Ca2+-dependent synaptotagmin binding to SNAP-25 is essential for Ca2+-triggered exocytosis

Synaptotagmin is a proposed Ca2+ sensor on the vesicle for regulated exocytosis and exhibits Ca2+-dependent binding to phospholipids, syntaxin, and SNAP-25 in vitro, but the mechanism by which Ca2+ triggers membrane fusion is uncertain. Previous studies suggested that SNAP-25 plays a role in the Ca2+ regulation of secretion. We found that synaptotagmins I and IX associate with SNAP-25 during Ca2+-dependent exocytosis in PC12 cells, and we identified C-terminal amino acids in SNAP-25 (Asp179, Asp186, Asp193) that are required for Ca2+-dependent synaptotagmin binding. Replacement of SNAP-25 in PC12 cells with SNAP-25 containing C-terminal Asp mutations led to a loss-of-function in regulated exocytosis at the Ca2+-dependent fusion step. These results indicate that the Ca2+-dependent interaction of synaptotagmin with SNAP-25 is essential for the Ca2+-dependent triggering of membrane fusion.     

The R-SNARE motif of tomosyn forms SNARE core complexes with syntaxin 1 and SNAP-25 and down-regulates exocytosis

Tomosyn is a 130-kDa syntaxin-binding protein that contains a large N-terminal domain with WD40 repeats and a C-terminal domain homologous to R-SNAREs. Here we show that tomosyn forms genuine SNARE core complexes with the SNAREs syntaxin 1 and SNAP-25. In vitro studies with recombinant proteins revealed that complex formation proceeds from unstructured monomers to a stable four-helical bundle. The assembled complex displayed features typical for SNARE core complexes, including a profound hysteresis upon unfolding-refolding transitions. No stable complexes were formed between the SNARE motif of tomosyn and either syntaxin or SNAP-25 alone. Furthermore, both native tomosyn and its isolated C-terminal domain competed with synaptobrevin for binding to endogenous syntaxin and SNAP-25 on inside-out sheets of plasma membranes. Tomosyn-SNARE complexes were effectively disassembled by the ATPase N-ethylmaleimide-sensitive factor together with its cofactor alpha-SNAP. Moreover, the C-terminal domain of tomosyn was as effective as the cytoplasmic portion of synaptobrevin in inhibiting evoked exocytosis in a cell-free preparation derived from PC12 cells. Similarly, overexpression of tomosyn in PC12 cells resulted in a massive reduction of exocytosis, but the release parameters of individual exocytotic events remained unchanged. We conclude that tomosyn is a soluble SNARE that directly competes with synaptobrevin in the formation of SNARE complexes and thus may function in down-regulating exocytosis.     

SNAP-25 inhibits L-type Ca2+ channels in feline esophagus smooth muscle cells

We recently reported that non-secretory gastrointestinal smooth muscle cells also possessed SNARE proteins, of which SNAP-25 regulated Ca(2+)-activated (K(Ca)) and delayed rectifier K(+) channels (K(V)). Voltage-gated, long lasting (L-type) calcium channels (L(Ca)) play an important role in excitation-contraction coupling of smooth muscle. Here, we show that SNAP-25 could also directly inhibit the L-type Ca(2+) channels in feline esophageal smooth muscle cells at the SNARE complex binding synprint site. SNARE proteins could therefore regulate additional cell actions other than membrane fusion and secretion, in particular, coordinated muscle membrane excitability and contraction, through their actions on membrane Ca(2+) and K(+) channels.     

A transient N-terminal interaction of SNAP-25 and syntaxin nucleates SNARE assembly

