Molecular characterization of Leucoanthocyanidin reductase and Flavonol synthase gene in Arachis hypogaea.

Arachis hypogaea (peanut) is a potential source of bioactive compounds including flavonols and proanthocyanidins, which have gained particular interest of metabolic engineering owing to their significance in the growth, development and defense responses in plants. To gain insight of proanthocyanidins and flavonols production in A. hypogaea, Leucoanthocyanidin reductase (AhLAR) and Flavonol synthase (AhFLS) enzymes responsible for their production, have been structurally, transcriptionally and functionally characterized. Structural and functional analysis of putative protein sequence of AhFLS indicated two functional motifs 2OG-FeII_Oxy and DIOX_N, while six functional motifs belonging to the families of NAD-dependent dehydratase, 3, β hydroxysteroid dehydrogenase and NmrA-like family were observed in case of AhLAR. Promoter sequence analysis unraveled several promoter elements related to the development regulation, environmental stress responses and hormonal signaling. Furthermore, the expression analysis of AhFLS and AhLAR and accumulation pattern analysis of proanthocyanidins and flavonols in three selected cultivars of A. hypogaea under saline environment confirmed their role against salinity in genotype-dependent and stress level-dependent manner. Correlation studies revealed that AhFLS and AhLAR expression is not directly dependent on the antioxidant enzymes activity, biochemical and growth parameters but higher Pearson r value depicted some level of dependency. This detailed study of AhLAR and AhFLS can assist in the metabolic engineering of flavonoid biosynthetic pathway to produce stress tolerant varieties and production of proanthocyanidins and flavonols at an industrial scale.

Introduction

Flavonoids are phenolic secondary metabolites that perform diverse array of metabolic functions in higher plants (Saito et al., 2013). Owing to their structural diversity, most of the flavonoids exhibit biological activity in the field of organic synthesis, natural product chemistry, food chemistry, medicinal chemistry, biology, toxicology, and biochemistry (Rauter et al., 2018). Plants produce flavonoids as oviposition stimulants, visual attractors, phytoalexins, allelopathy, antimicrobials, feeding repellents, photoreceptors, and antioxidants (Nabavi et al., 2020). However, these functions are not well conserved within plant species (Pelletier et al., 1997).

Flavonoids are widely distributed in Kingdom Plantae and their biosynthetic pathway has been explored extensively, both molecularly and biochemically (Yonekura-Sakakibara et al., 2019). During evolution, flavonoids have diversified from its original functions i.e., regulation of phytohormone and UV irradiation (Rausher et al., 1999). Consequently, developing metabolic pathways that assist in acquiring diversity in specialized metabolites (Yonekura-Sakakibara et al., 2019). Mostly, the enzymes involved in secondary metabolism are derived from primary metabolism e.g., early pathway enzymes of flavonoid biosynthetic pathway, Chalcone synthase (CHS) and Chalcone isomerase (CHI), originated from fatty acid metabolism. However, the late pathway enzymes include members from superfamilies of short-chain dehydrogenase/reductase (SDR), CYP and 2OGD evolved as a result of gene duplication (Ngakiet al., 2012, Yonekura-Sakakibaraet al., 2019).

The early pathway genes happen to be regulated in coordination, whereas the late pathway genes such as Flavonol synthase (FLS) and Leucoanthocyanidin reductase (LAR) exhibit a distinctive expression pattern in a species or tissue specific manner. FLS and LAR belong to two distinct branches of flavonoid biosynthetic pathway and lead to the production of flavonols and proanthocyanidins, respectively (Nabavi et al., 2018; Zhang et al., 2020). Both enzymes are of particular interest in the field of metabolic engineering, owing to the significance of their products. Additionally, flavonols and proanthocyanidins have imperative role in the defense response against both biotic and abiotic stress in plants due to their free radical scavenging and antioxidant activity (Andersen and Markham, 2005, Muhlemannet al., 2018).

Oxidative stress generated by reactive oxygen species (ROS) initiates flavonoid biosynthesis, since severe stress conditions hinder the activity of antioxidant enzymes such as superoxide dismutase (SOD), catalase (CAT) and ascorbate peroxidase (APX) (Nabavi et al., 2020). Flavonols and proanthocyanidins are naturally occurring antioxidants that not only scavenge ROS but also inhibit their production (Jiaet al., 2012, Martinezet al., 2016). Various studies have suggested that flavonoids establish a secondary ROS scavenging system because profuse excitation energy is being directed to photosynthesis due to severe oxidative stress, however, these studies are not conclusive (Martinez et al., 2016).

The presented study is designed to gain insight of genetic and functional diversity of AhFLS and AhLAR in Arachis hypogaea (Peanut) that harbors a plethora of diverse secondary metabolites. In-silico functional analysis has been performed on AhFLS and AhLAR gene isolated from Arachis hypogaea to verify its genetic and functional diversity. Furthermore, expression pattern analysis of AhFLS and AhLAR in three selected cultivars of A. hypogaea has been determined to evaluate their genotype dependent expression. Comparative analysis was carried out to assess the association of flavonols and proanthocyanidins production with growth, biochemistry and antioxidant responses in A. hypogaea under the influence of salinity induced oxidative stress. Although FLS and LAR have been characterized in numerous plant species, but this study is the first ever detailed report on FLS and LAR in A. hypogaea. The findings of this study would assist in understanding the accumulation pattern of flavonols and proanthocyanidins under salinity stress and would aid in the development of a salt tolerant variety through metabolic engineering.

Materials and methods

Plant material and salt treatments

Seeds of three commercial cultivars of A. hypogaea (A. hypogaea cv. Bari 2011, A. hypogaea cv. BARD 92 and A. hypogaea cv. Golden) were collected from National Agriculture Research Centre (NARC), Islamabad, Pakistan based on the yield potential. Among them, Bari 2011 is a high yielding cultivar (yield potential = 6300 Kg/ha), Golden has a moderate yield potential of 4100 Kg/ha and BARD 92 has a yield potential of 2500 Kg/ha. (Naeem-Ud-Dinet al., 2012, Saeed and Hassan, 2009). Dehusked seeds were surface sterilized by a method 70% v/v ethanol for 2 to 3 min and kept at 4 °C for two days to break dormancy. Seeds were sown in an autoclaved mixture of soil and coconut husk (1:1 ratio) in small pots and kept in a glass house chamber at 28 °C under 16/8 h light/dark condition. Plants were watered according to field capacity measured through the difference in saturated wet and completely dried weight of potting media (Lopez and Barclay, 2017). Salt treatment was given to three groups of two-week-old seedlings. NaCl solution was added at concentrations of 0 mM, 60 mM, 120 mM, and 240 mM, respectively, for two weeks on alternate days according to the field capacity (Chavan and Karadge, 1980). The tissue samples were collected in plastic tubes on day 15 of the treatment and frozen immediately in liquid nitrogen followed by storage at −80 °C for evaluation of antioxidant enzymes activities, biochemical and expression analysis. Remaining plant parts were harvested for measuring the difference in growth and moisture content among the treatments.

