In silico design of novel hERG-neutral sildenafil-like PDE5 inhibitors

Cyclic nucleotide phosphodiesterase enzymes (PDEs) have functions in regulating the levels of intracellular second messengers, 3′, 5′-cyclic adenosine monophosphate (cAMP) and 3′, 5′-cyclic guanosine monophosphate (cGMP), via hydrolysis and decomposing mechanisms in cells. They take essential roles in modulating various cellular activities such as memory and smooth muscle functions. PDE type 5 (PDE5) inhibitors enhance the vasodilatory effects of cGMP in the corpus cavernosum and they are used to treat erectile dysfunction. Patch clamp experiments showed that the IC₅₀ values of the human ether-à-go-go-related gene (hERG1) potassium (K) ion channel blocking affinity of PDE5 inhibitors sildenafil, vardenafil, and tadalafil as 33, 12, and 100 μM, respectively. hERG1 channel is responsible for the regulation of the action potential of human ventricular myocyte by contributing the rapid component of delayed rectifier K⁺ current (I_Kr) component of the cardiac action potential. In this work, interaction patterns and binding affinity predictions of selected PDE5 inhibitors against the hERG1 channel are studied. It is attempted to develop PDE5 inhibitor analogs with lower binding affinity to hERG1 ion channel while keeping their pharmacological activity against their principal target PDE5 using in silico methods. Based on detailed analyses of docking poses and predicted interaction energies, novel analogs of PDE5 inhibitors with lower predicted binding affinity to hERG1 channels without loosing their principal target activity were proposed. Moreover, molecular dynamics (MD) simulations and post-processing MD analyses (i.e. Molecular Mechanics/Generalized Born Surface Area calculations) were performed. Detailed analysis of molecular simulations helped us to better understand the PDE5 inhibitor–target binding interactions in the atomic level. Results of this study can be useful for designing of novel and safe PDE5 inhibitors with enhanced activity and other tailored properties.     

Investigation of PDE5/PDE6 and PDE5/PDE11 selective potent tadalafil-like PDE5 inhibitors using combination of molecular modeling approaches,molecular fingerprint-based virtual screening protocols and structure-based pharmacophore development

The essential biological function of phosphodiesterase (PDE) type enzymes is to regulate the cytoplasmic levels of intracellular second messengers, 3′,5′-cyclic guanosine monophosphate (cGMP) and/or 3′,5′-cyclic adenosine monophosphate (cAMP). PDE targets have 11 isoenzymes. Of these enzymes, PDE5 has attracted a special attention over the years after its recognition as being the target enzyme in treating erectile dysfunction. Due to the amino acid sequence and the secondary structural similarity of PDE6 and PDE11 with the catalytic domain of PDE5, first-generation PDE5 inhibitors (i.e. sildenafil and vardenafil) are also competitive inhibitors of PDE6 and PDE11. Since the major challenge of designing novel PDE5 inhibitors is to decrease their cross-reactivity with PDE6 and PDE11, in this study, we attempt to identify potent tadalafil-like PDE5 inhibitors that have PDE5/PDE6 and PDE5/PDE11 selectivity. For this aim, the similarity-based virtual screening protocol is applied for the “clean drug-like subset of ZINC database” that contains more than 20 million small compounds. Moreover, molecular dynamics (MD) simulations of selected hits complexed with PDE5 and off-targets were performed in order to get insights for structural and dynamical behaviors of the selected molecules as selective PDE5 inhibitors. Since tadalafil blocks hERG1 K channels in concentration dependent manner, the cardiotoxicity prediction of the hit molecules was also tested. Results of this study can be useful for designing of novel, safe and selective PDE5 inhibitors.     

Effect of O2-monobutyryl guanosine 3', 5'-cyclic monophosphate on hepatic metabolism of free fatty acids.

Livers from fed male rats were perfused $\text{in vitro}$ with O^2′-monobutyryl guanosine 3′,5′-cyclic monophosphate. The output of triglyceride was reduced, while output of ketone bodies and glucose was stimulated by 10^?4M monobutyryl guanosine 3′,5′-cyclic monophosphate. No effect was observed with 10^?5 M nucleotide. Monobutyryl guanosine 3′,5′-cyclic monophosphate did not affect uptake of free fatty acids. In these respects, monobutyryl guanosine 3′,5′-cyclic monophosphate mimics the effects of dibutyryl adenosine 3′,5′-cyclic monophosphate, although the guanylic nucleotide seems to be less potent than the adenosine 3′,5′-cyclic monophosphate derivative.     

