Redox signaling-driven modulation of microbial biosynthesis and biocatalysis – Nature Communications

Materials

Germanium oxide (GeO₂, 99.99%), zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O), manganese nitrate (Mn(NO₃)₂), manganese powder, manganese(II) chloride (Mn(Cl)₂), manganese dioxide (MnO₂), iron (III) chloride hexahydrate (FeCl₃·6H₂O), iron(II) chloride (FeCl₂), O-Phenanthroline were purchased from Sinopharm Chemical Reagent Co., Ltd. Concentrated nitric acid (HNO₃), concentrated sulfuric acid (H₂SO₄), sodium hydroxide (NaOH), lactate, sodium phosphate dibasic heptahydrate (Na₂HPO₄·7H₂O), potassium dihydrogen phosphate (KH₂PO₄), sodium chloride (NaCl), ammonium chloride (NH₄Cl), sodium fumarate, humic acid (HA), anthraquinone-2,6-disulfonate (AQDS), acetonitrile, methanol, chloroform, glutathione disulfide (GSSG) and glutathione (GSH) were purchased from Aladdin Biochemical Technology Co., Ltd. Hydroxylamine hydrochloride, sodium acetate, acetic acid, ammonium iron (II) sulfate (Fe(NH₄)₂·(SO₄)₂·6H₂O) were purchased from Macklin Biochemical Co., Ltd. Glycerin, protonophore carbonyl cyanide m-chlorophenyl hydrazone (CCCP), rotenone, formamide were purchased from Sigma-Aldrich Trading Co., Ltd. Peptone, yeast extract, paraformaldehyde, 10% SDS solution, agar, PBS buffer (PH = 7.2–7.4), Tris-HCl buffer solution (PH = 7.5) were purchased from Solarbio Science & Technology Co., Ltd. The bacteria strains S. putrefaciens (BNCC 337021) and R. palustris (BNCC 232041) were obtained from Bnbio Co., Ltd. G. Soli (CCTCC AB 2014145) was provided by China Center for Type Culture Collection. The FISH probes for S. putrefaciens, R. palustris and G. Soli with the sequences of 5′-FAM- AGCTAATCCCACCTAGGTTCATC-3′, 5′-FAM-AGCTAATCCCACCTAGGTTCATC-3′ were synthesized by Shanghai Sangon Biotech Co., Ltd. NADP⁺/NADPH Assay Kit with WST-8, NAD⁺/NADH Assay Kit with WST-8 and Enhanced ATP Assay Kit were purchased from Beyotime Biotech. Inc. The ultra-pure water was obtained using a Millipore water purification system.

Characterization

The morphology of ZGO:Mn nano verifier was measured by a transmission electron microscope (TEM) (JEOL, JEM-2100, Japan) with a working voltage of 200 kV. Powder X-ray diffraction (XRD) patterns of ZGO:Mn nano verifier was measured on an X-ray diffractometer (Burker, D8 Advance, Germany) with Cu-Kα radiation (λ = 1.5406 Å) to determine the crystal structure. Absorbance was measured and recorded on a BioTeK Synergy H1 Hybrid Reader (BioTek, USA). Persistent luminescence performance of ZGO:Mn nano verifier was measured on a Hitachi FL4600 fluorescence spectrometer (Hitachi, Japan). The morphologies of S. putrefaciens and R. palustris were observed with a field emission scanning electron microscope (SEM) (Zeiss, SIGMA,Germany). Microbial consortia were imaged on the Olympus Fluoview FV3000 confocal microscope (Olympus, Japan). Current intensity of microbial cells was measured by a digital source meter (Keysight B1500A, Germany) connected to a probe station (PRCBE LAB, China). Fluorescence-activated cell sorting (FACS) was performed by BD FACSAria III Cell Sorter (BD Biosciences, USA). The Agilent 1260 Infinity HPLC system (Agilent, USA) was used to measure the concentration of lycopene.

Scanning electron microscopy characterization

The morphologies of S. putrefaciens signal router and R. palustris bio-actuator were observed with a field emission scanning electron microscope (Zeiss, Germany). Specifically, bacteria solution with OD₆₀₀ of 0.6 was centrifuged and then washed twice with ultra-pure water. The washed bacteria were fixed in 10 mL 4% paraformaldehyde for 1.5 h. After the fixation process, the bacteria were washed twice with the rotation speed of 4000 rpm. Subsequently, 5 mL of 25, 50, 75, 100% ethanol solution was used in order to dehydrate the bacteria. Each dewatering time was 15 min. Finally, the dehydrated bacteria were centrifuged and dissolved in anhydrous ethanol for preparing SEM samples.

