Changes in pulmonary adenosine triphosphate-binding proteins detected by nucleotide photoaffinity labeling following treatment of mice with the tumor-modulatory agent butylated hydroxytoluene.

Protein kinases play a central role in the regulation of normal and neoplastic growth. A novel approach to evaluating this role is to apply the photoaffinity analogue of adenosine 5'-triphosphate (ATP), 8-azido[gamma-32P]adenosine 5'-triphosphate (8-N3-[gamma-32P]ATP), to extracts obtained from nondividing and dividing tissues in order to label high affinity ATP receptors, including protein kinases. A pattern of ATP-binding proteins was observed on autoradiograms prepared from lung extracts which were reproducible upon repeated experimentation and was identical among extracts prepared from several inbred strains. Seven specifically photolabeled proteins were detected with molecular weights of 120,000, 84,000, 53,000, 51,000, 43,000, 41,000, and 33,000. The pneumotoxicity that results from treating mice with butylated hydroxytoluene (BHT) is followed by a stimulation of lung cell proliferation as part of the regenerative repair phase. Increased photoincorporation of 8-N3-[gamma-32P]ATP into proteins with molecular weights of 41,000, 37,000 and 33,000 resulted from BHT treatment. The following correlations between increased photoincorporation and lung repair, as defined by an increased lung weight/body weight ratio, were observed: (a) these changes shared a similar BHT dose dependency; (b) both followed the same time course and were reversible; (c) both lung damage and enhanced photolabeling were abolished if the mice were pretreated with cedrene, an inducer of drug-detoxifying enzymes; (d) the effect of BHT on ATP-binding proteins was agent specific, since urethan treatment did not affect photoincorporation. These results demonstrate that the specific photolabeling of ATP receptors in tissue extracts is a sensitive indicator of a particular growth-related repair process.

Changes in cyclic adenosine 3':5'-monophosphate-dependent protein kinases during the progression of urethan-induced mouse lung tumors.

The cyclic adenosine 3':5'-monophosphate (cAMP)-dependent protein kinases in lung adenomas are functionally different from those of normal lung. The relevance of this change to neoplastic conversion was examined by comparing tumor kinases with those obtained from the normal cell of origin and by studying the kinases at different stages of tumor growth. Lung tumors were collected from A strain mice at different times after a single injection of urethan. These tumors are predominantly of alveolar type two cell origin, and cAMP-binding proteins in extracts from isolated type two cells and from lung adenomas at various stages of tumor progression were compared. Both the incorporation of the cAMP photoaffinity analogue, cyclic 8-azidoadenosine 3':5'-[32P]monophosphate (8-N3-[32P]cAMP), into the regulatory subunits of the type I (RI) and type II (RII) cAMP-dependent protein kinases and the autophosphorylation of RII were similar in extracts from whole normal lung and from type two cells. Altered protein kinases are thus not characteristic of normal type two cells. Lung tumors showed a decrease in photodetectable RII which correlated in degree with tumor size and extent of anaplasticity. This decreased RII photolabeling during tumor growth was associated with increased RII autophosphorylation. In contrast, decreased RII photolabeling in extracts from neonatal lung is accompanied by a substantial decrease in RII autophosphorylation. The characteristics of RII during normal development thus clearly differ from those during neoplastic development. An increase in the amount of an Mr 37,000 proteolytic fragment derived from R-subunits was also noted as a function of tumor progression. DEAE-cellulose chromatography of tumor cytosol showed that the increase in the amount of Mr 37,000 protein was accompanied by increased subunit dissociation of the type I isozyme. The dissociated RI subunit has been shown to be more sensitive to cleavage by a Ca2+-dependent neutral protease than when RI was in the holoenzyme form. This protease is present in both normal lung and lung adenomas, and its activity increases during the later stages of tumor progression. A comparison of cAMP binding and the light-induced covalent incorporation of 8-N3-[32P]cAMP showed that, for both RI and RII, photoincorporation was about 75% as efficient as noncovalent binding. In contrast, although the Mr 37,000 fragment can be photolabeled with low concentrations of 8-N3-[32P]cAMP, noncovalent cAMP binding to the endogenous Mr 37,000 fragment could not be demonstrated with a standard filtration assay. Such altered cAMP binding characteristics following Ca2+-dependent proteolysis of R-subunits would all

Autoradiographic localization of specific [3H]dexamethasone binding in fetal lung.

