Brain-derived acidic fibroblast growth factor: complete amino acid sequence and homologies.

Bovine brain-derived acidic fibroblast growth factor (aFGF) is a protein mitogen originally identified in partially purified preparations of whole brain. The protein was purified to homogeneity and shown to be a potent vascular endothelial cell mitogen in culture and angiogenic substance in vivo. The homology of aFGF to human interleukin-1 beta was inferred from partial sequence data. The complete amino acid sequence of aFGF has now been determined and observed to be similar to both basic FGF and interleukin-1's. A neuropeptide-like sequence, flanked by basic dipeptides, was observed within the aFGF sequence.

Ca2+, calmodulin and cyclic AMP-dependent modulation of actin-myosin interactions in aorta.

Incubation of bovine aortic native actomyosin with cyclic AMP and bovine aortic cyclic AMP-dependent protein kinase produced a rightward shift in the relation between free Ca2+ and both superprecipitation and actomyosin ATPase activity. The relation between free Ca2+ and phosphorylation of myosin light chains was also shifted to the right. The concentration of free Ca2+ required for half-maximal activation of both ATPase activity and myosin light chain phosphorylation was approximately 1.0 microM for control actomyosin and 2.5 microM for actomyosin incubated with cyclic AMP-protein kinase. Neither basal nor maximal activities were significantly affected by incubation with cyclic AMP-protein kinase. Addition of e microM calmodulin to cyclic AMP-protein kinase-treated actomyosin relieved inhibition of both superprecipitation and myosin light chain phosphorylation. These findings suggest that cyclic AMP-protein kinase-mediated inhibition of actin-myosin interactions in vascular smooth muscle involve a shift in the Ca2+ sensitivity of the system. This shift probably involves Ca2+-calmodulin interactions and the control of phosphorylation of the myosin light chains.

Acidic fibroblast growth factor accelerates dermal wound healing in diabetic mice.

Acidic fibroblast growth factor (aFGF) is a potent mitogenic and chemotactic agent for vascular endothelial cells, dermal fibroblasts, and epidermal keratinocytes, the principal cellular constituents of skin. To explore its potential to heal chronic dermal wounds, we applied pure recombinant human aFGF topically to full-thickness excisional injuries in healing-impaired genetically diabetic mice. Transformation of the nonlinear percent initial wound areas as a function of time to linear rates of tissue ingrowth from the original wound edges showed that aFGF increased wound closure in a dose-dependent manner. Optimal 3-micrograms/cm2 doses of aFGF nearly tripled the linear rate of healing. The median time to complete closure decreased from 46 d in vehicle-treated wounds to only 16 d in those treated with aFGF. Histomorphometric analyses established that aFGF increased granulation tissue formation and reepithelialization throughout healing. Vehicle- and aFGF-treated wounds appeared to be histologically equivalent by the time of closure. Therefore, aFGF has potential therapeutic applications for promoting healing of dermal ulcers, especially in healing-impaired individuals.

Protein phosphorylation: quantitative analysis in vivo and in intact cell systems.

Protein phosphorylation-dephosphorylation appears to be an essential component in the regulation of many cellular processes by hormones and drugs. This concept has developed primarily from in vitro biochemical studies in which various purified proteins have been phosphorylated and dephosphorylated by distinct protein kinases and phosphoprotein phosphatases. However, the more difficult, but essential, task of demonstrating the physiological occurrence of these reactions in intact tissue or cell preparations in many cases has not been undertaken in a quantitative manner. There are 4 basic approaches for assessing the extent of protein phosphorylation in vivo and in intact cell systems, each having particular advantages and disadvantages. These are summarized in Table 2. The applicability of any one procedure will be highly dependent upon the protein under investigation. For instance, chemical measurements of total protein-bound phosphate may provide only limited information for proteins which are phosphorylated at multiple sites but could be highly useful for those proteins such as glycogen phosphorylase which are phosphorylated at single sites. The relative ease and the high sensitivity of measuring 32P incorporation into proteins will tempt many investigators to rely heavily on this approach. It is a very powerful procedure, particularly for the initial identification of phosphoproteins, but ultimately quantitative conclusions regarding 32P incorporation must be corroborated by one or more of the other procedures. There is no simple, single experimental approach that may be used under all circumstances, but by integrating these procedures firm conclusions may be drawn regarding the physiological importance of phorphorylation of specific proteins.

Characterization of specific binding of [125I]L-762,459, a selective alpha1A-adrenoceptor radioligand to rat and human tissues.

