A low glycaemic load breakfast can attenuate cognitive impairments observed in middle aged obese females with impaired glucose tolerance.

BACKGROUND AND AIMS: There has been no systematic investigation of the individual and combined effects of impaired glucose tolerance (IGT) and obesity on cognitive function in the absence of ageing. The aims were to examine the effects of IGT and increased waist circumference on cognitive function in ostensibly healthy adults, and to investigate whether a low glycaemic load (GL) breakfast can attenuate cognitive impairments in these populations. METHODS AND RESULTS: Sixty five females aged 30-50 years were classified into one of four groups following waist circumference (WC) measurements and an oral glucose tolerance test: NGT/low WC (n = 25), NGT/high WC (n = 22), IGT/low WC (n = 9), IGT/high WC (n = 9). Memory, psychomotor and executive functions were examined 30 and 120 min after consuming low GL, high GL and water breakfasts according to a randomised, crossover, counterbalanced design. IGT was associated with impairment of verbal and spatial memory, and psychomotor function relative to females with NGT, independent of waist circumference. Increased waist circumference was associated with impairment of verbal memory and executive function relative to females with low WC, independent of IGT. Consumption of the LGL breakfast attenuated verbal memory impairment in the IGT/high WC group relative to the HGL breakfast and no energy control. CONCLUSION: Increased central adiposity and abnormalities in glucose tolerance preceding type 2 diabetes can have demonstrable negative effects on cognitive function, even in ostensibly healthy, middle-aged females. The potential for GL manipulations to modulate glycaemic response and cognitive function in type 2 diabetes and obesity merits further investigation.

The chemical composition of the cell wall of Chlamydomonas gymnogama and the concept of a plant cell wall protein.

Cell walls of Chlamydomonas gymnogama, shed during sexual mating, were collected and analyzed. Ultrastructural examination indicates that the walls are free of cytoplasmic contamination and that they exhibit a regular lamellate structure. The walls are composed of glycoprotein rich in hydroxyproline. The hydroxyproline is linked glycosidically to a mixture of heterooligosaccharides composed of arabinose and galactose. Altogether, the glycoprotein complex accounts for at least 32% of the wall. The amino acid composition of the walls is extraordinarily similar in widely different plant species. The implications of these similarities as well as the widespread occurrence of these glycoproteins are discussed.

Hydroxyproline heterooligosaccharides in Chlamydomonas.

Most of the hydroxyproline in Chlamydomonas reinhardtii is glycosidically linked to oligosaccharides and a monosaccharide that are different from the arabinosides found in hydroxyproline-containing plant cell walls previously examined. Of particular interest is the presence of hydroxyproline-O-galactose. These differences may be common to the volvocalean green algae and may be related to lower tensile strength of the cell walls of this group of plants.

A study of glycaemic effects following acute anthocyanin-rich blueberry supplementation in healthy young adults.

The postprandial response to ingested carbohydrate is recognised as a marker of metabolic health. Postprandial hyperglycaemia is observed in type 2 diabetes mellitus and is a significant risk factor for cardiovascular disease. Cognitive deficits are also associated with type 2 diabetes. Therefore interventions which moderate postprandial glucose profiles are desirable. Here we investigated the impact of anthocyanin-rich wild blueberries on postprandial glucose response. Seventeen healthy young adults consumed a range of doses of freeze-dried wild blueberry powder, in smoothie form, in both sugar-matched and no-added-sugar conditions. Plasma glucose was determined by a capillary sampling method at baseline and at regular intervals up to 2.5 hours postprandially. Blueberries were observed to significantly extend the postprandial glucose response beyond the period observed for a sugar-matched control, characteristic of a beneficial glycaemic response. Furthermore, blueberries were observed to reduce peak postprandial glucose levels, although statistical significance was not achieved. The findings suggest a tempering of the postprandial glucose response in the presence of anthocyanin-rich blueberry, and are discussed with reference to likely glucoregulatory mechanisms of action and their implications for cognitive and type 2 diabetes research.

Development and characterization of breast cancer reactive monoclonal antibodies directed to the core protein of the human milk mucin.

