Arabidopsis

Roepke, Jonathon & Bozzo, Gale G., 2015, Arabidopsis thaliana β-glucosidase BGLU 15 attacks flavonol 3 - O-β-glucoside- 7 - O- - rhamnosides, Phytochemistry 109, pp. 14-24 : 15-20

2.1. Evidence for flavonol 3-O- β -glucoside-7-O- α -rhamnoside hydrolysis in Arabidopsis during abiotic stress recovery

Pioneering research by Olsen et al. (2009) showed dramatic and rapid losses in flavonol bisglycosides occur in Arabidopsis during the recovery of rosette stage plants from synergistic abiotic stress. However, the biochemical mechanism driving these losses is unclear. Here, older plants (at 29 days after planting) relative to the aforementioned study were subjected to 0 mM nitrate at 10 ° C (nitrogen deficiency and low temperature, NDLT) for 7 days. Thereafter, NDLT plants were fertilized with 10 mM nitrate and transferred to 21 ° C (nitrogen sufficiency and high temperature, NSHT). Acidified methanol extracts of whole plants were profiled for flavonol bisglycosides levels, specifically K 3 G 7 R (1) and Q 3 G 7 R (2). K 3 G 7 R (1) and Q 3 G 7 R (2) concentrations were augmented 2.4 and 29.3-fold, respectively, by the end of the NDLT treatments ( Fig. 2A ). Within 2 days of the transfer of NDLT-treated plants to NSHT, K 3 G 7 R (1) and Q 3 G 7 R (2) levels were reduced by 33% and 56%, respectively. After an additional 3 days of NDLT recovery, concentrations of these bisglycosides declined to levels available prior to application of synergistic abiotic stress. Similar changes in the concentrations of putative flavonol 3- O -α- rhamnoside-7- O -α- rhamnosides, kaempferol and quercetin 3- O -β- rutinoside-7- O -α- rhamnosides with respective molecular masses of 740.3 and 756.3 were apparent (Supplementary Fig. 1 ). Notably, peaks identified as flavonol 3- O -β- rutinoside-7- O -α-rhamnosides are consistent with nomenclature devised by Olsen et al. (2009), although the possibility remains for their identification as flavonol 3- O -β- neohesperidoside-7- O -α- rhamnosides ( Saito et al., 2013; Yonekura-Sakakibara et al., 2008) since UHPLC-DAD-MS n technology yields identical ion fragments for either conjugate.

Kim et al. (2014) established that minor degradation of kaempferol bisglycosides occurs during the development of the rosette habit stage in Arabidopsis when cultivated under non-stressed conditions. No such phenomena were observed here, as flavonol bisglycoside levels remained low and stable under constant NSHT (control; Fig. 2B ). On the whole, the findings herein confirm the degradative phenomena described by Olsen et al. (2009). More importantly, disappearance of K 3 G 7 R (1) and Q 3 G 7 R (2) coincided with a 244% transient increase in Q 3 G 7 R (2) hydrolase activity within 2 days of NDLT recovery ( Fig. 2A ). Within 3 days of the transfer of NDLT-acclimated plants to NSHT, average Q 3 G 7 R (2) hydrolysis activity approximated 39 pmol mg protein — 1 min — 1 or 56 nmol per day; these activities are in excess of the apparent reduction in Q 3 G 7 R (2; 80 nmol) pools during this period (based on assumption 1 g of fresh matter yields 1 mg of extractable protein). Hydrolytic activity was relatively stable during NDLT and in NSHT control plants, and never higher than 22 pmol mg protein — 1 min — 1 under the precise assays conditions outlined here ( Fig. 2B ). HPLC analysis of the in vitro Q 3 G 7 R (2) hydrolysis indicated a product eluting at a retention time of 13.3 min, which was not evident in assays performed without the Arabidopsis cellfree enzyme preparation ( Fig. 3A ). No quercetin 3- O -β- glucoside was produced in these in vitro reactions suggesting the absence of α- rhamnosidase activity under the aforementioned assay conditions. Fungal-derived α- rhamnosidases hydrolyse flavonoids with α- linked sugars, like naringin ( Manzanares et al., 2007) but these activities are relatively unknown in plants.

