Seed-Specic Expression of Apolipoprotein A-I Milano Dimer in Rice (Oryza Sativa L.) Transgenic Lines

Apolipoprotein A-I Milano (ApoA-I M ) has been shown to signicantly reduce coronary atherosclerotic plaques. However, the preparation of cost-effective pharmaceutical formulations of Apo A-I Milano can be limited by the high cost and diculty of purifying the protein and producing the highly effective dimeric form. The aim of this study was the characterization of rice plants expressing the ApoA-I Milano protein primarily in a dimeric form, specically in the seeds. The seed-specic 13 kDa rice-prolamin promoter drives the expression of ApoA-I Milano to different caryopsis tissues such as seed coat, aleurone cells, and endosperm, particularly into amyloplast and storage vacuoles. A plant-based ApoA-I Milano production system offers numerous advantages over current production systems, including direct production of the most therapeutically effective dimeric ApoA-I Milano forms, long-term storage of protein in seeds, and ease of protein production by simply growing plants. Therefore, seeds could provide a cost-effective source of ApoA-I Milano for use as a therapeutic. formulations ApoA-I demonstrated rapid on plaque regression stabilization


Introduction
Apolipoprotein A-I (ApoA-I) is a component of high-density lipoprotein (HDL). It constitutes approximately 70% of HDL. HDL is a molecule that transports cholesterol and certain fats called phospholipids through the bloodstream from the body's tissues to the liver. Once in the liver, cholesterol and phospholipids are redistributed to the other tissues or removed from the body 1 . HDL is often called "good cholesterol" because high levels of this substance reduce the chance of developing heart and cardiovascular disease 2 . ApoA-I is the subject of intense research for its anti-atherogenic properties. Infusion of ApoA-I mimetic peptides has been shown to stimulate cholesterol e ux from tissues into plasma 3  Although the therapeutic effects of apolipoproteins have been proven, there are numerous di culties in their preparation as pharmaceuticals, mainly related to the high cost of production.
The production of recombinant proteins in bacterial systems is attractive for their capacity to produce high quantities but also shows di culties such as the presence of undesirable a nity signals at the end of the puri cation 8 . Further, the endotoxins of E. coli are known to form strong complexes with apolipoproteins 9 . The elimination of these toxins for pharmaceutical products is necessary and technically possible but requires complex and costly methods 10 . The production of the more therapeutically effective dimer form of Apo A-I M can be accomplished by puri cation of monomers from E. coli, as described in patent US6617134B1 with a subsequent dimerization process, however, this is a timely and expensive procedure that is not without risk.
The therapeutic use of apolipoproteins is therefore limited by the scarcity of methods to prepare the protein in a su cient quantity, and a form optimized for administration. In particular, the production of Apo A-I with recombinant methods has been shown to be very di cult due to its amphiphilic character, autoaggregation, and degradation 11 . Thus, there is a need for a system of production or expression that easily, safely, and cost-e ciently produces apolipoproteins in dimeric form.
The ability to introduce foreign genes into plant species by techniques such as agroinfection or by direct gene transfer has opened up the possibility of using transgenic plants as host organisms for the production of heterologous proteins 12 . Plants have several advantages that make them more attractive for this purpose: i) they represent an economic system, compared to mammalian cell cultures and microbial fermentation, ii) the expression of heterologous proteins can be induced in speci c organs as seeds, tubers, etc, where proteins are more stable 13,14 , iii) therapeutic proteins derived from plants, whether puri ed or not, are less likely to be contaminated with human pathogenic microorganisms than those derived from animal cells because plants are not hosts for human or animal infectious agents 15,16 . iv) Plants can perform most of the post-translational modi cations (glycosylation, prenylation) required for protein stability, and bioactivity in the same way as other higher eukaryotes17; v) proteins can be expressed in the edible organs of plants in order to be consumed raw as an edible vaccine eliminating the need for downstream processes 17,18 . Downstream processing when required is easy and less expensive, particularly when proteins are expressed in speci c organs like seeds.
A previous study showed that rice plants transformed for Apo A-I M expression provide rice milk with valuable therapeutic effects in protecting mice from atherosclerotic plaques, suggesting the use of these molecules in cardiovascular diseases 19 .
Apo A-I M transformed rice plants have been further characterized in the present work and the localization of the protein was studied in two different stages of seed ripening in transgenic plants.

