Induction and variants of neuronal nitric oxide synthase type I during synaptogenesis. PATRICIA. OGILVIE, KARL SCHILLING,*. MELVIN L BILLINGSLEY,t. AND.
RESEARCH
Induction synthase
and variants of neuronal nitric type I during synaptogenesis
PATRICIA OGILVIE, KARL SCHILLING,* HARALD H. H. W. SCHMIDT’
MELVIN
L BILLINGSLEY,t
COMMUNICATION
oxide AND
Medizinische UniversitAtsklinik, Klinische Biochemie und Pathobiochenue, D-97078 Wurzburg, Germany; *.t&hteilung Anatomie und Zeilbiologie der UniversitAt Ulm, D-89081 Ulm, Germany; and tDepartment of Pharmacology and Center for Cell and Molecular Biology, Pennsylvania Hershey, Pennsylvania 17033, USA
In the adult central nervous system, nitric oxide (NO) is formed from L-arginine by the so-called constitutive or type I NO synthase (NOS1155). However, expression of NOS-1155 immunoreactivity and activity was low or not detectable in developing mouse and rat brain. NOS-1155 was sharply induced coincident with the onset of synaptogenesis in specific brain regions. This was followed by a second phase in which total NOS-1155 expression decreased both in specific cell populations and in the total synaptosomal subcellular fraction. Furthermore, two putative variants of NOS-I were transiently observed: an NOS-I-inununoreactive protein with increased electrophoretic mobility (NOS-1144) and a transient hypersensitivity of NOS-liss to the competitive suhstrate inhihitor NWnitroLarginine. It is concluded that NOS-I expression is not constitutive but locally induced. In the central nervous system, this regionally specific, biphasic pattern of postnatal NOSI induction is consistent with a role for NO in synaptogenesis and synaptic plasticity.-Ogilvie, P., Schilling, K., Billingsley, M. L., Schmidt, H. H. H. W. Induction and variants of neuronal nitric oxide synthase type I during synaptogenesis. FASEB J. 9,799-806(1995) ABSTRACT
Key Worch:
expression
.
development
.
nilro-L-arginine
#{149} synap-
togenesia
Nitric oxide (NO)2 formation from L-arginine is used as a means for intercellular communication by a variety of organisms and has wide implications for biology and medicine (1, 2). Being a gas and small in molecular mass, highly diffusible, and short-lived, NO represents the first of a novel class of mammalian inter- and intracellular messenger molecules. Its biological actions range from signal transduction to cell killing and neurodegeneration. The molecular mechanisms that lead to its double-edged biological role are not fully understood. Low biophase concentrations of NO (2 M) and discontinuous, Ca2-regu1ated formation and release seem to mainly subserve messenger functions with the heme protein soluble guanylyl cyclase as a major molecular target (3, 4). Within the central nervous
0892-6638/95/0009-0799/$01
.50. © FASEB
State University College of Medicine,
system, NO has been suggested to be involved in synaptic plasticity of the mature brain, such as long-term potentiation (5-7) and long-term depression (8), as well as in synaptogenesis of the developing brain (9, 10). To make use of NO as a neuronal messenger molecule and avoid its neurotoxic potential, tight regulation of its biosynthesis is crucial. At least three distinct isozymes and corresponding genes of NO synthase (NOS; L-arg nine, NADPH:oxygen oxidoreductase [nitric oxide-forming], E.C. 1.14.13.39) exist (1). In neurons, the most abundant isoform of NOS is type I (NOS-I or ncNOS, also termed nNOS in ref 11), which has a cDNA-predicted molecular mass of 160 kDa (12, 13) and migrates on SDS-PAGE with an apparent molecular mass of 155 kDa (NOS-liss; ref 14). The key regulatory molecule of electron flow within all NOS isoforms (15, 16) is the Ca2”binding protein calmodulin (14, 17, 18), which binds to different NOS isoforms in either a Ca2”-dependent (NOS-I and III) or Ca2”-independent (NOS-Il) manner. NOSs are further differentiated by their mode of expression (19), which is termed either constitutive (NOS-I and III) or inducible (NOS-JI). Several other regulator mechanisms have been suggested for NOS-I based on in vitro findings (20, 21); however, none has been elucidated at the molecular level and shown to be of physiological relevance. Thus, [Cal+]i remains the only established in vivo regulator of neuronal NO formation. In light of the suggested role for NO in synaptogenesis (9, 10), we sought to investigate the developmental regulation of NOS-1155, which may also be of relevance for nonneuronal functions of NOS-I (2). We here demonstrate the physiologic postnatal induction of NOS-1155 during rodent brain development with maxima and a temporal and spatial pattern that correlates with synaptogensis. More-
1To whom correspondence should be addressed, at: Medizinische Universitatsklinik, Klinische Biochemie und Pathobiochemie, Versbacher Sir. 5, D-97078 Wurzburg/Germany. 2+ 2Abbreviations: [Ca2+]i, intracellular free Ca concentration; E, embryonal day; NO, nitric oxide; NO2Arg, )V-nitro-L-arginine; NOS, NO synthase;
P. postnatal
day.