A mucin molecule, which has a molecular weight of greater than 400,000 and which carries tumor associated epitopes recognized by monoclonal antibodies HMFG-1 and HMFG-2, has been purified from human skimmed milk by affinity chromatography followed by passage through a size exclusion column. While treatment of the mucin with hydrogen fluoride for 1 h at 4 degrees C removed the peripheral oligosaccharides, treatment with HF for 3 h at room temperature removed all of its lectin binding ability and revealed a dominant polypeptide of about 68,000. This appears to be the size of the mucin core protein. Monoclonal antibodies have been developed that react with the stripped and partially stripped molecule but not with the intact mucin. From the initial screening on histological sections one of these antibodies, SM-3, reacts with 91% of breast carcinomas but shows little or no reactivity on benign mammary tumors, normal resting, pregnant, or lactating breast. It appears that this monoclonal antibody is reacting with an epitope that is usually masked by oligosaccharide moieties in normal cells but which is exposed, perhaps due to aberrant glycosylation, in malignant cells.

Relationship of pancreatic cancer apomucin to mammary and intestinal apomucins.

Pancreatic cancer mucins have several carbohydrate antigens that are potentially useful in the detection of pancreatic cancers, but little is known about the core polypeptides of pancreatic cancer mucins. In this study, purified mucin from SW1990 pancreatic cancer xenografts was deglycosylated by treatment with hydrogen fluoride to give pancreatic cancer apomucin. Consistent with near-complete removal of carbohydrate, the apomucin had 10- to 70-fold decreased binding of lectins and, unlike the native mucin, served as an acceptor for polypeptidyl N-acetylgalactosaminyl transferase. Antibodies prepared against the apomucin did not bind to native mucin, and antibodies that bound to native mucin did not bind to apomucin. On the basis of cross-reaction with deglycosylated colon cancer mucin and intestinal mucin repeat peptide, apomucins from SW1990 pancreatic cancer xenografts contain the intestinal mucin repeat peptide. On the basis of binding of breast cancer-reactive monoclonal antibodies 139H2, DF3, and HMFG-2, apomucins from SW1990 pancreatic cancer xenografts also have the mammary mucin repeat peptide. Using complementary DNA probes specific for intestinal mucin and breast mucin sequences, both types of apomucin mRNA were detected in nude mouse xenografts of SW1990 cells. In immunohistochemical staining, antibody against deglycosylated SW1990 mucin stained normal breast and pancreas but not normal colon. Some pancreatic and mammary cancers and most colonic cancers, however, were stained by antibodies against both intestinal apomucin and mammary apomucin. We conclude that pancreatic cancers can produce mucins with the intestinal repeat peptide as well as those with mammary repeat peptide sequences.

Impact of postprandial glycaemia on health and prevention of disease.

Postprandial glucose, together with related hyperinsulinemia and lipidaemia, has been implicated in the development of chronic metabolic diseases like obesity, type 2 diabetes mellitus (T2DM) and cardiovascular disease (CVD). In this review, available evidence is discussed on postprandial glucose in relation to body weight control, the development of oxidative stress, T2DM, and CVD and in maintaining optimal exercise and cognitive performance. There is mechanistic evidence linking postprandial glycaemia or glycaemic variability to the development of these conditions or in the impairment in cognitive and exercise performance. Nevertheless, postprandial glycaemia is interrelated with many other (risk) factors as well as to fasting glucose. In many studies, meal-related glycaemic response is not sufficiently characterized, or the methodology with respect to the description of food or meal composition, or the duration of the measurement of postprandial glycaemia is limited. It is evident that more randomized controlled dietary intervention trials using effective low vs. high glucose response diets are necessary in order to draw more definite conclusions on the role of postprandial glycaemia in relation to health and disease. Also of importance is the evaluation of the potential role of the time course of postprandial glycaemia.

Life behind cell walls: paradigm lost, paradigm regained.

This review of the living cell wall and its protein components is in two parts. The first is anecdotal. A personal account spanning over 40 years research may perhaps be an antidote to one stereotypical view of scientists as detached and humorless. The second part deals with the meaning of function, particularly as it applies to hydroxyproline-rich glycoproteins. Function is a difficult word to define objectively. However, with help from such luminaries as Humpty Dumpty: "A word means what I want it to mean, neither more nor less," and Wittgenstein: "Giving examples of usage ... is the only way to talk about meaning," it is possible to construct a ziggurat representing increasingly complex levels of organization from molecular structure to ecology. Forty years ago I suggested that hydroxyproline-rich structural proteins played a key role in cell wall functioning. But because the bulk of the wall is carbohydrate, there has been an understandable resistance to paradigm change. Expansins, paradoxically, contribute greatly to this resistance because their modus operandi as cell-wall-loosening proteins is based on the idea that they break hydrogen bonds between polysaccharide chains allowing slippage. However, this view is not consistent with the recent discovery [Grobe et al. (1999) Eur. J. Biochem 263: 33-40] that beta-expansins may be proteases, as it implies that the extensin network is not a straightjacket but a substrate for expansin in muro. Such a direct role for extensins in both negative and positive regulation of cell expansion and elongation may constitute a major morphogenetic mechanism operating at all levels of plant growth and development.

