生物强化赖氨酸论文

Plant Mol Biol (2015) 87:235–248DOI 10.1007/s11103-014-0272-z

Biofortification of rice with lysine using endogenous histones

H. W. Wong · Q. Liu · S. S. M. Sun

Received: 24 June 2014 / Accepted: 4 December 2014 / Published online: 17 December 2014 © The Author(s) 2014. This article is published with open access at http://wendang.chazidian.com

Abstract Rice is the most consumed cereal grain in the world, but deficient in the essential amino acid lysine. Therefore, people in developing countries with limited food diversity who rely on rice as their major food source may suffer from malnutrition. Biofortification of stable crops by genetic engineering provides a fast and sustainable method to solve this problem. In this study, two endogenous rice lysine-rich histone proteins, RLRH1 and RLRH2, were over-expressed in rice seeds to achieve lysine biofortifica-tion. Their protein sequences passed an allergic sequence-based homology test. Their accumulations in rice seeds were raised to a moderate level by the use of a modified rice glutelin 1 promoter with lowered expression strength to avoid the occurrence of physiological abnormalities like unfolded protein response. The expressed proteins were

Electronic supplementary material The online version of this article (doi:10.1007/s11103-014-0272-z) contains supplementary material, which is available to authorized users.

H. W. Wong · S. S. M. Sun (*)

State Key Laboratory of Agrobiotechnology and School of Life Sciences, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong, China

e-mail: ssun@cuhk.edu.hk

H. W. Wong

e-mail: gundamwong2003@http://wendang.chazidian.com

Present Address:

H. W. Wong · S. S. M. Sun

SCG90, Science Center, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong, China

Q. Liu

Key Laboratory of Plant Functional Genomics of the Ministry of Education, College of Agriculture, Yangzhou University, Yangzhou, China

e-mail: qqliu@http://wendang.chazidian.com

further targeted to protein storage vacuoles for stable stor-age using a glutelin 1 signal peptide. The lysine content in the transgenic rice seeds was enhanced by up to 35 %, while other essential amino acids remained balanced, meet-ing the nutritional standards of the World Health Organiza-tion. No obvious unfolded protein response was detected. Different degrees of chalkiness, however, were detected in the transgenic seeds, and were positively correlated with both the levels of accumulated protein and lysine enhance-ment. This study offered a solution to the lysine deficiency in rice, while at the same time addressing concerns about food safety and physiological abnormalities in biofortified crops.

Keywords Lysine · Biofortification · Rice

(Oryza sativa L.) · Histone · Food safety · Chalkiness

Introduction

Rice is an important staple food, supplying 20 % of the world’s dietary energy, as well as serving as the primary food source of 17 Asian and Pacific, nine North and South American, and eight African countries (FAO 2004). It is also the sole stable food source in many developing coun-tries (Pellett and Ghosh 2004) where food availability and diversity is limited (Sautter et al. 2006; Zhu et al. 2007). However, rice provides insufficient vitamin A, iron, and lysine, an essential amino acid, resulting in serious mal-nutrition in these countries (Sautter et al. 2006). Industrial supplementary and fortification programs have been pro-posed as remedial measures, but these methods are often not sustainable in developing countries because of chemi-cal instability of supplements, costs, political instability, and the logistic challenge of reaching scattered populations

1 3

236Plant Mol Biol (2015) 87:235–248

(Sautter et al. 2006, Zhu et al. 2007; Mayer et al. 2008). Biofortification through agricultural biotechnology has been proposed as a more sustainable alternative, develop-ing stable crops with enhanced nutritional value to fulfill the daily nutritional requirements of humans (Sautter et al. 2006; Zhu et al. 2007; Mayer et al. 2008; Hirschi 2009).To biofortify rice with lysine, three major approaches can be used: (1) increase the accumulation of free lysine; (2) manipulate the seed storage proteins (SSPs); and (3) overexpress lysine-rich proteins in seeds. The two key The identification of seed and endosperm-specific pro-moters in crop plants will facilitate the third approach. With these promoters, proteins with desired properties can be accumulated in crop seeds for biofortification. Examples in rice include the seed-specific overexpression of the artifi-cial nutrient-rich protein Asp-1 gene (Potrykus 2003) and the expression of a fusion gene encoding the winged bean lysine-rich protein and the rice lysine-rich glutelin 1 (GT1), which increased total lysine content in transgenic seeds by 58 % (Sun and Liu 2008). The lysine-rich protein approach enzymes in lysine biosynthesis, aspartate kinase (AK) and dihydrodipicolinate synthase (DHPS), are feedback-inhib-ited by lysine (Galili et al. 2002), so for the first approach efforts have been made to elevate lysine content by express-ing lysine feedback-insensitive forms of these two enzymes in crops. For example, expression of native feedback-insensitive AK (lysC) from E. coli and DHPS (dapA) from Corynebacterium glutamicum, Falco et al. (1995) increased the lysine content up to five-fold in canola and soybean seeds. Huang et al. (2005) successfully doubled the lysine content in corn seeds by over-expressing lysine feedback-insensitive DHPS from C. glutamicum while reducing the accumulation of zein. Another strategy is to suppress the expression of lysine ketoglutarate reductase/saccharo-pine dehydrogenase (LKR/SDH), the key enzymes in the lysine degradation pathway, using antisense or RNA inter-ference (RNAi) methods (Zhu and Galili 2004; Hournard et al. 2007). Synergistic manipulation of both lysine bio-synthesis and catabolic enzymes could further enhance the free lysine levels in transgenic maize by up to 4,000 p.p.m. (Frizzi et al. 2008) and in rice by up to 60-fold (Long et al. 2012).

