# Source paper archive - Source title: CO2-concentrating nanoparticles drive progressive photosynthetic and cell-wall remodelling in Arabidopsis thaliana cells - Author(s): Routier, C., Pagliara, D., Karlsson, A., Lisica, L., Boutonnet, L., Bogdziewiez, L., Colaprico, E., Garcia Fleitas, A., Jin, X., Strand, A., Mantione, D., Verger, S., Hudson, E. P., Stavrinidou, E. - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.10.30.763161v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.10.30.763161v1.full.pdf - PDF SHA-256: 61fe56a0126cab2bc0fe0b8a717f03211d9b26ef03f751017289da43229fb6a0 - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1 **CO** **2** **-concentrating nanoparticles drive progressive photosynthetic and cell-wall** 2 **remodelling in** _**Arabidopsis thaliana**_ **cells** 3 4 5 _Cyril Routier_ _[1^]_ _, Denise Pagliara_ _[1^]_ _, Anna Karlsson_ _[2]_ _, Lucija Lisica_ _[3]_ _,_ _Léo Boutonnet,_ _[1]_ _Léa_ 6 _Bogdziewiez_ _[3]_ _, Erica Colaprico_ _[1]_ _, Ariel García Fleitas_ _[1]_ _, Xu Jin_ _[4]_ _, Åsa Strand_ _[4]_ _, Daniele_ 7 _Mantione_ _[5,6,]_ _, Stéphane Verger_ _[3,4]_ _, Elton P. Hudson_ _[2]_ _, Eleni Stavrinidou._ _[1,3 7]_ _*_ 8 9 1 Laboratory of Organic Electronics, Department of Science and Technology, Linköping University, SE-60174, 10 Norrköping, Sweden. 11 2 School of Engineering Sciences in Chemistry, Biotechnology and Health, Science for Life Laboratory, KTH 12 Royal Institute of Technology; Stockholm, Sweden 13 3 Umeå Plant Science Centre, Department of Forest Genetics and Plant Physiology, Swedish University of 14 Agricultural Sciences, SE-90183 Umeå, Sweden. 15 4 Umeå Plant Science Centre, Department of Plant Physiology, Umeå University, SE 901-87 Umeå, Sweden 16 5 POLYMAT, University of the Basque Country UPV/EHU, 20018 San Sebastián, Spain 17 6 Ikerbasque, Basque Foundation for Science, Bilbao, Spain 18 7 Wallenberg Initiative Materials Science for Sustainability, Linköping University, SE-60174, Norrköping, 19 Sweden. 20 21 22 ^equal contribution 23 *corresponding author: eleni.stavrinidou@liu.se 24 Keywords: plant cell biohybrids, CO 2 capture, polyethyleneimine nanoparticles, 25 photosynthesis, cell wall biosynthesis 26 27 28 29 30 31 32 33 34 35 36 1 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 37 **Abstract** 38 39 Plant cells offer a distinctive platform for biohybrid engineering because they combine growth 40 and proliferation with photosynthetic carbon acquisition and the continuous fabrication of 41 structurally complex biomass. Harnessing these capabilities could enable living systems in 42 which synthetic functionality develops together with cellular physiology and material 43 organization. Yet nanomaterial engineering of proliferating plant cell populations remains 44 largely unexplored. Here, we establish a biohybrid system combining photosynthetically active 45 _Arabidopsis thaliana_ suspension cells with chitosan-modified polyethyleneimine (gPEI–Chi) 46 nanoparticles, a synthetic CO 2 -concentrating nanomaterial. By supplying fresh nanoparticles 47 at each subculture, newly generated cell populations were successively functionalized, driving 48 progressive changes in cellular and multicellular phenotypes. Nanoparticle functionalization 49 reduced net headspace CO 2 accumulation, increased chlorophyll content and induced stage 50 dependent reorganization of photosynthetic function and carbon-handling pathways. These 51 physiological changes were accompanied by altered cell wall regeneration and progressive 52 remodelling of cell wall composition, spatial organization and ultrastructure, ultimately 53 promoting more cohesive multicellular aggregates. Our findings establish successive 54 nanomaterial functionalization as a strategy to steer the evolving physiology and structural 55 organization of proliferating plant cell populations, opening opportunities for plant biohybrid 56 systems and engineered living materials. 57 58 59 60 61 62 63 64 65 66 67 68 69 2 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 70 **Introduction** 71 Living cells offer capabilities difficult to reproduce in conventional engineered systems as they 72 grow, multiply, sense their environment and continuously remodel their own structures. 73 Integrating synthetic functionality into proliferating cell populations could therefore enable 74 biohybrid systems that develop as their living components grow. Plant cells are particularly 75 attractive because they combine photosynthetic carbon acquisition with the ability to convert 76 assimilated carbon into biomass and structurally complex cell walls [1], providing an opportunity 77 to engineer both cellular physiology and biological material formation. 78 Plant nanobionics integrates engineered nanomaterials with living plants to introduce or 79 enhance biological functions through nano–bio interactions [2,3] . Nanomaterials have enabled 80 molecular delivery [4,5], sensing [6,7] and modulation of photosynthetic processes [8,9] despite the 81 transport barriers imposed by plant cell walls and membranes [10,11] . These approaches have 82 largely focused on intact plants and differentiated tissues, whereas proliferating plant cells 83 remain comparatively unexplored. Suspension cells continuously divide and can be propagated 84 through successive subcultures, creating an opportunity to functionalize newly generated cell 85 populations and investigate whether nanomaterial interactions can progressively reshape 86 cellular and multicellular phenotypes. 87 Carbon handling provides a particularly relevant target for such engineering. CO 2 is the 88 principal carbon source for photosynthesis, yet its fixation is constrained by the catalytic 89 properties of RuBisCO, including competing oxygenase activity, and by limitations in CO 2 90 transport to the site of carboxylation [12,13] . Natural carbon-concentrating mechanisms increase 91 CO 2 availability around RuBisCO and suppress photorespiration [14,15], while efforts to engineer 92 such mechanisms in other photosynthetic organisms have largely relied on genetic and 93 metabolic modification [16–18] . Synthetic materials capable of concentrating CO 2 offer an 94 alternative strategy to modify the carbon environment of photosynthetic cells without editing 95 their genome. Although nanomaterials have been used to modulate light harvesting [19,20] and 96 photosynthetic electron transport [8,21], comparatively little attention has been directed towards 97 engineering inorganic carbon availability [22] . 98 Building on the ability of polyamines to capture CO 2 [23–25], we previously developed green 99 fluorescent chitosan-modified polyethyleneimine nanoparticles (gPEI–Chi NPs) as a synthetic 100 CO 2 -concentrating nanomaterial for plants [26] . Combining the CO 2 affinity of polyethyleneimine 101 with the phytocompatibility of chitosan [27,28], gPEI–Chi NPs reversibly capture inorganic carbon 102 and increased RuBisCO carboxylation by 20% in enzymatic assays. Following infiltration in 103 planta, the nanoparticles (NPs) distributed throughout tobacco leaf tissues, including 3 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 104 chloroplast-rich regions, while maintaining phytocompatibility [29] . These findings established 105 the feasibility of manipulating the local inorganic carbon environment of photosynthetic tissues 106 using nanomaterials. However, how such functionality interacts with proliferating plant cells 107 as they divide and are successively propagated, and whether functionalization of newly 108 generated cell populations can progressively reshape their phenotype, remain unknown. 109 This question extends beyond photosynthesis because assimilated carbon also provides the 110 building blocks for the polysaccharide-rich plant cell wall. Changes in carbon handling could 111 therefore influence both cellular metabolism and the structural organization of multicellular 112 assemblies. This coupling is particularly relevant to engineered living materials (ELMs), where 113 living cells provide adaptive and regenerative functionality [30,31] . Plant cells are attractive 114 building blocks for such systems because they photosynthetically acquire carbon and convert 115 it into structurally complex cell walls. Recently, we demonstrated that plant cell-based ELMs 116 can exploit cellular metabolic plasticity and cell wall biochemistry to generate adaptive and 117 programmable material properties [32] . Engineering proliferating plant cells with nanomaterials 118 could therefore provide a route to regulate both cellular physiology and structural organization 119 as living populations develop. 120 Here, we establish a biohybrid system combining photosynthetically active _Arabidopsis_ 121 _thaliana_ suspension cells with CO 2 -concentrating gPEI–Chi NPs and investigate how NP 122 functionalization across successive subcultures reshapes cellular and multicellular phenotypes. 123 We show that successive functionalization alters carbon handling and photosynthetic function 124 while progressively remodelling cell wall regeneration, composition and organization, 125 ultimately promoting a more cohesive multicellular phenotype. These findings demonstrate 126 that nanomaterial functionalization of proliferating plant cell populations can couple changes 127 in cellular physiology with structural reorganization, providing a platform for plant biohybrid 128 systems and engineered living materials. 129 130 131 **gPEI–Chi NPs functionalize photosynthetic Arabidopsis suspension cells** 132 _Arabidopsis thaliana_ suspension cultures were established from root-derived calli and adapted 133 to photomixotrophic growth under constant illumination and reduced sucrose supplementation 134 (Fig. 1a), as previously demonstrated [33] . These conditions promoted chloroplast development 135 while maintaining suspension growth, with confocal imaging revealing robust chlorophyll 136 autofluorescence throughout the cultures (Fig. 1b). 4 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 137 Green-fluorescent chitosan-modified polyethyleneimine nanoparticles (gPEI–Chi NPs), 138 previously developed as CO 2 -capturing nanomaterials for in planta applications [26], were 139 synthesized by reductive amination of polyethyleneimine with oligochitosan followed by 140 fluorescein isothiocyanate (FITC) labelling (Fig. 1c). NPs were added at subculture at a final 141 concentration of 0.1 mg mL⁻¹. Evans Blue staining revealed no detectable increase in cell death 142 following NP treatment in either non-photosynthetic or photosynthetic cultures 143 (Supplementary Fig. S1). Notably, membrane-compromised cells lacked detectable NP 144 fluorescence, suggesting that intracellular NP accumulation depends on cellular integrity. 145 Upon addition, gPEI–Chi NPs rapidly associated with the cell periphery, consistent with PEI 146 affinity for plant cell wall components (Supplementary Fig. S2) [26,34,35] . NP-associated 147 fluorescence subsequently redistributed towards intracellular compartments, with substantial 148 internalization observed after approximately 3 days (Fig. 1d). Spectroscopic analysis supported 149 progressive cellular association and uptake as FITC-associated absorbance and fluorescence in 150 NP-containing medium decreased substantially faster in the presence of cells, approaching 151 baseline by day 5 (Supplementary Fig. S3). Consistently, confocal quantification showed 152 increasing NP-associated fluorescence within the cellular area and progressive redistribution 153 from the cell surface towards intracellular compartments over 6–7 days (Supplementary Fig. 154 S4). Together, these measurements demonstrate progressive NP uptake and intracellular 155 accumulation during culture. 156 To determine whether NP functionalization affected culture growth, cell volume after 157 sedimentation (CVS; Fig. 1e-i and Supplementary Fig. S5) and biomass dry weight (Fig. 1e-ii) 158 were quantified over 14 days. NP treatment did not substantially alter the characteristic growth 159 kinetics of either photosynthetic or non-photosynthetic cultures, although a modest transient 160 delay was observed during early growth (Fig. 1e-i and Supplementary Fig. S6a). Photosynthetic 161 cultures required medium renewal after 7 days, when NPs were replenished at the same 162 concentration. 163 Despite similar CVS profiles, NP-treated photosynthetic cultures showed transient changes in 164 dry biomass. Dry weight was 43% and 38% higher than controls at days 3 and 5, respectively, 165 before returning to control levels by days 6–7. Following medium renewal, biomass increased 166 again by 9–11% at days 8–9, decreased by approximately 24% at day 11 and subsequently 167 stabilized near control levels. The divergence between CVS and dry weight indicates that NP 168 functionalization transiently alters biomass accumulation and/or cellular composition without 169 substantially affecting culture proliferation. 