Purification and Partial Characterization of a Hydroxyproline-Rich Glycoprotein in a Graminaceous Monocot, Zea mays.

Graminaceous monocots generally contain low levels of hydroxyproline-rich Glycoproteins (HRGPs). As HRGPs are often at the cell surface, we used the intact cell elution technique (100 millimolar AlCl(3)) to isolate soluble surface proteins from Zea mays cell suspension cultures. Further fractionation of the trichloroacetic acid-soluble eluate on the cation exchangers phospho-cellulose and BioRex-70 gave several retarded, hence presumably basic fractions, which also contained hydroxyproline (Hyp). One of these fractions yielded a pure HRGP after a final purification step involving Superose-6 gel filtration. As this HRGP was unusually rich in threonine, (25 mole%) we designated it as a threonine-hydroxyproline-rich glycoprotein (THRGP); it contained about 27% carbohydrate occurring exclusively as arabinosylated Hyp, predominantly as the monosaccharide (15%), and trisaccharide (25%) with 48% Hyp nonglycosylated-a characteristically graminaceous monocot profile. Amino acid analysis confirmed the basic character, and gave a low alanine content. Reaction with Yariv artificial antigen was negative. These characteristics show that the THRGP is not an arabinogalactan protein. On the other hand, antibodies raised against tomato extensin P1 cross-reacted significantly with the THRGP; this cross-reactivity and the above analytical data provide the best evidence to date for the presence of extensin in a graminaceous monocot.

Hydrogen fluoride saccharification of wood: lignin fluoride content, isolation of alpha-D-glucopyranosyl fluoride and posthydrolysis of reversion products.

Wood chips from bigtooth aspen (Populus grandidentata Michx.) were saccharified by reaction with liquid hydrogen fluoride either anhydrous or containing up to 10% v/v water. The reaction products were separated into a solid lignin fraction and a water-soluble saccharide fraction. The fluoride content of the lignin (determined after alkaline fusion) was initially about 1 mg/g wood, but was lowered to 0.1 mg/g wood by grinding and washing. Thus little or no chemical binding of fluoride to lignin occurred during hydrogen fluoride (HF) solvolysis. Analysis of the water-soluble fraction by gel filtration on Biogel P2 columns showed a range of low-molecular-weight oligosaccharides and only 10-20% sugar monomers. Thus considerable reversion occurred during HF evacuation. Posthydrolysis conditions were optimized for these reversion products by varying temperature and acid concentration. Optimal conditions at 1 h were 140 degrees C with 100mN sulfuric acid or 225mN Hydrofluoric acid resulting in monomer yields of > 90% for 0.5% sugar solutions and > 80% for 10% sugar solutions. After reaction of pure cellulose (Filter paper) with hydrogen fluoride in the absence of water, and terminating the reaction with calcium carbonate, the reaction intermediate alpha-D-glucopyranosylfluoride was isolated with a maximal yield of 0.2 g/g paper. Upon purification via paper chromatography glucosylfluoride was identified by its specific rotation and also by gas chromatography-mass spectrometry of its tetra-O-trimethylsilyl derivative.

A microapparatus for liquid hydrogen fluoride solvolysis: sugar and amino sugar composition of Erysiphe graminis and Triticum aestivum cell walls.

The assembly and use of a simple and safe apparatus for HF solvolysis of microgram amounts of cell walls, polysaccharides, or glycoproteins are described. Using this apparatus the cell wall composition of Erysiphe graminis was compared with that of its wheat host. The HF solvolysis combined with TFA posthydrolysis considerably increased sugar yields compared with TFA hydrolysis alone, due mainly to increased yields of glucose from wheat, and glucosamine from Erysiphe, corresponding to cellulose and chitin, respectively. A potentially useful method for determining amounts of fungal hyphae in plant tissue is also provided.