Adenosine 3' , 5'-cyclic monophsphate phosphodiesterase activities in the x-irradiation induced rat small bowel adenocarcinoma.

The adenosine 3′, 5′-cyclic monophosphate phosphodiesterase (PDE) activities were evaluated in X-irradiation induced Holtzman rat small bowell adenocarcinoma and age-matched normal small intestine. Within normal small intestine, PDE activity was optimal at pH 7.4, and highly dependent upon the addition of Mg²⁺ or Mn²⁺. Analyses of the rat small bowel adenocarcinoma revealed significantly elevated PDE activities above the normal small bowel which were found to be relatively constant throughout the length of the ileum and jejunum. These findings suggest that the diminished intracellular adenosine 3′, 5′-cyclic monophosphate levels observed in this lesion (1) may be the consequence of elevated PDE activities.     

Cyclic nucleotide esterification: relationship to phosphate ring geometry and growth regulation in synchronized human lymphocytes.

Infrared spectra of neutral aqueous solutions of nucleoside 3′,5′-cyclic monophosphates indicate an increase in the antisymmetric phosphoryl stretching frequency to 1236 cm^?1 from 1215 cm^?1 in trimethylene cyclic phosphates. A further increase to 1242 cm^?1 accompanies esterification of the 2′-ribose hydroxyl. The O²′-esterified and 2′-deoxy cyclic nucleotides examined display both reduced kinase binding and altered phosphoryl stretching frequencies, suggesting that modification of the phosphate ring represents a common feature in decreased kinase activation. Reversible inhibition of mitosis in thymidine-synchronized human lymphocytes by 2 mmN⁶,O²′-dibutyryladenosine 3′,5′-cyclic monophosphate and N⁶-monobutyryladenosine 3′,5′-cyclic monophosphate was observed. However, adenosine 3′,5′-cyclic monophosphate, O²′-monobutyryladenosine 3′,5′-cyclic monophosphate, butyric acid, and ethyl butyrate had no effect on mitosis when present at 2 mm concentrations during S and G2. These results are consistent with hydrolysis of O²′-monobutyryladenosine 3′,5′-cyclic monophosphate and adenosine 3′,5′-cyclic monophosphate by esterase and phosphodiesterase enzymes and suggest that modification of the N⁶ amino group is necessary for the antimitotic activity of N⁶,O²′-dibutyryladenosine 3′, 5′-cyclic monophosphate.     

Partial purification and properties of a multifunctional 3′,5′-cyclic nucleotide phosphodiesterase from Lactuca cotyledons

The extraction, partial purification and properties of a 3′,5′-cyclic nucleotide phosphodiesterase from lettuce cotyledons is described. Purification involved fractional precipation with (NH₄)₂SO₄, chromatography on Sephadex G-200, affinity chromatography on Affi-Gel Blue and non-denaturing polyacrylamide gel electrophoresis. The behaviour of the final enzyme preparation on SDS-polyacrylamide gel electrophoresis was examined and inidcated an M, of ca 62 000. The enzyme from 3′,5′-cyclic nucleotide phosphodiesterases previously isolated from plant tissues in that it exhibits activity towards pyrimidine as well as purine cyclic nucleotides. Furthermore, it hydrolyses cyclic CMP at a comparable rate to that with which it hydrolyses cyclic AMP and cyclic GMP. Both 3′- and 5′-AMP were released, with the 5′-nucleotide being the major product. Whereas the K_m with all three substrates remained constant during the purification procedure, V_max with cyclic AMP was lower than that for cyclic CMP but increased as purification proceeded. The effects were examined of a range of di- and trivalent metal ions on the enzyme activity. Fe³⁺ significantly stimulated the activity, more so when cyclic GMP was the substrate. Cu²⁺ inhibited the activity.     

