(Received for publication, November 15, 1996)
From the Department of Pharmacology, University of Innsbruck, A-6020 Innsbruck, Austria and the ¶ Department of Developmental and Molecular Biology, Albert Einstein College of Medicine, Bronx, New York 10461
The chromogranins comprise a class of acidic
proteins that are secreted from large dense core vesicles and expressed
in neuronal and endocrine tissues. We describe here the molecular
characterization of NESP55 (neuroendocrine secretory protein of
Mr 55,000), a novel member of the
chromogranins. Several NESP55 cDNA clones were isolated from bovine
chromaffin cell libraries. The cDNA sequence of NESP55 totals 1499 nucleotides. All of the clones that were isolated contained in their
3-untranslated mRNA a sequence that was homologous to exon 2 of
the G-protein Gs
. The open reading frame encodes for an
acidic and hydrophilic protein of 241 amino acids with a predicted
molecular mass of 27,494 Da. An antiserum directed against the C
terminus of NESP55 labeled a band of Mr 55,000 with an acidic pI ranging from 4.4 to 5.2 in one- and two-dimensional immunoblots of secretory proteins from chromaffin granules. NESP55 is
localized within the cell to the large dense secretory vesicles and is
expressed, apart from the adrenal medulla, in the anterior and
posterior pituitary and various regions of the brain. For the
physiological function, one interesting factor has emerged. NESP55 is
proteolytically processed within the chromaffin granule to smaller
peptides that might be physiologically active. One tetrapeptide,
Leu-Ser-Ala-Leu (LSAL), present in the NESP55 sequence and flanked by
arginine residues suitable for cleavage by prohormone convertases, has
been identified recently as an endogenous antagonist of the
serotonergic 5-HT1B receptor subtype. Alterations in the serotonergic system are thought to play an important role in mental disorders, especially depression, and might be related to abnormal ethanol consumption. It is tempting to speculate that increased expression of NESP55 or its proteolytically derived peptide LSAL might
contribute to the pathophysiology of the serotonergic transmission.
Synaptic transmission is mediated via exocytotic release of
neurotransmitters from presynaptic terminals followed by binding of
these substances to specific receptors located on the postsynaptic plasma membrane. In addition to classical neurotransmitters like glutamate, -aminobutyric acid, acetylcholine, noradrenaline, dopamine, or serotonin, several peptides are secreted. These peptides are stored together with neurotransmitters in specialized vesicular containers, i.e. large dense core vesicles. In recent years
we characterized the content of large dense core vesicles to a great extent with the intention of identifying novel peptidergic substances involved in chemical signaling across neurons (1, 2). The majority of
the proteins found in the content of large dense core vesicles have an
acidic pI of 4-5 and were collectively named chromogranins (3). These
proteins typically consist of 200-700 amino acids with glutamic acid
as the most abundant individual amino acid. In their primary amino acid
sequence, multiple pairs of consecutive basic amino acid residues known
as cleavage sites for trypsin-like endoproteases are present. The
family is composed of several members: the structurally related
chromogranins A and B, secretogranins II and III, and vgf
(1, 2, 4-6). The chromogranins are widely expressed in neuronal and
endocrine tissues, which has made them very useful in identifying
elements of the diffuse neuroendocrine system. Both intracellular
(involvement in sorting of peptidergic components to the large dense
core vesicles (5)) as well as extracellular functions (representing
precursors of small biologically active neuropeptides like
pancreastatin, vasostatin, or secretoneurin) have been proposed for
chromogranins (1, 2, 4, 7).
We report here the molecular and cellular characterization of NESP55 (neuroendocrine secretory protein with a Mr of 55,000), a novel member of this class of proteins. NESP55 is expressed in the adrenal medulla, pituitary, and brain. It is an acidic protein that is processed within the secretory vesicles to smaller peptides. A tetrapeptide (present in the primary amino acid sequence flanked by Arg residues) suitable for endoproteolytic processing has been identified recently as an endogenous inhibitor of the serotonergic 5-HT1B receptor.
