Dimensionality-dependent photoelectric conversion and cellular internalization
In this study, we first investigated the influence of dimensionality on photoelectric conversion and cellular internalization rates performance using CdS nanocrystals (NCs) with different dimensions. We synthesized 0D QDs, 1D NRs, and 2D NPLs following previously reports27, 28. The morphology of these nanomaterials was characterized by transmission electron microscopy (TEM), confirming the dimensions: QDs with a diameter of 4.8 nm, NRs with dimensions of 20×5 nm, and 2D NPLs (4 ML) with dimensions of 40×40×2.2 nm (Fig. 2a). UV-visible (UV-vis) spectroscopy revealed maximum absorption peaks at 424 nm, 455 nm, and 424 nm for QDs, NRs, and NPLs, respectively (Supplementary Fig. 1).
Photoelectric conversion efficiency was assessed via current-density-time (i-t) curves measured under blue light (470 nm) illumination. This method reflects the photocatalytic activity of the nanomaterials, a critical parameter in evaluating their performance. Photocurrent measurements under intermittent lighting conditions showed negligible current during “dark” periods, but significant photocurrent during “light” periods (Fig. 2b). The photocurrent data, as depicted in Fig. 2b, c, demonstrated that under blue light, NPLs exhibited a maximum photocurrent of 1.3 µA cm-2, which was 3.79 times and 1.96 times higher than that of QDs and NRs, respectively (Fig. 2c). This enhancement can be attributed to the extensive 2D surface area and strong 1D quantum confinement in NPLs, which exceed 0D QDs and 1D NRs in promoting rapid charge transfer and effective charge separation29, 30.
We next explored the uptake rate of nanomaterials by the bacterial cells, a critical step to achieve semiconductor biohybrid construction22. After introducing nanomaterials (1.5 mg L-1 of each) into a bacterial cell suspension and culturing in minimum medium (detail in the methods), cell growth was monitored by measuring the optical density at 600 nm (OD600). The growth rates were comparable across all nanomaterial types and a control group without nanomaterials, indicating good biocompatibility of the 1.5 mg L-1 0D QDs, 1D NRs, and 2D NPLs with bacterial cells (Supplementary Fig. 2). Inductively coupled plasma mass spectrometry (ICP-MS) was used to measure the residual concentration of CdS nanomaterials in the minimum medium after 3 h of culture. Surprisingly, the residual concentrations of 1D NRs and 2D NPLs were notably lower at 20.74 ppb and 13.89 ppb, respectively, compared to 84.11 ppb for 0D QDs, indicating a higher uptake of NPLs by the bacteria (Fig. 2d). We attributed this high penetration rate to the unique flat structure of NPLs, which allowed better access to bacteria due to their ultra-thin longitudinal thickness and adaptability to various bacterial surface topographies under different mechanical conditions31, 32. Taking together, we speculated the superior photocurrent and better internalization of 2D NPLs may achieved a higher light-driven chemical production by photosensitizing bacteria.
The atomic-layer thickness of 2D NPLs has been shown to control the optical and electronic properties, such as the carrier dynamics of two-dimensional NPLs33, thereby enhancing light absorption and facilitating effective charge separation34. Here, we investigated the effect of NPLs thickness (including 3, 4, and 5 ML) on photo-induced current efficiency (Fig. 3a, Supplementary Figs. 3 and 4). As shown in Fig. 2e, f, the photocurrent of 5 ML NPLs was 1.82 µA cm-2, a 32.43% increase compared to 4 ML NPLs and a 1.16-fold increase compared to 3 ML NPLs, reflecting a decrease in two-dimensional bandgap with increasing thickness, which aids photogenerated electron mobility, particularly in 5 ML NPLs35. We also investigated the effect of NPLs thickness on internalization efficiency by measuring the residual concentration of Cd2+ in the minimum medium (Fig. 2g). All thicknesses showed effective absorption, with the lowest residual concentration observed in 5 ML NPLs (8.14 ppb), albeit not significantly different from other thicknesses. Those results indicated the thicker 2D NPLs enhance photoelectrical efficiency while maintain the similar levels of bacterial cells internalization efficiency.