Hydrolysis of 2',3'-cyclic adenosine monophosphate and 3',5'-cyclic adenosine monophosphate in subcellular fractions of normal and neoplastic mouse spleen

A comparison has been made between the capacity to hydrolyse 2′,3′-cyclic adenosine monophosphate and 3′,5′-cyclic adenosine monophosphate in subcellular fractions of normal and neoplastic (lymphosarcoma) spleen of C₅₇BL mice. The effect of X-irradiation on these activities was tested. Subcellular fractionation of normal and lymphosarcoma spleen points to a different overall localization of the enzymes. The 2′,3′-cyclic nucleotide phosphodiesterase (2′,3′-cAMPase) has its highest specific activity in the particulate fractions of the cell, while the data on 3′,5′-cyclic nucleotide phosphodiesterase (3′,5′-cAMPase) show the highest activity in the soluble fraction. The 2′,3′-cAMPase activity is higher in the tumor as compared to the normal tissue, while the opposite holds for 3′,5′-cAMPase. Total body irradiation of normal mice with a dose of 600 rads of X-rays, results in a clear drop in 2′,3′-cAMPase 48 hours after the exposure. The 3′,5′-cAMPase is hardly affected at this time. Neither imidazol nor Mg⁺⁺ has any influence on the 2′,3′-cAMPase. The pH optimum for 3′,5′-cAMPase and 2′,3′-cAMPase appears to be 7.7 and 6.2 respectively. This report suggests a no-identity of the two enzymes in mouse spleen, a situation different from that found in certain plants.     

Triazole‐Pyridine Dicarbonitrile Targets Phosphodiesterase 4 to Induce Cytotoxicity in Lung Carcinoma Cells

Phosphodiesterase 4 (PDE4) is a key enzyme involved in the hydrolysis of cyclic adenosine monophosphate (cAMP) and widely expressed in several types of cancers. The inhibition of PDE4 results in an increased concentration of intracellular cAMP levels that imparts the anti‐inflammatory response in the target cells. In the present report, two series of triazolo‐pyridine dicarbonitriles and substituted dihydropyridine dicarbonitriles were synthesized using green protocol (TBAB in refluxed water). We next evaluated the title compounds for their cytotoxicity towards lung cancer (A549) cells and identified 7′‐[4‐(methylsulfonyl)phenyl]‐5′‐oxo‐1′,5′‐dihydrospiro[cyclohexane‐1,2′‐[1,2,4]triazolo[1,5‐a]pyridine]‐6′,8′‐dicarbonitrile ( 5h ) and 7′‐(1‐methyl‐1H‐imidazol‐2‐yl)‐5′‐oxo‐1′,5′‐dihydrospiro[cyclohexane‐1,2′‐[1,2,4]triazolo[1,5‐a]pyridine]‐6′,8′‐dicarbonitrile ( 5j ) as lead analogs with the IC₅₀ values of 15.2 and 24.1 μm , respectively. Furthermore, all the new compounds were tested for PDE4 inhibitory activity and 5j showed relatively good inhibitory activity towards PDE4 with inhibition of 50.9 % at 10 μm . In silico analysis demonstrated the favorable interaction of the title compounds with the target enzyme. Taken together, the present study introduces a new scaffold for the development of novel PDE4 inhibitors to fight against inflammatory diseases.     

The effect of anti-inflammatory agents on human synovial fibroblast prostaglandin synthetase

Human synovial fibroblast prostaglandin synthetase activity is inhibited by many different non-steroidal anti-inflammatory agents. Aspirin, indomethacin and phenylbutazone significantly inhibit both PGE₁, PGE₂ and PGF_1α and PGF_2α synthesis; whereas penicillamine and aurothioglucose are more potent inhibitors of the F prostaglandins. Histidine and antimalarials do not inhibit, to a significant degree, human synovial prostaglandin synthetase activity. Hydrocortisone has no direct effect on prostaglandin synthetase activity. No changes in synthetase activity are observed when synovial cells are incubated with hydrocortisone, and the prostaglandin synthetase system subsequently isolated and assayed. The proposed inhibitory effects of hydrocortisone on prostaglandin production by synovium may be the result of an alteration of enzyme substrate or cofactor concentration rather than a direct effect on prostaglandin synthetase.     