799
RESEARCH COMMUNICATION over, two putative transient contribute to the regulation
variants of NOS-I may further of neuronal NO formation.
for 20 mm. After extensive washing with PBS, sections were for 30 mm in PBS containing 2% normal goat serum. Subsections were reacted overnight at room temperature with anti-NOS-P(batch 6763-5, diluted 1:300 - 1:1000; ref 14). The im1% H202 incubated sequently,
munoreaction nique (24).
was visualized
METHODS Materials
Subcellular
fractionation
The materials used in this study were obtained from the following sources: FAD, FMN, GSH, and DNA molecular weight marker X were from Boehringer (Mannheim, Germany); SuperScript preamplification system from Life Technologies (Eggenstein, Germany); L-[2,3,4,5-3H]arginine.HCI (specific activity 2.85 TBq/mrnol) and enhanced chemiluminescence (ECL) Western blot detection kit, Amershain (Braunschweig, Germany); prestained molecular mass standards and Dowex AG 50 WX8 (100-200 mesh), Biorad Laboratories (M#{252}nchen, Germany); (oR)5,6,7,8-tetrahydro-L-biopterin, from Dr. B. Schircks Laboratories (Jona, Switzerland); horseradish peroxidase coupled to goat polyclonal antibody directed against rabbit immunglobulin (anti-rabbit Ig), phosphate-buffered saline (PBS), phosphodiesterase 3’,S’-cyclic nucleotide activator (calmodulin) and NO2Arg, Sigma Chemicals (Deisenhofen, Germany); and NADPI-1, Pharma Waldhof (D#{252}sseldorf,Germany). All other chemicals, reagents, and solvents were of analytical grade and provided either by Merck AC (Darmstadt, Germany) or by Sigma Chemicals (Deisenhofen, Germany). Polyclonal, monospecific antiserum to NOS-liss from adult rat cerebellum (22) was prepared as described (14). Monoclonal antibody to NOS-Ill was a kind gift from Dr. J. S. Pollock (Abbott Laboratories, Abbott Park, Ill.; ref 23). Water was deionized to 18
The collected brain tissues were thawed in 3 ml ice-cold homogenization buffer (20 mM Hepes, 0.25 M sucrose, 0.5 mM EDTA, 10 JiM leupeptin, I JIM pepstatin A, 0.2 mM phenylmethyl sulfonyl fluoride, 7 mM glutathione, pH 7.5) per 1 g tissue wet weight and homogenized on ice using a Potter-Elvehjem Teflon glass tissue homogenizer (20 strokes; Wheaton; Renner, Darmstadt, Germany). All subsequent procedures were performed at 4#{176}C. Homogenates were centrifuged at 1000 X g for 15 mm; the nuclear pellet was discarded and the resulting supernatant was centrifuged a second time at 33,000 X g for 30 mm. The resulting supernatant (crude cytosolic fraction) was rapidly frozen in liquid nitrogen; the pellet (crude synaptosomal fraction) was resuspended in homogenization buffer and also rapidly frozen in liquid nitrogen.
M#{252}.cm (Mili-Q, Milipore,
J.tM
Animal
Eschborn).
breeding
C57B116 mice of known age and either sex were obtained from the central animal facility of the University of VIm, where they were maintained under standard conditions on a 12 h light and dark cycle and fed ad libitum. Embryonic tissue was obtained from timed-pregnant animals (day of insemination equals day E0). Sprague-Dawley rats were housed in groups of 6, kept on a 12 h light and dark cycle, and fed ad libitum. Female breeders were housed in single cages at the time of mating and time of estrus was determined by vaginal lavage and microscopy. Male rats were introduced for 12 h after establishment of proestrus, and presence of sperm or sperm plug was verified; this day was then taken as gestational day 0 (E0). On E17, female breeders were transferred to a delivery room and placed in solid-bottom cages. All breeders were monitored twice daily. Rat pups were weaned on postnatal day 20 (P20). Animals undergoing cesarean section for purposes of isolating fetal brain were initially anesthetized with sodium pentobarbital (50 mg/kg; i.p.). After surgery, animals were killed using sodium pentobarbital. Pups (P0-P60) were anesthetized with sodium pentobarbital (30 mglkg; i.p.).