A gt11 expression
library was constructed according to the manufacturer's (Invitrogen)
protocol using mRNA isolated from cultured bovine chromaffin cells
with the guanidinium thiocyanate method. The library was screened with
an antibody directed against secretogranin II. One positive clone was
plaque-purified, subcloned into pGEM-5Zf(+), and then sequenced into
M13mp19 and M13mp19 (Pharmacia Biotech, Uppsala, Sweden) with the
dideoxy chain termination method. Based on the sequence, two
oligonucleotides were synthesized. These oligonucleotides were 5
end-labeled and used to screen two different Okayama-Berg libraries by
colony hybridization. Both libraries were prepared from cultured bovine
chromaffin cells and were a generous gift from Drs. Jeff Erickson, Lee
Eiden, and Anna Iacangelo (Laboratory of Cell Biology, National
Institute of Mental Health). Thirty positive clones of each
library, identified by autoradiography of filters, were rescreened
twice. The five clones containing the largest inserts were sequenced
with the Sequenase kit. Sequences were assembled and analyzed using the Genetics Computer Group software package of the University of Wisconsin, release 8.1. The signal peptide cleavage site was predicted by computer analysis (8).
The peptides
GAIPIRRH and GAIPI were synthesized by standard Fmoc
(N-(9-fluorenyl)methoxycarbonyl) chemistry and purified by
reversed phase high performance liquid chromatography. The peptide
structure was verified by amino acid analysis and mass spectrometry.
Both peptides were coupled via an additional N-terminal cysteine to
maleimide-activated keyhole limpet hemocyanin (Pierce), and these
conjugates were used to generate peptide antisera in chinchilla bastard
rabbits (Ivanovas, Kisslegg, Germany) with a standard immunization
protocol (9). YGAIPIRRH was iodinated by the chloramine-T oxidation
method and purified by reversed phase
HPLC.1 For RIA, peptide and iodinated
peptide were used as standard and trace, respectively. The antibody
(final dilution, 1:22,500) was incubated with standards and samples for
24 h at 4 °C in RIA buffer (150 mM
NaH2PO4, pH 7.4, 150 mM NaCl,
0.02% NaN3, 0.0006% phenol red, 0.1% bovine serum
albumin, 0.06% gelatin, 0.02% bacitracin). The trace (104
cpm) was added to standards and samples that were then incubated for an
additional 24 h at 4 °C. Bound and free fractions were separated by adding 1 ml of dextran-coated charcoal. After a 15-min incubation at 4 °C, samples were centrifuged at 3200 × g for 15 min at 4 °C, and the supernatant was counted in
a counter. The RIA was linear between 2.5 and 250 fmol of peptide.
The intra- and interassay coefficients of variation were 5.5 and 8.8%,
respectively. The RIA did not cross-react (tested up to 3000 fmol) with
the following other neuropeptides: calcitonin gene-related peptide, cholecystokinin 8, enkephalin, galanin, neurokinins A and B,
neuropeptide Y, neurotensin, secretoneurin, and vasoactive
intestinal polypeptide.
Chromaffin granules were isolated over 1.8 M sucrose as described (9). Soluble proteins were isolated by hypotonic lysis of chromaffin granules (9). One-dimensional electrophoresis was performed on 12% polyacrylamide slab gels under reducing conditions (10). Molecular weight standards were obtained from Novex (San Diego, CA).
Two-dimensional electrophoresis was performed as described (11) with minor modifications. In all solutions Nonidet P-40 was replaced by Triton X-100 (Serva, Heidelberg, Germany). Mercaptoethanol was added to the sample buffer at a concentration of 0.33%. Proteins were dissolved in sample buffer (11) and applied to the focusing system at the anodic (acidic) end of the gels. Gels (ampholines, pH 3-10, Bio-Rad) were electrophoresed at 15 °C for 10 min at a constant voltage of 100 V, then from 250 to 800 V at constant current, and finally at a constant voltage of 800 V for another 3 h.
