Towards chloroplastic nanofactories: formation of proteinaceous scaffolds for metabolic engineering

Abstract

The evolution of eukaryotic lipid-bound organelles allows for specialized metabolism to occur within spatially distinct metabolic landscapes within the same cell. However, this strategy of compartmentalization is not unique to eukaryotic organisms. Many bacteria, spread across 45 different phyla, contain loci that encode for specialized bacterial microcompartments (BMCs) (Sutter et al., 2021). The formation of a BMC involves self-assembly from three families of shell proteins to form the outer shell membrane, in addition to the encapsulation of an enzymatic core packaged within the lumen of a BMC (Kerfeld et al., 2018). These proteinaceous organelles provide a competitive growth advantage by enabling organisms to process inaccessible substrates by sequestering metabolic intermediates (i.e. aldehydes) that would otherwise be toxic within the cytoplasm. While BMCs perform specific metabolic functions in their native organism, synthetic, empty BMC shells can form without the requirement of native cargo inside (Doron and Kerfeld, 2024). This provides a transferable and tunable platform of protein scaffolding for guiding metabolic engineering within a host organism of choice (Raba and Kerfeld, 2022).

To date, efforts have been made in transiently producing cyanobacterial beta-carboxysomal BMC proteins in the model plant species, Nicotiana benthamiana, in aiming to reconstitute a functional carboxysome for enhanced carbon fixation (Lin et al., 2014; Nguyen et al., 2024). Targeting various shell protein combinations to the chloroplasts resulted in carboxysome-like structures, but the system did not have a functional carbon-fixing machinery. Furthermore, the gene components required for a functional beta-carboxysome present challenges for heterologous expression in plant systems (Nguyen et al., 2024). To that note, a minimal alpha-carboxysome shell system encapsulating Rubisco was established using genes from Cyanobium in a tobacco host, allowing autotrophic growth at high CO₂ (Long et al., 2018). Similar observations were concluded from a minimal alpha-carboxysome system heterologously produced from Halothiobacillus neapolitanus in tobacco, but the resulting carboxysomes displayed slightly abnormal morphologies (Chen et al., 2023; Nguyen et al., 2024). Concurrent with efforts in expressing genes encoding microcompartments for enhancing crop yields are methods using direct nanoparticle delivery of nucleic acids and various cargo. In recent years, there has been great progress in delivering nanoparticles with proof-of-concept reporters and cargo, resulting in physiological changes in plant growth. However, direct uptake of nanoparticles has challenges: the physical application, aggregation of nanoparticles, plant physiological incompatibilities, lack of host genome integration efficiency and the long-term cytotoxicity of the nanoparticles (Raba and Kerfeld, 2022; Yadav et al., 2023). To overcome these hurdles, we used a nuclear transformation approach for stable genomic integration of genes encoding a minimal, two-shell protein component scaffold, namely a synthetic BMC shell, with the long-term goal of assembling nanofactories within the chloroplast for metabolic engineering.

To establish a synthetic BMC shell-based scaffold within Arabidopsis thaliana chloroplasts, we transformed plants with a plasmid encoding chloroplast-targeted hexamer-forming (BMC-H, pfam00936, 10.1 kDa) and trimer-forming (BMC-T1, tandem pfam00936 domains, 21.9 kDa) shell proteins from the bacterium Haliangium ochraceum (HO). 60 BMC-H and 20 BMC-T1 proteins self-assemble into icosahedral, >5 MDa 40 nm diameter shells with 12 six nanometre gaps at the vertices (lacking BMC-Pentamer, pfam03319), termed ‘minimal wiffle balls’ (Doron and Kerfeld, 2024; Kerfeld et al., 2018; Sutter et al., 2017). To accomplish this, we modified the genes encoding HO BMC-H and BMC-T1 by adding the coding region of the transit peptide derived from Rubisco small subunit 1B (At5g38430) to the 5′ region of both HO BMC-T1 and BMC-H genes. Additionally, we added C-terminal His- and HA-tag epitopes, thus giving rise to cDNAs that encoded the modified proteins: tpSSU-BMC-H_HA and tpSSU-BMC-T1_His under the control of a 35S promoter. Finally, we linked these two modified shell proteins by inserting the hybrid linker, LP4/2A, which upon co- and post-translational processing, yields free monomeric shell proteins allowing assembly of the shell (Figures S1A,B and S2A).

