Peptides in rat brain immunoreactive for the amino terminus of cholecystokinin 33: distribution and chromatography

Antisera directed against the amino-terminus of porcine CCK 33 detects related immunoreactivity in rat brain extracts, the distribution of which follows that of CCK 8. Sephadex chromatography indicates that several immunoreactive peptides are present with a molecular weight range of 2600-3500. These peptides are likely to be CCK 39 or CCK 33 and the amino terminal segments of CCK 39/33 without the CCK 8 sequence. The presence of CCK 39/33 and its amino-terminal fragments without CCK 22 and its amino-terminal fragments confirms the absence of CCK 22 in the rat brain. This cleavage at CCK 22 is one of the major differences between the processing of CCK in rat brain and gut and may reflect differences in their physiological roles.     

Peptides in rat brain immunoreactive for the carboxyl terminal extension of cholecystokinin: distribution and chromatography

Utilizing an antiserum raised against a peptide fragment identical to part of the carboxyl terminal extension of cholecystokinin (CCK) predicted by the sequence of CCK mRNA [7], an antiserum has been generated which does not detect CCK 39, CCK 33, CCK 8, CCK 4 or gastrin 171. This antiserum detects several peptides in rat brain, one similar in size to CCK 33 and another slightly larger than CCK 8. These peptides may represent carboxyl-terminally extended forms of CCK, though their chemical structure has not been determined. These peptides are present in all brain regions where CCK 8 can be detected. The abundance of these peptides, their localization in CCK terminal regions, and their enrichment in synaptosome preparations [1] imply that the tryptic cleavage and amidation reaction occur late in the processing of CCK (as has been observed for other biologically active peptides), and probably occur in the synaptic vesicle.     

Immunoreactive cholecystokinin in human and rat plasma: correlation of pancreatic secretion in response to CCK

Immunoreactive cholecystokinin (CCK) levels in human and rat plasma are described using a radioimmunoassay specific for the biologically active sulfated end of CCK. This assay detected significant changes in plasma cholecystokinin levels during intrajejunal administration of amino acids and intravenous infusions of CCK-8 which were followed by increased pancreatic secretion. In humans, the concentration (pg/ml) of plasma cholecystokinin increased from 10.8 to 18.9 following intrajejunal amino acid instillation and from 15.4 to 31.1 during CCK infusion, while pancreatic trypsin secretion increased more than 15 fold. Ingestion of a test meal also caused a rapid and significant elevation (P less than 0.05) in both plasma CCK (14.5-21.7 pg/ml) and gastrin (50-160 pg/ml) levels. In the rat, an injection of 46 ng of CCK-8 produced a 300% increase in immunoreactive plasma CCK levels (2 min) and caused peak pancreatic protein secretion within 5 min; 4 fold lower doses (11.5 ng) elevated plasma CCK by 38% and pancreatic protein secretion to a small but significant extent. The ability of this assay to detect various forms of sulfated CCK in human plasma was also determined. Following gel chromatography on Sephadex G-50, at least three different immunoreactive peaks were found in plasma from fasted subjects and after intrajejunal amino acid stimulation. While the lower molecular weight CCK peptides (CCK-8 and CCK-12) were detected in plasma from both fasted and stimulated subjects, the larger form (CCK-33) was only present in measurable concentrations after amino acid infusion. The simultaneous measurement of increased plasma CCK levels and pancreatic secretion and the changes in the distribution of CCK peptides following amino acid infusion provides strong support that this assay detects physiologically relevant changes in biologically active CCK peptides.     

The molecular nature of vascularly released cholecystokinin from the isolated perfused porcine duodenum

Using sequence-specific radioimmunoassays, the quantities and molecular nature of cholecystokinin (CCK) have been determined in extracts of porcine duodenal mucosa and in the vascular perfusate from the isolated porcine duodenum. The basal concentration of CCK in the perfusate was 84 pM equiv. CCK-8 (mean; range: 32–173 pM, n = 5). After intraluminal stimulation with amino acids, acidified fat emulsions and hydrochloric acid, the concentrations increased 2–5-fold. Both in the basal and stimulated state the concentrations of the related hormone, gastrin, were below 5 pM equiv. gastrin-17. CCK in the perfusate was concentrated by affinitychromatography using antibodies directed against the bioactive C-terminus. Subsequent gel chromatography revealed a form with a size like or slightly larger than the C-terminal dodecapeptide (CCK-12), a predominant form resembling the C-terminal octapeptide (CCK-8), and a form resembling the C-terminal tetrapeptide (CCK-4). The duodenal mucosa contained in addition CCK-33, -39 and CCK-peptides with further N-terminal extensions. The results suggest that small CCK peptides are the principal circulating forms, while CCK-33 and larger forms are biosynthetic precursors.     