Growth measurement and moisture content analysis

The length (cm), fresh and dry weight (g) of both shoot and root were recorded for control and treated plants after harvest. The fresh weight was measured instantly at the time of harvest. For evaluating dry weight, same plant samples were wrapped in aluminum foil and placed in a drying oven for three days at 70 °C. The dry and fresh weight of both root and shoot were used to determine the percentage of moisture content in control and treated groups by using the formula (dry weight - fresh weight) / fresh weight × 100 (Pokhrel and Dubey, 2013). All experiments were performed in triplicate.

Biochemical assays

The effect of salinity on chlorophyll, soluble sugar and soluble protein contents was evaluated through spectrophotometric methods from leaf samples. The ethanol method was used to determine leaf chlorophyll content (Lichtenthaler, 1987) and the optical density (OD) was measured at 666 nm using UV Visible Spectrophotometer (AE-S70-2U). The final estimation of chlorophyll content in mg/mg FW was calculated by the formula (OD of chlorophyll − 0.01) × 10 / (92.6474 × Dry weight of sample). The concentration of soluble sugar was evaluated by using a glucose standard curve after measuring OD of the sample at 490 nm following the methodology of Dubois (Duboiset al., 1956, Zhang and Huang, 2013). Total soluble protein was determined by recording OD at 595 nm using the Bradford assay (Bradford, 1976) after protein extraction (Elavarthi and Martin, 2010). The final concentration was evaluated through bovine serum albumin (BSA) standard curve analysis.

Enzymatic assays

The extracted proteins were further analyzed for the antioxidant activity of CAT, APX and SOD. CAT activity was estimated by measuring the decrease in the absorbance (Ab) at 240 nm by using diluted leaf extract according to Abei’s protocol (Aebi, 1984). Final CAT activity was calculated by extinction coefficient (EC) of H2O2 (39.4 mM/cm at 240 nm) using the formula (change in Ab × total volume) / (sample volume × EC × min × Fresh weight). APX activity for each sample was recorded in mM/gFW by using EC (2.8 mM/cm), using the same formula as for CAT after taking OD of 3 mL mixture prepared by using the methodology of Nakano (Nakano and Asada, 1981) at 290 nm for 2 min with an interval of 30 s. However, SOD activity was measured by following the methods described by Beauchamp (Beauchamp and Fridovich, 1971). OD of both non-illuminated and illuminated sets of Ab was recorded at 560 nm and SOD activity was calculated through (Ab of sample - Ab of blank) / Ab of control (50/100).

RNA isolation and cDNA synthesis

Fresh leaves of A. hypogaea were subjected to total RNA extraction through TRIzol method (Jaakola et al., 2001). The pellet of RNA was dissolved in nuclease free water. The integrity and quality of RNA was evaluated through agarose gel electrophoresis, while it was quantified using Nanodrop spectrophotometer (Thermo Scientific, Wilmington, DE, USA). 1 µg of extracted RNA was used for cDNA synthesis using oligodT primer through RevertAid First Strand cDNA Synthesis Kit by Thermo Fisher Scientific.

Identification and isolation of AhFLS and AhLAR

The full-length coding sequences (CDS) of AhFLS (XM_025812267.1) and AhLAR (XM_025846016.1) was retrieved from National Center for Biotechnology Information (NCBI) and confirmed by genome sequences of A. hypogaea from PeanutBase (Bertioli et al., 2015). The nucleotide and amino acid sequences of AhFLS and AhLAR were analyzed through BLAST, respectively. Best hit published genes sequences were retrieved from GeneBank and aligned with AhFLS and AhLAR, respectively, through multiple sequence alignment tool of Geneious version 9.1.5 at default settings (Kearse et al., 2012). Reverse Transcription-polymerase chain reaction (RT-PCR) primers were designed from conversed region of both AhFLS and AhLAR, respectively (Table S1). Successfully amplified products were cleaned through ExoSAP-IT™ PCR Product Cleanup Reagent (Thermo Scientific, Wilmington, DE, USA) and transformed with pCR™2.1 Vector of TA Cloning™ Kit (Thermo Scientific, Wilmington, DE, USA) into the competent cells of E. coli DH5α through electroporation. Confirmed transformed clones were subjected to Sanger sequencing. Newly obtained sequences of AhFLS and AhLAR were submitted to NCBI for accession number and used for in silico analysis.

In silico analysis

Newly obtained AhFLS and AhLAR gene sequences were translated through ExPASy translate tool. Both nucleotide and amino acid sequence of AhFLS and AhLAR were analyzed by BLAST. Best hit published genes sequences were retrieved from GeneBank and aligned with AhFLS and AhLAR through multiple sequence alignment tool of Geneious version 9.1.5 (Kearse et al., 2012) at default setting. Phylogenetic analysis was performed through the Geneious tree builder tool using maximum likelihood method. Furthermore, AhFLS and AhLAR gene sequences were analyzed to determine the protein family using motif finder server (https://www.genome.jp/tools/motif/). While active site was predicted from online tool COFACTOR (). PeanutBase (Bertioli et al., 2015) were used to predict promoter region sequences approximately 1500 bp upstream to 5′ end of AhFLS and AhLAR gene and then analyzed for cis-regulatory elements using PlantCARE database (Lescot et al., 2002).

