Inhibitors of plant invertases do not affect the structurally related enzymes of fructan metabolism


Author for correspondence:
Steffen Greiner
Tel: +49 (0)6221 545786
Fax:+49 (0)6221 545859


  • • Plant fructan active enzymes (FAZYs), including the enzymes involved in inulin metabolism, namely sucrose:sucrose 1-fructosyltransferase (1-SST; EC, fructan:fructan 1-fructosyltransferase (1-FFT; EC and fructan 1-exohydrolase (1-FEH; EC, are evolutionarily related to acid invertases (AIs), that is, plant cell wall invertase (CWI) and vacuolar invertase (VI). Acid invertases are post-translationally controlled by proteinaceous inhibitors. Whether FAZYs are subject to similar controls is not known.
  • • To probe their possible interactions with invertase inhibitors, we transiently expressed chicory (Cichorium intybus) FAZYs, as well as several previously characterized invertase inhibitors from nonfructan species, and the C. intybus cell wall/vacuolar inhibitor of fructosidase (CiC/VIF), a putative invertase inhibitor of a fructan-accumulating plant, in leaves of Nicotiana benthamiana.
  • • Leaf extracts containing recombinant, enzymatically active FAZYs were used to explore the interaction with invertase inhibitors. Neither heterologous inhibitors nor CiC/VIF affected FAZY activities. CiC/VIF was confirmed as an AI inhibitor with a stronger effect on CWI than on VI. Its expression in planta was developmentally regulated (high in taproots, and undetectable in leaves and flowers). In agreement with its target specificities, CiC/VIF was associated with the cell wall.
  • • It is concluded that subtle structural differences between AIs and FAZYs result in pronounced selectivity of inhibitor action.


The most prominent storage carbohydrates in higher plants are starch and sucrose. In addition, linear and branched fructans occur as reserve carbohydrates in c. 15% of flowering plants (Hendry, 1993). In addition to the obvious function as carbohydrate storage compounds, fructans are also involved in adaptation to drought (Hendry, 1993; Pilon-Smits et al., 1999) and cold (Pollock et al., 1988) stress. Fructans are defined as oligo- or polysaccharides that contain at least two adjacent fructose units. Among the five structurally different types of fructans occurring in higher plants (Van Laere & Van den Ende, 2002), inulin, a linear exclusively β(2-1) fructosyl-fructose linked fructose polymer with a terminal glucose unit, is widely used in human nutrition, for example as a low-calorie sweetener, as dietary fibre or as a fat substitute (Niness, 1999), and it has been claimed to have several health-promoting effects, including stimulation of beneficial Bifidobacteria (Niness, 1999), enhanced calcium absorption (osteoporosis prevention) (Coudray et al., 1997) and a positive role in colon and breast cancer prevention (Taper et al., 1995, 1997). On an industrial scale, chicory (Cichorium intybus) roots are the main source for inulin extraction. Thus, exploring the regulation of fructan accumulation and its turnover in chicory serves two goals: to improve our understanding of how fructan metabolism is affected by plant development and during adaptation to stressful environments, and, based on this knowledge, to develop new strategies to improve the yield and/or quality of inulin via breeding efforts or transgenic approaches.

The starting substrate for inulin biosynthesis is sucrose. Based on the model of Edelman & Jefford (1968), sucrose:sucrose 1-fructosyltransferase (1-SST; EC catalyzes the first step, transferring one fructose moiety between two sucrose molecules to produce glucose and the trisaccharide 1-kestotriose. Fructan:fructan 1-fructosyltransferase (1-FFT; EC then transfers fructose moieties from and to 1-kestotriose or larger fructans (Koops & Jonker, 1996; Luscher et al., 1996; Van den Ende & Van Laere, 1996), resulting in inulin and oligofructose of varying degrees of polymerization. Fructan breakdown into hexoses is catalyzed by fructan 1-exohydrolases (1-FEHs; EC, with several isoforms found in chicory (1-FEH I, IIa and IIb), which remove terminal fructose units (Claessens et al., 1990), and finally by acid invertases (AIs; EC, which hydrolyze sucrose to produce glucose and fructose.

A comparison of reaction mechanisms and protein structures has revealed a close relationship between fructan active enzymes (FAZYs) and AIs, indicating that the former have evolved from the latter (Smeekens et al., 1991). AIs, that is, cell wall invertases (CWIs) and vacuolar invertases (VIs), and FAZYs have been grouped together in glycosyl hydrolase (GH) family 32 in the carbohydrate active enzyme database ( based on amino acid sequence similarities (Henrissat, 1991), with the implication that they are likely to share a common three-dimensional fold. Chicory 1-FEH IIa was recently found to have the typical GH 32 five-bladed β-propeller structure (Verhaest et al., 2005b), as was also found for a CWI from Arabidopsis thaliana, Arabidopsis thalania cell wall invertase 1 (AtcwINV1) (Verhaest et al., 2005a; Lammens et al., 2008). While 1-SST and 1-FFT are more closely related to VI, the fructan exohydrolases are more closely related to CWI (Van den Ende et al., 2000; Van den Ende et al., 2001; Van Laere & Van den Ende, 2002). In contrast to AIs with their cell wall and vacuolar isoforms, all FAZYs are thought to be localized to the vacuole (Pollock, 1988; Van Laere & Van den Ende, 2002), the putative site of fructan synthesis, storage and breakdown.

Regulation of FAZY expression at the transcriptional level during different developmental stages and in response to cold stress exposure is well documented (Van Laere & Van den Ende, 2002). Furthermore, control of FAZY expression by sucrose and other sugars (Müller et al., 2000; Martinez-Noël et al., 2001), calcium (Martinez-Noël et al., 2006) and nitrate (Morcuende et al., 2004) has been demonstrated. However, additional mechanisms for FAZY regulation beyond transcriptional and translational control may exist, as discrepancies between transcript amounts and enzyme activities have been observed: in chicory, 1-SST activity dropped to zero at the end of the field growing season, while mRNA concentrations remained clearly detectable (Van Laere & Van den Ende, 2002). This observation, and the relatedness of FAZYs to AIs, has motivated us to address the question of whether FAZYs are, like CWI and VI, also under post-translational control via proteinaceous inhibitors (Jaynes & Nelson, 1971; Burakhanova et al., 1987; Greiner et al., 1998; Greiner et al., 1999; Rausch & Greiner, 2004).

