Characterization of Carotenoid-protein Complexes and Gene Expression Analysis Associated with Carotenoid Sequestration in Pigmented Cassava (Manihot Esculenta Crantz) Storage Root

Plant Material and Tissue Preparation

Cassava Storage Root (CSR) from field-grown plants at EMBRAPA Genetic Resources and Biotechnology Research Unit (Brasilia, DF. Brazil) were used in this study. For protein extraction, quantification and protein blot analysis we used landrace BGM019 (intense yellow root) and the cv. IAC12.829 (white root). Protein sequence from carotenoid-protein complex was obtained from landrace BGM019. For gene expression analysis we used landrace BGM019 and cv. IAC12-829. Fresh storage roots (cylinders of 30-40 cm length x 4-6 cm diameters) were manually sliced and immediately frozen in liquid nitrogen, freeze-dried and stored at -80°C until used. For the chromoplast-enriched suspension studies, fresh intact storage roots were peeled off and processed immediately after harvest.

Protein Extraction, Quantification and Blot Analysis

One gram of freeze-dried CSR tissue was rehydrated and added with 2 ml of extraction buffer (pH 6.8, 5% SDS, 10% glycerol, 80 mM Tris, and 25 mM DTT), and 20 ml acetone and incubated at –20°C for more than 1h to obtain a protein precipitate free of pigments. After centrifugation (30000 rpm @ 4°C for 20 min) the supernatant was drained, the pellet dried by blowing N₂ air and suspended with 5 ml extraction buffer (EB). The re-suspended solution was centrifuged again (30000 rpm @ 4°C for 10 min) to collect 1ml aliquot of the supernatant that was stored at –80°C until use. Aliquots (100 µl) were treated with DOC (0.15%) and TCA (12%) to precipitate soluble proteins. After centrifugation (13000 rpm @ 4°C for 20 min) the pellet was suspended in 100 µl of AB, and 50 µl was used for tissue protein estimation by the BCA Micro Kit method (PIERCE) according to the manufacturer protocol. Proteins from crude extract (CE), peak1 and 2 (30 µg each) were separated by 12% SDS-PAGE and blotted onto nitrocellulose as previously described [13]. Transblots were blocked with nonfat dry milk and incubated with primary antibody to Or-protein or Fibrillin-protein at a 1:200 dilution and secondary AP-conjugated goat-anti-rabbit IgG antibody (Bio-Rad) at 1:3000 dilution. Western Blue Stabilized Substrate (Promega) was used for chromogenic detection.

Preparation of Chromoplast Enriched Suspensions

Fresh CSR was harvested, washed with tap water, sliced, and homogenized with Homogenate Buffer (HB) adjusted to pH 8.2 (100 mM Tris-HCl, 8 mM EDTA, 10 mM KCl, 2 mM MgCl₂, 400 mM sucrose, and 1 mM PMSF) using a household blender to obtain a paste that was filtered through three layers of cheesecloth. The filtrate was centrifuged (600 rpm @ 4^oC for 30 min) and the resulting supernatant was centrifuged (2000 rpm @ 4^oC for 30 min) a second time. The resulting pellet was suspended by brushing the surface with 50 mL HB, washed once with HB, followed by centrifugation (2000 rpm @ 4^oC for 30 min). The final pellet was collected by surface brushing with 50 mL of suspension buffer (SB) adjusted to pH 6.8 (100 mM Tris-HCl, and 250 mM NaCl), named Non-denatured Chromoplast-Enriched Suspension (NDCES), and used for separation of carotenoid-protein complex and non-carotenoid associated proteins by Size Exclusion Chromatography (SEC).

Size Exclusion Chromatography of NDCES

Chromoplast-enriched suspensions were sonicated for 10 min and 5 mL was loaded onto a column (1 m length x 1.5 cm internal diameter) filled with a resin-bed of Sepharose CL 6B-200 and equilibrated with SB. Fractions (1 mL) were collected at a flow rate of 1.5 mL min^-1 and OD readings @ 280 nm and 461 nm were recorded using a spectrophotometer (Spectra max Plus_384, Molecular Devices) using UV transparent microplate. Fractions from two observed peaks were pooled separately, named Peak 1 (fractions 33-49, carotenoid-protein complexc - CPC) and Peak 2 (fractions 108-122, non carotenoid-protein complex - NCPC), concentrated, and stored at –80°C until further analyzes.