The conserved domains of AhFLS and AhLAR were identified through Conserved Domain Database (CDD), NCBI. The protein sequence of AhFLS and AhLAR were used to calculate molecular weight and theoretical isoelectric points through Compute pI/Mw tool of ExPASy protparam tool () (Gasteiger et al., 2005). Furthermore, putative target localization of AhFLS and AhLAR was predicted by using WoLF PSORT (Horton et al., 2007). Secondary structure of AhFLS and AhLAR were predicted through PSI-blast based secondary structure prediction (PSIPRED) and Chou and fasman secondary structure prediction (CFSSP) server (http://www.biogem.org/tool/chou-fasman/). The three-dimensional (3D) protein structures were predicted through SWISS-MODEL ExPASY (Schwede et al., 2003) and demonstrated with the PyMOL viewer. The predicted protein model of AhFLS and AhLAR were evaluated and verified from ERRAT, QMEAN and Ramachandran plot. The model in the specified (.pdb) format was submitted to Protein Model Database (PMDB).

Expression analysis of AhFLS and AhLAR

Relative expression of AhFLS and AhLAR were determined in leaf samples of control and salt treated plants of three cultivars of A. hypogaea using real-time polymerase chain reaction (RT-qPCR). Maxima SYBR Green/ROX qPCR Master Mix by Thermo Fisher Scientific was used in Real-Time PCR in 7300 Fast Real-Time PCR system (Applied Biosystem, U.S.A.). Livak method (2-ΔΔCt method) was used for calculating fold change in expression and actin was used as a housekeeping gene (Livak and Schmittgen, 2001). Each sample was analyzed in triplicate. Primer sequences used in Real-Time PCR has been given in Table S1.

Fluorescence spectroscopy

Fluorescence spectra of both control and salt treated samples of the three cultivars of A. hypogaea on day 15 of the salt treatment were obtained using the spectrofluorometer system (FluoroMax-4, Horiba Scientific, USA) operated via FluorEssence software version 3.5, Horiba Scientific, USA. 150 W ozone free, xenon arc-lamp was a source of continuous wavelength light and an R928P photomultiplier tube with photon-counting detector as the signal detector (Ali et al., 2018). Synchronous spectra (Excitation range: 200–700 nm, offset: 60) for evaluating flavonoids in freshly plucked leaf samples from both control and treated groups were used for assessing spectral deviation. Three leaves from each of the replicates were used to analyze each study group. Flavonoids in root were detected at 380 nm as excitation wavelength and emission spectra range: 395–550 nm for all studied groups. Thoroughly washed tap roots were used for fluorescence spectroscopy immediately after their uprooting from the soil. The fluorescence spectrum for tap roots was recorded twice for each of the replicates. Spectra generated from leaves and roots were processed and analyzed through Origin software (OriginLab Corporation, USA) and GraphPad Prism® version 5.01 (GraphPad Software Inc., San Diego, CA, USA). Fluorescence spectra of both leaves and root for each variety were analyzed separately for spectral deviation between control and treated groups. Preprocessing of obtained fluorescence spectra involve removal of outliers and normalization while ANOVA was performed for inferential statistics of the obtained data.

Data analysis

Data collected from growth measurement, moisture content analysis, antioxidant enzyme activity, biochemical analysis and expression analysis were subjected to descriptive and inferential statistics. Microsoft® office 365 Excel and GraphPad Prism® version 5.01 (GraphPad Software Inc., San Diego, CA, USA) were used to organize and analyze data. Analysis of variance (ANOVA) followed by Bonferroni post hoc test was used to evaluates variance among control and treated groups. p-value smaller than 0.05 was considered as statistically significant in the given study. To check the dependability of expression of AhFLS and AhLAR with other variables, Pearson correlation test were performed using GraphPad Prism® version 5.

Results

Sequence analysis of AhFLS and AhLAR

cDNA sequence of Flavonol synthase and Leucoanthocyanidin reductase from A. hypogaea designated as AhFLS (GenBank accession no. MT994314) and AhLAR (GenBank accession no. MH823866), respectively. The nucleotide sequence of AhFLS was 1343 bp that contained open reading frame (ORF) of 1004 bp encoding a predicted protein sequence of 334 amino acid. On the other hand, the nucleotide sequence of AhLAR was 1083 bp that contained ORF of 1056 bp encoding a predicted protein sequence of 351 amino acid. The calculated physiochemical parameters of predicted AhFLS and AhLAR protein in comparison to FLS and LAR protein from other plants have been listed in Table 1 and Table 2, respectively.

Table 1

Physiochemical properties of Flavonol synthase (FLS)from different plant species.

Gene	Protein accn.No.	Gene accn. No.	Aa	Mw	pI	II	AI	EC	GRAVY
A. hypogaea FLS	RYQ99263.1	XM_025812267.2	334	38.18	6.03	39.30	80.99	58,455	−0.543
A. duranensis FLS	XP_015931954.1	XM_016076468.2	334	38.31	5.82	41.97	81.29	58,455	−0.544
A. ipaensis FLS	XP_016167304.1	XM_016311818.2	334	38.18	6.03	39.30	80.99	58,455	−0.543
S. guianensis FLS	QHQ74395.1	MN165118.1	334	38.30	6.29	38.52	80.42	62,465	−0.539
G. max FLS	NP_001237419.1	NM_001250490.2	334	38.15	5.81	46.37	81.32	55,600	−0.513

Number of amino acids (aa), Molecular weight (Mw), Theoretical pI, Extinction coefficient (EC), Instability index (II), Aliphatic index (AI), Grand average of hydropathicity (GRAVY).

Table 2

Physiochemical properties of Leucoanthocyanidin reductase (LAR)from different plant species.

Gene	Protein accn.No.	Gene accn. No.	Aa	Mw	pI	II	AI	EC	GRAVY
A. hypogaea LAR		MH823866	351	38.72	5.94	34.26	89.72	21,930	−0.126
A. duranensis LAR	XP_015967593.1	XM_016112107.2	351	38.73	5.80	32.68	89.72	21,430	−0.119
A. ipaensis LAR	XP_016203013.1	XM_016347527.2	351	38.75	5.94	34.29	90.83	21,430	−0.113
L. uliginosus LAR	AAU45392.1	AY730617.1	348	38.34	7.07	27.40	92.47	26,400	−0.056
L. corniculatus LAR 1–2	ABC71325.1	DQ349101.1	349	38.56	8.08	31.25	90.50	26,400	−0.095
C. cajan LAR	XP_020233704.1		368	41.20	5.59	28.58	89.48	27,390	−0.213
G. max LAR	NP_001352050.1	NM_001365121.1	365	40.32	5.78	36.93	95.01	22,920	−0.049
V. angularis LAR	XP_017413424.1	XM_017557935.1	348	38.72	5.89	30.27	95.20	26,400	−0.054

Number of amino acids (aa), Molecular weight (Mw), Theoretical pI, Extinction coefficient (EC), Instability index (II), Aliphatic index (AI), Grand average of hydropathicity (GRAVY).

