Redox and energy homeostasis enabled by photocatalytic material-microbial interfaces

Abstract

Material-microbial interfaces offer a promising future in sustainable and efficient chemical-energy conversions, yet the impacts of these artificial interfaces on microbial metabolisms remain unclear. Here, we conducted detailed proteomic and metabolomic analyses to study the regulations of microbial metabolism induced by the photocatalytic material-microbial interfaces, especially the intracellular redox and energy homeostasis which are vital for sustaining cell activity. Firstly, we learned that the materials have a heavier weight in perturbing microbial metabolism and inducing distinctive biological pathways, like the expression of the metal-resisting system, than light stimulations. Furthermore, we observed that the materials-microbe interfaces can maintain the delicate redox balance and the energetic status of the microbial cells, since the intracellular redox cofactors and energy currencies show stable levels as naturally inoculated microbes. These observations assure the possibility of energizing microbial activities with artificial materials-microbe interfaces for diverse applications, and also, provide guides for future designs of materials-microbe hybrids to guard microbial activities.

Material-microbial interfaces are constructed by incorporating microorganisms with artificial materials for applications mostly in light/electricity-driven reduction of CO₂/N₂ gases,^1–5 energy production from wastewater,^6–8 and other photocatalytic chemical production,^9–12 offering sustainable and efficient solutions to energy-chemical conversions. Tremendous efforts have been devoted to studying the charge transfer at the abiotic-biotic interfaces^13–16 from the view of material for deepened fundamental understandings and efficiency improvements. However, recent studies have demonstrated that these active material-microbial interfaces could alter the microbial metabolism therefore resulting in higher efficiency for energy-chemical conversions¹⁷ or unique reactivities.¹⁸ Thus, understanding these intricate metabolic regulations induced by the artificial interfaces will facilitate the development of microbial-material hybrids and advance their utilities across various future applications.

The regulations of microbial metabolisms are typically investigated through various omics analyses by comparing the level of transcriptomes,¹⁹ metabolites,^{17, 19–22} and proteins,^{17, 22–24} with or without the presence of abiotic-biotic interactions. Despite the limited studies about metabolic regulations, current analyses mainly focus on product-related biochemical pathways, such as the CO₂ reduction pathway in hybrids aiming for CO₂ fixations^{22, 25}. These studies help understand the biochemical processes directly related to the target chemical chemical-energy conversion, while it is insufficient to understand the overall microbial activity which would greatly affect the system’s performance. Since the redox status and energy pool are two crucial parameters for maintaining microbial activities, it is crucial to investigate the perturbations of redox status and energy pool in microbial cells supported by these materials-microbe interfaces for better assessing the system’s potential for long-term operations.

The redox state determined by the ratio between the reduced and oxidized forms of the redox cofactor – such as the ratio of reduced nicotinamide adenine dinucleotide phosphate (NADPH) and nicotinamide adenine dinucleotide phosphate (NADP⁺) – plays an important role in cellular processes related to oxidative stress,²⁶ aging,²⁷ and cell survival,²⁸ therefore, it is essential to maintain the redox homeostasis for these cellular processes. Meanwhile, the energy currency in cells, namely the adenosine triphosphate (ATP), is involved in the majority of metabolic pathways related to chemical synthesis and energy transformation;²⁹ whereas its biosynthesis occurs from the adenosine diphosphate (ADP) and adenosine monophosphate (AMP), making the ADP and AMP also important components of energy flow in biological systems.³⁰ Typically, the energetic status of cells can be evaluated by the adenylate energy charge, [(ATP + 0.5 ADP)] / [(ATP + ADP + AMP)], and a stable adenylate energy charge is required to maintain a stable microbial activity.³¹