Extensin: repetitive motifs, functional sites, post-translational codes, and phylogeny.

Homologous hydroxyproline-rich glycoproteins (HRGPs) of the plant extracellular matrix include extensins, repetitive proline-rich proteins (RPRPs), some nodulins, gum arabic glycoprotein (GAGP), arabinogalactan-proteins (AGPs), and chimeric proteins such as potato lectin which contain an extensin module fused to a lectin. The key to the role of HRGPs in cell wall self-assembly and cell extension lies in their chemistry, which is dependent on extensive post-translational modifications (PTMs): hydroxylation, glycosylation, and cross-linking. Repetitive peptide motifs characterize HRGPs. One or more repetitive peptide motifs and their variants, singly or in combination, may constitute functional sites involved in various aspects of cell wall assembly, as follows: (i) X-Hypn including Ser-Hyp4 (arabinosylation site, molecular rigidity, and reptation). (ii) Pro-Hyp-Val-Tyr-Lys and variants (putative intermolecular cross-links, adhesion, cohesion, and possible beta-turns). (iii) Tyr-X-Tyr-Lys (intramolecular isodityrosine [IDT] cross-links increase molecular rigidity and hydrophobicity). (iv) (Glyco)peptide palindromes (centrosymmetric domains: putative self-assembly nucleation sites). (v) Ionic interaction sites (protein-protein and protein-carbohydrate cross-links). (vi) Hyp and Ser glycosylation sites (enhance conformational stability and molecular recognition). (vii) Extensin modules in chimeric proteins (e.g. solanaceous lectins). Rules for the post-translational modifications are emerging: (i) Hydroxylation of proline residues may depend on multiple, sequence-specific prolyl hydroxylases rather than on a single (polyproline-II) conformation-dependent enzyme. Furthermore, Lys-Pro, Tyr-Pro, and Phe-Pro are not hydroxylated, while Pro-Val is always. (ii) Contiguity of Hyp residues probably determines the extent of Hyp glycosylation, blocks of tetrahydroxyproline (Hyp4) being the most highly arabinosylated, while single non-contiguous Hyp residues are rarely arabinosylated, although they are likely attachment sites for the larger arabinogalactan substituents of gum arabic glycoprotein and arabinogalactan-proteins. (iii) While intramolecular cross-links involve IDT, unidentified intermolecular cross-links most likely involve the Val-Tyr-Lys motif (perhaps also Val-Lys-Pro-Tyr-His-Pro), probably as an adduct between Tyr and Lys catalyzed in vitro by a pI 4.6 extensin cross-linking peroxidase. Thus, we can classify HRGPs functionally as either cross-linking or non-cross-linking, i.e. CL- or NCL-extensins. Their protistan origin obscures the phylogenetic affinities of a single extensin-HRGP family due to their sequence divergence. We propose a phylogenetic series ranging from the minimally glycosylated basic RPRPs to the highly glycosylated acidic AGPs. Furthermore, based on similarities between dicots and gymnosperm extensins, and their marked difference from graminaceous monocot extensins, graminaceous monocot and dicot lines may have diverged as early as the progymnosperms.(ABSTRACT TRUNCATED AT 400 WORDS)

An Accounting of Horseradish Peroxidase Isozymes Associated with the Cell Wall and Evidence that Peroxidase Does Not Contain Hydroxyproline.

Isopycnic equilibrium centrifugation techniques were used to determine whether any horseradish (Amoracia lapathifolia) peroxidase isozymes were associated with hydroxyproline containing moieties. Purified peroxidase, horseradish root extracts, and peroxidase isozymes released from horseradish root cell walls were tested. In no case could any peak of peroxidase activity be found to band with hydroxyproline.A fluorimetric method for measurement of peroxidase activity was used to determine quantitatively the amount of total peroxidase located on horseradish root cell walls. Twenty per cent of the total peroxidase is found in the cell wall fraction after extraction; 93% of this cell wall associated peroxidase can be removed by washing with 2 m NaCl. Some peroxidase isozymes released by salt washing are not found in the cytoplasmic extract. This indicates that not all of the ionically bound peroxidase represents cytoplasmic contamination. The 1.4% of the total peroxidase activity can thus be considered tightly bound to the cell wall. Of this portion, 75% can be solubilized by treatment with a cellulase preparation. One isozyme is released which was not present in the original cytoplasmic extract.