Performance measurements of the ZGO:Mn nano verifier

To investigate the relationship between the persistent luminescence intensity and Fe species concentration, standard solution with different Fe²⁺/Fe_total ratio was prepared with FeCl₂ and FeCl₃·6H₂O (Fe_total was fixed at 2 mM, and the final concentration of ZGO:Mn nano verifier was 0.5 mg/mL). Then, the persistent luminescence intensity of the different solution was measured with a fluorescence spectrometer (Hitachi, FL4600, Japan). The excitation wavelength was 254 nm and the data mode was set as phosphorescence. Similarly, we measured the persistent luminescence intensity upon the addition of Mn (0.110 mg/mL), MnCl₂ (2 mM), MnO₂ (0.174 mg/mL), GSSG (2 mM), GSH (2 mM), supernatant and medium with fluorescence spectrometer (Hitachi, FL4600, Japan).

Monitoring of the Fe species in biological LAN with the nano verifier

To have a better understanding of the Fe signal transduction between S. putrefaciens and R. palustris, ZGO:Mn nano verifier was adopted. Specifically, ZGO:Mn with final concentration of 0.5 mg/mL was added into 5 mL sterilized tubes containing 2 mL LB medium and 2 mM Fe³⁺. Subsequently, 20 uL S. putrefaciens and 20 uL R. palustris with the OD₆₀₀ around 0.8 was added. Then, the centrifuge tubes with mixed bacteria solution were placed into a shaker with a speed of 220 rpm at 30 °C. The persistent luminescence intensity of the mixed solution was measured at different culture time with a fluorescence spectrometer (Hitachi, FL4600, Japan). The excitation wavelength was 254 nm and the data mode was set as phosphorescence.

Cell viability tests

The colony-forming unit (CFU) was performed to determine the viability of S. putrefaciens signal router and R. palustris bio-actuator after incubation with the optical ZGO:Mn nano verifier. Specifically, diluted S. putrefaciens-ZGO:Mn and R. palustris-ZGO:Mn co-culture solution at different culture times was transferred into the Luria-Bertani (LB) solid medium respectively. After 24 h growth at 30 °C, white circular colonies can be observed. The CFU per milliliter can be determined by counting the colonies.

Fe signal transduction and bacterial growth model in biological LAN

Fe redox signal transduction processes in the biological LAN were modelled with Simulink 9.7 (R2019b). The systems of ordinary differential equations below were solved using ode45 solver. The mathematical model here can be used to investigate in silico the redox signal transduction kinetic behaviors in continuous cultures. In the constructed model, it is supposed that there are no delays in responses to signals to simplify the actual system. In addition, 1 OD was assumed to be 10⁹ cells. The model parameters were calculated using a combination of the experimental data and the model.

Specifically, the concentrations of Fe³⁺ and Fe²⁺ were calculated based on the recorded luminescence intensity of the verifier according to the following equation (Supplementary Fig. 9d) in which Fe_Total was 2 mM:

Then, based on the calculated concentrations, the Fe²⁺ signal transduction from Fe³⁺ in bare S. putrefaciens system was simulated with Eq. (2) to fit and abstain the constants:

Similarly, the Fe³⁺ signal transduction from Fe²⁺ in bare R. palustris system was simulated with Eq. (3) to fit and abstain the constants:

Based on the above models of the bare S. putrefaciens and bare R. palustris system, the redox signal transduction and bacterial growth models of the co-culture system were established. Considering that S. putrefaciens transduces Fe³⁺ signal into Fe²⁺ while the Fe²⁺ can be transduced back to Fe³⁺ by R. palustris, a redox cycle module was introduced in the co-culture model to simulate Fe redox communication. The initial Fe³⁺ concentration was set at 2 mM which is consistent with the experimental data. In addition, Fe redox communication influence on bacterial growth was also considered. Overall, the differential equations were obtained as bellowing:

S. putrefaciens density in co-culture system:

R. palustris density in co-culture system:

Fe²⁺ concentration in co-culture system:

Bacteria culture

S. putrefaciens, R. palustris, S. putrefaciens–R. palustris co-culture and S. putrefaciens–R. palustris–G. Soli microbial consortia were all cultured in Luria-Bertani (LB) medium that contains (per litre) 10.0 g peptone, 5.0 g yeast extract and 5.0 g NaCl. G. Soli was cultured in N₂-flushed LM medium that contains (per litre) 2.0 g lactate, 0.2 g yeast extract, 12.8 g Na₂HPO₄·7H₂O, 3 g KH₂PO₄, 0.5 g NaCl and 1.0 g NH₄Cl. Additional sodium acetate (20 mM) as the electron donor and sodium fumarate (50 mM) as the electron acceptor were added into the above LM medium. All the bacteria were cultured at 30 °C with a shaking rate of 220 rpm. R. palustris was photosynthetic bacteria and a full-spectrum LED (100 W) was used as its light source.

Preparation of the optical ZGO:Mn verifier

The optical ZGO:Mn verifier was synthesized by a hydrothermal method. Typically, a mixture solution was obtained by dissolving 0.01 mmol Ga(NO₃)₃, 2 mmol Zn(NO₃)₂, 0.005 mmol Mn(NO₃)₂ and 300 µL HNO₃ into in 11 mL deionized water. Then 1 mmol Na₂GeO₃ solution was prepared by dissolving GeO₂ powder in 2 mol L^–1 NaOH solution. Subsequently, the prepared Na₂GeO₃ solution was added drop by drop into the above mixture solution. Ammonium hydroxide (28 wt%) was used to adjust the pH value of the mixture solution to 9.5. Then, the reaction was left at room temperature under stirring for 1 h. After that, the solution was transferred into a Teflon-lined autoclave and reacted at 220 °C for 6 h. The resultant ZGO:Mn nanorods were then obtained by centrifugation and washed three times with deionized water.

Construction of redox based biological LAN

The initial inoculation of S. putrefaciens and R. palustris were cultured in LB medium. To construct redox based biological LAN, S. putrefaciens signal router, optical verifier and R. palustris bio-actuator were introduced. Specifically, 1% of the seed culture of S. putrefaciens and R. palustris with initial OD₆₀₀ around 0.8 were transferred respectively into sterilized centrifuge tubes with LB medium. Then, optical ZGO:Mn verifier (0.5 mg/mL) and Fe³⁺ (2 mM) were added. The biological LAN including the S. putrefaciens–G. Soli dual-band router was constructed as the above procedure except for the addition of G. Soli.

Redox signal verification in the biological LAN

The redox signal transduction can be monitored in real-time with the optical ZGO:Mn verifier. Briefly, the persistent luminescence intensity of the mixed solution at different time was measured with a Hitachi FL4600 fluorescence spectrometer (Hitachi, Japan). The excitation wavelength was 254 nm and the data mode was set as phosphorescence.

Construction of ΔCymA and ΔMtrC mutants

Mutants with MtrC-encoding gene deleted (ΔMtrC) and CymA-encoding gene deleted (ΔCymA) were constructed respectively as previous report. Specifically, an upstream and downstream fragment were amplified separately from the S. putrefaciens genomic DNA by PCR using the primer pair Mtrc-UP-smaI-F/Mtrc-UP-R, and the pair Mtrc-down-F/Mtrc-down-kpnI-R, respectively (Table S1). Then, the PCR products were purified using a QIAquick Gel Purification Kit (QIAGEN Inc.) and cloned into the SmaI/KpnI sites of pRE112 vector. Subsequently, the recombinant plasmid transformed in E. coli S17-1/λpir was transferred into S. putrefaciens via the E. coli–Shewanella conjugation. Finally, the Mtrc deletion mutants were screened and its genotype was confirmed by PCR using the primer pair Mtrc-JD-F/Mtrc-JD-R (Table S1) and DNA sequencing. Similarly, the CymA deletion mutants can be constructed according to the above protocol with its own primer pair (Table S1).

