Identification, subcellular localization and biochemical characterization of water-soluble heteroglycans (SHG) in leaves of Arabidopsis thaliana L.: distinct SHG reside in the cytosol and in the apoplast

Water-soluble heteroglycans (SHG) were isolated from leaves of wild-type Arabidopsis thaliana L. and from two starch-deficient mutants. Major constituents of the SHG are arabinose, galactose, rhamnose, and glucose. SHG was separated into low (<10 kDa; SHG(S)) and high (>10 kDa; SHG(L)) molecular weight compounds. SHG(S) was resolved into approximately 25 distinct oligoglycans by ion exchange chromatography. SHG(L) was further separated into two subfractions, designated as subfraction I and II, by field flow fractionation. For the intracellular localization of the various SHG compounds several approaches were chosen: first, leaf material was subjected to non-aqueous fractionation. The apolar gradient fractions were characterized by monitoring markers and were used as starting material for the SHG isolation. Subfraction I and SHG(S) exhibited a distribution similar to that of cytosolic markers whereas subfraction II cofractionated with crystalline cellulose. Secondly, intact organelles were isolated and used for SHG isolation. Preparations of intact organelles (mitochondria plus peroxisomes) contained no significant amount of any heteroglycan. In isolated intact microsomes a series of oligoglycans was recovered but neither subfraction I nor II. In in vitro assays using glucose 1-phosphate and recombinant cytosolic (Pho 2) phosphorylase both SHG(S) and subfraction I acted as glucosyl acceptor whereas subfraction II was essentially inactive. Rabbit muscle phosphorylase a did not utilize any of the plant glycans indicating a specific Pho 2-glycan interaction. As revealed by in vivo labeling experiments using 14CO2 carbon fluxes into subfraction I and II differed. Furthermore, in leaves the pool size of subfraction I varied during the light-dark regime.     

Analysis of cytosolic heteroglycans from leaves of transgenic potato (Solanum tuberosum L.) plants that under- or overexpress the Pho 2 phosphorylase isozyme

During starch degradation, chloroplasts export neutral sugars into the cytosol where they appear to enter a complex glycan metabolism. Interactions between glycans and glucosyl transferases residing in the cytosol were studied by analyzing transgenic potato (Solanum tuberosum L.) plants that possess either decreased or elevated levels of the cytosolic (Pho 2) phosphorylase isoform. Water-soluble heteroglycans (SHGs) were isolated from these plants and were characterized. SHG contains, as major constituents, arabinose, rhamnose, galactose and glucose. Non-aqueous fractionation combined with other separation techniques revealed a distinct pool of the SHG that is located in the cytosol. Under in vitro conditions, the cytosolic heteroglycans act as glucosyl acceptor selectively for Pho 2. Acceptor sites were characterized by a specific hydrolytic degradation following the Pho 2-catalyzed glucosyl transfer. The size distribution of the cytosolic SHG increased during the dark period, indicating a distinct metabolic activity related to net starch degradation. Antisense inhibition of Pho 2 resulted in increased glucosyl and rhamnosyl contents of the glycans. Overexpression of Pho 2 decreased the content of both residues. Compared with the wild type, in both types of transgenic plants the size of the cytosolic glycans was increased.     

A transglucosidase necessary for starch degradation and maltose metabolism in leaves at night acts on cytosolic heteroglycans (SHG)

The recently characterized cytosolic transglucosidase DPE2 (EC 2.4.1.25) is essential for the cytosolic metabolism of maltose, an intermediate on the pathway by which starch is converted to sucrose at night. In in vitro assays, the enzyme utilizes glycogen as a glucosyl acceptor but the in vivo acceptor molecules remained unknown. In this communication we present evidence that DPE2 acts on the recently identified cytosolic water-soluble heteroglycans (SHG) as does the cytosolic phosphorylase (EC 2.4.1.1) isoform. By using in vitro two-step ¹⁴C labeling assays we demonstrate that the two transferases can utilize the same acceptor sites of the SHG. Cytosolic heteroglycans from a DPE2-deficient Arabidopsis mutant were characterized. Compared with the wild type the glucose content of the heteroglycans was increased. Most of the additional glucosyl residues were found in the outer chains of SHG that are released by an endo- α -arabinanase (EC 3.2.1.99). Additional starch-related mutants were characterized for further analysis of the increased glucosyl content. Based on these data, the cytosolic metabolism of starch-derived carbohydrates is discussed.     