The discovery of the opaque-2 (o2) mutant (Mertz and Bates 1964) in maize prompted the very different SSP approach to enhance lysine content in cereal crops. The o2 mutation significantly reduced the levels of 22-kDa α-zein in corn, which was compensated by other lysine-rich pro-teins, thus increasing the lysine level (Mertz and Bates 1964; Schmidt et al. 1990; Segal et al. 2003). The reten-tion of endogenous 22 and 19-kDa α-zeins in the rough ER of the maize mutants floury2 (Coleman et al. 1997) and De*-B30 (Kim et al. 2004) induced strong unfolded pro-tein response (UPR) and enhanced the level of high-lysine ER chaperones and binding proteins, such as ER chaper-one luminal binding protein (BiP). In rice, the knockdown of 13-kDa prolamin could elevate the total lysine content up to 56 % (Kawakatsu et al. 2010a) as a result of com-pensatory increases in lysine-richer glutelin, globulin, and BiP; however, it led to smaller protein bodies (PBs) with modified structures. Over-accumulation of BiP could also increase the total lysine content up to 2.9-fold (Kawakatsu et al. 2010b) but was accompanied by severe decreases in starch content and rice seed weight.1 3

has great potential for lysine biofortification, since the pool of protein-bound amino acids is larger than that of free amino acids in crop seeds (Galili and Amir 2013).

While research using these three approaches has led to fruitful progress, two concerns remain: (1) the potential allergenicity of candidate transgene products and (2) the occurrence of abnormalities in transgenic crops. Regard-ing allergenicity, the candidate genes used for biofortifi-cation are often foreign to the host crop (Shaul and Galili 1992; Falco et al. 1995; Huang et al. 2005) and may have unknown function (Yu et al. 2004), raising concerns for consumer acceptance and food safety (Weale 2010; Bawa and Anilakumar 2013). An historical example is the transfer of a methionine-rich protein gene from Brazil nut to soy-bean for methionine biofortification, resulting in an aller-genic product (Altenbach et al. 1989; Nordlee et al. 1996). Unfortunately, the potential allergenicity of the transgenic product is often not considered in biofortification research.The other concern is the frequent presence of physiolog-ical abnormalities in transgenic biofortified crops. Over-accumulation of free lysine in tobacco was shown to affect its vegetative growth and floral and seed development (Shaul and Galili 1992). Similar results were observed in free lysine-biofortified transgenic canola and soybean (Falco et al. 1995), which showed decreased germination rates. UPR (Urade 2007) was another abnormality in sev-eral important high-lysine maize mutants, including o2, floury2, and De*-B30 (Coleman et al. 1997; Hunter et al. 2002; Kim et al. 2004). Over-expression or suppression of BiP also triggered strong UPR in transgenic rice seeds (Kawakatsu et al. 2010b; Wakasa et al. 2011). In these cases, UPR strongly affected host gene expression pro-files, inducing the accumulation of ER chaperones (e.g., BiP) and other protein processing enzymes (e.g., protein disulfide isomerase, PDI) in the ER of seed cells and atten-uating storage protein translation (Gething and Sambrook 1992; Urade 2007; Fanata et al. 2013), resulting in reduced protein content, abnormal PBs and protein storage vacuoles (PSVs), decreased grain weight and starch content, and increased chalkiness of the crop seeds. UPR and chalki-ness were also observed in rice overexpressing the winged bean lysine-rich protein-rice GT1 fusion gene (Sun and Liu 2008; Yu 2008), in which two- to three-fold increases

Plant Mol Biol (2015) 87:235–248 in BiP and PDI levels and abnormal PBs and PSVs were detected.

