Hybrid Electrophototroph Enables High-Efficiency Carbon Dioxide 1 Valorization to Fuel Molecules

Nature’s biocatalytic processes are driven by photosynthesis, whereby photosystems I 13 and II are connected in series for light-stimulated generation of fuel products or electricity. Externally supplying electricity directly to the photosynthetic electron transfer chain (PETC) has numerous potential benefits, although strategies for achieving this goal have remained elusive. Here we report an integrated photo-electrochemical architecture which 17 shuttles electrons directly to PETC in living cyanobacteria. The cathode of this architecture electrochemically interfaces with cyanobacterial cells lacking photosystem II 19 activity that cannot perform photosynthesis independently. Illumination of the cathode 20 channels electrons from external circuit to intracellular PETC through photosystem I, 21 ultimately fueling CO 2 conversion to acetate, a model fuel molecule with 9.32% energy 22 efficiency, exceeding the efficiency of natural photosynthesis in higher plants (<1%) and 23 cyanobacteria (~4-7%). The resulting “Electrophototrophic” bio-electrochemical hybrid 24 has the potential to produce fuel chemicals with numerous advantages over standalone 25 natural and artificial photosynthetic approaches.

We next use site-specific redox inhibitors to demonstrate that PETC components 122 downstream of PSII (see Fig. 1a) play a central role in electron flow from extracellular 123 circuit to cyanobacteria. Supplementation of the herbicide (3-3,4-dichlorophenyl)-1,1-124 dimethylurea (DCMU), a specific inhibitor that blocks the binding site of QB in the 125 photosystem 30 (Fig. S3) did not diminish the light-dependent electrical response in ΔPSII 126 cells (Fig. 2b), suggesting that exogenous electrons can flow into the PETC downstream 127 of QB. Either blocking cytochrome b6f activity with 2,5-dibromo-3-methyl-6- Motivated by global demand in CO2 recycling and energy production, we ask 139 whether the exogenous electrons in our hybrid "electrophototrophic" system is able to 140 energize CO2 fixation and conversion to hydrocarbon fuels or fuel feedstocks. To answer 141 this question, we incubated ΔPSII cultures and applied electrical potential with 142 amperometric characterization. Light and electrical bias were systematically investigated 143 as two key variables, and we observe photosynthetic CO2 fixation and carbon product 144 formation only when supplying both illumination and exogenous electron supply ( Fig. S4-145 S6, more discussion in Supplementary Text). Shown in Fig. 3a, illumination on the 146 cathode (typical white LED for plant growth, 55 μmol m -2 s -1 on FTO glass) led to a 3.5-147 fold increment of acetate production compared to its initial value with applied electrical 148 bias (-0.7 V vs. Ag/AgCl, intermittent supply, Fig. S4). In the dark, acetate concentrations 149 in the culture slightly decreased (from ~270 μM initial residual to ~100 μM), presumably 150 due to non-photoexcited PSI which cannot reduce NADP and fuel carbon metabolism.
The viability determined by optical density (OD730) measurements indicate slight increase 152 under illumination, while the OD730 gradually declined ~40% in dark after 8 days (Fig. S5). 153 In terms of exogenous electron supply, without negative electrical bias, no acetate 154 production was detected even though cells were illuminated constantly (Fig. S6). 155 As shown in Fig. 3b, acetate production by illuminated ΔPSII was not found within 156 the first 5 days for application of either no bias or -0.5 V vs. Ag/AgCl. In comparison, once 157 more negative bias (-0.7 V vs. Ag/AgCl, intermittent supply, Fig. S4) was applied (day 6-158 10), acetate production resumed. Fig. S7 displays acetate yield as a function of various 159 potentials (-0.15 to -0.7V) and indicates that potentials more negative than -0.6 V vs. 160 Ag/AgCl can drive acetate production. Consistently, this threshold potential of -0.6 V (vs. 161 Ag/AgCl) is near the standard reducing potential of electrons in photoexcited PSII (Fig. 162 S7). This correlation implies a thermodynamic overpotential which could favor exogenous 163 electrons flowing into the bacteria downstream of PETC. Fig. 3b demonstrates that 164 acetate concentration in medium increased steadily