Invertase inhibitors (cell wall and/or vacuolar inhibitors of fructosidases (C/VIFs)) belong to the larger family of pectin methylesterase inhibitor-related proteins (PMEI-RPs). Members of this protein family share only moderate sequence homology but similar structural scaffolds (Hothorn et al., 2004). At present, it is not possible to assign putative PMEI-RPs to their target enzymes, AI or pectin methylesterase (PME), respectively, based on sequence information alone (Rausch & Greiner, 2004). Interactions of FAZYs with inhibitory proteins have thus far not been described. If FAZYs are, like AIs, post-translationally regulated via proteinaceous inhibitors, the question arises as to whether they have specific inhibitor counterparts, or, alternatively, interact with invertase inhibitors. While PMEIs do not interact with AIs and invertase inhibitors do not interact with PMEs, the possibility of cross-reactions of invertase inhibitors with FAZYs cannot be excluded. Therefore, as a first step we have addressed the question of whether, as a result of their structural similarities with AIs, FAZYs also interact with invertase inhibitors. If they do so, this would have far-reaching consequences for the regulation of fructan metabolism in planta, and also for any attempt to engineer post-translational control of FAZYs. Here we report results concerning the effects of several invertase inhibitors, including the first invertase inhibitor-like protein from a fructan-accumulating plant, on enzyme activities of FAZYs and AIs from chicory.

Materials and Methods

Plant material and cultivation

For Agrobacterium tumefaciens leaf infiltration, 4–6-wk-old Nicotiana benthamiana L. plants grown in a growth chamber under a 16-h light period (300 µmol m−2 s−1) were used. Cichorium intybus L. var. Zoom was grown under glasshouse conditions (RNA isolation for cDNA synthesis), or in a field at the University of Heidelberg botanical garden (2006 and 2007) for C. intybus cell wall/vacuolar inhibitor of fructosidase (CiC/VIF) expression analysis. Musa acuminata and Helianthus annus plants were grown under glasshouse conditions and in the field, respectively.

cDNA cloning of CiC/VIF, FAZYs and invertase inhibitors

A 155-bp fragment of the CiC/VIF coding region was amplified from C. intybus taproot cDNA with the primers LsInhF and LsInhR (all primers are shown in Table 1), designed using expressed sequence tag (EST) data for a putative invertase inhibitor of the closely related Lactuca sativa (gi:22233301). Using this sequence information, the primers CiInh5′RACE (Rapid Amplification of cDNA Ends) and CiInh3′RACE were designed for 5′RACE and 3′RACE, respectively. With the obtained sequence information for 565 bp, which was 99% identical to an EST sequence (gi:124608900) deposited later in the National Centre for Biotechnology Information database, it was possible to clone the full-length CiC/VIF with primers CiInhF and CiInhR. Using Gateway® technology (Invitrogen, Karlsruhe, Germany), constructs were mobilized into the entry vector pDONRTMZeo (Invitrogen). For ectopic expression in plant tissue under control of the Cauliflower mosaic virus (CaMV) 35S promoter, full-length CiC/VIF was subcloned by recombination into the destination vector pMDC32 (Curtis & Grossniklaus, 2003) according to the manufacturer's instructions (Invitrogen). For bacterial expression in Escherichia coli Rosetta gami cells, CiC/VIF without the signal peptide was cloned with primers CiInhFragF and CiInhR into the entry vector pDONR201 and subcloned by recombination into the destination vector pETG10A according to the manufacturer's instructions. Full-length putative M. acuminata (MaC/VIF) and H. annus (HaC/VIF) inhibitors were cloned from cDNA with the primers shown in Table 1, designed according to the EST sequences gi:83722841 (M. acuminata) and gi:90465856 (H. annus), respectively. Cloning and subcloning were performed as described for CiC/VIF.

Table 1.  List of primers used for cDNA cloning
cDNACommentPrimer namePrimer sequence (5′→3′)
  1. Included are primers for full-length constructs (full), for partial cDNAs for probe synthesis (fragment), and for heterologous expression of protein domains used for raising polyclonal antisera (i.e. exon 3 for sucrose:sucrose 1-fructosyltransferase (1-SST) and fructan 1-exohydrolase II (1-FEH II); mature protein for Cichorium intybus cell wall/vacuolar inhibitor of fructosidase (CiC/VIF)). In primer sequences for Gateway cloning (see the Materials and Methods), attB1 (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCT-3′) and attB2 (5′-GGGGACCACTTTGTAAAGAAAGCTGGGT-3′) sites are indicated as (GW-fw) and (GW-rv), respectively. 1-FFT, fructan:fructan 1-fructosyltransferase; C/VIF, cell wall/vacuolar inhibitor of fructosidase; At, Arabidopsis thaliana; Ha, Helianthus annus; Ma, Musa acuminata; RACE, Rapid Amplification of cDNA Ends.


Full-length coding sequences of chicory FAZYs were amplified by polymerase chain reaction (PCR) from C. intybus root cDNA coding for 1-SST, 1-FFT, 1-FEH I and 1-FEH IIa/IIb (for primers used, see Table 1). With the primers designed initially for amplification of 1-FEH IIa, both 1-FEH II isoforms could be amplified. Using Gateway® technology, constructs were cloned into the entry vector pDONRTMZeo and, for expression under control of the CaMV 35S promoter, subcloned by recombination into the destination vector pMDC32 according to the manufacturer's instructions. As the two 1-FEH II isoforms are 94% identical (at the protein level), only 1-FEH IIa was selected for ectopic expression under control of the CaMV 35S promoter. Cloning of invertase inhibitors from sugar beet (Beta vulgaris) root cDNA (BvC/VIF) (Eufinger, 2006), N. tabacum leaf cDNA (NtCIF and NtVIF) (Greiner et al., 1998; Greiner et al., 1999) and A. thaliana flower cDNA (AtC/VIF1 and AtC/VIF2) (Link et al., 2004) has been described previously. For expression of full-length proteins in N. benthamiana leaves, the expression vectors pBinar_NtCIF, pBinar_NtVIF and pK2GW7_BvC/VIF were used and the full-length expression constructs pMDC32_AtC/VIF1 and pMDC32_AtC/VIF2 were constructed by cloning into the entry vector pDONRTMZeo (the primers used are shown in Table 1) and subcloning by recombination according to the manufacturer's instructions.

