Autologous lacrimal-lymphoid mixed-cell reactions induce dacryoadenitis in rabbits

Autoimmune dacryoadenitis, such as occurs in Sjögren's syndrome, is a frequent cause of lacrimal insufficiency, which in turn can cause dry eye. Rabbits are used frequently to test ocular therapies. Our goal is to develop a rabbit model of autoimmune dacryoadenitis to identify and test candidate therapies. Our approach arises from the observations that lacrimal gland epithelial cells stimulate proliferation in cultured autologous lymphocyte preparations and that an anti-MHC II antibody blocks this proliferation. The purpose of this study was to determine if injecting this proliferating autologous mixed cell reaction could induce dacryoadenitis in rabbits. After establishing that irradiated lacrimal gland epithelial cells stimulate proliferation in autologous peripheral blood lymphocytes, irradiated cells from a single lacrimal gland were co-cultured with autologous lymphocytes and after 5 days the mixed cell reaction, or components of the reaction, were injected into the contralateral lacrimal gland of the donor rabbit. After 2 weeks, the injected glands were removed and lymphocytic infiltration quantitated using digital image analysis of immunostained histological sections. Injecting an autologous mixed cell reaction of co-cultured irradiated lacrimal gland epithelial cells and lymphocytes reliably induced abundant periductal foci of >200 lymphocytes expressing CD18 and/or a rabbit thymic lymphocyte antigen (RTLA). Injection of medium or autologous lymphocytes alone elicited little response; injections of lymphocytes cultured with lysates of lacrimal gland epithelial cells elicited variable, modest responses. These lysates did not stimulate proliferation in the mixed cell reaction and proliferation was not observed if a porous membrane separated co-cultured lacrimal gland cells and lymphocytes. The results demonstrate that injecting an autologous mixed cell reaction of lacrimal gland epithelial cells and lymphocytes reliably creates a model of autoimmune dacryoadenitis. The relative ineffectiveness of components of the reaction to do the same supports the hypothesis that lacrimal gland epithelial cells trigger or exacerbate lacrimal autoimmune disease by presentation of autoantigens via MHC II. This experimental system can aid efforts to further understand mechanisms of diseases, and to identify and test candidate therapies.

Lacrimal gland epithelial cells stimulate proliferation in autologous lymphocyte preparations

Autoimmune dacryoadenitis is a frequent cause of lacrimal insufficiency. In order to test hypotheses regarding mechanisms that can trigger this syndrome, we developed a method to obtain a preparation of rabbit lacrimal gland epithelial cells essentially free of immune-system cells. The method relies on controlled digestion to disperse lacrimal acini, and recovers acini by filtration through various sizes of nylon mesh. Purity and integrity of the preparation were established qualitatively using light and electron microscopy. Contamination by immune-system cells was quantitated by immunohistochemistry using anti-CD18, and -RTLA (rabbit thymic lymphocyte antigen) antibodies. The novel method produced preparations of highly-purified lacrimal gland epithelial cells (pLGEC) with expected morphological characteristics with less than 1.5% of the cells staining for CD18 or RTLA. The method also yielded preparations of lacrimal gland interstitial cells (LGIC) enriched for lymphocytes; in these preparations either CD18 or RTLA were detected on nearly 10% of the cells. pLGEC promoted proliferation in preparations of autologous splenic lymphocytes (SPL) that was blocked by anti-MHC class II but not anti-MHC class I antibodies. This observation, combined with the apparent requirement that pLGEC must contact the autologous lymphocyte preparation to promote proliferation, supports the hypothesis the proliferation arises from antigen-presentation via MHC class II by pLGEC.

Sjögren's autoimmunity: how perturbation of recognition in endomembrane traffic may provoke pathological recognition at the cell surface

CD4 T cell antigen recognition requires presentation by major histocompatibility complex Class II molecules (MHC II). B cell surface immunoglobulins recognize antigens independently of MHC II, but activation typically requires CD4 cell cytokines as accessory signals. Plasma membrane-endomembrane traffic in lacrimal gland acinar cells, targets of autoimmune activity in Sjögren's syndrome, may satisfy both requirements. The Golgi protein galactosyltransferase and the lysosomal proteins cathepsin B and cathepsin D appear at the plasma membranes during sustained secretomotor stimulation. The RNA transcription termination factor La, a frequent target of Sjögren's autoantibodies, appears in the acinar cell cytoplasm and plasma membranes during viral infection and during in vitro exposure to cytokines. MHC II cycle through endomembrane compartments which contain La, galactosyltransferase, cathepsin B and cathepsin D and which are sites of proteolysis. This traffic may permit trilateral interactions in which B cells recognize autoantigens at the surface membranes, CD4 T cells recognize peptides presented by MHC II, B cells provide accessory signals to CD4 T cells, and CD4 T cells provide cytokines that activate B cells. Acinar cells stimulate lymphocyte proliferation in autologous mixed cell reactions, confirming that they are capable of provoking autoimmune responses.

A method to study induction of autoimmunity in vitro: co-culture of lacrimal cells and autologous immune system cells

Co-culturing autologous lacrimal gland cells and immune system cells can lead to spleen cell proliferation with a time course similar to that for proliferation in a typical heterologous MLR. Although these results are consistent with the hypothesis that lacrimal acinar cells are a source of antigen, and may or may not serve in part as an APC, future studies of this preparation are required to test these hypotheses. We are unaware of reports demonstrating that co-culturing control epithelial tissue and autologous splenic lymphocytes from apparently healthy animals leads to lymphocytic proliferation. Our results suggest that the appropriate co-culture of tissues and immune cells from healthy animals, perhaps such as detailed above, should help identify mechanisms contributing to the induction of autoimmune disease. Knowledge regarding such mechanisms should help efforts to prevent such disease, and perhaps reverse it.