HPLC Analysis of Peak 1 Carotenoids

Pooled fractions 33-49 from SEC were treated with acetone to isolate the carotenoid constituent, which was further separated on a C30 column according to [14] to identify the composition of carotenoids present in Peak 1.

Protein Preparation for LC_MS/MS Analysis

Pooled protein fractions from peaks 1 and 2 were separated by one-dimensional SDS_PAGE as previously described [10,15,16] and gels zones were excised and used for in-gel protein digestion with trypsin. Equivalent protein amount (100 mg) from both peak 1 and 2 were individually transferred to a siliconized eppendorf tube and dried in a speed-vac devise. The dried pellet was re-suspended in 200 mL of TCEP solution (10 mM) prepared fresh in ambicard and incubated at 56°C for one hour and cooled to room temperature. Proteins were then precipitated with 4 volume of acetone after incubation at -80°C for one hour. After centrifugation at maximum speed for 30 min at 4°C, the pellet was collected and added to 200 mL of Iodacetamide (55 mM) in ambicarb, and incubated for one hour at room temperature in the dark. The sample was incubated at -80°C for one hour with 4 volumes acetone and then centrifuged (maximum speed @ 4°C for 30 min). The resulting pellet was incubated with 100 mL of ambicarb by vortexing and sonicating for one hour. A trypsin solution (20 mg mL^-1 in ambicarb) was added to have a protease:protein ration of 1:50, incubated at 37°C for the first two hours, after which an additional 100 mL of the same trypsin solution was added for incubation over night. After cooling to room temperature the sample was used for mass spectrometry analysis.

LC-MS/MS Analysis

After initial protein digestion, peptides were ionized, detected, isolated, fragmented and sequenced in a qTOF MALDI MS/MS. The operating software generated MS and MS/MS spectra, and the peptide sequences were determined using the PROTEOMIC facilities and software available at USDA/ARS-National Animal Disease Center in Ames, Iowa.

Bioinformatics Analysis of Peptide and Protein Sequence Data

Peptide sequences were established using a MASCOT server database (http://www.matrixscience.com/cgi/search_form.pl?FORMVER=2&SEARCH=PMF) and statistical validation as well as protein sequence matching identification. Peptide sequences that produced no protein hits were further analyzed using tblastx and translation tools available in HExPASy at Proteomics Server (http://www.expasy.org/) using four cassava EST databases located at CIAT (http://webapp.ciat.cgiar.org/biotechnology/bioinformatics.htm), ESTIMA (http://titan.biotec.uiuc.edu/cgi-bin/ESTWebsite/estima_start?seqSet=cassava), EMBRAPA (http://www.genoma.embrapa.br/), and RIKEN (http://www.brc.riken.go.jp/lab/epd/Eng/species/manihot.shtml). Finally, the protein fragment sequences derived from the EST matching results were mapped to an Arabidopsis protein database [http://www.arabidopsis.org] and to cassava PROTEOME and GENOME databases [http://www.phytozome.net/cassava]. Information on cassava proteome loci, for identified protein, and genome scaffold (for their corresponding gene) were obtained to characterize their location and distribution of identified proteins in the carotenoid-protein complex.

RNA Isolation, Quantification, and cDNA Synthesis

Total RNA was extracted twice using the phenol chloroform procedure described previously [17] and treated with RNAse free DNAse. Total RNA was quantified using the Quant-It RiboGreen RNA Kit according to the manufacturer (Molecular Probe). Total RNA (400 ng) was reverse transcribed in a 20 μL reaction volume using a SuperScript III Platinum Two-Step qRT-PCR Kit (Invitrogen). The reactions were terminated by heat inactivation and the cDNA products were treated with RNase H and stored at –20°C.

Gene-Specific Primers, Certified Standard and Housekeeping Gene Primers

Fluorogenic primers (FAM labeled LUX primer) and corresponding unlabeled primers were designed using the LUX Designer-Desktop version (Invitrogen) for each sequence of the cDNA fragment coding for genes as shown in Table S1 (see supplementary data). Standard certified primers for 18S ribosomal (Invitrogen Cat. 115HM-02) and Gus (Invitrogen Cat. 112H-02), and qPCR plasmid Standards (Invitrogen Cat. 11741-100) with Gus ORF were used as internal controls and quantitative standards to generate standard curves respectively. All primers were synthesized and purchased from Invitrogen.

