Advertisement

Modulation of the Degree and Pattern of Methyl-esterification of Pectic Homogalacturonan in Plant Cell Walls

IMPLICATIONS FOR PECTIN METHYL ESTERASE ACTION, MATRIX PROPERTIES, AND CELL ADHESION*
      Homogalacturonan (HG) is a multifunctional pectic polysaccharide of the primary cell wall matrix of all land plants. HG is thought to be deposited in cell walls in a highly methyl-esterified form but can be subsequently de-esterified by wall-based pectin methyl esterases (PMEs) that have the capacity to remove methyl ester groups from HG. Plant PMEs typically occur in multigene families/isoforms, but the precise details of the functions of PMEs are far from clear. Most are thought to act in a processive or blockwise fashion resulting in domains of contiguous de-esterified galacturonic acid residues. Such de-esterified blocks of HG can be cross-linked by calcium resulting in gel formation and can contribute to intercellular adhesion. We demonstrate that, in addition to blockwise de-esterification, HG with a non-blockwise distribution of methyl esters is also an abundant feature of HG in primary plant cell walls. A partially methyl-esterified epitope of HG that is generated in greatest abundance by non-blockwise de-esterification is spatially regulated within the cell wall matrix and occurs at points of cell separation at intercellular spaces in parenchymatous tissues of pea and other angiosperms. Analysis of the properties of calcium-mediated gels formed from pectins containing HG domains with differing degrees and patterns of methyl-esterification indicated that HG with a non-blockwise pattern of methyl ester group distribution is likely to contribute distinct mechanical and porosity properties to the cell wall matrix. These findings have important implications for our understanding of both the action of pectin methyl esterases on matrix properties and mechanisms of intercellular adhesion and its loss in plants.
      HG
      homogalacturonan
      IDA
      immuno-dot assay
      PME
      pectin methyl esterase
      RG
      rhamnogalacturonan
      N
      newton(s)
      pPME
      plant pectin methyl esterase
      fPME
      fungal pectin methyl esterase
      DE
      degree of methyl-esterification
      TBS
      Tris-buffered saline
      BSA
      bovine serum albumin
      PBS
      phosphate-buffered saline
      PL
      endo-pectin lyase
      PG II
      endo-polygalacturonase II
      ELISA
      enzyme-linked immunosorbent assay
      ciELISA
      competitive inhibition enzyme-linked immunosorbent assay
      PIPES
      1,4-piperazinediethanesulfonic acid
      CDTA
      cyclohexanediamine N,N,N′,N′-tetraacetic acid
      The load-bearing components of primary cell walls, principally the cellulose and hemicellulose polysaccharide network, are embedded in a pectic matrix that is structurally complex and heterogeneous. The pectic matrix contributes to both the physical integrity and physiological status of cell walls, but the functional implications of the structural complexity of this matrix are poorly understood. Typically, heterogeneous populations of pectic polymers are present in primary cell walls, and our present understanding is that all essentially consist of galacturonan backbones with or without various side chain additions (
      • Pilnik W.
      • Phillips G.O.
      • Williams P.A.
      • Wedlock D.J.
      ,
      • O'Neill M.A.
      • Albersheim P.
      • Darvill A.
      • Dey P.M.
      ,
      • Mohnen D.
      • Barton D.
      • Nakanishi K.
      • Meth-Cohn O.
      ). Backbone domains consist of either contiguous 1,4-linked α-d-galacturonic acid (homogalacturonan, HG)1 or repeats of the disaccharide (→4)-α-d-GalA-(1→2)-α-l-Rha-(1→) (rhamnogalacturonan, RG). GalA residues in HG may be methyl-esterified, acetylated, and/or substituted with xylose or apiose (
      • O'Neill M.A.
      • Albersheim P.
      • Darvill A.
      • Dey P.M.
      ,
      • Mohnen D.
      • Barton D.
      • Nakanishi K.
      • Meth-Cohn O.
      ,
      • Schols H.A.
      • Bakx E.J.
      • Schipper D.
      • Voragen A.G.J.
      ). Oligosaccharide side chains may be attached to both RG and HG domains to form the branched domains known as RG-I and RG-II, respectively (
      • O'Neill M.A.
      • Albersheim P.
      • Darvill A.
      • Dey P.M.
      ,
      • Mohnen D.
      • Barton D.
      • Nakanishi K.
      • Meth-Cohn O.
      ,
      • O'Neill M.A.
      • Warrenfeltz D.
      • Kates K.
      • Pellerin P.
      • Doco T.
      • Darvill A.G.
      • Albersheim P.
      ,
      • Vidal S.
      • Doco T.
      • Williams P.
      • Pellerin P.
      • York W.S.
      • O'Neill M.A.
      • Glushka J.
      • Darvill A.G.
      • Albersheim P.
      ).
      The HG domain of the pectic network is implicated in influencing a range of cell wall properties that impact upon cell expansion, cell development, intercellular adhesion, and defense mechanisms. Stretches of HG with un-esterified GalA residues can associate by calcium cross-linking (
      • Jarvis M.C.
      ). Such association promotes the formation of supramolecular pectic gels, which are important in controlling the porosity and mechanical properties of cell walls and contribute to the maintenance of intercellular adhesion (
      • Knox J.P.
      ,
      • Carpita N.C.
      • Gibeaut D.M.
      ). Plant cells are adhered by contact across middle lamellae, which are HG-rich regions of the cell wall developed from cell plates formed at cytokinesis. In plants, cell adhesion is a default state and cell separation an active process that is under developmental control (
      • Knox J.P.
      ,
      • Roberts J.A.
      • Whitelaw C.A.
      • Gonzalez-Carranza Z.H.
      • McManus M.T.
      ). In addition to the roles of the HG polysaccharide domain, HG-derived oligogalacturonides generated by pectinolytic cleavage are involved in signaling processes during development and in defense responses to plant pathogens (
      • Eberhard S.
      • Doubrava N.
      • Marfà V.
      • Mohnen D.
      • Southwick A.
      • Darvill A.
      • Albersheim P.
      ,
      • Dumville J.C.
      • Fry S.C.
      ,
      • Moerschbacher B.M.
      • Mierau M.
      • Graeβner B.
      • Noll U.
      • Mort A.J.
      ).
      It is thought that HG is highly methyl-esterified when exported into cell walls and is subsequently de-esterified by the action of pectin methyl esterases (PMEs) in the cell wall (
      • Mohnen D.
      • Barton D.
      • Nakanishi K.
      • Meth-Cohn O.
      ,
      • Zhang G.F.
      • Staehelin L.A.
      ). PME genes occur in multigene families and encode isoforms with differing action patterns with respect to the removal of methyl esters. However, the specific functions of PME populations in the context of cell expansion and other processes are not well understood (
      • Bordenave M.
      • Breton C.
      • Goldberg R.
      • Huet J.C.
      • Perez S.
      • Pernollet J.C.
      ,
      • Catoire L.
      • Pierron M.
      • Morvan C.
      • Hervé du Penhoat C.
      • Goldberg R.
      ,
      • Micheli F.
      • Holliger C.
      • Goldberg R.
      • Richard L.
      ). Methyl esters can be distributed in diverse patterns along HG chains and it is clear that the action patterns of plant PMEs (pPMEs) can be influenced by local cell wall pH, the existing balance of methyl and free carboxyl groups on HG substrates, and metal ion concentration (
      • Catoire L.
      • Pierron M.
      • Morvan C.
      • Hervé du Penhoat C.
      • Goldberg R.
      ,
      • Nari J.
      • Noat G.
      • Diamantidis G.
      • Woudstra M.
      • Ricard J.
      ,
      • Moustacas A.M.
      • Nari J.
      • Borel M.
      • Noat
      • Ricard J.
      ,
      • Denès J.-M.
      • Baron A.
      • Renard C.M.G.C
      • Péan C.
      • Drilleau J.-F.
      ,
      • Mort A.J.
      • Qiu F.
      • Maness N.O.
      ,
      • Daas P.J.H.
      • Voragen A.G. J
      • Schols H.A.
      ). In addition to plants, some fungi and bacteria also produce PMEs, and it is significant that the action patterns of plant and fungal PMEs are thought to be different. It is generally proposed that pPMEs remove methyl esters in a processive blockwise fashion (single chain mechanism), giving rise to long contiguous stretches (blocks) of un-esterified GalA residues in HG domains of pectin (
      • Massiot P.
      • Perron V.
      • Baron A.
      • Drilleau J.F.
      ,
      • Limberg G.
      • Körner R.
      • Bucholt H.C.
      • Christensen T.M.I.E
      • Roepstorff P.
      • Mikkelsen J.D.
      ,
      • Limberg G.
      • Körner R.
      • Christensen T.M.I.E
      • Buchholt H.C.
      • Roepstorff P.
      • Mikkelsen J.D.
      ). In contrast, the action of fungal PMEs (fPMEs) is generally regarded as random (or multiple chain mechanism), resulting in the de-esterification of single GalA residues per enzyme/substrate interaction (
      • Massiot P.
      • Perron V.
      • Baron A.
      • Drilleau J.F.
      ,
      • Limberg G.
      • Körner R.
      • Bucholt H.C.
      • Christensen T.M.I.E
      • Roepstorff P.
      • Mikkelsen J.D.
      ,
      • Limberg G.
      • Körner R.
      • Christensen T.M.I.E
      • Buchholt H.C.
      • Roepstorff P.
      • Mikkelsen J.D.
      ). However, the precise nature of the action patterns of fPME and pPME are far from clear, and some pPMEs appear to have the capacity to remove a limited number of methyl esters per reaction, giving rise to short un-esterified blocks (
      • Denès J.-M.
      • Baron A.
      • Renard C.M.G.C
      • Péan C.
      • Drilleau J.-F.
      ). Moreover, recently developed procedures for the enzymatic fingerprinting of pectic fragments have indicated that the action patterns of fPMEs are non-blockwise, but probably not in fact random (
      • Limberg G.
      • Körner R.
      • Bucholt H.C.
      • Christensen T.M.I.E
      • Roepstorff P.
      • Mikkelsen J.D.
      ). In this report, the action pattern of an orange pPME is referred to as blockwise, whereas the action pattern of an Aspergillus fPME and the chemical process of base catalysis are referred to as non-blockwise.
      Previous investigations of the action patterns of PMEs have focused on the action of isolated PMEs on pectin substrates in vitro(
      • Catoire L.
      • Pierron M.
      • Morvan C.
      • Hervé du Penhoat C.
      • Goldberg R.
      ). However, the use of anti-HG monoclonal antibodies with appropriate binding specificities allows the products of de-esterification processes to be analyzed in planta in the context of cell wall architecture and cell development. We have used monoclonal antibodies with differing binding requirements with respect to the distribution of methyl ester groups to investigate the spatial regulation of HG with blockwise and non-blockwise methyl group distributions in pea. The monoclonal antibody PAM1 (which binds specifically to long stretches of un-esterified HG produced most readily by blockwise de-esterification) and JIM5 (which binds to a range of partially methyl-esterified HG domains) have been described previously (
      • Knox J.P.
      • Linstead P.J.
      • King J.
      • Cooper C.
      • Roberts K.
      ,
      • Willats W.G.T.
      • Gilmartin P.M.
      • Mikkelsen J.D.
      • Knox J.P.
      ,
      • Willats W.G. T
      • Limberg G.
      • Buchholt H.C.
      • van Alebeek G.J.
      • Benen J.
      • Christensen T.M.I.E.
      • Visser J.
      • Voragen A.
      • Mikkelsen J.D.
      • Knox J.P.
      ). Non-blockwise de-esterification of HG has been analyzed using a monoclonal antibody (LM7) described for the first time in this report. In addition, we have investigated the functional implications of varying degrees and patterns of methyl-esterification of HG in the context of the physical and physiological properties of calcium-HG gels. Taken together, the results indicate that wall-based pPMEs with a range of action patterns can produce HG with non-blockwise and blockwise distributions of methyl esters at discrete cell wall microdomains, resulting in distinct spatially regulated matrix properties.