In this study, we aimed to generate transgenic rice bio-fortified with protein-bound lysine while also addressing food safety and plant physiology concerns. To address food safety and allergenicity, we first surveyed the GenBank database to identify rice endogenous proteins that were rich in lysine (>10 mol%) to lessen both the possible inter-ference of a “foreign” protein (Chao and Krewski 2008) and ethical concerns. We then carried out sequence-based homology tests of these proteins against the databases of known allergens through AllergenOnline and Allermatch, as suggested by FAO and WHO (2001); this approach was recently shown to have about 94 % accuracy (Verma et al. 2011). Two endogenous histone proteins met our criteria. To avoid triggering physiological abnormalities by UPR and histone interference with normal seed physiology, we (1) carefully regulated the expression levels of the candi-date proteins to enhance the lysine level in balance with those of other amino acids (Joint WHO/FAO/UNU Expert Consultation 2007) and (2) targeted the transgene protein products to seed PSVs for stable storage to avoid possible interference with other cellular functions. Through these strategies, we were able to generate transgenic rice lines with up to 35 % more lysine than the wild type (WT) and with no significant UPR detected.

Materials and methods

Identification of candidate proteins for lysine biofortification in rice seeds

We surveyed the GenBank database to identify potential protein candidates with the following traits: (1) endogenous to rice; (2) high lysine content (>10 mol%); (3) known function or with high homology to proteins of known func-tion; (4) complete cDNA sequences; and (5) nonallergenic-ity, as determined by subjecting the protein sequences to 8-mer, 80-mer, and full FASTA searches in the allergen databases AllergenOnline and Allermatch for homology to known allergens, as suggested by the WHO/FAO guide-lines for genetically modified foods (FAO and WHO 2001).Cloning of candidate genes and functional removal of potential nuclear localization signals

To clone RLRH1 (GenBank:Os05g0113900) and RLRH2 (GenBank:Os01g0502900), the primer pair A5NS (5′-G

GGGATCCATGGACGTCGGCGTCGGCGG-3′) and A3 (5′-GGGAATTCCTAGGAGCGCGCCTGCTTC-3′) and the

primer pair B5NS (5′-GGGGATCCATGGCGCCCAAGGC AGAG-3′) and B3 (5′-GGGAGCTCCTAGATCTCGCGG 237

GAGGTGG-3′) were designed based on the respective cDNA sequences.

RT-PCR was carried out to clone the two candidate genes. Leaves of Oryza sativa ssp. japonica cv. 9983 were used to extract total RNA as described by Zheng et al. (1993). Each RT-PCR reaction used 1 μg total RNA sample and followed the protocol of SuperScript? II reverse transcriptase (Invitro-gen, Carlsbad, CA, USA). Platinum® Taq DNA Polymerase High Fidelity (Invitrogen, Carlsbad, CA, USA) were used for second-strand synthesis. The PCR products were purified for T-vector ligation and subsequent DNA sequencing.

The potential nuclear localization signals (NLSs) of the two candidate proteins were predicted using PSORT (http://psort.hgc.jp/form.html). The cloned genes were amplified by PCR (Supplementary Fig. 5) using specific primers bearing mutations to change the amino acids in the predicted signals to either glycine or alanine and so alter the signal function (Supplementary Fig. 6).Vector construction and plant transformation

Vectors p1017, containing the regular rice GT1 promoter–GT1SP–GT1 terminator expression cassette in the super binary vector pSB130M (Sun and Liu 2008), and p1011, containing the modified rice GT1 promoter pmGT1, were provided by Prof. Q. Liu. pmGT1 was cloned from p1011 to replace the original GT1 promoter in p1017. In pA1, pA2, pB1, and pB2, BamHI and SacI were used to insert the target genes into vector p1017. NcoI and SacI were used in the remaining constructs.

All constructs were made with the super binary vector pSB130M. The second T-border set of the vector contains a hygromycin R selectable marker gene. The constructs were transformed into Agrobacterium tumefaciens EHA105 by the heat-shock method and then introduced via the Agro-bacterium into primary calli derived from mature seeds of O. sativa ssp. japonica cv. 9983. Calli transformation, selection and regeneration were performed as previously described (Liu et al. 1998).Plant cultivation

Transformed rice plants were grown in a greenhouse for further analyses and identification, then propagated to the T2 generation in experimental fields at Yangzhou Univer-sity, Jiangsu Province, China, with the approval of the Min-istry of Agriculture, PRC. The field plots were randomly arranged, and T2 seeds were collected for further analyses.