5 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 170 Together, these results demonstrate that gPEI–Chi NPs efficiently associate with and are 171 internalized by photosynthetic _Arabidopsis thaliana_ suspension cells while preserving cellular 172 viability and proliferative capacity. 173 174 175 176 177 178 179 6 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 180 181 **Fig. 1│ gPEI-Chi NPs are internalized by photosynthetic** _**Arabidopsis thaliana**_ **cells in** 182 **suspension cultures without inducing cytotoxic effects. a**, Schematic representation of the 183 generation of photosynthetic suspension cultures derived from _Arabidopsis thaliana_ root 184 explants (created in BioRender.com). **b,** Confocal microscopy images of cells showing 185 brightfield and chlorophyll autofluorescence, confirming the establishment of 7 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 186 photosynthetically active cells. Scale bar 10µm. **c**, Synthesis pathway of green-fluorescent 187 chitosan-modified polyethyleneimine nanoparticles (gPEI-Chi NPs). **d,** Time-course confocal 188 imaging of NP internalization showing progressive association with the cell wall and 189 intracellular localization, with significant uptake observed after approximately 3 days. Scale 190 bar 20µm. **e**, Growth kinetics of control and gPEI-Chi-treated cultures over 14 days, assessed 191 by cell volume sedimentation (CVS, i) and dry weight measurements (ii), demonstrating no 192 significant impact of NP treatment on cell proliferation. 193 194 **gPEI–Chi NP functionalization across successive subcultures induces progressive** 195 **phenotypic changes** 196 197 Despite efficient intracellular association, a single NP functionalization cycle produced no 198 major macroscopic phenotypic changes beyond transient differences in biomass accumulation. 199 Because cellular responses to sustained abiotic stimuli can develop beyond acute responses to 200 encompass longer-term metabolic, structural and population-level changes, we investigated 201 whether repeated gPEI–Chi NP functionalization across successive subcultures would 202 progressively alter the phenotype of photosynthetic cultures [36–38] . 203 Photosynthetically active _Arabidopsis thaliana_ suspension cultures were subcultured weekly 204 with fresh NPs supplied at the same concentration at each subculture to maintain exposure as 205 the cells proliferated (Fig. 2a). Gen0 denotes the initial functionalization whereas Gen3 and 206 Gen6 represent cultures maintained under repeated NP functionalization for three and six 207 successive weekly subcultures, respectively. 208 Over successive subcultures, NP-functionalized cultures progressively developed a distinct 209 macroscopic phenotype characterized by denser, more cohesive cell clusters relative with 210 control and Gen0 cultures, which remained more homogeneous and exhibited a smoother 211 suspension appearance (Fig. 2b). Confocal imaging confirmed the increased aggregation of 212 Gen3 and Gen6 cultures (Supplementary Fig. S7 and S8). Chlorophyll autofluorescence also 213 appeared more pronounced and broadly distributed in Gen3 and Gen6 cells compared with 214 control and Gen0 cultures, suggesting increased chloroplast development (Fig. 2c). NP 215 associated fluorescence remained closely associated with chlorophyll-rich intracellular regions 216 in both Gen3 and Gen6 cells, consistent with localization of gPEI-Chi NPs in chloroplast-rich 217 regions. In addition, highly fluorescent shell-like structures were frequently observed in NP 218 functionalized cultures, consistent with cell wall remnants retaining NP-associated 219 fluorescence after loss of cellular contents (Supplementary Fig. S8b). 8 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 220 221 Transmission electron microscopy revealed corresponding ultrastructural changes across 222 successive NP functionalization cycles (Fig. 2d, Supplementary Fig. S9). Control cells 223 displayed regular vacuolar and organelle organization, including mitochondria and chloroplasts 224 containing plastoglobules and thylakoid membranes. Gen0 cells exhibited increased 225 aggregation, enlarged vacuoles, more abundant mitochondria and starch-enriched chloroplasts. 226 In Gen3 and Gen6 cultures, cellular aggregation became increasingly pronounced, while 227 vacuoles regained a more regular morphology and chloroplasts appeared more abundant 228 throughout the cytoplasm. Chloroplasts also exhibited increasingly organized thylakoid 229 membranes including grana formation, particularly in Gen6 cells. These observations reveal 230 progressive changes in cellular and chloroplast ultrastructure. 231 To determine whether successive functionalization altered NP uptake dynamics, Gen3 and 232 Gen6 cultures were monitored following NP replenishment (Supplementary Fig. S10). In both 233 cultures, NP-associated fluorescence rapidly accumulated intracellularly, indicating faster 234 uptake than during the initial Gen0 functionalization. Consistently, FITC-associated 235 absorbance and fluorescence signals appeared in the culture medium immediately after NP 236 replenishment but rapidly declined over subsequent days. 237 Together, these results demonstrate that gPEI–Chi NP functionalization across successive 238 subcultures induces faster NP uptake and progressively alters culture morphology and cellular 239 ultrastructure resulting in more cohesive cell aggregates and increased chloroplast 240 development. 241 242 243 9 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 244 245 **Fig. 2│ gPEI-Chi NPs functionalization over generations of cells induces cellular** 246 **aggregation and alters cell morphology and chloroplast development.** **a**, Schematic 247 representation of the experimental workflow for multi-generational NP functionalization, in 10 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 248 which suspension cultures are repeatedly exposed to gPEI-Chi NPs across successive 249 subcultures (Gen1, Gen2, etc.). **b**, Representative pictures showing the progressive 250 macroscopic appearance of cultures over generations, highlighting increased aggregation and 251 changes in culture morphology upon repeated NP exposure. **c**, Confocal microscopy images of 252 NP-treated cultures (Gen6) revealing enhanced cellular aggregation and increased chlorophyll 253 autofluorescence, indicative of more developed chloroplast structures, colocalized with gPEI 254 Chi NP fluorescence in the case of Gen3 and Gen6 cultures. **d,** Transmission electron 255 microscopy (TEM) images of NP treated and non-treated cultures, underlying the 256 ultrastructural differences between the conditions. Control cultures present chloroplasts with 257 abundance of stromal thylakoids and plastoglobuli (black dots). Gen0, Gen3 and Gen6 present 258 high content of starch-enriched chloroplasts and their number increases particularly in Gen3 259 and Gen6. In Gen6, granal thylakoids can be also detected together with stromal thylakoids. In 260 the images C=chloroplasts, S=starch, ST=stromal thylakoids, PG=plastoglobuli, GT=granal 261 thylakoids. 262 263 **NP functionalization reorganizes carbon handling and photosynthetic function** 264 Given the sustained association of gPEI–Chi NPs with chlorophyll-rich intracellular regions, 265 we next investigated whether NP functionalization alters CO 2 dynamics and photosynthetic 266 function. Net CO 2 exchange was monitored using a closed-loop system that continuously 267 measured headspace CO 2 concentrations under isolated conditions (Fig. 3a,b). To assess the 268 intrinsic CO 2 -capturing capacity of the NPs, gPEI–Chi NP suspensions were first monitored in 269 the absence of cells. Headspace CO 2 remained between 300 and 400 ppm over 7 days, 270 consistently below atmospheric levels, confirming CO 2 capture by the NPs under the 271 experimental conditions (Supplementary Fig. S11a). 272 Control cultures showed an initial increase in headspace CO 2, reaching approximately 1500 273 ppm by day 4, followed by a gradual decline (Fig. 3b). Gen0 cultures accumulated substantially 274 less CO 2, reaching a maximum of approximately 750 ppm at day 6. In Gen3, CO 2 peaked 275 earlier, at days 4–5, and in Gen6 at days 3–4, both remaining at or below approximately 750 276 ppm before declining. Thus, NP functionalization consistently reduced net headspace CO 2 277 accumulation potentially through changes in cellular respiration or CO 2 consumption, NP 278 mediated carbon retention, or a combination of these processes. Notably, by Gen6 the temporal 279 CO 2 profile became more similar to that of untreated controls while remaining consistently 280 lower in absolute concentration, suggesting progressive reorganization of carbon handling 281 across successive subcultures. 11 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 282 Medium pH showed a stage-dependent response (Supplementary Fig. S11b,c). Gen0 cultures 283 exhibited an initial alkaline shift followed by delayed acidification, whereas Gen3 and Gen6 284 progressively recovered more control-like dynamics. Although these changes broadly tracked 285 headspace CO 2 trends, extracellular pH is also influenced by ion, nutrient and metabolite 286 exchange and therefore cannot be interpreted as a direct proxy for CO 2 assimilation. Instead, 287 the pH response is consistent with broader changes in extracellular metabolic regulation during 288 successive NP functionalization (Supplementary Note 1). 289 290 We next examined whether altered CO 2 dynamics were accompanied by changes in the 291 photosynthetic apparatus. Chlorophyll accumulation increased progressively with successive 292 subculture NP functionalization (Fig. 3c, Supplementary Table 1). Gen0 cultures showed a 293 modest increase in chlorophyll content relative to untreated controls, reaching approximately 294 3% higher levels after 1 week and 26% after 2 weeks. The response was substantially greater 295 in repeatedly functionalized cultures. After 1 week, chlorophyll content was approximately 296 120% and 160% higher in Gen3 and Gen6, respectively, compared with age-matched untreated 297 controls, and remained elevated after 2 weeks of culture. 298 Pulse-amplitude-modulated (PAM) fluorescence was used to determine whether increased 299 chlorophyll accumulation was accompanied by changes in photosynthetic performance (Fig. 300 3d and Supplementary Fig. S12-S13). The maximum quantum yield of photosystem II (Fv/Fm) 301 showed modest reductions in Gen0 and Gen3 (-11% and -15%), consistent with an initial 302 perturbation of photosystem II (PSII) function, whereas Gen6 recovered to control-like levels 303 (-6% than control condition) (Fig. 3d-i). Gen0 also exhibited reduced maximum electron 304 transport rate (ETRmax) (Fig. 3d-ii) and lower initial light-response slope (α) (-23% and -18% 305 with respect to the control cultures, respectively), together with higher photoinhibition 306 parameter (β) (+78% than control), (Supplementary Fig. S13a,b), indicating reduced 307 photosynthetic performance following the initial NP functionalization. 308 In contrast, Gen3 displayed increased ETRmax (+46%) and lower β (-73%) compared to 309 control, indicating an enhanced capacity to sustain electron transport across increasing 310 irradiance with reduced high-light inhibition with respect to control cells. By Gen6, both ETR 311 at saturation (ETRs) and ETRmax were constant compared with Gen3, while α recovered 312 reaching control conditions, suggesting improved low-light efficiency and sustained maximum 313 electron transport capacity between higher generations. PAM quantifications were performed 314 on cultures exhibiting their characteristic aggregated phenotype, which may contribute by 315 modifying light exposure within the cell population (Supplementary Fig. S12b, S13c). 12 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 316 However, the clear differences between the aggregated Gen3 and Gen6 suggested that the 317 presence of aggregation alone is insufficient to explain the observed response and the 318 photosynthetic response to repeated gPEI-Chi NP functionalization is stage dependent. Gen3 319 exhibits greater resistance to high-light exposure, while Gen6 reflects further reorganization of 320 photosynthetic function, with higher low-light efficiency and maximum PSII quantum yield, 321 rather than a broad upregulation of photosynthesis. 