Serena REGGI, Elisabetta ONELLI, Alessandra MOSCATELLI, Nadia STROPPA, Matteo DELL’ANNO,Kiril PERFANOV, Luciana ROSSI

(1Department of Veterinary Medicine and Animal Science, University of Milan, Lodi 26900, Italy; 2Department of Biosciences,University of Milan, Milan 20133, Italy; 3Plantechno Srl, Casalmaggiore 26040, Italy)

Abstract: Apolipoprotein A-IMilano (ApoA-IM) has been shown to significantly reduce coronary atherosclerotic plaques. However, the preparation of cost-effective pharmaceutical formulations of ApoA-IM is limited by the high cost and difficulty of purifying the protein and producing the highly effective dimeric form. The aim of this study was to create an expression cassette that specifically drives the expression of dimeric ApoA-IM in the protein bodies of rice seeds. The ApoA-IM protein under control of the 13 kDa prolamin promoter is expressed exclusively in its dimeric form within the seeds, and immunocytochemical and immunogold analyses confirmed its expression in different caryopsis tissue such as seed coat, aleurone cell and endosperm, particularly in amyloplast and storage vacuoles. A plant-based ApoA-IM production system offered numerous advantages over current production systems, including the direct production of the most therapeutically effective dimeric ApoA-IM forms, long-term protein storage in seeds, and ease of protein production by simply growing plants. Therefore, seeds had the potential to serve as a costeffective source of therapeutic ApoA-IM.

Key words: apolipoprotein A-IMilano; engineered plant; immunofluorescence; immunogold analysis; rice;seed-specific promoter

Apolipoprotein A-I (ApoA-I) is a component of highdensity lipoprotein (HDL). It constitutes approximately 65%-70% of the protein mass in HDL. HDL is a molecule that transports cholesterols and phospholipids through the bloodstream from body tissues to livers,where they can be metabolized or removed from the body (Ben-Aicha et al, 2020). HDL is often called‘good cholesterol’ due to its ability to reduce the risk of heart and cardiovascular diseases (Tardif et al,2007). ApoA-I is the subject of intense research for its anti-atherogenic properties. The infusion of ApoA-I mimetic peptides stimulates cholesterol efflux from tissues into the bloodstream (Smith, 2010).

The naturally occurring variant ApoA-IMilano(ApoA-IM)contains a cysteine residue at position 173 (R173C)that promotes disulphide bonds andthe subsequent formation of dimers. Dimers increase protein stability(Alexander et al, 2009) and promote a greater interaction with lipids and enzyme lecithin-cholesterol acyltransferase, thereby increasing the therapeutic effect of the protein (Chiesa and Sirtori, 2003). ApoA-IMdimers can form HDL particles with the same efficiency as ApoA-I (Petrlova et al, 2014). Chyu and Shah(2015) indicated that ‘HDL Therapy’ with infusions of a formulation (ETC-216: Synthetic HDL phospholipids complex) of ApoA-IMsignificantly reduces coronary atherosclerotic plaques.

Although the therapeutic effects of apolipoproteins have been proven, there are numerous difficulties in their pharmaceutical preparation, 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 faces difficulties such as the presence of undesirable affinity signals at the end of the purification progress (Rosano and Ceccarelli, 2014). Furthermore, the endotoxins ofEscherichia. coliare known to form strong complexes with apolipoproteins (Emancipator et al, 1992). The elimination of these toxins for pharmaceutical products is necessary and technically possible but requires complex and costly methods (Mamat et al,2015). Producing the more therapeutically effective dimeric form of ApoA-IMcan be accomplished by purifying monomers fromE. coli, as described in patent US6617134B1, with a subsequent dimerization process. However, this is a time-consuming and expensive procedure that is not without risks.

To overcome these challenges, there is a need for an expression system that can easily, safely, and costefficiently produce apolipoproteins in their dimeric form. The ability to introduce foreign genes into plant species through techniques such as agro-infection or direct gene transfer has opened up the possibility of using engineered plants as host organisms for the production of heterologous proteins (Schmidt et al,1997; Streatfield, 2007). Plants offer several advantages for this purpose: i) they represent an economical system, compared with mammalian cell cultures and microbial fermentation; ii) the expression of heterologous proteins can be induced in specific organs such as seeds, and tubers, where proteins are more stable(Torres et al, 2001; Onelli et al, 2017); iii) therapeutic proteins derived from plants, whether purified 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 (Rossi et al, 2014b; Oluwayelu and Adebiyi, 2016); iv) plants can perform most of the post-translational modifications (eg., glycosylation and prenylation) required for protein stability and bioactivity in the same way as other higher eukaryotes;and v) proteins can be expressed in the edible organs of plants to be consumed raw as edible vaccines,eliminating the need for downstream processes (Rossi et al, 2014a; Gunasekaran and Gothandam, 2020).When downstream processing is required, it is simpler and less expensive, especially when proteins are expressed in specific organs like seeds.