Regulation of cytotoxic T cells by ecto-nicotinamide adenine dinucleotide (NAD) correlates with cell surface GPI-anchored/arginine ADP-ribosyltransferase

This report demonstrates that incubation of cytotoxic T cells with NAD causes suppression of their ability to proliferate in response to stimulator cells or to lyse targets. Effects are evident after incubation for 3 h with concentrations of NAD as low as 1 microM and are sustained for many hours after removal of NAD from culture media. Suppression is a result of the failure of CTL to form specific conjugates with targets as well as a lower level of activation in response to TCR-mediated stimulation, although TCR-mediated transmembrane signaling is demonstrable. Metabolites of NAD such as nicotinamide, ADP-ribose, and cyclic-ADP-ribose have no detectable effect, indicating that NAD-glycohydrolase or ADP-ribose cyclase do not mediate suppression. Incubation of intact CTL with [32P]NAD leads to incorporation of 32P into a particulate, subcellular fraction, a reaction that is not inhibitable by ADP-ribose. Hydroxylamine, but not mercuric ion releases [32P]ADP-ribose, whereas phosphodiesterase releases [32P]AMP from the particulate subcellular fraction, suggesting that labeling is a result of enzymatic mono-ADP-ribosylation of arginines. In support of this, treatment of intact CTL with phosphatidylinositol-specific phospholipase C releases an arginine-specific ADP-ribosyltransferase and causes insensitivity to ecto-NAD suppression. These results suggest that a GPI-anchored ADP-ribosyltransferase uses ecto-NAD to ADP-ribosylate proteins that regulate CTL function.

Regulation of cytotoxic T cells by ecto-nicotinamide adenine dinucleotide (NAD) correlates with cell surface GPI-anchored/arginine ADP-ribosyltransferase

This report demonstrates that incubation of cytotoxic T cells with NAD causes suppression of their ability to proliferate in response to stimulator cells or to lyse targets. Effects are evident after incubation for 3 h with concentrations of NAD as low as 1 microM and are sustained for many hours after removal of NAD from culture media. Suppression is a result of the failure of CTL to form specific conjugates with targets as well as a lower level of activation in response to TCR-mediated stimulation, although TCR-mediated transmembrane signaling is demonstrable. Metabolites of NAD such as nicotinamide, ADP-ribose, and cyclic-ADP-ribose have no detectable effect, indicating that NAD-glycohydrolase or ADP-ribose cyclase do not mediate suppression. Incubation of intact CTL with [32P]NAD leads to incorporation of 32P into a particulate, subcellular fraction, a reaction that is not inhibitable by ADP-ribose. Hydroxylamine, but not mercuric ion releases [32P]ADP-ribose, whereas phosphodiesterase releases [32P]AMP from the particulate subcellular fraction, suggesting that labeling is a result of enzymatic mono-ADP-ribosylation of arginines. In support of this, treatment of intact CTL with phosphatidylinositol-specific phospholipase C releases an arginine-specific ADP-ribosyltransferase and causes insensitivity to ecto-NAD suppression. These results suggest that a GPI-anchored ADP-ribosyltransferase uses ecto-NAD to ADP-ribosylate proteins that regulate CTL function.

Construction and characterization of recombinant Vibrio cholerae strains producing inactive cholera toxin analogs

The catalytic A subunit of cholera toxin (CT-A) is capable of ADP-ribosylating the guanine nucleotide-binding protein, which regulates cell adenylyl cyclase, leading to the life-threatening diarrhea of cholera. Amino acids involved in the enzymatic activity of CT-A have previously been identified. By means of site-directed mutagenesis, an analog of the CT-A subunit gene was created with codon substitutions for both Arg-7 and Glu-112, each of which has been shown to produce subunits lacking ADP-ribosyltransferase activity. The mutated gene fragment was exchanged for the wild-type copy in the previously cloned ctxAB operon from El Tor biotype, Ogawa serotype Vibrio cholerae strain 3083, which produces CT-2. Further, the zonula occludens toxin gene, zot, was inactivated by an insertional mutation to create the new plasmid construct pCT-2*. Additionally, a DNA fragment encoding the B subunit of CT-1 (CT produced by classical biotype, Inaba serotype V. cholerae strain 569B) was exchanged for the homologous part in pCT-2*, resulting in the creation of pCT-1*. These plasmid constructs were introduced into the CT-negative V. cholerae mutant strain JBK70 (E1 Tor biotype, Inaba serotype); CT-A-B+ derivatives CVD101 and CVD103 of classical biotype Ogawa and Inaba serotype strains 395 and 569B, respectively; El Tor biotype Inaba and Ogawa serotype strains C6706 and C7258, respectively, recently isolated in Peru; and O139 (synonym Bengal) strain SG25-1 from the current epidemic in India. Recombinant toxins (CT-1* and CT-2*), partially purified from culture supernatants of transformed JBK70, were shown to be inactive on mouse Y1 adrenal tumor cells and in an in vitro ADP-ribosyltransferase assay. CT-1* and CT-2* reacted with polyclonal and monoclonal antibodies against both A and B subunits of CT. The toxin analogs reacted with antibodies against CT-A and CT-B on cellulose acetate strips and in a GM1 enzyme-linked immunosorbent assay; they reacted appropriately with B-subunit epitype-specific monoclonal antibodies in checkerboard immunoblots, and they formed precipitin bands with GM1-ganglioside in Ouchterlony tests. However, the reactions of the modified proteins with anti-A-subunit monoclonal antibodies were weaker than the reactions with wild-type holotoxins. V, cholerae strains carrying ctxA*, with either ctxB-1 or ctxB-2, and inactivated zot genes were created by homologous recombination. The recombinant strains and the purified toxin analogs were inactive in the infant rabbit animal model.(ABSTRACT TRUNCATED AT 400 WORDS)