322 323 Untargeted proteomics supported this stage-dependent response (Fig. 3e, Supplementary Table 324 2, Fig. S14). In Gen0, several proteins associated with light harvesting and photosynthetic 325 electron transport, including LHCB2.1, psbA, psbC, PSBO1 and PSAF, showed lower 326 abundance relative to controls, together with proteins involved in Calvin–Benson cycle 327 regeneration, including SBPase, TKL-1 and RPI3. Combined with the PAM measurements and 328 the reduced representation of photosynthesis-related processes identified by gene set 329 enrichment analysis a (GSEA; Supplementary Fig. S15), these changes are consistent with an 330 initial perturbation of photosynthetic function following NP exposure. 331 In Gen3 and Gen6, proteins associated with light harvesting and energy production (LHCB2.1, 332 RFNR1 and atpB), Calvin–Benson cycle regeneration (SBPase, TKL-1, RPI3 and TIM), and 333 RuBisCO abundance and activation (rbcL, RBCS-1A and RCA) generally recovered relative 334 to Gen0. These changes were consistent with the progressive development of starch-enriched 335 chloroplasts observed by ultrastructural analysis. Enrichment of proteins associated with 336 responses to high light was also consistent with the reduced photoinhibition observed following 337 repeated functionalization. Recovery of LHCB2.1 coincided with the pronounced increase in 338 chlorophyll content in Gen3 and Gen6 and may reflect changes in the abundance or 339 organization of chlorophyll-containing antenna complexes. 340 However, because several PSII core proteins remained below control levels, the response is 341 better characterized as a selective reorganization of the photosynthetic machinery rather than a 342 uniform upregulation of photosynthesis. Changes in photorespiratory proteins, together with 343 increased abundance of mitochondrial proteins, further indicate coordinated remodelling of 344 photosynthetic carbon metabolism, photorespiratory recycling and mitochondrial energy 345 metabolism in Gen3 and Gen6. 346 Carbonic anhydrases (CAs) abundance provided further evidence for adaptation of intracellular 347 inorganic carbon handling. Whereas Gen0 showed relatively limited changes in CA abundance, 348 Gen3 and Gen6 displayed a reorganization of the plastidial β-CAs. BCA1 and BCA2 were 349 reduced in Gen3 and partially recovered in Gen6, whereas BCA5 increased significantly in 13 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 350 Gen6. Together with the preserved PSII competence, reduced photoinhibition and persistently 351 lower headspace CO 2 concentrations, these changes are consistent with altered regulation of 352 intracellular CO 2 /HCO₃⁻ pools during repeated NP functionalization. Importantly, RuBisCO 353 subunits rbcL and RBCS-1A remained broadly comparable to controls in Gen3 and Gen6, 354 indicating that the long-term response is not explained simply by increased RuBisCO 355 abundance. Instead, the combined physiological and proteomic data point to broader 356 reorganization of inorganic carbon handling, RuBP regeneration, photosynthetic energy 357 production and carbon recycling. 358 Overall, gPEI–Chi NP functionalization over successive subcultures consistently reduced net 359 headspace CO 2 accumulation and progressively increased chlorophyll content, while inducing 360 a stage-dependent reorganization of photosynthetic function. Gen3 exhibited the greatest 361 photosynthetic electron transport capacity, whereas Gen6 combined recovered PSII quantum 362 efficiency and sustained maximal electron transport. Together with the accompanying 363 proteomic changes, these findings indicate that sustained NP functionalization progressively 364 reorganizes carbon handling and photosynthetic metabolism rather than uniformly enhancing 365 a single carbon-fixation pathway. 366 14 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 367 368 369 **Fig. 3│gPEI-Chi NPs enhance CO** **2** **capture and modulate photosynthetic activity in plant** 370 **cell suspension cultures. a**, Schematic of the closed loop system allowing to measure the CO 2 371 concentration in the headspace of sealed culture flasks (created in BioRender.com). **b**, CO 2 372 concentration measured in the headspace of sealed culture flasks over 7 days for control (Ctrl), 15 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 373 NP-treated (Gen0), and multi-generational cultures (Gen3 and Gen6). NP-treated samples 374 exhibit a greater reduction in headspace CO 2 compared to controls, while NP-only controls 375 show minimal variation. **c**, Quantification of total chlorophyll content in control and NP-treated 376 cells, showing increased chlorophyll accumulation in functionalized cultures, indicative of 377 enhanced photosynthetic capacity. **d**, (i) Maximum quantum yield of photosystem II (Fv/Fm) 378 and (ii) maximum electron transport rate (ETR max ) measurement for control cultures and NP 379 treated cells (Gen0, Gen3 and Gen6), demonstrating unimpaired basal PSII photochemical 380 performance in all the conditions and improved maximum photosynthetic electron transport 381 capacity in Gen3. **e**, Heatmap of protein level in control and NP-treated samples quantifying 382 protein abundance involved in carbon handling, RuBisCO and Calvin cycle, electron transport, 383 photorespiration and TCA. Comparisons between treated and treated cultures (Gen0 vs Ctrl, 384 Gen3 vs Ctrl, Gen6 vs Ctrl) and between early functionalized cells and high generation cells 385 (Gen3 vs Gen0, Gen6 vs Gen0, Gen3 vs Gen6) were performed. Each box reports a color 386 coded log2 fold-change between the analysed and reference condition, for n=3 biological 387 replicates. Red indicates the protein is upregulated compared to the reference condition; blue 388 indicates the protein is downregulated compared to the reference condition; white indicates no 389 change or protein not detected. Comparisons between conditions were performed and p-values 390 after multiple hypothesis testing are reported in the heatmap as level of significance: * p<0.05, 391 ** p<0.01, *** p<0.001. 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 16 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 407 **NP functionalization alters cell wall regeneration dynamics** 408 The plant cell wall is a major sink for photosynthetically derived carbon. We therefore 409 investigated whether gPEI–Chi NP functionalization across successive subcultures altered cell 410 wall regeneration using protoplasts as a single-cell model [39] . Control and NP-treated cultures 411 were enzymatically digested after 4 days of growth, corresponding to active proliferation and 412 high NP internalization, and transferred to cell wall regeneration medium. Viability, 413 chlorophyll content and cell wall regeneration were monitored over 72 h (Fig. 4). 414 Control and NP-treated protoplasts showed comparable viability immediately after isolation 415 and subsequently underwent morphological recovery and cell division (Fig. 4b,c and 416 Supplementary Figs. S16-S17). Viable cell density increased over time across conditions and 417 remained predominant despite an increase in dead cells at 72 h. No significant differences in 418 viability were detected between control and NP-treated groups at individual time points, 419 indicating that NP functionalization did not compromise protoplast viability or regenerative 420 capacity (Supplementary Table 3-6, Note 2). 421 422 Cell wall-associated fluorescence progressively increased in all conditions, but with distinct 423 regeneration dynamics (Fig. 4d,e and Supplementary Figs. S18–S22; Supplementary Table 7 424 8, Note 3). Immediately after isolation, control and Gen0 protoplasts showed minimal wall 425 associated fluorescence, whereas Gen3 and Gen6 retained higher and more heterogeneous 426 signals, consistent with less efficient digestion of the more aggregated cultures. During 427 regeneration, control and Gen0 showed pronounced increases in wall-associated fluorescence 428 and wall coverage, whereas changes were smaller and delayed in Gen3 and Gen6. Broad 429 fluorescence distributions across conditions further revealed substantial cell-to-cell 430 heterogeneity in wall regeneration. Together, these results indicate altered and slower wall 431 regeneration in Gen3 and Gen6 cells. 432 433 Chlorophyll fluorescence also differed between conditions (Fig. 4d,f). Gen3 and Gen6 434 protoplasts initially exhibited higher and more homogeneous chlorophyll autofluorescence, 435 consistent with their higher chlorophyll content before isolation, and remained comparatively 436 stable during regeneration. In contrast, control and Gen0 populations progressively shifted 437 towards higher chlorophyll fluorescence. Intracellular NP-associated fluorescence declined 438 during regeneration, particularly in Gen3 and Gen6, while protoplast area remained comparable 439 across conditions (Supplementary Fig. S23). 440 17 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 441 Overall, these single-cell analyses demonstrate that gPEI–Chi NP functionalization preserves 442 protoplast viability and regenerative capacity while altering the kinetics and heterogeneity of 443 cell wall regeneration. These distinct regeneration trajectories suggest that functionalization 444 across successive subcultures establishes a cellular state with altered cell wall regeneration 445 processes. We therefore investigated whether these changes were also reflected in the 446 composition and organization of mature cell walls. 18 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 447 448 **Fig. 4│NPs functionalization alters cell wall regeneration dynamics in dedifferentiated** 449 **plant cells. a,** Schematic representation of the protoplasting workflow to generate 450 dedifferentiated cell wall-free protoplasts from suspension cultures (created in 19 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 451 BioRender.com). **b,** Protoplasts viability was tracked over three days of culture based on Evans 452 blue staining for control and Gen0 cells and multi-generational NP-treated cell cultures (Gen3 453 and Gen6). **c,** Quantification on the brightfield images showed that NP functionalization does 454 not impair short-term viability. **d,** An experimental pipeline for cell wall regeneration and 455 chlorophyll analysis used chlorophyll autofluorescence and Calcofluor White (CFW) staining 456 to image protoplasts for 3 days in control, Gen0, Gen3 and Gen6 protoplasts to track, 457 respectively, cell wall deposition and chlorophyll dynamics in regenerating protoplasts. In the 458 pipeline, from fluorescence images, cell wall **(e)** and chlorophyll **(f)** mean intensity of 459 fluorescence were quantified per protoplast among the 3-days culture period and represented 460 as scatter and violin plots, highlighting the median (filled square), mean (empty diamond) and 461 first and third quartiles (Q1 and Q3). 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 20 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 485 486 **NP functionalization progressively remodels cell wall architecture** 487 488 Monosaccharide profiling revealed progressive changes in cell wall composition following NP 489 functionalization (Fig. 5a and Supplementary Fig. S24). After one week, Gen0 cultures showed 490 increased arabinose (+65.3%) and decreased rhamnose (−16.2%) and galacturonic acid 491 (−6.5%), indicating an early shift in pectin composition. After two weeks, increased arabinose 492 (+36%) and decreased galacturonic acid (−9.3%) persisted, accompanied by increased xylose 493 (+19.3%). Gen3 and Gen6 similarly showed increased arabinose and xylose and decreased 494 galactose relative to age-matched controls, while rhamnose and galacturonic acid remained 495 largely unchanged. Glucose was also reduced in NP-treated cultures, particularly in Gen3 496 (−20.9%) and Gen6 (−17.2%), indicating broader remodelling of wall polysaccharide 497 composition. 498 499 Cellulose quantification supported these compositional changes (Fig. 5b). Total cellulose 500 remained relatively stable, with a modest reduction in Gen0 after one week and a gradual 501 decrease at later stages, particularly in Gen6. Calcofluor White imaging further revealed 502 transient cellulose-rich puncta following NP addition, persisting until day 3 in Gen0 and day 1 503 in Gen3 and Gen6, but absent from controls (Supplementary Figs. S25, S26), indicating 504 localized reorganization of cellulose-rich wall domains. 