In this study, we conducted a detailed proteomic and metabolomic analysis of the Xanthobacter autotrophicus with the perturbation of CdTe quantum dots (QDs) under the photocatalytic condition of N₂ and CO₂ fixations. Among six deliberately designed conditions, a distinctive metabolism was observed specifically responding to the photocatalytic material-microbial interfaces with maintained redox and energetic homeostasis compared to naturally-grown microorganisms. The principal component analysis (PCA) and the pseudo-Venn diagram showed a greater influence of the material in perturbing the microbial metabolism in comparison to the light solely, for example, cation diffusion facilitators for resisting metal toxicity were particularly expressed with CdTe additions. Especially, the redox status of the microbial cells was investigated by examining the relative abundances and normalized ratio of intracellular redox couple (NADP⁺/NADPH), and redox homeostasis was maintained with the presence of photo-generated electrons on the abiotic-biotic interfaces compared to naturally-grown microbes fed with hydrogen. Simultaneously, the normalized adenylate energy charge was found to be stable regardless of the presence of materials-microbe interfaces, indicating a steady energetic status despite a relatively smaller adenylate pool of ATP, ADP, and AMP under photocatalytic condition (Scheme 1). This comprehensive study with detailed proteomic and metabolic results support the capability of material-microbial interfaces in maintaining microbial redox and energy homeostasis for adequate microbial activities, despite its distinctive metabolic behaviors. The findings in this study fill the knowledge gap regarding microbial redox status and energy pool when incorporated with material electron suppliers. Particularly, other microbial species possessing same metabolic pathways may expect similar regulations in the material-microbial hybrid systems as this work showed.

Scheme 1.

Regulations of redox balance and energetic status in response to photocatalytic material-microbial interfaces.

RESULTS AND DISCUSSIONS

Regulations of microbial metabolism

The naturally grown X. autotrophicus (ATCC 35674^™)^{32, 33} was prepared by inoculating in an inorganic minimal medium with a gas mixture of N₂: H₂: CO₂: O₂ (60: 20: 17: 3), whereas the hydrogen gas serves as the electron donor for N₂ and CO₂ fixations (treatment 0, T0 in Figure 1a). The X. autotrophicus | CdTe QDs interfaces (T3 in Figure 1a) were assembled by mixing the microbes with the commercially available CdTe QDs³⁴ (Sigma-Aldrich 777943) following the reported method,¹⁷ and the autotrophic fixation of N₂/CO₂ was driven by the photo-generated electrons on the material with a light illumination of 505 nm (0.09 mW/cm²) in a gas environment of N₂: O₂: CO₂ (79: 3: 18). Control groups that study the independent influence of materials (T1 in Figure 1a) or light (T2 in Figure 1a) to the naturally grown X. autotrophicus were assembled by mixing the microbes with QDs or exposing the microbes to light illumination as in the photocatalytic condition; the light (T4 in Figure 1a) or materials (T5 in Figure 1a) was removed independently comparing to the photocatalytic condition (T3 in Figure 1a) to further distinguish the synergistic effect of the light and material achieving by the material-microbial interfaces.

Figure 1.

Regulation of microbial metabolisms under different inoculation conditions. (a) Inoculation conditions (group T0 – T5) of the microbes for the proteomic and metabolomic analyses (1 – true, 0 – false in the table). (b) Experimental procedures of the omics studies. Principal component analyses (PCA) of (c) the proteomes and (d) the metabolomes from microbes inoculated under different conditions. The percentages demonstrated the explained variance ratios. LC, liquid chromatography; MS, mass spectrometry; PC, principal component.

Proteins and intracellular metabolites were extracted and characterized by the state-of-the-art proteomic and metabolomic techniques (see details in Supplementary Methods, Figure 1b), and the proteome was initially annotated with the UniProt database. A total of 2,743 proteins and 110 metabolites were detected and identified (Table S1 and S2). Principal component analyses (PCA) were performed on the identified proteomes (Figure 1c) and metabolomes (Figure 1d) with three principal components (PC) explaining >50% variance of the data. Proteomes in the score plot of PCA (Figure 1c) were mostly separated into distinctive groups, indicating distinguished expression of proteins under different inoculation conditions. However, the metabolomes in the score plot of PCA were separated into 3 sets – the entries with hydrogen as the electron donor (T0, T1, and T2 in Figure 1d), the photocatalytic entry with photo-generated electrons supplied by material-microbial interfaces (T3 in Figure 1d), and the others lacking electron supply (T4, T5 in Figure 1d) – which indicates that the metabolomes are quite sensitive to the sources of electron donors compared to other individual environmental stimuli like light or material additions solely.