o-Phenanthroline spectrophotometric method for Fe²⁺ quantification

The well-known o-phenanthroline spectrophotometric method was adopted to quantify the concentration of Fe²⁺. Typically, 0.7 g Fe(NH₄)₂·(SO₄)₂·6H₂O was dissolved in 1 M H₂SO₄ solution to prepare the Fe²⁺ standard solution (100 mg·L^–1). 20 μL 10% (w/w) hydroxylamine hydrochloride solution, 50 uL of o-Phenanthroline (0.5 wt%) and 50 uL of sodium acetate buffer (pH = 5) were added into a 96-well plate. Subsequently, 0, 2, 4, 6, 8, 10, and 12 µL Fe²⁺ standard solution was added respectively into the 96-well plate. Each well was fixed to 200 µL with ultra-pure water. The above 96-well plate was left at room temperature for 15 min and the absorbance was measured at 510 nm to obtain the standard curve. Subsequently, the Fe²⁺ concentrations of different samples were measured. Specifically, 50 uL of o-Phenanthroline (0.5 wt%) and 50 uL of sodium acetate buffer (pH = 5) were added into a 96-well plate. Then, sample solution was added into the 96-well plate. Each well was fixed to 200 µL with ultra-pure water. The above 96-well plate was left at room temperature for 15 min and the absorbance was measured at 510 nm. Hence, the concentration of Fe²⁺ can be quantified according to the obtained standard curve.

Confocal fluorescence microscopy imaging

S. putrefaciens–R. palustris co-culture and S. putrefaciens–R. palustris–G. Soli microbial consortia were imaged with a Olympus Fluoview FV3000 confocal microscope (Olympus, Tokyo, Japan). Specifically, 10 mL cultured bacteria were first washed twice and resuspended in 4% paraformaldehyde solution for 1.5 h to fix the bacteria. The samples were then washed twice with PBS and resuspended in ethanol-PBS (V/V = 1:1) solution. Subsequently, the samples were storage at –20 °C for over 24 h. Subsequently, the samples were washed twice and resuspended in a hybridization buffer (30% formamide, 0.01% SDS, 20 mM Tris and 0.9 M NaCl). The FISH probes of S. putrefaciens (5′-FAM-AGCTAATCCCACCTAGGTTCATC-3′) and R. palustris (5′-Cy5-CCTCTGACTTAGAAACCCGC-3′) were then added. Both the FISH probes had a final concentration of 5 μg/mL. The mixed solution was then incubated overnight at 50 °C. After the hybridization, the samples were washed with washing buffer (0.01% SDS, 20 mM Tris and 0.9 M NaCl). Finally, the samples were prepared for the confocal imaging.

Fluorescence-activated cell sorting

As described above, fixed bacterial samples of S. putrefaciens–R. palustris co-culture and S. putrefaciens–R. palustris–G. Soli microbial consortia were resuspended in a hybridization buffer (30% formamide, 0.01% SDS, 20 mM Tris and 0.9 M NaCl). Then, The FISH probe of R. palustris (5′-Cy5-CCTCTGACTTAGAAACCCGC-3′) was added into the above prepared solution with a final concentration of 5 μg/mL. The mixed solution was then incubated overnight at 50 °C. After the hybridization, the samples were washed with washing buffer (0.01% SDS, 20 mM Tris and 0.9 M NaCl). Finally, the samples were prepared for flow cytometry to isolate the R. palustris bio-actuator cells.

Lycopene extraction and titration

The lycopene yields at different conditions were measured. Specifically, the bacteria cells cultured for 96 h at different conditions were harvested by centrifugation at 5000 × g for 5 min. The collected bacteria were washed twice with 1.0 g L^–1 NaCl. The cells were dried at 70 °C overnight to remove the water. Subsequently, lycopene in R. palustris bio-actuator was extracted with 3 mL mixed reagent of n-hexane and methanol (1:1 v/v). The mixed solution was vortexed for 5 min and then centrifugated at 7500 × g for 10 min at 4 °C. Finally, the exacted lycopene in supernatant was collected. The concentration of lycopene was measured by the Agilent 1260 Infinity II HPLC system (Agilent, Palo Alto, CA, USA). This involved a C18 column (4.6 × 100 mm, Agilent) with the mobile phase consisted of acetonitrile/methanol/chloroform (42.5:42.5:15 v/v/v). The flow rate was set at 1.0 mL min^–1 and the injection amount was set at 10 uL. The final concentrations of CCCP, rotenone, AQDS and HA used in this manuscript are 1 μM, 100 μM, 250 μM and 0.5 mg/mL respectively.