L-762,459 ((+/-)1-(3-¿[5-carbamoyl-2-2-[(4-hydroxy-3-iodobenzimidoyl)-amino] -ethoxy-methy¿-6-methyl-4-(4-nitropheny)-1,4-dihydropyridine -3-carbonyl]-amino¿-propyl)-4-phenyl-1-piperidine-4-carboxylic acid methyl ester), an analog of a series of dihydropyridines previously reported to be selective alpha1A-adrenoceptor subtype antagonists was found to have alpha1A-adrenoceptor subtype selectivity (Ki (nM), la = 1.3, lb = 240, Id = 280). Specific [125I]L-762,459 binding was detected in rat cerebral cortex, hippocampus, vas deferens, kidney, heart and prostate tissues known to contain the alpha1A-adrenoceptor subtype, but not in tissues known to contain alpha1B-adrenoceptor (spleen, liver) and alpha1D-adrenoceptor (aorta). Scatchard analysis of [125I]L-762,459 binding in rat cerebral cortex and prostate indicated a single binding site with a Kd of 0.7 nM and Bmax of 11 (cerebral cortex) and 1 (prostate) pmole/g tissue. Specific and saturable [125I]L-762,459 binding was also found in human cerebral cortex, liver, prostate and vas deferens (Kd = 0.2-0.4 nM, Bmax = 0.4-4 pmole/g tissue). The specific binding in rat and human tissues was competed by non-selective alpha1-adrenoceptor compounds (Ki values in nM: prazosin (0.14-1.2), terazosin (1.8-5.9) and phentolamine (2.4-11)) and selective alpha1A-adrenoceptor compounds [Ki values in nM: (+) niguldipine (0.04-1.2) and SNAP 5399 ((+/-)-2-((2-aminoethyl)oxy)methyl-5-carboxamido-6-ethyl-4-(4-nitropheny l)-3-N-(3-(4,4-diphenylpiperidin-1-yl)propyl)carboxamido-1,4-dihyd ropyridine hydrate (0.5-4.8)]. The results were consistent with the selective binding of [125I]L-762,459 to the alpha1A-adrenoceptor. The specific labeling of the alpha1A-adrenoceptor subtype by [125I]L-762,459 may make it a useful tool to localize the distribution of the alpha1A-adrenoceptor.

In vitro studies on L-771,688 (SNAP 6383), a new potent and selective alpha1A-adrenoceptor antagonist.

L-771,688 (SNAP 6383, methyl(4S)-4-(3, 4-difluorophenyl)-6-[(methyloxy)methyl]-2-oxo-3-[(¿3-[4-(2-pyridin yl)-1-piperidinyl]propyl¿amino)carbonyl]-1,2,3, 4-tetrahydro-5-pyrimidine carboxylate) had high affinity (Ki less than or = 1 nM) for [3H]prazosin binding to cloned human, rat and dog alpha1A-adrenoceptors and high selectivity (>500-fold) over alpha1B and alpha1D-adrenoceptors. [3H]Prazosin / (+/-)-beta-[125I]-4-hydroxy-phenyl)-ethyl-aminomethylteralone ([125I]HEAT) binding studies in human and animal tissues known to contain alpha1A and non-alpha1A-adrenoceptors further demonstrated the potency and alpha1A-subtype selectivity of L-771,688. [3H]L-771,688 binding studies at the cloned human alpha1A-adrenoceptors and in rat tissues indicated that specific [3H]L-771,688 binding was saturable and of high affinity (Kd=43-90 pM) and represented binding to the pharmacologically relevant alpha1A-adrenoceptors. L-771,688 antagonized norepinephrine-induced inositol-phosphate responses in cloned human alpha1A-adrenoceptors, as well as phenylephrine or A-61603 (N-[5-4,5-dihydro-1H-imidazol-2yl)-2-hydroxy-5,6,7, 8-terahydro-naphthlen-1-yl] methanesulfonamide hydrobromide) induced contraction in isolated rat, dog and human prostate, human and monkey bladder neck and rat caudal artery with apparent Kb values of 0.02-0.28 nM. In contrast, the contraction of rat aorta induced by norepinephrine was resistant to L-771,688. These data indicate that L-771,688 is a highly selective alpha1A-adrenoceptor antagonist.

Contractile responses to epinephrine, serotonin, and potassium in arterial smooth muscle.