Quadrupole TOF-MS/ MS analysis of the Q 3 G 7 R (2) hydrolysate identified a molecular mass of 448.2, which is 162 mass units less than that of Q 3 G 7 R (2) ( Roepke and Bozzo, 2013) indicating the loss of a non-reducing terminal glucose moiety; collision-induced dissociation spectra of this parent ion yielded an [ M — H] — fragment of 301.1, which corresponded to the loss of a rhamnose moiety ( Fig. 3B ). Together, these data indicate quercetin 7- O -α- rhamnose (4) was released from Q 3 G 7 R (2) in vitro in the presence of an enzyme preparation from abiotic stress recovering Arabidopsis . Quercetin 7- O -α- rhamnose (4) is a rare natural product, occurring in the herbal medicine plant Hypericum japonicum and most recently detected in field-grown Arabidopsis ( Ishiguro et al., 1991; Nakabayashi et al., 2009). Quercetin 7- O -α- rhamnose (4) is a bona fide catabolite as there is no evidence for its direct formation from quercetin in planta. The only known flavonol 7- O -rhamnosyltransferase, AtUGT89 C 1, displays no preference for quercetin; moreover, flavonols are attacked primarily by flavonol 3- O -glycosyltranferases yielding glycoside intermediates, like Q 3 G and quercetin 3- O -α- rhamnoside ( Jones et al., 2003; Toghe et al., 2005; Yonekura-Sakaibara et al., 2007). These findings prompted investigations to identify the gene and the enzyme governing flavonol 3- O -β- glucoside-7- O -α- rhamnoside hydrolysis in Arabidopsis .

2.2. Phylogenetic and transcript analyses of Arabidopsis BGLUs

As these data implicated the involvement of BGLU activity in the hydrolysis of Q3G7R, the Arabidopsis genome was queried for BGLUs with similarity to flavonoid-utilizing enzymes. An updated and comprehensive phylogenetic comparison of glycoside hydrolase family 1 translated sequences (including biochemically characterized enzymes) using a maximum likelihood approach confirmed a clade consisting of BGLU12 (At 5g 42260), BGLU13 (At 5g 44640), BGLU14 (At 2g 25630), BGLU15 (At 2g 44450), BGLU16 (At 3g 60130), and BGLU17 (At 2g 44480) clustered with hydrolases exhibiting a pronounced specificity for flavonoids, including a Thai rosewood seed dalcochinase, an enzyme specific for isoflavonoid conjugates, chemicals that are structurally somewhat similar to flavonols ( Fig. 4 ). Although most family 1 glycoside hydrolases target monoglucosides or disaccharides, an Arabidopsis enzyme with catalytic preference for the 3-hydroxy conjugated glucose of flavonol 3- O -β- glucoside-7- O -α- rhamnosides seems likely as evidence exists for a Costus speciosus BGLU attacking bisglycosides, like furostanol glycosides, yielding spirostanol glycosides ( Inoue and Ebizuka, 1996). The current phylogenetic analysis confirmed the Arabidopsis clade was highly related to the C. speciosus furostanol (bis)glycoside 26- O -β- glucosidase, a pattern consistent with a previous phylogenetic analysis of Arabidopsis glycoside hydrolase family 1 enzymes ( Xu et al., 2004). Moreover, of all 47 Arabidopsis BGLUs, BGLU12, -13, -14, -15, -16, and -17 were the most proximal to recently characterized isoflavone BGLUs from soybean and barrel clover ( Naoumkina et al., 2007; Suzuki et al., 2006); these putative BGLUs were phylogenetically distant from lineages consisting of BGLUs specific for monolignol glucosides, abscisic acid conjugates, glucosinolates, alkaloids and scopolin.