The SNARE proteins syntaxin, SNAP-25, and synaptobrevin play a central role during Ca(2+)-dependent exocytosis at the nerve terminal. Whereas syntaxin and SNAP-25 are located in the plasma membrane, synaptobrevin resides in the membrane of synaptic vesicles. It is thought that gradual assembly of these proteins into a membrane-bridging ternary SNARE complex ultimately leads to membrane fusion. According to this model, syntaxin and SNAP-25 constitute an acceptor complex for synaptobrevin. In vitro, however, syntaxin and SNAP-25 form a stable complex that contains two syntaxin molecules, one of which is occupying and possibly obstructing the binding site of synaptobrevin. To elucidate the assembly pathway of the synaptic SNAREs, we have now applied a combination of fluorescence and CD spectroscopy. We found that SNARE assembly begins with the slow and rate-limiting interaction of syntaxin and SNAP-25. Their interaction was prevented by N-terminal but not by C-terminal truncations, suggesting that for productive assembly all three participating helices must come together simultaneously. This suggests a complicated nucleation process that might be the reason for the observed slow assembly rate. N-terminal truncations of SNAP-25 and syntaxin also prevented the formation of the ternary complex, whereas neither N- nor C-terminal shortened synaptobrevin helices lost their ability to interact. This suggests that binding of synaptobrevin occurs after the establishment of the syntaxin-SNAP-25 interaction. Moreover, binding of synaptobrevin was inhibited by an excess of syntaxin, suggesting that a 1:1 interaction of syntaxin and SNAP-25 serves as the on-pathway SNARE assembly intermediate.     

Suppression of Ca2+ syntillas increases spontaneous exocytosis in mouse adrenal chromaffin cells

A central concept in the physiology of neurosecretion is that a rise in cytosolic [Ca²⁺] in the vicinity of plasmalemmal Ca²⁺ channels due to Ca²⁺ influx elicits exocytosis. Here, we examine the effect on spontaneous exocytosis of a rise in focal cytosolic [Ca²⁺] in the vicinity of ryanodine receptors (RYRs) due to release from internal stores in the form of Ca²⁺ syntillas. Ca²⁺ syntillas are focal cytosolic transients mediated by RYRs, which we first found in hypothalamic magnocellular neuronal terminals. (scintilla, Latin for spark; found in nerve terminals, normally synaptic structures.) We have also observed Ca²⁺ syntillas in mouse adrenal chromaffin cells. Here, we examine the effect of Ca²⁺ syntillas on exocytosis in chromaffin cells. In such a study on elicited exocytosis, there are two sources of Ca²⁺: one due to influx from the cell exterior through voltage-gated Ca²⁺ channels, and that due to release from intracellular stores. To eliminate complications arising from Ca²⁺ influx, we have examined spontaneous exocytosis where influx is not activated. We report here that decreasing syntillas leads to an increase in spontaneous exocytosis measured amperometrically. Two independent lines of experimentation each lead to this conclusion. In one case, release from stores was blocked by ryanodine; in another, stores were partially emptied using thapsigargin plus caffeine, after which syntillas were decreased. We conclude that Ca²⁺ syntillas act to inhibit spontaneous exocytosis, and we propose a simple model to account quantitatively for this action of syntillas.     

Synthesis of vesicle cargo determines amplitude of Ca(2+)-sensitive exocytosis

Here we examine the potential coupling between the synthesis of secretory proteins and the sensitivity of exocytosis to the concentration of free Ca(2+) in the cytosol ([Ca(2+)](i)) in plant cell. We therefore monitor in tobacco protoplasts the excursion of the membrane capacitance in response to an elevation of [Ca(2+)](i) as a measure for exocytotic activity. The data show that a ramp like elevation of [Ca(2+)](i) generates in protoplasts from wild type plants and from transgenic plants, which overexpress the secreted α-amylase, an exocytotic burst with an initial steep and a subsequent slow phase. The largest capacitive burst is obtained in α-amylase producing plants and the amplitude of the [Ca(2+)](i) evoked C(m) excursion is a function of the amylase synthesis of the plants. The data support a model according to which plant cells have at least two serial [Ca(2+)](i) sensitive processes in the final steps of their exocytotic pathway. The overproduction of a secreted cargo does not affect the kinetics of this process but the number of vesicles in pools upstream of the [Ca(2+)](i) sensitive steps.     

An inhibitory effect of extracellular Ca2+ on Ca2+-dependent exocytosis

Aim

Neurotransmitter release is elicited by an elevation of intracellular Ca²⁺ concentration ([Ca²⁺]_i). The action potential triggers Ca²⁺ influx through Ca²⁺ channels which causes local changes of [Ca²⁺]_i for vesicle release. However, any direct role of extracellular Ca²⁺ (besides Ca²⁺ influx) on Ca²⁺-dependent exocytosis remains elusive. Here we set out to investigate this possibility on rat dorsal root ganglion (DRG) neurons and chromaffin cells, widely used models for studying vesicle exocytosis.