A previous study demonstrated that rice plants transformed for ApoA-IMexpression produce rice milk with valuable therapeutic effects, protecting mice from atherosclerotic plaques and suggesting the potential use of these molecules in cardiovascular diseases (Romano et al, 2018). Rice seeds are an ideal platform for molecular farming due to their high grain yield, genetic purity maintained through self-pollination,ease of transformation, and suitability for oral delivery.Rice endosperm contains starch particles and proteins:60%-80% of storage proteins are glutelins, and 20%-40% are prolamins. Protein bodies (PBs) such as PB-I and PB-II, in rice seeds are specialized storage organs for protein in seeds. Recombinant proteins bioencapsulated in PBs demonstrate higher resistance to digestion than proteins located in other intracellular compartments and remain highly stable for one year at room temperature.The administration of recombinant proteins expressed in rice seeds is safe and economically attractive as oral drugs or minimally processed functional nutrients(Takaiwa et al, 2017; Zhu et al, 2022).

The aim of this study was to evaluate an expression cassette driving the specific expression of dimeric ApoA-IMwithin the PBs of rice seeds. In addition,ApoA-IMtransformed rice plants were characterized,and the localization of the protein was investigated at different stages of seed ripening in these engineered plants.

RESULTS

Expression vector and molecular analysis of putative transformed plants

An expression vector for rice, designated pCAMBIAPROL-ApoA-IM(Fig. S1-A), was developed to carry the ApoA-IMgene under the control of the putative 13 kDa prolamin promoter. In order to confirm the correct insertion of the ApoA-IMgene and the promoter into the expression vector pCAMBIA-PROL-ApoA-IM,the plasmid was digested with different restriction enzymes and subjected to electrophoresis on agarose gels. The sizes of the digested fragments matched the expected size according to the map of recombinant plasmid (Fig. S1-B). The fusion of the promoter and gene, as well as the correct reading frame for translation,were confirmed by Sanger sequencing. A total of 100 calli derived from mature embryos of the rice variety Rosa Marchetti were transformed by agro-infection,and approximately 32 independent hygromycin-resistant T0transformed plants were obtained and then grown under the greenhouse conditions for evaluating morphology and fertility. Considering the seeds production, about 93% of the transformed plants were fertile and displayed morphological characteristics similar to the controls (untransformed plants).

PCR analysis was conducted on all independently derived engineered plants to ascertain the presence of the ApoA-IMgene using specific primers (Apo-fw and Apo-rv, Table S1). Plants harboring the ApoA-IMgene yielded a band of the expected size, corresponding to PCR product of 732 bp (Fig. S1-C). No amplification was detected in the genome of wild type seeds (negative control). Approximately 72% of the rice plants were confirmed as transformed. The absence of detectable PCR products in some plants, despite undergoing hygromycin selection, could be attributed to factors such as the development of antibiotic resistance,genetic instability after post-integration, and partial integration of the expression vector (Gelvin, 2003).

The status of the engineered plants was characterized in detail by Southern blot, and the number of transgene integration sites into the genome of engineered plants was estimated. We analyzed 16 independently derived T0engineered plants by Southern blot. Unique and complex hybridization patterns were revealed, indicating that these plants indeed originated from independent events (Fig. S2). All 16 engineered plants showed positive hybridization, with most hybridization bands between 3 kb and 8 kb. The majority of engineered plants carried multiple copies of the transgenes ApoAIM, while only four plants carried a single copy of the engineered gene (Fig. S2). No hybridization bands were found in the untransformed rice plants (used as controls). After transplanting to greenhouse, all T0engineered rice plants grew well and produced normal seeds similar to untransformed rice plants. Segregation for hygromycin resistance was observed in all independently transformed rice lines. The results indicated that the segregating ratios in plant lines (3,15, 23, and 25) followed the classic Mendelian inheritance pattern of 3:1 (resistant to sensitive) and co-segregated with hygromycin resistance, confirming a single insertion site, as shown by a single band (Fig.S2). T1plants (including both multi-copy and singlecopy plants) germinated from F1seeds were analyzed by Southern blot, demonstrating the stable integration and inheritance of exotic genes.