Glucose-1-phosphate transport into protoplasts and chloroplasts from leaves of Arabidopsis

Almost all glucosyl transfer reactions rely on glucose-1-phosphate (Glc-1-P) that either immediately acts as glucosyl donor or as substrate for the synthesis of the more widely used Glc dinucleotides, ADPglucose or UDPglucose. In this communication, we have analyzed two Glc-1-P-related processes: the carbon flux from externally supplied Glc-1-P to starch by either mesophyll protoplasts or intact chloroplasts from Arabidopsis (Arabidopsis thaliana). When intact protoplasts or chloroplasts are incubated with [U-(14)C]Glc-1-P, starch is rapidly labeled. Incorporation into starch is unaffected by the addition of unlabeled Glc-6-P or Glc, indicating a selective flux from Glc-1-P to starch. However, illuminated protoplasts incorporate less (14)C into starch when unlabeled bicarbonate is supplied in addition to the (14)C-labeled Glc-1-P. Mesophyll protoplasts incubated with [U-(14)C]Glc-1-P incorporate (14)C into the plastidial pool of adenosine diphosphoglucose. Protoplasts prepared from leaves of mutants of Arabidopsis that lack either the plastidial phosphorylase or the phosphoglucomutase isozyme incorporate (14)C derived from external Glc-1-P into starch, but incorporation into starch is insignificant when protoplasts from a mutant possessing a highly reduced ADPglucose pyrophosphorylase activity are studied. Thus, the path of assimilatory starch biosynthesis initiated by extraplastidial Glc-1-P leads to the plastidial pool of adenosine diphosphoglucose, and at this intermediate it is fused with the Calvin cycle-driven route. Mutants lacking the plastidial phosphoglucomutase contain a small yet significant amount of transitory starch.     

Cytosolic heteroglycans in photoautotrophic and in heterotrophic plant cells

In plants several ‘starch-related’ enzymes exist as plastid- and cytosol-specific isoforms and in some cases the extraplastidial isoforms represent the majority of the enzyme activity. Due to the compartmentation of the plant cells, these extraplastidial isozymes have no access to the plastidial starch granules and, therefore, their in vivo function remained enigmatic. Recently, cytosolic heteroglycans have been identified that possess a complex pattern of the monomer composition and glycosidic bonds. The glycans act both as acceptors and donors for cytosolic glucosyl transferases. In autotrophic tissues the heteroglycans are essential for the nocturnal starch-sucrose conversion. In this review we summarize the current knowledge of these glycans, their interaction with glucosyl transferases and their possible cellular functions. We include data on the heteroglycans in heterotrophic plant tissues and discuss their role in intracellular carbon fluxes that originate from externally supplied carbohydrates.     

Subcellular localization of the starch degradative and biosynthetic enzymes of spinach leaves

The subcellular localization of the starch biosynthetic and degradative enzymes of spinach leaves was carried out by measuring the distribution of the enzymes in a crude chloroplast pellet and soluble protein fraction, and by the separation on sucrose density gradients of intact organelles, chloroplasts, peroxisomes, and mitochondria of a protoplast lysate. ADP-Glucose pyrophosphorylase, starch synthase, and starch-branching enzymes are quantitatively associated with the chloroplasts. The starch degradative enzymes amylase, R-enzyme (debranching activity), phosphorylase, and D-enzyme (transglycosylase) are observed both in the chloroplast and soluble protein fractions, the bulk of the degradative enzyme activities reside in the latter fraction. Chromatography of a chloroplast extract on diethylaminoethyl-cellulose resolves the R- and D-enzymes from amylase and phosphorylase activities although the two latter enzyme activities coeluted. The digestion pattern of amylase with amylopectin as a substrate indicates an endolytic activity but displays properties unlike the typical α-amylase as isolated from endosperm tissue.     