Results

Using photolysis of caged Ca²⁺ and caffeine-induced release of stored Ca²⁺, we found that extracellular Ca²⁺ inhibited exocytosis following moderate [Ca²⁺]_i rises (2–3 µM). The IC₅₀ for extracellular Ca²⁺ inhibition of exocytosis (ECIE) was 1.38 mM and a physiological reduction (∼30%) of extracellular Ca²⁺ concentration ([Ca²⁺]_o) significantly increased the evoked exocytosis. At the single vesicle level, quantal size and release frequency were also altered by physiological [Ca²⁺]_o. The calcimimetics Mg²⁺, Cd²⁺, G418, and neomycin all inhibited exocytosis. The extracellular Ca²⁺-sensing receptor (CaSR) was not involved because specific drugs and knockdown of CaSR in DRG neurons did not affect ECIE.

Conclusion/Significance

As an extension of the classic Ca²⁺ hypothesis of synaptic release, physiological levels of extracellular Ca²⁺ play dual roles in evoked exocytosis by providing a source of Ca²⁺ influx, and by directly regulating quantal size and release probability in neuronal cells.     

Localized Ca2+ uncaging reveals polarized distribution of Ca2+-sensitive Ca2+ release sites: mechanism of unidirectional Ca2+ waves

Ca2+-induced Ca2+ release (CICR) plays an important role in the generation of cytosolic Ca2+ signals in many cell types. However, it is inherently difficult to distinguish experimentally between the contributions of messenger-induced Ca2+ release and CICR. We have directly tested the CICR sensitivity of different regions of intact pancreatic acinar cells using local uncaging of caged Ca2+. In the apical region, local uncaging of Ca2+ was able to trigger a CICR wave, which propagated toward the base. CICR could not be triggered in the basal region, despite the known presence of ryanodine receptors. The triggering of CICR from the apical region was inhibited by a pharmacological block of ryanodine or inositol trisphosphate receptors, indicating that global signals require coordinated Ca2+ release. Subthreshold agonist stimulation increased the probability of triggering CICR by apical uncaging, and uncaging-induced CICR could activate long-lasting Ca2+ oscillations. However, with subthreshold stimulation, CICR could still not be initiated in the basal region. CICR is the major process responsible for global Ca2+ transients, and intracellular variations in sensitivity to CICR predetermine the activation pattern of Ca2+ waves.     

Raising the cytosolic Ca2+ concentration increases the membrane capacitance of maize coleoptile protoplasts: Evidence for Ca2+-stimulated exocytosis

Enhanced elongation of coleoptile cells has been proposed to be related to a rise in secretory activity. Therefore, to obtain a direct measurement of exocytotic events in maize (Zea mays L.) coleoptile protoplasts we used the patch-clamp method to record changes in membrane capacitance (Cm) as a parameter proportional to fluctuations of the membrane surface area. The secretory activity of protoplasts was correlated with the cytosolic free Ca²⁺ concentration ([Ca²⁺]_cyt): dialyzing protoplasts with 1 M [Ca²⁺]_cyt caused a steady rise in Cm of 3.3 ± pF·s^–1. In contrast, dialysis with a solution containing <20 nM Ca²⁺ produced a small and persistent decrease in Cm. This demonstrates that secretory activity in coleoptile cells can be controlled by factors which modulate [Ca²⁺]_cyt.Abbreviation Cm membrane capacitance This work was made possible by a visiting grant from the Research Council of Slovenia and financial support of the Deutsche Forschungsgemeinschaft to G.T. We are grateful to Dr. W. Diekmann (University of Göttingen) for teaching us the preparation of coleoptile protoplasts.     

The effect of troponin C removal on the Ca2+-sensitive binding of Mg2+ AMPPNP to myofibrils

It was previously shown that when rabbit skeletal myofibrils are titrated with Mg2+ AMPPNP under conditions that result in the dissociation of cross-bridges from the thin filaments (i.e. 50% ethylene glycol, 0 degrees C), Ca2+-sensitive, biphasic binding is observed. These titrations have been repeated using myofibrils from which the troponin C has been selectively removed. The disappearance of both Ca2+ sensitivity and biphasic binding is taken as evidence that the Ca2+ sensitivity is due to Ca2+ binding to troponin C and the biphasic binding of Mg2+ AMPPNP observed in intact myofibrils is not due to packing constraints or steric hindrance.     