Expression analysis of ApoA-IM in engineered plants

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE) analysis in the absence or presence of a reducing agent (dithiothreitol or β-mercaptoethanol)was used to detect cys-cys linked covalently dimers as reported by Petrlova et al (2014). All proteins from F1seeds of PCR-positive lines were separated via SDSPAGE in the presence and absence of the reducing agent (+/- β-mercaptoethanol) followed by Western blot with antibodies specific for human ApoA-I protein. Approximately 64% of the PCR-positive lines produced the monomeric ApoA-IM, visualized as a specific band at variable levels of 28 kDa, corresponding to the molecular weight of the monomer ApoA-I (Fig.1-A and -C). Results of gene expression showed different transcription levels of ApoA-I mRNA in engineered plants (P≤ 0.05; Fig. 1-B), confirming the different band intensities observed in Western Blot. At the same time, the F1seeds producing ApoA-IMmonomers were analyzed under non-reducing conditions,and 75% of the rice seeds exhibited a second band at about 56 kDa, corresponding to the molecular weight of the ApoA-IMdimer, a feature not present in wild type rice seeds. The relative expression levels of monomeric and/or dimeric forms of the ApoA-IMprotein in seed extracts differed among the analyzed plant lines. The ApoA-IMquantity was quantified by Enzyme-linked immunosorbent assay, and the expression level was resumed in Table S2. There was no cross reactivity between the ApoA-I polyclonal antibody and any endogenous proteins in untransformed rice seeds. ApoA-IMaccumulated in rice seeds in variable quantity, with the highest recovery reaching 82 mg/kg of rice seeds. The number of gene copies of the introduced ApoA-IMgenes could not be strictly correlated with the expression levels in the same plants,suggesting that the variation in expression resulted from other factors, such as chromosomal position effects of the integrated genes in theOryzagenome.

The glycosylation prediction study showed one possible site at position 16 (NASA) located in the signal peptide of the promoter (Fig. 1-D). Prediction studies of promoter cleavage confirmed a 99.3% possibility of cleavage occurring at the same position (16), generating a free signal peptide and a mature protein. In order to investigate the presence of different isoforms of ApoA-I due to post-translational modification, transformed seed proteins were separated by two-dimensional gel electrophoresis. This technique allowed the separation of polypeptides with a isoelectric point ranging from 3.6 to 9.3 and molecular masses from 14 to 120 kDa(Fig. S3). Western blot analysis (under reducing conditions) using anti-ApoA-I antiserum showed a faint reaction with a single spot, suggesting that the protein was not post-translationally modified.

Fig. 1. Western blot under reducing and non-reducing conditions, and the model of ApoA-I dimer by Gogonea et al (2013).

Temporal and spatial expression of ApoA-IM protein

Prolamins are important storage proteins in rice, and they are classified by size as 10, 13, and 16 kDa(Muench et al, 1999). The temporal expression of the 13 kDa prolamin promoter was initially confirmed by differential display experiments conducted at different days after flowering (DAF) in rice seeds.

To investigate the role of the 13 kDa prolamin in the temporal expression of ApoA-IM, 80 μg total proteins, isolated from engineered seeds at different ripening stages (4, 8, 12, 16, 20, and 25 DAF), were loaded onto a 15% acrylamide gel under denaturing conditions. After blotting onto a nitrocellulose membrane,hybridization with antiserum ApoA-I was performed.As shown in Fig. 2-A, the expression of the ApoA-IMprotein followed a time-course pattern during seed ripening. Western blot experiments revealed that the ApoA-IMprotein was detected at 12 DAF and rapidly accumulated to high levels during seed ripening. The level of expression remained stable until at complete seed maturation, consistent with Saito et al (2012). To confirm the expression pattern of the prolamin promoter during ripening stages, qRT-PCR was performed. As shown in Fig. 2-B, the ApoA-IMtranscript began at 12 DAF and peaked from 20 to 25 DAF until complete seed maturation.

Simultaneously, a Western blot analysis was performed on different rice tissues to verify seed-specific expression.Equal amounts (80 μg) of protein extracted from seeds,leaves, culms, and roots from ApoA-IMengineered plants were blotted and hybridized under the same conditions. As shown in Fig. 2-C, the ApoA-IMsignal was only detected in the seeds, while leaves, culms,and roots showed an absence of positive hybridization,possibly due to prolonged exposure. RT-PCR confirmed that the 13 kDa prolamin promoter was active in directing ApoA-IMexpression only in the seeds (Fig.2-D), indicating that the 13 kDa prolamin promoter in association with the NOS terminator is sufficient for seed specificity (Qu and Takaiwa, 2004).

Fig. 2. Temporal and spatial expression of ApoA-IM gene.

Immunolocalization of ApoA-IM in seed tissues

The localization of ApoA-IMwas further analyzed by immunocytochemical and immunogold observations.Within the same panicle we find caryopses at two developmental stages. These different caryopses(referred to as I and II) were collected from the same panicle (25 DAF) and characterized by fluorescence and electron microscopy.