Unique cholecystokinin peptides isolated from guinea pig intestine

Fractionation on Sephadex G50 gel of methanol extracts of guinea pig intestine reveals two molecular forms of cholecystokinin (CCK) of about equal abundance. One elutes at the position of CCK8 while the other elutes at a position intermediate between CCK33 and CCK8. Purification and sequencing of these peptides identify them as CCK8 and CCK22, respectively. Guinea pig CCK8 differs from other mammalian CCK octapeptides isolated thus far in that there is a valine substituted for methionine at position 6 from the C-terminus. In addition to the substitution in CCK8, serine is substituted for asparagine in position 22, glycine for serine in position 19, and asparagine for serine in position 15 from the C-terminus compared to the pig sequence. HPLC separation on a C18 column yields two peaks each of CCK8 and of CCK22 in pig intestinal tissue obtained from a commercial supplier. The two CCK8 peptides have identical amino acid sequences as do the two CCK22 peptides. The CCK22 peptides are equally bioactive in the guinea pig pancreatic acinar cell assay but are about 10-fold less potent than synthetic CCK8(s). One of the guinea pig CCK8 peptides is fully bioactive whereas the other is about 50-fold less potent compared to synthetic CCK8(s).     

Cleavage-site mutagenesis alters post-translation processing of Pro-CCK in AtT-20 cells

Cholecystokinin (CCK) is expressed in the central and peripheral nervous systems and functions as a neurotransmitter and neuroendocrine hormone. The in vivo forms of CCK include CCK-83, -58, -39, -33, -22, -12, and -8. Tissues in the periphery produce the larger forms of CCK, such as CCK-58, whereas the brain primarily produces CCK-8. The different biologically active forms of CCK observed in vivo may result from cell-specific differences in endoproteolytic cleavage during post-translational processing. Evidence suggests that cleavages of pro-CCK occur in a specific sequential order. To further delineate the progression of cleavages during pro-CCK maturation, mutagenesis was used to disrupt putative mono- and dibasic cleavage sites. AtT-20 cells transfected with wild-type rat prepro-CCK secret CCK-22 and -8. Mutagenesis of the cleavage sites of pro-CCK had profound effects on the products that were produced. Substitution of basic cleavage sites with nonbasic amino acids inhibits cleavage and leads to the secretion of pathway intermediates such as CCK-83, -33, and -12. These results suggest that CCK-58 is cleaved to both CCK-33 and -22. Furthermore, CCK-8 and -12 are likely derived from cleavage of CCK-33 but not CCK-22. Alanine substitution at the same site completely blocked production of amidated products, whereas serine substitution did not. The cleavages observed at nonbasic residues in this study may represent the activity of enzymes other than PC1 and carboxypeptidase E, such as the enzyme SKI-1. A model for the progression of pro-CCK processing in AtT-20 cells is proposed. The findings in this study further supports the hypothesis that pro-CCK undergoes parallel pathways of proteolytic cleavages.     

Biosynthesis and post-translational processing of site-directed endoproteolytic cleavage mutants of Pro-CCK in AtT-20 cells.