After electrophoresis, proteins were transferred electrophoretically to
nitrocellulose sheets (Schleicher & Schuell) at 150 mA for 24 h as
described (9, 12). After the transfer, blots were then incubated for
3 h with antiserum (1:200). After several washes, nitrocellulose
sheets were treated with 125I-protein A (5 × 105 cpm/ml, DuPont NEN). Then the nitrocellulose sheets
were washed to removed unbound 125I-protein A, blotted dry,
and exposed to x-ray film (Cronex, DuPont) at 70 °C. Quantitation
was done by a Fuijx Bas 1000 phosphorimager analyzer.
Bovine adrenal medulla was homogenized in 0.3 M sucrose, and subcellular organelles were isolated by differential centrifugation. A large granule fraction was centrifuged (130,000 × g for 90 min at 4 °C) over a sucrose gradient ranging from 1.3 to 2.0 M sucrose as described in detail (9). After centrifugation, 10 different fractions were collected, diluted four times with 0.3 M sucrose (20%, v/v), and recentrifuged at 110,000 × g for 1 h at 4 °C. The pellets were resuspended in lysis buffer (5 mM Tris/sodium succinate, pH 5.9), and soluble proteins were analyzed by RIA and Western blotting.
High Performance Gel Filtration ChromatographyBovine tissues were obtained from the local slaughterhouse. After dissection of the organs, sample tissues were immediately chilled on ice. Extractions were performed by 10 s of sonication in distilled water (1 ml/100 mg, w/w) followed immediately by boiling for 10 min to inactivate any protease activity. Boiled samples were centrifuged at 14,000 × g for 20 min at 4 °C, and the supernatants were loaded on a Superose 12 HR 10/30 gel filtration column (Pharmacia) at a flow rate of 0.4 ml/min. As the column buffer, 75 mM sodium phosphate buffer (pH 7.4) containing 75 mM NaCl and 0.02% NaN3 was used. 1-min fractions were collected and analyzed by RIA.
A gt11
expression library prepared from cultured bovine chromaffin cells was
screened with a polyclonal antiserum against secretogranin II. One
immunoreactive clone (L311) was plaque-purified, subcloned into the
pGEM-5Zf(+) vector, and then sequenced with the dideoxy chain
termination method. It soon became clear that clone L311 was a false
positive since it was different from other clones obtained by screening
of an Okayama-Berg library with radioactively labeled degenerative
oligonucleotides synthesized according to peptides obtained by CNBr
cleavage of secretogranin II (13). In a study by Sigafoos et
al. (14) on peptides occurring in the secretory content of
chromaffin granules, five novel peptides were identified by Edman
sequencing among several already known chromogranin-derived peptides.
One of these putative novel peptides, GAIPIRRH, was encoded by clone
L311. Thus, additional clones were isolated from two different
Okayama-Berg libraries prepared from cultured bovine chromaffin cells
by homologous hybridization with two oligonucleotides that were
synthesized according to the sequence of clone L311. Sixty clones were
isolated, and the five containing the longest inserts were further
characterized. The overall structure of these clones is given in Fig.
1. All clones shared an identical sequence in the
5
-untranslated mRNA and the mRNA encoding for the open reading
frame. In the 3
-untranslated mRNA, variations were found among the
different clones. In all clones a sequence encoding for exon 2 of the
G-protein Gs
was present (Fig. 1). Further 3
downstream, clones L311, 81, and 82 shared a novel unique sequence,
whereas in clones 14, 31, and 51 a sequence corresponding to exons
3-13 of Gs
was found. Further differences between
clones resulted from deletions in the 3
-untranslated mRNA upstream
or downstream of the sequence homologous to exon 2 of Gs
(see Fig. 1).