The cellular and subcellular localization of specific [3H]dexamethasone binding was examined in fetal mouse lung at various stages of development and in human fetal lung at 8 weeks of gestation using a rapid in vitro steroid incubation technique followed by thaw-mount autoradiography. Competition studies with unlabeled steroids demonstrate the specificity of [3H]dexamethasone labeling, and indicate that fetal lung mesenchyme is a primary glucocorticoid target during lung development. Quantitative binding studies, involving incubation of intact tissue with competing ligand and subsequent subcellular fractionation, show this to be specific, nuclear binding characteristic of glucocorticoid receptors. Autoradiographs of [3H]dexamethasone binding in lung tissue at early stages of development demonstrate that the mesenchyme directly adjacent to the more proximal portions of the bronchiolar network is heavily labeled. In contrast, the epithelium which will later differentiate into bronchi and bronchioles, is relatively unlabeled. Distal portions of the growing epithelium, destined to become alveolar ducts and alveoli, do show nuclear localization of [3H]dexamethasone. Because of the known importance of the mesenchyme in controlling lung development and the ability of glucocorticoids to stimulate lung development, these results suggest that many of the growth-promoting effects of glucocorticoids may be mediated through the mesenchyme. In addition, by utilizing a technique which allows the simultaneous examination of extracellular matrix components and [3H]dexamethasone binding, a relationship is observed between extensive mesenchymal [3H]dexamethasone binding and extensive extracellular matrix accumulation. Since glucocorticoids stimulate the synthesis of many extracellular matrix components, these results suggest a role for these hormones in affecting mesenchymal-epithelial interactions during lung morphogenesis.

Strain differences among inbred mice in protein kinase C activity.

A protein kinase was identified in mouse organs whose activity is strictly dependent on the presence of Ca++ and phosphatidylserine when assayed at pH 6, and thus has the characteristics of protein kinase C. The relative order of specific activities was brain greater than spleen greater than lung greater than heart, an order similar to that found previously for rat organs. Mice from seven strains had the same level of protein kinase C activity, but strain A/J had half as much activity in each organ as did the other strains. F1 hybrid mice resulting from a cross between A/J and BALB/cByJ mice had levels of activity intermediate to the parental strains, indicating additive inheritance of this genetic difference.

Functional changes in the regulatory subunit of the type II cyclic adenosine 3':5'-monophosphate-dependent protein kinase isozyme during normal and neoplastic lung development.

The abilities of cyclic adenosine 3':5'-monophosphate (cAMP) and cyclic 8-azidoadenosine 3':5'-[32P]monophosphate (8-N3-[32P]cAMP) to bind to the regulatory subunit (RII) of the type II cAMP-dependent protein kinase isozyme and to cause subsequent dissociation of the holoenzyme were compared in extracts from adult and neonatal mouse lung and lung adenoma. RII in extracts from adult lung exhibits equal numbers of high- (Kd 15 nM) and low- (Kd 230 nM) affinity 8-N3-[32P]cAMP binding sites. In the neonate, the proportion of high-affinity sites is reduced to 20% while, in lung adenoma, only low-affinity RII binding is observed. Low-affinity RII binding is correlated with an inability of cAMP to dissociate the type II holoenzyme completely. Sucrose gradient sedimentation of adult lung cytosol in the presence of cAMP shows complete dissociation of the type I isozyme, while only some of the type II holoenzyme is dissociated. This is in contrast to the case with lung tumor cytosol, in which only low-affinity binding is observed and no apparent dissociation of the type II isozyme occurs. cAMP does promote RII dephosphorylation within the holoenzyme, however, suggesting that cAMP can bind to RII without dissociating the tetramer. Consistent with this interpretation, photoincorporation of 8-N3-[32P]cAMP prior to sucrose gradient sedimentation results in the formation of a photolabeled RII complex which sediments at the same rate as does the holoenzyme. Two-dimensional gel electrophoresis of RII photolabeled at low and high concentrations of 8-N3-[32P]cAMP suggests that these altered binding and dissociation characteristics of the type II isozyme are not due to the presence of a structurally altered RII molecule. After DEAE-cellulose chromatography of lung cytosol, only high-affinity RII binding is observed, and all of the RII can now be dissociated with cAMP. Low-affinity binding may thus reflect either an altered conformational state of RII or the interaction of the type II kinase with other cytosolic molecules which can affect RII binding and dissociation without altering the functional properties of the type I isozyme.

Co-elution of type I and type II cAMP-dependent protein kinase holoenzymes in high salt fractions from DEAE-cellulose.