Ca2+ signaling and exocytosis in pituitary corticotropes

The secretion of adrenocorticotrophin (ACTH) from corticotropes is a key component in the endocrine response to stress. The resting potential of corticotropes is set by the basal activities of TWIK-related K(+) (TREK)-1 channel. Corticotrophin-releasing hormone (CRH), the major ACTH secretagogue, closes the background TREK-1 channels via the cAMP-dependent pathway, resulting in depolarization and a sustained rise in cytosolic [Ca(2+)] ([Ca(2+)](i)). By contrast, arginine vasopressin and norepinephrine evoke Ca(2+) release from the inositol trisphosphate (IP(3))-sensitive store, resulting in the activation of small conductance Ca(2+)-activated K(+) channels and hyperpolarization. Following [Ca(2+)](i) rise, cytosolic Ca(2+) is taken into the mitochondria via the uniporter. Mitochondrial inhibition slows the decay of the Ca(2+) signal and enhances the depolarization-triggered exocytotic response. Both voltage-gated Ca(2+) channel activation and intracellular Ca(2+) release generate spatial Ca(2+) gradients near the exocytic sites such that the local [Ca(2+)] is ~3-fold higher than the average [Ca(2+)](i). The stimulation of mitochondrial metabolism during the agonist-induced Ca(2+) signal and the robust endocytosis following stimulated exocytosis enable corticotropes to maintain sustained secretion during the diurnal ACTH surge. Arachidonic acid (AA) which is generated during CRH stimulation activates TREK-1 channels and causes hyperpolarization. Thus, corticotropes may regulate ACTH release via an autocrine feedback mechanism.     

Myosin-Va regulates exocytosis through the submicromolar Ca2+-dependent binding of syntaxin-1A

Myosin-Va is an actin-based processive motor that conveys intracellular cargoes. Synaptic vesicles are one of the most important cargoes for myosin-Va, but the role of mammalian myosin-Va in secretion is less clear than for its yeast homologue, Myo2p. In the current studies, we show that myosin-Va on synaptic vesicles interacts with syntaxin-1A, a t-SNARE involved in exocytosis, at or above 0.3 microM Ca2+. Interference with formation of the syntaxin-1A-myosin-Va complex reduces the exocytotic frequency in chromaffin cells. Surprisingly, the syntaxin-1A-binding site was not in the tail of myosin-Va but rather in the neck, a region that contains calmodulin-binding IQ-motifs. Furthermore, we found that syntaxin-1A binding by myosin-Va in the presence of Ca2+ depends on the release of calmodulin from the myosin-Va neck, allowing syntaxin-1A to occupy the vacant IQ-motif. Using an anti-myosin-Va neck antibody, which blocks this binding, we demonstrated that the step most important for the antibody's inhibitory activity is the late sustained phase, which is involved in supplying readily releasable vesicles. Our results demonstrate that the interaction between myosin-Va and syntaxin-1A is involved in exocytosis and suggest that the myosin-Va neck contributes not only to the large step size but also to the regulation of exocytosis by Ca2+.     

The distribution of [3H]kainate binding sites in primate hippocampus is similar to the distribution of both Ca2+-sensitive and Ca2+-insensitive [3H]kainate binding sites in rat hippocampus

The distribution of [³H]kainate binding sites was determined by quantitative autoradiography in three vertebrate species: rat, monkey, and human. These animals displayed a similar pattern of binding site density in the hippocampus. Highest levels were found within the stratum lucidum and moderate levels in the inner portion of the dentate gyrus molecular layer. Although the distribution is similar, there is a lower density of binding sites in the stratum lucidum of primates than in rodents. Experiments using rat brain synaptic plasma membrane fractions indicated that inclusion of Ca²⁺ ions results in a selective reduction in binding at the high affinity sites. The Ca²⁺-inhibited and Ca²⁺-insensitive binding sites in the rat hippocampus exhibited a similar distribution. Together, these results suggest that in a variety of mammalian species kainate receptors exhibit similar regional distributions, and that the high anf loe affinity kainate binding sites also exhibit similar regional distributions.Special Issue dedicated to Prof. Eduardo De Robertis.     