Confirmation of gene integrationPrimers GT1S (5′-GAACAACACAATGCTGCGTC-3′) and A3S (5′-CTAGGAGCGCGCCTGCTTC-3′) were used for

1 3

238PCR screening of plants harboring RLRH1 and primers GT1S and B3S (5′-CTAGATCTCGCGGGAGGTG-3′) were used for RLRH2. Seeds of positive lines were germi-nated to obtain materials for amino acid analysis (AAA), Southern blot, western blot, and transmission electron microscopy (TEM).

For Southern blot analysis, genomic DNA samples were extracted from green leaves using the CTAB method (Yu et al. 2005). For RLRH1 lines, 20 μg genomic DNA from each sample was digested by NdeI (NEB) for 24 h, while in RLRH2 lines, BamHI (NEB) was used. Electrophoresis, blotting, hybridization, and detection were carried out as described by Li et al. (2009) using gene-specific digoxi-genin-labeled probes.Amino acid analysis

Seeds of WT and T2 lines harvested in the experimental fields were used for AAA. Rice samples without husks were ground to powder and dried overnight in a 55 °C oven to remove moisture. For each sample, two technical rep-licates of 0.01 g rice powder were weighed. Each sample was hydrolyzed with 1 mL 6 N HCl (H0636; Sigma, St. Louis, MO, USA) in Sarstedt 2-mL screw-cap tube, and 10 nmol L(+)-norleucine (140-07291; Wako Pure Chemi-cals, Osaka, Japan) was added. The samples were heated at 110 °C for 24 h, then the HCl was evaporated for 6 h at 65 °C. Dried samples were dissolved in 1 mL Na–S buffer (2 % sodium citrate, 1 % HCl, 0.1 % benzoic acid) and fil-tered with an Acrodisc® 0.45 μm nylon membrane syringe filter (4426T; Pall Life Sciences, Port Washington, NY, USA) for injection and analyses using amino acid analyzer L8900 (Hitachi, Tokyo, Japan).

Data obtained from HPLC were normalized with the level of norleucine for each sample. Cysteine and methionine lev-els were excluded as they are unstable in hydrolysis and were difficult to detect. Fifteen amino acids were investigated, which included: aspartic acid, threonine, serine, glutamic acid, proline, glycine, alanine, valine, isoleucine, leucine, tyrosine, phenylalanine, histidine, lysine, and arginine.Antibodies production and western blot analysis

The amino acid sequences KPAKASKDKAAKSPK-KQARS and QEAAHLARYNKKPIA were chosen to produce synthetic peptides and subsequently antibod-ies against the two histone proteins RLRH1 and RLRH2, respectively. Anti-BiP antibody was from Stressgen® (SPA-818; Enzo Life Sciences, Inc., Farmingdale, NY, USA) while anti-PDI polyclonal antibodies were from our labora-tory. For western blot analysis, 10–20 T2 seeds were ran-domly selected from each transgenic line, and their husks were removed. Total protein was extracted with buffer

1 3

Plant Mol Biol (2015) 87:235–248

containing 0.125 M Tris, pH 7, 4 M urea, 4 % SDS, and 5 % β-mercaptoethanol. The proteins in the total protein extract were separated by size using the tricine SDS-PAGE system (Schagger and von Jagow 1987) and blotted onto PVDF membrane (BioRad, Hercules, CA, USA) using the Towbin buffer system (Towbin et al. 1979). The concen-tration of primary antibodies was 1:2,000 and of second-ary antibodies 1:30,000 (A-3687; Sigma). Signals were detected using the Aurora? kit (ICN Biomedicals, Costa Mesa, CA, USA).

Transmission electron microscopy (TEM)

Immature (10-DAF) T2 seeds were harvested from inde-pendent transformants harboring WT, pA1, pA2, pA3, pA4, pB3, and pB4 constructs respectively for TEM analysis. Half of each immature seed was subjected to total protein extraction and western blot, while the other half was fixed at 4 °C overnight in a fixation solution (0.1 M Na3PO4, 0.1 % gluteraldehyde, 4 % paraformaldehyde, pH 7). After wash-ing, it was dehydrated in an ethanol series (30, 50, and 70 % for 10 min each; then 85, 95, 100 and 100 % for 20 min each) and finally infiltrated with LR White (London Resin Company, UK) overnight at 4 °C. Polymerization of LR White was finished at 60 °C in gelatin capsules after 24 h. Ultrathin sections (70–80 nm) were cut using a Reichert UltracutS microtome (Leica Microsystems, Wetzlar, Ger-many) and mounted on formvar-coated copper grids.