Agrobacterium tumefaciens-mediated transient expression of FAZYs and invertase inhibitors in N. benthamiana leaves

Agrobacterium tumefaciens C58C1 cells were transformed by electroporation with the respective expression constructs. For leaf infiltration, A. tumefaciens cultures were grown overnight in 0.1% yeast extract, 0.5% beef extract, 0.5% sucrose, 0.0493% MgSO4 heptahydrate; pH 7.5 medium containing rifampicin, carbenicillin and kanamycin (100, 50 and 50 µg ml−1, respectively), harvested by centrifugation at 3400 g and re-suspended to an optical density (OD600) = 1 in 5 mM 2-(N-morpholino)-ethansulfonic acid (pH 5.6), 10 mM MgCl2 and 150 µM acetosyringone. After an additional 1 h of incubation with slight rotation at room temperature, N. benthamiana leaves were infiltrated with the A. tumefaciens cell suspensions using a 2-ml syringe without a needle by applying gentle pressure on the lower surface of the leaf at various points until the bacterial suspension filled the whole leaf, as described elsewhere (Sparkes et al., 2006). Infiltrated leaves were harvested 48 h after infiltration.

Generation of antisera and immunoblot analysis

For detection of 1-FEH IIa, an antiserum against 1-FEH IIb was used, as the two isoforms are 97% identical in the protein fragment recombinantly expressed for antibody production. For antiserum production, fragments of 1-SST and 1-FEH IIb (‘exon3’, corresponding to exon 3 of Nicotiana tabacum cell wall invertase) and full-length CiC/VIF (the mature protein without the signal peptide) were amplified by PCR with the primers shown in Table 1 and, for expression of recombinant proteins in E. coli, cloned into the vector pQE30 (1-SST and 1-FEH IIb) using SacII/XbaI (1-SST) or SphI/HindIII (1-FEH IIb) restriction sites or pETG10A (CiC/VIF) using Gateway® technology according to the manufacturer's instructions. His6-tagged recombinant proteins were purified via Ni-NTA (Qiagen, Hilden, Germany) (1-SST and 1-FEH IIb) or Ni-TED (Macherey-Nagel, Düren, Germany) (CiC/VIF) columns according to the manufacturer's instructions under denaturing conditions with 6 M guanidine hydrochloride and sent to Eurogentec (Seraing, Belgium) for polyclonal antibody production in a rabbit. For immunoblot analysis, protein extracts (fresh weight equivalents) were blotted on Immobilon P (Millipore, Schwalbach, Germany) transfer membrane. ImmunoPure Peroxidase Conjugated Goat Anti-Rabbit IgG (H + L) (Pierce, Rockford, IL, USA) was used as secondary antibody (1 : 20 000) for chemiluminescent detection (Super Signal® West Dura Extended Duration Substrate; Thermo Scientific, Rockford, IL, USA). Even protein loading and blotting were analyzed by subsequent amido black staining (0.1% in 45% methanol/10% acetic acid) of the membrane.

Preparation of biotinylated probes and northern blot analysis

Gene-specific biotinylated probes for northern blot analysis were generated by PCR against 850–870-bp fragments of 1-SST, 1-FFT, 1-FEH I and 1-FEH IIa coding sequences (see Table 1‘fragment’), respectively, and against the full-length coding sequence (without the signal peptide) of CiC/VIF. RNA from 500 mg of frozen leaf tissue was extracted according to the protocol of Logemann et al. (1987); for northern blot analysis, 10 µg of RNA was blotted on Roti-Nylon (0.2 µm) transfer membrane (Roth, Karlsruhe, Germany), hybridized with gene-specific probes, and analyzed by chemiluminescent detection (North2South® Chemiluminescent Substrate HRP Kit; Pierce). Even RNA loading and blotting were analyzed by subsequent methylene blue (0.02% in 0.3 M Na-acetate; pH 5.2) staining of the membrane.

Preparation of FAZY fractions, soluble and cell wall-associated invertase fractions and inhibitor fractions for activity assays and immunoblot analysis

Soluble proteins were extracted from 250 mg of frozen and ground leaf material in extraction buffer (Van den Ende et al., 1999) followed by 80% ammonium sulphate precipitation and dialysis against assay buffer (50 mM Na-acetate; pH 5). For extraction of cell wall-associated proteins for activity assays, remaining pellets were incubated overnight at 4°C (in an overhead shaker) in 1 M NaCl. The supernatant was subsequently precipitated with ammonium sulphate (80%) and dialysed against assay buffer. Protein concentrations were determined by the Bradford assay (Roti®-Quant; Roth). For immunoblot analysis, soluble protein extracts (without dialysis) and remaining pellets after two washing steps (the first with extraction buffer + 1% Triton X100 and the second with extraction buffer) were boiled in sodium dodecyl sulphate (SDS) sample buffer (final concentrations: 31.25 mM Tris-HCl, pH 6.8, 1% SDS, 2.5% 2-mercaptoethanol and 0.01% bromphenol blue) before sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE).