CTX-B inhibits CTL cytotoxicity and cytoskeletal movements

Binding of cytotoxic T lymphocytes (CTL) to specific targets induces cytoskeletal movements in the effector cell followed by delivery of the lethal hit which ultimately results in target cell lysis. The question whether movement of the cytoskeleton in CTL are obligatory for delivery of the lethal hit is not resolved. Here we report that the CTX-B subunit of cholera toxin which is devoid of the catalytic CTX-A subunit inhibits CTL function. Inhibition was found not to be due to interference with TCR expression, CTL-target conjugate formation, target induced transmembrane signalling or secretion of BLT-esterase. CTX-B however does interfere with F-actin patch formation at the effector target binding site and inhibits reorientation of the microtubule organizing center and Golgi apparatus towards the target binding site. It is concluded that interference with cytoskeletal movements is responsible for inhibition of cytolysis pointing to an important role of the cytoskeleton in the lytic reaction.

Pertussis toxin and target eukaryotic cells: binding, entry, and activation

Pertussis toxin, a protein virulence factor produced by Bordetella pertussis, is composed of an A protomer and a B oligomer. The A protomer consists of a single polypeptide, termed the S1 subunit, which disrupts transmembrane signaling by ADP-ribosylating eukaryotic G-proteins. The B oligomer, containing five polypeptides, binds to cell receptors (most likely containing carbohydrate) and delivers the S1 subunit. Current knowledge suggests that expression of ADP-ribosyltransferase activity in target eukaryotic cells arises after 1) nucleotides and membrane lipids allosterically promote the release of the S1 subunit; and 2) the single disulfide bond in the S1 subunit is reduced by reductants such as glutathione. This model suggests conditions for the proper use of the toxin as an experimental reagent.

Detection of antibodies inhibiting the ADP-ribosyltransferase activity of pertussis toxin in human serum

Bordetella pertussis produces a protein virulence factor termed pertussis toxin. Many candidate pertussis vaccines are based on the rationale that an immune response that neutralizes the virulence activities of this toxin, which are thought to arise from its catalytic ADP-ribosyltransferase activity, would be beneficial. The report describes two methods that quantify the inhibition of this activity by human serum. One, termed a direct assay, involves an initial incubation of toxin with serum, a second incubation that activates the toxin, and a third incubation that measures the ADP-ribosyltransferase activity of the mixture. The other assay, termed a plate assay, involves immobilization of the toxin, exposure of the immobilized toxin to serum and washing of the plate, and then activation and assay of the toxin's ADP-ribosyltransferase activity. The plate assay may be more selective than the direct assay in terms of identifying antibodies that neutralize the toxin in vivo. Sera from controls, selected patients presenting with cough, and vaccinated infants were first analyzed by the direct assay. In contrast to sera from controls, sera from several of the patients and vaccinated infants strongly inhibited activity. Dose-response curves of inhibition were determined for samples from three vaccinated infants by both the direct and plate assays. One of the samples had a dose-response curve of a different shape and thus differed not only in titer but also in functional characteristics. A comparison of inhibition of ADP-ribosyltransferase activity and neutralization in a CHO cell assay indicated that there was incomplete agreement between the two assays. Taken together, these results indicate that measurement of inhibition of ADP-ribosyltransferase activity by human serum is practical and may be useful in the evaluation of responses to pertussis vaccines.

Site-specific mutagenesis of the catalytic subunit of cholera toxin: substituting lysine for arginine 7 causes loss of activity

Cholera and pertussis toxins each contain a subunit with ADP-ribosyltransferase activity, sharing a region of nearly identical amino acid sequence near the NH2 terminus. Previous investigations have shown that substitution of a lysine residue for Arg-9 in the catalytic A subunit of pertussis toxin substantially eliminates its enzyme activity. We now report that substitution of lysine for the position-equivalent Arg-7 of cholera toxin subunit A leads to a similar loss of catalytic activity. This result suggests a correlation of function with structure between the sequence-related cholera and pertussis toxin A subunits and may contribute to the design of a vaccine containing an enzymatically inert analog of cholera toxin.