Hydroxyproline arabinosides in the plant kingdom.

The hydroxyproline-O-arabinosyl linkage is present in cell walls of selected tissues representing green plants from algae to angiosperms.

Detection of glycosylated and deglycosylated extensin precursors by indirect competitive ELISA.

A competitive indirect enzyme-linked immunosorbent assay (ELISA) was developed for the rapid quantitation of the glycosylated and deglycosylated forms of the monomeric soluble extensin precursor subunits P1 and P2. A log-linear response range for each kind of precursor in the competition curve was between 0.01 and 100 nanograms per milliliter.

Characterization of native and modified extensin monomers and oligomers by electron microscopy and gel filtration.

We isolated hydroxyproline-rich extensin precursors from suspension-cultured tomato, cucumber, and sycamore-maple by salt-elution of intact cells and cell wall preparations. Cation exchange chromatography and HPLC gel filtration resolved these precursors into monomeric and oligomeric fractions, confirmed by amino acid analysis, immunological cross-reactivity, and TEM visualization. After rotary shadowing monomers appeared as flexuous rods with a contour length of 70 to 100 nanometers and a ;persistence length' (maximum linear displacement) of 44 to 51 nanometers. Oligomers were larger branched assemblies with occasional pores. Native extensin monomers gave uniform gel filtration retention times (Rts), but the Rts of HF-deglycosylated monomers varied depending on concentration, implying ionic interaction between the highly basic deglycosylated monomers and a weakly cationic gel matrix. Succinylation of the deglycosylated monomers reversed the net charge, and restored the retention time to that of glycosylated monomers, confirming the ionic interaction. Succinylation enhanced visualization of the deglycosylated monomers, which previously were barely discernible flexuous rods. The persistence length:contour length ratios of succinylated deglycosylated monomers (tomato sdP2) and glycosylated monomers (sP2) were the same, implying a similar molecular flexibility for both glycosylated and deglycosylated monomers at room temperature. These molecular properties are consistent with suggestions that extensin monomers reptate into the wall as a transmural protein ;weft' which becomes progressively cross-linked forming a network penetrated by the cellulose ;warp.'

Trypsin cleaves lysylproline in a hydroxyproline-rich glycoprotein from Zea mays.

Although trypsin is highly specific for lysyl and arginyl bonds, some peptide bonds, such as lysylproline, are generally trypsin-resistant, with rare exceptions as reported here. Trypsin cleaved a specific Lys-Pro bond in the chymotryptic peptide: Thr-Hyp-Ser-Hyp-Lys-Pro-Hyp-Thr-Pro-Lys-Pro-Thr-Hyp-Hyp-Thr-Tyr isolated from a Zea mays hydroxyproline-rich glycoprotein (HRGP). The daughter peptides, Thr-Hyp-Ser-Hyp-Lys-Pro-Hyp-Thr-Pro-Lys and Pro-Thr-Hyp-Hyp-Thr-Tyr, show cleavage of only one of the two Lys-Pro bonds in the parent peptide. From these and other data we suggest that there are two prerequisites for Lys-Pro cleavage: First, an extended helix characteristically present in proline or hydroxyproline-rich proteins; second, flexibility in two residues flanking the Lys-Pro bond.

Structure of the Threonine-Rich Extensin from Zea mays.

Chymotryptic digestion of a threonine-rich hydroxyproline-rich glycoprotein (THRGP) purified from the cell surface of a Zea mays cell suspension culture gave a peptide map dominated by the hexadecapeptide TC5: Thr-Hyp-Ser-Hyp-Lys-Pro-Hyp-Thr-Pro-Lys-Pro-Thr-Hyp-Hyp-Thr-Tyr, in which the repetitive motif Ser-Hyp-Lys-Pro-Hyp-Thr-Pro-Lys is homologous with the dominant decamer of P1-type dicot extensins: Ser-Hyp-Hyp-Hyp-Hyp-Thr-Hyp-Val-Tyr-Lys, modified by a Lys for Hyp substitution at residue 3, a Val-Tyr deletion at residues 8 and 9, and incomplete post-translational modification of proline residues. One of the minor peptides (TC1) contained the 8-residue sequence: Thr-Hyp-Ser-Hyp-Hyp-Hyp-Hyp-Tyr corresponding to the C-terminal tail (judging from the recently isolated maize cDNA clone MC56) which is homologous with the major repetitive motif of the ;P3' class of dicot extensins. Direct peptide sequencing defined potential glycosylated regions on the THRGP corresponding to clone MC56 and showing that glycosylated and nonglycosylated domains alternate with high regularity. The THRGP is not in the polyproline-II conformation, judging from circular dichroic spectra, but nevertheless is an extended rod, from electron microscopic data. HF-solvolysis of cell walls from maize coleoptile, root, and root tip released deglycosylated THRGP detected on sodium dodecyl sulfate-polyacrylamide gel electrophoresis immunoblots with high titer rabbit polyclonal antibodies raised against the intact THRGP. In a quantitative enzyme-linked immunosorbent assay, these antibodies cross-reacted 20% with tomato P1 extensin, and 18% with anhydrous hydrogen fluoride-deglycosylated P1. These results, together with other previously published data, show that maize THRGP is homologous with the dicot P1 extensins and, as such, is the first extensin isolated from a graminaceous monocot.