Molecular analysis of putative transformed plants
The rice expression vector pCambia-PROL-ApoA-I M (Figure 1a) containing ApoA-I M gene under the control of the 13 KDa seed promoter was developed. In order to con rm the correct insertion of the ApoA-I M gene and promoter in the expression vector pCAMBIA-PROL-ApoA-I M , the plasmid was digested with different restriction enzymes and subjected to electrophoresis on an agarose gel. The digested fragments corresponding to the expected size according to the map of recombinant plasmid (Figure 1b). The fusion from promoter and gene and the correct frame of translation was con rmed by sequencing. A total of approximately 32 independent hygromycin-resistant T0 transformed plants were obtained and grown under greenhouse conditions and evaluated for morphology and fertility. Considering the seeds production, about 93% of the transformed plants were fertile and presented the same morphological aspects of the controls (untransformed plants).
We analysed all of the independently derived transgenic plants by the PCR for the presence of the ApoA-I M gene using Apo-fw and Apo-rv speci c primers (Table 1)  Temporal expression of the prolamin promoter was rst proved by differential display experiments considering different days after owering (DAF) stages in rice seeds (Oryza sativa cv. Ariete) (Availability of materials and data).
Total proteins were isolated from ApoA-I M transgenic seeds at different ripening stages (4, 8, 12, 16, 20, and 25 DAF), loaded on 15% acrylamide gel in denaturing conditions. After blotting on nitrocellulose membrane, the hybridization with antiserum ApoA-I was performed. As shown in g. 3a, ApoA-I M protein was detected early, 8 DAF, and rapidly reached the plateau between 12 and 16 DAF, remaining at the maximal level during the following ripening stages.
At the same time, western blotting was performed on different tissues of rice to the veri er the seed expression. Protein extracted from seeds, leaf, culm, and root from ApoA-I M transgenic plants was blotted and hybridized at the same condition. As shown in g. 3b, ApoA-I M signal was detected only in the seeds; leaf, culm, and roots showed absence of positive hybridization maybe after a prolonged exposition.