Partial (negative) purification of CiC/VIF and other invertase inhibitors by concanavalin A (Con A) chromatography

Soluble protein extracts of CiC/VIF-expressing leaves were partly purified by Con A chromatography (Sigma-Aldrich, Taufkirchen, Germany) at pH 6.8 (Con A buffer: 50 mM Na-acetate, 1 mM CaCl2, 1 mM MgCl2 and 1 mM phenylmethylsulphonyl fluoride (PMSF)) to separate CiC/VIF from glycosylated invertase proteins. The Con A-nonbound fraction (ConA−) was concentrated (Vivapore5, MWCO 7500; Sigma-Aldrich), dialysed against assay buffer and used as the inhibitor fraction. Proteins bound to the Con A column (ConA+) were eluted with Con A buffer containing 15% methyl-α-D-glucopyranoside. The distribution of CiC/VIF between different Con A fractions was analyzed by immunoblot analysis of all fractions. Soluble AI activity (i.e. monitoring the separation of inhibitor from invertase) was analyzed before Con A chromatography and subsequently in the ConA− and ConA+ fractions. CiC/VIF amounts were estimated by immunoblot analysis, calibrated with recombinant CiC/VIF protein expressed in E. coli. Negative purification of other inhibitors was carried out with the same protocol; however, quantification of inhibitor amounts was not possible. Therefore, inhibitor fractions for inhibition assays were adjusted to the maximum amount of CiC/VIF-containing extracts used (i.e. corresponding to 500 ng of CiC/VIF).

Determination of enzyme activities

Dialysed protein extracts (20 µg of total protein) were incubated in assay buffer (50 mM Na-acetate, pH 5.2, and 0.02% Na-azide) for 60 min at 30°C with the following substrates: 5 mM sucrose for invertase activity, 100 mM sucrose for 1-SST activity, 1.8 mM 1-kestotriose (GF2; Wako Chemicals, Neuss, Germany) for 1-FFT activity and 1.2 mM 1,1,1-kestopentaose (GF4; Wako Chemicals) for 1-FEH activity. As reaction products, fructose (for invertase and 1-FEH activity), 1-kestotriose (for 1-SST activity) and 1,1-kestotetraose (for 1-FFT activity) were identified (via co-chromatography) and quantified by high-performance anion-exchange chromatography with pulsed amperometric detection analysis (ICS-3000, ED Electrochemical Detector, Carbopac PA1 column; Dionex, Sunnyvale, CA, USA) with glucose, fructose, sucrose, 1-kestotriose, 1,1-kestotetraose and 1,1,1-kestopentaose as external standards. A stepwise gradient of eluent A (150 mM NaOH) and eluent B (150 mM NaOH containing 700 mM Na-acetate) was applied at a flow rate of 1 ml min−1 as follows: 0–3 min 3% B (constant), linear gradient to 29% B at 8 min, linear gradient to 97% B at 13 min. Specific activities were calculated in nmol product formed per mg total protein per minute. For inhibition assays, invertase inhibitor-containing fractions were preincubated with target enzyme-containing fractions for 20 min at room temperature in assay buffer before addition of substrate. All enzyme measurements were performed under conditions where activities were proportional to enzyme amount and incubation time.

Enzyme activities and remaining activities in inhibition assays are means ± SD of duplicate measurements of three biological replicates, except for the inhibitory effect of CiC/VIF on AI activities (Fig. 4b), which is given for a single experiment which was repeated three times with similar results.

Figure 4.

Functional characterization of recombinant Cichorium intybus cell wall/vacuolar inhibitor of fructosidase (CiC/VIF) protein ectopically expressed in leaves of Nicotiana benthamiana. (a) Separation of nonglycosylated CiC/VIF protein from glycosylated invertase via concanavalin A (Con A) chromatography. Upper panel: immunoblot analysis of recombinant CiC/VIF. (S), soluble protein fraction; (–), Con A-nonbound fraction; (+), Con A-bound fraction eluted with 15% methyl-α-D-glucopyranoside. Lower panel: corresponding acid invertase activities (nmol fructose min−1 mg−1 protein). (b) Inhibitory effect of CiC/VIF (i.e. Con A-nonbound fraction; see (a)) on vacuolar invertase (VI) activities (CiVI and NbVI: soluble protein fractions from chicory hairy root culture and N. benthamiana leaves, respectively) and cell wall invertase (CWI) activities (CiCWI and NbCWI: cell wall protein fractions from chicory hairy root culture and N. benthamiana leaves, respectively). The amount of CiC/VIF protein (ConA− fraction) was estimated from a comparison with recombinant CiC/VIF protein expressed in Escherichia coli (not shown). Values are shown for a single experiment which was repeated three times with similar results.


Ectopic expression of chicory 1-SST, 1-FFT, 1-FEH I and 1-FEH IIa enzymes in N. benthamiana

Full-length FAZY cDNAs were cloned from glasshouse-grown chicory taproots by reverse transcriptase–poymerase chain reaction (RT-PCR). As enzymatically active recombinant FAZYs could not be expressed in E. coli cells, probably because of incorrect folding or lack of glycosylation, we decided to transiently express chicory FAZYs (1-SST, 1-FFT, 1-FEH I and 1-FEH IIa) in N. benthamiana leaves under control of the 35S promoter (via A. tumefaciens leaf infiltration; see Materials and Methods) to generate recombinant FAZY preparations for inhibition studies with recombinant invertase inhibitors. Nicotiana benthamiana does not accumulate fructans and does not possess any of the above-mentioned FAZY genes. Therefore, this expression system provides an elegant method for the enzymatic analysis of individual FAZYs in relatively crude extracts, being particularly suitable for rapidly screening large numbers of inhibitor–target enzyme combinations. Also, this plant expression system has the additional advantage of mediating plant-type glycosylation, as FAZYs are, like AIs, glycoproteins (Van den Ende et al., 1996a,b, 2001). Only in the case of 1-FEH were significant background activities observed (originating from AI; see ‘Ectopic expression of chicory 1-SST, 1-FFT, 1-FEH I and 1-FEH IIa enzymes in N. benthamiana’ in the Results section), which could be corrected for.