Tissue
isolation
Tissues were obtained from two discrete brain regions (mouse cerebellum and rat neocortex), briefly washed in phosphate-buffered saline, pH 7.4 (PBS), and then either itnmediately used for immunocytochemistry, subcellular fractionation, RNA isolation, or rapidly frozen in liquid nitrogen and stored for up to 2 wk at -80#{176}C until use.
Immunocytochemistry Cerebella of 8-day-old and adult (P60) C57BI/J mice were used for immunocytochemical localization of NOS-l. Animals were anesthetized with 4% chloral hydrate and perfusion-fixed with phosphate buffer (100 mM, pH 7.2) containing 4% formaldehyde. Brains were postfixed in the same fixative for 3 h and vibratome sections were cut in the horizontal plane at 40 J.tm. Sections were permeabilized by incubation in PBS containing 0.5% Triton X-100 for 20 mm; endogenous peroxidase activity was quenched by incubation in PBS containing 10% methanol and
800
Vol. 9
june
1995
Assay
of NOS
using an avidin-biotin-peroxidase
tech-
activity
Specific NOS activity was assayed according to Bredt and Snyder (25) by conversion of L-arginine into L-citrulline with slight modifications. Aliquots of the crude cytosolic and synaptosomal fractions were incubated in assay buffer (50 mM triethanolamine.HCI, 0.5 IM calmodulin, 226 IIM CaCl2, 477 LtM MgCI2, 0.5 mM EDTA, 5 JiM FAD, 5 JIM FMN, 5
H4biopterin, up to 50 J.tM L-arginine, 0.22 nM L-[2,3,4,5-3H]arginine, 1 mM NADPH, 7 mM GSH, pH 7.0) for 30 mm at 37#{176}C in a total volume of 100 J.tl. To some reactions, either 2 mM EGTA, to test for Ca2”-independent NOS activity, or different concentrations of the NOS inhibitor NO2Arg (26) were added. The reactions were stopped by addition of 900 Ltl of slop buffer (20 mM sodium acetate, 2 mM EDTA, pH 5.5) and the total volume of 1 ml was then applied to a 0.8 ml cation exchange column Dowex AG 50 WX-8, Na” form) equilibrated with 2 ml of stop buffer. H-labeled citrulline was eluted twice with 1 ml of water and the radioactivity in the combined flow-through and eluate (3 ml) was measured by liquid scintillation counting. Each result is shown as the mean of n = 3 experiments ± the standard error of mean (vertical bars). The limit of detection was 1.49 pmol L-citrulline.
Protein
determination
Protein concentrations ford (27) using bovine
SDS-PAGE,
were determined in triplicate serum albumin as a standard.
Western
blotting,
according
to Brad-
and densitometry
Synaptosomal and cytosolic fractions of fetal brains and postnatal cerebral neocortex (100 JIg protein/lane) were separated by 1-dimensional SDS.-PAGE under reducing conditions using 7.5% gels (28) and then electroblotted to nitrocellulose membranes using a semidry transfer method (29). Protein blots were blocked for 12 h at 4#{176}C in Tris-buffered saline (TBS; 20 mM Tris.HC1, pH 7.5, 200 mM NaCl) containing 3 g/100 ml chicken egg albumin. NOS-I immunoreactive proteins were detected using anti-NOS-I (batch 6763-5, ref 14; 1:3000 in TBS, overnight, 4#{176}C,), followed by horseradish peroxidase-coupled goat polyclonal antibody directed against rabbit immunoglobulin (Sigma, Deisenhofen, Germany; 1:5000 in TBS, 1 h, room temperature). Between each step, blots were washed were
five times visualized
in TBS. They were developed using
an
enhanced
and immune complexes procedure as
chemiluminescence
(14) with a limit of detection for 125 ng NOS-I. Blots were further subjected to densitometric scanning using a Molecular Dynamics densitometer or flatbed laser scanner coupled to digital analysis software (PDI Inc., Huntington Station, N. Y.; NIH Image). The densities of the 144 and 155 kDa-NOS-l immunoreactive protein bands were quantitated as peak area over background, standardized to the peak area observed at described
The FASEB Journal
OGILVIE
Ef AL.
RESEARCH P60 (% of P60), experiments.