BGLU transcript levels during NDLT recovery relative to control plants were assessed using a RT-qPCR approach. Gene expression analysis identified a 300% greater transient peak in BGLU15 expression at 1 day after the transfer of NDLT-treatments to NSHT relative to control plants ( Fig. 5 ). Similarly, simultaneous upregulation of BGLU12 and BGLU16 was apparent in response to NDLT recovery, although respective transcript levels were 148% and 68% higher than control plants. For all aforementioned BGLUs, transcript abundance remained unchanged during the 7 day NDLT period relative to control plants. BGLU17 expression was minimally affected by NDLT or recovery therefrom; no BGLU13 and - 14 transcripts were detected although primers were able to amplify respective fragments from genomic DNA. It is noteworthy that the BGLU15 pattern of induction, albeit 1 day in advance, coincides with the pattern of in vitro Q3G7R BGLU activity in response to NDLT recovery; in addition, this transient accumulation of BGLU15 transcripts may coincide with the 90–99.9% depletion in flavonol biosynthesis transcripts described by Olsen et al. (2009). As negative controls, the expression pattern of biochemically-characterized BGLU37 and BGLU45 was monitored, as both are phylogenetically distinct from flavonoid hydrolases. BGLU45 is a monolignol glucoside-specific hydrolase ( Escamilla-Treviño et al., 2006). Interestingly, an approximate 180% increase in levels of BGLU45 transcripts was evident by day 7 of the NDLT period; monolignol glucoside hydrolysis is postulated to be a response to abiotic stress ( Wang et al., 2013). Transcript levels of BGLU37, a myrosinase, and BGLU45 were not enhanced by the 5 day NDLT recovery period ( Fig. 5 ). This is not unexpected as myrosinases cleave thioglycosides (e.g., glucosinolates) and not β- O -linked substrates ( Andersson et al., 2009); various mutants of BGLU45 display altered coniferin content with only minor changes in other phenylpropanoid pathway derivatives ( Chapelle et al., 2012). Moreover, both are confined to specialized cell or tissue boundaries not known to participate in flavonol metabolism ( Andréasson et al., 2001; Chapelle et al., 2012), thus it is unlikely that either is involved in flavonol 3- O -β- glucoside-7- O -α- rhamnoside degradation.

2.3. Recombinant BGLU15 displays preference for flavonol 3-O- β - glucoside-7-O- α -rhamnosides

Focus was thus placed upon biochemical characterization of BGLU15, as it was the most highly expressed BGLU in response to NDLT recovery according to the RT-qPCR approach. A diverse array of biochemically characterized BGLUs have been expressed in Escherichia coli including Rauvolfia serpentia alkaloid BGLUs, an aroma yielding β- primeverosidase from tea, a Pinus contorta coniferin utilizing enzyme, and most importantly the soybean isoflavone-conjugate hydrolyzing enzyme ( Dharmawardhana et al., 1999; Gerasimenko et al., 2002; Mizutani et al., 2002; Nomura et al., 2008; Suzuki et al., 2006; Warzecha et al., 2000). Recombinant BGLU15 was expressed in E. coli Origami 2 (DE3) cells with N-terminal thiroedoxin-His 6 tags and the soluble protein was purified by immobilized metal affinity chromatography (IMAC). SDS–PAGE and immunoblots probed with a His 6 -antibody indicated the final recombinant BGLU15 was purified to apparent homogeneity and contained a single major band containing a His 6 -tag, identical to the predicted molecular mass of 72.6 kDa ( Fig. 6 ). A final yield of 453 ± 27 µg (mean ± SE of three separate enzyme preparations) of purified thioredoxin-His 6 -BGLU15 was attained per liter of bacterial culture. Rice and soybean BGLUs fused to a thioredoxin molecule are highly active ( Opassiri et al., 2006; Suzuki et al., 2006). Although a recombinant rice salicylic acid β- D- glucoside BGLU fused to thioredoxin displays reduced hydrolytic activity relative to the native protein, both used similar substrates ( Himeno et al., 2013). Tag removal with an endoprotease can yield compromised activity due to long incubation periods and non-specific cleavage ( Arnau et al., 2006), thus thioredoxin was not cleaved from the recombinant BGLU15 preparation prior to its use in biochemical assays.