Effects of epinephrine, serotonin, and KCl on development of isometric tension were studied in strips of rat femoral and tail artery smooth muscle. Strkining differences were detected between the preparations. Tension development in tail artery smooth muscle was greatest with epinephrine, intermediate with serotonin, and least with K+. In femoral preparations tension developed was greatest with serotonin. Smaller but comparable contractions were elicited with epinephrine and K+. Theoretical and experimental dose-response curves for epinephrine and serotonin agreed closely, whereas curves for K+ differed markedly. The relationship between fractional response and fractional stimulus was hyperbolic for epinephrine and serotonin but sigmoid for K+. The Hill coefficient for serotonin was about 1, slightly smaller than 1 for epinephrine (0.5 to 0.8) and significantly greater for K+ (5 to 5). These findings show that responses elicited with either epinephrine or serotonin are mechanistically consistent with receptor occupancy theory. They suggest that specific receptors for K+ exist and that they may involve positive cooperative interactions similar those described for receptor mechanisms in electroplax.

Adenosine 3':5'-monophosphate-mediated inhibition of myosin light chain phosphorylation in bovine aortic actomyosin.

Ca2+-dependent phosphorylation of the myosin light chains in bovine aortic native actomyosin is markedly depressed in the presence of cyclic AMP and its dependent protein kinase. This inhibition occurs with either cardiac, skeletal, or aortic protein kinase plus cyclic AMP, while little or no inhibition occurs with either cyclic AMP or protein kinase alone. The extent of inhibition is related to the concentration of protein kinase and approaches a maximum of approximately 50%. Concomitant with the inhibition of myosin light chain phosphorylation is (a) an increased phosphorylation of a 100,000-dalton moiety which possibly corresponds to the myosin light chain kinase present in the native actomyosin preparation and (b) a decrease in the actomyosin Mg2+-ATPase activity. These findings suggest that modulation of actin-myosin interactions by the cAMP system directly at the level of the contractile proteins may represent a mechanism by which beta adrenergic relaxation occurs in mammalian vascular smooth muscle.

Concentration and time-dependent relationships between isosorbide dinitrate-induced relaxation and formation of cyclic GMP in coronary arterial smooth muscle.

This study is based on the hypothesis that isosorbide dinitrate (ISDN)-induced relaxation of coronary arterial smooth muscle is causally linked to formation of cyclic (c) GMP. The hypothesis requires the extent of relaxation to be correlated to both time-and concentration-dependent increases in coronary content of cGMP. Accordingly, studies were performed with bovine coronary arterial strips to determine the relationships among isometric force and coronary content of cGMP and cAMP with respect to time of exposure to and concentration of ISDN. Cyclic nucleotide levels were determined by radioimmunoassay. No change in cAMP levels was observed during ISDN-induced relaxation of KCl contracted strips. In sharp contrast, cGMP levels increased significantly with time of exposure and concentration of ISDN stimulation. Moreover, the addition of methylene blue, a reported inhibitor of guanylate cyclase, to the bathing medium significantly inhibited the relaxation and cGMP increase during ISDN stimulation. In addition, prolonged exposure to ISDN resulted in a redevelopment of force with a parallel decrease in cGMP content. The increase in cGMP during ISDN stimulation also occurs in the absence of depolarization by KCl and in an essentially Ca++-free medium. These data support the hypothesis that the relaxation of coronary arterial strips in response to ISDN stimulation is causally linked to cGMP.

Eicosonoid metabolism and beta-adrenergic mechanisms in coronary arterial smooth muscle: potential compartmentation of cAMP.

Beta-Adrenergic relaxation in bovine coronary arteries is enhanced by inhibition of eicosonoid metabolism and inhibited by its stimulation. We investigated the interaction between eicosonoid metabolism and beta-adrenergic mechanisms by studying the effect of perturbations of eicosonoid metabolism on vascular adenosine 3',5'-monophosphate (cAMP) content and the cAMP-dependent relaxation of isometric force and activation of glycogen phosphorylase. KCl (35 mM) elicited a contraction, activated phosphorylase, and slightly decreased cAMP content. Isoproterenol (10(-7) M) relaxed the KCl contraction, further increased phosphorylase activity, and increased cAMP. Neither indomethacin (5 X 10(-6) M) nor arachidonic acid (3 X 10(-5) M) affected the KCl contraction, but arachidonic acid increased both cAMP and phosphorylase activity and indomethacin decreased cAMP. Indomethacin potentiated the relaxation induced by isoproterenol but inhibited the activation of phosphorylase and had no effect on the isoproterenol-induced increase in cAMP. Arachidonic acid, on the other hand, inhibited the isoproterenol-induced relaxation but potentiated both the increases of phosphorylase activity and cAMP. Thus neither relaxation nor phosphorylase activity was related in a straightforward manner to the total cAMP content. A direct relation between cAMP, relaxation, and phosphorylase can be reconciled with the antiparallel effects of alterations of eicosonoid metabolism observed in this study by a proposed model in which the effects of cAMP are assumed to be functionally compartmentalized.