Proteolysis of synaptobrevin, syntaxin, and SNAP-25 in alveolar epithelial type II cells

Synaptobrevin-2, syntaxin-1, and SNAP-25 were identified in rat alveolar epithelial type II cells by Western blot analysis. Synaptobrevin-2 was localized in the lamellar bodies, and syntaxin-1 and SNAP-25 were found in 0.4% Nonidet P40-soluble and -insoluble fractions, respectively, of the type II cells. When the isolated type II cells were stimulated for secretion with calcium ionophore A23187 or with phorbol 12-myristate 13-acetate, these proteins were found to have been proteolyzed. Preincubation of cells with calpain inhibitor II (N-acetylleucylleucylmethionine), however, prevented the proteolysis. Treatment of the cell lysate with exogenous calpain resulted in a time-dependent decrease of these proteins. The data suggest that synaptobrevin, syntaxin, and SNAP-25 are subject to proteolytic modification by activated calpain in intact type II cells stimulated for secretion.     

Differential contribution of syntaxin 1 and SNAP-25 to secretion in noradrenergic and adrenergic chromaffin cells

We used botulinum neurotoxins (BoNT) to examine whether differences in the secretory activity of noradrenergic and adrenergic chromaffin cells are related to differences in the exocytotic machinery of these two types of bovine adrenal medulla cells. Cleavage of syntaxin and SNAP-25 by BoNT/C1 decreased in a dose-dependent way the release of both noradrenaline and adrenaline, but noradrenaline release was more sensitive to BoNT/C1. Cleavage of SNAP-25 by BoNT/A also had a larger inhibitory effect on noradrenaline release than on adrenaline release. Neither BoNT/C1 nor BoNT/A affected the intracellular Ca2+ responses induced by K+-depolarisation, and the extent of the inhibition of K+-evoked catecholamine release by selective blockers of voltage-gated Ca2+ channels was not affected by BoNT/C1. Therefore, our data do not support the hypothesis of a regulatory effect of syntaxin or SNAP-25 on the activity of Ca2+ channels. The lower sensitivity of adrenaline release to BoNT was not due to a reduced ability of the toxins to enter or to cleave their protein targets in adrenergic cells, since immunoblot analysis showed the cleavage of a larger fraction of syntaxin 1A in adrenergic cells, as compared to the cleavage in noradrenergic cells. The immunoblot analysis also showed larger amounts of syntaxin 1A in noradrenergic chromaffin cells than in adrenergic cells. Thus, in spite of a greater cleavage of syntaxin 1A in adrenergic cells by BoNT/C1, adrenaline release was less sensitive to BoNT/C1, suggesting that the release process in noradrenergic cells might be more dependent on syntaxin 1A and SNAP-25, as compared to adrenergic cells.     

Isolation and characterization of Ca2+-sensitive mutants of Saccharomyces cerevisiae

Thirty Ca2+-sensitive (cls: calcium sensitive) mutants of Saccharomyces cerevisiae were isolated by replica-plating. These mutants, which each had a single recessive chromosomal mutation, were divided into 18 complementation groups. Some cls mutants showed a phenotype of specific sensitivity to Ca2+, while others showed phenotypes of sensitivities to several divalent cations. From measurements of the calcium contents and initial rates of Ca2+ uptake of the cls mutants, 16 of the 18 cls complementation groups were classified into four types: type I mutants (cls5, cls6, cls13, cls14, cls15, cls16, cls17, and cls18) had both elevated calcium contents and increased uptake activities. A type II mutant (cls4) had a normal calcium content and normal uptake activity; type III mutants (cls1, cls2 and cls3) had elevated calcium contents but normal initial rates of Ca2+ uptake; type IV mutants (cls8, cls9, cls10 and cls11) had normal calcium contents but increased initial rates of Ca2+ uptake. Two of the mutants (cls7 and cls12) had intermediate biochemical properties. The primary defects of these four types of cls mutants were considered in terms of the Ca2+ transport system(s). Both type I and type III mutants, which had elevated calcium contents, simultaneously showed a trifluoperazine-sensitive phenotype, suggesting a close correlation of this phenotype with elevated calcium content. In addition, all type IV mutants were unable to utilize nonfermentable sugars. One CLS gene, CLS7, was located on the left arm of chromosome V.     