Monoclonal antibodies that inhibit ADP-ribosyltransferase but not NAD-glycohydrolase activity of pertussis toxin

Kenimer et al. (J. G. Kenimer, J. Kim, P. G. Probst, C. R. Manclark, D. G. Burstyn, and J. L. Lowell, Hybridoma 8:37-51, 1989) identified three classes of monoclonal antibodies, termed A, B, and C, that recognize the S1 subunit of pertussis toxin. This report presents data demonstrating that class A monoclonal antibodies (3CX4, 6D11C, and 3C4D), which block the ADP-ribosyltransferase activity and recognize the predominant neutralizing epitope on the S1 subunit of the toxin, do not inhibit the NAD-glycohydrolase activity of the toxin. In addition, alkylation of cysteine 41 of the S1 subunit, which may interact with NAD, inactivates the toxin but does not prevent binding by class A antibodies. Taken together, these results support the conclusion that proper alterations of amino acids that interact with NAD should allow for inactivation of the toxin without destruction of the predominant neutralizing epitope. The class A antibodies recognized control but not heat-treated pertussis toxin spotted onto nitrocellulose, indicating that class A antibodies do not recognize denatured S1 subunit. In contrast, a nonneutralizing class C antibody (X2X5) failed to bind to control toxin or S1 subunit in solution and recognized heat-treated pertussis toxin better than control toxin when spotted onto nitrocellulose. Thus, this type of analysis presents a heterogeneous mixture of fully or partially denatured and native S1 proteins and fails to distinguish between neutralizing and nonneutralizing antibodies.

Radiometric assays for glycerol, glucose, and glycogen

We have developed radiometric assays for small quantities of glycerol, glucose and glycogen, based on a technique described by Thorner and Paulus (1971, J. Biol. Chem. 246, 3885-3894) for the measurement of glycerokinase activity. In the glycerol assay, glycerol is phosphorylated with [32P]ATP and glycerokinase, residual [32P]ATP is hydrolyzed by heating in acid, and free [32P]phosphate is removed by precipitation with ammonium molybdate and triethylamine. Standard dose-response curves were linear from 50 to 3000 pmol glycerol with less than 3% SD in triplicate measurements. Of the substances tested for interference, only dihydroxyacetone gave a slight false positive signal at high concentration. When used to measure glycerol concentrations in serum and in media from incubated adipose tissue, the radiometric glycerol assay correlated well with a commonly used spectrophotometric assay. The radiometric glucose assay is similar to the glycerol assay, except that glucokinase is used instead of glycerokinase. Dose response was linear from 5 to 3000 pmol glucose with less than 3% SD in triplicate measurements. Glucosamine and N-acetylglucosamine gave false positive signals when equimolar to glucose. When glucose concentrations in serum were measured, the radiometric glucose assay agreed well with hexokinase/glucose-6-phosphate dehydrogenase (H/GDH)-based and glucose oxidase/H2O2-based glucose assays. The radiometric method for glycogen measurement incorporates previously described isolation and digestion techniques, followed by the radiometric assay of free glucose. When used to measure glycogen in mouse epididymal fat pads, the radiometric glycogen assay correlated well with the H/GDH-based glycogen assay. All three radiometric assays offer several practical advantages over spectral assays.

Alkylation of cysteine 41, but not cysteine 200, decreases the ADP-ribosyltransferase activity of the S1 subunit of pertussis toxin

Sulfhydryl-alkylating reagents are known to inactivate the NAD glycohydrolase and ADP-ribosyltransferase activities of the S1 subunit of pertussis toxin, a protein which contains two cysteines at positions 41 and 200. It has been proposed that NAD can retard alkylation of one of the two cysteines of this protein (Kaslow, H.R., and Lesikar, D.D. (1987) Biochemistry 26, 4397-4402). We now report that NAD retards the ability of these alkylating reagents to inactivate the S1 subunit. In order to determine which cysteine is protected by NAD, we used site-directed mutagenesis to construct analogs of the toxin with serines at positions 41 and/or 200. Sulfhydryl-alkylating reagents reduced the ADP-ribosyltransferase activity of the analog with a single cysteine at position 41; NAD retarded this inactivation. In contrast, sulfhydryl-alkylating reagents did not inactivate analogs with serine at position 41. An analog with alanine at position 41 possessed substantial ADP-ribosyltransferase activity. We conclude that alkylation of cysteine 41, and not cysteine 200, inactivates the S1 subunit of pertussis toxin, but that the sulfhydryl group of cysteine 41 is not essential for the ADP-ribosyltransferase activity of the toxin. These results suggest that the region near cysteine 41 contributes to features of the S1 subunit important for ADP-ribosyltransferase activity. Using site-directed mutagenesis, we found that changing aspartate 34 to asparagine, arginine 39 to lysine, and glutamine 42 to glutamate had little effect on ADP-ribosyltransferase activity. However, substituting an asparagine for the histidine at position 35 markedly decreased, but did not eliminate, ADP-ribosyltransferase activity. Chou-Fasman analysis predicted no significant modifications in secondary structure of the S1 peptide with the change of histidine 35 to asparagine. Thus, histidine 35 may interact with a substrate of the S1 subunit without being essential for catalysis.

Regulation of glycogen synthase in muscle and adipose tissue during fasting and refeeding

Using immunoblot analysis, we examined the electrophoretic mobility of glycogen synthase from rat skeletal muscle and adipose tissue. Extracts from muscle freeze clamped in situ contained at least three forms of synthase with different electrophoretic mobilities. Extracts from adipose tissue also contained multiple forms but lacked the form with greatest mobility found in the muscle extracts. Phosphorylation at multiple sites of glycogen synthase is known to deactivate the enzyme and retard its electrophoretic mobility in sodium dodecyl sulfate gels. These results suggest that there is very little or no dephosphorylated glycogen synthase in adipose tissue and that phosphorylated forms of glycogen synthase synthesize adipose tissue glycogen. Relative to control, it is known that fasting decreases and refeeding increases glucose incorporation into glycogen in rat epididymal adipose tissue but not skeletal muscle incubated in vitro in the presence of insulin. Fasting did not change the electrophoretic pattern of muscle synthase but decreased the relative amount of adipose tissue forms with greater mobility. Refeeding increased above control the relative amount of adipose tissue synthase with greater mobility. These results indicate that changes in the phosphorylations that retard mobility contribute to the effects of fasting and refeeding on adipose tissue glycogen metabolism.