Enhanced charge separation efficiency with core/crown heterojunction NPLs
In addition to the electron transfer efficiency, as indicated by higher current density, the charge separation efficiency of a nanomaterial is another important factor influencing its photoelectrochemical performance. To improve charge separation efficiency and electron utilization performance, we modulated the energy level distribution by designing and constructing heterojunctions based on 5 ML NPLs (Fig. 3a). In comparison to other heterojunction nanomaterials, core-crown (CC) heterojunctions are capable of achieve efficient electron-hole separation by modulating the energy level distribution without necessitating changes to the thickness of the atomic layers or altering the optical absorption characteristics of the core NPLs. In this regard, zinc selenide (ZnSe) was chosen for its narrow band gap, and a CdS/ZnSe CC NPLs was constructed by epitaxial growth to create a type-II heterostructure.
As shown in Fig. 3b, TEM images demonstrated the rectangular-shaped morphology of the NPLs with average lateral dimensions of 45×45×2.2 nm. In-situ energy-dispersive X-ray spectroscopy (EDS) mapping (Fig. 3c) provided further evidence for a core/crown composition in the CdS/ZnSe CC NPLs, with CdS (red and green) located in the central region and ZnSe (yellow and blue) at the outer edge. Raman spectroscopy measurements analyzed, the bonding composition of the heterojunction CC NPLs in detail, and identified a characteristic peak for the Cd-S bond at ~302 cm-1 (longitudinal optics, LO). X-ray diffraction (XRD) spectra demonstrated its crystalline characteristics of the zinc blende phase (Fig. 3e). The UV-vis and photoluminescence (PL) showed spectra shown a strong absorption peak at 440 nm and green emission peak at 510 nm (Fig. 3f). The energy differences between the valence (VB) and conduction (CB) bands reported in Fig. 3g, were based on UV photoelectron spectroscopy (UPS) measurements and calculations of the absorption and emission energies. The schematic representation in the insert of Fig. 3g described the band-edge alignment for CC NPLs. In the CC NPLs, charges were separated, with the electron wave function localized in the CdS core and the hole in the ZnSe crown. This type-II band alignment drives the photogenerated excitons at the ZnSe crown to migrate towards the core/crown interface, forming a charge-transfer (CT) exciton. The observed broader full width at half maximum (FWHM, 40 nm) was a common feature of type-II nanocrystals36. Subsequently, we measured the photocurrent of CC NPLs and the residual concentration of Cd2+ in the minimum medium. As shown in Fig. 2e, f, the photocurrent of CC NPLs is 2.63 µA cm-2, increased by 44.51% compared to 5 ML NPLs (1.82 µA cm-2), which illustrates that the core-crown heterojunctions (CdS/ZnSe CC NPLs) improve the charge separation efficiency. The residual Cd2+ concentration of CC NPLs (9.71 ppb) after inoculated with bacteria was similar to 5 ML NPLs (8.14 ppb).
Improved bioproduction of V. natriegens via CC heterojunction NPLs biohybrids
V. natriegens, recognized as a next-generation host for biotechnology with exceptionally high growth (doubling time of less than 10 minutes) and substrate consumption rate37, has demonstrated the capability for extracellular electron transfer which could facilitate the transfer of photo-electrons from semiconductor to bacterial cells in biohybrid systems38. The bioproduction of BDO is dependent on the supply of reducing power, thus serving as an excellent platform for evaluating the energy efficiency of biohybrids. Thus, the pET-RABC plasmid39, which contains the BDO biosynthetic pathway, was introduced into V. natriegens, resulting in strain XG211.
To test the hypothesis that enhanced internalization coupled with higher photocurrent of CC NPLs could improve bacteria photosensitization, we added the nanomaterials to the cell suspension (OD600 ~ 2.0) of strain XG211 in the minimum medium with 4 g L-1 glucose, electron sacrificial agent (cystine) and mediator (flavin mononucleotide, FMN). After incubation for 1 h under blue light (470 nm, 5 mW cm-2), the BDO production of strain XG211 were measured using gas chromatography (GC). We found that BDO production of strain XG211 with CC NPLs nanomaterials under illumination were all higher than their counterpart in absence of illumination or nanomaterials (Fig. 4b). Notably, BDO production in the strain with CC NPLs was superior to all other nanomaterials, achieving a final titer of 1.68 g L-1 (Fig. 4b), showing 1.28-fold increase compared with 5 ML NPLs without heterostructure (1.31 g L-1).