Expression of distinctive pathways induced by material

To further distinguish the expression of distinctive pathways, pseudo-Venn diagrams were constructed based on the detected proteomes and metabolomes to visualize the overlap among the different inoculation conditions. The pseudo-Venn diagrams depicted an overlap of 97/110 (88.2%) metabolites being detected among the six different conditions (T0 – T5, Figure S1), while only 1,993/2,743 (72.7%) proteins were detected and identified in all of the 6 conditions (Figure 2a). These set differences imply distinguished expressions of proteins under different inoculation environments, consistent with the results of PCA. Particularly, the distinctively expressed proteins among groups are concentrated in the topright of the pseudo-Venn diagram where the materials were added to those groups (Figure 2a), meanwhile, there are barely any exclusively expressed proteins in groups with light as the major environmental stimulus (T2 and T5), suggesting that the material has a greater impact on regulating the expressions of proteins than light.

Figure 2.

Microbial metabolism being sensitive to the stimulus of material. (a) The pseudo-Venn diagram of the proteome with distinctively expressed proteins concentrated in groups with material addition. n/d*, undefined. (b) Schematic illustration of a metal-resisting system in biology. (c) Relative abundance of statistically different expressed proteins among groups of T0 – T5 (the corresponding p-value < 0.01* or < 0.001** from one-way ANOVA test) related to the metal-resisting system, and data for heatmap has been scaled for representative purposes. N.A., unavailable in proteomic data. (d) Bar graph of the 47 distinctively detected proteins under photocatalytic hybrid condition (T3) categorized by protein functions based on the COG database.

Then, we performed analyses of biological pathways based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, and one-way analysis of variance (one-way ANOVA) was applied to gate the statistically different expressed proteins or metabolites among groups of T0 – T5 (the corresponding p-value < 0.01* or 0.001** from one-way ANOVA unless noted). Specifically, proteins (Figure 2b) related to the resistance of heavy metals were expressed in response to CdTe QDs (Figure 2c). The CzcCBA efflux protein complex – composed of CzcA protein from the resistance-nodulation-cell division family, CzcB protein from the membrane fusion protein family, and CzcC protein from the family of outer membrane factors – is responsible for exporting heavy metals directly to the outside with the proton motive force³⁵ from the cytoplasm, the cytoplasmic membrane or the periplasm, which has also been reported for exporting cadmium ions.³⁶ In our proteome results, while the expression of CzcA was observed in most of the groups (T0 – T4 in Figure 2c), the expressions of CzcB were exclusive to conditions with CdTe materials (T1, T3, and T4 in Figure 2c), indicating a microbial response to the addition of heavy metal materials. Moreover, the CzcC was uniquely detected in the hydrogen group with additional CdTe (T1 in Figure 2c) along with the highest abundance of CzcB even compared to the photocatalytic group (T3 in Figure 2c). Since the added CdTe nanomaterials have an averaged diameter of ~ 3 nm¹⁷, the microbes could easily uptake these nanoparticles unconsciously³⁷ thereby activating the metabolic response to export these heavy metals through the CzcCBA efflux protein complex. Interestingly, the expression of CzcCBA under the photocatalytic condition (T3 in Figure 2c) was not as prominent as under hydrogen conditions (T1 in Figure 2c), indicating a possible reluctance to export these materials as they come with the benefits of photo-generated electrons. Microbial tolerance to these potentially toxic materials could be counterintuitive under abiotic-biotic systems but not unique in this particular system³⁸. For the future design of material-microbial hybrid systems, the development of biocompatible materials remains the primary approach, however, it is also possible to incorporate biological components into ‘less biocompatible’ environments if these ‘toxicants’ can be managed or needed by the microbes. For instance, the expression of cation diffusion facilitators of CzcCBA, as discussed above, enables the microbes to resist the toxicity of heavy metals like Cd²⁺, Co²⁺, Zn²⁺, and Ni²⁺, not even the photo-generated electrons produced on the materials¹⁷. Additionally, exploring possible defensing pathways in biological systems through gene databases is worthwhile when designing abiotic-biotic systems, especially for applications like wastewater treatment.