In caryopsis I, the pericarp, seed coat and nucellus were clearly observed, with only weak autofluorescence observed in these tissues after anti-ApoA-IMantibody treatment (Fig. 3-A and -B). In contrast, a high fluorescence was observed in the endosperm cells,while ApoA-IMwas excluded from aleurone cells. In the endosperm cells, ApoA-IMwas localized within amyloplasts (Fig. 3-C and -D), particularly in the stroma,surrounding starch granules which in turn appeared dark (Fig. 3-C and -D; indicated by arrows). Negative controls showed only faint autofluorescence and no cross-reaction with seed tissues (Fig. 3-E and -F).

The pattern of ApoA-IMdistribution changed in caryopsis II, in which the pericarp appeared thinner due to the flattening of mesocarp and seed coat cells(Fig. 3-G to -J). The nucellus also appeared compressed due to endosperm growth and was hardly recognizable(Fig. 3-H, -J, and -N). Immunofluorescence analysis showed that ApoA-IMwas localized not only in endosperm cells as in caryopsis I, but also in seed coat and aleurone cells (Fig. 3-G to -L), suggesting that during seed ripening, the expression and distribution of this protein changed. High fluorescence was observed in seed coat cells, while in pericarp cells, close to the seed coat, autofluorescence occurred as observed in the negative control (Fig. 3-I, -J, -K, and -L). In the endosperm, round organelles attributable to amyloplasts were fluorescent (Fig. 3-G to -J), and, interestingly,ApoA-IMwas also localized in both large and small organelles of the aleurone cells (Fig. 3-K and -L,indicated by arrows). Negative control always showed autofluorescence in pericarp cells (Fig. 3-M and -N).

To better investigate the ApoA-IMlocalization inside cells, immunogold analysis was performed.Different anti-ApoAIMantibody concentrations were used to exclude nonspecific reactions with other cell components, and the observations confirmed the presence of ApoA-IMpredominantly in the endosperm,aleurone cells, and seed coat (Fig. 4). Interestingly,ApoA-IMappeared to be preferentially localized in plastids. In the seed coat, ApoA-IMwas localized in chloroplasts (Fig. 4-A), particularly in the grana (Fig.4-B). Aleurone cells were not fully differentiated, and some starch grains were also observed (Fig. 4-D).However, numerous storage vacuoles were stained with ApoA-IMantibodies which were observed as dark spots (Fig. 4-D, arrows). In the endosperm, amyloplasts appeared bigger compared with those observed in aleurone cells and contained several starch granules(Fig. 4-E). ApoA-IMwas localized in both storage vacuoles and amyloplasts at the periphery of starch granules (Fig. 4-E; see arrows for amyloplast). Some gold particles were also observed in the cell wall,particularly in the primary wall of seed coat and endosperm cells (Fig. 4-C and -F, arrows), as well as in Golgi-associated vesicles (Fig. 4-F, arrowheads).Negative control did not show any staining in plastids,vacuoles, or the cell wall (Fig. S4).

Fig. 3. Immunofluorescence and bright field images of caryopses I (A-F) and II (G-N) of ApoA-IM transformed plants.

DISCUSSION

HDL and its major protein ApoA-I are protective against atherosclerosis through several mechanisms,including the ability to mediate reverse cholesterol transport. ApoA-I formulations and ApoA-I mimetic peptides have demonstrated rapid effects on plaque regression and stabilization (Smith, 2010). The natural variant Milano has been the subject of numerous studies and intriguing clinical trials (Nissen et al, 2003)that have highlighted its significant antiatherogenic properties (Sirtori et al, 2019). Recombinant expression of ApoA-IMhas been achieved in different systems(e.g. bacteria, yeasts, insect cells, and mammalian cells). Plants have been utilized for over two decades for the production of numerous pharmaceutical proteins,with promising results also in the realm of edible vaccines (Kurup and Thomas, 2020).

ApoA-I production inE. colireported in the literatures ranges from 0.1 to 5.0 mg/L (Lorenzetti et al, 1986;Ryan et al, 2003), but in some cases, ApoA-I protein accumulates in inclusion bodies and requires refolding after purification. Nykiforuk et al (2011) expressed ApoA-IMin a fusion protein in engineered safflower seeds at very high quantities, but ApoA-IMrice offers certain advantages, including cost-effectiveness and largescale production of readily usable edible materials.

Fig. 4. Immunogold analysis on caryopsis II of ApoA-IM transformed plants.