Sequence Analysis and Protein Structure
The nucleotide
sequence of three clones (L311, 81, 51) was determined with the dideoxy
chain termination method. The sequence (Fig. 2) totals
1499 nucleotides. An in-frame stop codon is located at position 178, and the first ATG codon is present 358 nucleotides downstream of the 5
end. The open reading frame consists of 725 nucleotides and ends with a
stop codon TAA at position 1081-1083. A search of the
GenBankTM data base revealed no homology with other
published sequences for nucleotides 1-1230 and 1303-1499. Nucleotides
1231-1302 located in the 3
-untranslated mRNA (Fig. 1) were 100%
identical to the sequence of exon 2 of the bovine G-protein
Gs
(15, 16). A search of the Expressed Sequence Tag
division of GenBankTM revealed significant homologies with
sequences W49084[GenBank] and W16881[GenBank]. W49084[GenBank], a 403-base pair mouse brain cDNA clone generated by the Washington University-HHMI mouse
Expressed Sequence Tag project, is 79% identical to nucleotides
24-519 and most likely comprises a partial mouse homologue of NESP55
(Fig. 1). W16881[GenBank] comprises a 421-base pair human fetal lung mRNA
sequence derived from the Washington University-Merck Expressed Sequence Tag project. Nucleotides 41-118 of this clone are 91% homologous to nucleotides 119-194 of the 5
-untranslated mRNA of
NESP55 (see Fig. 1 for a schematic representation). Further downstream
(nucleotides 119-421), the sequence of this clone is identical to
exons 2 and 4-6 of human Gs
(16).
The open reading frame encodes for a hydrophilic and acidic protein of
241 amino acids with a predicted molecular mass of 27,494 Da. The
calculated pI of the protein is 4.8 and reflects the high abundance of
acidic amino acids (15.8% Glu, 5.4% Asp). The first 28 amino acids
have characteristic features of a signal peptide. Several sites in the
primary amino acid sequence meet the criteria as acceptor sites of
phosphorylation by cAMP-dependent protein kinase (Thr-150,
Thr-221), protein kinase C (Ser-49, Thr-156, Ser-183, Thr-221,
Ser-231), and casein kinase II (Ser-81, Thr-86, Ser-91, Ser-107,
Thr-109, Ser-111, Ser-113, Thr-119, Thr-123, Thr-131, Ser-154, Ser-177,
Thr-221). The primary amino acid sequence does not contain a potential
motif for N-glycosylation (Asn-X-Ser) but does
contain five pairs of basic amino acids (Fig. 3) that could be used for endoproteolytic cleavage. Two peptides found in the
sequence have been described previously, GAIPIRRH and LSAL. The
octapeptide GAIPIRRH (Fig. 3, 234-241, NESP55) has been
purified by HPLC from extracts of chromaffin granules, and its sequence was established by the Edman degradation (14). This peptide is located
at the C terminus of NESP55 and is preceded by a KR (Lys-Arg) cleavage
site at its N terminus. The tetrapeptide LSAL (Fig. 3,
159-162, NESP55) has recently been isolated from cow and
rat brain as an endogenous inhibitor of the 5-HT1B receptor (17). This peptide is flanked at its N and C termini by an Arg residue
suitable for endoproteolytic cleavage. Thus, NESP55 is a likely
candidate for the precursor of this novel endogenous serotonergic
receptor antagonist. The N-terminal amino acid sequence of bovine
NESP55 is highly conserved (Fig. 4, 91% similarity) to
the mouse brain homologue deduced from clone W49084[GenBank].
Molecular Forms of NESP55 in Bovine Adrenal Medulla
To
further characterize the NESP55 protein, antisera directed against
sequences corresponding to the C terminus of this protein were
generated. In one- and two-dimensional blots of the soluble content of
bovine chromaffin granules, an antiserum directed against both the
peptide GAIPIRRH (Fig. 5) and the shorter peptide GAIPI (not shown) labeled one prominent broad band corresponding to a protein
of Mr 55,000. In two-dimensional blots, a tilted
band with an acidic pI between 4.4 and 5.2 was labeled (Fig. 5). The appearance of this band in Western blots resembles that of the chromogranin A proteoglycan but is different from that of
N-glycosidically linked glycoproteins (dopamine
-hydroxylase and glycoproteins II and III), which is consistent with
the lack of a N-glycosylation signal in the primary amino
acid sequence. Thus, it is most likely that glycosaminoglycosylation
together with the high amounts of acidic residues (21%) lead to the
retardation of this protein on SDS-polyacrylamide gels that has already
been established for the other chromogranins (1, 2).
Putative proteolytic processing to smaller peptides was investigated by
separation of adrenal medullary extracts by gel filtration HPLC
followed by analysis of the individual fractions by RIA. In the adrenal
medulla, two main immunoreactive peaks were found (Fig.