505 506 To further investigate pectin organization, homogalacturonan (HG) immunolabeling revealed 507 predominantly low- or non-methylesterified HG (LM19), with negligible highly 508 methylesterified HG (LM20) across conditions (Fig. 5c and Supplementary Fig. S27). 509 Although LM19- and cellulose-positive regions showed no major differences at cell– 510 extracellular interfaces, both domains became broader at cell–cell interfaces in NP-treated 511 cultures (Fig. 5d and Supplementary Fig. S28, Table 9). In Gen3 and Gen6, LM19-positive 512 regions increased in apparent width by 26% and 29.1%, respectively, accompanied by increases 513 of 22.3% and 29.8% in cellulose-positive domains. Thus, rather than substantially altering HG 514 methylesterification, NP functionalization reorganized the spatial distribution of pectin- and 515 cellulose-rich domains, particularly at cell–cell contact sites. The abundance of free carboxyl 516 groups in low-methylesterified HG may also favour interactions with cationic gPEI–Chi NPs 517 and contribute to their initial association with the cell wall [40,41] . 518 21 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 519 520 Cell wall ultrastructure provided further evidence of remodelling (Fig. 5e, Supplementary Fig. 521 S29). Control cultures displayed smooth cell walls with homogeneous contrast and regular 522 thickness. Gen0 walls retained a smooth appearance and regular thickness but showed 523 increased contrast, whereas Gen3 and Gen6 developed a more granular ultrastructure. These 524 changes may result from altered reorganization of cell wall fibre arrangement together with 525 increased material accumulation, or changes in composition that enhance heavy metal staining. 526 527 Proteomic analysis supported extensive remodelling of cell wall biosynthesis and turnover 528 (Fig. 5f and Supplementary Table 10). Gen3 and Gen6 showed increased abundance of proteins 529 involved in sucrose metabolism and UDP-glucose production (SPS1 and UGP1), together with 530 changes in starch metabolism (SS1 and PHS1), consistent with broader reorganization of 531 carbon metabolism. Proteins involved in pectin biosynthesis and remodelling were also altered. 532 Changes in GAUT proteins indicated reorganization of HG synthesis, while reduced abundance 533 of RG-I-remodelling enzymes, including BXL1, ARAF1 and BGAL2/5/6, suggested decreased 534 side-chain turnover. Together with the monosaccharide profiles, these changes support 535 progressive reorganization of pectic domains rather than a general increase in pectin synthesis. 536 537 Cellulose-associated CESA3 showed only limited changes, consistent with the absence of a 538 major increase in total cellulose. In contrast, proteins involved in xyloglucan remodelling 539 (XTH4, XYL1), wall loosening (EXPA4) and polysaccharide rearrangement (BXL4), were 540 reduced in NP-treated cultures. These changes potentially favour a denser and less dynamic 541 architecture through stronger interactions between cellulose microfibrils and matrix 542 polysaccharides, consistently with the increased contrast and granular wall ultrastructure 543 observed in Gen3 and Gen6. 544 545 Changes in cytoskeletal organization and intracellular trafficking may further contribute to the 546 altered wall regeneration observed in these cultures (Fig. 4d,e). Microtubule-associated 547 proteins involved in nucleation and stabilization (TUBG1, TUBG2 and SPR1) increased 548 without corresponding increases in proteins required for higher-order microtubule organization 549 (AIR9 and MOR1), while myosin-associated motor activity decreased in Gen3 and Gen6 550 (Supplementary Fig. S30 and Table 11), potentially affecting the spatial organization and 551 delivery of wall components during regeneration. 552 22 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 553 Together, these results demonstrate that gPEI–Chi NP functionalization progressively 554 remodels cell wall composition and architecture. Changes in pectin biosynthesis and turnover, 555 xyloglucan remodelling and wall-loosening processes are accompanied by broader pectin- and 556 cellulose-rich domains at cell–cell interfaces and increasingly granular wall ultrastructure. This 557 progressive wall remodelling provides a structural basis for the altered regeneration dynamics 558 and increased cohesiveness of NP-functionalized cell aggregates. 559 23 560 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 24 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 561 **Fig.5 | NP functionalization changes cell wall composition and architecture. a**, 562 Monosaccharide composition of extracted cell wall from control, Gen0, Gen3 and Gen6 563 cultures represented as mean ± standard deviation of moles percentage (n=3). Ara, Rha, Xyl, 564 Gal and GalA represent arabinose, rhamnose, xylose, galactose, galacturonic acid, respectively. 565 **b,** Cellulose content was evaluated as the percentage of mg of cell wall dry weight and 566 represented as mean ± standard deviation (n=3). **c,** Immunofluorescence confocal images were 567 acquired from cryo-thin sections of cells treated and non-treated with NPs, followed by 568 immunostaining with monoclonal antibodies associated to low methylesterified pectin 569 homogalacturonan (LM19) and Calcofluor White (CFW) for cellulose labelling in the cell wall. 570 **d,** Full width at half maximum (FWHM) was quantified for cellulose and pectin fluorescence 571 signal in the cell wall at the cell-cell interface to evaluate the effect of NP functionalization on 572 cell wall thickness. **e,** Transmission electron microscopy images of control, Gen0, Gen3 and 573 Gen6 cells showing ultrastructural changes both at the level of intracellular starch and 574 chloroplasts and cell wall contrast. **f,** Heatmap of protein level in control and NP-treated 575 samples quantifying protein abundance involved in biomass production and carbohydrate 576 metabolism, pectin synthesis and remodelling, cellulose, hemicellulose and wall expansion. 577 Comparisons between treated and treated cultures (Gen0 vs Ctrl, Gen3 vs Ctrl, Gen6 vs Ctrl) 578 and between early functionalized cells and high generation cells (Gen3 vs Gen0, Gen6 vs Gen0, 579 Gen3 vs Gen6) were performed. Each box reports a color-coded log2 fold-change between the 580 analysed and reference condition. Red indicates proteins upregulated in the analysed condition; 581 blue indicates proteins upregulated in the reference condition; white indicates no change or 582 protein not detected. Comparisons between conditions were performed and p-values after 583 multiple hypothesis testing are reported in the heatmap as level of significance: * p<0.05, ** 584 p<0.01, *** p<0.001. 585 586 587 **Conclusion** 588 Here, we show that synthetic nanomaterials can be used to progressively engineer proliferating 589 plant cell populations across successive subcultures. gPEI–Chi NPs were efficiently 590 internalized by photosynthetic Arabidopsis suspension cells without compromising viability or 591 proliferation. Supplying fresh NPs as the cultures were propagated enabled newly generated 592 cells to be successively functionalized, leading to progressive changes in cellular and 593 multicellular phenotype that were not apparent after the initial functionalization cycle. 25 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 594 This process extended beyond NP uptake. Successive functionalization reorganized carbon 595 handling and photosynthetic physiology, producing persistently lower net headspace CO 2 596 accumulation, increased chlorophyll content and stage-dependent changes in photosynthetic 597 electron transport and associated metabolic pathways. In parallel, NP functionalization altered 598 cell wall regeneration and progressively remodelled mature wall composition, spatial 599 organization and ultrastructure. Broader pectin- and cellulose-rich domains at cell–cell 600 interfaces, together with changes in wall biosynthesis and remodelling pathways, provide a 601 structural basis for the increasingly cohesive multicellular aggregates that emerged across 602 successive subcultures. Thus, the response to the NPs develops across multiple levels of 603 biological organization, linking metabolic and photosynthetic changes with remodelling of the 604 extracellular architecture. 605 These findings establish proliferating plant cell cultures as a platform in which nanomaterial 606 functionality can be introduced as new cells are generated, allowing nano–bio interactions to 607 shape the evolving phenotype of a living population. More broadly, this approach provides a 608 route to engineer plant cells beyond transient modulation of individual physiological processes, 609 coupling synthetic functionality to the intrinsic capacity of living cells to proliferate, remodel 610 their metabolism and construct their own structural environment. Such population-level 611 nanobionic engineering could provide a foundation for plant biohybrid systems and engineered 612 living materials in which cellular physiology and material organization develop together over 613 time. 614 615 616 26 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 617 **Materials and Methods** 618 619 _Preparation of gPEI-Chi nanoparticles:_ The full synthesis procedure for gPEI-Chi 620 nanoparticles and their characterization can be found in the supporting information of our 621 previous study [26] . 622 623 _Cell culture preparation and subculture:_ All culture work and manipulations were conducted 624 in sterile conditions. Calli for the preparation of stable cell lines and cell suspension cultures 625 were obtained from root explants of Arabidopsis thaliana wild-type (Col-0), following the 626 procedure described by _Pesquet et al._ _[42]_ . The calli were grown on Murashige & Skoog (MS) 627 medium (4.4 g/L, Duchefa Biochemie B.V., Haarlem, Netherlands) supplemented with 0.8% 628 ( _w/v_ ) agar (Sigma Aldrich, St. Louis, MO, USA) under dark conditions at 22°C. Once 629 sufficiently grown, calli pieces measuring approximately 0.5x0.5cm were transferred to 6-well 630 plates containing 3 mL of MS medium with 3% ( _w/v_ ) sucrose (Sigma Aldrich, St. Louis, MO, 631 USA) adjusted to a pH of 5.7. The calli pieces were dispersed into smaller fragments using 632 sterile tweezers to facilitate cell release into the suspension medium. Sealed well plates were 633 then placed back in the incubator (Pol-Eko aparatura, Model ST 700 CM, Poland) under dark 634 conditions at 22°C with constant rotational agitation at 120 rpm. The plates were monitored 635 daily, and once cell density was sufficiently high, 3 mL of cell suspension was transferred to 636 100mL Erlenmeyer flasks containing 27 mL of 3 % ( _w/v_ ) sucrose MS medium. The flasks were 637 sealed with cotton and aluminium foil and incubated in the dark at 22°C under continuous 638 shaking at 120 rpm. After an adaptation period, 8 mL of cell culture was transferred to larger 639 250mL Erlenmeyer flasks containing 72 mL of MS medium with 3 % ( _w/v_ ) sucrose. These 640 flasks were again sealed with cotton and aluminium foil and maintained in the dark at 22°C 641 while shaking at 130 rpm. Liquid cultures were subcultured weekly by directly performing a 642 1:10 dilution of cells in fresh culture medium (MS, 3 % ( _w/v_ ) sucrose, pH 5.7) to promote 643 continuous cell proliferation and maintain culture health. 644 Photosynthetic cell cultures were generated starting by established cell cultures kept in dark 645 conditions. 8 mL of cells cultured in dark incubator were transferred in a 250 mL Erlenmeyer 646 flask supplemented with 72 mL of MS medium with 1% (w/v) sucrose (pH 5.7) and then kept 647 under continuous exposure to white light (150 μmol photons m [-2] s [-1] ) and constant shaking at 648 130 rpm and 22°C, with the final aim to induce the development of chloroplasts, following an 649 established protocol from Dubreuil et al. [33] . The liquid cultures were subcultured weekly, by 27 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 650 replacing the medium with fresh MS 1% sucrose first and then transferring the photosynthetic 651 cells in the new flask with a 1:10 dilution in fresh MS 1% sucrose. 652 _Nanoparticle incubation in the cell culture:_ The gPEI-Chi nanoparticles were dissolved in 653 sterile MS medium supplemented with 3 % ( _w/v_ ) sucrose for cells in dark conditions and 1% 654 (w/v) sucrose for cells under light conditions, at a concentration of 0.3 mg/mL and agitated 655 until completely dissolved, following by a filtering step with a 0.22 µm sterile filter for ensuring 656 sterile conditions. During the weekly subculture step, the cell culture medium was replaced 657 with fresh medium and gently mixed. In a new flask, the 0.3 mg/mL gPEI-Chi NPs solution 658 was diluted in fresh medium at a final concentration of 0.1 mg/mL in a total volume of 72 mL. 659 Therefore, 8 mL of refreshed cell suspension was transferred to the new flask and gently mixed. 