Specifically, 47 proteins are exclusively detected in the presence of photocatalytic material-microbial interfaces (T3 in Figure 1a), and the Clusters of Orthologous Genes (COGs) database was applied to categorize these proteins based on the protein function (Figure 2d). The functions of the categorized proteins range from (J) translation, ribosomal structure and biogenesis, (M) cell wall/membrane/envelope biogenesis, (O)posttranslational modification, protein turnover, chaperones, (C) energy production and conversion, (E) amino acid transport and metabolism, (F) nucleotide transport and metabolism, (H) coenzyme transport and metabolism, and (P) inorganic ion transport and metabolism. However, more than half of the proteins remain with unknown functions or are even uncategorized, likely due to limitations in gene annotation for the X. autotrophicus species. Notably, these proteins are unlikely to be expressed under natural conditions (T1 in Figure 1a) but are distinctly induced by the material-microbial interactions (T3 in Figure 1a). Though the functions of distinctively expressed proteins require further investigations, it is evident that these material-microbial interfaces did regulate microbial metabolism in a certain way. From here, we set to study the regulations of the redox balance and energy pool of the microbes in response to the artificial interfaces of microbes and materials based on KEGG pathways.

Maintained redox balance with photo-generated electrons

Since the abiotic-biotic interfaces are assumed for transferring charges between the material and the microbial moieties, the regulations of the electron transport chain and the redox balance (Figure 3a) were studied in response to these sophisticated material-microbial interfaces. NADH dehydrogenase (NuoA-N, Ndh), succinate dehydrogenase (SdhABCD), cytochrome bc1 complex (FbcH, PetAB), and cytochrome c oxidase (CcoNOPQ, CoxABC, COX11) were characterized, confirming the activation of the electron transport chain in the microbes under all inoculation conditions (T0 – T5 in Figure 1a). Expressions of four proteins – NuoE, Ndh, SdhC, and CcoN – were statistically perturbed (Figure 3b), whereas the Ndh and SdhC were undetected in some groups and maintained similar relative abundances among the detected groups. Interestingly, the relative abundance of the NuoE protein was observed to be more than doubled specifically with the presence of photocatalytic material-microbial interfaces (T3 in Figure 3b) compared to all other conditions. Though we failed to detect the reduced nicotinamide adenine dinucleotide (NADH) in the metabolome data, the relative concentrations of nicotinamide adenine dinucleotide (NAD⁺) were found to be stable (Figure S2), especially among groups of T0 – T3 with electron supplies (the corresponding p-value = 0.7424 among T0 – T3). It is possible that the upregulated proportion of the NuoE takes additional duties as assisting electron transport at the abiotic-biotic interfaces.

Figure 3.

Maintained redox balance with the photocatalytic material-microbial interfaces. (a) Schematic illustration of electron transport chain. (b) Heatmap of proteins in the electron transport chain with statistically different relative abundances (the corresponding p-value < 0.01* or < 0.001** among groups of T0 – T5, calculated from one-way ANOVA test), and data for heatmap has been scaled for representative purposes. N.A., unavailable in proteomic data. Box plots of (c) NADPH, (d) NADP⁺, the normalized ratios of (e) NADPH/NADP⁺ and (f) GSH/GSSG from microbes inoculated under different conditions (n = 6 biological replicates for T0, n = 3 biological replicates for T2, and n = 4 biological replicates for the other groups). The metabolite levels used for calculating normalized ratios were normalized to the mean peak intensity of group T0. In each box, the central line represents the median value, the box represents the upper and lower quartiles and the whiskers extend up to 1.5 times the interquartile range beyond the box range, and the black diamonds represent the outliers.