Moreover, a series of control experiments were conducted to compare the light-driven BDO production with variations in nanomaterial dimensions. Initially, the BDO production of strain with 2D (4 ML) NPLs was 1.05 g L-1, showing 35.4% and 36.3% increase when compared to 1D NRs and 0D QDs under illumination, respectively (Fig. 4a). Furthermore, 5 ML NPLs biohybrids achieve higher BDO production compared to 4 ML NPLs biohybrids, while the 3 ML NPLs biohybrids (0.94 g L-1) was similar to 4 ML NPLs biohybrids. As aforementioned mentioned, 2D nanomaterials offer better internalization coupled with enhanced photocurrent compared to 0D and 1D nanomaterials, resulting in higher BDO production. This trend is also the same regarding the layers of the nanomaterials. Taking together, our results demonstrated that all low-dimensional CdS nanomaterials were capable of photosensitizing V. natriegens, with 2D CC NPLs showing the highest efficiency in light-driven chemical production.
After confirming the BDO production and the superior catalytic performance in 2D CC NPLs the biohybrids, we hypothesize a direct interaction at the nanomaterials-bacteria interface. To this end, we prepared cross-sectional slices of CC NPLs-XG211 biohybrids samples using microtome sections. High angle angular dark field-scanning transmission electron microscopy (HAADF-STEM) image and EDS mapping showed that internalized CC NPLs were composed of Cd, S, Zn and Se, with highly correlated locations (Fig. 4c and Supplementary Fig. 5). Those results clearly revealed the successful transportation of CC NPLs into the cytoplasm of V. natriegens. Previous studies have reported that intracellular photosensitizers enhanced the efficiency of photoelectron transfer and energy transduction by avoiding the energy loss during photoelectrons transmembrane transfer40, 41. We studied the charge-transfer kinetics at this internalized CC NPLs-XG211 biohybrids interface. The photo-excited electrons lifetime and photocurrent of both CC NPLs-XG211 biohybrids and pure CC NPLs were measured (Fig. 4d). The average photogenerated state lifetime of biohybrids was only 0.26 ± 0.01 ns, showing a 280-fold shorted compared with the CC NPLs alone (72.8 ± 5.2 ns) (Fig. 4e). Similarly, upon the addition of bacteria, the photocurrent of biohybrid systems decreased by about 69.96% compared to that of CC NPLs alone (Fig. 4f). These results indicated that the assembly of CC NPLs into bacteria results in rapid photo-induced charge transfer between CC NPLs and the bacterial cells.
Photoelectron-induced regulation of energy metabolism in biohybrids
The semiconductor harvested light for suppling reducing equivalent instead of sugar oxidation for microbial cells in the biohybrid systems, therefore increasing the carbon yield by reducing/eliminating carbon loss during chemical production (Fig. 5a). We firstly detail analyzed the biomass and glucose-to-BDO of strain XG211 with and without CC NPLs under illumination. The biomass of illuminated biohybrid systems was only slightly higher (~ 28.9% improvement) than all other conditions (Fig. 5b). However, the carbon yield of illuminated biohybrid systems reach 0.457 g g-1, showing a 2.69-fold increase compared with XG211 with no nanomaterials, and a 2.01-fold increase compared to its counterpart under dark condition (Fig. 5c). Additionally, the production of acetoin (Fig. 1b), a direct precursor of BDO, was increased by 96.8% under light conditions (Supplementary Fig. 6). The mechanism of this enhanced production yield was further studied as followings.
We initially measured the changes of intracellular energy pools including NAD(P)+, NAD(P)H, ADP and ATP. The NADH/NAD+ ratio, represent the redox energy state of cells, in illuminated biohybrid systems surpassed all other conditions, with a ratio of 0.877 (Fig. 5d). This ratio indicated a 3.06-fold increase compared with its counterpart with no illumination and a 1.19-fold increase compared with bacterial system in absence of CC NPLs (Fig. 5d). While, the NADPH/NADP+ ratio of illuminated biohybrid systems was 23% higher than that of dark condition and was 16% higher than that of bacterial system (Fig. 5d). The highest intracellular ATP concentration was also achieved in the illuminated biohybrid systems, increased 24% compared with the dark condition (Fig. 5e). Those results confirmed two-dimensional semiconductor biohybrid systems absorbed light to power the bacterial energy metabolism, thereby promoted the flux of BDO synthetic pathway which required quantity of reducing energy.