660 The flask was sealed with cotton and aluminium and returned to the dark or light incubator 661 under shaking and at 22°C. Cells were then monitored daily and utilized for all subsequent 662 experiments as needed. The cells treated in the previously described way are reported as 663 “Gen0” cultures, indicating a generation of cells exposed for the first time to the NPs. 664 _Multi-generational NPs-treated cultures_ : To underscore the long-term effect of continuous NP 665 exposure on photosynthetic cell cultures, cells were subcultured by weekly adding fresh NP 666 solution. In particular, starting with the Gen0 flask and every 7 days, the cell culture medium 667 was replaced with fresh medium supplemented with NPs at a final concentration of 0.1 mg/mL 668 and gently mixing them. 8 mL of cell suspension supplemented with NPs was transferred in a 669 new flask keeping the regular 1:10 ratio in fresh medium supplemented with 0.1 mg/mL of 670 NPs. This process was repeated for several weeks, reaching namely the “Gen3” after 3 weeks 671 and the “Gen6” after 6 weeks, used as middle and final generations, respectively, for 672 subsequent characterization experiments for long-term NP exposure. 673 _Growth curve measurements and cell volume after sedimentation (CVS):_ To assess cell volume 674 after sedimentation, a custom plastic vial holder was fabricated using a 3D printer (Ultimaker, 675 model 2+ Extended, Utrecht, Netherlands). This holder was specifically designed to 676 accommodate Erlenmeyer flasks containing the cultures, allowing them to be tilted at a 60° 677 angle following a previously reported method [43,44] . After tilting, the flasks were held in position 678 for five minutes to facilitate sedimentation of the cells. A ruler integrated into the apparatus 679 enabled measurement of the height of the cell mass along the glass wall of the flasks and the 680 cell volume was determined by referencing a calibration curve established using known 681 volumes of water and their corresponding heights measured on the glass wall for flasks of 28 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 682 similar volumes and tilt angles. The resulting volume of the cells, defined as the cell volume 683 after sedimentation (CVS), was then expressed per litre of suspension. 684 685 _Dry weight measurement_ : Paper filter discs with a porosity of 100 mm were previously dried 686 in oven at 60°C to remove the accumulated humidity and then weighed. 1 mL of cells was 687 collected by each flask and filtered on the pre-weighed paper filter disc by using a Büchner 688 vacuum funnel. The filtered cells on the paper disc were transferred in oven at 60°C for 24 h. 689 The filter with dried cells was then weighed, registering the stable weight. The dry weight of 690 the cells was calculated by subtracting the weight of the filter paper to the weight of cells plus 691 filter. 692 693 _Evans Blue staining and viability:_ Evans Blue staining was carried out by first transferring 694 1mL of cell culture (with or without gPEI-Chi nanoparticles) into a small 2mL vial. Evans blue 695 dye (DirectBlue 53, Sigma Aldrich, St. Louis, MO, USA) stock solution was then added to 696 give a final concentration of 0.025% ( _w_ / _v_ ). The cells were stained at room temperature for 5 697 min before immediately being mounted for microscopy for further analysis. In light microscopy 698 the living cells then appear unstained while dead cells appear blue stained. 699 700 _Cell wall Calcofluor White staining:_ Cell wall of plant cells was staining by using Calcofluor 701 White stain. 1 mL of cell suspension was incubated with the staining solution at a concentration 702 of 2.5 mM for 10 min before mounting for confocal microscopy inspection. 703 704 _Confocal Microscopy Imaging:_ Utilizing an inverted Zeiss LSM 980 confocal microscope 705 equipped with an Airyscan2 detection unit, images were captured and processed using Zen 706 Blue software version 3.4. Micrographs were acquired with a EC Plan-Neofluar 5x/0.16 M27, 707 Plan-Apochromat 20x/0.8 M27 and Plan-Apochromat 63x/1.4 Oil DIC M27 objectives. 708 Fluoresceine isothiocyanate (FITC) from the gPEI-Chi nanoparticles was excited with a 488 709 nm laser and detected using the AG520 filter (emission wavelength range 517-534 nm). 710 Chlorophyll autofluorescence was acquired by exciting the sample with a 639 nm laser and 711 collecting the signal by using a Chlorophyll A filter in the emission wavelength range 667-694 712 nm. Evans Blue fluorescence was detected by exciting it with a 639 nm laser and detected using 713 the Alexa Fluor 680 filter (emission wavelength 702 nm). Calcofluor White stained cell wall 714 was acquired exciting at 405 nm and collecting by using the filter CFW in the wavelength range 29 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 715 410-481 nm. LM19 and LM20 expression in the cell wall was detected by exciting with a 639 716 nm laser and collecting the emission within the range 650-700 nm. 717 718 _Absorption and fluorescence spectroscopy for NPs internalization:_ Absorption spectra for 719 culture medium both with and without gPEI-Chi NPs were recorded on a microplate reader 720 Biotek Synergy H1 in the range 300-700 nm with a step of 3 nm by using a Quartz 96 well 721 plate. Fluorescence endpoint of culture medium with and without NPs was recorded at 722 excitation 485 nm and emission 520 nm with gain fixed to 70. 723 724 _CO_ _2_ _and pH measurements:_ A closed-loop CO 2 measurement system was built to directly 725 deliver the CO 2 contained in the cell culture flask headspace to a CO 2 sensor (Sensirion I2C 726 SCD30) and an Arduino microcontroller board, pumped with a diaphragm gas micropump 727 (NMP05KPDC-B 6V, Knf), as represented in Fig. 3a. In particular, the flask was sealed and 728 connected to a water trap by means of impermeable tubes, after which a hydrophobic filter was 729 used to maintain sterile condition during the measurement period. The diaphragm pump 730 delivered the headspace CO 2 from the flask to the sensor, which was sealed in a hermetic Petri 731 dish to ensure any gas permeation during the measurement. CO 2 was recycled back to the flask 732 to measure the CO 2 consumption/release during the observation time. The recording lasted 7 733 days and a registration step of 5 min was used for all the measured cultures. 734 Along with CO 2 detection, pH was also measured every day for one week by using a Thermo 735 Scientific Orion Star A211 pH Benchtop Meter. 20 mL cell culture medium was removed every 736 day from the flask, filtered and used for measurement. A different flask was used every day to 737 avoid altering the cell culture condition and the CO 2 -pH balance during handling and medium 738 removal. 739 740 _Chlorophyll extraction and measurements:_ Chlorophyll was extracted from the control and 741 NP-treated photosynthetic cell cultures at the seventh and fourteenth day of culture. For the 2 742 weeks conditions, after one week of culture, cell medium was refreshed without subculturing, 743 keeping the culture for another week until the chlorophyll extraction. The extraction protocol 744 was readapted from an optimized procedure [45] . The cell suspension was collected from the flask 745 and dried with a paper filter disc by using a Büchner vacuum funnel. 500 mg of dry sample 746 was weighed out for the extraction and transferred to a pestle, previously cooled down with 747 liquid nitrogen. 260 mg of magnesium sulphate heptahydrate was added to the cells in the 748 pestle, along with few droplets of liquid nitrogen and 15 mL of acetone. The mixture was 30 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 749 grinded until complete homogenization and transferred and left for 15 min in a sealed beaker, 750 after which it was filtered with a strainer with 20 μm mesh size and left to settle for other 10 751 min. The solution was centrifuged at 200 g for 5 min and the supernatant was used for 752 absorbance measurement. A microplate reader Biotek Synergy H1 was used to record the 753 absorbance spectra of the chlorophyll extracts between 350-700 nm with 3-nm step by using a 754 Quartz 96 well. Absorbance at 647 nm and 665 nm, respectively was also collected and used 755 for the chlorophyll concentration calculation. Chlorophyll a and b concentrations were 756 calculated using the extinction coefficient reported by Inskeep et al [46] : 757 𝐶ℎ𝑙 𝑏 = 20.47 ∙𝐴 647 −4.73 ∙𝐴 665 758 𝐶ℎ𝑙 𝑎 = 12.63 ∙𝐴 664 −2.52 ∙𝐴 647 759 𝐶ℎ𝑙 𝑇𝑜𝑡𝑎𝑙 = 𝐶ℎ𝑙 𝑎 + 𝐶ℎ𝑙 𝑏 760 761 The final concentration of chlorophyll was computed as: 762 763 764 𝑉 𝑎𝑐𝑒𝑡𝑜𝑛𝑒 ∙ 𝐶 𝐶ℎ𝑙 = 𝐶ℎ𝑙 𝑇𝑜𝑡𝑎𝑙 𝑚 𝑑𝑟𝑦 𝑠𝑎𝑚𝑝𝑙𝑒 765 Where 𝑉 𝑎𝑐𝑒𝑡𝑜𝑛𝑒 is the total volume of acetone used for the extraction and 𝑚 𝑑𝑟𝑦 𝑠𝑎𝑚𝑝𝑙𝑒 is the 766 mass of the dried cell culture. 767 768 _Photosynthetic characterization:_ The photosynthetic performance of the samples was 769 evaluated using a portable photosynthesis system (LI-6800, LI-COR Biosciences) equipped 770 with a Multiphase Flash Fluorometer (6800-01A). The fluorometer was maintained in a 771 horizontal position, and a custom 3D-printed black chamber was used to hold a Petri dish 772 containing each liquid sample. Samples were dark-acclimated for 30 min before measurement. 773 Illumination consisted of 90% red light (625 nm) and 10% blue light (475 nm). 774 Measurements were first performed under dark-adapted conditions to determine the minimum 775 fluorescence (F 0 ) and maximum fluorescence (F m ). F m was obtained by applying a rectangular 776 saturating pulse with a duration of 1 s and an intensity of 6000 µmol photons m [-2] s [-1] . 777 Subsequently, the samples were exposed to increasing actinic light intensities of 10, 20, 50, 778 125, 150, 175, 300, 600, 900, 1200, and 1500 µmol photons m [-2] s [-1] . At each light intensity, the 779 steady-state fluorescence (F s ) and the maximum fluorescence under illumination (F m [’] ), 780 obtained following application of the saturating pulse, were recorded. 31 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 781 From the fluorescence measurements, several photosynthetic parameters were calculated [47] . 782 The maximum quantum yield of photosystem II (PSII), ( 𝐹 𝑣 𝐹 𝑚 ~~[)]~~ [, was calculated under dark-adapted ] 783 conditions as: 784 𝐹 𝑣 𝐹 𝑚 [=] 𝐹 𝑚 −𝐹 𝑜 785 𝑣 𝑚 𝑜 (Eq.1), 𝐹 𝑚 [=] 𝐹 𝑚 786 787 788 The effective quantum yield of PSII, (𝜙 𝑃𝑆𝐼𝐼 ), at each actinic light intensity was calculated as: 𝑚 𝑠 789 𝜙 𝑃𝑆𝐼𝐼 = (Eq.2) 𝐹 𝑚 [′] = 𝜙 𝑃𝑆𝐼𝐼 𝐹 𝑚 [′] −𝐹 𝑠 790 The electron transport rate (ETR) was subsequently estimated as: 791 𝐸𝑇𝑅= 𝜙 𝑃𝑆𝐼𝐼 × 𝐸 𝑃𝐴𝑅 × 𝑓 × 𝐴𝐹 (Eq.3) 792 where (PAR) is the incident photon flux density integrated over the photosynthetically active 793 radiation region (400–700 nm), 𝑓 is the fraction of absorbed light allocated to PSII, assumed 794 to be 0.5, and AF is the absorption factor, assumed to be 0.83. The latter should be considered 795 an approximation, as the fraction of absorbed light can vary depending on parameters such as 796 pigment content and cell density within the sample. 797 The resulting ETR vs light intensity curves were fitted using a custom python code and the 798 model described by Platt and Gallegos (1980) [48] : − − ~~(~~ ~~)~~ ~~(~~ ~~)~~ 799 𝐸𝑇𝑅= 𝐸𝑇𝑅 𝑠 (1 −𝑒 𝐸𝑇𝑅𝑠 ) × (𝑒 𝐸𝑇𝑅𝑠 ) (Eq.4) − ~~(~~ [𝛽][ × 𝐸][𝑃𝐴𝑅] 𝐸𝑇𝑅= 𝐸𝑇𝑅 𝑠 (1 −𝑒 − ~~(~~ [𝛼 × 𝐸][𝑃𝐴𝑅] ~~)~~ 𝐸𝑇𝑅𝑠 ) × (𝑒 800 where 𝐸 𝑃𝐴𝑅 is the incident light intensity, 𝐸𝑇𝑅 𝑠 represents the theoretical maximum ETR in 801 the absence of photoinhibition, α describes photosynthetic efficiency under low-light 802 conditions and determines the initial slope of the ETR curve, and β describes the degree of 803 photoinhibition at high light intensities. Thus, larger values of α indicate a stronger increase in 804 ETR with increasing irradiance in the light-limited region, whereas larger values of β indicate 805 stronger photoinhibition at high irradiance. 806 The maximum value of ETR was calculated after fitted as followed: 𝛼 𝛽 𝛽 807 𝐸𝑇𝑅 𝑚𝑎𝑥 = 𝐸𝑇𝑅 𝑠 × ( ~~[)]~~ [ × (] ~~[)]~~ 𝛼 (Eq.5) 𝐸𝑇𝑅 𝑚𝑎𝑥 = 𝐸𝑇𝑅 𝑠 × ( 𝛼+𝛽 ~~[)]~~ [ × (] 𝛼 𝛼+𝛽 ~~[)]~~ 𝛽 32 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 808 809 _Proteomic sample preparation:_ Photosynthetic cell cultures, treated and non-treated with 810 gPEI-Chi NPs, were collected and pelleted, after which they were snap-frozen in liquid 811 nitrogen. Pelleted cells were ground to a fine powder using a mortar and pestle in liquid 812 nitrogen. The powdered pellet was thawed on ice and shortly equilibrated to room temperature. 