Structural requirements for the binding of phenylglycosides to the surface of protoplasts

A variety of phenylglycosides have been synthesized and tested for binding to the surface of protoplasts from suspension-cultured cells of “Paul's Scarlet” rose (Rosa sp.). Multivalent phenylglycosides in the form of Yariv antigens (1,3,5,-tri-[p-glycosyloxyphenylazo]-2,4,6,-trihydroxybenzene) agglutinated the protoplasts. Fluorescence-labeled derivatives of other monovalent and polyvalent phenyl-β-glycosides did not bind to the protoplast surface. Agglutination was induced by Yariv antigens only if these probes contained β-anomeric, O-glycosidic linkages. Yariv antigens containing α-anomeric or thio-glycosidic linkages did not agglutinate protoplasts. These same structural features of Yariv antigens were also required for the precipitation of gum arabic-Yariv antigen complexes. The results suggest that plasma membranes of “Paul's Scarlet” rose protoplasts contain arabinogalactan-proteins that interact with phenyl-β-glycosides. The results further show that binding at these plasma membrane sites is not solely dependent upon the carbohydrate portion of single phenylglycosides, but may also require specific spatial orientations of adjacent phenylglycosides.     

Starch phosphorylase: Role in starch metabolism and biotechnological applications

The α-glucan phosphorylases of the glycosyltransferase family are important enzymes of carbohydrate metabolism in prokaryotes and eukaryotes. The plant α-glucan phosphorylase, commonly called starch phosphorylase (EC 2.4.1.1), is largely known for the phosphorolytic degradation of starch. Starch phosphorylase catalyzes the reversible transfer of glucosyl units from glucose-1-phosphate to the nonreducing end of α-1,4-d-glucan chains with the release of phosphate. Two distinct forms of starch phosphorylase, plastidic phosphorylase and cytosolic phosphorylase, have been consistently observed in higher plants. Starch phosphorylase is industrially useful and a preferred enzyme among all glucan phosphorylases for phosphorolytic reactions for the production of glucose-1-phosphate and for the development of engineered varieties of glucans and starch. Despite several investigations, the precise functional mechanisms of its characteristic multiple forms and the structural details are still eluding us. Recent discoveries have shed some light on their physiological substrates, precise biological functions, and regulatory aspects. In this review, we have highlighted important developments in understanding the role of starch phosphorylases and their emerging applications in industry.     

The role of cytosolic alpha-glucan phosphorylase in maltose metabolism and the comparison of amylomaltase in Arabidopsis and Escherichia coli

Transitory starch of leaves is broken down hydrolytically, making maltose the predominant form of carbon exported from chloroplasts at night. Maltose metabolism in the cytoplasm of Escherichia coli requires amylomaltase (MalQ) and maltodextrin phosphorylase (MalP). Possible orthologs of MalQ and MalP in the cytosol of Arabidopsis (Arabidopsis thaliana) were proposed as disproportionating enzyme (DPE2, At2g40840) and alpha-glucan phosphorylase (AtPHS2, At3g46970). In this article, we measured the activities of recombinant DPE2 and AtPHS2 proteins with various substrates; we show that maltose and a highly branched, soluble heteroglycan (SHG) are excellent substrates for DPE2 and propose that a SHG is the in vivo substrate for DPE2 and AtPHS2. In E. coli, MalQ and MalP preferentially use smaller maltodextrins (G(3)-G(7)) and we suggest that MalQ and DPE2 have similar, but nonidentical, roles in maltose metabolism. To study this, we complemented a MalQ(-) E. coli strain with DPE2 and found that the rescue was not complete. To investigate the role of AtPHS2 in maltose metabolism, we characterized a T-DNA insertion line of the AtPHS2 gene. The nighttime maltose level increased 4 times in the Atphs2-1 mutant. The comparison of maltose metabolism in Arabidopsis with that in E. coli and the comparison of the maltose level in plants lacking DPE2 or AtPHS2 indicate that an alternative route to metabolize the glucan residues in SHG exists. Other plant species also contain SHG, DPE2, and alpha-glucan phosphorylase, so this pathway for maltose metabolism may be widespread among plants.     