      EXPERIMENTAL PROCEDURES

       Monoclonal Antibodies

      Monoclonal antibody LM7 was generated using hybridoma technology subsequent to the immunization of a group of rats with lime pectin with a degree of methyl-esterification (DE) of 22.9%, a degree of amidation of 27.3% and an average molecular mass of 84 kDa. Immunization, hybridoma preparation, and cloning procedures were performed as described previously (
      • Willats W.G. T
      • Marcus S.E.
      • Knox J.P.
      ). LM7 was selected by differential immuno-dot assay (IDA) screens of a series of pectin samples differing in DE with blockwise and non-blockwise patterns of de-esterification (see below). Monoclonal antibodies JIM5 and PAM1 were produced by hybridoma and phage display technologies, respectively (
      • Knox J.P.
      • Linstead P.J.
      • King J.
      • Cooper C.
      • Roberts K.
      ,
      • Willats W.G.T.
      • Gilmartin P.M.
      • Mikkelsen J.D.
      • Knox J.P.
      ,
      • Willats W.G. T
      • Limberg G.
      • Buchholt H.C.
      • van Alebeek G.J.
      • Benen J.
      • Christensen T.M.I.E.
      • Visser J.
      • Voragen A.
      • Mikkelsen J.D.
      • Knox J.P.
      ).

       Production of Lime Pectins with Defined DE

      A series of lime pectins with different patterns (blockwise and non-blockwise), and defined DE were prepared by enzymatic and chemical treatments of a commercial highly methyl-esterified (81%) lime pectin (E81, GRINDSTED™ Pectin URS 1200) as described previously (
      • Limberg G.
      • Körner R.
      • Bucholt H.C.
      • Christensen T.M.I.E
      • Roepstorff P.
      • Mikkelsen J.D.
      ). Briefly, one series was produced by blockwise de-esterification of E81 with a pPME isolated from orange peel (P-series), while another series was produced by non-blockwise de-esterification of E81 with a fPME fromAspergillus niger (F-series). A further set of samples was also produced by non-blockwise de-esterification of E81 by base catalysis (B-series). A sample of completely de-esterified pectin (E0) was prepared by treatment with fPME followed by base catalyzed de-esterification.

       Digestion of Pectin B34 with endo-Pectin Lyase (PL) and endo-Polygalacturonase II (PG II)

      Pectin sample B34 (prepared by base catalysis and with a DE of 34%) was digested with PL (EC3.2.1.15) or PG II (EC 4.2.2.10), both from Aspergillus niger. B34 was dissolved in 50 mm NaOAc (pH 5.0 for PL and pH 4.2 for PG II) at a concentration of 5 mg/ml by overnight rocking at room temperature. 0.1 unit of PL or 0.2 unit of PG II was added to 1 ml of the above pectin solution and in both cases incubated at room temperature for 20 h. The reaction was stopped by boiling for 5 min.

       Competitive Inhibition ELISAs (ciELISAs)

      The effect of PL and PG II digestion of B34 on the binding of LM7 was assessed by ciELISAs with untreated B34 as the immobilized antigen. Untreated B34 (50 μg/ml in Tris-buffered saline (TBS)) was coated (100 μl/well) overnight at 4 °C onto microtiter plates (Maxisorp, Nunc, Denmark). The coating solution was removed, and plates were blocked at room temperature with 3% bovine serum albumin in TBS (3%BSA/TBS) for 2 h (200 μl/well). Following washing, competitor solutions (untreated B34, B34 digested with PL, and B34 digested with PG II) were applied (100 μl/well) as 5-fold serial dilutions in 3%BSA/TBS. All competitor solutions also contained LM7 at a final level of 1/100 dilution of hybridoma supernatant (corresponding to ∼90% maximal binding on antibody capture ELISAs). After 2 h of incubation, plates were washed and secondary antibody (anti-rat IgG horseradish peroxidase conjugate, Sigma, Poole, United Kingdom) diluted 1/1000 in 3%BSA/TBS applied (100 μl/well) and incubated for 2 h at room temperature. After washing, plates were developed with a tetramethyl benzidine-based substrate (150 μl/well). After stopping the reaction with 2 m H2SO4 (35 μl/well), absorbances were read at 450 nm. Concentrations of competitors resulting in 50% inhibition (IC50) of antibody binding were determined. Values from controls with no competitor were taken as 0% inhibition of antibody binding, and values from controls with no LM7 antibody represented 100% inhibition of binding. In some cases, CaCl2 or MgCl2 were added to competitor solutions to a maximum level of 1 mm.