Fructose effect to suppress hepatic glycogen degradation

The effect of fructose on glycogen degradation was examined by measuring the flux of 14C from prelabeled glycogen in perfused rat livers. During 2-h refeeding of 24-h-fasted rats, newly synthesized hepatic glycogen was labeled by intraperitoneal injection of [U-14C] galactose (0.1 mg and 0.02 microCi/g of body weight). The livers of refed rats were then perfused in a nonrecirculating fashion for an initial 30 min with glucose alone (10 mM) for the following 60 min with glucose (10 mM) without (n = 5) or with fructose (1, 2, or 10 mM; n = 5 for each). When livers were exposed to fructose, release of label into the perfusate immediately declined and remained markedly suppressed through the end of perfusion (p less than 0.05). The suppression was dose-dependent; at steady state (50-70 min), label release was suppressed 45, 64, and 72% by 1, 2, and 10 mM fructose, respectively (p less than 0.0001). Suppression was not accompanied by significant changes in the activities of glycogen synthase or phosphorylase assessed in vitro. These results suggest the existence of allosteric inhibition of phosphorylase in the presence of fructose. Fructose 1-phosphate (Fru-1-P) accumulated in proportion to fructose (0.11 +/- 0.01 without fructose, 0.86 +/- 0.03, 1.81 +/- 0.18, and 8.23 +/- 0.60 mumol/g of liver with 1, 2, and 10 mM fructose, respectively; p less than 0.0001). Maximum inhibition of label release was 82%; the Fru-1-P concentration for half inhibition was 0.57 mumol/g of liver, well within the concentration of Fru-1-P attained during refeeding. We conclude that fructose enhances net glycogen accumulation in liver by suppressing glycogenolysis and that the suppression is presumably caused by allosteric inhibition of phosphorylase by Fru-1-P.

Sulfhydryl-alkylating reagents inactivate the NAD glycohydrolase activity of pertussis toxin

The combination of ATP, CHAPS (3-[(3-cholamidopropyl)dimethylammonio]-1-propane-sulfonate), and DTT (dithiothreitol) is known to promote the expression of the NAD glycohydrolase activity of pertussis toxin, which resides in the toxin's S1 subunit. By monitoring changes in electrophoretic mobility, we have found that ATP and CHAPS act by promoting the reduction of the disulfide bond of the S1 subunit. In addition, ATP, CHAPS, and DTT allowed sulfhydryl-alkylating reagents to inactivate the NAD glycohydrolase activity. In the presence of iodo[14C]acetate, the combination of ATP, CHAPS, and DTT increased 14C incorporation into only the S1 subunit of the toxin, indicating that alkylation of this subunit was responsible for the loss of activity. If iodoacetate is used as the alkylating reagent, alkylation can be monitored by an acidic shift in the isoelectric point of the S1 peptide. Including NAD in alkylation reactions promoted the accumulation of a form of the S1 peptide with an isoelectric point intermediate between that of native S1 and that of S1 alkylated in the absence of NAD. This result suggests that NAD interacts with one of the two cysteines of the S1 subunit. In addition, we found the pH optimum for the NAD glycohydrolase activity of pertussis toxin is 8, which may reflect the participation of a cysteine in the catalytic mechanism of the toxin.

Structure-activity analysis of the activation of pertussis toxin

Bordetella pertussis, the causative agent of whooping cough, releases pertussis toxin in an inactive form. The toxin consists of an A protomer containing one S1 peptide subunit and a B oligomer containing several other peptide subunits. The toxin binds to cells via the B oligomer, and the S1 subunit is activated and expresses ADP-ribosyltransferase and NAD glycohydrolase activities. Treatment of purified toxin with dithiothreitol (DTT) in vitro increases both activities. ATP and the detergent 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) synergistically reduce the A0.5 (activation constant) for DTT from greater than 100 mM to 200 microM. We studied the structure-activity relationships of activators of the toxin. In the presence of CHAPS (1%) and DTT (10 mM) the following compounds increased the NAD glycohydrolase activity of the toxin with the following A0.5's in microM and fraction of the ATP effect in parentheses: ATP, 0.2 (1.0); ADP, 6 (0.8); UTP, 15 (0.7); GTP, 35 (0.6); pyrophosphate, 45 (0.7); triphosphate, 60 (0.6); tetraphosphate, greater than or equal to 170 (greater than or equal to 0.4). Thus, the polyphosphate moiety is sufficient to stimulate the toxin, and the adenosine moiety confers upon ATP its extraordinary affinity for the toxin. Phospholipid and detergents could substitute for CHAPS in the activation of the toxin. Glutathione substituted for DTT with an A0.5 of 2 mM, a concentration within the range found in eucaryotic cells. Thus, membrane lipids and cellular concentrations of glutathione and ATP are sufficient to activate pertussis toxin without the need for a eucaryotic enzymatic process.