Under the control of an RP5 prolamin promoter,ApoA-IMwas expressed in rice seeds at a maximum level of 49 mg/kg seeds, with its activity efficiently determined by Romano et al (2018). In this study, rice plants expressing ApoA-IMunder the control of a 13 kDa prolamin promoter were further investigated,particularly regarding seed expression and its cellular localization. The 13 kDa prolamin promoter (with its signal peptide) was demonstrated as an excellent candidate for ApoA-IMexpression in rice due to its seed-specificity and stability. The maximum recovery of expression was 82 mg/kg of seeds, which is more efficient than the RP5 prolamin promoter. The promoter plays a fundamental role in determining spatial expression.The subcellular localization of a recombinant protein influences its stability, accumulation, and expression level. Protein targeting also plays a pivotal role in post-translational modifications. Glycosylation is the most relevant modification in molecular farming because glycan chains affect protein structure, biological function, and immunogenicity. Targeting to a specific compartment cannot influence the correct folding, and the accumulation site has a direct impact on protein stability and protein recovery.

Both immunocytochemical (Fig. 3) and immunogold(Fig. 4) analyses showed that the recombinant protein ApoA-IMappeared to be localized in storage vacuoles.In rice endosperm, two different protein bodies were identified: PB-I characterized by the presence of prolamin while PB-II containes glutein and globulin(Tanaka et al, 1980; Saito et al, 2012). Since ApoA-IMis an exogenous animal protein under the control of the 13 kDa prolamin promoter, it is expressed in seed tissues as expected, but its localization within organelles appeared different compared with prolamin.In liver cells, ApoA-IMis synthesized in the endoplasmic reticulum lumen and then directed to the secretory pathway as a soluble protein (Stoffel, 1984).In plant cells, the expressed ApoA-IMprotein does not possess a plant-specific signal for correct addressing in post-Golgi secretion. In the absence of specific receptors, soluble proteins may be collected in Golgiderived secretory vesicles, as observed in Fig. 4, and randomly sorted towards vacuoles (presumably both PB-I and PB-II), the plasma membrane/cell wall and plastids (Kim and Brandizzi, 2014; Baslam et al, 2016).Interestingly, the presence of ApoA-IMin amyloplasts and chloroplasts suggests that a post-Golgi vesiclemediated route is responsible for the delivery of proteins into plastids (Baslam et al, 2016). The mechanisms regulating vesicle transport to plastids are poorly understood. Three different hypotheses have been formulated: fusing/budding model, invagination model, and pass-through model (Baslam et al, 2016).As the ApoA-IMprotein is sorted by a specific mechanism, it is not possible to ascribe it to one of these hypotheses. As observed in engineered tobacco seeds, obtained by the insertion of exogenous genes coding for seed-based oral vaccines (F18 and VT2eB),modified sorting of endogenous proteins can occur in transformed plants (Onelli et al, 2017). The sorting of ApoA-IMto different organelles appeared similar at two ripening stages. The obtained transformed plants offer a promising option for large-scale ApoA-IMproduction. Further investigations are required to better establish its effect after oral administration by edible transformed rice.

METHODS

Rice material and culture conditions

The seeds ofOryza sativavariety Rosa Marchetti were used for regeneration and transformation. They were surface-sterilized with a 5% sodium hypochlorite solution for 5 min, followed by 70% ethanol for 1 min, and washed 5 times with sterile distilled water. Embryo explants were cultured on 0.5× MS (Murashige and Skoog) solid medium. All cultures were maintained under controlled laboratory conditions at (28 ± 2) °C under a 16 h light/8 h dark photoperiod with a cool white fluorescent lamp in a growth chamber. The ApoA-IMrice plants were cultivated in a biosafety greenhouse under ambient light conditions and the seeds were collected at the 4LAB Diagnostics Srl Laboratory(Casalmaggiore, Italy).

Expression vector

The 13 kDa seed-specific promoter, along with its relative 5′-UTR (untranslated region) and transit peptide sequence (NCBI Acc. No. D63901 from -674 to +57), was amplified fromOryza sativavariety Ariete using forward and reverse primers(PROL-fw and PROL-rv, Table S1). The 729 bp promoter fragment was inserted into the pGEM-T vector (Promega, WI,USA), and the accuracy of the amplified promoter region was verified by DNA sequencing (Watts and MacBeath, 2001).PstI andXbaI sites were respectively inserted in the forward and reverse primers to facilitate subsequent DNA cloning into a plant expression vector (pCAMBIA1302, NCBI Acc. No.AF234298) atPstI andXbaI restriction enzyme cutting sites.The cDNA of mature ApoA-IM, obtained from the biological collection at 4LAB Diagnostics Srl Laboratory (Casalmaggiore,Italy), was placed under the control of the rice 13 kDa prolamin promoter and cloned into theBamHI andSacI restriction enzyme cutting sites (blunted) (Fig. S1-A).