       Immuno-dot Assays

      Pectins were dissolved in water to a concentration of 5 mg/ml or 10 mg/ml and applied as 1-μl aliquots to nitrocellulose (Scheicher & Schuell) in a 5- or 10-fold dilution series. Nitrocellulose membranes were air-dried at room temperature for at least 30 min. After blocking for 1 h in phosphate-buffered saline (PBS) containing 5% (w/v) fat-free milk powder (5%M/PBS), membranes were incubated for 1 h in primary antibodies diluted in 5%M/PBS. JIM5 and LM7 were used as 1/10 dilutions of hybridoma supernatants while PAM1 was used at a concentration of ∼1 × 1011 phage particles/ml (∼1/100 dilution of phage prepared by polyethylene glycol precipitation; Ref.
      • Willats W.G.T.
      • Gilmartin P.M.
      • Mikkelsen J.D.
      • Knox J.P.
      ). In all cases, membranes were incubated in primary antibodies for 1.5 h. After washing, membranes were incubated for 1.5 h in secondary antibody (anti-rat horseradish peroxidase conjugate (for JIM5 and LM7) (Sigma, Poole, United Kingdom) or anti-M13 horseradish peroxidase conjugate (for PAM1) (Amersham Pharmacia Biotech) diluted 1/1000 in 5%M/PBS. Membranes were briefly washed prior to development in substrate solution (25 ml of de-ionized water, 5 ml of MeOH containing 10 mg/ml 4-chloro-1-naphthol, 30 μl 6% (v/v) H2O2). In some cases, immobilized pectin samples were chemically de-esterified by incubation of membranes in 0.1 mNa2CO3 for 1 h prior to processing.

       Immunolabeling of Plant Material

      Pea (Pisum sativum L. cv. Avola) seeds were imbibed overnight in tap water, sown in sterile vermiculite, and grown for 7–15 days. Regions (0.5 cm long) of stem, petiole, or root were excised and sectioned by hand to a thickness of ∼100–300 μm. Sections were placed immediately in fixative consisting of 4% paraformaldehyde in 50 mm PIPES, 5 mm MgSO4, and 5 mm EGTA. Following 30 min of fixation, sections were washed in the PIPES buffer and then incubated for 1 h in primary antibody diluted in 5%M/PBS. JIM5 and LM7 were used as 10- and 3-fold dilutions of hybridoma supernatants, respectively. PAM1 was used at a concentration of ∼5 × 1011 phage particles/ml (∼ 1/20 dilution of phage prepared by polyethylene glycol precipitation; Ref.
      • Willats W.G.T.
      • Gilmartin P.M.
      • Mikkelsen J.D.
      • Knox J.P.
      ). Sections were washed by gently rocking in PBS prior to incubation for 1 h in secondary antibody. For visualization of LM7 and JIM5 binding, the secondary antibody was anti-rat IgG coupled to fluorescein isothiocyanate (Sigma). For visualization of PAM1 binding, a secondary antibody was prepared by conjugating an anti-M13 antibody (Amersham Pharmacia Biotech) to fluorescein isothiocyanate using a protein conjugation kit (Sigma). All secondary antibodies were used at dilutions of 1/100 in 5%M/PBS. After washing in PBS, sections were mounted in anti-fade agent (Citifluor, Agar Scientific) and examined on a microscope equipped with epifluorescence illumination (Olympus BH-2). In some cases, hand sections were chemically de-esterified by incubation in 0.1 m Na2CO3 for 1 h prior to processing.
      In certain cases plant material was embedded in resin for electron microscopy. Regions (2 mm long) of stem were fixed in 2.5% (w/v) glutaraldehyde in 0.1 m sodium phosphate buffer, pH 7.2, for 2 h at 4 °C, then washed extensively in 0.1 msodium phosphate buffer. Material was then post-fixed in 0.1% (w/v) osmium tetroxide in 0.1 m sodium phosphate buffer, for 1 h at 4 °C, washed extensively with 0.1 m sodium phosphate buffer, and then dehydrated in an ethanol series. Dehydrated material was infiltrated with resin (LR White) (London Resin, Reading, United Kingdom), then placed in gelatin capsules containing resin and allowed to polymerize at 37 °C for 5 days. Material used for transmission electron microscopy but not immunogold labeling, was stained en bloc with 4% (w/v) uranyl acetate in distilled water overnight at 4 °C, then washed extensively with water before being dehydrated and embedded as described above.
      For immunogold labeling, sections obtained from resin-embedded material (∼0.1 μm thick) were incubated in 3%BSA/TBS for 30 min. Sections were then incubated in a solution containing JIM5 or LM7 diluted 1/10 in 3%BSA/TBS for 1.5 h. The sections were washed with 3%BSA/TBS, and then incubated in secondary antibody (anti-rat monoclonal antibody conjugated to 10 nm colloidal gold (Sigma) diluted 1/40 in 3%BSA/TBS for 1.5 h. Section were washed with TBS and post-stained with 4% uranyl acetate in distilled water for 15 min, and then washed extensively with distilled water. Sections were observed with an electron microscope (1200ex, JEOL, Tokyo, Japan). All incubations were at room temperature.

       Calcium-mediated Gelation of Lime Pectin Samples

      A subset of the P-, F-, and B-series of lime pectins that contained the epitopes recognized by LM7 or PAM1 to various levels were selected for analysis of the effects of the degree and pattern of methyl-esterification on the physical properties of calcium-mediated gels. Gels were prepared from E0, F11, F31, P41, F43, and B34, essentially as described elsewhere (
      • MacDougall A.J.
      • Needs P.W.
      • Rigby N.M.
      • Ring S.G.
      ). Pectin solutions (2% (w/v) in de-ionized water) were prepared with gentle rocking at 4 °C for at least 18 h. For casting gels, 900 μl of pectin solution was transferred to a 2-ml syringe (with an internal diameter of 8.6 mm) from which the nozzle end had been removed. 67 μl of 500 mm CaCl2 was added as a layer to the top of the pectin solution (to give a final CaCl2 concentration of 35 mm) and the cut end of the syringe sealed with tape. Gels were left to equilibrate for 24–48 h at 4 °C. Polymerized gels were removed using the syringe plunger and cut to a uniform height using a custom-made nylon cutting block with guide slots.

       Rheological Testing of Calcium-mediated Pectin Gels

      Gel samples prepared as described above were subjected to compressive tests to determine their elasticity under low strain and their yield points under high strain. Compression tests were performed using an in-house texture measuring device, known as the “Ministron,” constructed in the University of Leeds Department of Food Science instrument workshop. This device was used to analyze the controlled compression of gel samples between two parallel stainless steel plates at a precisely defined speed. The lower plate, on which the sample rested, is mounted on a high precision load cell (Maywood Instruments, Basingstoke, United Kingdom). The gap between the plates and the force, F, on the load cell are electronically logged throughout the experiment at a suitable frequency. For all experiments, at least four samples from different gel preparations were analyzed.
      Gels were compressed by reducing the gap between the plates at a rate of 0.1 mm s1. The elasticity, E, of the gels in the low strain region was calculated from Equation1.
      E=F/(A·s)
      Equation 1


      s is the strain, and A is sample cross-sectional area. The strain is given by Equation 2.
      s=Δh/h0
      Equation 2


      h 0 is the initial height of the sample, and Δh is the change in height due to compression. The line of best fit to the plot of F versus s, in the region 0.1 < s < 0.2, was used to calculate E. Below this range samples did not always compress uniformly, due to not having exactly parallel sides, whereas at higher strains there were significant increases in cross-sectional area and/or water loss from the sample, contributing to non-linearF versus s plots. Up tos = 0.2 the change in A was negligible for all samples, and for 0.1 < s < 0.2 the line of best fit to the data always had a regression coefficient of at least 0.95.