Laminin-mediated process formation in neuronal cells involves protein dephosphorylation.

Laminin mediates neural adhesion and process formation. A possible signal transduction pathway for laminin was investigated in both NG108-15 and PC12 neuronal cells using radiolabeling studies as well as various stimulators and inhibitors of phosphatases and kinases. Using [32P]-ortho-phosphate, laminin caused a decrease in the TCA-precipitable counts. Further, laminin stimulated dephosphorylation of laminin binding proteins of 110 kDa, 67 kDa, and 45 kDa and this dephosphorylation was blocked by the phosphatase inhibitor, okadaic acid, and the protein kinase C stimulator, TPA. The phosphatase inhibitors okadaic acid and vanadate, as well as the protein kinase C stimulators, TPA and DAG, blocked laminin-mediated process formation. Inhibitors of kinase activity such as H-7, H-8, and H-9 increased laminin-mediated neural process formation. Since phosphate incorporation into laminin-binding proteins is decreased by laminin and because both phosphatase inhibitors and kinase stimulators inhibit laminin-mediated process formation, we conclude that dephosphorylation events promote the neural cell response to laminin.

cGMP and cAMP inhibit tension development in skinned coronary arteries.

The effects of physiological concentrations of cGMP and cAMP on tension development in skinned coronary arteries (Triton X-100) were studied. cGMP inhibited tension elicited at intermediate Ca2+ concentrations at pH 7.0 but not at more acidic or alkaline pH values. cAMP, on the other hand, decreased submaximal tension development independent of pH (from pH 6.5 to pH 7.2). Neither nucleotide affected tension development at maximally activating Ca2+ concentrations.

Adenosine-mediated relaxation and activation of cyclic AMP-dependent protein kinase in coronary arterial smooth muscle.

Conflicting evidence exists regarding the participation of cyclic AMP (cAMP) in adenosine-induced relaxation of the coronary vasculature. Because the mechanism of action of cAMP is thought to involve activation of its dependent protein kinase, the purpose of this study was to determine if cytosolic cAMP protein kinase was activated in response to adenosine stimulation and to determine if such activation was correlated to the extent of relaxation in intact coronary arterial strips. Adenosine produced increases in cAMP protein kinase activity in both main trunk and branch circumflex bovine arterial strips. However, both the relaxant and kinase effects were greater in branch strips. Concentration and time-dependent increases in adenosine-induced relaxation of contracted branch strips were tightly coupled to concomitant increases in cAMP protein kinase activity (r = 0.93). Moreover, this increase in kinase activity was ascribable to the cAMP-dependent kinase, as the specific inhibitor of the cAMP protein kinase attenuated these increases. In contrast, relaxation produced by sodium nitroprusside was associated with an increase in a cAMP-independent kinase. In additional experiments, cumulative dose-response curves (10(-7) to 10(-3) M) for relaxation by adenosine and nine of its analogs showed that all agents were more effective in branch strips. Adenine-9-beta-D-arabinofuranoside, the least potent analog, did not produce relaxation or increase kinase activity. In contrast, 2-chloroadenosine, the most effective relaxant analog, also increased cAMP protein kinase activity. These findings suggest that adenosine-induced relaxation may involve cAMP and activation of cAMP protein kinase in coronary arterial smooth muscle.

Enhanced carotid sinus baroreflex responses in dogs with chronic A-V block.