Rice transformation and analysis of engineered plants

Embryonic calli derived from mature zygotic embryos of Rosa Marchetti were inoculated withAgrobacteriumtumefaciensstrain EHA105 transformed by electroporation with the plasmid pCAMBIA-PROL-ApoA-IM. Callus and bacterial induction,transformation, selection, and regeneration of engineered tissues were performed as described by Hiei et al (1994) with minor modifications. A number of independently transformed (hygromycinresistant) plants (T0plants) were potted in peat and hardened in a greenhouse together with WT and self-pollinated to produce F1seeds. These F1seeds were then subsequently planted to produce T1plants (n= 3-5 of each line) and obtained F2seeds.During the entire experimental period, the plants were evaluated for the principal agronomic characteristics.

PCR analysis

Before flowering, T0and T1selected plants were subjected to PCR analysis to verify the presence of the transgene. DNA was extracted from leaves of the hygromycin-resistant rice lines,using the method described by Doyle and Doyle (1987). The primers Apo-fw and Apo-rv (Table S1) were used for PCR,amplifying the entire ApoA-I gene. Successful transformation would result in the visualization of a 732 bp fragment.

Southern blot analysis

Independent engineered rice plants from the T0and T1generations,positive in PCR, were subjected to Southern blot analysis to verify the number of transgene integration sites in each plant.Genomic DNA was extracted from young leaves according to Doyle and Doyle (1987). Approximately 10 μg of DNA was digested withXbaIwhich cuts only once in pCAMBIA-PROLApoA-IM, and then separated on a 0.8% agarose gel. It was subsequently transferred to a positively charged nylon membrane,and hybridized with a probe corresponding to the ApoA-IMcoding sequences, labelled with the PCR DIG Probe Synthesis Kit (Roche Molecular Biochemicals, Penzberg, Germany) as per the manufacture’s instructions. The primer set used in PCR analysis was also used to generate the DIG-labeled probe. The positive control included the vector (pGEMT-APOA-IM) linearized with the same restriction enzyme. Detection was carried out with CDP-Star?(GE Healthcare, IL, USA) following the manufacturer’s instructions. Segregation was studied by Southern blot analysis in selected plants (numbers 2-5) of the T1generation (Fig. S2-A).

Evaluation of protein expression

Western blot analysis

Seeds from the F1and F2generations collected respectively from the T0and T1PCR-positive plants were used to verify protein expression by Western blot analysis. Briefly, total proteins from rice seeds (100 mg) were extracted using an extraction buffer [50 mmol/L Tris-HCl pH 8.0, 5 mmol/L ethylene diamine tetraacetic acid (EDTA), 200 mmol/L NaCl,0.1% Triton X-100, 1 mmol/L phenylmethanesulfonyl fluoride(PMSF)] in a 1:10 ratio, mixed with SDS-loading buffer (with or without β-mercaptoethanol, to discriminate from monomer/dimer, respectively). After Bradford protein quantification, 80μg total soluble proteins were subjected to electrophoretic separation on an acrylamide gel and transferred using electroblotting (a solution of 25 mmol/L Tris, 192 mmol/L glycine, 20% methanol, 30 V at 4 °C overnight) onto a nitrocellulose membrane (Hybond ECL, GE Healthcare, IL,USA). The membrane with the bound protein was placed in a solution of PBS-T and 5% skim milk, agitated for 1 h, washed,and then exposed to the primary antibody Anti ApoA-I Goat 1:5 000 (Acris, CA, USA) and subsequently to the secondary antibody anti-goat peroxidase conjugate, at 1:12 000 (Sigma Aldrich, MO, USA). The membrane was washed several times and placed in a chemiluminescent detection solution, ECL (GE Healthcare, IL, USA). The same extraction protocol was performed for protein extraction from leaves, culms, and roots.