Purification and Some Properties of Adenosine (Phosphate) Deaminase from the Liver of the Squid Todarodes pacificus

An enzyme that catalyzed the deamination of adenosine 3′-phenylphosphonate was purified from squid liver to homogeneity as judged by SDS-PAGE. The molecular weight of the enzyme was estimated to be 60,000 by SDS-PAGE and 140,000 by Sephadex G-150 gel filtration. The enzyme deaminated adenosine, 2′-deoxyadenosine, 3′-AMP, and 2′,3′-cyclic AMP, but not adenine, 5′-AMP, 3′,5′-cyclic AMP, ADP, or ATP. The apparent K_m and V_max at pH 4.0 for these substrates were comparable (0.11-0.34mM and 179-295 μmol min^?1 mg^?1, respectively). The enzyme had maximum activity at pH 3.5-4.0 for adenosine 3′-phenylphosphonate, at pH 5.5 for adenosine and 2′-deoxyadenosine, and at pH 4.0 for 2′,3′-cyclic AMP and 3′-AMP when the compounds were at concentration of 0.1 mM. The K_m at 4.0 and 5.5 for each substrate varied, but the Vmax were invariant. These results indicated that the squid enzyme was a novel adenosine (phosphate) deaminase with a unique substrate specificity.     

Control of ribosomal protein phosphorylation in HeLa cells

The effects of a large series of hormones, cyclic nucleotides and metabolic inhibitors on phosphorylation of ribosomal protein S6 in HeLa cells suggest that at least two metabolic pathways are involved. One responds to insulin and epidermal growth factor; the other responds to adenosine 3′, 5′-cyclic monophosphate. Some phosphodiesterase inhibitors can suppress the phosphorylation of S6 that ordinarily is stimulated by insulin.     

CYCLIC NUCLEOTIDES IN MURINE BRAIN: THE TEMPORAL RELATIONSHIP OF CHANGES INDUCED IN ADENOSINE 3',5'-MONOPHOSPHATE AND GUANOSINE 3',5'-MONOPHOSPHATE FOLLOWING MAXIMAL ELECTROSHOCK OR DECAPITATION

–Adenosine 3′,5′-cyclic monophosphate (cyclic AMP) levels increase about 5-fold in the cerebral cortex and 2-fold in the cerebellum following electroconvulsive shock (ECS). The peak levels of cyclic AMP occur at 45 s after ECS in the cerebral cortex, and at 15 s in the cerebellum. In the cerebral cortex, ECS produces twice the cyclic AMP accumulation as does decapitation in a comparable time period; however, the relative effect of a number of neurotropic agents on the cyclic AMP accumulation is essentially the same, whether stimulated by decapitation or by ECS. In the cerebellum, the levels of guanosine 3′,5′-cyclic monophosphate (cyclic GMP) also increase following ECS. The cyclic GMP levels are greatest at 60 s after ECS during the postictal depression. An association between elevated cerebellar cyclic GMP and depression seems unlikely, since CNS depressants either lowered or had no effect on cyclic GMP levels. From these results, cyclic nucleotide profiles following treatments such as ECS or decapitation may be useful in elucidating the molecular events involved in seizures, brain injury and ischemia.     

The cyclic nucleotide phosphodiesterases of spinach chloroplasts and microsomes

Cyclic nucleotide phosphodiesterase was extracted from intact chloroplasts and partially purified. Peak 1_c activity from Sephadex G-200 was resolved by electrophoresis into two major bands (MWs 1.87 × 10⁵ and 3.7 × 10⁵). Both also possessed acid phosphatase, ribonuclease, nucleotidase and ATPase. The chloroplast peak 1_c cyclic nueleotide phosphodiesterase was located in the envelope. Peak 1_m cyclic nucleotide phosphodiesterase obtained from the microsomal fraction had a MW of 2.63 × 10⁵. Electrophoresis separated 1_m into two bands of cyclic nucleotide phosphodiesterase activity (MWs 2.63 × 10⁵ and 1.28 × 10⁵). Both contain ATPase, ribonuclease, nucleotidase, but not acid phosphatase. Peak 1_c has high activity towards 3′:5′-cyclic AMP and 3′:5′-cyclic GMP but little towards 2′:3′-cyclic nucleotides. Peak 1_m showed most activity towards 2′:3′-cyclic AMP, 2′:3′-cyclic GMP and 2′:3′-cyclic CMP with little activity towards 3′:5′-cyclic nucleotides. With 1_c, 3′:5′-cyclic AMP and 3′:5′-cyclic GMP exhibit mixed-type inhibition towards one another. The 2′:3′-cyclic AMP phosphodiesterase 1_m was competitively inhibited by 2′:3′-cyclic GMP. p-Chloromercuribenzoate inhibits 1_c but not 1_m. Electrophoresis after dissociation indicates that 1_c and 1_m are both enzyme complexes. After dissociation, the 1_c complex but not that of 1_m could be reassociated. The ribonuclease of the 1_m complex hydrolyses RNA to yield 2′:3′-cyclic nucleotides as the main products. These results are compatible with the 1_c cyclic nucleotide phosphodiesterase complex being involved in the metabolism of 3′:5′-cyclic AMP, and the 1_m complex being concerned with RNA catabolism.     