Site-directed mutagenesis in which individual cleavage site P1 amino acids were changed to Ala was performed to delineate their importance in the processing of pro-CCK in mouse pituitary tumor AtT-20 cells. Individual substitution of cleavage sites on pro-CCK, viz., CCK 58 cleavage site R/A to A/A, CCK 33 cleavage site R/K to A/K, CCK 22 cleavage site K/N to A/N, and CCK 8 cleavage site R/D to A/D, did not inhibit pro-CCK expression or the production of some form of amidated CCK. Wild-type CCK cDNA expression in these cells results in production and secretion of CCK 8 and CCK 22. Substitution of the 58R/A cleavage site with A/A produces only CCK 33; 33A/K and 22A/N produce only CCK 8, whereas 8A/D produces CCK 12 and some CCK 22. Where the GRR residues on the C-terminus of CCK 8 were mutated to GAA, no amidated CCK was produced. Significant amounts of the pro-CCK, C-terminal peptide S9S was found in the medium of cells transfected with GAA mutant cDNA, indicating that this pro-CCK was cleaved at the GAA site probably by a nonprohormone convertase enzyme. Further analysis of the cells expressing the GAA mutant demonstrated that it is not extensively cleaved at other sites to produce CCK 8 GAA or larger peptides. In the mutant where the entire pro-CCK, C-terminal S9S was deleted, CCK 8 is processed and secreted normally. Thus, the cleavage at the C-terminal GRR site is essential for subsequent cleavages, and modification of other cleavage sites (58, 33, 22, and 8) has a major impact on pro-CCK processing. These results suggest that there is a temporal order of cleavages, and the structure of pro-CCK has a strong influence on where and whether pro-CCK is processed.     

Purification and sequencing of a rat intestinal 22 amino acid C-terminal CCK fragment

Fractionation on Sephadex G50 gel of methanol extracts of rat intestine revealed two molecular forms of cholecystokinin (CCK) of about equal immunopotency: one form has an elution volume between CCK33 and CCK12; the other elutes in the salt region as does authentic CCK8. Purification and sequencing have demonstrated that the smaller molecular form is CCK8 with a sequence identical to the pork and sheep CCK8's that had previously been sequenced. Purification and sequencing of the larger molecular form reveals that it is a 22 amino acid C-terminal CCK fragment identical with pig CCK22 except that glycine instead of serine is present at the nineteenth residue from the C-terminus. This sequence is consistent with that predicted by cloned cDNA encoding preprocholecystokinin from a rat medullary thyroid carcinoma. CCK22 has not previously been reported to be a prominent molecular form in either pig or dog intestines.     

Characterization of a cholecystokinin 8-generating endoprotease purified from rat brain synaptosomes.

An endoproteolytic activity that specifically cleaves CCK 33, producing CCK 8, has been purified from a rat brain synaptosome preparation. The purification, which included anion exchange, chromatofocusing, hydroxyapatite, and gel filtration chromatography, resulted in a greater than 3000-fold increase in specific activity. This neutral endoprotease (pH optimum 8) exists as a 90-kDa species, which can be dissociated into active 40-kDa species. The enzyme is a non-trypsin serine protease, which is inhibited by diisopropyl-fluorophosphate and p-aminobenzamidine but not by soybean trypsin inhibitor, phenylmethylsulfonyl fluoride, aprotinin, or a number of thiol or metalloprotease inhibitors. It is highly substrate-specific and cleaves neither trypsin, enteropeptidase, kallikrein substrates, nor analogues of mono- or dibasic cleavage sites of prohormones other than pro-CCK. The endoprotease will not cleave CCK 12 desulfate or CCK (20-29), although these peptides contain common sequences with CCK-33. The protease does cleave [Glu27]CCK (20-29), a peptide in which the glutamate mimics the negative charge normally present on tyrosine sulfate. This suggests that the negative charge at position 27 is important in substrate recognition. The enzyme will also cleave CCK 33 and CCK (1-21) on the carboxyl-terminal side of a single lysine residue in position 11. The subcellular location and specificity of this endoprotease make it a good candidate for a CCK-processing protease.     

New molecular forms of cholecystokinin. Microsequence analysis of forms previously characterized by chromatographic methods

Cholecystokinin previously has been characterized by chromatographic and immunological techniques. Analysis of cDNA has revealed the structure of preprocholecystokinin. The predicted processing sites of preprocholecystokinin cannot account for the multiple forms of cholecystokinin detected. This report details the isolation and characterization of cholecystokinin peptides containing 58, 39, 33, 25, 18, 8, 7, and 5 amino acids. None of the cleavages that occurred to form these peptides was at the carboxyl side of double basic residues. Cholecystokinin 25, 18, and 7 have not previously been isolated and identified by sequence analysis. The processing that forms these peptides includes cleavage after single basic residues for the 58, 39, 33, and 8 amino acid peptides. The 25, 18, 7, and 5 amino acid peptides must be formed by other endopeptidases or combinations of endo- and exopeptidases. The analysis of this series of peptides provides the chemical basis for structural differences in cholecystokinin molecules previously demonstrated by chromatographic and immunological methods. These structures provide insight into tissue-specific processing that occurs for this important regulatory peptide in intestine and brain.     