Two-dimensional (2D) electrophoresis

Two-dimensional gel electrophoresis was performed with 50μg extracted seed proteins precipitated with cold ethanol (1:5)and incubated for 30 min on ice. The pellet was obtained by centrifugation at 4 °C for 36 min at 15 000 ×gin an ALC A21-C rotor. Ethanol was removed, and the pellet was resuspended with 200 μL of rehydration buffer [8 mol/L urea, 4% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate hydrate(CHAPS), 65 mmol/L dithioerythritol, 0.5% bromophenol blue]supplemented with 2% IPG Buffer pH 3-10 NL (Cytiva Sweden AB, Uppsala, Sweden). This mixture was loaded onto 7-cm non-linear pH 3-10 strips and passively rehydrated overnight at room temperature. Isoelectric focusing was performed with the Multifor II system (GE Healthcare, IL, USA) according to Moscatelli et al (2015) (200 V for 1 h, 2 000 V for 3 h, and 3 000 V for 3 h and 30 min). Focused strips were equilibrated in Buffer I (0.5 mol/L Tris-HCl, pH 6.8, 2% SDS, 6 mol/L urea,30% glycerol, 2% DTE) for 12 min and then for another 5 min in Buffer II (composition the same as Buffer I, but with 2.5%iodoacetamide instead of DTE) at room temperature. SDSPAGE was conducted on a 10% polyacrylamide gel (MiniVertical Electrophoresis System, GE Healthcare, PA, USA), as previously described (Laemmli, 1970), and the gel was silver-stained(Sinha et al, 2001). After 2D electrophoresis, polypeptides were transferred onto a PVDF membrane (AmershamTMHybond,GE Healthcare, PA, USA) and tested for ApoA-I following protocol previously described for Western blot experiments.

The sequence related to the fusion of the signal peptide of the prolamine promoter and the mature ApoA-IMgene was analyzed by NetNGlyc 1.0 Server for determination possible glycosylation sites. The glycosylation site prediction analysis was performed on the primary structure according to Gupta and Brunak (2002) for estimating possible post-translational modifications of ApoA-IM.

Protein quantification in rice seeds

An indirect competitive ELISA (IC-ELISA) was used to detect ApoA-IMprotein in F1rice seeds. In particular, recombinant hApoA-I (Sigma Aldrich St. Louis, MO, USA) was coated at a concentration of 600 ng/mL onto a micro-well plate overnight at 4 °C. The plate was then washed three times with 0.01 mol/L phosphate buffered saline (PBS, pH 7, Sigma Aldrich, MO,USA) and blocked with 200 μL of 5% BSA for 2 h at 37 °C.Subsequently, the plate was washed with 0.01 mol/L PBS containing 0.05% Tween 20 (PBS-T). 100 μL of goat polyclonal anti-ApoA-I antibody (Acris, CA, USA) diluted 1 :6000 was mixed with serial dilution (1:1, 1:10 and 1:50) of rice seed protein extraction (100 μL). From each mixture, a total of 100 μL solution was added to each well coated, and the plate was incubated at 37 °C for 1 h. After incubation, the plate was washed with PBS-T and then a 1:5000 diluted solution (100 μL)of anti-goat IgG-HRP antibody (Thermo Fisher Scientific,Waltham, MA, USA) was added to each well and incubated at 37 °C for 1 h. After that, the plate was rewashed with PBS-T,and a 50 μL solution of TMB (3,3′,5,5′-tetramethylbenzidine)was added to each well followed by incubation at 37 °C for 15 min. To stop the reaction, 150 μL of 0.4 mol/L hydrochloric acid (HCl) was added to each well, and then absorbance was measured at 450 nm using an ELISA plate reader (Bio-Rad, CA,USA). Each experiment was performed in triplicates, and a standard curve using recombinant hApoA-I (Sigma Aldrich,MO, USA) protein was constructed.

RNA extraction, cDNA synthesis, RT-PCR, and qRT-PCR

Total RNA was isolated from leaves, culms, roots and seeds at different ripening stages in ApoA-IMplants for gene expression analysis. Briefly, frozen tissue was grinded with an equal volume of RNA extraction buffer (0.1 mol/L Tris-HCl with pH 8, 50 mmol/L NaCl, 1 mmol/L EDTA, 0.5% EDTA and a 1 : 1 mixture of chloroform and phenol). RNA was precipitated with a LiCl at a final concentration of 2 mol/L (Takaiwa et al, 1987).The integrity of RNA was confirmed by electrophoresis on a 1.2% formaldehyde agarose gel.

cDNA synthesis was performed by the iScript cDNA synthesis kit (Bio-Rad, CA, USA) using a mixture of random primers and oligo dT, along with 2 μg of total RNA. Subsequently,2 μL of cDNA generated from RNA extraction from leaves,seeds, culms, and roots were amplified in a PCR reaction with Apo-fw and Apo-rv primers (Table S1), as well as β-TUB primer (Jain et al, 2018). β-TUB mRNA was amplified with 20 cycles, and ApoA-IMcDNA was amplified with 30 cycles using the standard PCR program following the instruction of the manufacturer of RedTaq ReadyMix PCR (Sigma-Aldrich, MO,USA). The cDNA transcript from RNA at different ripening stages of ApoA-IMseeds was used in a qRT-PCR. qRT-PCR was performed using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, CA, USA) following the manufacturer’s instructions. The reaction was performed on the CFX OPUS system (Bio-Rad, CA, USA) with a setting block standard, and initial denaturation at 98 °C for 2 min, followed by 40 cycles at 98 °C for 10 min and annealing/extension at 60 °C for 30 s.Apo forward and reverse primers were used. The expression level was calculated using the 2-ΔΔCTmethod (Livak and Schmittgen, 2001), withbeta-tubulinas the housekeeping gene(Jain et al, 2018). The calibrator was selected as the sample with the highest CT (cycle threshold) value or the lowest gene expression for relative expression.