Calmodulin and calmodulin-binding proteins in liver cell nuclei

Three nuclear subfractions were prepared from isolated hepatocytes nuclei. The calmodulin content in whole nuclei was 79 ng/mg of protein. The soluble fraction obtained after digestion of the nuclei with DNase I and RNase A (S1 fraction) contained 252 ng of calmodulin/mg of protein. The pellet obtained after the digestion with nucleases was treated with 1.6 M NaCl, and the soluble fraction and the residual structures obtained after the treatment were called S2 fraction and nuclear matrix, respectively. The calmodulin contents of the S2 fraction and of the nuclear matrix were 68 and 190 ng/mg of protein, respectively. If nuclei were digested only with DNase I, the calmodulin content in the soluble fraction increased to 703 ng/mg of protein, indicating that part of the nuclear calmodulin is associated with active DNA. Five nuclear calmodulin-binding proteins were identified. Two, having apparent molecular masses of 240 and 150 kDa were only found in the nuclear matrix, whereas the other three, having molecular masses of 120, 65, and 40 kDa were found in different proportions in all nuclear subfractions. A calmodulin-dependent inhibition of protein phosphorylation in the S1 fraction was discovered. Purification attempts on the calmodulin-binding proteins of the S1 subfraction by calmodulin affinity chromatography yielded four major polypeptides with apparent molecular masses of about 41, 46, and 120 (two products) kDa. These polypeptides retained the ability to inhibit protein phosphorylation but not the sensitivity to calmodulin.     

Maltose metabolism by pea chloroplasts

The aim of this work was to investigate the origin of maltose formed during starch breakdown in the dark by chloroplasts of Pisum sativum. The maximum catalytic activities of maltose phosphorylase and maltase in pea leaves were shown to be low, relative to those of enzymes known to be involved in starch breakdown. Fractionation of pea leaves indicated that the chloroplasts lack maltase but have enough maltose phosphorylase to synthesize the amounts of maltose formed when isolated chloroplasts breakdown starch. The absence of exogenous phosphate markedly reduced starch breakdown and maltose accumulation by isolated chloroplasts. When [¹⁴C]glucose was supplied to chloroplasts that were breaking down starch in the dark, maltose was labelled and most of the label was in the glucose moeity. It is suggested that maltose phosphorylase, using glucose-1-phosphate formed from starch by α-glucan phosphorylase, is responsible for, at least some of, the synthesis of maltose during starch breakdown by pea chloroplasts in vitro.     

Activity and expression of banana starch phosphorylases during fruit development and ripening

Two main forms of starch phosphorylase (EC 2.4.1.1) were identified and purified from banana (Musa acuminata Colla. cv. Nanic?o) fruit. One of them, designated phosphorylase I, had a native molecular weight of 155 kDa and subunit of 90 kDa, a high affinity towards branched glucans and an isoelectric point around 5.0. The other, phosphorylase II, eluted at a higher salt concentration from the anion exchanger, had a low affinity towards branched glucans, a native molecular weight of 290 kDa and subunit of 112 kDa. Kinetic studies showed that both forms had typical hyperbolic curves for orthophosphate (Pi) and glucose-1-phosphate, and that they could not react with substrates with a blocked reducing end or alpha-1,6 glucosidic bonds. Antibodies prepared against the purified type-II form and cross-reacting with the type-I form showed that there was an increase in protein content during development and ripening of the fruit. The changes in protein level were parallel to those of phosphorylase activity, in both the phosphorolytic and synthetic directions. Considering the kinetics, indicating that starch phosphorylases are not under allosteric control, it can be argued that protein synthesis makes a contribution to regulating phosphorylase activity in banana fruit and that hormones, like gibberellic acid and indole-3-acetic acid, may play a regulating role. For the first time, starch phosphorylases isoforms were detected as starch-granule-associated proteins by immunostaining of SDS-PAGE gels.     

SUBCELLULAR DISTRIBUTION OF NORADRENALINE IN GUINEA-PIG CEREBRAL CORTEX AND SPLEEN

Abstract— The distribution of noradrenaline (NA) in subcellular fractions of guinea-pig cerebral cortex and spleen was determined by differential and density gradient centrifugation. Of the primary fractions, the microsomal fraction from both tissues was enriched in NA, that of the spleen having the higher specific activity. Microsomal fractions were therefore placed on gradients and NA determined in the subfractions since these fractions appeared suitable preparations in which to search for discrete populations of vesicles. So that the non-occluded micro-particulate bound noradrenaline (MPBNA) content of gradient subfractions could be measured, [³H]NA was used to control for the diffusion and or adsorption of free NA, and occluded lactate dehydrogenase was used to estimate the amount of entrapped MPBNA and soluble NA. Non-occluded MPBNA on gradients from microsomal fractions of cerebral cortex formed a single peak mainly in subfraction F (0.6-0.8 m -sucrose). Spleen microsomal fractions, however yielded two peaks of MPBNA. one in sub-fractions D to G (0.4-1.0 m -sucrose) and the other in sub-fraction J (1.4 m -sucrosc); electron microscopy showed that the latter subfraction contained large vesicles.
Since there were unexpectedly small amounts of MPBNA in microsomal subfractions D and E of cerebral cortex, the synaptosome fraction was investigated. Following water treatment of synaptosomes. MPBNA formed a peak in subfraction E (0.4-0.6 m -sucrose) with smaller amounts in subfractions D and F (0.4 and 0.6 0.8 m -sucrose).     