Quantitative Real Time PCR Analysis

Gene sequences coding for proteins identified from the carotenoid-protein complexes were used to quantify transcript abundance in white and intense yellow CSR using (qRT-PCR). The qRT-PCR assays were performed in triplicate for each extraction using a Bio-Rad system (BioRad model iCycler). Ribosomal RNA (18S) was used as an internal control and PCR efficiency was evaluated using GUS certified primers to obtain a quantitative standard curve for absolute expression analysis of target transcript level in samples. Each experiment was replicated at least three times. Alternatively, a relative expression assay using MEC 1 (gene coding for Pt2L4 protein, 17) was used as an internal control. qRT_PCR products were sequenced at the EMBRAPA Genetic Resources and Biotechnology Genome facilities.

Data Analysis

To access the intrinsic property for the level of each transcript corresponding to the target gene in all samples, qRT-PCR amplification procedures consisted of absolute quantification expression using total RNA as normalizer. These methods provided an absolute quantitative standard curve for a dilution series based on the qPCR plasmid standards kit protocol (Invitrogene). Preliminary experiments with two unknown samples were performed to set up qRT-PCR amplification optimization conditions, including reliable exponential phase of amplification, qRT-PCR efficiency ranging from 90% to 105%, Ct threshold and baseline, and evaluation of each primer sets for target gene within an unknown sample. After setting up the optimized conditions, Ct values for all unknown samples and replications were interpolated from the standard curve estimating the level of transcript in unknown samples. In addition, a five point curve made with each amplicon was performed to obtain the efficiency of each primer (needed for the ΔΔct formula) and to know the range of amplification. After setting up the optimized conditions, Ct values for all unknown samples and replications were collected and normalized with housekeeping genes (internal controls) using the ΔΔct formula. Statistical parameters, such as mean and mean standard deviation, were performed in Excel.

Carotenoid-protein Complex Isolation and Characterization

Proteins from non-denatured chromoplast-enriched suspension (CES) present in white and intense yellow cassava, containing chromoplast-associated proteins (CAP) were separated by SEC Fig. (1A and 1B), respectively.

Size exclusion chromatography profiles monitored with Abs₄₆₁ (carotenoids) and Abs₂₈₀ (proteins) indicated the presence of two major peaks. Peak 1 (fractions 33-49) eluted within the void volume of the column, as determined with blue dextran (MW 2000 kDa) and overlap between the carotenoid and protein profiles indicates that Peak 1 represents a carotenoid-protein complex (CPC) with an estimated size close to 400 KDa. Peak 2 (fractions 108-122) eluted later in the SEC profile and no observed overlap between carotenoids and proteins in Peak 2 suggests it represents non-carotenoid-associated proteins (NCPC). Proteins present in CPC and NCPC were separated by SDS-PAGE to visualize protein complexity and to fractionate proteins for LC_MS/MS analysis. From the gel image (see Supplementary data complementing Table S2 and S5), a greater number of proteins were observed in fractions forming NCPC than in fractions forming CPC. Peptide and protein sequences from Peaks 1 and 2 identified 83 and 106 peptides that assembled into 34 and 38 proteins respectively. Carotenoids from pooled fractions of CPC, separated by HPLC, indicate that β-carotene is the major carotenoid of the CPC isolated from both white (Fig. 2A) and intense yellow cassava (Fig. 2B) phenotypes. These results confirm the presence of carotenoid-protein complexes in cassava storage root that is similar to the one reported for carrot root [18-20] in terms of size, but different in the type of carotenoid and proteins types that includes β-carotene, α-carotene, and lutein as the major carotenoids and a single major protein of 18kDa in the CPC of carrot.