The analysis of amino acid composition indicated that predicted protein sequences of AhFLS contained high content of Val (9.6%), Glu (9%), Lys (8.7%), Leu (7.5%), Pro (7.5%), Gly (6.9%), Ser (5.7%) and Ile (5.4%), while predicted protein sequences of AhLAR contained high content of Ile (8.3%), Leu (7.7%), Asp (7.4%), Val (7.3%), Lys (6.8%), Ala (6.8%), Gly (6.6%), Ser (6.6%), and Thr (6.3%). The comparative amino acid composition analysis of LAR and FLS protein from various plants have been graphically presented in Fig. 1b and Fig. 2b. Secondary structure analysis revealed 76% α-helix, 45.2% sheets, 13.8% turns in predicted AhFLS protein and 67.8% α-helix, 61% sheets, 12.3% turns in predicted AhLAR as compared to FLS and LAR from other plant species, respectively (Fig. 1e and Fig. 2e).

Fig. 1

Phylogenetic analysis of AhFLS and secondary structure analysis of putative protein sequence of AhFLSa Phylogenetic tree based on maximum likelihood b Amino acid composition analysis of AhFLS in comparison with FLS from other plant species c, d Secondary structure analysis of AhFLSe Percentage of helixes, sheets and turns of FLS in different plants.

Fig. 2

Phylogenetic analysis of AhLAR and secondary structure analysis of putative protein sequence of AhLARa Phylogenetic tree based on maximum likelihood b Amino acid composition analysis of AhLAR in comparison with LAR from other plant species c, d Secondary structure analysis of AhLARe Percentage of helixes, sheets and turns of LAR in different plants.

AhFLS showed 100% similarity with the retrieved AhFLS (XM_025812267.2) sequence and AhLAR (MH823866.1) showed 100% similarity with the retrieved AhLAR (XM_025846016.1) sequence (Fig. S3 and Fig. S5). Protein-protein blast revealed that the predicted AhLAR protein sequence shared 72.63%, 73.35%, 73.07%, 73.64%, and 72.91% identity with LAR proteins from Cajanus cajan (XP_020233704.1), Glycine max LAR1(NP_001352050.1), Lotus corniculatus LAR 1 (ABC71325.1), Lotus uliginosus (AAU45392.1) and Vigna angularis (XP_017413424.1), respectively. It also showed identity with its wild parents Arachis ipaensis (XP_016167304.1, 99.43%) and Arachis duranensis (XP_015931954.1, 99.43%). Whereas, predicted AhFLS protein sequence shared 91.32%, 83.53%, 83.93%, 82.62%, 82% and 79% identity with FLS proteins from Stylosanthes guianensis (SgFLS, QHQ74395.1), Glycine max (GmFLS, NP_001237419.1), Clitoria ternatea (CtFLS, BAF49296.1), Onobrychis viciifolia (OvFLS, AEF14416.1), Medicago truncatula (MtFLS, XP_003614780.1). It also showed identity with its wild parents Arachis ipaensis (XP_016167304.1, 100%) and Arachis duranensis (XP_015931954.1, 99.7%) (Fig. S4 and Fig. S6).

Phylogenetic analysis

Phylogenetic analysis of AhFLS resolved two distinct groups with the closest relationship to its wild parents: A. ipaensis (Araip FLS) and A. duranensis (Aradu FLS) (Fig. 1a). It also showed homology with C. ternatea (CtFLS), G. max (GmFLS) and O. viciifolia (OvFLS). This analysis also predicts that AhFLS evolved later in history. The analysis of the phylogenetic tree revealed that AhLAR gene in A. hypogaea had an early origin in the evolutionary history as compared to similar genes in other plant species. The candidate gene is closely related to A. duranensis, A. ipaensis, Stylosanthes guianensis, Glycine max, Clitoria ternatea, Onobrychis viciifolia and Medicago truncatula (Fig. 2a).

3D protein model of AhFLS and AhLAR

Three-dimensional model of AhFLS and AhLAR were predicted through homology modeling using SWISS-MODEL Expasy (Fig. 3 and Fig. 4). AhLAR depicted the highest model homology with Leucoanthocyanidin reductase from Vitis vinifera:3i52.1 (67.67%). The oligostate of the predicted protein model was a monomer with resolution of 2.28 Å. The predicted model of AhLAR (PMDB ID: PM0083389) has QMEAN4 and QMEAN6 scores are −0.53 and −0.74, respectively. SAVES server verified 3D score to be more than 80%. The 3D-model showed 92.50% of residues have an average 3D-1D score of ≥ 0.2. Therefore, the model is acceptable. At least 80% amino acid scored ≥ 0.2. This protein model showed the ERRAT value of 95.6332 for both chains. Analysis of Ramachandran plot showed that 95.8% residues resided in the favored region whereas 3.8% of residues were in allowed regions and the remaining 0.4% residues were present in generously allowed or outer region (Fig. S12).

Fig. 3

Predictive three-dimensional structure of AhFLS and its assessment a Tertiary structure showing clear secondary arrangements and disulfides (Cyan Blue = Helix, Pink = Sheet, Grey = Loop) b Presumed catalytic sites with ligand AKG (2-Oxoglutaric Acid) and MES 2-(N-morpholino)ethanesulfonic acid) are indicated in blue and green respectively c Hydrophobic surface of the AhFLSd Detailed atomic structure of the AhFLS.

Fig. 4

Predictive three-dimensional structure of AhLAR and its assessment a Tertiary structure showing clear secondary arrangements and disulfides (Cyan Blue = Helix, Pink = Sheet, Grey = Loop) b Presumed catalytic sites with ligand NAP (nicotinamide-adenine-dinucleotide phosphate) and KXN (2R,3S)-2-(3,4-Dihydroxyphenyl)-3,4-dihydro-2H-chromene-3,5,7-triol; tetrahydrate are indicated in red and yellow, respectively C Hydrophobic surface of the AhLARd Detailed atomic structure of the AhLAR.