Stimulation of the thiol-dependent ADP-ribosyltransferase and NAD glycohydrolase activities of Bordetella pertussis toxin by adenine nucleotides, phospholipids, and detergents

Pertussis toxin catalyzed ADP-ribosylation of the guanyl nucleotide binding protein transducin was stimulated by adenine nucleotide and either phospholipids or detergents. To determine the sites of action of these agents, their effects were examined on the transducin-independent NAD glycohydrolase activity. Toxin-catalyzed NAD hydrolysis was increased synergistically by ATP and detergents or phospholipids; the zwitterionic detergent 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) was more effective than the nonionic detergent Triton X-100 greater than lysophosphatidylcholine greater than phosphatidylcholine. The A0.5 for ATP in the presence of CHAPS was 2.6 microM; significantly higher concentrations of ATP were required for maximal activation in the presence of cholate or lysophosphatidylcholine. In CHAPS, NAD hydrolysis was enhanced by ATP greater than ADP greater than AMP greater than adenosine; ATP was more effective than MgATP or the nonhydrolyzable analogue adenyl-5'-yl imidodiphosphate. GTP and guanyl-5'-yl imidodiphosphate were less active than the corresponding adenine nucleotides. Activity in the presence of CHAPS and ATP was almost completely dependent on dithiothreitol; the A0.5 for dithiothreitol was significantly decreased by CHAPS alone and, to a greater extent, by CHAPS and ATP. To determine the site of action of ATP, CHAPS, and dithiothreitol, the enzymatic (S1) and binding components (B oligomer) were resolved by chromatography. The purified S1 subunit catalyzed the dithiothreitol-dependent hydrolysis of NAD; activity was enhanced by CHAPS but not ATP. The studies are consistent with the conclusion that adenine nucleotides, dithiothreitol, and CHAPS act on the toxin itself rather than on the substrate; adenine nucleotides appear to be involved in the activation of toxin but not the isolated catalytic unit.

L-type glycogen synthase. Tissue distribution and electrophoretic mobility

We previously reported (Kaslow, H.R., and Lesikar, D.D.FEBS Lett. (1984) 172, 294-298) the generation of antisera against rat skeletal muscle glycogen synthase. Using immunoblot analysis, the antisera recognized the enzyme in crude extracts from rat skeletal muscle, heart, fat, kidney, and brain, but not liver. These results suggested that there are at least two isozymes of glycogen synthase, and that most tissues contain a form similar or identical to the skeletal muscle type, referred to as "M-type" glycogen synthase. We have now used an antiserum specific for the enzyme from liver, termed "L-type" glycogen synthase, to study its distribution and electrophoretic mobility. Immunoblot analysis using this antiserum indicates that L-type glycogen synthase is found in liver, but not skeletal muscle, heart, fat, kidney, or brain. In sodium dodecyl sulfate-polyacrylamide gels of crude liver extracts prepared with protease inhibitors, rat L-type synthase was detected with electrophoretic mobility Mapp = 85,000. In contrast, the M-type enzyme in crude skeletal muscle extracts with protease inhibitors was detected with Mapp = 86,000 and 89,000. During purification of L-type synthase, apparent proteolysis can generate forms with increased electrophoretic mobility (Mapp = 75,000), still recognized by the antiserum. These M-type and L-type antisera did not recognize a protein with Mapp greater than phosphorylase. The anti-rat L-type antisera recognized glycogen synthase in blots of crude extracts of rabbit liver, but with Mapp = 88,000, a value 3,000 greater than that found for the rat liver enzyme. The anti-rat M-type antisera failed to recognize the enzyme in blots of crude extracts of rabbit muscle. Thus, in both muscle and liver, the corresponding rat and rabbit enzymes are structurally different. Because the differences described above persist after resolving these proteins by denaturing sodium dodecyl sulfate electrophoresis, these differences reside in the structure of the proteins themselves, not in some factor bound to the protein in crude extracts.

Adenine nucleotides directly stimulate pertussis toxin

Both cholera toxin and pertussis toxin catalyzed ADP-ribosylation of purified bovine brain tubulin. The effect of cholera toxin was evident in the absence or presence of nucleotides. In contrast, pertussis toxin required adenine nucleotides for its ADP-ribosylating activity. ATP, ATP gamma S, App(NH)p, deoxy-ATP, and ADP all supported pertussis toxin-catalyzed ADP-ribosylations in the absence or presence of EDTA, suggesting that nucleotide hydrolysis was not involved. Adenine nucleotides also promoted pertussis toxin-catalyzed ADP-ribosylation of heat-treated bovine serum albumin. This result suggests that adenine nucleotides directly affect pertussis toxin. ATP stimulation of pertussis toxin-catalyzed hydrolysis of NAD to ADP-ribose supports this hypothesis.

Fasting and diabetes alter adipose tissue glycogen synthase

In a previous report (J. Biol. Chem. 254: 4678-4683, 1979), we showed that fasting blunted the ability of insulin to promote glucose incorporation into glycogen in vitro. In addition, we showed that glycogen synthase activity was altered in two ways: the concentration of glucose 6-P causing half-maximal activation increased, and positive cooperativity appeared in the glucose 6-P activation of the enzyme. We now show that streptozotocin-diabetes causes the same changes in glucose incorporation and glycogen synthase activity. We show that these changes in glycogen synthase activity persist during enzyme purification; thus it is likely the changes are a result of a structural alteration of the enzyme. Because glycogenolysis of a glycogen particle from rabbit skeletal muscle also caused the appearance of positive cooperativity, we propose that both phosphorylation and glycogenolysis are involved in the appearance of positive cooperativity.