Then, we studied the redox couple of NADPH and NADP⁺, crucial in maintaining redox balance and supporting the biosynthesis of fatty acids and nucleic acids.³⁹ The relative concentrations of the NADPH (the corresponding p-value = 0.0836 among T0 – T5, Figure 3c) and NADP⁺ (the corresponding p-value = 0.0154 among T0 – T5, Figure 3d) were not statistically perturbed, especially for groups with the supply of reducing power (the corresponding p-value = 0.896 and 0.664 for NADPH and NADP⁺, respectively, among T0 – T3), however, the normalized NADPH/NADP⁺ ratio that represents the cell’s redox status varied (Figure 3e). The photocatalytic group (T3 in Figure 3f) maintained a similar normalized NADPH/NADP⁺ ratio as the groups fed with hydrogen gas (T0 – T2 in Figure 3e) (the corresponding p-value = 0.979 among T0 – T3), indicating a stable redox balance within the cells supplied with hydrogen or photo-generated electrons and being exempted from the interruption of material or light. For groups of T4 and T5, the balances of NADPH and NADP⁺ were disrupted (T4 – T5 in Figure 3e) possibly due to the lack of electron donors therefore suffering from limited growth¹⁷. Moreover, the normalized ratio of the reduced glutathione and the oxidized glutathione²⁶ (GSH and GSSG, respectively), which is sensitive to the oxidative stress,⁴⁰ was estimated and found to be stable (the corresponding p-value = 0.198 among T0 – T5, Figure 3f). Meanwhile, we observed an insignificant expression of glutathione reductase that catalyzes the interchange between the GSH and GSSH (Table S1).⁴¹ These results further confirm redox homeostasis in microbial cells with the cooperation of materials-microbe interfaces. Overall, the redox homeostasis was observed in the photocatalytic microbial-material hybrids similar to the hydrogen-fed microbes, indicating the capability of artificial abiotic-biotic interfaces for providing an adequate reducing power without perturbing the microbial redox balance.

Stable energy status with photocatalytic abiotic-biotic interfaces

In the photocatalytic microbial-material system, microbes are energized by the photo-generated electrons instead of hydrogen gas, therefore, it is necessary to investigate how would the energetic status of the microbes be changed by the switching of feeding sources. In microbial systems, the tricarboxylic acid (TCA) cycle (Figure 4a) provides the main energy source in microbial cells by harnessing the chemical energy in acetyl-coenzyme A (acetyl-CoA) to reducing power of NADH, and the generated NADH creates the necessary proton motive force for the synthesis ATP through electron transport chain (Figure 3a). For all the six experimental groups, proteins necessary for the TCA cycle were observed (Table S1), illustrating the activation of the TCA cycle. Compared to the natural condition with hydrogen (T0 in Figure 4b), the addition of material or light (T1 and T2 in Figure 4b, respectively) barely affects the expression of proteins involved in the TCA cycle. However, among the proteins with statistical differences, unique upregulations of proteins were specifically with the photocatalytic abiotic-biotic groups (T3 in Figure 4b), including 2 oxoglutarate dehydrogenase (SucAB), succinyl-CoA synthetase beta subunit (SucC), succinyl-CoA: acetate CoA-transferase (AacC), succinate dehydrogenase cytochrome b subunit (SdhC), fumarate hydratase (FumC), and malate dehydrogenase (Mdh). Noteworthily, most of these upregulated proteins are involved in processes generating NADH or ATP, and it is likely that these upregulated proteins related to ATP synthesis are to compensate for the smaller size of the adenylate pool therefore maintaining a stable energetic status (vide infra).

Figure 4.

Energy conservation in microbes. Schematic illustration of the (a) TCA cycle and the (c) ATP synthesis. Relative abundances of proteins involved in (b) TCA cycle and (d) ATP synthesis that are statistically different (the corresponding p-value < 0.01* or < 0.001** among groups of T0 – T5, calculated from one-way ANOVA test), repeated protein labeling represent different proteins with same functions sharing the same protein name and data for heatmaps has been scaled for representative purposes. Box plots of (e) ATP, (f) ADP, (g) AMP, and normalized adenylate energy charge from microbes inoculated under different conditions (n = 6 biological replicates for T0, n = 3 biological replicates for T2, and n = 4 biological replicates for the other groups). The metabolite levels used for calculating normalized ratios were normalized to the mean peak intensity of group T0. In each box, the central line represents the median value, the box represents the upper and lower quartiles and the whiskers extend up to 1.5 times the interquartile range beyond the box range, and the black diamonds represent the outliers.