Whereas AhFLS showed the highest model homology with Flavonol synthase from Arabidopsis thaliana:1gp4 (40.61%). The oligostate of the predicted protein model was a monomer with the resolution of the structure was 2.28 Å. The predicted model of AhFLS (PMDB ID: PM0083390) has QMEAN4 score of −1.08 and 3D score to be more than 80%. The 3D-model showed 87.31% of residues have an average 3D-1D score of ≥ 0.2. Therefore, the model is acceptable. At least 80% amino acid scored ≥ 0.2. This protein model showed the ERRAT value of 87.1473 for both chains. Analysis of Ramachandran plot showed that 93.4% residues resided in the favored region whereas 5% of residues were in allowed regions and the remaining 0.6% residues were present in generously allowed or outer region (Fig. S11).

Functional analysis of AhFLS and AhLAR

A similar pattern was observed in both protein and nucleotide sequences after multiple sequence alignment. CDD search revealed that the deduced AhFLS protein has putative conserved domains belonging to the 2OG-Fe(II) oxygenase superfamily domain (pfam03171, 200–295), a highly conserved N-terminal region of proteins with 2-oxoglutarate/Fe(II.)-dependent dioxygenase activity and the AhFLS domain (PLN02704, 2-191) (Fig. S1). However, NADB_Rossmann Superfamily domain has been identified for AhLAR (Fig. S2).

The active site amino acid residues of AhLAR were predicted to be Cys and Asp while as that of AhFLS were Leu, Lys, Cys, Glu, Gln, Ile and His as determined by COFACTOR. Due to the unavailability of any direct method for identifying the active site amino acids of AhLAR and AhFLS, the sequence information of closest homology putative leucoanthocyanidin reductase 1 from Vitis vinifera (PDB ID: 3I52.1) and Anthocyanidin synthase from Arabidopsis thaliana (PDB ID: 1gp4.1) were used for identifying active site amino acid, respectively. Then, sequence alignment analysis was performed for aligning AhLAR with 2gasA and AhFLS with 1gp6A. Six functional motifs were detected from the functional study in case of AhLAR, which were found to be a member of NmrA-like family, NAD-dependent dehydratase family and 3, β hydroxysteroid dehydrogenase family (Fig. S8). On the other hand, two functional motifs 2OG-FeII_Oxy and DIOX_N were detected from the functional study in case of AhFLS, which were found to be a member of 2OG-Fe(II) oxygenase superfamily (Fig. S7).

Promoter sequence analysis

Cis-acting regulatory elements predicted using plantCARE database showed that both gene promoters contained typical eukaryotic regulatory elements including CAAT and TATA, Table 3 and Table 4. Furthermore, several other stress related promoter elements were identified in the promoter sequences of AhFLS and AhLAR which were classified into four groups based on the cellular mechanisms that modulate the development regulation, environmental stress responses, light-mediated responses and hormonal signaling. MYB binding site and hormonal responsive cis-regulatory element were also observed in the promoter regions of AhFLS and AhLAR.

Table 3

Putative Cis regulatory elements in the promoter region of AhFLS.

Class	Cis Regulatory Element Name	Position	Core Sequence	Putative Function
Environmental stress responsive	5UTR Py-rich stretch			Involved in Biotic Stress
	ARE	496,1461	AAACCA, AAACCA	Involved in anaerobic induction
Hormone signalling	TCA-element	1195	CCATCTTTTT	Involved in the salicylic acid responsiveness
	TGACG-motif	1092	TGACG	Involved in the MeJA-responsiveness

Table 4

Putative Cis regulatory elements in the promoter region of AhLAR.

Class	Cis Regulatory Element Name	Position	Core Sequence	Putative Function
Development responsive	RY-element	604	CATGCATG	Cis acting regulatory element involved in seed-specific regulation
Light Responsive	G box	628	CACGTC	Light responsiveness
	AE-box	1391	AGAAACAA	Light responsiveness
	Box 4	254, 267	ATTAAT, ATTAAT	Light responsiveness
Hormone signalling	CGTCA-motif	630	CGTCA	MeJA-responsiveness
	TATC-box	531	TATCCCA	Gibberellin-responsiveness
MYB binding site	CCAAT-box	1396	CAACGG	MYBHv1 binding site

Relative AhFLS and AhLAR expression in response to salinity

Expression of AhFLS and AhLAR against salinity stress has been evaluated to confirm their role in abiotic stress in peanut. Three distinct cultivars of A. hypogaea (Bari 2011, BARD 92 and Golden) were selected for analysis at various level of stress. Both AhFLS and AhLAR depicted differential expression against various concentration of salinity stress and confirmed their role in response to salinity (Fig. 5). Highest upregulation of AhFLS (36.7 fold increase) has been observed in Golden at highest concentration of salt (240 mM) while highest upregulation of AhLAR (3.5 fold increase) has been observed in Bari 2011 at 60 mM and 240 mM concentration of salt. AhLAR has been found to be positively regulated at lower concentrations of salt while negatively regulated at higher concentration of salt. However, AhFLS expression was quite contrary in different variety. In Golden, the gradual increase of AhFLS expression has been observed with the increase in salt concentration. However, Bari 2011 and BARD 92 showed highest upregulation of AhFLS at 60 mM.

Fig. 5

Expression analysis of aAhFLS (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = < 0.0001, Row factor p-value = < 0.0001) and bAhLAR (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = < 0.0001, Row factor p-value = < 0.0001) in three selected varieties of A. hypogaea against salinity stress. Here, control represents internal comparison of individual cultivar.