Immunolabeling was performed by first incubating the samples in blocking solution (5 % BSA in 0.1 M Na3PO4 and 0.1 % Tween 20, pH 7) for 1.5 h and then in anti-sera (1:250 for RLRH1, 1:100 for RLRH2; both in blocking solution) for 1 h, followed by washing and incubating in secondary antibody solution containing goat anti-rabbit IgG immunogold reagent (EMS 25104 and 25109; Elec-tron Microscopy Sciences, Hatfield, PA, USA; 1:750 for RLRH1, 1:600 for RLRH2; in blocking solution) for 45 min. After washing, the grids were counter-stained with uranyl acetate (2.5 %) and lead citrate solutions. Signals were observed in the cells of aleurone (with nuclei but very few starch granules) and endosperm (no nucleus but more starch granules) using an H7650 transmission elec-tion microscope with AMT XR40 side-mount CCD camera (Hitachi, Tokyo, Japan).

Results

Identification of endogenous rice genes encoding high-lysine proteins with low allergenic potential

We surveyed the GenBank database for rice proteins rich in lysine (>10 mol%) and carried out 8-mer, 80-mer, and full

Plant Mol Biol (2015) 87:235–248 Fig. 1 Constructs for the expression of RLRH1 and

RLRH2 in transgenic rice seeds and their predicted sub-cellular locations. pmGT1, modi-fied rice glutelin 1 promoter (1.3 kb); tGT1, glutelin 1

terminator; GT1-SP, glutelin 1 signal peptide; RLRH1-NLS and RLRH2-NLS, modified RLRH1 and RLRH2 in which predicted nuclear localization signal sequences were functionally

removed

239

FASTA searches using the allergen databases AllergenOn-line and Allermatch to the lysine-rich proteins’ sequence homology with known allergens according to the WHO/FAO guidelines for genetically modified food (FAO and WHO 2001). Two candidates, RLRH1 (NP_001054458) and RLRH2 (EAZ12048), were selected. Both showed high homology to the rice histone H2 family and were expressed at low levels in rice seeds (Supplementary Fig. 1). RLRH1 had a lysine content of 14.7 mol% and passed the 8-mer, 80-mer, and full FASTA searches (AllergenOnline results summarized in Supplementary Fig. 2). RLRH2 had higher lysine content (20.6 mol%) and passed the 8-mer and 80-mer searches, but had a marginal hit with a latex aller-gen (35.4 % identity) in the FASTA search (see Supple-mentary Fig. 3).

We cloned the two corresponding genes from total RNA of japonica rice cv. 9983 leaves by RT-PCR (Sup-plementary Fig. 4). The potential NLSs of the two candi-date proteins were located using PSORT and modified by PCR (Supplementary Figs. 5 and 6, respectively) to (1) use as complements to compare their sub-cellular locations among the expressed constructs; (2) prevent the expressed histone proteins from re-entering the nuclei and potentially interfering with normal cell physiology and function; and (3) decrease the homology of RLRH2 to the latex allergen. The two modified proteins passed all three FASTA searches (AllergenOnline results summarized in Supplementary Figs. 7 and 8, respectively).

Strategy of expressing the candidate genes

Figure 1 shows the structures of the gene constructs used in this study. Plasmids pA1 and pB1 contained a modi-fied GT1 promoter (pmGT1), a GT1 3′ UTR (tGT1), and the cDNAs NP_001054458 (RLRH1) and EAZ12048 (RLRH2), respectively. The pmGT1 drives a moderate level of expression in the aleurone and endosperm cells of rice seeds compared to the original GT1 promoter (Liu 2002). It was used in this study to reduce the extent of UPR/ER stresses and chalkiness in transgenic rice seeds. The expressed histones of the two constructs carrying the tar-geting signals are expected to enter the nucleus. Plasmids pA2 and pB2 were similar to pA1 and pB1, except that the two histone genes were modified to remove the NLS func-tion (Supplementary Figs. 5 and 6), and the resulting pro-teins were predicted to cytolocate to the cytoplasm. A GT1 signal peptide was included in constructs pA3, pA4, pB3, and pB4 to direct the expressed proteins to the ER. Simi-lar to pA2 and pB2, the potential NLS function was also removed from RLRH1 and RLRH2 in constructs pA4 and pB4, respectively.

Rice plants regenerated from transformants with posi-tive genomic PCR results were propagated to generate T1 seeds. At this stage, we found that the T1 seeds from con-structs pB1 and pB2 failed to germinate, implying that the accumulation of RLRH2 or RLRH2-NLS in the nucleus and cytosol, respectively, might cause physiological

1 3