Immunoblot analysis of extracts from 1-SST- and 1-FEH IIa-expressing N. benthamiana leaves detected proteins with the expected molecular weights of 49 kDa (and a 28-kDa fragment; not shown) and 65 kDa (and a 32-kDa fragment; not shown), respectively (Fig. 1b), in agreement with the molecular weights of native enzymes in extracts from chicory taproots. For 1-FFT and 1-FEH I, no antisera were available, but expression at transcript level was comparable to that of 1-SST and 1-FEH IIa (Fig. 1a). FAZY activities measured in dialyzed extracts from transformed leaves were 201.2 ± 54 nmol min−1 mg−1 for 1-SST and 11.4 ± 1.9 nmol min−1 mg−1 for 1-FFT, with background below 5% in each case (Fig. 1c). 1-FEH I and 1-FEH IIa activities were also clearly detectable at 12.2 ± 1.6 and 20.5 ± 0.2 nmol min−1 mg−1, respectively. However, here the background activity of 3.8 nmol min−1 mg−1 was not negligible. The observed background activity is presumably derived from native N. benthamiana AIs, as it was previously shown that AIs can also hydrolyze substrates other than sucrose, such as 1-kestotriose, raffinose or stachyose (Tymowska-Lalanne & Kreis, 1998; De Coninck et al., 2005). Corroborating this assumption, background activity could be strongly inhibited by the recombinant invertase inhibitor NtCIF (up to −90%; data not shown).

Figure 1.

Ectopic expression of fructan active enzymes (FAZYs) in leaves of Nicotiana benthamiana via Agrobacterium tumefaciens-mediated transformation and validation of corresponding leaf extracts as target enzyme preparations. SST, sucrose:sucrose 1-fructosyltransferase (1-SST); FFT, fructan:fructan 1-fructosyltransferase (1-FFT); FEH1, fructan 1-exohydrolase I (1-FEH I); FEH2, fructan 1-exohydrolase I (1-FEH IIa). (a) Northern blot showing expression of different FAZYs; (b) immunoblot with specific polyclonal antisera directed against recombinant 1-SST and 1-FEH II protein domains (see the Materials and Methods; immunosignals at 49 kDa (1-SST) and 65 kDa (1-FEH IIa) are shown); nd, not determined. (c) Specific FAZY activities after ectopic expression in leaves of N. benthamiana for 48 h in nmol product × min−1 × mg−1 protein; products measured: 1-kestotriose (1-SST), 1,1-kestotetraose (1-FFT) and fructose (1-FEH). Controls (C) for 1-SST and 1-FFT: extracts of leaves expressing 1-FEH IIa; controls (C) for 1-FEH I and 1-FEH IIa: extracts of leaves expressing 1-SST.

FAZY activities are not affected by invertase inhibitors from nonfructan plant species

Using the same transient expression system as for FAZYs, extracts containing recombinant invertase inhibitors from the following nonfructan species were generated: N. tabacum (cell wall (NtCIF) and vacuolar (NtVIF) isoforms) (Greiner et al., 1998, 1999), B. vulgaris (BvC/VIF) (Eufinger, 2006) and A. thaliana (AtC/VIF 1 and 2) (Link et al., 2004). To separate recombinantly expressed inhibitors from endogenous AIs, the latter were removed by binding to a Con A matrix (see the Materials and Methods section), whereas inhibitors remained in the nonbound fraction (ConA−) as only a minor fraction of inhibitor remains bound to AIs at pH 6.8. All five PMEI-RPs have previously been expressed in E. coli and functionally characterized as bona fide invertase inhibitors. To confirm these results with recombinant inhibitors synthesized in planta and to validate the plant expression system, the effects of recombinant inhibitor-containing plant extracts on VI and CWI activities of N. benthamiana leaf and C. intybus hairy root extracts, respectively, were determined (Table 2); for chicory, hairy root cultures were chosen as they provided a tissue with consistently high CWI and VI activities. As expected, NtCIF, an apoplastic invertase inhibitor from tobacco with broad specificity against AIs from many plant species (S. Greiner, unpublished), caused the strongest inhibition of CWI and VI activities. All other inhibitors were consistently effective against NbVI and CiCWI, but did not inhibit NbCWI and CiVI activities. Conversely, an apparent enzyme activation (up to 54%) was observed for several inhibitor–AI combinations. Such effects have been previously observed but the underlying mechanism has remained unexplained. One possibility for this mechanism is the replacement of endogenous AI-bound (active) inhibitor(s) by excess of the recombinant inhibitor, the latter not being active against the AI isoforms present in the plant extracts. Antisera were not available for all tested invertase inhibitors, preventing exact quantification of inhibitor amounts; however, based on comparison with recombinant CiC/VIF, the inhibitor-containing extracts contained c. 500 ng per assay of recombinant inhibitor protein.

Table 2.  Effect of plant invertase inhibitors from nonfructan species on chicory (Cichorium intybus) fructan active enzyme (FAZY) activities as compared with their effect on cell wall invertase (CWI) and vacuolar invertase (VI) preparations extracted from Nicotiana benthamiana leaves (Nb) and chicory hairy root cultures (Ci), respectively
 Specific activityRemaining enzyme activity as % of control
  1. For characterization of chicory FAZY preparations see Fig. 1. Enzyme activities after preincubation with inhibitor preparations are given as a percentage of the control value. Specific activities of control extracts are given in nmol product min−1 mg−1 protein. Combinations with significant inhibition of enzyme activity are shaded. Note that in the case of fructan 1-exohydrolase I (1-FEH I) and IIa (1-FEH IIa), the significant inhibition of background activities (i.e. originating from endogenous invertase activities; see Fig. 1) was accounted for using the corresponding controls. Values are means ± SD (n = 3). At, Arabidopsis thaliana; Nt, Nicotiana tabacum; Bv, Beta vulgaris; C/VIF, cell wall/vacuolar inhibitor of fructosidase; 1-SST, sucrose:sucrose 1-fructosyltransferase; 1-FFT, fructan:fructan 1-fructosyltransferase.