A comparison of hydrolysis activity across a broad pH range determined that recombinant BGLU15 had optimal activity at pH 5 (Supplementary Fig. 2 ). The pH optimum is typical of other plant BGLUs, including scopolin hydrolases, a dalcochinase from Dalbergia cochinchinensis and a soybean isoflavone conjugate hydrolase ( Ahn et al., 2010; Srisomsap et al., 1996; Suzuki et al., 2006); acidic pH optimum of the latter fits with its localization to the apoplast. In fact, proteome analyses of suspension cell cultures identified BGLU15 as a loosely bound cell wall protein ( Borderies et al., 2003), suggesting it too may be confined to the apoplast. All remaining biochemical assays were performed at pH 5.

In order to identify aglycone specificity of BGLU15, comparison of in vitro activities of the recombinant enzyme towards various natural plant derived compounds was carried out as indicated in Fig. 7 . Discontinuous in vitro assays at a fixed substrate concentration (500 µM) demonstrated recombinant BGLU15 hydrolyzed K3G7R (1) and Q3G7R (2) with high rates relative to all other substrates tested, followed by 29–54% lower activities in the presence of the monoglucosides isorhamnetin 3- O -β- glucoside, kaempferol 3- O -β- glucoside, and quercetin 3- O -β- glucoside. As expected, activity was displayed in the presence of the artificial substrate, p -nitrophenyl-β- D- glucoside, albeit 34% lower relative to K3G7R (1), a phenomenon consistent with other plant BGLUs ( Ahn et al., 2010; Escamilla-Treviño et al., 2006; Inoue and Ebizuka, 1996; Naoumkina et al., 2007; Nomura et al., 2008; Suzuki et al., 2006; Zhou et al., 2012). The rate of hydrolysis of the coumarin esculin was 15% that of K3G7R; this is not unexpected as high specificity for esculin is rare among plant BGLUs, including for three hydrolases known to cleave the related plant-derived chemical scopolin ( Ahn et al., 2010). The dihydrochalcone conjugate, phloridzin or phloretin 2 0 -O -glucoside, was not hydrolysed by recombinant BGLU15. Phloridzin is prominent in apple trees, including fruit, and is attacked by mammalian BGLUs ( Gosch et al., 2010; Ketudat Cairns and Esen, 2010), although not widely used by plant enzymes, including a Thai rosewood seed dalcochinin 8- O -β- D- glucoside BGLU ( Hösel and Barz, 1975; Svasti et al., 1999). Together, biochemical screening of various natural products revealed BGLU15 preferred flavonols.

To determine the relative substrate utilization for glycoside linkages, BGLU15 activity was tested towards β- and α- linked conjugates of quercetin, as a compromise for not testing other bisglycosides due to the unavailability of milligram quantities of pure flavonol bisrhamnosides (e.g., quercetin 3- O -α- rhamnoside-7- O -α- rhamnoside) and flavonol 3- O -β- rutinoside-7- O -α- rhamnoside. BGLU15 displayed no preference for non-glucose conjugates as there was negligible activity for the 3- O -β- galactoside, 3- O -β-rutinoside (rutin) and 3- O -α- rhamnoside of quercetin ( Fig. 7 ). The lack of activity for substrates consisting of α- rhamnose or β-rutinose conjugated to the 3-hydroxy position of quercetin suggests flavonol bisrhamnosides or flavonol 3- O -β- rutinoside-7- O - α- rhamnosides are unlikely physiological substrates. This is not uncommon, as monolignol glucoside hydrolases display specificity for β- linked O -glucosides of o - and p -nitrophenyl and natural plant-derived compounds, but not for other glycosides ( Escamilla-Treviño et al., 2006).