       Analysis of the Water Holding Capacity of Calcium-mediated Pectin Gels under Compression

      It was observed that there were significant differences in the water expelled from gel samples during compression. The final percentage of water lost was calculated from the change in the volume of the gels after compression up to a force of 9.82 N (E0, F11, and F31) or by the force that resulted in yielding (P41, F43, and B34). The yield point was taken as the point when the gradient of F versus s became negative. Yielding was sometimes also accompanied by visible splitting of the gel piece.

       Determination of the Porosity of Calcium-mediated Pectin Gels

      The porosity of gels to protein was determined by incubating gels in a solution of BSA and measuring the incorporation of protein into gels over time. For each time point, two gel blocks (prepared as described previously) above were incubated with gentle rocking in 2 ml of BSA solution (5 mg/ml in de-ionized water). Following incubation, the two gel blocks were washed briefly in de-ionized water, briefly blotted dry on filter paper to remove surface liquid, and incubated with gentle rocking in 2 ml of 50 mm calcium chelator (CDTA) (pH 7) until gels were completely dissolved (∼30 min). The protein concentration of the solutions were then analyzed using Bradford protein assays. For each pectin gel sample, four replicates of the protocol described were analyzed for each time point.

      RESULTS

       Monoclonal Antibody LM7 Recognizes an Epitope of HG Produced by Non-blockwise De-esterification

      Monoclonal antibody LM7 was generated subsequent to immunization with a lime pectin (containing 88.3% GalA) and selected by IDA screening on the basis of its specific binding to a subset of F-series and B-series pectins that have non-blockwise patterns of methyl-esterification as shown in Fig.1 a. The binding of the previously characterized anti-HG JIM5 (
      • Knox J.P.
      • Linstead P.J.
      • King J.
      • Cooper C.
      • Roberts K.
      ,
      • Willats W.G. T
      • Limberg G.
      • Buchholt H.C.
      • van Alebeek G.J.
      • Benen J.
      • Christensen T.M.I.E.
      • Visser J.
      • Voragen A.
      • Mikkelsen J.D.
      • Knox J.P.
      ) and the phage antibody PAM1 (
      • Willats W.G.T.
      • Gilmartin P.M.
      • Mikkelsen J.D.
      • Knox J.P.
      ) to the same pectin samples are shown for comparison (Fig. 1,b and c). Both the degree and the pattern of de-esterification influence the capacity of these antibodies to bind to HG domains. LM7 did not bind at the highest level tested (1 μg) to any of the P-series pectins or to F-series pectins with DE of 76%, 69%, or 11% (Fig. 1 a). LM7 did bind to F-series pectins with DE from 58% to 31% with increasing avidity, and LM7 bound to F31 with a detection limit of <0.2 μg (Fig. 1 a). LM7 bound weakly, at the highest level tested (1 μg), to a B-series sample with a DE of 43% and bound to B-series pectins with DEs of 34 and 15% with detection limits of <0.2 μg. The binding of LM7 to all samples was abolished when blots were de-esterified by treatment with 0.1m Na2CO3 prior to labeling (data not shown). The binding profile of LM7 indicates that it is specific for a partially methyl-esterified domain of HG and that its epitope is most readily produced by the non-blockwise de-esterification processes such as that produced by fPME action and base catalysis. In contrast, PAM1 binds to long (>30 residues) contiguous stretches of de-esterified GalA residues produced by the blockwise action of pPME as shown in Fig. 1 b. However, un-esterified blocks are also produced if enough non-blockwise de-esterification occurs and the PAM1 epitope is also produced by extensive de-esterification by fPME or base catalysis as indicated by binding to F11 and B15, respectively (Fig.1 b). The optimal binding requirements of JIM5 are not fully defined, and JIM5 has the capacity to bind to a wide range of HG epitopes with varying degrees and patterns of methyl-esterification as shown in Fig. 1 c and as discussed elsewhere (
      • Willats W.G. T
      • Limberg G.
      • Buchholt H.C.
      • van Alebeek G.J.
      • Benen J.
      • Christensen T.M.I.E.
      • Visser J.
      • Voragen A.
      • Mikkelsen J.D.
      • Knox J.P.
      ).
      Figure thumbnail gr1
      Figure 1The monoclonal antibody LM7 binds to a subset of F-series and B-series model pectins. Figure shows IDA of LM7 (a), PAM1 (b), and JIM5 (c) binding to a series of lime pectin samples with defined degrees and different patterns of methyl-esterification. The series was produced by the de-esterification of a common high ester pectin sample with a DE of 81% by digestion with a pPME (P-series), an fPME (F-series), or by base catalysis (B-series). A completely de-esterified pectin sample (E0) was produced by digestion of E81 with fungal pectin methyl esterase followed by base catalysis. pPME removes contiguous methyl groups from relatively long stretches of HGHG, resulting in a blockwise de-esterification, while treatment with fPME and base results in non-blockwise de-esterification. All samples were applied to nitrocellulose in dilution series as indicated.

       The Partially Methyl-esterified Epitope Recognized by LM7 Is Degraded by the Action of Both Endo-polygalacturonase and Pectin Lyase

      In order to explore the structure of LM7 epitope further, its susceptibility to digestion by PG II and PL was assessed. The products of enzymatic digestion were analyzed by ciELISAs using untreated B34 as the immobilized antigen as shown in Fig.2. The use of ciELISAs allowed the binding of LM7 to digest fragments to be analyzed in solution, rather than in an immobilized state, as would be the case for IDAs. This is important because the binding of pectic fragments to nitrocellulose sheets is related to fragment size. For example, oligogalacturonides with degrees of polymerization less than 15 are not immobilized effectively onto nitrocellulose sheets (data not shown).
      Figure thumbnail gr2
      Figure 2The LM7 epitope is degraded by polygalacturonase and pectin lyase. Competitive inhibition ELISAs of the effects of pectinolytic digestion on the binding in solution of LM7 to pectin sample B34. Untreated B34 was used as the immobilized antigen. The binding of LM7 to B34 was assessed for untreated B34 (B34 NT), B34 completely digested with endo-polygalacturonase II (B34 PG II) and for B34 treated with pectin lyase (B34 PL).
      When untreated B34 was used as a soluble competitor in competitive inhibition ELISAs, 20 μg/ml was required to achieve a 50% inhibition (IC50) of LM7 binding. LM7 binding to B34 was abolished entirely following complete digestion with PG II and PL, and the digestion fragments failed to produce any significant inhibition of LM7 binding even at the highest level used (1 mg/ml) as shown in Fig. 2. The PG II used from A. niger has an absolute requirement for de-esterified GalA residues to be present at both sides of the cleavage position (subsites +1 and −1). Moreover, optimal cleavage occurs where subsite +2 is de-esterified, whereas whether or not subsites −2 and −3 are de-esterified appears to be less critical (
      • Limberg G.
      • Körner R.
      • Bucholt H.C.
      • Christensen T.M.I.E
      • Roepstorff P.
      • Mikkelsen J.D.
      ). In contrast, PL cleaves optimally in regions of HG that are fully methyl-esterified (
      • Limberg G.
      • Körner R.
      • Bucholt H.C.
      • Christensen T.M.I.E
      • Roepstorff P.
      • Mikkelsen J.D.
      ). The susceptibility of the epitope recognized by LM7 to both PGII and PL cleavage, and the profile of LM7 binding in IDAs (Fig. 1), indicate that the LM7 epitope contains both contiguous methyl-esterified GalA residues and contiguous un-esterified GalA residues.
      Confirmation of epitope structure can be obtained by demonstration of binding to oligosaccharides in ciELISAs, and this generally indicates carbohydrate epitope sizes of 4 to 6 sugars (
      • Willats W.G.T.
      • Gilmartin P.M.
      • Mikkelsen J.D.
      • Knox J.P.
      ,
      • Willats W.G. T
      • Limberg G.
      • Buchholt H.C.
      • van Alebeek G.J.
      • Benen J.
      • Christensen T.M.I.E.
      • Visser J.
      • Voragen A.
      • Mikkelsen J.D.
      • Knox J.P.
      ). Fully methyl-esterified and fully un-esterified oligogalacturonides with degrees of polymerization up to 8, prepared as described elsewhere (
      • Willats W.G. T
      • Limberg G.
      • Buchholt H.C.
      • van Alebeek G.J.
      • Benen J.
      • Christensen T.M.I.E.
      • Visser J.
      • Voragen A.
      • Mikkelsen J.D.
      • Knox J.P.
      ), were not effective inhibitors of LM7 binding in ciELISAs (data not shown). In order to characterize the LM7 epitope fully, a range of oligogalacturonides with intermediate degrees and defined patterns of methyl-esterification would be required, and currently these are not available. Nonetheless, it is clear that LM7 binds to a HG domain of pectin that contains both un-esterified and methyl-esterified GalA residues, that this structure is generated most readily by non-blockwise de-esterification processes, and that the epitope is distinct from any other previously characterized HG epitope.