Immunolocalization of ApoAI M in seed tissues
Localization of ApoAI M was further analysed by immunocytochemical and immunogold observations. Caryopsis at two different developmental stages (called I and II respectively) were collected in the mid maturation phase (around 25 DAF) from the same panicle and observed by uorescence and electron microscope.
In caryopsis I ( Figure 4) the pericarp, seed coat, and nucellus were clearly observed ( Figure 4b) and only a weak auto uorescence was observed in these tissues after anti-ApoAI M antibody treatment (Figure 4a mammalian cells) has been achieved, even if with not high level of expression. Many different expression systems are available for the production of pharmaceutical proteins and the selection of the appropriate system depends on a variety of factors including the host's ability to express the desired product (e.g. processing and post-translational modi cation) and the complexity and scalability of the product. Bacterial expression systems are commonly used for therapeutical heterologous protein production as these systems provide an economical route for protein expression and require minimal technical expertise to set up a lab production system. Nevertheless, prokaryotic cells are not capable of the post-translational modi cations for biological activity and in the Gram-negative bacterial systems, such as E. coli, endotoxin contamination can pose a signi cant problem in puri cation. On the contrary, eukaryotic cells, tissues, or organisms, available as expression hosts, can produce complex proteins with the correct folding and post-translational changes but they are expensive to operate, have a longer growth period, and present contamination risks. Plants have been utilized for more than two decades for the production of many pharmaceuticals proteins and many encouraging results were obtained also in the eld of edible vaccine. Many plant-based vaccines have been developed in different plant expression systems and evaluated against various life-threatening diseases.
In our study, the production of ApoA-I M in rice offers an attractive alternative for e cient and inexpensive recombinant protein production while minimizing the risk for contamination by human pathogens.
In our study, the 13 KDa prolamin promotor allowed the seed-speci c expression of the recombinant ApoA-I M protein in transformed rice plants. Seeds have proven to be effective tools for recombinant protein production, including peptides or short and long polypeptides as well as complex like mammalian immunoglobulins 21 . The extraction and recovery of recombinant proteins from seeds is greatly assisted by their dormancy properties because this allows for the long-term stability of stored heterologous proteins and decoupling of processing from the growth and harvest cycles. Furthermore, the low water content reduces the possibility of microorganism's secondary contamination (e.g. bacteria and fungi), possessing possess more storage opportunities 22 .
Moreover, ApoA-I M was expressed in rice seeds in dimeric form and its activity was e ciently determined by Romano et al. (2018). In this study, the rice plants expressing APOA-I M were further investigated, particularly seed expression and its localization at the cellular level. in association with NOS terminator is su cient for seed speci city; on the contrary, it has been reported that the 10kDa prolamin promoter is not completely effective for the seed-speci c expression due to the association in the expression cassettes with NOS terminator 25 .
The expression of Apo A-I M protein follows a time course pattern during seed ripening; western blotting experiments revealed the ApoA-I M protein accumulated early at 8 DAF and rapidly reached the maximum expression at 18 DAF (mid-maturation). The level of expression remained stable just at complete maturation of the seeds, according to Saito et al. (2012) 26 .
The subcellular localization of a recombinant protein strongly in uences its stability, accumulation, and ultimately the yield on recovery. Protein targeting also plays a key role in many types of post-translational modi cations, among which glycosylation is the most relevant for molecular farming because the glycan chains can affect protein structure, biological function, and immunogenicity. Targeting to a speci c compartment may therefore in uence the interrelated processes of folding, assembly, and ultimate accumulation site (deposition), which have a direct impact on protein stability and yield. Seeds, therefore, offer several alternative subcellular destinations for recombinant proteins including Protein Bodies (PBs) derived from the reticulum endoplasmatic (ER), protein storage vacuoles (PSVs), starch granules, and the surface of oil bodies.
Immunocytochemical experiments showed a different localization of apolipoproteins in caryopsis at two different stages of ripening (mid maturation phase), from the same panicle. In the early stage, the Apo AI is expressed in the amyloplasts of endosperm cells. Later in addition to endosperm, the recombinant protein also localized in pericarp and aleurone cells. Electron microscopy observation on caryopsis in a later stage showed the distribution of ApoA-I in different cell organelles; as reported for 13 kDa prolamins 26 , the recombinant proteins were observed in PBs. The presence of Apo AI occurred in the middle and surrounding layer of PBs. In fact, the 10 KDa prolamine was mainly detected in the core of PBs, whereas the 13kDa and 16 KDa proteins were detected in rings surrounding the 10 kDa prolamin-rich core.
The distribution pattern of ApoA-I M in mid and external layers of PBs, re ects the expression timing of 13 KDa prolamin promoter. Indeed, the 10 KDa prolamin accumulated in the seed very early, from 4 DAF, on the contrary, the accumulation of 13 KDa prolamin begins later 26 .
Interestingly, the ApoA-I M was also observed in plastids in both stages of seed ripening. Moreover, in the late phase, ApoA-I M is present in the cell wall and as a soluble protein in the cytoplasm.
These apparent mislocalizations of ApoA-I M with respect to endogenous seed reserve proteins suggests that protein folding in rough endoplasmic reticulum (RER) or post-translational modi cations in Golgi apparatus could be responsible for changes in Post-Golgi tra cking, leading to secretion of recombinant proteins into cell wall instead of to storage vacuoles (REFs). Intriguingly was observed in the cytoplasm and within plasmodesmata, suggesting that mRNAs for ApoA-I M could be in part translated by cytoplasmic ribosomes instead ribosomes localized on RER and that the recombinant protein could be transferred among cells by symplastic route.
ApoA-I M accumulated in the rice seeds in variable quantity, maximum recovery visualized was 8,2 mg/100g of rice seeds. ApoA-I production in E.coli recovered in literature was from 0,1 mg/L to 5 mg/L 27,28 , but in some cases, the ApoA-I protein accumulated in inclusion bodies and required a refolding after puri cation. Nykiforuk et al. (2011) 29 expressed ApoA-I M in a fusion protein in transgenic sa ower seeds in very high quantity, but the ApoA-I M rice offers some advantages as an inexpensive and a large production of edible materials ready to use.
Obtained transformed plants offer a promising option for large-scale APOA-I M production. Further investigations are required in order to better establish its effect after oral administration by edible transformed rice.

Expression vector
The 13 KDa seed-speci c promoter together with relative 5' UTR and transit peptide sequence (NCBI acc. N° D63901) was ampli ed from Oryza sativa var. Ariete using a forward and reverse primer (PROL-fw and PROL-rv, table 1). The 729 bp promoter fragment was inserted in pGEM-T (Promega, Madison WI USA), and the accuracy of the ampli ed promoter region was veri ed by DNA sequencing 30 . A PstI and XbaI site was inserted respectively in the forward and reverse primer to ease the subsequent DNA cloning in a plant expression vector (pCAMBIA1302, NCBI Acc. N° AF234298) in PstI and XbaI. The cDNA of mature ApoA-I M obtained according to Romano et al.(2018) 19 was put under the control of the rice 13 KDa prolamin promoter and cloned respectively in BamHI and SacI (blunted) (Figure 1a).