For the most effectively utilized kaempferol and quercetin substrates, kinetic parameters of their hydrolysis by recombinant BGLU15 were probed further, along with the artificial substrate p -nitrophenyl-β- D- glucoside; in all cases, the relationship between substrate concentration and velocity was hyperbolic (see Supplementary Fig. 3 for Michaelis–Menten and Hanes–Woolf plots of K3G7R and Q3G7R hydrolysis). BGLU15 demonstrated a low apparent Km for flavonol glycosides in the range of 36–60 µM ( Table 1 ). This is not without precedent as other plant BGLUs display high affinities for physiologically relevant substrates including the soybean isoflavone conjugate hydrolase with Km s for 7- O -(6 00 -O -malonyl-β- D- glucosides of daidzein and genistein in the range of 19– 25 µM ( Suzuki et al., 2006), and a flavonol 3- O -β- heterodisaccharidase from dried aerial tissues of common buckwheat with Km s for kaempferol 3- O -β glucoside and K3G7R of 50 and 60 µM, respectively ( Baumgertel et al., 2003). BGLU15 demonstrated preference for flavonol 3- O -β- glucoside-7- O -α- rhamnosides, as kinetic parameters (kcat / Km) for K3G7R (1) and Q3G7R (2) were 67–200% higher relative to their 3- O -β- monoglucoside counterparts. QTOF-MS/MS confirmed that recombinant BGLU15 action on K3G7R (1) and Q3G7R (2) released kaempferol 7- O -α- rhamnoside (3) and quercetin 7- O -α- rhamnoside (4), respectively; the latter was similar to the Q3G7R product identified from in vitro assays of cell free Arabidopsis enzyme preparations. Preferential hydrolysis of flavonol 3- O -β- glucoside-7- O -α- rhamnosides suggests a biological role for BGLU15 as the initial step promoting the loss of these compounds in response to abiotic stress recovery, potentially resulting in the formation of smaller catabolite molecules. However, the flavonol bisglycosides are protected from futile degradation as they tend to accumulate in the central vacuole ( Zhao and Dixon, 2010), while BGLU15 is loosely associated with the cell wall ( Borderies et al., 2003). This implies a requirement for vacuolar efflux of flavonol bisglycosides, a hitherto unknown mechanism that is also postulated for isoflavone conjugate hydrolysis ( Suzuki et al., 2006).

Catabolism of flavonol glycosides (i.e., rutin) is well documented in fungi whereby quercetin formed via hydrolysis yields phloroglucinol carboxylic and protocatechuic acids following the concerted action of an oxygen-dependent quercetinase and esterase activities ( Tranchimand et al., 2010). Similarly, accumulation of quercetin oxidation products, 3,4-dihydroxybenzoic acid and 2,4,6-trihydroxyphenylglyoxylic acid occurs during drying of the outer scales of senescing onions, a phenomenon dependent upon sequential hydrolysis of quercetin 3,4 0 bisglucoside by BGLU activities ( Takahama and Hirota, 2000). Thus, the possibility remains that flavonol 7- O -rhamnosides produced by BGLU15 are hydrolyzed further to flavonol by an unknown α- rhamnosidase, and subsequently oxidized during abiotic stress recovery. Alternatively, BGLU15 mediated turnover of flavonol bisglycosides may have important consequences for plant growth and development. Hyper-accumulation of kaempferol bisglycosides, including K3G7R, in an Arabidopsis mutant deficient in the lignin biosynthesis enzyme p -coumaroyl shikimate 3 0 -hydroxylase is associated with dwarfism; transformation of the mutant with a chemically inducible form of this enzyme confers a rapid loss of flavonol bisglycosides, followed by renewed stem inflorescence development due to increased lignin deposition ( Kim et al., 2014).