Isozymes of glycogen synthase

Antisera to rat skeletal muscle glycogen synthase failed to recognize liver glycogen synthase by electroblot analysis. The antisera recognized the enzyme in skeletal muscle, heart, fat, kidney, and brain. The results support the hypothesis that there are at least two isozymes of glycogen synthase, and that most tissues contain a form similar or identical to the skeletal muscle type. There is a virtual absence of the muscle-type enzyme in adult rat liver.

A direct radioassay for pyruvate kinase activity

Pyruvate kinase catalyzes the conversion of phosphoenolpyruvate (PEP) to pyruvate. A direct radioassay for this enzyme using [14C]PEP as substrate has been developed. The product, [14C]pyruvate, can be separated from the substrate rapidly and easily by applying the mixture to a hydroxyapatite column, and eluting the [14C]pyruvate directly into a scintillation vial. The [14C]PEP is bound to the column which can be regenerated and used indefinitely. The assay is sensitive, rapid, and particularly well suited for the simultaneous assay of large numbers of samples.

Identification by direct photoaffinity labeling of an altered phosphodiesterase in a mutant S49 lymphoma cell

Extracts of a mutant S49 lymphoma cell line, termed K30a, hydrolyze cAMP and cGMP at rates much faster than do wild type S49 extracts. This elevated phosphodiesterase activity, called K-PDE, elutes as a single peak of activity on DEAE-cellulose columns (Brothers, V. M., Walker, N., and Bourne, H. R. (1982) J. Biol. Chem. 257, 9349-9355). Direct photoaffinity labeling of K30a extracts with [32P]cGMP results in radiolabeling of a unique polypeptide, not observed in wild type extracts, which migrates in sodium dodecyl sulfate polyacrylamide gels with an Mr = 106,000. The 106-kDa band was identified as the catalytic K-PDE polypeptide based on the following observations: competitive inhibitors and substrates of K-PDE inhibit photolabeling of the 106-kDa band, indicating that [32P] cGMP photolabels the enzyme at its catalytic site; on DEAE-cellulose chromatography the polypeptide that is susceptible to photolabeling co-elutes with K-PDE activity; the 106-kDa band is detectable in extracts of WT X K30a hybrids (where WT denotes wild type) in amounts proportional to the K-PDE activity in the hybrids, but is undetectable in wild type. The hybrid phenotype strongly suggests that the K30a phenotype is not due to mutations that affect either a diffusible regulator of transcription or an enzyme that modifies K-PDE. Although wild type cells contain a minor cGMP phosphodiesterase activity distinct from the major cAMP phosphodiesterase, the wild type cGMP phosphodiesterase is not susceptible to radiolabeling with [32P]cGMP; this rules out the possibility that the K30a phenotype is caused by overexpression of a wild type phosphodiesterase. We conclude that the K30a mutation produced expression of a new species of phosphodiesterase molecule that is not detectably expressed in the parental S49 wild type cell line.

Cholera toxin can catalyze ADP-ribosylation of cytoskeletal proteins

Cholera toxin catalyzes transfer of radiolabel from [32P]NAD+ to several peptides in particulate preparations of human foreskin fibroblasts. Resolution of these peptides by two-dimensional gel electrophoresis allowed identification of two peptides of Mr = 42,000 and 52,000 as peptide subunits of a regulatory component of adenylate cyclase. The radiolabeling of another group of peptides (Mr = 50,000 to 65,000) suggested that cholera toxin could catalyze ADP-ribosylation of cytoskeletal proteins. This suggestion was confirmed by showing that incubation with cholera toxin and [32P]NAD+ caused radiolabeling of purified microtubule and intermediate filament proteins.

Fibroblast defect in pseudohypoparathyroidism, type I: reduced activity of receptor-cyclase coupling protein

Erythrocytes of many patients with pseudohypoparathyroidism, type I (PHP-I), exhibit reduced activity of the N protein, a guanine nucleotide-binding regulatory component of hormone-sensitive adenylate cyclase. We compared N and adenylate cyclase activities and the accumulation of cAMP in fibroblasts propagated from skin biopsies of six normal subjects and seven PHP-I patients. N activities were reduced by approximately 40% in fibroblasts as well as erythrocytes of five PHP-I patients. N activities in fibroblasts from two PHP-I patients with normal erythrocyte N activities were within the normal range. These results are consistent with the hypothesis that N deficiency is generalized in tissues of most PHP-I patients and is the primary defect responsible for their resistance to metabolic effects of hormones that work by stimulating adenylate cyclase. Fibroblast N deficiency was not associated with decreases in hormone-stimulated adenylate cyclase or cAMP accumulation in fibroblasts, probably because these activities involve many potentially regulable cellular components in addition to the N protein.

Defect of receptor-cyclase coupling protein in pseudohypoparathyroidism

Hormone-sensitive adenylate cyclase contains a recently discovered protein component that is required for stimulation of cyclic AMP synthesis by hormones and guanine nucleotides; the component presumably couples the membrane receptor to the cyclase. We studied this protein (termed "N") in erythrocyte membranes of patients with pseudohypoparathyroidism, using assays of the protein's biochemical activity and of its susceptibility to radiolabeling in the presence of [32P]NAD and cholera toxin. By both assays, the protein's activity was reduced by 40 to 50 per cent in erythrocytes of five of 10 patients with Type I pseudohypoparathyroidism as compared with those of normal and hypoparathyroid subjects and one patient with Type II pseudohypoparathyroidism. If activity of the N protein is reduced in other tissues, this deficiency could cause the resistance of target organs in pseudohypoparathyroidism to parathyroid hormone and other hormones that work via cyclic AMP. Erythrocytes of five patients with Type I pseudohypoparathyroidism, all in one family, showed no defect in activity of the N protein; the biochemical defect of this family remains undefined.