Determination of cellular redox state and ATP concentration

R. palustris bio-actuator was isolated from the S. putrefaciens–R. palustris co-culture and S. putrefaciens–R. palustris–G. Soli microbial consortia through flow cytometry by labeling FISH probe of R. palustris (5′- Cy5-CCTCTGACTTAGAAACCCGC-3′). The collected R. palustris bio-actuator through fluorescence-activated cell sorting were centrifuged and washed with PBS solution (0.01 M, PH = 7.2–7.4). In particular, NAD⁺/NADH and NADP⁺/NADPH ratios were measured with purchased NAD⁺/NADH Assay Kit with WST-8 and NADP⁺/NADPH Assay Kit with WST-8. According to the protocol, the NAD⁺/NADH and NADP⁺/NADPH ratios were determined with a colorimetric assay under a microplate reader by detecting wavelength of 450 nm. For ATP concentration measurements, an Enhanced ATP Assay kit was adopted. The ATP concentration can be obtained by detecting chemiluminescence with a luminometer plate reader.

Shake-flask fermentation

Shake-batch fermentations were conducted in 50.0 mL tube with the working volume of 30.0 mL. Firstly, S. putrefaciens and R. palustris were cultured in O₂-free LB medium respectively. Then, 1% (v/v) seed culture of S. putrefaciens and R. palustris solution with the culture OD₆₀₀ around 0.8 were transferred respectively into the tubes containing 30 mL LB medium with or without 2 mM Fe³⁺. The fermentation temperature was set at 30 °C and the agitation speed was controlled at 220 rpm. In addition, a full-spectrum LED (100 W) was used for external light illumination.

Fed-batch fermentation

Fed-batch fermentations were conducted in 5.0 L fermenter with the working volume of 3.0 L. Firstly, S. putrefaciens and R. palustris were cultured in O₂-free LB medium respectively. Then, both the S. putrefaciens and R. palustris solution with the culture OD₆₀₀ around 0.8 was transferred into sterilized culture bottle with 400 mL O₂-free LB medium containing 2 mM Fe³⁺. The S. putrefaciens-R. palustris co-culture seed was then incubated at 30 °C with a shaking rate of 220 rpm for 96 h, and a full-spectrum LED (100 W) was used as its light source. Subsequently, 1% (v/v) seed culture was transferred into sterilized fermenter, and the fed-batch fermentation started with feeding. An anaerobic condition inside fermenter was created by flushing argon gas until concentration of dissolved oxygen dropped to zero. In order to maintain anaerobic conditions inside fermenter, the feed medium (LB medium containing 2 mM Fe³⁺) was also flushed with argon gas before being fed into fermenter. 3 g/L HCl and 3 g/L NaOH were used to control the pH of bacterial fermentation broth at ~7.0. The fermentation temperature was set at 30 °C and the agitation speed was controlled at 220 rpm. In addition, a full-spectrum LED (100 W) was used for external light illumination.

16S rRNA gene sequencing

Total genomic DNA was extracted from the S. putrefaciens–R. palustris co-culture and S. putrefaciens–R. palustris–G. Soli microbial consortia. Three independent biological replicates were performed. Specifically, V3 and V4 regions of 16S rRNA were amplified by PCR with primers of 341 F:5′-CCTA CGGG NGGC WGCA G-3′ and 805 R: 5′-GACT ACHV GGGT ATCT AATC C-3′. The amplified DNA was then processed for gene sequencing with Illumina MiSeq-PE250 system. QIIME2 was adopted to analyze the gene sequences and classify the operational taxonomic units (OTUs) on the basis of 97% sequence identity. Subsequently, the functional potential of the bacterial community was predicted by using PICRUSt2 based on the 16 S rRNA gene sequencing data. The predictive genes were then used to construct Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways based on KEGG database. Differential functional gene expression was analyzed with STAMP 2.0.