Gastrointestinal peptides in the brain.

Both immunoreactive intact cholecystokinin (CCK33) and its COOH-terminal octapeptide (CCK8) are detected in brain and gut extracts of monkey, dog, and pig using an antiserum with equivalent sensitivities for detecting CCK8 in the free form or when incorporated in the intact molecule. The failure to detect intact cholecystokinin in extracts from monkey or dog by using an antiserum developed by immunization with porcine CCK33 is due to marked species differences in the NH2-terminal portion of the molecule. Immunohistochemical staining reveals the presence of CCK peptides in rabbit cerebral cortical tissue neurons. Subcellular fractionation of rat cerebral cortical tissue demonstrates that CCK immunoreactivity is concentrated in the pellet identified by electron microscopy to contain a high proportion of synaptic vesicles. A converting enzyme that differs from trypsin has been partially purified from canine and porcine cerebral cortical extracts. It converts porcine CCK to smaller immunoreactive forms, but fails to convert big gastrin to heptadecapeptide gastrin. This enzyme differs from trypsin not only in substrate specificity but also in several physicochemical properties. Cerebral cortical extracts from hyperphagic ob/ob mice have strikingly lower contents of CCK than those from their lean littermates and other normal mice. These studies taken together are consistent with a role for CCK as a neurotransmitter involved in the overall regulation of appetite.     

Large and small forms of cholecystokinin in human plasma: measurement using high pressure liquid chromatography and radioimmunoassay

Defining the role of cholecystokinin (CCK) in gut physiology and disease has proved difficult because of problems with development of radioimmunoassays and because CCK exists in several different molecular forms which have different biological actions. In order to measure small (8 amino acid residues, CCK 8) and large (33 and 39 residues) forms of CCK in plasma we have developed high pressure liquid chromatographic (HPLC) fractionation of plasma prior to radioimmunoassay. Fasting plasma CCK 8 and CCK $\text{33}\text{39}$ levels were usually undetectable (< 3 and < 6 pmol/1, respectively). After a liquid fat meal both CCK 8 and CCK $\text{33}\text{39}$ levels were significantly elevated at 5 min (11.3±3.3 and 11.6±2.6 pmol/1, respectively). Peak CCK 8 levels occurred at 30 min (15.0±4.4 pmol/1) while peak CCK $\text{33}\text{39}$ levels occurred at 120 min (16.7±4.9 pmol/1. Total CCK levels showed a biphasic response to the meal. These CCK 8 and CCK $\text{33}\text{39}$ responses to oral fat are consistent with a role for these hormones in the regulation of gallbladder emptying and pancreatic secretion.     

Evidence against autocrine feedback regulation of cholecystokinin secretion in man.

To determine whether exogenous cholecystokinin (CCK) inhibits endogenous CCK release, cholecystokinin-8S (CCK-8S) was infused intravenously during continuous intraduodenal stimulation of endogenous CCK by a meal. CCK was measured in plasma by 2 region-specific radioimmunoassays employing antibodies T204 and 1703. AB T204 binds to carboxy-terminal CCK peptides containing the sulphated tyrosyl region, including CCK-8S, and AB 1703 to carboxy-terminal CCK peptides containing at least 14 amino acid residues. Meal-stimulated plasma CCK concentrations remained elevated during the entire infusion period. CCK-8S infusion further increased meal-stimulated plasma CCK concentrations, when measured with AB T204, while meal-stimulated plasma CCK concentrations were not suppressed by CCK-8S infusion, when measured with AB 1703. It is concluded that meal-stimulated endogenous CCK release is not affected by the effects of intravenously administered CCK-8S. These data suggest that autocrine feedback regulation of CCK release is not operative in man.     