for 5 days during incubation under -165 0.7 V, eventually reaching 650µM. Cell counts for ΔPSII, inferred by OD730 measurements, 166 decreased unless a certain bias was applied (Fig. S8). These results support our 167 hypothesis that the primary metabolic processes such as metabolite production and cell 168 maintenance can be energized by highly reductive exogenous electrons, flowing through 169 the PETC. 170 To further investigate the metabolic activities that can be driven by this 171 electrophototrophic system, we performed an isotope tracer analysis by adding 13 C-172 sodium bicarbonate into the ΔPSII culture on cathode. Bicarbonate can be converted to 173 CO2 by cyanobacterial carbonic anhydrase 33 . This CO2 can then drive carbon product 174 formation (acetate) and/or be fixed into biomass via cell metabolism. We first examined 175 the labeling fraction of acetate excreted into the medium. The GC-MS revealed the 176 production of 13 C-acetic acid, indicating that newly fixed carbons end into this C2 product 177 (Fig. 3c). 1 H-NMR spectra demonstrate that acetate was labeled in both methyl and 178 carboxyl carbons (Fig. S9) and allow us to evaluate the energy conversion efficiency in 179 the electro-photosynthetic process. Similar to the faradaic efficiency of conventional 180 electrochemical processes, the exogenous electrons involved in electrophototrophic 181 synthesis of acetate can be quantified by defining the exogenous electrons uptake efficiency (EEUEacetate). Over half (61.8%) of exogenous electrons were utilized by ΔPSII 183 for selective acetic acid generation. Taking the incident photon flux into account, the 184 overall energy conversion efficiency is approximately 9.32% (see Supplementary Text 185 and Table S2). Even though this estimation only reflects the fixed carbons in acetate and 186 does not those fixed into biomass (vide infra), the value still exceeds typical natural energy 187 conversion efficiency of higher plants (<1%) and cyanobacteria (~4-7%). 34,35 188 We next analyzed the labeling patterns of seven proteinogenic amino acids that 189 are digested from cell biomass and are directly produced from the central carbon 190 metabolism (Fig. 3d). After four days incubation of ΔPSII with 13 C-bicarbonate under 191 constant white-light illumination, the cathodically biased cultures demonstrate partial 13 C-192 labeling in proteinogenic amino acids and display significantly higher fractional labeling 193 (FL, denoting the proportion of labeled carbons) than the negative control cultures without 194 applied bias. Serine, which can be synthesized from 3-phosphoglycerate, the first CO2-195 fixation product of the CBB cycle, demonstrated a 3% FL in comparison with 1% in the 196 negative control. This moderate 13 C-accumulation is real because we indeed detected 197 significant increase of the m+1 13 C-pattern in the carboxylic group of serine, consistent 198 with the reaction skeleton of Ribulose-1,5-bisphosphate carboxylase/oxygenase 199 (RuBisCO) ( Table S1). As another major CO2 entry point, 13 C-bicarbonate can be fixed 200 by amphibolic reactions (e.g. phosphoenolpyruvate carboxylase) to generate 201 oxaloacetate which is the precursor of aspartate and threonine. Consistently, biased 202 cultures have much higher FL (7%) in these two amino acids than those in the unbiased 203 cultures (1%). 204 Interestingly, we observed a new CO2 fixation pathway activated in cyanobacteria 205 via glycine cleavage system which was found in Synechocystis 36 but with no detailed in 206 vivo characterization. The metabolic activity of this CO2-fixing pathway can be reflected 207 by the extremely high fractional labeling in glycine over 30% when incubated ΔPSII with 208 13 C-bicarbonate under constant white-light illumination for four days. Through this 209 pathway, CO2 will enter the one-carbon (C1) metabolism via formate which then forms 210 the methylene group of glycine. The GC-MS fragment of glycine (Gly_85) represents this 211 methylene group, demonstrating high FL (30%) consistently. Our 13 C-tracer analysis as 212 well as extracellular metabolite analysis support that exogenous electron supply to cyanobacterial PETC may lead to CO2 fixation and conversion, demonstrating a 214 functional bioenergetics system that fuels endergonic metabolism. 215 216