       LM7 Binding Is Not Dependent on, nor Reduced by, Calcium-mediated Chain Association of HG Domains

      The possible effects of the formation of conformational HG structures on LM7 binding were investigated by addition of divalent cations to ciELISA assays. The presence of calcium or magnesium at up to 1 mm had no effect on LM7 binding to either immobilized or soluble pectin sample F31, as shown in Fig. 3. Levels of calcium above 1 mm resulted in gel formation and disruption of the assay. Surface labeling of calcium pectate gel blocks (prepared as described previously) indicated that LM7 binding was retained when HG was cross-linked via calcium (data not shown). Taken together, these results indicate that LM7 binding is neither dependent on, nor abolished by, calcium-mediated HG chain association of HG domains.
      Figure thumbnail gr3
      Figure 3LM7 binding is not dependent on, nor abolished by, calcium-mediated HG chain association. The binding of LM7 to F31 was assessed in the presence of CaCl2 and MgCl2 (LM7+CaCl2+F31 andLM7+MgCl2+F31, respectively). Control samples (LM7+CaCl2 and LM7+MgCl2) without F31 in the solution were included to assess the effects of calcium and magnesium ions on LM7 binding to the immobilized antigen.

       The Partially Methyl-esterified Epitope Recognized by LM7 Occurs in Discrete Micro-domains of Primary Cell Walls

      The partially methyl-esterified epitope recognized by LM7 was found to be abundant in plant tissues. The epitope was most readily visualized in plant materials that had undergone preparations that maintained maximum antigenicity, i.e. hand-cut sections or cryosections of non-embedded material. The epitope recognized by LM7 appears to be unstable in pectin preparations and in plant material. The capacity of F31 and B34 pectin samples to be recognized by LM7 was gradually lost when these samples were stored as frozen solutions (data not shown). Similarly, the epitope was lost from plant materials when they had undergone extensive preparation, such as resin-embedding for immunocytochemistry. As discussed, LM7 binding is critically dependent on DE being with a certain range, and the instability of the epitope is probably due to de-esterification occurring during freeze/thawing or processing.
      The distribution of the LM7 epitope was examined most extensively in pea seedlings (a system that is amenable to hand-sectioning) and immunofluorescent labeling of sections indicated that the LM7 epitope was restricted to discrete regions of cell walls in the roots, stems, and leaves of seedlings. The distribution of the LM7 epitope in pith parenchyma cells in a transverse section of a pea stem internode is shown in Fig. 4 a. In this region, most cell junctions are expanded to some extent to form intercellular spaces. The LM7 epitope was restricted to the region of the cell wall lining the intercellular spaces between the parenchyma cells, and no LM7 epitope was detected in other regions of the cell walls in this tissue. Fig. 4 d shows a higher magnification of an individual intercellular space and shows that the LM7 epitope was particularly abundant at the corners of the intercellular space,i.e. the point between adherent and separated cell walls. For comparison, the labeling patterns of JIM5 and PAM1 on equivalent sections are shown in Fig. 4 (panels b ande and panels c and f, respectively). JIM5 bound to all primary cell walls (Fig.4 b). At higher magnification (Fig. 4 e), the increased abundance of the JIM5 epitope in the region of the cell wall lining the intercellular spaces and in the region of the wall closest to the plasma membrane was evident. The blockwise de-esterified HG epitope recognized by PAM1 occurred in discrete cell wall domains (Fig.4, c and f). Like LM7 and JIM5, PAM1 bound to material in the cell wall lining intercellular spaces and, like JIM5, did not bind to the central regions of the cell walls (including the middle lamellae). In contrast to LM7, PAM1 also bound to the region of the wall closest to the plasma membrane, but, unlike JIM5, PAM1 labeling was absent from inner regions of the cell wall adjacent to the intercellular space (Fig. 4 f). The reason for the apparently thicker cell walls and more diffuse nature of labeling obtained using PAM1 is due to the fact that PAM1 is a phage display antibody and large (∼800 nm long) intact phage particles were used as the primary stage antibody.
      Figure thumbnail gr4
      Figure 4The HG epitope recognized by LM7 occurs at cell junctions in plant tissues. Figure shows micrographs of the binding of LM7 (a and d), JIM5 (b and e), and PAM1 (c andf) to the cell walls of pea stem cortical cells as visualized by immunofluorescent labeling. All sections were hand-cut transverse sections of pea stem. Arrowheads andarrows in d–f indicate the corners and linings of intercellular spaces, respectively. The double arrowhead in f indicates a region of cell wall close to the plasma membrane. Scale bars ina-c = 100 μm; scale bars ind–f = 20 μm.
      The localization of the LM7 epitope at intercellular spaces was consistent throughout all tissues in the pea stem, and the relationship between the occurrence of the LM7 epitope and the formation of intercellular space was examined in more detail by examination of cortical parenchyma tissue. The immunolabeling of a small non-expanded intercellular space without air, occurring in the cortical parenchyma closer to the epidermis, and larger intercellular air spaces that occur between larger parenchyma cells toward the center of the stem are shown in Fig. 5 (panel aand panels b and c respectively. Fig.5 (a–c) shows dual labeling of sections with LM7 and the cellulose-binding fluorescent probe, calcofluor. At non-expanded junctions, LM7 bound to all of the developing space (which at this stage is filled with expanded middle lamellae) but did not bind to any other regions of the cell wall (Fig. 5 a). At larger, air-filled junctions, the LM7 epitope was most abundant at the corners of the triangular intercellular spaces as viewed in the sections shown in Fig. 5 (b and c). The epitope recognized by LM7 is therefore most abundant in regions of the expanded middle lamellae at the point of separation of cell walls. The ultrastructure of intercellular spaces was investigated by transmission electron microscopy. However, as discussed above, the epitope recognized by LM7 is prone to instability and, when stem material was resin-embedded for immunofluorescent or immunogold detection, the LM7 epitope could not be detected. However, the fine structures of comparable intercellular junctions between parenchyma cells are shown in Fig. 5(d–f). These microgaphs indicate that, as air spaces form, there is an accumulation of darkly staining material at the corners of the junction (Fig. 5, e and f). The position of this material appears to correspond to regions of the wall containing the LM7 epitope as localized by immunofluorescent labeling of non-embedded material (Fig. 5, b and c). Pectic material typically stains darkly in the transmission electron microscopy staining protocol used, and immunogold labeling with JIM5 indicated the presence of HG at this position (Fig. 5 f,inset).
      Figure thumbnail gr5
      Figure 5The LM7 epitope is present throughout intercellular space formation in pea stem parenchyma. a–c, micrographs showing the dual labeling of the cell walls of pea stem cortical parenchyma cells with LM7 (immunofluorescent labeling, green) and the cellulose-binding fluorophor calcofluor (blue) showing the positions HG epitopes and cellulose in relation to intercellular spaces at different stages of formation. df, transmission electron micrographs of approximately equivalent cell junctions to those in a–c. The junction shown ind consists entirely of cell wall material, whereas junctions at positions near the middle of the cortex are progressively separated resulting in the formation of intercellular spaces (e and f). The darkly stained material accumulated at the corners of expanding air spaces (e andf) contains HG, as indicated by immunogold labeling with JIM5 (inset to f). Scale bars: a–c = 5 μm; d–f = 1 μm, inset to f = 500 nm.
      The LM7 epitope was also found to occur in regions of the thickened outer cell walls of stem epidermal cells as shown in Fig.6 a. The distribution of the LM7 epitope was restricted to discrete regions of the outer epidermal cell wall that were aligned with radial cell walls between epidermal cells but was absent entirely from the radial cell walls themselves (Fig. 6 a). In contrast, the JIM5 epitope occurred only sparsely in most of the thickness of the outer tangential epidermal cell walls but abundantly in radial epidermal cell walls (Fig.6 b).
      Figure thumbnail gr6
      Figure 6The LM7 epitope occurs at regions of outer epidermal cell walls and at intercellular spaces throughout pea seedlings. Figure shows micrographs of the binding of LM7 (a and c–e) and JIM5 (b) to the cell walls of pea stem epidermal cells (a and b), leaf epidermal and parenchyma cells (c and d, respectively), and root parenchyma cells (e). All sections were hand-cut transverse sections. Arrowheads ina–c indicate the position of a region of outer epidermal cell wall overlying cell junctions. Arrows in cindicate the limits of the thick outer epidermal wall.Scale bars: a andd = 50 μm; b, c, e, and f = 10 μm.
      In addition to its occurrence in pea stem tissue discussed above, the LM7 epitope was detected at comparable locations in epidermal cell walls and intercellular spaces in all pea seedling organs examined. These included domains within the outer cell wall (just under the cuticle) of the leaf epidermis in line with radial cell walls as shown in Fig. 6 c, leaf parenchyma near the midrib (Fig.6 d), and root cortical parenchyma (Fig. 6 e). Furthermore, the LM7 epitope was found in equivalent regions of developing intercellular spaces and epidermal cell junctions in a range of angiosperms including carrot (Apiaceae), maize (Poaceae),Silene latifolia (Caryophyllaceae), Kalanchoe daigremontiana (Crassulaceae), and Nicotiana tabacum(Solanaceae) (data not shown).