LOCALIZATION OF STARCH BIOSYNTHETIC AND DEGRADATIVE ENZYMES IN MAIZE LEAVES

The cellular distribution of the starch biosynthetic and degradative enzymes in protoplasts prepared from maize leaf mesophyll and bundle sheath cells was investigated. In conformity with the cellular distribution of starch, starch biosynthetic enzymes (soluble starch synthase, ADPglucose pyrophosphorylase, branching enzyme and starch Phosphorylase) were exclusively localized in the bundle sheath cells. In contrast, starch degradative enzymes (α-amylase, β-amylase and debranching enzyme) were present in both types of leaf cells. Isolated chloroplasts from bundle sheath cells were shown to contain 100% of the starch biosynthetic enzymes. However, approximately 60% of the activity of degradative enzymes and 67% of the activity of starch Phosphorylase was localized in bundle sheath chloroplasts.     

Enzymes of carbohydrate metabolism in the developing endosperm of maize

A number of enzymes presumably implicated in starch synthesis were assayed at various stages of endosperm development ranging from 8 days to 28 days after pollination. Activity for invertase, hexokinase, the glucose phosphate isomerases, the phosphoglucomutases, phosphorylase I, uridine diphosphate glucose pyrophosphorylase, and the starch granule-bound nucleoside diphosphate glucose-starch glucosyltransferase was present at the earliest stage of development (8 days) studied. Activity was detectable for phosphorylase III, the soluble adenosine diphosphate glucose-starch glucosyltransferase, adenosine diphosphate glucose pyrophosphorylase, and sucrose-uridine diphosphate glucosyltransferase at 12 days. For phosphorylase II and cytidine diphosphate glucose pyrophosphorylase, activity was first detectable at the 14- and 16-day stages, respectively. Rapid increases in starch content are observed prior to detectable activity for adenosine diphosphate glucose pyrophosphorylase, the soluble adenosine diphosphate glucose-starch glucosyltransferase and phosphorylases II and III. For all enzymes, except invertase, activity per endosperm rises to a peak at 22 or 28 days. Greatest activity for invertase is found at 12 days with a steady decline thereafter. The pattern of invertase activity in comparison with that of sucrose-uridine diphosphate glucosyltransferase supports previous suggestions, that the latter plays a key role in the conversion of sucrose to starch. In addition to phosphorylases I, II, and III, multiple forms of glucosephosphate isomerase and phosphoglucomutase were detected.     

Uridinediphosphate-glucose: Isovitexin 7-O-glucosyltransferase from barley protoplasts: Subcellular localization

Protoplasts isolated from 6-d-old primary leaves of barley (Hordeum vulgare L.) contain an enzyme which transfers the glucosyl moiety of uridine-diphosphateglucose to isovitexin, resulting in the formation of saponarin, the major flavonoid of barley. Purified chloroplasts isolated from protoplasts contained less than 2% of the total glucosyltransferase activity. These chloroplasts were 97% intact, based on ribulose-bisphosphate-carboxylase activity. Similarly low levels of glucosyltransferase activity were found in mitochondria and microbody or microsomal preparations from protoplasts. The soluble fraction (cytosol) contained at least 93% of the isovitexin 7-O-glucosyltransferase activity.     