Effects of carotid sinus pressure on arterial pressure, atrial rate, and ventricular rate were examined in anesthetized normal dogs and in dogs with chronic complete A-V block. Change in arterial pressure per mmHg change in sinus pressure was 0.8 plus or minus 0.2 mmHg for controls but increased (P is less than 0.001) to 1.6 plus or minus 0.1 mmHg in A-V blocked dogs. Arterial pressure was 140-145 mmHg at low sinus pressure in both groups, but at high sinus pressure arterial pressure was significantly lower in A-V blocked dogs (44 plus or minus 4 mmHg) than in controls (92 plus or minus 8 mmHg). These differences were virtually abolished after vagotomy. Heart rate increased in normal dogs as sinus pressure was increased before vagotomy, but decreased after vagotomy. In blocked dogs atrial and ventricular rates decreased at high sinus pressure either before or after vagotomy. The results show that reflex circulatory responses to changes in carotid sinus pressure are enhanced in dogs with A-V block. This enhancement may involve attenuation of buffering influences exerted from other baroreceptors whose afferents are in the vagus nerves.

Contractile responses to angiotensin I and angiotensin II in hamster uterine smooth muscle.

Angiotensin I (A-I) and angiotensin II (A-II) produced dose-dependent increases in isometric tension in isolated strips of uterine smooth muscle prepared from ovariectomized golden hamsters treated with estrogen. Responses to A-II were consistent with receptor--occupancy theory of agonist--receptor interactions. Inhibition of angiotensin-converting enzyme virtually abolished responses to A-I but not those to A-II. Blockade of A-II receptors inhibited responses to both A-I and A-II. Cholinergic or alpha-adrenergic blockade did not alter uterine responses to either A-I or A-II. These findings suggest that contractile responses elicited in the isolated uterus of the hamster are due to its local conversion to A-II and subsequent interactions with specific A-II receptors. Such conversion occurs at least to the extent of 14 to 27 %.

X-ray crystal structure of human acidic fibroblast growth factor.

Fibroblast growth factors (FGFs) are mitogenic and chemotactic agents for a wide variety of cell types and play a primary role in the regulation of angiogenesis. Angiogenesis is involved in a variety of critical physiological events including organogenesis, wound healing, ischemic collateral circulation, and solid tumor growth. High-resolution structural information is key to understanding the mechanism of action of these growth factors. We report here the X-ray crystal structure of human acidic FGF (aFGF), with data extending to 2.0 angstroms resolution. The crystal contains four independent molecules in the asymmetric unit. Each molecule contains a single bound sulfate ion, in similar juxtapositions. The bound sulfate is stabilized through hydrogen-bond interactions with residues Asn 18, Lys 113, and Lys 118 and defines a potential heparin binding site. The hydrogen bond with the N delta 2 moiety of Asn 18 appears to be the most conserved interaction, being similar to those observed for sulfate ion bound to human basic FGF (bFGF) and similar but not identical to interactions observed for bovine aFGF with heparin analogs. Of the added solvent groups, five ordered water molecules are conserved in each of the four independent structures of human aFGF. These water molecules, located at buried positions, provide hydrogen bonding partnerships with several buried polar groups in the core of the protein. A central interior cavity exists in each of the four structures, with sizes ranging from approximately 20 to 50 angstroms3. The cavity sizes appear to be significantly smaller than that observed in the related protein interleukin-1 beta. The region comprising the high affinity FGF receptor binding site is structurally very similar to the corresponding region from human bFGF, whereas the low affinity site is structurally quite different. The results provide a structural basis for the role of the low affinity binding site in FGF receptor discrimination.

beta-Adrenergic relaxation and cAMP kinase activation in coronary arterial smooth muscle.

If beta-adrenergic relaxation of smooth muscle is partly mediated by the adenosine 3',5'-cyclic monophosphate (cAMP) system, then beta-stimulation should be correlated to activation of cAMP-dependent protein kinase (cPK). Studies were performed with bovine coronary arterial strips to identify isozymic forms of cPK and to determine if beta-relaxation is correlated to activation of cPK (reflected by elevated ratios of cPK activity without cAMP to cPK activity with cAMP). Both ion exchange chromatography and a new electrophoretic technique revealed two cPK isozymes (types I and II). No change in cPK activity occurred in strips contracted with 30 mM KCl. In contrast, dose- and time-dependent relaxation during beta-stimulation with isoproterenol was highly correlated to parallel increases in cPK activity. Increased cPK activity was inhibited in assays performed with a specific inhibitor of cPK. Both relaxation and activation of cPK were abolished during beta-adrenergic blockade with propranolol. Relaxation by KCl removal or the ionophore R02-2985, unlike beta-mediated relaxation, did not increase cPK activity. These findings show that beta-mediated relaxation of isolated coronary arterial strips specifically activates cPK, and they support the hypothesis that beta-induced relaxation of vascular smooth muscle involves the cAMP system.