegans confirms that the CdSe QDs were partially oxidized to Na2SeO3 IV.

Conclusion

The transformation and distribution of CdSe QDs during their transfer from S. onedensis to C. elegans through food-chain was systematically examined by using advanced spectrum techniques including fluorescence microscopy, Raman mapping. μ-XRF and XAFS spectroscopy. Our results show that the ingested QDs were initially accumulated in the alimentary system, mainly the gut, of C. elegans. Later, the distribution area of both the QDs and the cytochrome C increased over time, and partial excretion of the QDs was observed at later stage. Meanwhile, a gradual transformation of CdSe QDs to Se0 and Na2SeO3 IV during the food-chain transport was also identified.

Materials and Methods

Procedures of QDs transferred from S. onedensis to C. elegans

The C. elegans (Bristol strain N2, wild type) was obtained from Caenorhabditis Genetics Center (Minneapolis, USA) and were grown in Petri dishes on nematode growth medium (17 g/L agar, 2.5 g/L peptone, 3 g/L NaCl, 1 mM CaCl2, 1 mM MgSO4, 25 mM KPO4 buffer and 5 mg/L ethanol). The Escherichia coli OP50 strain was used as food source and cultured at 20 °C. The age-synchronized nematode was obtained using hydroxide/hypochlorite treatment according to the description in a previous study25 and grown at 20 °C until L4 larval stage. Then, the plates were washed with M9 buffer (3 g/L KH2PO4, 6 g/L Na2HPO4, 5 g/L NaCl, 1 mM MgSO4) to obtain L4 larval stage worms. The obtained worms were washed three times with M9 buffer to remove excess E. coli bacteria. Then transferred them into 200 mL S-complete medium (10 mM potassium citrate, 10 ml trace metals solution, 3 mM CaCl2, 3 mM MgSO4, 5.85 g/L NaCl, 1 g/L K2HPO4, 6 g/L KH2PO4) in 1 L erlenmeyer flask. S. oneidensis MR-1 with QDs inside the cells was also added as an extra food source to constitute the food chain. The concentrations of Se and Cd in S. onedensis were 101 ± 1.7 and 257 ± 2.0 mg/g protein, respectively. The content of QDs-containing S. onedensis fed to C. elegans was 280 µg protein/mL. The erlenmeyer flask was put on a shaker with 200 rpm at 20 oC in dark.

Fluorescence microscopy observation

The fluorescence images of the bacteria and nematodes were obtained using a BX-51 microscope (Olympus Co., Japan) equipped with a 120 W mercury lamp (X-Cite 120 Q). During the observation wideband MWU2 filter (Ex 330-385 nm) and oil immersion objective (100×) were used. DP2-BSW (Olympus Co., Japan) Software was used for the image processing.

Raman measurements

The Raman spectral analysis and Raman mapping of the cells were conducted using a Thermo ScientificTM DXRTM xi spectrometer equipped with a 532 nm Ar+ ion laser and 100× objective. Silicon wafer was used to automatically calibrate the instrument wavelength. In order to improve the signal quality, the samples were placed on a tinfoil plate. The Raman mapping parameters were 0.5 μm diameter spot, 0.3 μm × 0.3 μm aperture and 10 s integration of per Raman spectrum at each location of the laser spot.

Micro-X-ray fluorescence (μ-XRF) spectroscopy analysis

The distribution of Se in C. elegans was monitored using μ-XRF spectroscopy. Before analysis, the samples were washed three times with 10 mM Tris-HCl (pH 7.6) and immediately fixed with liquid nitrogen on metal-free polyimide film, then frozen dried using vacuum freeze dryer (FreeZone 2.5, Labconco Co., USA). The μ-XRF spectroscopy analysis was performed at the 4W1B beamline of Beijing Synchrotron Radiation Facility (BSRF), which can provide 2.5 GeV storage ring and 150 to 250 mA current. The incident X-ray energy was monochromatized by W/B4C Double-Multilayer-Monochromator (DMM) at 15 keV and was focused down to 50 μm in diameter by the polycapillary lens. The two-dimensional mapping was acquired by step-mode: the sample was held on a precision motor-driven stage. The Si (Li) solid state detector was used to detect X-ray fluorescence emission lines with live time of 30 s. The data reduction and process were performed using PyMca package26.

X-ray absorption fine structure (XAFS) data collection and analysis

Prior to XAFS analysis, the samples and standards (CdSe, Se) were frozen dried using a vacuum freeze dryer (FreeZone 2.5, Labconco Co., USA). The XAFS analysis was performed at the beamline BL14W1 in SSRF (Shanghai Synchrotron Radiation Facility) with 3.5 GeV storage ring and 300 mA maximum current. The spectra were acquired at room temperature under fluorescence mode using a Si (111) double-crystal monochromator. Before the samples analysis, we first calibrated the energy using Se foil, then the standard ion chamber and Lytle-type detector were used to monitor the samples signal. The acquired XAFS data were background-subtracted, normalized, and Fourier transformed according the standard procedures with the ATHENA program27, 28.