813 sodium deoxycholate (SDC) was added to 0.5 mL pellet to a final concentration of 3% (w/v) 814 SDC. Cells were lysed using a FastPrep-5G homogenizer for six cycles of 45 s each. Samples 815 were diluted to a final concentration of <1% SDC, and cell debris was removed by 816 centrifugation at 21,000 × _g_ for 5 min. Protein concentration was determined by Bradford assay 817 with measured aliquots diluted to <0.1% SDC to avoid interference. For proteolytic digestion, 818 25 μg protein was reduced with 10 mM dithiothreitol (DTT) at 55 °C for 45 min and alkylated 819 with 17 mM iodoacetamide (IAA) for 30 min in the dark. Proteins were then digested overnight 820 (16 h) at 37 °C using a Trypsin/Lys-C Protease Mix (Thermo Scientific) at a protease-to-protein 821 ratio of 1:100. Peptides were then desalted using StageTips packed with six layers of C18 822 Empore SPE disks (Merck) and dried in a SpeedVac at 45 °C for 1 h. 823 824 _Liquid chromatography–mass spectrometry (LC–MS) analysis:_ Samples were resuspended in 825 0.1% formic acid and analysed using a Q Exactive HF Hybrid Quadrupole-Orbitrap Mass 826 Spectrometer operated in positive mode coupled to an UltiMate 3000 RSLCnano system 827 equipped with an EASY-Spray ion source. Peptides were loaded onto a C18 Acclaim PepMap 828 100 trap column (75 μm × 2 cm, 3 μm particle size, 100 Å pore size) at a flow rate of 7 μL/min 829 using solvent A (3% acetonitrile and 0.1% formic acid in water). Peptides were separated on 830 an ES802 EASY-Spray PepMap RSLC C18 analytical column (75 μm × 25 cm, 2 μm, 100 Å) 831 at 0.7 μL/min using a 40 min linear gradient from 1% to 32% solvent B (95% acetonitrile, 0.1% 832 formic acid in water). 833 Peptides were analysed using data-dependent acquisition (DDA), with full MS1 spectra 834 acquired at a resolution of 60,000 (at _m/z_ 200) over an _m/z_ range of 300–1200, followed by 835 MS2 scans of the 10 most abundant precursor ions. MS2 spectra were acquired at a resolution 836 of 60,000 using a 2 _m/z_ isolation window and 20 s dynamic exclusion. The MS1 and MS2 837 maximum injection times were 205 and 105 ms, with AGC targets of 1 × 10⁶ and 2 × 10⁵, 838 respectively. Precursor ions were fragmented by HCD at a normalised collision energy of 26. 839 840 _Proteomic data analysis:_ Raw MS data was analysed using MaxQuant (version 2.6.7.0). 841 Spectra were searched against the _A. thaliana_ FASTA reference proteome UP000006548 33 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 842 limited to reviewed entries. Trypsin/P digestion was used, allowing up to two missed cleavages, 843 and match-between-runs enabled. Carbamidomethylation of cysteine was set as a fixed 844 modification, while methionine oxidation and protein N-terminal acetylation were set as 845 variable modifications. Identifications were filtered at a 1% false discovery rate (FDR). 846 Proteins were median-normalized, and differential expression was calculated using MSstats 847 (version 4.4.1) in R [49] . Gene set enrichment analysis (GSEA) was performed in R using the 848 ClusterProfiler package (version 4.15.1) [50] . 849 850 _Protoplasts extraction:_ The protoplasts extraction protocol is an adaptation of the protocols 851 from Yoo et al. [51] and Bogdziewiez et al. [39] and it is summarized below. A protoplasting buffer 852 (PB) was prepared adding in deionized water 20 mM MES, 0.4 M mannitol, 20 mM KCl, 10 853 mM CaCl 2, 0.1% (w/v) BSA and adjusting the pH to 5.7. An enzymatic solution was prepared 854 for the cell wall lysis and protoplasts extraction by warming up the PB solution to 55°C under 855 stirring and adding 1.5% (w/v) Cellulase R-10 and 0.4% (w/v) Macerozyme R-10. Under sterile 856 conditions, 5 mL of 4-days-old control photosynthetic cell suspension and 15 mL of 4-days 857 old photosynthetic cells treated with gPEI-Chi NPs (Gen0, Gen3 and Gen6) were transferred 858 and the medium was replaced with 20 mL of the filtered enzymatic solution. Different volumes 859 of cell suspensions were used between treated and non-treated conditions, balanced over the 860 differences in CVS detected at day 4 of culture (Fig. 1e-i). Cells were gently mixed with the 861 digestion solution and 10 mL of the cell suspension was transferred to glass non-treated Petri 862 dishes, followed by incubation in dark conditions for 4 hours under mild rotation (74 rpm) at 863 22°C. A high-salt solution (W5 solution) was prepared to inhibit the enzymatic reaction, by 864 adding 2 mM MES, 154 mM NaCl, 125 mM CaCl 2, 5 mM KCl in deionized water and adjusting 865 its pH to 5.7. To stop the enzymatic digestion, the sterile W5 solution was added to the digested 866 cell culture in a ratio 1:1 and incubated back for 5 min under mild rotation (30 rpm). The cell 867 suspension was then filtered with a 70 µm cell strainer to isolate single protoplasts from bigger 868 clusters of cells. To remove debris and digested cell wall material, the protoplasts were 869 centrifuged at 200g for 5 min, the supernatant was removed and protoplasts were resuspended 870 in a cell wall regeneration medium (CWR) made by 0.316% (w/v) Gamborg B5, 20 mM MES, 871 0.25 M mannitol, 0.2 M glucose, 5 µM NAA in deionized water and calibrated to achieve pH 872 5.7. Protoplasts were counted, seeded in non-treated plastic Petri dishes at a density of 2 · 10 [5] 873 cells/mL in CWR and incubated at 22°C in dark conditions without shaking. Independent 874 samples were used every day for viability evaluation through Evans Blue assay, cell wall 34 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 875 regeneration detection with Calcofluor White staining and chlorophyll content quantification 876 at 0, 24, 48 and 72 h. 877 878 _Protoplasts’ wall regeneration - Data analysis:_ Every day starting by the day of protoplasting 879 (day 0), protoplasts were stained with Evans blue and loaded in prefabricated microchannel 880 chips (IBIDI µ-Slide VI Flat, non-treated) for imaging. Protoplasts were acquired with a 881 fluorescence microscope Nikon Eclipse Ni-E equipped with a 20x objective in brightfield with 882 a color-coded camera. At least 30 images per microchannel were acquired for each condition. 883 Dead cells were counted as blue-stained circular protoplasts, while live cells were considered 884 as round-shaped, intact and non-stained protoplasts. The count of live and dead cells was 885 normalized to the total acquired area in mm [2] . 886 In parallel, protoplasts were stained with Calcofluor White stain and loaded in a parallel 887 microchannel. Multichannel images were acquired with the Nikon Eclipse fluorescence 888 microscope with a 20x objective, acquiring consecutively brightfield, NP-associated 889 fluorescence (FITC filter, 488 nm), chlorophyll autofluorescence (Cy5 filter, 550 nm) and cell 890 wall-associated Calcofluor signal (DAPI filter, 370 nm). At least 20 images per microchannel 891 were acquired for each condition. A custom Python code was used for collecting the 892 multichannel images, creating 2-by-2 tiles of the separate 4 channels to create an hyperstack 893 for successive data analysis. Following the Q-Warg pipeline procedure previously developed 894 by Bogdziewiez et al. [39], a mask was generated by segmentation of the protoplasts’ brightfield 895 tiles by using CellPose, a deep learning-based ImageJ (Fiji) macro to automate objects 896 segmentation. A cyto4 pretrained model developed by Stringer and Pachitariu [52] was used to 897 obtain the labelled mask for each segmented protoplast. The Q-Warg pipeline was followed by 898 using an ImageJ macro extracting the morphometric parameters from the segmented labels, 899 including cell size and circularity, as well as to measure the mean intensity of fluorescence 900 related to chlorophyll, cell wall and gPEI-Chi NPs for each segmented protoplast, by using the 901 MorphoLibJ library. A custom Python script was instead used for previous data plotting 902 through an interactive GUI for manual selection of the parameters of interest for the analysis. 903 In particular, the mean fluorescence intensity for each detected protoplast was matched to the 904 corresponding morphometric and chlorophyll measurements using the label associated during 905 segmentation. Protoplasts were initially filtered according to morphology and chlorophyll 906 thresholding, retaining objects with a circularity greater than 0.88 and chlorophyll fluorescence 907 intensity greater than 100, aimed to represent only protoplasts classified as alive when they 908 retain the circular shape and possess chlorophyll. The data on cell wall and chlorophyll mean 35 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 909 fluorescence intensity for each protoplast were then normalized to the corresponding averaged 910 mean fluorescence of the protoplasts’ control population at day 3 for each experiment. 911 Differently, NPs mean fluorescence intensity for each protoplast was normalized to NPs 912 maximum fluorescence. These normalization steps were included to account for differences 913 between experimental sets and represent the data as relative increase/decrease of fluorescence 914 intensity with respect to maximum conditions of cell wall, chlorophyll content and NP-related 915 fluorescence. All the data related to normalized cell wall mean fluorescence, normalized 916 chlorophyll mean fluorescence, normalized NP mean fluorescence and protoplasts’ area were 917 visualized using violin plots of the distribution of the single protoplasts mean fluorescence 918 intensity and area for each experimental condition, underlying the mean, median and 919 interquartile range of the 25 [th] and 75 [th] percentile (Q1-Q3). 920 Cell wall coverage progression during the protoplasts culture in CWR for the 4-day observation 921 time was quantified using a custom ImageJ macro as the fraction of protoplast area positive for 922 cell wall staining. First, the fluorescence images of protoplasts were background-subtracted 923 and segmented following the CellPose procedure previously described. The background 924 corrected cell wall channel was converted into a binary mask by Otsu thresholding, classifying 925 pixels as either cell wall-positive or background. For each segmented cell, the mean intensity 926 of the binary mask was extracted and the percentage of the cell area occupied by cell wall 927 fluorescence was computed as: 928 𝐶𝑊 𝑐𝑜𝑣𝑒𝑟𝑎𝑔𝑒 [%] = [𝑀𝑒𝑎𝑛] 255 [∙100 ] 929 The mean ± standard deviation of cell wall coverage between the protoplasts’ population was 930 calculated for each condition and time point. 931 932 _Cell wall extraction and monosaccharides quantification_ 933 Photosynthetic cell cultures (control, Gen0, Gen3 and Gen6) were freeze-dried and 934 subsequently homogenized to a fine powder using a ball mill. Alcohol insoluble residue (AIR) 935 was prepared from 35 mg of ground material per sample by first incubating it with 80% ethanol 936 for 30 min at RT which was followed by 30 min incubation with chloroform: methanol (1:1). 937 The resulting pellet was dried and treated with 1000 units of α-amylase in 0.1M potassium 938 phosphate buffer to remove starch. After overnight incubation, the samples were successively 939 washed with Milli-Q water and acetone then dried overnight in a vacuum desiccator. 940 For monosaccharide composition analysis, around 500 μg of dry weight (DW) of the processed 941 AIR was derivatized via trimethylsilylation (TMS) and analysed by GC/MS (7890A/5975C; 36 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 942 Agilent Technologies) according to previously described protocols [53] . The degree of methyl 943 esterification was determined using around 1.5 mg DW of AIR. Samples were first saponified 944 with 1 M NaOH for 1 h at RT, neutralized with 1 M HCl, and approximately 180 µL of the 945 supernatant was collected for methanol quantification. Reactions were prepared in ELISA 946 plates: 50 µL of alcohol oxidase in 0.1 M sodium phosphate buffer was added to each sample 947 and incubated for 15 min at RT. Then, 100 µL of a developer solution (containing 948 acetylacetone, acetic acid, and ammonium acetate in water) was added, plates were sealed and 949 incubated for 10 min at 68°C. Absorbance was measured at 412 nm. The analysis of the 950 absorbances was done as previously described [54] . 951 Crystalline cellulose was isolated using the previously described Updegraff method [55,56] . The 952 obtained glucose was measured using the anthrone assay [57] . 953 Data was further analysed and the graphs were generated with the help of RStudio. 