NbVI11 (± 3)80 (± 5)78 (± 8)6 (± 5)75 (± 3)23 (± 11)
NbCWI60 (± 13)92 (± 6)105 (± 6)19 (± 5)137 (± 8)119 (± 7)
CiVI13 (± 5)104 (± 14)122 (± 16)59 (± 9)154 (± 10)135 (± 9)
CiCWI143 (± 35)69 (± 9)67 (± 14)28 (± 11)52 (± 9)57 (± 15)
1-SST201 (± 54)109 (± 10)106 (± 7)101 (± 10)104 (± 7)109 (± 5)
1-FFT11 (± 2)105 (± 8)97 (± 8)104 (± 10)94 (± 7)103 (± 6)
1-FEH I12 (± 2)96 (± 9)94 (± 10)86 (± 19)95 (± 2)87 (± 7)
1-FEH IIa21 (± 1)106 (± 5)105 (± 6)111 (± 7)108 (± 1)96 (± 18)

Having confirmed the inhibitory effect of invertase inhibitors ectopically expressed in planta on AI target enzymes, the same inhibitor preparations were incubated with extracts from N. benthamiana leaves ectopically expressing FAZYs (Table 2). None of the inhibitor–target combinations revealed any influence on enzyme activity, neither inhibition nor activation (as observed for some AIs). The above results were corroborated by inhibition assays with recombinant NtCIF and BvC/VIF proteins expressed in E. coli and purified to homogeneity. Recombinant NtCIF protein (150 ng) and recombinant BvC/VIF protein (1 µg) inhibited VI activity in N. benthamiana leaf extracts by 92 and 99%, respectively, with both inhibitors showing no significant effect on FAZY activities. Thus, under experimental conditions where recombinant invertase inhibitors showed pronounced effects on AI activities, FAZYs did not appear to interact with these inhibitors. This observation was further supported by the subsequent cloning and functional analysis of the first PMEI-related protein from chicory.

Molecular cloning of CiC/VIF, the first PMEI-related protein of a fructan-accumulating species

We started our search for a PMEI-related protein in chicory using sequence information from an EST clone of L. sativa (GI:22233301). By a combination of RT-PCR and 5′/3′-RACE-PCR, we cloned a full-length inhibitor-encoding cDNA, tentatively named CiC/VIF, using total RNA extracted from glasshouse-grown chicory taproots, a tissue that expresses both AI activities and FAZY activities. The predicted mature CiC/VIF protein has a molecular weight of 16 kDa, and a calculated isoelectric point of 8.3. Figure 2(a) shows an alignment of the translated mature CiC/VIF protein sequence without the signal peptide with the previously described invertase inhibitors NtCIF, NtVIF, BvC/VIF and AtC/VIF 1 and 2. It is well known that protein sequences are only moderately conserved among members of the PMEI-RP family (Rausch & Greiner, 2004). PMEI-RPs contain only a few conserved residues, among them four conserved cysteine residues also found in CiC/VIF. Although it is not possible to predict with certainty whether a novel PMEI-RP candidate belongs to the PMEI subgroup or the invertase inhibitor subgroup, a cladogram of CiC/VIF with several confirmed invertase inhibitors and two functionally characterized PMEIs from A. thaliana (AtPMEI 1 and 2) (Wolf et al., 2003) suggested that CiC/VIF is an invertase inhibitor rather than belonging to the PMEI subgroup (Fig. 2b). Attempts to express the CiC/VIF protein in E. coli resulted in insoluble protein aggregates with no inhibitory effect on AI preparations (results not shown). However, enough recombinant CiC/VIF protein was purified under denaturing conditions to generate a polyclonal antiserum.

Figure 2.

Molecular cloning of the first pectin methylesterase inhibitor-related protein (PMEI-RP) from a fructan-accumulating species: Cichorium intybus cell wall/vacuolar inhibitor of fructosidase (CiC/VIF) from chicory, a putative invertase inhibitor. (a) Protein sequence alignment (ClustalW; 2.0.3, of CiC/VIF with invertase inhibitors from tobacco (Nicotiana tabacum) (NtCIF, cell wall isoform; NtVIF, vacuolar isoform), sugar beet (Beta vulgaris) (BvC/VIF) and Arabidopsis thaliana (AtC/VIF 1 and 2), and (b) a corresponding cladogram which includes two previously characterized A. thaliana PMEI proteins (Wolf et al., 2003).

Ectopic expression of CiC/VIF in N. benthamiana leaves and expression analysis of CiC/VIF during plant development in C. intybus

To functionally characterize recombinant CiC/VIF under the same conditions as used for other invertase inhibitors (see ‘FAZY activities are not affected by invertase inhibitors from nonfructan plant species’ in Results section), CiC/VIF was transiently expressed in leaves of N. benthamiana. Forty-eight hours after N. benthamiana leaf infiltration with A. tumefaciens cells carrying the pMDC32_CiC/VIF construct, soluble and cell wall-associated proteins were extracted from CiC/VIF-expressing leaves and control (i.e. 1-FFT transformed) leaves. Strong expression of the CiC/VIF protein was detected in both soluble and cell wall-associated protein extracts (Fig. 3a). With an apparent molecular weight of 18–20 kDa, recombinant CiC/VIF protein showed a slightly higher molecular weight than predicted (16 kDa), but this was in agreement with immunosignals from endogenous CiC/VIF protein in chicory taproot extracts. During plant development, endogenous CiC/VIF was expressed in first-year and second-year (flowering stage) chicory taproots, being almost exclusively recovered in the cell wall-associated protein fraction (Fig. 3a; the presence of some CiC/VIF protein in the soluble protein fraction was detected after a longer exposure time; results not shown). Conversely, CiC/VIF was not expressed in chicory leaves or flowers. CiC/VIF expression in chicory taproots was maintained throughout the entire field growing season (Fig. 3b), with a significant increase towards the end of the growing season. Interestingly, some (but not all) analyzed taproot extracts revealed two CiC/VIF species of 18 and 20 kDa. The same result was recently obtained for BvC/VIF in sugar beet taproot (Eufinger, 2006). Presently, it is not known whether several related CiC/VIF isoforms are present in chicory taproots or whether the CiC/VIF protein undergoes different modifications. Similar to other PMEI-RPs, CiC/VIF does not contain glycosylation sites and therefore does not bind to a Con A matrix (Fig. 4a).