       Functional Implications of Blockwise and Non-blockwise De-esterification of HG Domains of Pectin

      A series of model pectin samples containing the epitopes recognized by PAM1 or LM7 were used in a series of in vitro assays to investigate the implications for a range physical properties of blockwise or non-blockwise distributions of methyl esters on HG domains. Calcium gels formed from pectins E0, F11, F31, P41, F43, and B34 all maintained a stable shape under gravity but differed in their opacities, as shown in Fig. 7 a. The different opacities of the gels suggested that the degree and pattern of de-esterification of HG domains influenced the structure and pore size of the gel and that this was reflected in differences in the light scattering properties.
      Figure thumbnail gr7
      Figure 7The degree and pattern of methyl-esterification of HG domains influences the response of calcium-pectin gels to compression. a, 2% (w/v) gels were formed from a subset of the P-, F-, and B-series pectins by the addition of calcium chloride to a final concentration of 35 mm and equilibration for at least 24 h. Gels were cast in 8.6-mm diameter syringes and cut to a height of 4 mm. band c, the appearance of calcium-mediated pectin gels (prepared as described in a) following compression by the application of a linearly increasing force. Compression was stopped when a force of 9.82 N was reached (E0, F11, and F31), or when gels yielded (F43, B34, and P41). d, force versusstrain curves for pectin gels under compression. Gels were prepared as described in a and compressed at a rate of 0.1 mm s1 to a maximum force of 9.82 N.
      The degree and pattern of methyl-esterification was found to effect the elasticity of the gels and their response to compressive strain. There were significant differences in the extent and manner of deformation of gels when compressed by a force of 9.82 N or to yield point as shown in Fig. 7 (b and c). Compression testing of gels indicated that both the degree and pattern of methyl-esterification were important in determining the yield point and elasticity of gels, as shown in Figs. 7 d and 8. The mean values of the yield strain, yield force, and elasticities are given in the tables of Figs. 7 d and 8 a, with standard deviations given in parentheses. These values were obtained from averages from at least two pairs of measurements, each pair being made on a gel formed from a completely separate solution. The plots of F versus s shown are those of the average values for the samples.
      Figure thumbnail gr8
      Figure 8The degree and pattern of methyl-esterification of HG domains influences the elasticity of calcium-pectin gels. The elasticity of calcium-mediated pectin gels prepared described as in Fig. a. a, elastic moduli were calculated from the gradients of force/strain curves at low strain. b, the graph shows the relationship between elasticity and degree of methyl-esterification for calcium-mediated pectin gels.
      F31 and F43 have the same distribution pattern of methyl groups and differ only in DE. Although F31 formed a strong gel that did not yield at a force of 9.82 N, gels formed from F43 were relatively weak under compression and yielded at a mean force of 0.92 N (Fig. 7 d). Samples F43 and P41 have different distribution patterns of methyl groups but differ in DE by only 2%. However, there was nearly a 3-fold difference in the yield point of gels formed from P41 and F43 (Fig.7 d). The outstanding feature of gels formed from P41 was the fact that, although the gels did yield due to the development of fractures in the gel below F = 9.82 N, on removal of the force the sample recovered its original dimensions almost immediately, as seen in Fig. 7 (b and c). In complete contrast, and as shown in Fig. 7 (b andc), gels formed from F43 showed practically no recovery on removal of the compressive force and behaved more like a plastic material, irreversibly deformed by a stress above its yield stress. Similarly, samples F31 and B34 have similar DE, but gels prepared from these samples differed greatly in their response to compression. The force at which gels formed from B34 yielded was at least 10 times less than the yield force of gels formed from F31. These results strongly suggest that, although methyl groups on both F31 and B34 are distributed in a non-blockwise fashion in both cases, the distribution patterns are not identical. This was also suggested by their differing capacities to be degraded by PL (
      • Limberg G.
      • Körner R.
      • Bucholt H.C.
      • Christensen T.M.I.E
      • Roepstorff P.
      • Mikkelsen J.D.
      ).
      With respect to the elasticity of gels, decreasing DE was broadly correlated with increasing elasticity as shown in Fig. 8 (aand b), as would be expected if free carboxyl groups are required for calcium cross-linking to occur. However, as was the case for the yield force, both the degree and pattern of methyl-esterification appeared to be important in determining the elasticity of gels, as not all the samples fell exactly on the same curve.