Rice transformation and analysis of transgenic plants.
Embryonic calli derived from mature zygotic embryos of the rice (Oryza sativa var. Rosa Marchetti) were inoculated with Agrobacteriumtumefaciens strain EHA105 transformed by electroporation with the plasmid pCAMBIA-PROL-ApoA-I M . Callus and bacterial induction, transformation, selection, and regeneration of transgenic tissues were performed as described in Hiei et al. (1994) 31 , with minor modi cations. Putatively transformed (hygromycin-resistant) plants (T0 plants) were potted in peat and hardened in a greenhouse together with controls (WT rice) to produce T1 seeds. These T1 seeds were then subsequently planted to produce T1 plants and T2 seeds. During the entire experimental period, the plants were evaluated for the principal agronomic characteristic. PCR Analysis. Before owering, T0 plants were subjected to PCR analysis to verify the presence of the transgene. The PCR was done using DNA extracted 32 from leaves of the rice lines resistant to hygromycin, using the primers Apo-fw and Apo-rv (Table 1), which amplify the entire Apo A-I gene, visualized, in the case of successful transformation, by a fragment of 732 bp. Southern Analysis. Independent transgenic rice plants expressing the ApoA-I M gene were subjected to Southern Blot analysis to verify the copy number and insertions of the transgene present in each plant.
The total genomic DNA was extracted from young leaves according to Doyle and Doyle (1987) 33 . Approximately 10 µg of DNA was digested with XbaI which cuts only once in pCAMBIA-PROL-ApoA-I M , fractionated on a 0,8% (W/v) agarose gel, transferred to a positively charged nylon membrane, and hybridized with a probe corresponding to ApoA-I M CDS labelled with the PCR DIG Probe Synthesis Kit (Roche Molecular Biochemicals Penzberg Germany) following the manufacture's instruction. The primer set used in PCR analysis was also used to generate the DIG-labelled probe. The positive control consisted of the vector (pGEMT-APOA-I M ) linearized with the same restriction enzyme. The detection was realized with CDP-Star® (GE Healthcare, Chicago Illinois USA) following the manufacturers' instructions.

Evaluation of protein expression
The seeds collected from the T0 and T1 plants PCR positive were then used to verify the expression by Western Analysis. Brie y, the total protein from rice seeds (100 mg) was extracted from the seeds using extraction buffer (50 mM Tris-HCl pH 8.0, 5 mM EDTA, 200 mM NaCl, 0.1% Triton X-100, 1mM PMSF) in ratio 1:10 (w/v), mixed with SDS-loading buffer (with or without β-mercaptoethanol, to discriminate from monomer/dimer respectively), subjected to electrophoretic separation on an acrylamide gel, and transferred using electroblotting (solution of 25 mM Tris, 192 mM glycine, 20% methanol, 30 V at 4°C overnight) to a nitrocellulose membrane, hybond ECL (GE Healthcare, Chigaco Illinois USA). The membrane with the bound protein was placed in a solution of PBS-T and 5% skim milk, agitated for 60 minutes, washed, then exposed to the primary antibody Anti ApoA-I Goat 1:5000 (Acris, San Diego CA, USA) and subsequently to the secondary antibody Anti-Goat peroxidase conjugate, 1:12.000 (Sigma Aldrich, St. Louis Missouri USA). The membrane was washed several times and placed in a chemiluminescent detection solution, ECL (GE Healthcare, Chicago Illinois USA).
Protein quanti cation in rice seeds. An indirect competitive ELISA (IC-ELISA) was used to detect ApoA-I M protein in rice seeds. In particular, recombinant hApoA-I (Sigma Aldrich St. Louis Missouri USA) was coated (600ng/ml) onto a micro-well plate overnight at 4°C, after which the plate was washed three times with 0,01 M PBS (pH 7) and then blocked with 200 μl of 5%(W/V) BSA for 2 h at 37°C. The plate was then washed with 0,01 M PBS containing 0,05% (v/v) Tween 20 (PBST). 100 µl of goat polyclonal anti-ApoA-I antibody (Acris, San Diego CA USA) diluted 1:6000 was mixed with serial dilution (1:1, 1:10, and 1:50) of rice seeds protein extraction (100 μl). From each mixture, 100 μl solution was added to each well coated, and the plate was incubated at 37°C for 1 h. The plate was washed with PBST and then a 1:5000 diluted solution (100 μl) of anti-goat IgG-HRP antibody (Thermo sher Scienti c, Waltham Massachusetts USA) was added to each well and incubated at 37°C for 1 h. After that, the plate was rewashed with PBST, and a 50 μl of TMB was added to each well followed by incubation at 37°C for 15 min. To stop the reaction, 150 μl of 0,4 N hydrochloric acid (HCl) was added to each well, and then absorbance was measured at 450 nm using an ELISA plate reader (BIORAD, Hercules California USA). Each experiment was performed in triplicates. The standard curve using recombinant hAPOA-I (Sigma Aldrich, St. Louis Missouri USA) protein was constructed. (Supplementary Figure S3) Immuno uorescence and Immunogold analyses Rice seeds (mid-maturation) were collected, reduced in small pieces, and xed in 50mM Hepes, pH 7,4, 2% formaldehyde, and 0,2% glutaraldehyde, overnight at 4°C. Samples were repeatedly rinsed in 50mM Hepes, pH 7.4, dehydrated with increasing concentrations of ethanol, and embedded in LR Gold resin (Sigma Aldrich St. Louis Missouri USA) at -20°C. Semi-ne sections (2µm) and ultra-thin sections (80 nm) were obtained using a Reichert Jung Ultracut E microtome. Ultra-thin sections were collected with nickel grids (Agar Scienti c, Stansted United Kingdom) for immunogold experiments.