Figure 3.

Ectopic expression of recombinant Cichorium intybus cell wall/vacuolar inhibitor of fructosidase (CiC/VIF) protein in Nicotiana benthamiana leaves and expression of endogenous CiC/VIF during plant development. (a) Ectopic expression of recombinant CiC/VIF protein in leaves of N. benthamiana (2) as compared with control leaves ectopically expressing fructan:fructan 1-fructosyltransferase (1-FFT) (1), and expression of native CiC/VIF in chicory taproots (3 and 5), leaves (4 and 6) and flowers (7). S, soluble proteins; CW, cell wall-associated proteins; 3 and 4, first-year chicory plants; 5–7, second-year chicory plants (flowering stage). (b) Developmental changes in CiC/VIF expression in first-year taproots of field-grown chicory (extracts of cell wall-associated proteins). Harvest dates (2007): 1, 14 June; 2, 20 June; 3, 9 July; 4, 27 July; 5, 13 August; 6, 28 August; 7, 12 September; 8, 24 September; 9, 4 October; 10, 17 October; 11, 12 November. For each harvest, taproot extracts from two individual plants were compared. CiC/VIF protein was detected via immunoblot with an antiserum directed against the recombinant CiC/VIF protein.

Functional analysis of ectopically expressed CiC/VIF protein

To investigate the expected inhibitory function of CiC/VIF on invertases, CiC/VIF-containing extracts from N. benthamiana leaves were passed over a Con A matrix (pH 6.8), thereby separating inhibitor (ConA−; flow through) from AI proteins (ConA+). Successful separation of CiC/VIF from endogenous AIs was confirmed by monitoring AI activities in the ConA+ and ConA− fractions. While AI activity was exclusively found in the ConA+ fraction, immunoblot analysis confirmed that CiC/VIF was predominantly present in the ConA− fraction (Fig. 4a). The amount of CiC/VIF in the concentrated ConA− fraction was quantitatively estimated by immunoblot analysis with the CiC/VIF antiserum (calibrated with recombinant CiC/VIF protein expressed in E. coli; data not shown). VI and CWI activities in crude extracts from N. benthamiana leaves were only moderately inhibited by CiC/VIF (Fig. 4b). Similarly, VI activity in soluble protein extracts from chicory hairy root culture was only moderately inhibited by up to 300 ng of inhibitor, but showed a further drop in activity at a higher CiC/VIF concentration (80% inhibition at 500 ng CiC/VIF). Conversely, CWI activity in the salt-eluted protein fraction from the chicory hairy root culture was very sensitive to CiC/VIF, being inhibited by 87 and 93% at 125 and 500 ng CiC/VIF, respectively (Fig. 4b).

Finally, CiC/VIF was investigated for a possible interaction with FAZYs from chicory. When 500 ng of CiC/VIF was used (i.e. the maximum amount of CiC/VIF tested against invertase preparations), activities of FAZYs, i.e. 1-SST, 1-FFT, 1-FEH I and 1-FEH IIa, were not affected (Table 3). This observation confirms the results obtained with invertase inhibitors from nonfructan species (see ‘FAZY activities are not affected by invertase inhibitors from nonfructan plant species’ in Results section), and characterizes CiC/VIF as a bona fide invertase inhibitor, which, based on its localization in taproots (Fig. 3) and its higher activity against CWI than against VI, is likely to represent a cell wall-localized isoform.

Table 3.  Effect of recombinant Cichorium intybus cell wall/vacuolar inhibitor of fructosidase (CiC/VIF) protein on chicory fructan active enzyme (FAZY) activities
 + CiC/VIF
Remaining enzyme activity as % of control
  1. Remaining enzyme activities after preincubation with Nicotiana benthamiana leaf extracts containing 500 ng of CiC/VIF protein are given as a percentage of the control value. The amount of CiC/VIF protein was estimated from a comparison with recombinant CiC/VIF protein expressed in Escherichia coli (not shown). For specific FAZY activities of control incubations, see Table 2. Note that in the case of fructan 1-exohydrolase I (1-FEH I) and IIa (1-FEH IIa), the significant inhibition of background activities (i.e. originating from endogenous invertase activities; see Fig. 1) was accounted for using the corresponding controls. Values are means ± SD (n = 3). 1-SST, sucrose:sucrose 1-fructosyltransferase; 1-FFT, fructan:fructan 1-fructosyltransferase.

1-SST103 (± 2)
1-FFT98 (± 9)
1-FEH I102 (± 14)
1-FEH IIa102 (± 21)

By means of database searches, we were able to clone two additional putative invertase inhibitors from fructan-accumulating species, H. annus and M. acuminata, which have 53 and 24% protein sequence identity with CiC/VIF, respectively. Ectopically expressed HaC/VIF inhibited CiVI and NbVI by 23 and 63%, respectively, under the assay conditions used for CiC/VIF. Again, all FAZY preparations remained unaffected, confirming the results obtained with CiC/VIF. Conversely, MaC/VIF was inactive against all AI and FAZY preparations, possibly as a result of the lower sequence conservation in this monocot species.


Heterologous expression of functional, glycosylated FAZYs in N. benthamiana leaves

It was the ambitious goal of this study to address the general question of whether FAZYs interact with invertase inhibitors, based on their structural relatedness to CWI and VI. For this purpose, we chose to include the entire FAZY spectrum, that is, 1-SST, 1-FFT, 1-FEH I and 1-FEH IIa, from the fructan-accumulating species C. intybus. FAZYs from this species have been extensively studied at the biochemical level (Claessens et al., 1990; Van den Ende & Van Laere, 1996; De Roover et al., 1999) and 1-FEH IIa has been crystallized, providing the first FAZY protein structure (Verhaest et al., 2005b). To cover a representative range of invertase inhibitors, we included five experimentally confirmed inhibitors from nonfructan plants (B. vulgaris BvC/VIF (Eufinger, 2006), N. tabacum NtCIF and NtVIF (Greiner et al., 1998, 1999) and A. thaliana AtC/VIF 1 and 2 (Link et al., 2004)), and, in addition, cloned novel invertase inhibitors from fructan-accumulating species. In expressing all these proteins in the same transient plant expression system, namely A. tumefaciens-mediated transformation of N. benthamiana via leaf infiltration, we opted for a robust system with high expression rates while assuring plant-type glycosylation of recombinant FAZYs. Because of the low or absent FAZY background activities in this nonfructan species (Fig. 1), recombinant enzymes could be used in virtually crude extracts. This novel experimental approach allowed us to perform a fairly comprehensive survey.