RNA extraction, RNA-Seq and transcriptomic analysis

Three independent biological replicates of S. putrefaciens–R. palustris co-culture with/without Fe redox communication were subjected to RNA-seq analysis. According to the manufacturer’s protocols, the total RNA was extracted using the Trizol reagent (Invitrogen, California, USA). RNA quantity was measured with the Qubit 2.0 fluorometer (ThermoFisher Scientific, Waltham, MA, USA). Obtained RNA was quality and integrity checked using the RNA 6000 Nano kit with an Agilent Bioanalyzer 2100 (Agilent Technologies). Then the rRNA was depleted with Ribo-Zero rRNA Removal Kit (bacteria) (Illumina, San Diego, CA, USA) and cDNA libraries were constructed with an Illumina NEBNext Ultra RNA Library Prep Kit (New England Biolabs, Ipswich, MA, USA). Subsequently, the cDNA libraries were pair-end-sequenced on an Illumina Novaseq 6000 platform (Illumina, San Diego, CA, USA). Afterward, Cutadapt software was adopted to clean the raw reads by removing adaptor sequences, ambiguous based annotated as N, and low-quality sequences (Q_phred ≤ 20 bases, accounting for >50% of the total read length). The clean reads were obtained with the FastQC software and were then mapped to prjna881029 (S. putrefaciens) and prjna683609 (R. palustris) reference genome using STAR software. The read counts and fragments per kilobase of transcripts per million mapped reads were calculated using featureCounts (version 1.5.0-p3) and kallisto (version 0.46.1) respectively. Mapped reads were normalized with fragments per kilobase per million mapped reads (FPKM). Differential expression analysis of two groups was performed using DESeq2 (version 1.16.1), which is based on a negative binomial distribution model. The standardized method was DESeq. Genes with adjusted P value < 0.05 and Log₂ fold change ≥ |0.58| were considered significantly differentially expressed.

Bioelectrochemical detection

The gold interdigitated microelectrode was adopted to measure electron transfer function of S. putrefaciens–R. palustris co-culture with or without Fe redox communication. The gold interdigitated microelectrode was fabricated on a silicon wafer with 300 nm thermal silicon oxide (SiO₂/Si) by photolithography and metal evaporation⁴⁷. To be specific, a 4 inch silicon wafer was washed sequentially in acetone, ethanol and deionized (DI) water. Subsequently, the silicon wafer was dried with a compressed nitrogen gun. The residual moisture and solvent were removed by baking the wafer at 180 °C for 5 min. After cooling down, 3 ml of S1813 resist was dispensed at the wafer center and the wafer was then baked on the hot plate at 115 °C for 1 min. Then, a direct-write optical lithography system (ABM. Inc, America) was adopted for the preparation of the gold interdigitated microelectrode. After the photolithography process, the wafer was then loaded face down onto the sample holder of a thermal evaporator system (Jiashuo JSD-300, China) to deposit the active material (The thickness of chromium and gold are set as 5 nm and 40 nm respectively). S. putrefaciens–R. palustris co-culture with or without Fe redox communication incubated at 96 h were washed twice with 1×PBS buffer. Then, we fixed the OD₆₀₀ of S. putrefaciens–R. palustris co-culture to 3.0 with 1 × PBS buffer. Subsequently, the real-time current intensities of the co-culture were measured by injecting 10 µL bacterial suspension into the gold interdigitated microelectrode with the input voltage set at 0.2 V. Reported current densities (μA cm^–2) were then obtained by averaging the ratio of the stable current intensity to the effective working-area (0.80 × 0.88 mm) of the gold interdigitated microelectrode.

Determination of CO₂ fixation rate

The CO₂ fixation rates in S. putrefaciens–R. palustris co-culture and S. putrefaciens–R. palustris–G. Soli microbial consortia were measured respectively. The bacteria were cultured in LB medium with the addition of NaHCO₃ (30 mM). Then, the suspension was collected by centrifugation at 5000 × g for 5 min. H₂SO₄ solution (98 wt%) was used to prepare a 5 mM solution for the quantification of inorganic carbon by recording the mass variation. The CO₂ fixation rate was calculated as: bicarbonate consumption (mM)/time interval (96 h)/DCW(g).

Statistics and reproducibility

All data are presented as mean ± standard deviation (SD) and statistical analyses and graphs were performed using OriginLab 9.0 and GraphPad Prism 8.0.1. For comparison of two groups, significance was determined by the two-tailed unpaired Student’s t test. Each n indicates the number of biologically independent samples. Data in Figs. 2c, f, h, 3c–e, 5a–d and 6d–g and Supplementary Figs. 2, 9b, e, f, 10a, b, 11, 22 and 26 were successfully replicated in two independent experiments.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
• Source: https://www.nature.com/articles/s41467-023-42561-3