Furthermore, we examined proteins catalyzing the ATP synthesis. For groups supplied with hydrogen, the expressions of the F-type ATPases (AtpA and AtpH in Figure 4c) seemed to be suppressed by the additional material or light (T1 and T2 in Figure 4c, respectively) compared to the pristine inoculation condition (T0 in Figure 4c); while these proteins were found to have higher relative abundances in groups in the absence of hydrogen (T3 – T5 in Figure 4c), especially under the photocatalytic condition (T3 in Figure 4c). Normally, a higher relative abundance of a certain protein occurs with the observation of a higher concentration of the corresponding catalyzed products. Conversely, the relative abundances of the ATP (Figure 4e), ADP (Figure 4f), and AMP (Figure 4g) were found be lower in groups with higher abundances of the F-type ATPases (T3 – T5) compared to the others (T0 – T2), together with a lower relative abundance of polyphosphate kinase (Ppk2) (Figure 4d) that involves in phosphate production for ATP synthesis⁴². The decreased AMP, ADP, and ATP could resulted from the insufficient proton gradient caused by the lack of electron donors (T4 and T5), while it is complicated why the adenylate pool has a smaller size under photocatalytic conditions (T3) whereas photo-generated electrons are presented. Considering the adenylate pool as the central pool, a shrunken pool size could be contributed by a smaller flux going into the central pool from its upstream chemical pool or a larger flux leaving from the central pool to its downstream chemical pool. In the photocatalytic condition (T3), the observed downregulation of Ppk2 could lead to a smaller influx of phosphate into the adenylate pool⁴²; meanwhile, the previous study showed increased concentrations of uracil, adenine, and guanine¹⁷, suggesting a larger efflux to the downstream formation of ribonucleic acids. The combined smaller influx and larger efflux to/from the central adenylate pool together leads to the smaller size of the adenylate pool in group T3, possibly stimulated by the photocatalytic material-microbial interfaces. Despite this, microbes under photocatalytic conditions still managed a stable energetic status (vide infra) with such a smaller adenylate pool, which is also commonly seen in other biological systems and seemingly innocuous to biological activities³¹.

To better evaluate the energetic status of the microbial populations, we then calculated the normalized adenylate energy charge, representatives of the energetic state in microbial cells. Considering the microbes under pristine inoculation condition (T0 in Figure 4h) to have the ‘usual’ energetic state of the X. autotrophicus, the normalized adenylate energy charge was found to be stable with the addition of material (T1 in Figure 4h) and the addition of light (T2 in Figure 4h). More importantly, stable adenylate energy charge (p-value = 0.1115 between T0 and T3 from twotailed t-test) was observed in microbes inoculated under the photocatalytic hybrid condition (T3 in Figure 4h), indicating a maintained energetic status for microbial activities ensured by the photogenerated electrons despite a smaller size of adenylate pool. The normalized ADP/ATP and AMP/ATP ratios were also calculated (Figure S3) and found to have a similar trend as the normalized adenylate energy charge, corresponding to the stable adenylate energy charge in the abiotic-biotic system.

CONCLUSIONS

In summary, we studied microbial responses to the X. autotrophicus | CdTe interfaces of photocatalytic CO₂/N₂ fixations via detailed proteomic and metabolomic analyses. The results from the PCA and Venn diagram depicted distinctive metabolisms of microbes undergoing different environmental stimuli involved in photocatalytic abiotic-biotic networks. Particularly, we successfully decoupled the effects of material addition and light illumination, finding a greater impact of materials in regulating microbial metabolism than light. More importantly, we examined the redox and energy homeostasis of the microbial cells with the photocatalytic material-microbial interfaces. Compared to the microbes feeding with hydrogen, the steady status of NADPH/NADP⁺ normalized ratio, GSH/GSSG normalized ratio, and normalized adenylate energy charge under photocatalytic conditions illustrate the stable redox and energy status enabled by the photocatalytic abiotic-biotic interfaces with sufficiency reducing power. We did notice a smaller adenylate pool of ATP, ADP, and AMP, resulting from the combined smaller influx and larger efflux to/from the adenylate pool. Future research could focus on the impacts of such a shrunken adenylate pool. Overall, these photocatalytic material-microbial interfaces are capable of maintaining normal cell activity with significant metabolomic regulations. Particularly, the metabolomic regulations induced by the artificial interfaces could improve microbial tolerance to material toxicity. These characteristics not only benefit future designs of abiotic-biotic hybrid systems in tackling the near energy-food-water challenges but also provide additional routes in engineering microbial metabolisms compared to traditional methodologies like genome editing, particularly comparable to biological systems possessing similar metabolic pathways discussed in this work.