Discussion 217
This work provides electrochemical and biochemical evidence to support a proof-218 of-concept hybrid electro-photosynthetic system that leverages exogenous electrons to 219 supplement photosynthetic energy conversion for driving CO2 fixation and conversion. 220 Cyanobacterial cells without PSII can sustain their metabolic viability on an electrode 221 surface and produce acetate, the primary excreting product (Fig. 3). Growing 222 photosystems-modified cyanobacteria in a photo-electrochemical architecture allows us 223 to expand the means by which photosynthetic organisms produce fuels and chemicals. Another merit for this hybrid photosynthesis approach arises from the fact that 238 inactivated PSII does not evolve O2 as the photosynthetic byproduct. Suppressed O2 239 evolution minimizes the propensity for RuBisCO to fix O2 as a competitive substrate for 240 CO2. In natural photosynthesis, substrate competition initiates an energy-intensive 241 recovery process of photorespiration 38 that can consume up to 25% of the initially stored 242 energy 39 , a substantial source of inefficiency. Interestingly, although photorespiration also plays a biosynthetic role in metabolic processes, e.g. supplying glycine as an essential 244 metabolite 38 , this role in hybrid photosynthesis seems to be substitutable with redundant 245 pathways, such as glycine cleavage system 36 . This notion is strongly supported by the 246 presence of pathway genes in cyanobacterial genome in line with isotope labeling 247 patterns as we provided here. Decrease in photorespiration thus underlie new 248 opportunities in the hybrid system to raise theoretical limits of photosynthesis. 249 More importantly, the hybrid system introduces a unique strategy for managing 250 photosynthetic outcomes. In natural photosynthesis, linear electron flow occurring 251 between two photochemical systems produces ATP and NADPH as energetic currency, 252 and their proportions are regulated for various biosynthetic purposes. Phototrophs 253 containing only PSI implement electron transport whereby electrons can be recycled from 254 either reduced ferredoxin or NADPH to PQ, and subsequently to the cytochrome b6f 255 complex 40 . Such cyclic flow generates a pH gradient (and thus ATP), but without the 256 accumulation of reduced species for biosynthesis 41 . However, this study shows that 257 Synechocystis carrying single PSI can be electrically energized to fix CO2 and generate 258 building blocks of biomass, evinced by labeled proteinogenic amino acids from 13 C-259 bicarbonate. This study further indicates that the hybrid photo-electrochemical process 260 demonstrated here could enable on-demand control over the proportion of linear versus 261 cyclic electron flow to tailor the stoihiometric ratios of ATP and NADPH and ultimate 262 photosynthetic products. To achieve this goal, Nature evolved complicated regulatory 263 mechanisms to tune the ratio of PSI to PSII 42,43 . In the photo-electrochemical hybrid 264 demonstrated here, the ratios of energetic currency and products could instead be 265 regulated through the injection of exogenous electrons, which creates an artificial linear 266 electron flux that can be varied on-demand relative to cyclic electron flow by tuning the 267 cathodic current density and/or incident photon flux. Since this hybrid approach is not 268 tailored by evolution, it will be less constrained by the natural needs/environments to 269 implement. Instead, the hybrid can be optimized in well-designed conditions for targeted 270 ATP/NADPH ratio. Reengineering the system, for example on the biotic-abiotic interface, 271 is expected to improve overall efficiency for tunable electron transfer. 272 Taken together, the hybrid electrophototroph as we demonstrate, drives 273 exogenous electrochemical energy to replenish the universal energy and redox currency 274 in living cyanobacteria for biosynthesis. Considering its functionality and a number of 275 advantages over pure natural/artificial photosynthesis, we posit that the development of 276 this bio-electrochemical platform will pave a new avenue to couple renewable electricity 277 with photobiological activities, a practical approach for production of hydrocarbon fuels 278 from sun and CO2. Cyanobacteria-electrode hybrid system 294 The PSII deficient Synechocystis (Δslr0906) was first inoculated and cultured 295 photoheterotrophically in BG11 medium with addition of 5mM glucose, under 30-50 296 µE/m 2 /s illumination at 30°C. Exponentially growing cells were collected for further 297

applications. 298
In the following procedure, the tailored electrochemical H-cell with three-electrode 299 configuration was applied for the electrochemical process. The reference and counter 300 electrodes were silver/silver chloride electrode and Pt, respectively. The working 301 electrode and reference electrode (CH Instruments, Inc.) were in the bottom chamber and 302 the Pt wire counter electrode was in the top chamber. A Nafion 117 membrane (Sigma-Aldrich) separates the two chambers. Each chamber has an inlet/outlet. The exponentially 304 growing culture was centrifuged, separated from the supernatant and re-dispersed in the 305 medium (BG11+ bicarbonate, pH = 7.8). A ~7 ml culture was transferred to the cathode 306 chamber of the H-cell, where the culture was illuminated from the bottom transparent 307 window. The device was air-tight and maintained at 30 °C for the duration of the 308 electrochemical characterization. 309