Alterations in cytosolic glucose-phosphate metabolism affect structural features and biochemical properties of starch-related heteroglycans

The cytosolic pools of glucose-1-phosphate (Glc-1-P) and glucose-6-phosphate are essential intermediates in several biosynthetic paths, including the formation of sucrose and cell wall constituents, and they are also linked to the cytosolic starch-related heteroglycans. In this work, structural features and biochemical properties of starch-related heteroglycans were analyzed as affected by the cytosolic glucose monophosphate metabolism using both source and sink organs from wild-type and various transgenic potato (Solanum tuberosum) plants. In leaves, increased levels of the cytosolic phosphoglucomutase (cPGM) did affect the cytosolic heteroglycans, as both the glucosyl content and the size distribution were diminished. By contrast, underexpression of cPGM resulted in an unchanged size distribution and an unaltered or even increased glucosyl content of the heteroglycans. Heteroglycans prepared from potato tubers were found to be similar to those from leaves but were not significantly affected by the level of cPGM activity. However, external glucose or Glc-1-P exerted entirely different effects on the cytosolic heteroglycans when added to tuber discs. Glucose was directed mainly toward starch and cell wall material, but incorporation into the constituents of the cytosolic heteroglycans was very low and roughly reflected the relative monomeric abundance. By contrast, Glc-1-P was selectively taken up by the tuber discs and resulted in a fast increase in the glucosyl content of the heteroglycans that quantitatively reflected the level of the cytosolic phosphorylase activity. Based on (14)C labeling experiments, we propose that in the cytosol, glucose and Glc-1-P are metabolized by largely separated paths.     

Identification of a novel heteroglycan-interacting protein, HIP 1.3, from Arabidopsis thaliana

Plastidial degradation of transitory starch yields mainly maltose and glucose. Following the export into the cytosol, maltose acts as donor for a glucosyl transfer to cytosolic heteroglycans as mediated by a cytosolic transglucosidase (DPE2; EC 2.4.1.25) and the second glucosyl residue is liberated as glucose. The cytosolic phosphorylase (Pho2/PHS2; EC 2.4.1.1) also interacts with heteroglycans using the same intramolecular sites as DPE2. Thus, the two glucosyl transferases interconnect the cytosolic pools of glucose and glucose 1-phosphate. Due to the complex monosaccharide pattern, other heteroglycan-interacting proteins (HIPs) are expected to exist.Identification of those proteins was approached by using two types of affinity chromatography. Heteroglycans from leaves of Arabidopsis thaliana (Col-0) covalently bound to Sepharose served as ligands that were reacted with a complex mixture of buffer-soluble proteins from Arabidopsis leaves. Binding proteins were eluted by sodium chloride. For identification, SDS-PAGE, tryptic digestion and MALDI-TOF analyses were applied. A strongly interacting polypeptide (approximately 40 kDa; designated as HIP1.3) was observed as product of locus At1g09340. Arabidopsis mutants deficient in HIP1.3 were reduced in growth and contained heteroglycans displaying an altered monosaccharide pattern. Wild type plants express HIP1.3 most strongly in leaves. As revealed by immuno fluorescence, HIP1.3 is located in the cytosol of mesophyll cells but mostly associated with the cytosolic surface of the chloroplast envelope membranes. In an HIP1.3-deficient mutant the immunosignal was undetectable. Metabolic profiles from leaves of this mutant and wild type plants as well were determined by GC-MS. As compared to the wild type control, more than ten metabolites, such as ascorbic acid, fructose, fructose bisphosphate, glucose, glycine, were elevated in darkness but decreased in the light. Although the biochemical function of HIP1.3 has not yet been elucidated, it is likely to possess an important function in the central carbon metabolism of higher plants.     

Dynamic phosphorylation of RNA polymerase III holoenzyme subunits from human epidermoid carcinoma cells A431 cultivated under different conditions

Dynamic phosphorylation in vivo of individual subunits of holoenzyme RNA polymerase III subfractions IIIa and IIIb on serine/threonine and tyrosine aminoacid residues has been shown in human epidermoid carcinoma cells A431. Cells cultivation under starvation on embryonic serum caused a prolongation of the cell cycle, while cultivation at low concentration of epidermal growth factor (EGF) activated cell proliferation. Subunit 45 kDa is phosphorylated on serine/threonine residues in subfraction IIIa only in starving cells. This subunit of subfraction IIIb is unphosphorylated. Phosphorylation of this particular subunit is restored under induction by EGF at low concentration (0.1 ng/ml). The level of phosphorylation on tyrosine aminoacid residues of subunits of holoenzyme subfractions IIIa and IIIb is high in cells cultivated under starvation. Subunit 60 kDa has a higher level of phosphorylation as compared with subunits 45 and 38 kDa. Induction by EGF at low concentrations increases the level of phosphorylation of subunits 60 and 38 kDa in both subfractions of holoenzyme.