6). One main peak corresponds to the uncleaved NESP55
eluting at a molecular weight of 55,000 (fractions 29-33), and a
second peak (fraction 46) elutes in the position corresponding to the C-terminal octapeptide GAIPIRRH (Fig. 6).
Subcellular Distribution of NESP55
The subcellular
distribution of NESP55 in the bovine adrenal medulla was investigated
after separation of subcellular elements by differential
centrifugation. In the large granule fraction, 74.5 ± 1.6% of
the total immunoreactivity was present, comparable to that of the
established secretory protein secretogranin II (64.8 ± 2.6%).
The microsomal fraction contained only 5.8 ± 0.7% of the total
immunoreactivity. The large granule fraction containing mainly
chromaffin granules, lysosomes, and mitochondria was further separated
by equilibrium centrifugation on sucrose gradients ranging from 1.3 to
2.0 M. NESP55 immunoreactivity as measured by RIA was found
in fractions 3-5 of the gradient corresponding to chromaffin granules
as shown by a distribution comparable to that of the established
granule marker secretogranin II (Fig. 7). An identical distribution of NESP55 in the gradient was found with immunoblots as
with the RIA (not shown). A lysosomal enzyme marker (cathepsin D)
peaked in fractions 6-8, whereas mitochondria and microsomes equilibrated in fraction 9 (not shown).
Tissue Distribution of NESP55
Several bovine tissues were analyzed for the presence of NESP55 with our specific RIA directed against the C terminus of this protein. High amounts of NESP55 immunoreactivity were detected in the adrenal medulla (5029 ± 338 fmol/mg (w/w, ±S.E., n = 4), the anterior (204 ± 35 fmol), and posterior (69 ± 25 fmol) pituitary. In the bovine brain, NESP55 immunoreactivity was detected in the hypothalamus (54 ± 7 fmol), hippocampus (27 ± 3 fmol), caudate nucleus (23 ± 0.6 fmol), thalamus (21 ± 3 fmol), and in significantly lower amounts in the cerebellum (5 ± 0.5 fmol).
We describe here the molecular characterization of the cDNA encoding for NESP55, a novel chromogranin-like protein. The primary amino acid sequence of NESP55 was established by sequencing six independent clones obtained from three different libraries. NESP55 comprises a 241-amino acid-long protein with a primary amino acid sequence unrelated to any other protein deposited previously in the GenBankTM data base. It resembles in several aspects, however, the class of proteins called chromogranins. (i) Like other chromogranins, i.e. chromogranins A and B and secretogranin II, NESP55 is characterized by a high abundance of acidic amino acids (21%) in its primary amino acid sequence. (ii) NESP55 is heat-stable (results not shown). (iii) It is localized to the soluble content of the large dense core vesicles of the adrenal medulla, the chromaffin granules. (iv) Like other chromogranins, NESP55 is synthesized not only in the adrenal medulla, but also it is expressed in certain regions of the brain and other endocrine tissues (anterior and posterior pituitary). Preliminary immunohistochemical experiments demonstrate a strong reactivity in the raphe and locus coeruleus of the rat brain, but in general a much lower abundance of NESP55 is found in the rat brain when compared with chromogranins A or B (not shown). (v) In the primary amino acid sequence, several pairs of basic amino acids are found that are used as cleavage sites by endopeptidases.
After translation, the protein is modified post-translationally as judged by the difference in molecular weight calculated from the primary sequence and that obtained from immunoblots on SDS-polyacrylamide gels. The continuous tilted pattern occurring in the two-dimensional gel is reminiscent of the proteoglycan form of chromogranin A (18), indicating a putative glycosaminoglycation of NESP55. Further work is required on this point. The N terminus of NESP55 was deduced from the sequence assuming that the first in-frame methionine located 180 nucleotides downstream from the bordering stop codon is used. It is corroborated further by comparison with the partial mouse sequence obtained from the translation of clone W49084[GenBank]. The first in-frame methionine is located in this cDNA only 15 nucleotides downstream of the bordering stop codon, and the amino acid sequence starting at this methionine, but not preceding it, is highly conserved between both cow and mouse.