       The Effect of the Degree and Pattern of Methyl-esterification on the Water Holding Capacity and Porosity of Calcium-mediated Pectin Gels

      Gels formed from E0, F11, F31, P41, F43, and B34 differed significantly in their water holding capacities and porosity to protein as shown in Fig. 9. For the F-series samples, there was some correlation between DE and water holding capacity as shown in Fig. 9 a. However, for the samples as a whole, the correlation between DE and water holding capacity was not strong. For example, there were significant differences in the water holding capacities of gels formed from F31 and B34 and in the water holding capacities of gels formed from P41 and F43. It is worth noting that the pectin gels that showed the largest volume loss (E0 and F11) were also the ones that appeared to collapse under compression with relatively little expansion, or increase in the lateral dimension. For these samples this implies a collapse of the gel structure as water is lost, unlike F43 mentioned previously, which maintains a coherent structure, loses relatively little water, and exhibits the pronounced broadening of an incompressible, plastic material.
      Figure thumbnail gr9
      Figure 9The degree and pattern of methyl-esterification of HG domains affects the water holding capacity and porosity of calcium-pectin gels. a, the water holding capacity of gels was assessed by the change in volume of gels following compression by a force of 9.82 N (E0, F11, and F31) or by a force at which gels yielded (F43, B34, and P41). The graphshows the relationships between volume loss under compression and degree of methyl-esterification. b, the porosity of pectin gels to protein was assessed by the rate of incorporation of BSA into gels. Gels were incubated in a solution of BSA in de-ionized water (5 mg/ml) and removed at selected time points (loading time). The amount of BSA that had entered the gels was determined by dissolving gels in 50 mm CDTA and assaying the amount of protein in the resulting solutions.
      For all the samples, the rate of incorporation of BSA into gels decreased over time, but there were significant differences in the amount of protein that had been incorporated into gels after 20 h as shown in Fig. 9 b. Although the gels formed from F31 and B34 had similar porosities, the gels formed from P41 and F43 had the highest and lowest porosities, respectively. Additionally, the porosity of gels formed from F11 was more similar to that of gels formed from F43 than it was to gels formed from F31. From these results, it appears that the pattern rather the degree of methyl-esterification of HG may be the more important factor in determining both the water holding capacity and the porosity of calcium-mediated pectin gels.

      DISCUSSION

      The formation of the pectic network in the primary cell wall matrix is contingent upon polysaccharide synthesis in the Golgi apparatus, its deposition and assembly in the cell wall, and subsequent modification by cell wall-based enzymes in response to functional requirements. The modulation of the degree and pattern of methyl-esterification of HG is one aspect of this functional fine tuning, and the work reported here indicates that HG domains with distinctive physical properties are produced in discrete microdomains of primary cell walls. The fact that LM7 bound to cell walls in a range of organs and species indicates that a non-blockwise pattern of methyl-esterification is a widespread aspect of HG modification in plants.
      It is likely that the abundance and pattern of methyl ester groups varies along an HG chain, and, although some epitope structures, such as that recognized by LM7, have distinct locations within the intercellular matrix, they do not necessarily occur exclusively. For example, the epitopes recognized by LM7, JIM5, and PAM1 were all present at the lining of intercellular spaces of pea stem parenchyma. Therefore, discrete microdomains of the cell wall matrix are likely to contain HG with a mixture of HG methyl ester distribution patterns resulting in complex combinations of physical properties. In vitro analysis of calcium-mediated model pectin gels indicated that the compressive strength, elasticity, water holding capacity, and the porosity of gels was significantly influenced by both the pattern as well as the degree of methyl-esterification of HG domains. Although it is possible that some variation of the degree and pattern of methyl-esterification of HG may be generated during synthesis, it is thought that HG is usually highly methyl-esterified prior to insertion into cell walls (
      • Mohnen D.
      • Barton D.
      • Nakanishi K.
      • Meth-Cohn O.
      ,
      • Zhang G.F.
      • Staehelin L.A.
      ). It is therefore likely that the activity of PMEs with varying de-esterification action patterns is an important mechanism for modifying matrix properties in planta.
      The results reported here demonstrate that blockwise and non-blockwise distribution patterns of methyl groups on HG can significantly influence the physical properties of calcium-mediated pectin gels. Details of PME action and its wider consequences on the cell wall environment are far from clear. One aspect of PME action may be to generate extracellular pH gradients that in turn orchestrate cell wall loosening via pH-sensitive processes. For example, partial inhibition of the expression of a pea PME gene (rcpme1) has been correlated with changes in extracellular pH and in cell development (
      • Wen F.
      • Zhu Y.
      • Hawes M.C.
      ). Moreover, PME activity can itself modulated by pH. Kinetic analysis of PMEs from mung bean hypocotyl demonstrated the coexistence of three isoforms with different pH and ion sensitivities and differentKm and V max (
      • Bordenave M.
      • Breton C.
      • Goldberg R.
      • Huet J.C.
      • Perez S.
      • Pernollet J.C.
      ).In vitro analysis of the mung bean PME isoforms indicated that different action patterns could be generated by different pH conditions and the DE of the substrate (
      • Catoire L.
      • Pierron M.
      • Morvan C.
      • Hervé du Penhoat C.
      • Goldberg R.
      ). These observations raise the possibility of feedback mechanisms modulating HG structure and hence cell wall matrix properties. An another aspect of possible feedback is that HG backbones with differing methyl-esterification patterns are differentially susceptible to subsequent enzymatic cleavage. PME action on a highly esterified HG is required before polygalacturonases can cleave effectively. Plant polygalacturonases occur in large multigene families, and members are likely to have a range of functions in plant growth and development although these are as yet uncertain (
      • Hadfield K.A.
      • Bennett A.B.
      ,
      • Torki M.
      • Mandaron P.
      • Mache R.
      • Falconet D.
      ). The generation of HG substrates susceptible to polygalacturonase cleavage is likely to be greatly influenced by PME activities.
      The LM7 epitope is the first spatially regulated HG epitope to be found to occur at a consistent location within cell walls and intercellular matrices across a range of plant organs and species. This consistency suggests a specific functionality for the LM7 epitope-rich pectin at these locations. The occurrence of the LM7 epitope within the cell walls of epidermal cell junctions and intercellular spaces indicates that a pectic HG domain with a non-blockwise distribution of methyl esters (and with consequent distinctive properties) has a role associated with cell adhesion at cell junctions. Of the model pectin samples tested, F31 and B34 contained the LM7 HG epitope at the highest abundance. These samples have similar DE and are both the products of (different) non-blockwise de-esterification procedures, but gels formed from the B34 and F31 differ significantly in their physical properties. This suggests that even subtle differences in the distribution patterns of methyl groups (revealed by enzymatic fingerprinting in the case of B34 and F31; Ref.
      • Limberg G.
      • Körner R.
      • Bucholt H.C.
      • Christensen T.M.I.E
      • Roepstorff P.
      • Mikkelsen J.D.
      ) may have profound effects on matrix properties. The occurrence of the LM7 epitope in plant material indicated the presence of HG with non-blockwise distributions of methyl groups, but not the precise distribution pattern. However, as more PME genes are cloned and their products characterized in developmental contexts, these details may become clear.
      The driving force for intercellular space formation appears to be turgor pressure producing tensile forces in cell walls and inducing a tendency toward reduced volume and hence spherical cell shapes (
      • Jarvis M.C.
      ). Three-way cell junctions are therefore subjected to forces tending toward cell separation (
      • Jarvis M.C.
      ). To initiate space formation, a region of primary cell wall must first be dismantled to allow the middle lamellae to link up (
      • Knox J.P.
      ) and as shown schematically in Fig.10. Intercellular space then results from controlled splitting at the middle lamellae. As space develops, the stresses are greatest at regions of adhered walls bordering the separated cell walls and the intercellular space. The LM7 epitope-rich pectin appears to be present from the earliest stages of intercellular space formation and to be most abundantly maintained at the points of cell to cell contact. Pectin containing this epitope may have a direct role in maintaining cell wall to cell wall links at these points through calcium-mediated cross-linking. Discrete electron-dense regions within middle lamellae of pea cotyledon parenchyma have been proposed to be involved in limiting cell separation in this tissue (
      • Kollöffel C.
      • Linssen P.W.T.
      ). In the outer thickened cell wall of epidermal cells, the LM7 epitope occurs in discrete regions that are associated with cell junctions and may be involved in maintaining the integrity of the outer cell layer. An accumulation of calcium in equivalent regions of outer epidermal cell walls (together with an accumulation at the corners of intercellular spaces) has been reported in mung bean hypocotyl (
      • Liberman M.
      • Mutaftschiev S.
      • Jauneau A.
      • Vian B.
      • Catesson A.M.
      • Goldberg R.
      ). The common aspect between the corners of an intercellular space and points of outer epidermal cell wall at the plant surface is that they are both points of contact between two cells, although in the former case cell separation occurs to some extent. LM7 epitope-rich pectin may provide an appropriate environment (porosity, ionic status, etc.) for processes that directly maintain cell to cell contacts (or indeed for enzymes involved in dismantling such contacts, although this seems less likely as cell separation is not generally a feature of epidermal cell junctions). An additional possibility is that the LM7-binding pectin may have a defensive role at points of intercellular attachment. For example, a subtly altered pattern of HG methyl ester groups may alter both the capacity to be degraded by microbial pectinases and also the precise nature and properties of any oligogalacturonide products released.
      Figure thumbnail gr10
      Figure 10Schematic diagram showing occurrence of the LM7 epitope in relation to cell junctions and the formation of intercellular space. a, intercellular space forms at the junction between old (o) and new (n) cell walls and involves the linking up of middle lamellae. This requires the dismantling of a region of the older cell wall (indicated by thedotted circle). b, a non-expanded intercellular junction with no air space. The gray triangle indicates the position of expanded middle lamella material that occupies the space completely and corresponds to the position of LM7 labeling. c, an expanded cell junction with a large intercellular space (is). The arrowsindicate the forces generated by intracellular turgor pressure that drive cell separation. The gray trianglesindicate the regions of the cell that correspond to the position of LM7 labeling. In all cases, c indicates the interior of cells. In all cases, thick lines indicate the plasma membrane face of cell walls and thin lines the position of middle lamellae. Figure was adapted from Jarvis (
      • Jarvis M.C.
      ).
      In conclusion, the observations reported here demonstrate that modulations of the pattern and degree of methyl-esterification of pectic HG occur within discrete regions of primary cell walls and, in particular, that a non-blockwise pattern of methyl esters of HG is an abundant feature of HG. We also show that the pattern and degree of methyl group distribution significantly affect the mechanical and physiological properties of calcium-mediated pectin gels and are therefore likely to influence the in vivo functionalities of pectic HG domains. In this way, a highly methyl-esterified HG polysaccharide that is deposited in the cell wall can potentially be modified in different ways to generate distinct functional properties. Understanding the cell biological context of the products of PME action will be crucial for determining the functions of PME multigene family members.