The final construct, tpSSU-BMC-T1_His-(LP4/2A)-tpSSU-BMC-H_HA (Figures S1C and S2A) was used to transform Arabidopsis plants. We characterized four independently transformed lines by either Coomassie staining or western blotting of SDS-PAGE gels using antibodies against either His-tag or HA-tag epitopes (Figure S2B). Western blot analyses confirmed the expression of both BMC-T1_His and BMC-H_HA within our transformed Arabidopsis lines (Figure S2B). Coincidently, the molecular weights of both BMC-T1_His and BMC-H_HA identified by western blotting (Figure S2B) indicated that the transit peptide attached to these two-shell proteins had been removed, suggesting that these shell components were indeed imported into the stroma of chloroplast. However, SDS-PAGE and western blots also confirmed the presence of both unprocessed (~50 kDa) and processed chloroplast-targeted fused shell proteins (Figure S2B), requiring optimization of the construct for more efficient peptide processing. Higher peptide processing efficiency, that is more shell protein released to self-assemble in hexamers or trimers, will yield more fully intact purified shells to be extracted from the stroma. Regardless, production of both BMC-T1_His and BMC-H_HA proteins in transformed Arabidopsis lines did not alter the overall growth of these plants when compared to wild-type plants (Figure S2B), either due to low expression of the gene products and/or the compatibility of these microcompartments with plant growth. Thus, this approach is a suitable system for metabolic engineering with no negative impact on plant health, as demonstrated by the introduction of these two-shell proteins.

To confirm the presence of assembled synthetic shells within the chloroplasts of these transformed Arabidopsis lines encoding both BMC-H_HA and BMC-T1_His proteins (Figure S2B), we isolated the minimal wiffle ball shells from ground up whole rosettes (Figure 1A) and subsequently chloroplast-enriched (Figure S3) leaf fractions using His-tag affinity purification. Clarified whole leaf lysate was applied to a nickel column, and elution fractions (Figure 1B) were pooled and concentrated with a 100 and subsequent 1000 kDa size exclusion filter to eliminate protein structures smaller than assembled wiffle ball shells. SDS-PAGE and western blot analyses of the purification pipeline demonstrated the presence of BMC-T1_HIS and the co-eluted BMC-H_HA in the final size-filtered eluate (Figure 1C). While the antibody signal was weak for the presence of monomeric shell proteins, the 1000 kDa size-filtered elution was spotted on a grid for TEM analysis, revealing shell structures approximately 40 nm in size indicating enough free BMC-H and BMC-T proteins were present to self-assemble in the expected stoichiometry (Figure 1D). Ground whole leaf rosettes from wild-type Arabidopsis were subjected to the same HisTrap purification pipeline, and the presence of shells was not detected as determined by TEM (Figure S4). For chloroplast-enriched leaf samples, a higher abundance of monomeric BMC-H_HA and BMC-T1_His in the stroma fractions was observed by western blot (Figure S3A). Subsequently, BMC-H_HA with BMC-T1_His co-eluted from a His-affinity nickel purification (Figure S3B), demonstrating the presence of shell complexes. Furthermore, the elution sample was concentrated with a 1000 kDa size exclusion filter then spotted on a grid for TEM analysis. Similar to the shells extracted from whole rosette tissue (Figure 1), 40-nm shell particles were present in this fraction (Figure S3C), supporting in vivo shell assembly within the chloroplast. The 40 nm diameter shells purified from Arabidopsis in this study phenocopy the shells purified from the well-established shell production system in E. coli (Doron and Kerfeld, 2024; Sutter et al., 2017), demonstrating no plant host interference of shell protein self-assembly.

In summary, we constructed wiffle ball shells in Arabidopsis chloroplasts and were able to purify an intact bacterial microcompartment, demonstrating the potential of a new sub-compartmental synthetic scaffold for plant metabolic engineering inside the chloroplast. In E. coli, this tunable synthetic shell system has been modulated for shell protein composition or cargo attachments, that is enzymes, small molecules or fluorescent reporters (Doron and Kerfeld, 2024). With this newly established platform, the same technology will be used to target and enhance photosynthetic and metabolic processes to harvest the biochemical potential of chloroplasts.

All authors declare no conflict of interest.