Immunolocalization

Rice seeds were collected, cut into small pieces, and fixed in a solution containing 50 mmol/L Hepes with pH 7.4, 2%formaldehyde, and 0.2% glutaraldehyde, overnight at 4 °C.Samples were then repeatedly rinsed with 50 mmol/L Hepes with pH 7.4, dehydrated using increasing concentrations of ethanol, and embedded in LR Gold resin (Sigma Aldrich, MO,USA) at -20 °C. Semi-fine sections (2 μm) and ultra-thin sections (80 nm) were obtained using an Ultracut E microtome(Reichert Jung, Buffalo, USA). Ultra-thin sections were collected on nickel grids (Agar Scientific, Stansted, the United Kingdom) for immunogold experiments.

Immunofluorescence analysis

Semi-fine sections were hydrated with TBS (Tris-buffered saline, 20 mmol/L Tris, 150 mmol/L NaCl with pH 7.6) for 5 min, blocked with 1% bovine serum albumin (BSA, Sigma Aldrich, MO, USA) in TBS for 1 h at room temperature, and then incubated with primary goat polyclonal anti-ApoA-I antibody (Acris, CA, USA) diluted 1:500 overnight at 4 °C.After three rinses in TBS, sections were incubated with the FITC-conjugated antigoat secondary anti-body (Santa Cruz Biotechnology, TX, USA) at a 1:200 dilution in TBS at room temperature for 2 h, followed by rinsing three times with TBS,and finally mounting the section in Cityfluor (Citofluor AF1, a triethylenediamine 5% glycerol/PBS solution) (Agar-Scientific,London, UK). Negative controls were performed using only the secondary antibodies. Fluorescence observations were performed with a Leica DMRB microscope (set filter BP450-490, RKP 410, long pass 515), and images were captured using a Leica MC170 HD camera (Leica, Wetzlar, Germny).

Immunogold analysis

Ultra-thin sections were hydrated with TBS for a few minutes,blocked with 1% BSA in TBS for 30 min at room temperature,and incubated with primary goat polyclonal anti-ApoA-I antibody (Acris, San Diego CA, USA) at different dilutions(ranging from 1:500 to 1:200) in TBS at room temperature for 2 h. After three rinses in TBS, the sections were incubated with a 20 nm gold-conjugated rabbit anti-goat secondary antibody(BBA International, NY, USA) diluted 1:100 at room temperature for 1 h, and then rinsed three times with TBS, and post-fixed with 1% glutaraldehyde for 15 min at room temperature. Subsequently, the sections were that stained with 3% uranyl acetate for 20 min in the dark, rinsed with distilled water, dried, and observed using an EFTEM LEO 912AB transmission electron microscope (Zeiss, Oberkochen, Germany)operating at 100 kV.

Statistical analysis

The results of qRT-PCR were analyzed using GrapPad Prism software (version 9.0, GraphPad Software, CA, USA). Gene expression data from different rice plant clones were evaluated with analysis of variance (ANOVA) for unpaired samples. To assess statistical differences among different ripening days, a repeated measure ANOVA was used. The normality of residuals was verified using the D’Agostino-Pearson (K2) test.Multiple pairwise comparisons were evaluated with Tukey’s honestly significant differences test. Means were considered statistically different forP≤ 0.05.

ACKNOWLEDGMENT

The authors want to thank 4LAB diagnostics, Casalmaggiore,Italy, for providing the greenhouse facilities.

SUPPLEMENTAL DATA

The following materials are available in the online version of this article at http://www.sciencedirect.com/journal/rice-science;http://www.ricescience.org.

Fig. S1. Expression vector and molecular analysis of putative transformed plants.

Fig. S2. Southern blot analysis of engineered rice plants.

Fig. S3. Two-dimensional gel electrophoresis (A) and Western blot of bidimensional gel (B) of 50 μg of rice polypeptides extracted from ApoA-I transformed seeds.

Fig. S4. Negative control of seed tissues did not show antibody staining.

Table S1. Oligonucleotides used in this study for amplification of the ApoA-I gene and the 13 kDa prolamin promoter.

Table S2. T0rice plants obtained from transformation, relative controls (PCR, Western blotting, and Sothern analysis), and ApoA-IMquantification by indirect competitive enzymelinked immunosorbent assay (ELISA).