Characterization of cholecystokinin receptors in bullfrog (Rana catesbeiana) brain and pancreas

The binding of biologically active 125I-Bolton-Hunter-CCK-33 to bullfrog brain and pancreatic membrane particles was characterized. Both tissues exhibited time-dependent, saturable, reversible, and high affinity binding without evidence for cooperative interaction. Both bullfrog CCK receptors resembled their mammalian counterparts in having acidic pH optima for tracer binding and a Kd of about 0.5 nM. However, the receptors differed from their mammalian counterparts in that (1) the bullfrog brain membranes bound more tracer per mg protein than did the pancreatic membranes, (2) both bullfrog CCK receptors were relatively insensitive to dibutyryl cGMP, and (3) both bullfrog brain and pancreatic CCK receptors exhibited the same general specificity toward a variety of CCK and gastrin peptides. For both tissues, the relative order of receptor binding potency was CCK-8 greater than caerulein = CCK-33 greater than gastrin-17-II greater than CCK-8-ns = gastrin-17-I greater than caerulein-ns greater than gastrin-4 with the sulfated CCK peptides being 1000-fold more potent than their nonsulfated analogs. Sulfated gastrin was also relatively potent, being only 10-fold weaker than CCK-8. Gastrin-4 was 20 000-fold weaker than CCK-8 in interacting with the brain CCK receptor. The latter finding is in sharp contrast to the mammalian brain CCK receptor. We conclude that the bullfrog brain and pancreas contain similar CCK receptors of probable physiological significance and may represent an ancestral condition from which the two distinct CCK receptors present in mammalian brain and pancreas have evolved.     

Feeding-suppressive mechanism of sulfated cholecystokinin (26-33) in chicks

The anorexigenic effect of cholecystokinin (CCK) is well documented in mammals, but documentation in neonatal chicks is limited. Thus, the present study investigated the mechanism underlying the anorexigenic effect of CCK in neonatal chicks. Intraperitoneal (IP) injection of sulfated CCK(26-33) (CCK8S) significantly decreased food intake in chicks at 60 and 300 nmol/kg. Non-sulfated CCK(26-33) (CCK8) also significantly decreased food intake, but its anorexigenic effect was observed only at the highest dose (300 nmol/kg) and short-lived. However, CCK(30-33) (CCK4) had no effect on food intake. Also, the intracerebroventricular (ICV) injection of CCK8S (0.2 and 1 nmol) significantly decreased food intake in chicks. Similar to IP administration, the anorexigenic effect of CCK8 was weak and CCK4 did not affect food intake. IP and ICV injections of CCK8S caused conditioned aversion and increased plasma corticosterone concentrations, suggesting that their anorexigenic effects might be related to stress and/or malaise. This might be true in ICV-injected CCK8S because co-injection of astressin, a corticotropin-releasing hormone receptor antagonist, tended to attenuate the effect of CCK8S. The present study revealed that N-terminal amino acids and the sulfation of Tyr are important for the anorexigenic effect of CCK8S after IP and ICV administered in chicks. Additionally, the effect of central CCK8S might be related to stress and/or malaise.     

Isolation and characterization of biologically active and inactive cholecystokinin-octapeptides from human brain

Cholecystokinin-like immunoreactivity (CCK-LI) in 0.9 kg human brain was extracted by 2% trifluoroacetic acid at 4 degrees C. Sephadex G50 gel filtration of crude extract revealed one main molecular form of CCK, detected by a carboxy-terminal antibody (5135), that eluted in the position of CCK8. When the CCK-LI in the extract was purified by affinity chromatography using another carboxyl-terminal CCK antibody followed by several steps of reverse phase high pressure liquid chromatography (HPLC), a component was isolated that was found by sequence analysis to be identical to the carboxyl-terminal CCK-octapeptide of porcine CCK33, isolated from intestinal mucosa, and to CCK-octapeptide, isolated from sheep brain. This component possessed comparable biological potencies to synthetic sulfated CCK8 in eliciting amylase release and in competitively displacing radioiodinated CCK33 from isolated mouse pancreatic acini. Furthermore, it exhibited a similar binding characteristic to CCK8 in binding to specific receptors on mouse brain cortical particulate preparations. On high pressure liquid chromatography another minor, earlier eluting immunoreactive peak was observed, which had the same amino acid composition and sequence as CCK8. These findings suggested that this material was oxidized CCK8. This earlier eluting component, exhibiting CCK8-like immunoreactivity, did not induce amylase release from acini and had no or minimal effect in inhibiting tracer CCK33 binding to receptors on isolated acini or on mouse brain cortical particulate preparations at the concentrations tested.     