To reveal the impacts of intracellular CC NPLs on bacterial metabolism during illumination, we performed metabolomics and transcriptomic analysis. We collected biohybrid samples with and without illumination for targeted metabolite quantification analysis using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Central metabolic pathways such as glycolysis, tricarboxylic acid (TCA) cycle and pentose phosphate pathway (PPP), which were primary sources of reducing power and ATP, were analyzed. As shown in Fig. 5f, the intracellular concentration of key metabolites such as 6-phosphate glucose (G6P), fructose 1,6-bisphosphate (FBP), 3-phosphoglycerate (3PG) and phosphoenolpyruvate (PEP) in the glycolysis, as well as Acetyl-CoA (AcCoA) and malate (MAL) in the TCA cycle, were increased under light compared to dark condition. Additionally, the concentration of 6-phosphogluconate (6-GPC) in the PPP pathway was also higher under light than dark condition (Fig. 5f). Those results suggested the nanomaterials induced photoelectrons likely improve the activity of central metabolism, which consistence with previous reports19, 42.
To explored global regulation of bacterial cells by illuminated nanomaterials, we further performed RNA-seq to compare gene expression and transcription activation at genome-wide level. Comparing biohybrid with versus without illumination, we identified 28 genes with significantly upregulation and 28 downregulation (log2 (fold change (FC)) > 1.0 or < -1.0 and P < 0.01) (Fig. 5g, h). Gene ontology (GO) term analysis of the upregulated genes highlighted a common feature: “sulfide metabolism” or “flavin binding” (Supplementary Fig. 7), which likely related to electron acceptors or mediators participating in bacterial electron transfer processes. As shown in Fig. 5g, h, the upregulated genes including those encoding proteins involved in cellular energy metabolism with illumination, such as the oxidative respiratory chain (Complex Ⅰ [NADH-quinone reductase], Complex Ⅱ [fumarate reductase], Complex Ⅳ [cytochrome ubiquinol oxidase]), FAD [FMN transferase] and RNF [RNF complex, electron transport complex], as well as the ATPase [ATP synthase]. Remarkably, 10 out of 28 upregulated genes were involved in thiamine biosynthesis, for production of TPP (the active form of thiamine), a key cofactor for enzymes involved in central metabolic pathway such as transketolase and dehydrogenase, which participate in PPP, TCA cycle and the link between glycolysis and TCA cycle43-45. Those enzymes are vital for the production of NADH, NADPH and ATP. The above up-regulation genes were further confirmed by the real-time fluorescence quantitative PCR (qRT-PCR) results, which consistence with the RNA-seq results (Supplementary Fig. 8).
Taking together, those results suggesting the illuminated biohybrid systems not only provide extra-energy from light but also promote both carbon metabolism and energy metabolism of bacteria, notably increasing carbon yield by reducing or eliminating carbon loss during chemical production.
Verification and application of photoelectron-induced regulation in biohybrids
In our biohybrid systems, semiconductors capture light to supply reducing equivalents for cellular metabolism, significantly enhancing the thiamine/TPP synthetic pathway. This pathway typically boosts the native energy machinery (central metabolism) to generate more reducing power. Observations of these effects led us to hypothesize that converting photoelectrons into reducing power might require the mediation of TPP. To explore this hypothesis, we analyzed the rate of NADH generation from CC NPLs with and without TPP under in vitro conditions (Fig. 6a, b). In a typical experiment, 1.5 mg mL-1 CC NPLs, 1 mM NAD+ with or without TPP was inoculated under 470 nm light (5 mW cm-2). NADH regeneration was monitored by the absorption at 340 nm (Abs 340 nm) and the final NADH product was confirmed by nuclear magnetic resonance (NMR, Fig. 6b). Adding TPP resulted in a 4.45-fold increase in NADH production under light conditions compared to controls without TPP, whereas NADH production remained unchanged under dark conditions (Fig. 6c). These findings were corroborated by NMR, affirming that the increase in Abs 340 nm was indeed attributable to heightened NADH production (Fig. 6b).
Additionally, the presence of thiamine (vitamin B1), commonly used as a substrate in cultures for TPP biosynthesis, led to a 9.45-fold increase in NADH production under light conditions, with no improvement noted under dark conditions (Supplementary Fig. 9). Based on these results and corresponding in vitro experiments, both TPP and thiamine not only enhance reducing power production from photoelectrons but also boost the native central metabolism. Recent studies also suggest that TPP/thiamine promotes NADH production and electron transfer in two bacterial species co-cultures46. We speculated the photoelectron transfer and solar-to-chemical efficiency in biohybrid systems might enhanced by the addition of TPP/thiamine.