FAZYs: evolved from AIs but not prone to regulation by AI inhibitors

FAZYs are structurally related to AIs, and it has been speculated that FAZYs evolved from AIs by a few mutational changes (Ritsema et al., 2006; Le Roy et al., 2007). Consequently, it was hypothesized that invertase inhibitors interact with FAZYs by complex formation and, possibly, by inhibition of enzyme activity. Our study has clearly demonstrated that AI inhibitors from both fructan-accumulating and nonfructan species do not affect FAZY activities (Tables 2, 3). Of the two additional putative invertase inhibitors from fructan-accumulating species (H. annus and M. acuminata) only the H. annus inhibitor showed significant inhibition of CiVI (by 23%) and NbVI (by 63%) under our assay conditions. Again, all FAZY preparations remained unaffected, confirming the results obtained with CiC/VIF. Whether the complete absence of effects on FAZY activities also indicates an absence of inhibitor binding remains to be shown.

Possible reasons for the observed lack of inhibitor activity may be subtle structural differences between AIs and FAZYs. Thus, a comparison of AtcwINV1 and chicory 1-FEH IIa structures revealed important differences, a single amino acid substitution in the active site region (Asp239) being sufficient to seriously affect the ability of AtcwINV1 to hydrolyze sucrose while retaining its enzymatic activity against 1-kestotriose (Le Roy et al., 2007). Similarly, an INV(W161Y; N166S) mutant of onion (Allium cepa) vacuolar invertase harboring a double mutation in the sucrose binding box shows increased transglycosylation activity (Ritsema et al., 2006). Therefore, a few amino acid substitutions are sufficient to change the donor substrate specificity of an AI from sucrose to fructan or to convert a hydrolyzing AI into a transglycosylating fructosyltransferase. Similarly, it may be speculated that small changes in the amino acid composition of inhibitor proteins (e.g. at contact sites with the target protein) might prevent target enzyme binding and inhibition.

It must be emphasized that protein sequence conservation within the PMEI-RP family (a minor part of which consists of invertase inhibitors) is very limited. First, invertase inhibitors from one species vary considerably in protein sequence, NtCIF and NtVIF sharing only 47% sequence identity at protein level (Rausch & Greiner, 2004); secondly, sequence identity is even less conserved between invertase inhibitors and bona fide PMEIs, with identities as low as 20% (Rausch & Greiner, 2004). Individual invertase inhibitors differentially affect VI or CWI (Link et al., 2004; Rausch & Greiner, 2004); however, there is a strict functional separation between invertase inhibitors and PMEIs with respect to their target enzyme class (AI versus pectin methylesterase) (Camardella et al., 2000; Wolf et al., 2003). Thus, the PMEI family includes at least the two subgroups of invertase inhibitors and PME inhibitors. However, as in A. thaliana the PMEI-RP group comprises more than 35 members, with only a limited number being functionally characterized, the possibility cannot be excluded that additional subgroups exist with specificity for other yet to be defined target enzymes, for example FAZYs in fructan-accumulating species. It is too early to speculate on the postulated co-evolution of PMEI-RPs with their target enzyme families, but, as outlined above for the evolutionary relatedness between AIs and FAZYs, FAZY inhibitors, should they exist, are expected to have evolved from invertase inhibitors.

AI inhibitors from fructan-accumulating species: their possible role in fructan biosynthesis

We additionally report here the first cloning of an invertase inhibitor from a fructan-accumulating plant, CiC/VIF from Cichorium intybus. CiC/VIF was experimentally confirmed as an inhibitor of VI and CWI in vitro, but it showed considerably higher affinity for CWI when tested against chicory crude protein extracts (Fig. 4b). Together with the observation that native CiC/VIF was almost exclusively detected in the cell wall-associated protein fraction, these observations strongly argue in favor of CWI being its in vivo target. The physiological role of CiC/VIF is presently unclear, as comprehensive information on CWI (multiple isoforms expected) and CiC/VIF expression and complex formation in vivo is not available. Developmental studies of NtCIF and its target NtCWI in suspension-cultured tobacco cells revealed co-expression of the two enzymes throughout the entire culture period, with persistent complex formation but only a transient inhibitory effect (Krausgrill et al., 1998). In adult tobacco plants, co-expression of NtCIF and NtCWI was only detected in leaves (Greiner et al., 1998). Strong expression of putative CIFs was also observed in cell cultures from other species such as Chenopodium rubrum and Daucus carota (Greiner et al., 2000). CiC/VIF was not expressed in chicory hairy root cultures (results not shown). However, the consistently high expression of CiC/VIF in chicory taproots during the entire growing season suggests that it may indeed play a role in preventing sucrose hydrolysis in the fructan-accumulating phase. Thus, it may be assumed that sucrose imported from the shoot is maintained uncleaved after phloem unloading in the taproot to maintain the high sucrose concentration in the vacuole required for continuous fructan biosynthesis (1-SST Km for sucrose c. 300 mM).


This study has demonstrated that FAZYs are not affected by bona fide invertase inhibitors, despite their close relatedness to AIs. This allows fructan-accumulating plants to post-translationally control their AIs via invertase inhibitors without interfering with FAZY activities. Whether in these plants a novel subgroup of PMEI-RPs has evolved, specifically adapted for FAZY control, remains to be investigated.


We gratefully acknowledge the contribution of Mirsada Kurtisi to the cDNA cloning of CiC/VIF. This research was supported by the BMBF (grants 0315050A and 0315050B to TR and KH, respectively) and Südzucker AG (grants to TR and SG).