Accumulation of flavonol and proanthocyanins in leaves

Leaf synchronous fluorescence spectra showed eight prominent fluorescent regions; low intensity bands at 330–350 nm, 378 nm, 579–605 nm, high intensity regions 390–425 nm, 681 nm, 725 nm and strong fluorescent regions between 449 and 466 nm, 510–552 nm with a sharp peak at 526 nm (Fig. 6). Fluorescent region between 330 and 350 nm has already been reported with the presence of catechins (Holser, 2014, Palet al., 2012), approximately 350 nm emission maxima for proanthocyanidin, fluorescence emission with maxima at 378 nm with apigenin (Parket al., 2013, Tuet al., 2016), regions with fluorescence emission in the range of 390–425 nm and 449–466 nm represent both non-acylated flavonol glucosides that are present in plant cuticle (Donaldson and Williams, 2018); 510–552 nm for flavonol, flavones (Guharayet al., 2001, Voicescuet al., 2014) 395 nm for Reversetrol (Figueiras et al., 2011), 522 nm for curcumin, 440 nm polyphenolics (Singh and Mishra, 2015) and 334 nm tocopherol (Demirkaya-Miloglu et al., 2013) and 579–605 nm for anthocyanin (Drabent et al., 1999). Chlorophyll b and a emitted fluorescence at 681 nm and 725 nm (Weber and Teale, 1957). The fluorescent region between 360 and 560 nm has been studied due to their association with flavonol peaks including kaempferol (λem ≈ 463 nm), quercetin (λem ≈ 535 nm), rutin (λem ≈ 545 nm), morin (λem ≈ 490 nm) and non-acylated flavonol glucosides (Condatet al., 2016, Donaldson and Williams, 2018, Höfeneret al., 2013, Tuet al., 2016, Xuet al., 2010). Higher fluorescence intensity was observed from 440 to 560 nm in salt treated groups as compared to the control in all three cultivars, which showed accumulation of flavonol during salt stress. The appearance of a fluorescence peak at 345 nm that represents catechins and tocopherol was only observed in Golden salt stressed plants at all salt concentrations whereas, that peak was only observed in Bari 2011 at 240 mM salt concentration.

Fig. 6

Synchronous fluorescence spectra of leaves of three cultivar of peanut (Bari 2011, BARD 92 and Golden) showing spectral deviation at different salt concentrations with excitation range: 200–700 nm, offset: 60 a Fluorescence spectra of Bari 2011 leaves b Spectral deviation in Bari 2011 leaves in the region of 330 to 430 nm c Spectral deviation in the leaves of Bari 2011 in region of 435 to 550 nm d Fluorescence spectra of BARD 92 leaves e Spectral deviation in the leaves of BARD 92 in region of 400 to 460 nm f Spectral deviation in the leaves of BARD 92 in region of 510 to 610 nm g Fluorescence spectra of Golden leaves h Spectral deviation in the leaves of Golden in region of 330 to 430 nm i Spectral deviation in the leaves of Golden in region of 430 to 550 nm.

Accumulation of flavonol and proanthocyanins in roots

The excitation of tap root at 380 nm yielded emission fluorescence spectra at a low intensity region (408–418 nm), two high intensity regions (435–443 nm and 450–480 nm) and a moderate intensity region (505–522 nm) (Fig. 7). Generally, all these fluorescent regions have been reported as emission spectra related to flavonol, chalcones, flavones and isoflavnoids (Asiriet al., 2014, De Rijkeet al., 2002, Guharayet al., 2001, Voicescuet al., 2014). In Bari 2011, significant spectral variation has been observed in the region of 418 to 443 nm region that generally represents polyphenolics (λem ≈ 440 nm) (Singh and Mishra, 2015), 390–425 nm non-acylated flavonol glucosides (Donaldson and Williams, 2018). In BARD 92, spectra representing polyphenolics, non-acylated flavonol glucosides, trans-resveratrol (λem ≈ 395 nm) (Figueiras et al., 2011) (Fig. 6e). In Golden, region representing trans-resveratrol appeared in the control and 120 mM treated plants while significant spectral variation recorded in between 510 and 545 nm region that represents flavonols (λem ≈ 535 nm) and curcumin (λem ≈ 522 nm). Generally, all the three cultivars showed statistically significant spectral deviation as compared to their respective controls.

Fig. 7

Fluorescence spectra of tap root of three cultivar of peanut (Bari 2011, BARD 92 and Golden) showing spectral deviation at different salt concentrations after excitation at 380 nm (emission spectra range: 395–550 nm) a Fluorescence spectra of Bari 2011 tap roots b Spectral deviation in the tap roots of Bari 2011 in region of 395 to 420 nm c Spectral deviation in the tap roots of Bari 2011 in region of 425 to 450 nm d Fluorescence spectra of BARD 92 tap roots e Spectral deviation in the tap roots of BARD 92 in region of 490 to 440 nm f Spectral deviation in the tap roots of BARD 92 in region of 475 to 550 nm g Fluorescence spectra of Golden roots h Spectral deviation in the tap roots of Golden in region of 395 to 410 nm i Spectral deviation in the tap roots of Golden in region of 490 to 550 nm.

Correlation of AhFLS and AhLAR expression with biochemical and growth parameters

All three peanut cultivars were grown under salinity stress (0 mM, 60 mM, 120 mM, and 240 mM) resulting in different seedling lengths, fresh weight and moisture content as compared to their respective control plants after 15 days of salt treatment (Fig. 8). Negative relationship has been observed in the expression of AhFLS and AhLAR with the root length. On contrary, positive relationship has been observed in the expression of AhFLS and AhLAR with the shoot length except AhFLS in Golden showed negative correlation with shoot length (r = -0.9209). Positive relationship has been observed in the expression of AhFLS and AhLAR with root fresh weight, shoot fresh weight and moisture content in BARD 92. However, Golden and BARD 92 showed negative relationship in the expression of AhFLS and AhLAR with root fresh weight, shoot fresh weight and moisture content, respectively (Table 5).

Fig. 8

Effects of salinity on growth and moisture content of three Pakistani cultivars of peanut (Bari 2011, BARD 92 and Golden) a Symptoms of salinity (wilting and browning) and gradual decease in overall growth of all three cultivars b Gradual decrease in moisture content with respect to increasing salt concentration (Two-way ANOVA: Interaction p-value = 0.6696, Column factor p-value = 0.0107, Row factor p-value = < 0.0001) c Relative shoot length of peanut cultivars under salinity (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = < 0.0001, Row factor p-value = < 0.0001) d Shoot fresh weight under salinity (Two-way ANOVA: Interaction p-value = 0.1196, Column factor p-value = 0.0010, Row factor p-value = < 0.0001) e Relative root length of peanut cultivars under salinity (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = 0.0006, Row factor p-value = < 0.0001) f Shoot fresh weight under salinity (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = 0.0001, Row factor p-value = < 0.0001) . Data presented as mean ± S.E. p-value smaller than 0.05 is considered significant. ‘*’ is depicting p-value<0.05, ‘**’ p-value<0.01, ‘***’ p-value<0.001.