Cyclic Purine Mononucleotides: Induction of Gibberellin Biosynthesis in Barley Endosperm

The de novo synthesis of α-amylase in barley endosperm and isolated aleurone layers is induced by 3′,5′-cyclic purine mononucleotides and gibberellic acid. The induction of α-amylase by cyclic purine mononucleotides is prevented by 2,4-DNP, inhibitors of RNA and protein syntheses, CCC, AMO-1618 and phosfon. The induction of α-amylase formation by 3′,5′-cyclic purine mononucleotides, but not by gibberellic acid, is also blocked by inhibitors of DNA synthesis. Extracts from cyclic AMP-treated endosperm halves exhibit a characteristic gibberellin-like activity which is detectable within 12 hours from the addition of the cyclic AMP. On paper chromatograms this gibberellin-like activity is located at the Rf typical for GA₃. Its formation is prevented by inhibitors of DNA synthesis, CCC and AMO-1618. Glucose inhibits the formation of α-amylase induced by gibberellic acid. Glucose has no effect on the cAMP-induced gibberellin biosynthesis. The evidence shows that the cyclic purine mononucleotides induce DNA synthesis, which results in gibberellin biosynthesis, which in turn activates the synthesis of α-amylase.     

The effect of N6-2'-O-dibutyryl-adenosine 3',5'-monophosphate on rat liver calciferol 25-hydroxylase.

Liver calciferol 25-hydroxylase activity of vitamin-D deficient rats was enhanced 24 hours following the intravenous injection of N⁶-2′-O-dibutyryl adenosine 3′,5′-monophosphate. Sodium butyrate administered in the same way had no effect on this enzyme system. Administration of actinomycin D with N⁶-2′-O-dibutyryl adenosine 3′,5′-monophosphate abolished the stimulatory effect of the cyclic nucleotide. Direct addition to the incubation medium of adenosine 3′,5′-cyclic monophosphate or of its dibutyryl derivative did not influence the hepatic conversion of cholecalciferol to 25-hydroxycholecalciferol. These results suggest a possible role for the cyclic nucleotide in the regulation of this enzyme system.     

On the presence of adenosine 3′, 5′ -cyclic monophosphate in moss (Funariahygrometrica)

Partially purified nucleotide fraction of moss containing [14C]-labelled putative adenosine 3′, 5′ -cyclic monophosphate (cAMP) and marker authentic [3H] -cAMP was characterized by chemical deamination and also by the enzymatic hydrolysis with beef heart cyclic nucleotide phosphodiesterase. A significant conversion of marker authentic [3H] -cAMP into [3H] -inosine 3′, 5′ -cyclic monophosphate (cIMP) and [3H] -5′ adenosine monophosphate was observed by respective treatments. In contrast, the [14C] -labelled putative cAMP from control and theophylline-treated moss tissue was insensitive to chemical deamination and enzymatic hydrolysis. Apparently, the [14C] -labelled product which comigrates with authentic [3H] -cAMP does not represent true cAMP. Both the methods employed for characterization of the labelled putative cAMP were sensitive enough to detect picomole quantities of authentic [3H] -cAMP. Lack of detectability of prelabelled [14C] -cAMP in our preparations implies that the tissue may contain authentic cyclic AMP below the picomole levels. Thus, the attributed physiological role to adenosine 3′, 5′ -cyclic monophosphate in moss tissue appears somewhat skeptical.     