954 955 _Cryosectioning for cell wall composition:_ Photosynthetic plant cells, treated and non-treated 956 with gPEI-Chi NPs, were prepared for cryosectioning and immunostaining, with the aim to 957 characterize the cellulosic and pectinic contribution to the cell wall composition before and 958 after NP functionalization. In more details, cells were collected after 1 week of culture in light 959 conditions, washed 3 times in 1x PBS and fixed in 4% PFA for 30 minutes. Cells were then 960 washed 4 times in PBS and embedded in gradient of sucrose for reducing the formation of ice 961 crystals and protect cell structures from low temperatures damage. In particular, PBS was 962 removed and cells were gradually incubated in 10% (w/v) sucrose for 25 min, 20% (w/v) 963 sucrose for 45 min and overnight incubated at 4°C with 30% (w/v) sucrose. Sucrose was 964 removed and cells were transferred to cryomolds (Simport Scientificinc M475), covered with 965 optimal cutting temperature (OCT) compound (CellPath KMA-0100-00A) and snap-frozen in 966 liquid nitrogen vapours for few minutes until OCT solidification was achieved. Samples were 967 then stored at -80°C. Cryosections of 5-10 µm were obtained transferring the embedded cells 968 on cryostat stabs and cutting them with cryostat microblades (MX35 ultra, Thermo Scientific) 969 using a cryostat (Epredia CryoStar NX70) at -15°C. Sections were then collected onto adhesion 970 slides (Epredia™ SuperFrost Plus™, 10149870, Fisher Scientific), air-dried for few hours and 971 stored at 4°C. 972 973 _Immunofluorescence for cell wall composition:_ The sections were washed 3 times in 1x PBS 974 and once in PBS-T (1x PBS and 0.1% Tween 20) for 5 min. They were then incubated in 975 blocking buffer consisting of 1% (w/v) BSA in PBS-T for 1 h to prevent non-specific labelling. 37 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 976 After 3 washes in PBS-T, the sections were incubated with the anti-pectic monoclonal primary 977 antibodies LM19 and LM20 (Agrisera) diluted in 0.1% (w/v) BSA in PBS-T at a ratio of 1:20 978 for 2 h at room temperature in humid chamber and darkness. After 3 washes in PBS and 3 979 washes in PBS-T, the secondary antibody Alexa Fluor 647 (Invitrogen™ Goat anti-Rat IgM, 980 Heavy chain, Fisher Scientific) diluted 1:100 in blocking buffer 0.1% (w/v) BSA in PBS-T for 981 2 h at room temperature or overnight at 4°C in humid chamber and darkness. After washing in 982 PBS-T, the sections were also stained with Calcofluor White stain diluted at a concentration of 983 2.5 mM in PBS-T for cellulose imaging. An anti-fade medium made by mixing PBS and 984 glycerol (1:1) was used, the sections were sealed and stored at 4°C in darkness until imaging. 985 986 _Data analysis for cell wall composition_ : Confocal images were acquired for the sectioned cells, 987 acquiring multichannel images of LM19 and Calcofluor White fluorescence and LM20 and 988 Calcofluor White fluorescence. Pectin- and cellulose-associated wall thickness was estimated 989 from fluorescence intensity profiles acquired across LM19+LM20 and Calcofluor positive 990 regions both at the boundaries of individual cells and at the cell-cell interface between adjacent 991 cells. Prior to analysis, the images were processed in ImageJ by subtracting the background. 992 Multiple line were manually drawn approximately perpendicular to the cell boundary and 993 crossing the fluorescent cell wall signal, while the corresponding intensity profiles were 994 extracted as grey-value intensity versus distance. The line profiles were analysed using a 995 custom Python script. Here, for each line profile the fluorescence baseline (𝐼 𝑏𝑎𝑠𝑒𝑙𝑖𝑛𝑒 ) was 996 defined as the minimum intensity value while the peak (𝐼 𝑝𝑒𝑎𝑘 ) was defined as the maximum 997 intensity. The signal amplitude was then calculated as the difference between the peak and 998 baseline, retrieving the half-maximum intensity (𝐼 𝐻𝑀 ) as: 999 𝐼 𝐻𝑀 = 𝐼 𝑏𝑎𝑠𝑒𝑙𝑖𝑛𝑒 + [𝐼] [𝑝][𝑒𝑎𝑘] [−𝐼] [𝑏𝑎𝑠𝑒𝑙𝑖𝑛𝑒] 2 1000 The values at which the fluorescence profile crossed the half maximum level on either left and 1001 right side of the peak were determined by linear interpolation. The full width at half maximum 1002 (FWHM) was then calculated as the distance between the left and right half maximum crossing 1003 points. This provided an estimation of the apparent width of the cellulose- and pectin-positive 1004 area in the cell wall of NP treated and non-treated photosynthetic cells. At least 15 lines were 1005 used for line profile detection for each cell (for boundary interface) or pair of cells (for cell 1006 cell interface) and the resulting FWHM values were averaged to obtain a single mean value per 1007 cell. At least 10 cells were analysed per condition and the values were reported as mean ± 1008 standard deviation. 38 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1009 1010 _Transmission electron microscopy (TEM):_ Cells from 1-week control and NP-treated cultures 1011 were transferred and incubated with fixative solution made by 2.5% glutaraldehyde (EM grade) 1012 in 0.1 M phosphate buffer for 10 minutes. The solution was removed letting the cells settle and 1013 replaced with 2h postfixing 1% osmium tetraoxide (OsO 4 ) in water. After that the samples were 1014 dehydrated in ethanol gradient followed by resin infiltration (3:1, 1:1, 1:3 resin-to-pure ethanol) 1015 and embedded in pure Spurr’s resin ( TAAB Laboratories Equipment Ltd, England). To prepare 1016 sections the samples in resin were trimmed, and the presence of the cells were confirmed under 1017 a light microscope. Ultra-thin sectioning of approximately 70 nm thickness was done using an 1018 Ultra microtome (Leica EM UC7) and the sections were placed on hexagonal copper grids, 1019 with a formvar film. The ultra-thin sections were stained with uranyl acetate (EMS England 1020 22451) and lead citrate and examined with the Thermo Scientific Talos L120C transmission 1021 electron microscope. 1022 1023 _Statistical analysis:_ For the chlorophyll and photosynthetic characterization through pulse 1024 amplitude modulated (PAM) chlorophyll fluorescence, three technical replicated were 1025 measured for each biological replicate. Technical replicates were first averaged to obtain a 1026 mean value for each biological replicate. Normality and homogeneity of variance were tested 1027 by using Shapiro-Wilk and Levene’s test. One-way ANOVA followed by Tukey’s multiple 1028 comparison test were subsequently performed using the biological replicates as independent 1029 observations. Data are reported as the mean ± standard deviation among biological replicates. 1030 Significance levels were indicated as * p<0.05, ** p<0.01, *** p<0.001. 1031 In the proteomic data analysis, proteins differential expression were statistically tested with p 1032 values adjusted for multiple hypothesis testing using the Benjamini-Hochberg method. Proteins 1033 detected in fewer than n-1 replicates were excluded from statistical analysis. The log2 fold 1034 change (log2FC) values in the heatmaps were defined as the abundance in the first condition 1035 minus that in the reference condition, with positive and negative values indicating increased 1036 and decreased protein abundance, respectively. Significance levels were indicated as * p<0.05, 1037 ** p<0.01, *** p<0.001. In the gene set enrichment analysis (GSEA) proteins were ranked 1038 according to their t-statistic from the differential expression analysis and analysed using the 1039 GSEA function. Default settings were used (gene set size range of 10–500 and Benjamini– 1040 Hochberg correction for multiple testing) with a p-value cutoff of 0.05. 1041 Protoplasts viability was assessed by analysing live and dead cell densities. Homogeneity of 1042 variance among groups was evaluated using Levene’s test. As the assumption of equal 39 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1043 variances was satisfied, data were analysed using ordinary two-way ANOVA, with 1044 experimental condition and time as fixed factors, including their interaction. Planned post-hoc 1045 comparisons were performed to compare each NP-treated group with the corresponding control 1046 at individual time points and to evaluate temporal changes within each condition relative to 1047 day 0 of protoplasting. These multiple comparisons were corrected using the Holm method and 1048 differences were considered statistically significant at p<0.05 and indicates as * p<0.05, ** 1049 p<0.01, *** p<0.001. 1050 For protoplasts cell wall and chlorophyll content evaluation, mean intensity of fluorescence 1051 distributions were non-normally distributed and non-parametric rank-based statistics were 1052 applied. Differences among multiple groups were assessed using the Kruskal-Wallis test, 1053 separately comparing conditions within each time point and then time points within each 1054 condition. When significant differences were detected, pairwise comparisons were performed 1055 using the Mann-Whitney U test followed by Holm correction for multiple comparisons. 1056 Statistical significance was defined as adjusted p<0.05 and indicates as * p<0.05, ** p<0.01, 1057 *** p<0.001. 1058 Protoplasts’ cell wall coverage was analysed testing the mean values with a two-way ANOVA 1059 with condition and time as fixed factors and their interaction. Before, homogeneity of variance 1060 was assessed using Levene’s test. Planned post-hoc comparisons were performed among 1061 conditions at each time point and comparisons of each time point with day 0 within the same 1062 condition. P-values were adjusted using Holm correction for multiple testing. Statistical 1063 significance was defined as adjusted p<0.05 and indicates as * p<0.05, ** p<0.01, *** p<0.001. 1064 For the monosaccharides quantification analysis, statistically significant differences between 1065 the samples were determined using Kruskal-Wallis with Dunn’s test for pairwise comparisons 1066 (P<0.05) and Wilcoxon rank-sum test (P<0.05) where applicable. Statistical significance was 1067 defined as adjusted p<0.05 and indicates as * p<0.05, ** p<0.01, *** p<0.001. 1068 Cellulose and pectin thickness quantification at single-cell level was tested by analysing the 1069 mean FWHM measured across cell boundary and cell-cell contact, by comparing control and 1070 NP-treated cell cultures. Data distribution within each condition was assessed using the 1071 Shapiro-Wilk test, while homogeneity of variance among groups was tested using Levene’s 1072 test. When normality and equal variance assumptions were satisfied, differences among 1073 conditions were analysed using one-way ANOVA followed by Tukey’s multiple comparison 1074 test. When parametric assumptions were not met, the Kruskal-Wallis test was used, followed 1075 by pairwise Mann-Whitney U test with Bonferroni correction for multiple comparisons. 40 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1076 Statistical significance was defined as adjusted p<0.05 and indicates as * p<0.05, ** p<0.01, 1077 *** p<0.001. 1078 1079 1080 1081 41 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1082 **Conflict of Interest** 1083 The authors declare no conflict of interest. 1084 1085 **Acknowledgments** 1086 The authors wish to thank Agnieszka Ziolkowska (Umeå University) for TEM imaging and 1087 support for the discussion for ultrastructural characterization of functionalized cells. 1088 1089 The authors thank Therése Klingstedt (Department of Physics, Chemistry and Biology (IFM) 1090 – Linköping University) for support for cryostat sectioning of functionalized cells. 1091 1092 The work was primarily supported by the European Union (ERC-2021-STG, 4DPhytoHybrid, 1093 101042148) and the Knut and Alice Wallenberg Foundation (KAW project FATE 2022.0029). 1094 Additional funding was provided by the Swedish Government Strategic Research Area in 1095 Materials Science on Advanced Functional Materials at Linköping University (Faculty Grant 1096 SFO-Mat-LiU No. 2009-00971). 1097 1098 The authors acknowledge the facilities and technical assistance of the Umeå Core Facility 1099 Electron Microscopy (UCEM) at the Chemical Biological Centre (KBC), Umeå University, a 1100 part of the National Microscopy Infrastructure NMI (VR-RFI 2019-00217). 1101 This work was supported by grants from the Novo Nordisk Foundation (NNF21OC0067282) 1102 and Åforsk Foundation (20-502). 1103 This work was also supported by Umeå Plant Science Center with grants from the Knut and 1104 Alice Wallenberg Foundation (KAW 2016.0352 and KAW 2020.0240); the Wallenberg 1105 Initiatives in Forest Research (WIFORCE) funded by the Knut and Alice Wallenberg 1106 Foundation, the Swedish Governmental Agency for Innovation Systems (VINNOVA 2016 1107 00504); and Bio4Energy, a Strategic Research Environment supported through the Swedish 1108 Government's Strategic Research Area initiative. 