Table 5

Correlation of AhFLS and AhLAR expression with antioxidant enzymes activity, biochemical and growth parameters.

Cultivars	FLS vs. RL		FLS vs. SL		FLS vs. RFW		FLS vs. SFW		FLS vs. MC		FLS vs. SC
	r	p-value	r	p-value	r	p-value	r	p-value	r	p-value	r	p-value
Bari 2011	0.1582	0.8418	0.6316	0.3684	0.1897	0.8103	0.3862	0.6138	0.3997	0.6003	−0.4181	0.5819
BARD 92	−0.5957	0.4043	0.4651	0.5349	0.6258	0.3742	0.6570	0.3430	0.4574	0.5426	0.4713	0.5287
Golden	−0.1197	0.8803	−0.9209	0.0791	−0.08499	0.9150	−0.7582	0.2418	−0.9345	0.0655	0.8276	0.1724
	FLS vs. SP		FLS vs. APX		FLS vs. CAT		FLS vs. SOD		FLS vs. Chlorophyll		FLS vs. LAR
	r	p-value	r	p-value	r	p-value	r	p-value	r	p-value	r	p-value
Bari 2011	−0.1324	0.8676	−0.3660	0.6340	0.6219	0.3781	−0.1154	0.8846	−0.3840	0.6160	0.6344	0.3656
BARD 92	0.1588	0.8412	−0.2759	0.7241	0.6422	0.3578	0.2347	0.7653	0.05184	0.9482	0.8802	0.1198
Golden	−0.4449	0.5551	0.8211	0.1789	0.9000	0.1000	−0.3991	0.6009	−0.8769	0.1231	−0.7155	0.2845
	LAR vs. RL		LAR vs. SL		LAR vs. RFW		LAR vs. SFW		LAR vs. MC		LAR vs. SC
	r	p-value	r	p-value	r	p-value	r	p-value	r	p-value	r	p-value
Bari 2011	−0.4761	0.5239	−0.1886	0.8114	−0.6386	0.3614	−0.4658	0.5342	−0.4549	0.5451	0.3763	0.6237
BARD 92	−0.1563	0.8437	0.8056	0.1944	0.9145	0.0855	0.8814	0.1186	0.8239	0.1761	0.3316	0.6684
Golden	−0.1756	0.8244	0.3903	0.6097	0.3465	0.6535	0.1056	0.8944	0.4254	0.5746	−0.2220	0.7780
	LAR vs. SP		LAR vs APX		LAR vs. CAT		LAR vs. SOD		LAR vs. Chlorophyll
	r	p-value	r	p-value	r	p-value	r	p-value	r	p-value
Bari 2011	−0.3983	0.6017	0.1053	0.8947	0.4556	0.5444	−0.3351	0.6649	−0.4637	0.5363
BARD 92	0.4899	0.5101	−0.6278	0.3722	0.5709	0.4291	0.2219	0.7781	0.2273	0.7727
Golden	0.7767	0.2233	−0.1891	0.8109	−0.3790	0.6210	0.6965	0.3035	0.2920	0.7080

FLS = Flavonol synthase; LAR = Leucoanthocyanidin reductase; RL = Root Length; SL = Shoot Length; RFW = Root Fresh Weight; SFW = Shoot Fresh Weight; MC = Moisture Content; SC = Sugar Contents; SP = Soluble Protein; APX = Ascorbate peroxidase; CAT = Catalase; SOD = Superoxide dismutase.

r = Pearson r that is showing correlation between two variables.

p-value less than 0.05 is considered to be statistically significant.

In response to salinity stress, significant differential accumulation of soluble protein, sugar and chlorophyll has been observed within selected peanut cultivars (Fig. 9 and Fig. 10). Correlation studies has been conducted to evaluate the dependability of AhFLS and AhLAR expression on soluble protein, sugar, and chlorophyll content in plant cells. High level of negative correlation has been observed between chlorophyll content and AhFLS expression in Golden (r = −0.8769) while BARD 92 showed no relationship between AhFLS and chlorophyll content. Fair degree of negative correlation has been observed between chlorophyll content and AhLAR expression in Bari 2011 (r = −0.4637). The expression of AhFLS was found to dependable on soluble sugar than soluble protein while the expression of AhLAR was observed to dependable on soluble protein than soluble sugar (Table 5).

Fig. 9

Salinity affecting accumulation of soluble protein and sugar in three cultivars of peanut (Bari 2011, BARD 92 and Golden) a Enhanced accumulation of soluble sugar in all the three verities of peanut (Two-way ANOVA: Interaction p-value = 0.0805, Column factor p-value = 0.0004, Row factor p-value = 0.0010) b Salinity induced accumulation of soluble protein in Golden (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = 0.0022, Row factor p-value = < 0.0001). Data presented as mean ± S.E. p-value smaller than 0.05 is considered significant. ‘**’ is depicting p-value<0.01, ‘***’ p-value<0.001.

Fig. 10

Synchronous fluorescence spectra and chlorophyll analysis in three cultivars of peanut (Bari 2011, BARD 92 and Golden) a Fluorescence spectrum of Bari 2011b Fluorescence spectrum of BARD 92c Fluorescence spectrum of Golden d Effect of salinity on chlorophyll content of peanut seedlings through biochemical analysis (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = 0.0022, Row factor p-value = < 0.0001). Each spectral line represents mean of replicates. Data of chlorophyll content presented as mean ± S.E. p-value smaller than 0.05 is considered significant. ‘**’ is depicting p-value<0.01, ‘***’ p-value<0.001.

Correlation of AhFLS and AhLAR expression with antioxidant enzymes activity

Expression of APX, SOD, CAT, AhFLS and AhLAR have been observed to be influenced by salinity stress (Fig. 5 and Fig. 11). To observe association of AhFLS and AhLAR expression with the activity of antioxidant enzymes, Pearson’s correlation was performed and revealed AhFLS has little or no relationship with SOD activity in Bari 2011 and BARD 92 (r = -0.1154 and r = 0.2347), fair relationship with SOD activity in Bari 2011 (r = −0.3991) and APX activity in Bari 2011 and BARD 92 (r = −0.3660 and r = −0.2759). Moderate relationship of AhFLS has been observed with CAT activity in Bari 2011 and BARD 92 (r = 0.6219 and r = 0.6422) while strong relationship of AhFLS has been observed with APX and CAT activity in Golden (r = 0.8211 and r = 0.9000). Moderate association of AhLAR has been observed with APX activity (r = −0.6278), CAT activity (r = 0.5709) in BARD 92 and SOD activity in Golden (r = 0.6965). While fair or little association of AhLAR has been with APX and CAT activity in Bari 2011 and Golden (Table 5).