Photoelectrochemical characterization 310
During the electrochemical incubation, a typical amperometry (i-t) procedure (CH 311 Instruments, Inc.) was conducted to check the ability of ΔPSII cyanobacteria as an 312 electron acceptor under illumination. It was conducted at different potentials (vs. Ag/AgCl). 313 A 0.15 ml culture was taken every day for OD730 and metabolite analysis. 314

Electron microscopy characterization 322
After the electrochemistry process, the carbon felt electrode was fixed in 2.5% 323 glutaraldehyde in phosphate buffer under 4 °C for 2 h. The samples then underwent a 324

MilliQ water postfix wash and dehydration (~ 24 h in a high vacuum desiccator). Scanning 325
Electron Microscopy (Hitachi S-4800 SEM) was applied to characterize the surface 326 morphology. Samples were imaged at 3 kV acceleration, 7-10 mm working distance. 327

Quantitative analysis of acetate 328
We measured the excretion of acetate from Synechocystis using the following 329 method. The culture samples were collected and the supernatant was separated from 330 cells by filtration through 0.2 μM-diameter nylon membrane (Acrodisc®). Acetate 331 concentration in each culture was analyzed with High Performance Liquid Chromatography (HPLC, Agilent Technologies1200 series) by injecting 25 μL samples 333 into an HPLC column (Bio-Rad Aminex HPX-87H), eluting with 5mM sulfuric acid at a 334 flow rate of 0.6 ml/min, and detecting by a refractive index detector (retention time for 335 acetate: 15.2 min). Standard samples with five different acetate concentrations (2.5, 5, 336 10, 25, and 50 mM) were used for quantification (R 2 = 0.99839). 337 13 C-isotope tracer analysis to track carbon fixation 338 13 C-bicarbonate was supplied during the electrochemical procedures to monitor 339 carbon metabolism in the photoelectrochemical environment. The 13 C-labeled fraction of 340 acetate and protein-bound amino acids were measured by NMR and gas 341 chromatography-mass spectrometry (GC-MS), respectively. Exponentially growing ΔPSII 342 cells were suspended in BG-11 medium supplemented with 100mM 13 C-labeled sodium 343 bicarbonate. The culture was applied in the electrochemical device under sunlight 344 simulated illumination (white LED, 55 μmol m -2 s -1 on FTO glass). Cultures were sampled 345 at 0 hour, 2 day, 4 day and 5 day. 346 The sample treatment and GC-MS analysis were performed as previous 347 reported. 44 Briefly, 5mL of sampled cultures were centrifuged at 10,000 g for 1 minute, 348 the cell pellets were digested in 500μL 6M HCl at 105°C for 12 hours. The hydrolysate 349 was dried under nitrogen gas flow at 65°C, dissolved in 50 μL water-free 350 dimethylformamide. For the GC-MS measurement the proteinogenic amino acids were 351 derivatized prior to analysis. The dried hydrolysate, dissolved in pyridine was derivatized 352 by N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide (TBDMS) with 1% tert-butyl-353 dimethylchlorosilane at 85°C for 60 min. 1 μL of the sample in the organic phase was 354 loaded on the Agilent GC-6890 gas chromatography equipped with a Agilent 19091J-413 355 column (30m×0.32mm×0.25μm) directly connected to a MS-5975C mass spectrometer. 356 Helium was used as the carrier gas. The oven temperature was initially held at 50°C for 357 2 min; then raised to 150°C at 5°C /min and held at that value for 2 min; finally, it was 358 raised to 320°C at 7°C /min, and held at that final value for 2 min. Other settings included 359 splitless and electron impact ionization (EI) at 70 eV. The FLs of alanine, aspartate, 360 glutamate, glycine, phenylalanine, serine, threonine was analyzed.  biological triplicates. c) 13 C-acetate production via fixation of 13 CO2 (derived from 13 C-571 bicarbonate) in illuminated ΔPSII, with or without application of external electrical bias (-572 0.7 V vs. Ag/AgCl). d) Fractional labeling of seven protein-bound amino acids that were 573 directly produced from the central carbon metabolism via fixation of 13 CO2 (derived from 574 13 C-bicarbonate) in illuminated ΔPSII, with or without application of external electrical bias 575 regions that cannot be utilized by photosynthetic organisms, may be converted to 589 electricity to energize electrophototrophic cyanobacteria with single PSI for CO2 590 valorization. Dash line represents the standard redox potential where the exocellular 591 electrons shuttled to the photosynthetic cells carrying a single photosystem. 592