Adaptations of glycogen metabolism in rat epididymal adipose tissue during fasting and refeeding

It is well documented that adipose tissue glycogen content decreases during fasting and increases above control during refeeding. We now present evidence that these fluctuations result from adaptations intrinsic to adipose tissue glycogen metabolism that persist in vitro: in response to insulin (1 milliunit/ml), [3H]glucose incorporation into rat fat pad glycogen was reduced to 10% of control after a 3-day fast; incorporation increased 6-fold over fed control on the 4th day of refeeding following a 3-day fast. We have characterized this adaptation with regard to alterations in glycogen synthase and phosphorylase activity. In addition, we found that incubation of fat pads from fasted rats with insulin (1 milliunit/ml) increased glucose-6-P content, indicating that glucose transport was not the rate-limiting step for glucose incorporation into glycogen in the presence of insulin. In contrast, feeding a fat-free diet resulted in dramatic increases in glycogen content of fat pads without a concomitant increase in glucose incorporation into glycogen in response to insulin (1 milliunit/ml). Thus, fasting and refeeding appeared to alter insulin action on adipose tissue glycogen metabolism more than this dietary manipulation.

Interconversion between multiple glucose 6-phosphate-dependent forms of glycogen synthase in intact adipose tissue

We have tested the hypothesis that interconversion between multiple glucose-6-P-dependent forms of glycogen synthase helps regulate glycogen synthesis in adipose tissue. Our results indicate that interconversion of glycogen synthase in adipose tissue involves primarily dependent forms and that these interconversions were measured better by monitoring the activation constant (A0.5) for glucose-6-P than measuring the -: + glucose-6-P activity ratio. Insulin decreased and epinephrine increased the A0.5 for glucose-6-P without significant change in the activity ratio. Insulin consistently decreased the A0.5 in either the presence or absence of glucose, indicating that the insulin-promoted interconversion did not require increased hexose transport. Isoproterenol increased the A0.5 for glucose-6-P, while methoxamine was without effect, indicating beta receptors mediate adrenergic control of interconversion between glucose-6-P-dependent forms. The changes in the A0.5 produced by incubations with insulin or epinephrine were mutually reversible. We conclude that 1) glycogen synthesis in adipose tissue is catalyzed by multiple glucose-6-P-dependent forms of glycogen synthase, 2) hormones regulate glycogen metabolism by promoting reversible interconversions between these forms, and 3) there is no evidence that a glucose-6-P-independent form of glycogen synthase exists in intact adipose tissue.

Genetic evidence that cholera toxin substrates are regulatory components of adenylate cyclase

Cholera toxin, using [32P]NAD+ as substrate, specifically radiolabels at least two proteins in plasma membranes of wild type S49 mouse lymphoma cells. The toxin-specific substrates are detectable by sodium dodecyl sulfate-polyacrylamide gel electrophoresis as bands corresponding to molecular weights of 45,000 and a doublet of 52,000 to 53,000. Membranes of two other cell types exhibit similar patterns of radiolabeled bands specifically produced by incubation with cholera toxin: the "uncoupled" variant S49 cell, which possesses adenylate cyclase activity unresponsive to hormones, and the HTC4 rat hepatoma cell, which lacks detectable catalytic adenylate cyclase activity but contains components of the cyclase system necessary for regulation by guanyl nucleotides and NaF. Little or no toxin-specific radiolabeling is observed in membranes of a fourth cell type, the adenylate cyclase activity-deficient S49 variant, which functionally lacks components of the cyclase system involved in cholera toxin action and regulation by guanyl nucleotides and NaF. The toxin-specific labeling pattern is not observed in membranes prepared from wild type S49 cells previously treated with cholera toxin in culture. One or both of the toxin substrates thus appears to be involved in regulation of adenylate cyclase by guanyl nucleotides and fluoride ion.

Reconstitution of cholera toxin-activated adenylate cyclase

Reconstitution of adenylate cyclase activity responsive to stimulation by guanylyl-5'imidodiphosphate or NaF may be achieved by mixing dilute Lubrol 12A9-solubilized extracts of wild-type S49 membranes with membranes of an adenylate cyclase-deficient variant. Experiments using N-ethylmaleimide to inactivate components of the adenylate cyclase system indicate that distinct components from both wild-type detergent extracts and adenylate cyclase-deficient membranes are essential for reconstitution. These results and conclusions confirm those of E. M. Ross and A. G. Gilman [J. Biol. Chem. (1977) 252, 6966-6969]. Detergent extracts of cholera toxin-treated wild-type membranes yield a reconstituted adenylate cyclase as responsive to GTP as to guanylyl-5'-imidodiphosphate whereas, in the absence of cholera toxin treatment, GTP has little or no effect. Cholera toxin-treated adenylate cyclase-deficient membranes and Lubrol 12A9 extracts from them, however, fail to yield a reconstituted adenylate cyclase that responds to GTP with an increase in cyclase activity. Because treatment of the adenylate cyclase-deficient variants with cholera toxin is without effect on the reconstituted cyclase, we propose that the cholera toxin substrate is absent or altered in the adenylate cyclase-deficient phenotype.