Figures

Fig. 2. Evidence for Q3G7R (2) hydrolyzing activity from cell-free Arabidopsis extracts is coincident with the loss of flavonol 3-O-β-glucoside-7-O-α-rhamnosides during synergistic abiotic stress recovery. (A) Plants were maintained under 0 mM nitrate at 10 °C (nitrogen deficiency and low temperature, NDLT; represented as the grey-shaded portion of the time course) for 7 days. Thereafter plants were repleted with 10 mM nitrate and maintained at 21 °C (nitrogen sufficiency and high temperature, NSHT; represented as the non-shaded portion of the time course) for 5 days. (B) As a control, plants of similar age were left under continual NSHT. K3G7R (1) and Q3G7R (2) concentrations were determined by UHPLC-DAD-MS n analysis of acidified methanolic extracts. A peak areas for Q3G7R (2; retention time = 16.3 min) and K3G7R 360 (retention time = 16.9 min) were compared to known amounts of Q3G and corrected for amount of fresh matter used for the extraction.MS n analysis confirmed the identity of K3G7R (1) with an [M—H] —1 parent ion of 593.2 and fragment ions of 447.2, 431.2 and 285.1; Q3G7R (2) analysis revealed a parent ion of 609.2 and fragment ions of 463.2, 447.2 and 301.1. Q3G7R (2) hydrolysis rate reflects the amount of product formed following 5 h incubation with Arabidopsis enzyme preparations using assay A and their analysis by HPLC-DAD (see Experimental). Reaction rates are expressed as pmol quercetin equivalents min—1 mg protein—1. For all plots, each data point represents the mean ± SE of three separate experiments.

Fig. 1. Hydrolysis of kaempferol 3-O-β-glucoside-7-O-α-rhamnoside (1) and quercetin 3-O-β-glucoside-7-O-α-rhamnoside (2) by β-glucosidase from Arabidopsis thaliana yielding kaempferol 7-O-α-rhamnoside (3) and quercetin 7-O-α-rhamnoside (4).

Fig. 3. Quercetin 3-O-β-glucoside-7-O-α-rhamnoside β-glucosidase activity in crude protein extracts obtained from Arabidopsis plants growing under NDLT for 7 days and NSHT for 2 days. (A) HPLC-DAD analysis of enzyme assay shows the formation of a reaction product (P) with the retention time 13.3 min. The reaction product (see magnification in the inset) was produced in the presence of the protein extract (grey line) but not in its absence (black line) (B) Analysis of the purified reaction product by negative ion quadrupole TOF-MS/MS produced a molecular mass of 448.2.