Inhibition of 3',5'-nucleotide phosphodiesterase in guinea pig cerebral cortical slices

Several compounds have been tested for their activity as inhibitors of 3′,5′-nucleotide phosphodiesterase in brain cortical slices from guinea pig. SQ 20,009 (1-ethyl-4-isopropylidenehydrazino)-1H-pyrazolo (3,4-b)pyridine-5-carboxylate, ethylester, hydrochloride), a very potent inhibitor of 3′,5′-nucleotide phosphodiesterase from rat and rabbit brain shows only moderate activity as 3′,5′-nucleotide phosphodiesterase inhibitor when tested in brain slices. It enhances cyclic AMP accumulation only when slices are stimulated by histamine. It does not affect cyclic AMP levels when histamine/norepinephrine are used as stimuli of cyclic AMP formation and decreases the activity of adenosine as stimulant slightly. Ro 20–1724 (4-(3-butoxy-4-methoxy)-2-imidazolidinone) a potent inhibitor of canine cerebral cortex PDE activity effectively augments the increase in cyclic AMP under all stimulating conditions mentioned, as does to a somewhat smaller extent the more water soluble Ro 20–2926 (4-(3-ethoxy-ethoxy-4-methoxy)-2-imidazolidinone). Dose-response curves for Ro 20–1724 under three stimulating conditions of increased cyclic AMP formation (0.1 mm histamine, 0.1 mm histamine/0.1 mm norepinephrine, 0.1 mm adenosine) yield an ED₅₀ of about 20 μm in all instances. A significant increase over respective controls is seen even at 1 μm Ro 20–1724 (histamine/norepinephrine). The drugs may be useful as tools for studying the regulation of cyclic AMP levels in the central nervous system.     

Establishment of a clonal strain of hepatoma cells which maintain in culture the five enzymes of the urea cycle

A clonal strain of epithelial cells has been established from the transplantable Morris hepatoma 7800 and is designated 7800C₁. The cells grow with a population doubling time of about three days in serum-supplemented synthetic medium. Cells of the 7800C₁ strain have maintained measurable activities of all the enzymes of the urea cycle during 17 months in continuous culture. The activity of argininosuccinate lyase is approximately that found in normal rat liver, while argininosuccinate synthetase, carbamoyl phosphate synthetase, arginase and ornithine carbamoyl transferase activities are, respectively, 40%, 28%, 6%. and 1% of normal values. Treatment of 7800C₁ cells with glucagon, dibutyryl 3′,5′-cyclic adenosine monophosphate or hydrocortisone did not increase the activity of any of the five enzymes.     

Detection of a 3′, 5′-cyclic-AMP phosphodiesterase activity in the lichenized fungus Evernia prunastri

Abstract

A 3′, 5′-cyclic-AMP phosphodiesterase (PDE) was detected and measured in the lichen Evernia prunastri. The percentage of hydrolysis of tritiated 3′, 5′-cyclic-adenosine monophosphate ([³H]-cAMP) and 3′, 5′-cyclic-guanosine monophosphate ([³H]-cGMP) by the PDE enzyme into tritiated 5′-adenosine-monophospahte ([³H]-AMP) and tritiated 5′-guanosine-monophospahte ([³H]-GMP) was measured by treating the PDE products with a 5′-nucleotidase enzyme present in snake venom. The lysate fraction (L) (plasma membranes and cell walls) and the supernatant (S) (soluble fraction of the cells) were tested. In both fractions, competition of unlabelled cAMP, but not unlabelled cGMP, was revealed. Specific competitive PDE inhibitors such as IBMX inhibited enzymatic activity. Although it is thought that in this species cAMP is regulated by red/far red light through PDE activity, this is the first report that seems to suggest the presence of a PDE activity specific for cAMP in lichenized fungi. However, this work is at a preliminary stage and despite the high levels of enzymatic activity with cAMP found in both fractions, data are still insufficient to state the absolute specificity for this nucleotide.     

Identification and characterization of the adenosine 3′,5′-cyclic monophosphate binding proteins appearing during the development of Dictyostelium discoideum

A photosensitive, radioactive analogue of cyclic adenosine monophosphate, 8-azido-adenosine 3′,5′-[³²P]monophosphate (8-N₃-cyclic AMP), was used to label the cyclic AMP binding proteins of Dictyostelium discoideum. During development cytosolic proteins appear which are specifically labeled by the photoaffinity agent. The proteins are developmentally regulated since they are only found in starved, developing cells. Unlabeled cyclic AMP competes specifically with the labeled analogue for protein binding sites in contrast to unlabeled 5′-AMP which does not compete. A mutant which develops spores but is deficient in stalk cell production produces a different set of cyclic AMP binding proteins from the parent strain.