Immuno uorescence analyses
Semi-ne sections were hydrated with TBS (Tris-buffered saline, 20 mM TRIS, 150 mM NaCl pH 7.6) for 5 minutes, blocked with 1% BSA in TBS for 1 hour at room temperature, and then incubated with primary goat polyclonal anti-ApoA-I antibody (Acris, San Diego California USA) diluted 1:500 overnight at 4 °C. After three rinses in TBS, sections were incubated in the FITC-conjugated antigoat secondary antibody (Santa Cruz Biotechnology, Dallas Texas USA) at 1:200 in TBS for 2 hours at room temperature, rinsed three times with TBS, and mounted in City uor. Negative controls were performed with only the secondary antibodies. Fluorescence observations were performed with a Leica DMRB microscope (set lter BP450-490, RKP 410, long pass 515). Images were acquired by a Leica MC170 HD camera.

Immunogold analyses
Ultra-thin sections were hydrated with TBS for few minutes, blocked in BSA 1% in TBS for 30 min at room temperature, and incubated with primary goat polyclonal anti-ApoA-I antibody (Acris, San Diego California USA) at different dilution (from 1:500 to 1:200) in TBS at room temperature for 2 h. After three rinses in TBS, sections were incubated with 20 nm gold-conjugated rabbit anti-goat secondary antibody (purchased by BBA International, USA) diluted 1:100 for 1h at room temperature, rinsed three times with TBS, and post-xed with 1% glutaraldehyde 15 min at room temperature. Sections were that stained with Uranyl acetate 3% for 20 min in the dark, rinsed with distilled water, dried, and observed with an EFTEM LEO 912AB transmission electron microscope (Zeiss, Oberkochen Germany) working at 100 kV.      a-i Immuno uorescence (a,c,e,g,i) and bright eld images (b,d,f,h) of caryopsis II of ApoAI M transformed plant. Pericarp, seed coat, nucellus, aleurons cells, and endosperm cells were observed (Pr, Sc, N, a, and se respectively). Pericarp and nucellus appeared thinner (a,b and cf respectively). ApoA-IM was localized in organelles in endosperm cells, seed coat, and aleurone cells (c,d; g-h). High uorescence was observed in the seed coat cell cytoplasm (a, c, i) and ApoA-IM is localized in the cell wall of pericarp cells (i, arrows) and aleuron cells (g,h; arrows). Only auto uorescence was observed in negative control (e, f). Magni cation bar: b: 20 µm, d, f, h, i: 10 µm. arrow; g: Golgi body) and in the cell wall, sometimes into plasmodesmata (f; arrow). In the endosperm,