Peptide Sequence and Protein Identification in Chromoplast Enriched Suspension (CES)

Reverse-phase capillary chromatography (capillary LC) with quadrupole time-of-flight tandem spectrometry (nano-ESI Q-TOF MS/MS) was used with in gel tryptic digestion of proteins from CPC and NCPC. The LC-MS/MS peptide sequences were validated by MASCOT (http://www.matrixscience.com/cgi/search_form.pl?FORMVER=2&SEARCH=PMF) and used to identify proteins. The proteins sequences identified in CPC and NCPC are listed in the Supplementary data files in Table S2 and Table S5. A total of 65 proteins were identified, with 39 identified in NCPC and 26 in CPC. The majority of the proteins identified in both CPC (Fig.3A) and NCPC (Fig. 3B) are heat shock proteins (HSP) that are differentially distributed in the two fractions , which are the focus of discussion in this work. Out of 10 HSPs observed in both CPC and NCPC, four of them are present only in CPC (HSP15.7, HSP17.4I, HSP18.1I, and HSP26.5), four are present only in NCPC (HSP17.4, HSP17.6II, HSP26.5I and HSP70.1) and the remaining six are present in both fractions. This differential distribution of HSPs between these two fractions may indicate (1) differential properties of these HSPs to complex with carotenoids, (2) differential solubility in the SDS-PAGE gel separation procedure used in the present work, or (3) aggregation dynamics of a particular HSP in vitro or in vivo as reported by others [21].

Occurrence of Fibrillin- and Or-protein in White and Intense Yellow Cassava

Protein-blot analysis was used to test for the presence of Fibrillin- and Or-proteins as markers for fibrillar structure of chromoplast occurrence in CSR as suggested in others plant systems [7,22], as well as chromoplast formation and differentiation [23,24] respectively. Fibrillin- (Fig. 4A) and Or-protein (Fig. 4B) are present in CES and NCPC of intense yellow cassava, but not in CPC isolated fractions or CES of the white cassava. These results suggest that the type of chromoplast structure accumulating carotenoid in cassava is of fibrillar-type . Since Fibrillin- and Or-proteins were not detected in the CES or in CPC of white cassava, these results also indicate that intense yellow root accumulate more chromoplasts than white roots of cassava. At the same time, these results suggest that further microscopic studies will be needed to confirm the form, number and the behavior of chromoplast in CSR, which was not addressed in the present work.

Assign Assessed CAP Sequences to Cassava PROTEOME and GENOME

List of peptide and identified proteins sequences present in CPC are showed in Supplementary data in Table S2 and Supplementary data in Table S3 respectively. List of peptide and identified proteins sequences present in NCPC are showed in Supplementary data in Table S4 and Supplementary data in Table S5 respectively. The identified proteins were assigned to cassava PROTEOME and GENOME databases and results are summarized in Tables 1 to 4.

It was observed that 30 identified proteins in CPC were distributed in 28 cassava-protein loci and 19 families while 33 proteins in NCPC are distributed in 44 loci and 25 families. The HSP group individuals for CPC are located in 9 loci. and distributed in 2 protein families. The gene coding for those proteins are distributed in 30 scaffolds for the CPC and 43 for the NCPC in the cassava GENOME. For the case of HSP, the genes coding for 10 HSP identified in CPC and in NCPC showed different distribution pattern, being 11 scaf folds for the CPC group and 15 scaffolds for the NCPC group. These results contribute to the overall observation about the functional complexity of HSPs in plants as well as the distribution and arrangement of their genes in the cassava genome.

Gene Expression Analysis

Expression of two genes, one related to chromoplast structure definition (Fibrillin gene) and another related to chromoplast development (Or-gene coding for Or-protein), were analyzed to gain insights on abundance and formation of chromoplast in CSR. Four identified gene sequences coding for proteins in the CPC (HSP18.1, HSP17.6, HSP17.4 and HSP21) were used to search for their possible roles in differential sequestration of β-carotene in white and intense yellow genotypes. Transcript levels for the above listed genes were more abundant in the intense yellow than in white CSR (Fig. 5) and followed the pattern of carotenoid accumulation.

The dissociation temperature curve patterns shown in Fig. (6) of the qRT-PCR amplification of the four genes listed above indicates a change in nucleotide coding sequence for HSP21 (Fig. 6D), which is associated with high amounts of carotenoid accumulation in the intense yellow landrace. The overall observation about the functional complexity of HSPs in plants as well as their distribution and arrangements of their genes in the cassava genome. Indeed, the sequence of the corresponding amplified fragment (Fig. 7) revealed four single point mutations in the gene coding for this protein that may favor carotenoid sequestration in the CSR of intense yellow phenotype.

In addition, these results corroborate those observed by protein blot analysis where Fibrillin- and Or-protein support a role for enhancement of protein carotenoid sink capacity (i.e. chromoplast formation and number) as well as a possible specific role for HSP21 in the carotenoid sequestration mechanism.