METHODS

Materials and chemicals.

All materials and chemicals were used as received. The mercaptosuccinic acid functionalized CdTe QDs (777943, Lot# mkch3110) and all other chemicals were purchased from Sigma-Aldrich. All deionized (DI) water was obtained from a Millipore Millipak Express 40 system. All gases were passed through a syringe filter with 0.2 μm pore size (VWR 28145-477) to remove the potential microbes before use.

Cultivation of Xanthobacter autotrophicus.

X. autotrophicus (ATCC 35674^™) was purchased from American Type Culture Collection (ATCC). The growth protocols follow the reported procedures.^{17, 32, 33} Briefly, the X. autotrophicus was initially inoculated on the autoclaved super nutrient broth agar plate. The super nutrient broth medium was prepared with 5 g/L nutrient broth, 4 g/L yeast extract, 3 g/L NaCl, 5 g/L sodium succinate, and 15 g/L agar for plates only. Then, an individual colony was picked from the agar plate and inoculated into the super nutrient broth solution for overnight growth at 30 °C with 200 rpm stirring. The pH of the culture was then adjusted to 10 ~ 11 by adding drops of 2 M NaOH. The microbial pellet was collected by centrifugation (4,830 × g, 5 min), followed by resuspending in an autoclaved inorganic minimal medium, which contains 1 g/L K₂HPO₄, 0.5 g/L KH₂PO₄, 2 g/L NaHCO₃, 0.1 g/L MgSO₄·7H₂O, 0.0316 g/L CaSO₄, 0.0115 g/L FeSO₄·7H₂O, and 1 ml/L trace mineral mix. The trace minimal mix was prepared with 2.8 g/L H₃BO₃, 2.1 g/L MnSO₄·4H₂O, 0.75 g/L Na₂MoO₄·2H₂O, 0.24 g/L ZnSO₄·7H₂O, 0.04 g/L Cu(NO₃)2·3H₂O, and 0.13 g/L NiSO₄·6H₂O. The culture in the inorganic minimal medium was incubated (200 rpm stirring) with a gas mixture of N₂: H₂: CO₂: O₂ (60: 20: 17: 3) at 30 °C for days before use. The identity of the as-grown X. autotrophicus was confirmed by 16S rRNA sequencing conducted by Laragen Inc.

Assembly of the sample cultures.

The sample cultures of treatment 3 – 5 (T3 – T5) were prepared following the reported method.¹⁷ Briefly, the autotrophically-grown X. autotrophicus was spun down (4,830 × g, 5 min) and resuspended in the fresh inorganic minimal medium. The CdTe QDs were suspended in the inorganic minimal medium and sonicated to make a uniform suspension. Cysteine hydrochloride was added as the hole scavenger by dissolving in the inorganic minimal medium and neutralized with sodium bicarbonate. The hybrids of X. autotrophicus and CdTe QDs (T3 and T4) were prepared by simply mixing the microbial culture (0.20 final OD₆₀₀), the CdTe suspension (0.027 g/L final concentration), and the cysteine solution (0.15 g/L final concentration). The sample culture involving the deletion of materials (T5) follows the procedure here with the deletion of QDs. Groups inoculated with hydrogen (T0 – T2) were prepared by spining down (4,830 × g, 5 min) and resuspending the autotrophically-grown X. autotrophicus in the fresh inorganic minimal medium (0.20 final OD₆₀₀), especially, the group with the addition of QDs (T1) were prepared by adding CdTe QDs similar to the preparation of the photocatalytic microbe-material hybrid.

Inoculation of microbes for omics studies.