Oxidation/reduction of methionine residues in CCK: a study by radioimmunoassay and isocratic reverse phase high pressure liquid chromatography

The study was undertaken to investigate the oxidation and reduction of cholecystokinin (CCK) both as pure standards and as endogenous porcine peptides. Furthermore an attempt was made to prevent oxidation of the endogenous porcine peptides in the extraction procedure. CCK-8 and CCK-33 standards were always oxidized in weak solutions, CCK-8 varying from 26% to 67% oxidized and CCK-33 from 18% to 70%. Similarly, tissue extracts of porcine brain and duodenum contained oxidized forms of the peptide. CCK standards were readily oxidized in the presence of hydrogen peroxide. Oxidized CCK-8 standard and CCK-8 in porcine brain was 90% reduced and oxidized CCK-33 standard and in duodenal extracts was reduced by 70% by a 40 hour incubation with 0.725 mol/l dithiothreitol at 37 degrees C. Extraction of CCK peptides in the presence of 65 mmol/l dithiothreitol resulted in almost complete prevention of oxidation with over 95% of the peptides being obtained in the reduced state. This additive is therefore recommended for all tissue quantitation studies.     

Expression of porcine cholecystokinin cDNA in a murine neuroendocrine cell line. Proteolytic processing, sulfation, and regulated secretion of cholecystokinin peptides

The cDNA for porcine preprocholecystokinin (pre-pro-CCK) was engineered for expression in mammalian cells under the control of the Rous sarcoma virus-long terminal repeat promoter. This expression construct was transfected into the murine anterior pituitary cell line, AtT-20. A stable cell line (AtT-20/CCK) was derived that expresses CCK mRNA indistinguishable from the CCK mRNA found in pig brain or gut. The AtT-20/CCK cells carry out proteolytic processing and sulfation reactions to generate authentic sulfated CCK8 from pro-CCK. The cells also store and secrete CCK-immunoreactive peptides. This secretion can be stimulated with corticotropin releasing factor, the natural secretagogue for anterior pituitary cells. In contrast, monkey kidney epithelial cells (COS cells), which are transiently transfected to express CCK, predominantly secrete nonsulfated pro-CCK into the medium. These studies show that a murine neuroendocrine cell line contains the complete processing machinery required to generate authentic porcine CCK8. The processing events include simultaneous proteolytic processing at one and two basic amino acid sites and sulfation of tyrosine residues. The cell line thus duplicates exactly the processing patterns found to occur in pig brain cortex.     

Cholecystokinin-33 inhibits meal size and prolongs the subsequent intermeal interval

There are various forms of the satiety gut-brain peptide cholecystokinin (CCK), a short, widely utilized form or CCK-8, and a long, putatively more effective form or CCK-33. The issue of which of these forms is a more effective satiety peptide is not resolved. Here, we compared the satiety responses, including the sizes of the first three meals (MS) and intermeal intervals (IMI) as well as their calculated satiety ratios (SR), evoked by both peptides. CCK-8 and 33 (1, 3 and 5 nmol/kg, i.p) reduced the size of the first meal similarly, only CCK-33 prolonged the first IMI and increased SR and both peptides failed to affect second and third MS and IMI. As such, CCK-33 is a more effective satiety peptide than CCK-8. The current results confirm previous findings which showed that both peptides reduce food intake by inhibiting meal size, whereas only CCK-33 reduces food intake by prolonging the intermeal interval.     

Ontogeny of molecular forms of CCK-peptides in rat brain and gut

COOH-terminal immunoreactive cholecystokinin (iCCK) in methanol and acid extracts of brain and gut in the developing rat between 3 and 28 days after birth and in the adult has been fractionated on Sephadex G50. The single peak observed in methanol extracts of brain has an elution volume identical to that of CCK8. Acid extracts of brain contain only 10% as much iCCK as methanol extracts in a molecular form that appears to be a larger precursor because its elution volume on Sephadex is earlier than that of CCK33 or 39. Methanol extracts of duodenum at all ages contain a molecular form which is larger than CCK12 and about equal in abundance to CCK8. Acid extracts of rat duodenum contain 3 peaks: one elutes in the region of a previously described precursor form, CCK58; another co-elutes with CCK33 or 39; the third peak appears to be lower in molecular weight and is most prominent in the 2 week rat duodenum. Whether the larger iCCK found in methanol extracts or the smaller form found in acid extracts is derived from a common CCK precursor is yet to be determined.