Fig. 4. Phylogenetic tree of amino acid sequences of Arabidopsis and other plant glycoside hydrolase family 1 BGLUs. The unrooted tree was constructed with maximum likelihood method in MEGA 6.06 (Tamura et al., 2013), following multiple protein sequence alignment with ClustalW (www.genome.jp/tools/clustalw; Larkin et al., 2007). The numbers shown at each node of the tree are representative of percent support values greater than 60 obtained by bootstrap analysis using 1000 replicates. BGLUs shown in black represent biochemically and/or functionally characterized enzymes; BGLUs shown in grey represent enzymes not yet biochemically characterized. Biochemically characterized flavonoid glycoside utilizing enzymes and are represented in blue. Arabidopsis (displayed in red font) and rice (Oryza sativa, Os) amino acid sequences are as designated by Xu et al. (2004) and Opassiri et al. (2006), respectively; where applicable, their biochemical function is provided in square brackets.For amino acid sequences of BGLUs from other plant species used for the alignment, their corresponding GenBank™ accession numbers are provided in parentheses as follows: isoflavone conjugatehydrolyzing BGLU (Glycine max, BAF34333); non-cyanogenic BGLU (Cicer arietinum, CAG14979); linamarase (Trifolium repens, CAA40058); BGLU G1 (M. truncatula, ABW76286); BGLU G2 (M. truncatula, ABW76287); BGLU G3 (M. truncatula, ABW76288); BGLU G4 (M. truncatula, ABW76289); prunasin hydrolase (Prunus serotina, AAF34650); amygdalin hydrolase (Prunus serotina, AAA93234); furostanol glycoside 26-O-β-glucosidase (Cheilocostus speciosus, BAA11831); furcatin hydrolase (Viburnum furcatum, BAD14925); β-primeverosidase (Camellia sinensis, BAC78656); dalcochinin 80-O-β-glucoside BGLU (Dalbergia cochinchinensis, AAF04007); linamarase (Manihot esculenta, AAB22162); alkaloid BGLU (Carapichea ipecacuanha, BAH02544); raucaffricine BGLU (Rauvolfia serpentina, AAF03675); strictosidine BGLU (Rauvolfia serpentina, CAC83098); cardenolide 16-O-glucohydrolase (Digitalis lanata, CAB38854); myrosinase (Brassica napus, CAA42775); myrosinase (Raphanus sativus, BAB17226); thioglucoside glucohydrolase (Sinapis alba; CAA42534); coniferin β-glucosidase (Pinus contorta, AAC69619); dhurrinase (Sorghum bicolour, AAC49177); cytokinin glucoside β-glucosidase (Zea mays, CAA52293); β-glucosidase (Secale cereale, AAG00614); avenacosidase (Avena sativa, AAD02839); β-glucosidase (Hordeum vulgare, AAA87339); hydroxynitrile glucoside-cleaving β-glucosidase D2 (Lotus japonicus, ACD65510); hydroxynitrile glucoside-cleaving β-glucosidase D4 (L. japonicus, ACD65509); hydroxynitrile glucosidecleaving β-glucosidase D7 (L. japonicus, ACD65511). Accessions numbers for re-annotated amino acid sequences of putative M. truncatula BGLUs described by Suzuki et al. (2006) are as follows (ascribed a lower case for putative gene with similar names): M. truncatula BGLU D2a (AES76414); M. truncatula BGLU D2b (XP_003597509); M. truncatula BGLU D4a (XP_003620203); M. truncatula BGLU D4b (AES67734); M. truncatula BGLU D4c (XP_003589430); M. truncatula BGLU (AES67732).The bar at the bottom of the diagram represents the scale (0.2 substitutions per site) implemented to draw the branches.

Fig. 5. Analysis of relative Arabidopsis BGLUs expression patterns. Plants were maintained under 0 mM nitrate at 10 °C (nitrogen deficiency and low temperature, NDLT; represented as the grey-shaded portion of the time course) for 7 days. Thereafter, plants were repleted with 10 mM nitrate and transferred to 21 °C (nitrogen sufficiency and high temperature, NSHT; represented as the non-shaded portion of the time course) for 5 days. Expression levels were determined by RTqPCR as indicated under Materials and Methods. For each BGLU gene, data is expressed as the fold change in NDLT treatments (before and after transfer to NSHT) relative to control (continuous NSHT) conditions (2 —ΔΔC). For each experimental T replicate, RT-qPCR reactions were performed in duplicate. For each time course data represents the mean ± SE of three separate experiments.

Fig. 6. SDS–PAGE (A) and immunoblot analysis (B) of the expression and purification of the mature recombinant BGLU15 from E. coli Origami 2 (DE3) cells. The left-hand lane on each gel contains the size in kDa of molecular mass standards. Each lane was loaded with 5 µL of the sample mixed with 5 µL of 2× sample dye. The immunoblot was probed with an anti-His6 antibody.

Fig. 7. Relative substrate utilization of recombinant BGLU15. For each substrate, in vitro activity was measured at 500 µM using Assay B; each assays were incubated for 10 min and products detected by HPLC-DAD. Reaction rates are expressed as percentage of that calculated for K3G7R (1) (100% = 0.96 µmol min—1 mg protein —1) ± SE of three separate recombinant protein preparations.