Fig. 11

Activity of enzymes involved in the antioxidation in three cultivars of peanut (Bari 2011, BARD 92 and Golden) a Effect of salinity on Ascorbate peroxidase (APX) activity of peanut seedlings (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = < 0.0001, Row factor p-value = < 0.0001) b Effect of salinity on Catalase (CAT) activity of peanut seedlings (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = < 0.0001, Row factor p-value = < 0.0001) c Effect of salinity on Superoxide dismutase (SOD) activity of peanut seedlings (Two-way ANOVA: Interaction p-value = < 0.0001, Column factor p-value = < 0.0001, Row factor p-value = < 0.0001). Data presented as mean ± S.E. p-value smaller than 0.05 is considered significant. ‘*’ is depicting p-value<0.05, ‘**’ p-value<0.01, ‘***’ p-value<0.001.

Discussion

FLS and LAR belong to two distinct branches of flavonoid biosynthetic pathway and lead to the production of flavonols and proanthocyanidins, respectively (Nabavi et al., 2018; Zhang et al., 2020). In the present study, FLS and LAR isolated from Arachis hypogaea have been structurally, transcriptionally, and functionally characterized. Furthermore, the oxidative stress generated by salinity has been used to evaluate the protective role AhFLS and AhLAR against stress in three selected cultivars of A. hypogaea (Bari 2011, BARD92 and Golden). cDNA sequence of AhFLS (MT994314) and AhLAR (MH823866.1) showed conservation of enzymes’ sequence and functionality within varieties of A. hypogaea as 100% similarity has been observed in these enzymes with already published AhFLS (XM_025812267.2) and AhLAR (XM_025846016.1) sequence, respectively. The role conservation of AhFLS and AhLAR has been further analyzed within species through multiple sequence alignment and phylogenetic analysis which revealed that both genes are closely related to A. duranensis, A. ipaensis, Stylosanthes guianensis, Glycine max, Clitoria ternatea, Onobrychis viciifolia and Medicago truncatula. Phylogenetic study also showed that AhFLS evolved later in history while AhLAR had an early origin in peanut.

AhFLS and AhLAR showed highest model homology with Flavonol synthase from Arabidopsis thaliana:1gp4 and Leucoanthocyanidin reductase from Vitis vinifera:3i52.1. Based on sequence and structural similarities, functional analysis of AhFLS and AhLAR revealed that AhFLS has two functional motifs 2OG-FeII_Oxy and DIOX_N. The presence of these two functional domains indicated that AhFLS has part in the biosynthesis of flavonols in peanut plant (Hou et al., 2020). Whereas, AhLAR has six functional motifs that are related to NmrA-like family, NAD-dependent dehydratase family and 3, β hydroxysteroid dehydrogenase family that are crucial for the catalytic activity of LAR and consequent production of proanthocyanidins (Maugé et al., 2010). Cys and Asp are predicted as active site amino acid residues of AhLAR and Leu, Lys, Cys, Glu, Gln, Ile and His as that of AhFLS. Presence of stress related promoter elements in the promoter sequences of AhFLS and AhLAR their role in the regulation growth, development and stress responses.

Flavonols and proanthocyanidins have imperative role in the defense response against both biotic and abiotic stress in plants due to their free radical scavenging and antioxidant activity (Andersen and Markham, 2005, Muhlemannet al., 2018). The expression analysis of AhFLS and AhLAR along with the accumulation pattern of flavonols and proanthocyanidins in A. hypogaea indicated their role against salinity stress. Moreover, their expression and accumulation patterns were found to be genotype and stress-level dependent. Hence, indicating that their regulation is multivariate dependent. Higher differential accumulation in the roots rather than in the leaves of A. hypogaea confirmed that they specifically accumulate in the roots of A. hypogaea. Many flavonoids are found to be accumulated specifically in the root regions such as nodule primordia and lateral roots of the clover (Trifolium subterraneum) (Djordjevicet al., 1997, Morris and Djordjevic, 2006) and in root primordia and root tip of A. thaliana (Bueret al., 2007, Peeret al., 2001).

Correlation of AhFLS and AhLAR expression with antioxidant enzymes activity, biochemical and growth parameters revealed negative, positive or no relation, but no statistically significant p-value has been observed for any correlation which showed that AhFLS and AhLAR expression is not directly dependable on these selected parameters, but higher Pearson r value depicted some level of dependency. Differential expression of AhFLS and AhLAR along with accumulation pattern of flavonols and proanthocyanidins in all three selected cultivars at various salt concentrations indicated vital role of flavonols and proanthocyanidins in salt tolerance. Such differential expression indicated the presence of precise regularity molecular mechanisms that are yet to explored.

Conclusion

Over the course of time, plants have adapted themselves to the flatulating environmental conditions by engineering their metabolic pathways. Detailed understanding of metabolic enzymes can assist in the metabolic engineering of such pathways to produce stress tolerant varieties and production of desirable metabolites. Flavonols and proanthocyanidins are specific targets for metabolic engineering owing to their significant role in plant stress response and biological activities. The characterization of AhFLS and AhLAR given in this study could be a potential source for understanding the structure, function and regulation of these enzymes under stress. Furthermore, the present study would assist in the metabolic engineering of flavonoid biosynthetic pathway to produce stress tolerant varieties and production of proanthocyanidins and flavonols at an industrial scale.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data and materials

The data and results of this study are represented in the enclosed table, figures and supplementary materials.

Funding

This research was supported by the Higher Education Commission (HEC), Pakistan and National University of Sciences and Technology (NUST), Pakistan.

Authors' contributions

RA, GK and SH conceived the project. All co-authors were involved in planning the experiments. GK, SH, JM, TI, and MK conducted the experiments and collected the data. GK, SH, HA, MK, FM and RA performed data analysis and drafting of the manuscript. RA, FM, AG and HA revised the final manuscript.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.