To investigate further, we cultured biohybrid systems with varying concentrations (0, 0.3, 0.6, 1, 3, 6 mM) of TPP or thiamine, analyzing the ratio of BDO production and carbon yield of illuminated biohybrids compared to control bacteria (Fig. 6d-6f), and also compared biohybrid performance under light versus dark conditions (Fig. 6g-6i). The ratio of BDO production and carbon yield of illuminated biohybrids increased by 11.73% and 8.57% respectively with 0.6 mM TPP compared to the conditions without TPP, showing a decrease when TPP concentrations exceeded 1 mM (Fig. 6e, f). Similarly, the ratio of BDO production and carbon yield of the biohybrid systems under light increased by 22.75% and 24.02% with 0.6 mM TPP compared to conditions without TPP, decreasing with higher TPP concentrations (Fig. 6h, i). However, the growth of biomass was increased 38.09% and 21.29% with 6 mM TPP under light and dark compared to the control without TPP, respectively (Supplementary Fig. 10). Similar results were observed with thiamine, where both BDO production and carbon yield increased with 0.6 mM thiamine (Supplementary Fig. 11). These results indicate that TPP/thiamine indeed promotes solar-to-chemical efficiency, but the concentration is critical for optimizing carbon partitioning between desired chemicals and biomass growth.
Verification and application of photoelectron-induced regulation in biohybrids
In our biohybrid systems, semiconductors capture light to supply reducing equivalents for cellular metabolism, significantly enhancing the thiamine/TPP synthetic pathway. This pathway typically boosts the native energy machinery (central metabolism) to generate more reducing power. Observations of these effects led us to hypothesize that converting photoelectrons into reducing power might require the mediation of TPP. To explore this hypothesis, we analyzed the rate of NADH generation from CC NPLs with and without TPP under in vitro conditions (Fig. 6a, b). In a typical experiment, 1.5 mg mL-1 CC NPLs, 1 mM NAD+ with or without TPP was inoculated under 470 nm light (5 mW cm-2). NADH regeneration was monitored by the absorption at 340 nm (Abs 340 nm) and the final NADH product was confirmed by nuclear magnetic resonance (NMR, Fig. 6b). Adding TPP resulted in a 4.45-fold increase in NADH production under light conditions compared to controls without TPP, whereas NADH production remained unchanged under dark conditions (Fig. 6c). These findings were corroborated by NMR, affirming that the increase in Abs 340 nm was indeed attributable to heightened NADH production (Fig. 6b).
Additionally, the presence of thiamine (vitamin B1), commonly used as a substrate in cultures for TPP biosynthesis, led to a 9.45-fold increase in NADH production under light conditions, with no improvement noted under dark conditions (Supplementary Fig. 9). Based on these results and corresponding in vitro experiments, both TPP and thiamine not only enhance reducing power production from photoelectrons but also boost the native central metabolism. Recent studies also suggest that TPP/thiamine promotes NADH production and electron transfer in two bacterial species co-cultures46. We speculated the photoelectron transfer and solar-to-chemical efficiency in biohybrid systems might enhanced by the addition of TPP/thiamine.
To investigate further, we cultured biohybrid systems with varying concentrations (0, 0.3, 0.6, 1, 3, 6 mM) of TPP or thiamine, analyzing the ratio of BDO production and carbon yield of illuminated biohybrids compared to control bacteria (Fig. 6d-6f), and also compared biohybrid performance under light versus dark conditions (Fig. 6g-6i). The ratio of BDO production and carbon yield of illuminated biohybrids increased by 11.73% and 8.57% respectively with 0.6 mM TPP compared to the conditions without TPP, showing a decrease when TPP concentrations exceeded 1 mM (Fig. 6e, f). Similarly, the ratio of BDO production and carbon yield of the biohybrid systems under light increased by 22.75% and 24.02% with 0.6 mM TPP compared to conditions without TPP, decreasing with higher TPP concentrations (Fig. 6h, i). However, the growth of biomass was increased 38.09% and 21.29% with 6 mM TPP under light and dark compared to the control without TPP, respectively (Supplementary Fig. 10). Similar results were observed with thiamine, where both BDO production and carbon yield increased with 0.6 mM thiamine (Supplementary Fig. 11). These results indicate that TPP/thiamine indeed promotes solar-to-chemical efficiency, but the concentration is critical for optimizing carbon partitioning between desired chemicals and biomass growth.