A chenopod extensin lacks repetitive tetrahydroxyproline blocks.

An extensin isolated from sugar beet (Beta vulgaris) cell suspension cultures fulfills all criteria for membership of the extensin family save one, notably, lack of the ;diagnostic' pentamer Ser-Hyp-Hyp-Hyp-Hyp. However, sequence analysis of the major tryptic peptides shows that sugar beet extensin shares a motif in common with tomato extensin P1 but differs by the position of an insertion sequence [X] or [Y] which, in sugar beet, splits the tetrahydroxyproline block: Ser-Hyp-Hyp-[X]-Hyp-Hyp-Thr-Hyp-Val-Tyr-Lys, where [X] is [Val-His-Glu/Lys-Tyr-Pro], while in tomato the insertion sequence [Y] = [Val-Lys-Pro-Tyr-His-Pro] and, when it occurs, immediately follows the tetrahydroxyproline block: Ser-Hyp-Hyp-Hyp-Hyp-[Y]-Thr-Hyp-Val-Tyr-Lys. Based on these data we reinterpret three highly repetitive cDNA sequences, including nodulin N75 from soybean and wound-induced P33 of carrot, as extensins with split tetra(hydroxy)proline blocks.

Gum arabic glycoprotein is a twisted hairy rope : a new model based on o-galactosylhydroxyproline as the polysaccharide attachment site.

Separation of the wound exudate from Acacia senegal (L.) Willd., "gum arabic," on a preparative Superose-6 column gave two major fractions: a high molecular weight gum arabic glycoprotein (GAGP) containing about 90% carbohydrate and a lower molecular weight heterogenous gum arabic polysaccharide fraction. Hydrogen fluoride-deglycosylation of GAGP gave a large ( approximately 400 residue) hydroxyproline-rich polypeptide backbone (dGAGP). Alkaline hydrolysis of GAGP showed that most of the carbohydrate was attached to the polypeptide backbone as small ( approximately 30 residue) hydroxyproline (Hyp)-polysaccharide substituents. After partial acid hydrolysis of the Hyp-polysaccharide fraction we identified O-galactosylhydroxyproline as the glycopeptide linkage, identical with that of hydroxyproline-rich arabinogalactan-proteins (AGPs). However, unlike the acidic alanine-rich AGPs, GAGP is basic and notably deficient in alanine. Thus, while the GAGP polypeptide backbone more closely resembles that of the Hyp-rich cell wall protein extensin, the GAGP polysaccharide sidechains resemble AGPs. Possibly all three proteins comprise a phylogenetically related extensin superfamily of extended rod-like macromolecules. The "wattle-blossom" model for AGP and gum arabic predicts a few large polysaccharide substituents along the polypeptide backbone of a spheroidal macromolecule. On the contrary, our data imply a rodlike molecule with numerous small polysaccharide substituents (attached to 24% of the Hyp residues), regularly arranged along a highly periodic polypeptide backbone based, hypothetically, on a 10 to 12 residue repetitive peptide motif. Thus, a simple statistical model of the gum arabic glycoprotein predicts a repeating polysaccharide-peptide subunit of about 7 kilodaltons. The small polysaccharide substituents will maximize intramolecular hydrogen bonding if aligned along the long axis of the molecule, forming in effect a twisted hairy rope. Electron micrographs of rotary shadowed GAGP molecules support that prediction and may also explain how such apparently large molecules can exit the cell by endwise reptation through the small pores of the primary cell wall.