Functional Aspects of NESP55For the physiological function of NESP55, at least one interesting factor has already emerged. In principle, both an intracellular function within the secretory vesicle or an extracellular one after secretion into circulation and the synaptic cleft seem conceivable. The amino acid sequence of NESP55 contains several pairs of basic amino acids that can be used for endoproteolytic cleavage by the subtilisin-like prohormone convertases. In fact, our data demonstrate a significant amount of proteolytic processing of NESP55 within the chromaffin granule. Thus, proteolytic processing of NESP55 might yield smaller physiologically active peptides. GAIPIRRH, one of these putative peptides, is located at the C terminus of NESP55. It was isolated in 1993 in a random search for novel peptides secreted from chromaffin granules (14) and has not yet been assigned with a specific function.
The tetrapeptide LSAL is located in the center of the NESP55 sequence. This peptide has been identified as an endogenous peptidergic inhibitor of the serotonergic 1B receptor subtype in a recent study (17). LSAL is flanked on both sides by Arg residues; the C-terminal Arg is followed by another Arg in position P2* (Arg-Leu-Arg) that is compatible with a type IV precursor protein cleavage site recognized by the subtilisin/kex-like proprotein convertases (19). Thus, NESP55 is most likely the protein precursor of this novel serotonergic receptor antagonist. The 5-HT1B receptor is localized mainly presynaptically and controls serotonin release from serotonergic nerve terminals (20). 5-HT1B receptors are also found on non-serotonergic terminals where they inhibit the release of, for example, acetylcholine from the hippocampus or noradrenaline from peripheral tissues (20). Alterations in the serotonergic system are thought to play an important role in mental disorders, especially in depression (21, 22). Specific 5-HT1B agonistic drugs have been used clinically with aggressive psychiatric patients (23). 5-HT1B receptor knock-out mice drink twice as much ethanol compared with wild-type mice (24) and show an altered release of serotonin from in vitro slice preparations (25). The characterization of the precursor of the novel peptidergic 5-HT1B receptor antagonist LSAL now facilitates co-localization experiments with the 5-HT1B receptor in vivo. The molecular nature of the perturbations of the serotonergic transmission leading to mental disorders or increased ethanol consumption are as yet still unknown. It is tempting to speculate that an increased expression of NESP55 or its proteolytically derived peptide LSAL might contribute to the pathophysiology of the serotonergic transmission.
Presence of GsIn the 3-untranslated mRNA of all of the NESP55
clones isolated, a sequence corresponding 100% to exon 2 of the bovine
Gs
is present. Since our clones have been isolated from
three independent libraries, a cloning artifact can be excluded. For
Gs
, several different mRNAs exist. Originally, four
forms of Gs
generated by alternative splicing of exons 3 and 4 were described (26, 27). In addition, mRNAs resulting from
the use of alternative promoters and alternative first exons were
reported (28-30). There is no in-frame AUG located between the stop
codon terminating the open reading frame of NESP55 and the exon 2 sequence, which is followed by a stop codon 16 nucleotides further
downstream. Thus, no protein can be translated from the mRNA
species NESP55 a and b (Fig. 1). From mRNA
species c and d (Fig. 1), in addition to NESP55,
a truncated form of a Gs
could be transcribed by usage of an AUG in exon 2 as described previously (29). The significance of
exon 2 mRNA in the 3
-untranslated mRNA of NESP55 is unclear. Most likely, the NESP55 gene locus is nested in close proximity to the
various alternative first exons of the Gs
gene. Further studies are warranted to investigate whether both NESP55 and
Gs
are regulated in a coordinated way due to usage of
common upstream promoter elements or whether they share a mutual
function.
It is worth emphasizing that probes generated according to the sequence
of exon 2 will not specifically hybridize to Gs mRNA but will also detect NESP55 mRNA. Furthermore, probes corresponding to exons 3-13 might as well detect both messages due to the presence of these exons in cDNA clones NESP55 c and d
(Fig. 1). Thus, the detection of Gs
protein by
immunological methods is necessary to establish the expression of this
G-protein in the various tissues.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U77614[GenBank].
We wish to thank Dr. R. Schneider for fruitful discussions.