1109 1110 **Author contributions** 1111 E.S., C.R. and D.P. conceived and designed the research. E.S. supervised the work. C.R. 1112 established the NP-treated cell culture and protoplasts models, performed initial experiments 1113 in NP internalization, set the CO 2 measurement apparatus and performed initial CO 2 42 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1114 experiments, performed CVS and dry weight measurements on both non- and photosynthetic 1115 cells, chlorophyll extraction and measurement. D.P. performed the NP internalization 1116 experiments and characterization, the CO 2 measurements and chlorophyll extraction and 1117 measurement, cell chlorophyll/wall and NP/chlorophyll colocalization imaging, pH 1118 measurements, protoplasts extraction and full characterization, LM19/20/cellulose 1119 immunofluorescence and analysis, proteomics data interpretation, and all the statistical 1120 analysis. A.K. and E.P.H. conducted the proteomics experiments and data analysis, providing 1121 support for the discussion. L.L. and S.V. performed the measurement and analysis on cell wall 1122 monosaccharides and cellulose content, and contributed to the discussion for cell wall 1123 composition. L.Boutonnet refined the pipeline for protoplasts cell wall and chlorophyll 1124 quantifications. L.Bogdziewiez and S.V. provided protocols and support for protoplasts 1125 extraction, maintenance and data analysis. X.J. and A.S. provided calli of root plant cells for 1126 starting the cell cultures and provided expertise in plant cell culture. E.C. established the 1127 protocol of transition from dark cell culture to light cell cultures and performed CVS and dry 1128 weight measurements of non-photosynthetic cells. A.G.F. performed the photosynthetic 1129 measurements and analysis. D.M. synthesized and provided the green fluorescent 1130 polyethyleneimine-chitosan nanoparticles (gPEI-Chi NPs) for all the experiments performed 1131 in the paper. E.S., D.P. and C.R. wrote the manuscript with input from all authors. 1132 1133 **Data availability** 1134 The data that support this study are all included in the manuscript and raw data are available 1135 upon request from the authors. 1136 1137 **References** 1138 1. Cosgrove, D. J. Structure and growth of plant cell walls. _Nat. Rev. Mol. Cell Biol._ **25**, 1139 340–358 (2024). 1140 2. Lew, T. T. S., Koman, V. B., Gordiichuk, P., Park, M. & Strano, M. S. The Emergence 1141 of Plant Nanobionics and Living Plants as Technology. _Adv. Mater. Technol._ **5**, (2020). 1142 3. Dufil, G., Bernacka-Wojcik, I., Armada-Moreira, A. & Stavrinidou, E. Plant 1143 Bioelectronics and Biohybrids: The Growing Contribution of Organic Electronic and 1144 Carbon-Based Materials. _Chem. Rev._ **122**, 4847–4883 (2021). 1145 4. Kwak, S. Y. _et al._ Chloroplast-selective gene delivery and expression in planta using 1146 chitosan-complexed single-walled carbon nanotube carriers. _Nat. Nanotechnol._ **14**, 1147 (2019). 43 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1148 5. Demirer, G. S. _et al._ High aspect ratio nanomaterials enable delivery of functional 1149 genetic material without DNA integration in mature plants. _Nat. Nanotechnol._ **14**, 1150 (2019). 1151 6. Wong, M. H. _et al._ Nitroaromatic detection and infrared communication from wild-type 1152 plants using plant nanobionics. _Nature Materials 2016 16:2_ **16**, 264–272 (2016). 1153 7. Lew, T. T. S. _et al._ Real-time detection of wound-induced H2O2 signalling waves in 1154 plants with optical nanosensors. _Nat. Plants_ **6**, (2020). 1155 8. Giraldo, J. P. _et al._ Plant nanobionics approach to augment photosynthesis and 1156 biochemical sensing. _Nature Materials 2014 13:4_ **13**, 400–408 (2014). 1157 9. Sompura, Y. _et al._ Nanoparticles and photosynthesis: a critical mechanistic review of 1158 molecular interactions and physiological impacts. _npj Science of Plants 2026 2:1_ **2**, 13 1159 (2026). 1160 10. Pérez-de-Luque, A. Interaction of nanomaterials with plants: What do we need for real 1161 applications in agriculture? _Front. Environ. Sci._ **5**, 259755 (2017). 1162 11. Parkinson, S. J. _et al._ Polymer nanoparticles pass the plant interface. _Nature_ 1163 _Communications 2022 13:1_ **13**, 7385- (2022). 1164 12. Raines, C. A. The Calvin cycle revisited. _Photosynth. Res._ **75**, 1–10 (2003). 1165 13. Shi, X. & Bloom, A. Photorespiration: The futile cycle? _Plants_ **10**, (2021). 1166 14. Badger, M. R. & Price, G. D. CO 2 concentrating mechanisms in cyanobacteria: 1167 Molecular components, their diversity and evolution. _J. Exp. Bot._ **54**, 609–622 (2003). 1168 15. Raven, J. A., Beardall, J. & Sánchez-Baracaldo, P. The possible evolution and future of 1169 CO 2 -concentrating mechanisms. _J. Exp. Bot._ **68**, (2017). 1170 16. Hanson, M. R., Lin, M. T., Carmo-Silva, A. E. & Parry, M. A. J. Towards engineering 1171 carboxysomes into C3 plants. _Plant J._ **87**, (2016). 1172 17. Von Caemmerer, S., Quick, W. P. & Furbank, R. T. The development of C4 rice: Current 1173 progress and future challenges. _Science (1979)._ **336**, (2012). 1174 18. Bar-Even, A., Noor, E., Lewis, N. E. & Milo, R. Design and analysis of synthetic carbon 1175 fixation pathways. _Proc. Natl. Acad. Sci. U. S. A._ **107**, (2010). 1176 19. Xu, X. _et al._ PVA-Coated Fluorescent Carbon Dot Nanocapsules as an Optical Amplifier 1177 for Enhanced Photosynthesis of Lettuce. _ACS Sustain. Chem. Eng._ **8**, (2020). 1178 20. Sai, L., Liu, S., Qian, X., Yu, Y. & Xu, X. Nontoxic fluorescent carbon nanodot serving 1179 as a light conversion material in plant for UV light utilization. _Colloids Surf. B_ 1180 _Biointerfaces_ **169**, (2018). 44 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1181 21. Chandra, S. _et al._ High throughput electron transfer from carbon dots to chloroplast: A 1182 rationale of enhanced photosynthesis. _Nanoscale_ **6**, (2014). 1183 22. Yasumoto, K. _et al._ Atmospheric CO 2 captured by biogenic polyamines is transferred as 1184 a possible substrate to Rubisco for the carboxylation reaction. _Sci. Rep._ **8**, (2018). 1185 23. Yasumoto, K. _et al._ Biogenic Polyamines Capture CO 2 and Accelerate Extracellular 1186 Bacterial CaCO3 Formation. _Marine Biotechnology_ **16**, (2014). 1187 24. Hoffman, J. R., Baumann, A. E. & Stafford, C. M. Thickness dependent CO 2 adsorption 1188 of poly(ethyleneimine) thin films for direct air capture. _Chemical Engineering Journal_ 1189 **481**, (2024). 1190 25. Singh, S., Alebrahim, T., Esmaeili, N., Jiao, Y. & Lin, H. Polyamines with reactive CO 2 1191 diffusion for carbon capture: the obvious and the unexpected. _J. Mater. Chem. A Mater._ 1192 **14**, 22903 (2026). 1193 26. Routier, C. _et al._ Chitosan-Modified Polyethyleneimine Nanoparticles for Enhancing 1194 the Carboxylation Reaction and Plants’ CO 2 Uptake. _ACS Nano_ **17**, (2023). 1195 27. Casper, J. _et al._ Core–Shell Structured Chitosan-Polyethylenimine Nanoparticles for 1196 Gene Delivery: Improved Stability, Cellular Uptake, and Transfection Efficiency. 1197 _Macromol. Biosci._ **23**, (2023). 1198 28. Zhao, X. _et al._ Alginate and chitosan surface coating reduces the phytotoxicity of CeO2 1199 nanoparticles to duckweed (Lemna minor L.). _Chemosphere_ **362**, (2024). 1200 29. Routier, C., Hermida-Carrera, C. & Stavrinidou, E. Investigating the Effect of Syringe 1201 Infiltration on Nicotiana tabacum (Tobacco). _ACS Agricultural Science and Technology_ 1202 **5**, (2025). 1203 30. Tang, T. C. _et al._ Materials design by synthetic biology. _Nat. Rev. Mater._ **6**, (2021). 1204 31. An, B. _et al._ Engineered Living Materials For Sustainability. _Chem. Rev._ **123**, (2023). 1205 32. Colaprico, E. _et al._ Plant cell-based 4D engineered living materials. Preprint at 1206 https://doi.org/10.64898/2026.10.19.760083 _bioRxiv_ (2026). 1207 33. Dubreuil, C. _et al._ Establishment of photosynthesis through chloroplast development is 1208 controlled by two distinct regulatory phases. _Plant Physiol._ **176**, (2018). 1209 34. Qi, J. _et al._ Rational design of ROS scavenging and fluorescent gold nanoparticles to 1210 deliver siRNA to improve plant resistance to Pseudomonas syringae. _J._ 1211 _Nanobiotechnology_ **22**, (2024). 1212 35. Zhang, J. _et al._ Expression of PEI-coated gold nanoparticles carrying exogenous gene in 1213 periwinkle mesophyll cells and its practice in huanglongbing research. _iScience_ **25**, 1214 (2022). 45 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1215 36. Goulas, E. _et al._ The chloroplast lumen and stromal proteomes of Arabidopsis thaliana 1216 show differential sensitivity to short- and long-term exposure to low temperature. _Plant_ 1217 _Journal_ **47**, (2006). 1218 37. Jiang, Z., van Zanten, M. & Sasidharan, R. Mechanisms of plant acclimation to multiple 1219 abiotic stresses. _Communications_ _Biology_ vol. 8 Preprint at 1220 https://doi.org/10.1038/s42003-025-08077-w (2025). 1221 38. Zhang, H., Zhu, J., Gong, Z. & Zhu, J. K. Abiotic stress responses in plants. _Nat. Rev._ 1222 _Genet._ **23**, 104–119 (2022). 1223 39. Bogdziewiez, L. _et al._ The Q-Warg Pipeline: A Robust and Versatile Workflow for 1224 Quantitative Analysis of Protoplast Culture Conditions. _Plant Direct_ **9**, (2025). 1225 40. Wu, H. C., Bulgakov, V. P. & Jinn, T. L. Pectin Methylesterases: Cell Wall Remodeling 1226 Proteins Are Required for Plant Response to Heat Stress. _Front. Plant Sci._ **9**, 1612 1227 (2018). 1228 41. Levesque-Tremblay, G., Pelloux, J., Braybrook, S. A. & Müller, K. Tuning of pectin 1229 methylesterification: consequences for cell wall biomechanics and development. _Planta_ 1230 _2015 242:4_ **242**, 791–811 (2015). 1231 42. Pesquet, E., Korolev, A. V., Calder, G. & Lloyd, C. W. The Microtubule-Associated 1232 Protein AtMAP70-5 Regulates Secondary Wall Patterning in Arabidopsis Wood Cells. 1233 _Current Biology_ **20**, (2010). 1234 43. Blom, T. J. M., Kreis, W., van Iren, F. & Libbenga, K. R. A non-invasive method for 1235 the routine-estimation of fresh weight of cells grown in batch suspension cultures. _Plant_ 1236 _Cell Rep._ **11**, (1992). 1237 44. Mustafa, N. R., De Winter, W., Van Iren, F. & Verpoorte, R. Initiation, growth and 1238 cryopreservation of plant cell suspension cultures. _Nat. Protoc._ **6**, (2011). 1239 45. Victor Rodriguez. Chlorophyll Extraction and Spectral Analysis with Spectrometer 1240 Calibration. dx.doi.org/10.17504/protocols.io.8eahtae (2019). 1241 46. Inskeep, W. P. & Bloom, P. R. Extinction Coefficients of Chlorophyll a and b in N,N 1242 Dimethylformamide and 80% Acetone . _Plant Physiol._ **77**, (1985). 1243 47. Ralph, P. J. & Gademann, R. Rapid light curves: A powerful tool to assess 1244 photosynthetic activity. _Aquat. Bot._ **82**, (2005). 1245 48. Platt, T., G. C. L. Modelling Primary Production. in _Primary Productivity in the Sea._ 1246 _Environmental Science Research_ (ed. Springer) vol. 19 339–362 ( Boston, MA, 1980). 1247 49. Choi, M. _et al._ MSstats: An R package for statistical analysis of quantitative mass 1248 spectrometry-based proteomic experiments. _Bioinformatics_ **30**, (2014). 46 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.30.763161; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.10.30.763161) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1249 50. Wu, T. _et al._ clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. 1250 _Innovation_ **2**, (2021). 1251 51. Yoo, S. D., Cho, Y. H. & Sheen, J. Arabidopsis mesophyll protoplasts: A versatile cell 1252 system for transient gene expression analysis. _Nat. Protoc._ **2**, (2007). 1253 52. Stringer, C. & Pachitariu, M. Cellpose3: one-click image restoration for improved 1254 cellular segmentation. _Nat. Methods_ **22**, (2025). 1255 53. Sweeley, C. C., Bentley, R., Makita, M. & Wells, W. W. Gas-Liquid Chromatography 1256 of Trimethylsilyl Derivatives of Sugars and Related Substances. _J. Am. Chem. Soc._ **85**, 1257 (1963). 1258 54. Klavons, J. A. & Bennett, R. D. Determination of Methanol Using Alcohol Oxidase and 1259 Its Application to Methyl Ester Content of Pectins. _J. Agric. Food Chem._ **34**, (1986). 1260 55. Updegraff, D. M. Semimicro determination of cellulose inbiological materials. _Anal._ 1261 _Biochem._ **32**, (1969). 1262 56. Saeman, J. F. Kinetics of Wood Saccharification - Hydrolysis of Cellulose and 1263 Decomposition of Sugars in Dilute Acid at High Temperature. _Ind. Eng. Chem._ **37**, 1264 (1945). 1265 57. Scott, T. A. & Melvin, E. H. Determination of Dextran with Anthrone. _Anal. Chem._ **25**, 1266 (1953). 1267 47