Post-mitochondrial supernatants from adult and neonatal mouse lung were each separated using DEAE-cellulose chromatography, and fractions were assayed for cAMP-dependent kinase activity and photoincorporation of the analog, 8-azido-cyclic adenosine 3':5'-monophosphate. Although there is no significant developmental differences in the amount of kinase activity which elutes in these high salt or DEAE peak II fractions, such fractions derived from neonatal lung contain only half as much photodetectable regulatory subunit of the type II protein kinase isozyme (RII) as do corresponding fractions from the adult. Much of the kinase activity associated with these neonatal DEAE peak II fractions may actually be due to the type I holoenzyme. Significant amounts of photodetectable regulatory subunit of the type I protein kinase isozyme (RI) co-elute with this peak of type II kinase activity, and several lines of evidence suggest that much of this RI exists in a holoenzyme state. Incubation of these peak II fractions with cAMP prior to a second elution from DEAE-cellulose causes much of the RI to elute at lower salt concentrations characteristic of the free RI subunit. After sucrose gradient sedimentation of peak II fractions, RI cosediments with the peak of cAMP-dependent kinase activity. Finally, kinase activation by cAMP is biphasic, suggesting that both the type I and II holoenzymes contribute to the kinase activity in the type II peak. Two-dimensional gel electrophoresis of photolabeled proteins does not suggest any structural difference between RI subunits present in various DEAE fractions. The elution of the type I holoenzyme over a range of salt concentrations may be due to structural variations in the C subunit of the type I isozyme or possibly to the stable association of the type I holoenzyme with other molecules in the cell.

Developmental changes in the endogenous Ca2+-stimulated proteolysis of mouse lung cAMP-dependent protein kinases.

Regulatory subunits (R subunits) of mouse lung cAMP-dependent protein kinases undergo age-dependent changes in endogenous proteolysis, with the greatest amount of the major Mr = 37,000 proteolytic fragment detectable during fetal and neonatal development. Homogenization of lung in the presence of various protease inhibitors does not affect this age-related difference, suggesting that the observed quantitative change in R subunit proteolysis occurs in vivo. Mechanisms were sought to account for this age-dependent change. The production of a Mr = 37,000 proteolytic fragment can be stimulated in lung extracts by the addition of exogenous calcium and is due to the action of an endogenous Ca2+-stimulated protease. Neonatal lung extracts show more Ca2+-stimulated proteolysis of R subunits than adult extracts, although only slight age-related differences in either the Ca2+-stimulated protease or its specific endogenous inhibitor were observed. Age-dependent differences in R subunits which may affect sensitivity to proteases were also examined. Analysis of the two-dimensional patterns of adult and neonatal 8-N3-[32P]cAMP-labeled R subunits before or after limited proteolysis with trypsin suggests that the R subunits are structurally similar. Differences are found, however, in the relative proportions of adult and neonatal Type I R subunits (RI) in the holoenzyme or dissociated forms. An increased proportion of neonatal R subunits exist in the dissociated state, whereas adult R subunits exist primarily in the holoenzyme form. Dissociated R subunits from mouse lung are more susceptible than the holoenzyme to limited proteolysis by the partially purified lung Ca2+-stimulated protease. Dissociation of the holoenzyme in vivo may be a major factor in the age-dependent proteolytic changes observed in mouse lung protein kinases.

Species specificity of brain cyclic AMP receptor proteins.

The cyclic AMP binding proteins present in mouse, rat, bovine, and sheep brains were compared. Extracts were isotopically labeled with 8-azido-cyclic [32P]AMP, a photoaffinity analog specific for cyclic AMP binding sites, and then subjected to two-dimensional gel electrophoresis. The resulting autoradiographic patterns were generally similar, but showed consistent species variations. Proteins identified by their size and phosphorylatability as regulatory subunits of Type II protein kinase isozymes were present in all species, but with slight variations in pI. A series of charge variants identified as regulatory subunits of Type I kinase isozymes on the basis of their size was also ubiquitously present, as were several smaller proteins postulated to be proteolytic fragments derived from the regulatory subunits. The major species difference was a series of labeled proteins found only in rodent brains and not in the brains of any ruminant, or in other rodent tissues. These proteins had a molecular weight of 54,000 and a pI range of 5.89--6.26, and could not be endogenously phosphorylated. The identities of these proteins and their relationship to the protein kinase regulatory subunits are unknown.

H-2 and non-H-2 determined strain variation in palatal shelf and tongue adenosine 3':5' cyclic monophosphate: a possible role in the etiology of steroid-induced cleft palate.

The concentration of cAMP was measured in palatal shelves and tongues of 14.5-day old foetuses, 14.5-day old foetuses from steroid treated mothers, and 15.5-day old foetuses from four inbred lines of mice which represent the four possible combinations of two H-2 alleles and two residual genetic backgrounds. The incidence of spontaneous and steroid-induced cleft palate in these four strains was also determined. Analyses of variance of the cAMP data reveal that both the H-2 region and residual genetic background determine cAMP concentrations in both tissues and on both days of development. Similar analyses of cAMP concentrations after steroid treatments of the mother indicate that the interaction between H-2 and residual genetic background is significantly different in the injected than in the uninjected mice in both palatal shelves and tongues. The incidence of steroid-induced cleft palate parallels the palatal shelf concentration of cAMP before steroid treatment of the mother with one exception. These data suggest that a portion of the H-2 controlled component of susceptibility to steroid-induced cleft palate is mediated through alterations in the metabolism of cAMP.