A highly Ca2+-sensitive pool of vesicles contributes to linearity at the rod photoreceptor ribbon synapse

Studies of the properties of synaptic transmission have been carried out at only a few synapses. We analyzed exocytosis from rod photoreceptors with a combination of physiological and ultrastructural techniques. As at other ribbon synapses, we found that rods exhibited rapid kinetics of release, and the number of vesicles in the releasable pool is comparable to the number of vesicles tethered at ribbon-style active zones. However, unlike other previously studied neurons, we identified a highly Ca(2+)-sensitive pool of releasable vesicles with a relatively shallow relationship between the rate of exocytosis and [Ca(2+)](i) that is nearly linear over a presumed physiological range of intraterminal [Ca(2+)]. The low-order [Ca(2+)] dependence of release promotes a linear relationship between Ca(2+) entry and exocytosis that permits rods to relay information about small changes in illumination with high fidelity at the first synapse in vision.     

Removal of Ca2+ channel beta3 subunit enhances Ca2+ oscillation frequency and insulin exocytosis

An oscillatory increase in pancreatic beta cell cytoplasmic free Ca2+ concentration, [Ca2+]i, is a key feature in glucose-induced insulin release. The role of the voltage-gated Ca2+ channel beta3 subunit in the molecular regulation of these [Ca2+]i oscillations has now been clarified by using beta3 subunit-deficient beta cells. beta3 knockout mice showed a more efficient glucose homeostasis compared to wild-type mice due to increased glucose-stimulated insulin secretion. This resulted from an increased glucose-induced [Ca2+]i oscillation frequency in beta cells lacking the beta3 subunit, an effect accounted for by enhanced formation of inositol 1,4,5-trisphosphate (InsP3) and increased Ca2+ mobilization from intracellular stores. Hence, the beta3 subunit negatively modulated InsP3-induced Ca2+ release, which is not paralleled by any effect on the voltage-gated L type Ca2+ channel. Since the increase in insulin release was manifested only at high glucose concentrations, blocking the beta3 subunit in the beta cell may constitute the basis for a novel diabetes therapy.     

Direct interaction of the Rab3 effector RIM with Ca2+ channels, SNAP-25, and synaptotagmin.

To define the role of the Rab3-interacting molecule RIM in exocytosis we searched for additional binding partners of the protein. We found that the two C(2) domains of RIM display properties analogous to those of the C(2)B domain of synaptotagmin-I. Thus, RIM-C(2)A and RIM-C(2)B bind in a Ca(2+)-independent manner to alpha1B, the pore-forming subunit of N-type Ca(2+) channels (EC(50) = approximately 20 nm). They also weakly interact with the alpha1C but not the alpha1D subunit of L-type Ca(2+) channels. In addition, the C(2) domains of RIM associate with SNAP-25 and synaptotagmin-I. The binding affinities for these two proteins are 203 and 24 nm, respectively, for RIM-C(2)A and 224 and 16 nm for RIM-C(2)B. The interactions of the C(2) domains of RIM with SNAP-25 and synaptotagmin-I are modulated by Ca(2+). Thus, in the presence of Ca(2+) (EC(50) = approximately 75 microm) the interaction with synaptotagmin-I is increased, whereas SNAP-25 binding is reduced. Synaptotagmin-I binding is abolished by mutations in two positively charged amino acids in the C(2) domains of RIM and by the addition of inositol polyphosphates. We propose that the Rab3 effector RIM is a scaffold protein that participates through its multiple binding partners in the docking and fusion of secretory vesicles at the release sites.     

Activation of store-operated Ca2+ current in Xenopus oocytes requires SNAP-25 but not a diffusible messenger.

Depletion of Ca2+ stores in Xenopus oocytes activated entry of Ca2+ across the plasma membrane, which was measured as a current I(soc) in subsequently formed cell-attached patches. I(soc) survived excision into inside-out configuration. If cell-attached patches were formed before store depletion, I(soc) was activated outside but not inside the patches. I(soc) was potentiated by microinjection of Clostridium C3 transferase, which inhibits Rho GTPase, whereas I(soc) was inhibited by expression of wild-type or constitutively active Rho. Activation of I(soc) was also inhibited by botulinum neurotoxin A and dominant-negative mutants of SNAP-25 but was unaffected by brefeldin A. These results suggest that oocyte I(soc) is dependent not on aqueous diffusible messengers but on SNAP-25, probably via exocytosis of membrane channels or regulatory molecules.