      REFERENCES

        • Pilnik W.
        • Phillips G.O.
        • Williams P.A.
        • Wedlock D.J.
        Gums and Stabilisers for the Food Industry. Oxford University Press, Oxford1990: 209-221
        • O'Neill M.A.
        • Albersheim P.
        • Darvill A.
        • Dey P.M.
        Methods in Plant Biochemistry. 2. Academic Press, London1990: 415-441
        • Mohnen D.
        • Barton D.
        • Nakanishi K.
        • Meth-Cohn O.
        Comprehensive Natural Products Chemistry. 3. Elsevier Science Publishers B. V., Amsterdam1999: 497-527
        • Schols H.A.
        • Bakx E.J.
        • Schipper D.
        • Voragen A.G.J.
        Carbohydr. Res. 1995; 279: 265-279
        • O'Neill M.A.
        • Warrenfeltz D.
        • Kates K.
        • Pellerin P.
        • Doco T.
        • Darvill A.G.
        • Albersheim P.
        J. Biol. Chem. 1996; 271: 22923-22930
        • Vidal S.
        • Doco T.
        • Williams P.
        • Pellerin P.
        • York W.S.
        • O'Neill M.A.
        • Glushka J.
        • Darvill A.G.
        • Albersheim P.
        Carbohydr. Res. 2000; 326: 277-294
        • Jarvis M.C.
        Plant Cell Environ. 1984; 7: 153-164
        • Knox J.P.
        Plant J. 1992; 2: 137-141
        • Carpita N.C.
        • Gibeaut D.M.
        Plant J. 1993; 3: 1-30
        • Roberts J.A.
        • Whitelaw C.A.
        • Gonzalez-Carranza Z.H.
        • McManus M.T.
        Ann. Bot. 2000; 86: 223-235
        • Eberhard S.
        • Doubrava N.
        • Marfà V.
        • Mohnen D.
        • Southwick A.
        • Darvill A.
        • Albersheim P.
        Plant Cell. 1989; 1: 747-755
        • Dumville J.C.
        • Fry S.C.
        Plant Physiol. Biochem. 2000; 38: 125-140
        • Moerschbacher B.M.
        • Mierau M.
        • Graeβner B.
        • Noll U.
        • Mort A.J.
        J. Exp. Bot. 1999; 50: 605-612
        • Zhang G.F.
        • Staehelin L.A.
        Plant Physiol. 1992; 99: 1070-1083
        • Bordenave M.
        • Breton C.
        • Goldberg R.
        • Huet J.C.
        • Perez S.
        • Pernollet J.C.
        Plant Mol. Biol. 1996; 31: 1039-1049
        • Catoire L.
        • Pierron M.
        • Morvan C.
        • Hervé du Penhoat C.
        • Goldberg R.
        J. Biol. Chem. 1998; 273: 33150-33156
        • Micheli F.
        • Holliger C.
        • Goldberg R.
        • Richard L.
        Gene (Amst.). 1998; 220: 13-20
        • Nari J.
        • Noat G.
        • Diamantidis G.
        • Woudstra M.
        • Ricard J.
        Eur. J. Biochem. 1986; 155: 199-202
        • Moustacas A.M.
        • Nari J.
        • Borel M.
        • Noat
        • Ricard J.
        Biochem. J. 1986; 279: 351-354
        • Denès J.-M.
        • Baron A.
        • Renard C.M.G.C
        • Péan C.
        • Drilleau J.-F.
        Carbohydr. Res. 2000; 327: 385-393
        • Mort A.J.
        • Qiu F.
        • Maness N.O.
        Carbohydr. Res. 1993; 247: 21-35
        • Daas P.J.H.
        • Voragen A.G. J
        • Schols H.A.
        Carbohydr. Res. 2000; 326: 120-129
        • Massiot P.
        • Perron V.
        • Baron A.
        • Drilleau J.F.
        Lebensm. Wiss. U. Technol. 1997; 30: 697-702
        • Limberg G.
        • Körner R.
        • Bucholt H.C.
        • Christensen T.M.I.E
        • Roepstorff P.
        • Mikkelsen J.D.
        Carbohydr. Res. 2000; 327: 293-307
        • Limberg G.
        • Körner R.
        • Christensen T.M.I.E
        • Buchholt H.C.
        • Roepstorff P.
        • Mikkelsen J.D.
        Carbohydr. Res. 2000; 327: 321-332
        • Knox J.P.
        • Linstead P.J.
        • King J.
        • Cooper C.
        • Roberts K.
        Planta. 1990; 181: 512-521
        • Willats W.G.T.
        • Gilmartin P.M.
        • Mikkelsen J.D.
        • Knox J.P.
        Plant J. 1999; 18: 57-66
        • Willats W.G. T
        • Limberg G.
        • Buchholt H.C.
        • van Alebeek G.J.
        • Benen J.
        • Christensen T.M.I.E.
        • Visser J.
        • Voragen A.
        • Mikkelsen J.D.
        • Knox J.P.
        Carbohydr. Res. 2000; 327: 309-320
        • Willats W.G. T
        • Marcus S.E.
        • Knox J.P.
        Carbohydr. Res. 1998; 308: 149-152
        • MacDougall A.J.
        • Needs P.W.
        • Rigby N.M.
        • Ring S.G.
        Carbohydr. Res. 1996; 293: 235-249
        • Wen F.
        • Zhu Y.
        • Hawes M.C.
        Plant Cell. 1999; 11: 1129-1140
        • Hadfield K.A.
        • Bennett A.B.
        Plant Physiol. 1998; 117: 337-343
        • Torki M.
        • Mandaron P.
        • Mache R.
        • Falconet D.
        Gene (Amst.). 2000; 242: 427-436
        • Jarvis M.C.
        Plant Cell Environ. 1998; 21: 1307-1310
        • Kollöffel C.
        • Linssen P.W.T.
        Protoplasma. 1984; 120: 12-19
        • Liberman M.
        • Mutaftschiev S.
        • Jauneau A.
        • Vian B.
        • Catesson A.M.
        • Goldberg R.
        Ann. Bot. 1999; 84: 225-233