In a typical experiment, 7 ml of the prepared sample culture in triplicates was placed in a customized Vacu-Quik anaerobic jar (Almore, 15000) at 30 ℃ under 200 rpm stirring. Groups fed with hydrogen (T0 – T2) were supplied with a gas mixture of N₂: H₂: CO₂: O₂ (60: 20: 17: 3); groups without hydrogen supply (T3 – T5) were inoculated with a gas mixture of N₂: CO₂: O₂ (80: 17: 3). Groups with light supply (T2, T3, and T5) were exposed to a uniform light illumination (0.09 mW/cm²) achieved by a light-emitting diode source with a collimator lens (MIGHTEX Systems, 505 nm), and the light intensity was calibrated by a commercial photodetector (Newport 1916-R) to be 0.09 mW/cm². Groups without light supply (T0, T1, and T4) were inoculated under a dark environment. The microbial culture was inoculated under the descripted experimental conditions separately for at least 2 – 4 days, providing sufficient time for microbes to reach metabolic steady state under respective conditions (Figure S4). The microbes with hydrogen or photogenerated electrons (T0 – T3) were under the mid-log phase of growth when sampling, the microbes without electron donors (T4 and T5) were also alive thanks to the commonly-seen energy granules stored in the Xanthobacter autrotophicus species³³.

Omics analyses.

The detailed experimental procedures of proteomics and metabolomics follow the previous published methods,¹⁷ which can also be found in the “Proteomics analysis” and “Metabolomics analysis” from the Supporting Methods. The principal component analyses (PCA) were processed using the Scikit-learn library (version 0.21.3) in Python. The pseudo-Venn diagrams were produced by the Venn library (version 0.1.3) in Python. The one-way analysis of variance (ANOVA) was conducted by using the Scipy library (version 1.3.1) in Python. The heatmaps and the box plots were generated by the Seaborn library (version 0.9.0) in Python. The pathway analyses were performed with the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, and the functional annotation of proteins was carried out based on the of Orthologous Groups of proteins (COG) database. The normalized ratios reported in this work were calculated with the normalized metabolite levels to the mean values in group T0.

Supplementary Material

The Supporting Information is available free of charge.

Supporting methods of proteomics and metabolomics analysis; pseudo-Venn diagram of metabolomes; box plot of NAD⁺, box plots of normalized ratios of ADP/ATP and AMP/ATP, representative growth curves of microbes under various inoculation conditions, and additional references 1–3 (PDF)

Supporting tables of proteomic and metabolomics data (excel)

ABBREVIATIONS

QDs

quantum dots

e ⁻

electron

hv

incident photon

NADH/ NAD⁺

reduced/ oxidized nicotinamide adenine dinucleotide

NADPH/ NADP⁺

reduced/ oxidized nicotinamide adenine dinucleotide phosphate

GSH/GSSG

reduced/oxidized glutathione

NuoA-N and Ndh

NADH dehydrogenase

SdhABCD

succinate dehydrogenase

FbcH and PetAB

cytochrome bc1 complex

CcoNOPQ, CoxABC, and COX11

cytochrome c oxidase

AMP/ADP/ATP

adenosine mono-/di-/tri-phosphate

PPPi/PPi/Pi

poly-/inorganic pyro-/inorganic phosphate

AtpA-H

F-type ATP synthase

Ppa

inorganic pyrophosphatase

Ppk1,2

polyphosphate kinase

acetyl-coA

acetylcoenzyme A

Glt

citrate synthase

AcnA

aconitate hydratase

Icd

isocitrate dehydrogenase

sucAB

2-oxoglutarate dehydrogenase

SucCD

succinyl-CoA synthetase

AarC

succinyl-CoA: acetate CoA-transferase

FrcA-D

succinate dehydrogenase

FumC

fumarate hydratase

Mdh

malate dehydrogenase

TCA cycle

tricarboxylic acid cycle

M²⁺

heavy metal ion

CzcCBA

proteins from heavy metal efflux system

LC

liquid chromatography

MS

mass spectrometry

PC(A)

principal component (analysis)

one-way ANOVA

one-way analysis of variance

N.A.

not available