Seed plastids: A novel platform for recombinant protein expression

Platforms for expression of recombinant proteins in plants are protein bodies in the seed (Kanai et al., 2023), and plant viral vectors and chloroplasts in leaves (Abrahamian et al., 2020; Bock, 2015; Maliga, 2021). Expression of recombinant proteins in chloroplasts has the advantages of high-level protein accumulation, expression of multiple genes in operons and natural biocontainment due to the lack of the transmission of transgenes by pollen. Recombinant proteins produced in leaves are perishable and need to be extracted for use or lyophilized for storage (Twyman et al., 2003). Seeds are natural storage organs for the accumulation of proteins. Proteins in the seed are well protected and may be stored for a long time. So far, no attempt has been reported on expressing recombinant proteins in seed plastids. We report that recombinant proteins can be expressed in seed plastids, where they remain stable for at least 1 year. To test seed-specific expression in plastids, we used a system based on an engineered PPR10 RNA-binding protein. PPR10 is a member of the pentatricopeptide repeat (PPR) family, whose members are defined by degenerate 35-amino acid helical repeats. In maize plastids, PPR10 binds in the atpI–atpH intergenic region and defines and stabilizes the 5′ and 3′ termini of the mRNAs. The minimal PPR10 binding site in the atpH 5′-UTR spans 17 nucleotides. When PPR10 binds its binding site (BS) upstream of the AUG translation initiation codon, it prevents the formation of an inhibitory stem structure that would sequester the GGAGG ribosome binding site, thereby facilitating mRNA translation. PPR10 target recognition specificity was changed by engineering the helical repeats so that the engineered PPR10 (PPR10GG) does not recognize the wild-type binding site (BSWT), only the GG site (BSGG) (Barkan et al., 2012). We employed the two-component PPR10GG/BSGG regulatory system for protein expression in seed plastids (Rojas et al., 2019; Yu et al., 2019) (Figure 1a). The first component is a plastid reporter gene with an engineered PPR10GG RNA-binding protein binding site (BSGG) upstream of the AUG translation initiation codon. We expressed three reporters: GFP, mScarlet and GUS. Transplastomic tobacco plants were obtained by the transformation of tobacco leaves with vectors shown in Figure 1c. The uniform transformation of plastid genomes was confirmed by DNA gel blot analyses (Figure S1). The second component is a nuclear transgene expressing the engineered PPR10GG RNA-binding protein from the napin A seed-specific promoter that selectively recognizes the engineered BSGG binding site (Figure 1b) (Barkan et al., 2012). To test seed-specific expression, transplastomic GFP plants carrying the BSGG:gfp construct (pAI5, Figure 1c) were transformed with the PnapA:PPR10GG gene (pAQ5, Figure 1b). Plants representing 17 independently transformed lines were transferred to the greenhouse, and GFP accumulation was tested in dry seed. GFP could be readily detected in two out of the 17 lines (#4 and #5). PPR10GG was also detectable in the same two lines, indicating that GFP accumulation is dependent on the PPR10GG protein (Figure 1d). In seed, the PnapA promoter transcribes PnapA:PPR10GG mRNA, which is exported to the cytoplasm, where it is translated on cytoplasmic ribosomes. The PPR10GG protein is then imported into seed plastids, where it facilitates the translation of GFP on plastid ribosomes. The level of GFP accumulation in the seed was 780 ± 50 μg/kg dry seed (Figure 1g). The active PnapA:PPR10GG alleles from lines #4 and #5 were introduced into the transplastomic mScarlet (pAQ2, Figure 1c) and GUS (pMP2, Figure 1c) maternal parents by crossing. The seed of transplastomic plants carrying the active PnapA:PPR10GG allele was then tested for accumulation of mScarlet (Figure 1e) and GUS (Figure 1f). Data are shown for fresh seed, collected from capsules that turned brown. The full-size mScarlet protein in the fresh seed accumulated at 770 ± 140 μg/kg seed (Figure 1g). However, in the seed, a significant amount of stable degradation product also accumulated. As to GUS, in the fresh seed, most GUS was already degraded, and only a small amount of full-size GUS was present (26 ± 4 μg/kg seed, Figure 1g). Recombinant seed protein levels obtained by transformation of the tobacco nucleus are significantly higher, for example 3 mg/g in case of the H5N1 hemagglutinin antigen (Ceballo et al., 2017). We also tested protein levels after storing the seed at room temperature for 1 year (25 °C). GFP in the seed was essentially unchanged after 1 year (Figure 1d). As to the stability of mScarlet, after storage for 1 year, only the ~26-kDa stable degradation product was present in the seed (Figure 1e). After 1 year, GUS was undetectable in the seed protein extracts (Figure 1f). After 1 year, PPR10GG was also present in the seed (Figure 1d–f). When the proteins are stable, seed-based bioreactors enable long-term storage of the proteins of interest at ambient temperature, although it may be necessary to remove protease-sensitive sites. Glycosylation is absent in plastids. Therefore, good candidates for seed-specific expression in plastids would be bacterial vaccine antigens that do not require glycosylation, such as the anthrax protective antigen, the porcine post-weaning diarrhoea vaccine and the tuberculosis antigen (Ahmad et al., 2016). More importantly, the non-photosynthetic plastids in the seed are prospective hosts for new metabolic pathways (Mellor et al., 2018). This research was supported by the National Science Foundation Grants NSF MCB 1716102, NSF IOS 2037155 and NSF IOS 2224861. We thank undergraduate Alifiya Quresh for constructing the pAQ2 and pAQ5 vectors and characterizing the Nt-pAQ2 plants. Undergraduates A.L., A.Q. and M.P. were recipients of a Waksman Undergraduate Summer Research Fellowship and were supported by the NSF Grants and the Rutgers Aresty Science Program. A.C. was supported by NSF grant MCB 1716102. The authors declare no competing interests. P.M. designed the experiments. K.A.L. and A.C. tagged the PPR10GG protein, constructed pKAL16 vector and obtained transgenic Nt-pKAL16 plants. M.P. constructed the pMP2 vector and obtained the transgenic plants under the supervision of A.M. M.M. and A.L. tested protein accumulation. P.M., M.M. and K.L. wrote the paper. The data that supports the findings of this study are available in the supplementary material of this article. Appendix S1 Materials and methods. Figure S1 DNA gel blot analyses confirms uniform transformation of plastid genomes in mScarlet (a) and GUS (b) independently transformed lines. Figure S2 Coomassie-stained SDS-PAGE of leaf protein extracts, the source of reference proteins. Table S1 DNA sequence of vectors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.