and
reported
as mean
± SEM (vertical
bars)
of n
=
3
COMMUNICATION
A cytosol
Isolation
of total RNA and RT-PCR
analysis
Total RNA was isolated from rat fetal and postnatal cerebral neocortex at specific developmental times using the guanidine isothiocyanate extraction method and CsCI density centnfugation protocol as outlined by Chirgwin et al. (30). Aliquots from each sample (2-7 tg RNA/5 Jil) were analyzed on ethidium bromide-stained agarose (1%)/formaldehyde gels; 28S and 18S rRNA bands were clearly visible. For RT-PCR analysis, first-strand cDNA was synthesized from total RNA using random hexamen and the SuperScript preamplification system (Life Technologies, Eggenstein, Germany). The RNase H-treated rat target DNA (equivalent to the reported mouse nNOS-1 and nNOS-2 RNAs; ref 11) was amplified by PCR from base 1315 to 2007 using two synthetic, rat-specific oligonucleotide primers (S3 and rAS3) based on the reported nucleotide sequence of rat NOS-I (12). Oligonucleotide primers were S3 (5’-TFCAACTACATCTgTAACCA-3’ from base 1315 to 1334), which is identical to the S3 sense primer used for mouse n-NOS eDNA (11), and rAS3 (5’-TCTGCAgCggTA1TCATFCTC-3’ from base 1987 to 2007), a modification of the AS3 antisense primer for mouse nNOS-1 incorporating one base exchange in the respective rat NOS-I sequence (12). Dcnaturation, annealing, and elongation temperatures for the PCR reaction were 94, 58, and 72#{176}C for 1, 2, and 3 mm and 30 cycles, respectively (11). To further analyze the entire open reading frame of NOS-!, the NH2-terminal primer rSl (5’-ggg’ITCAgCAgATCCA-3’), in combination with rAS3, and rAS1 (5’-CTCATCTgCgTC1TIT-3’), in combination with 53, were also designed; conditions were as described previously except for the annealing temperature, which was 54#{176}C. The presence and size of the obtained PCR products were analyzed on ethidium bromide-stained agarose (1-3%) gels.
155
-.
I
.
..
synaptosomes
E15 E17 E19 0
B
P2
P5
P7
P9
P12 P15 P20
150
f
0
CD
0 ‘
P60
100
a
. 50
a0
A
age, days RESULTS Induction
C25
of NOS-1155
After subcellular fractionation of late postnatal and adult brain tissue homogenates, NOS-I immunoreactivity migrated with an electrophoretic mobility equivalent to 155 kDa (NOS-liss; Fig. 1). NOS activity indicative of the type I isoform, i.e. Ca2-dependent and soluble (22), was recovered from both the crude cytosolic and synaptosomal fractions (Fig. 1 and Fig. 2). Ca2”-independent NOS activity was not detected. The ratio between cytosolic and synaptosomal NOS-I activity was between 6.2:1 (P20) and 104:1 (P60) in rat neocortex, and it was 5.4:1 (P60) in mouse cerebellum. When assayed in the presence of saturating concentrations of cofactors and calmodulin, NOS activity in both brain regions was dependent on free Ca2” levels. However, during embryonic stages of cortical as well as cerebellar development, neither NOS-1155 immunoreactivity nor NOS activity was expressed (Fig. 1 and Fig. 2). Rather, NOS-1155 expression commenced sharply after birth (P0) both in the crude cytosolic and synaptosomal fractions. The induction of NOS-1155 protein expression and NOS activity was particularly pronounced in the synaptosomal fraction, thus lowering the ratios of NOS activity in the crude cytosolic and synaptosomal fractions to 0.56:1 in rat neocortex and to 3.6:1 in mouse cerebellum. In the cerebellum, where specific cell types can be reliably identified, NOS-1155 expression was morphologically confined to the same cell types, i.e., granule cells and stel-
INDUCTION
AND
VARIANTS
OF NO
SYNTHASE
20
E 15
5
10
Cl) 0
Z
0 E15 E17 E19
0
P2 P5 P7
P9 P12 P15 P20 P60
developmentalday Figure
1. Induction and
variants
of NOS-I in synaptosomal
and cytosolic
fractions of developing rat neocortex. A) Representative Western blot; arrows indicate apparent molecular mass of NOS-l immunoreactive protein bands (144, NOS-I1; 155, N0S-I1). B) Densitometric quantification of NOS-l immunoreactive protein bands in cytosol (0, NOS-I1) and synaptosomes (A, NOS-I1; NOS-I5.5) expressed in % of the respective NOS-I1 value at P60. C) NOS activity in cytosol (open bars) and synaptosomes (closed bars).
#{149},
801
RESEARCH COMMUNICATION
C
E C) E
0
E 0.
E16 P0 P2 P5 P7 P9 P12 P15 P20 P60 developmental day Figure
2. Induction
of NOS-I
in cytosolic
(open
bars)
and
synaptosomal
(closed bars) fractions of developing mouse cerebellum.
late/basket
(Fig.
3).
neurons, In granule
throughout postnatal development cells, NOS-I immunoreactivity was
after the cells had reached their final position in the (internal) granule cell layer. Using a monoclonal antibody to endothelial NOS-lIl (23), we also detected a NOSIII immunoreactive protein band in Western blots of the crude particulate subcellular fraction of both neocortex and cerebellum at various time points of pre- and postnatal development (data not shown). However, in both cerebellum and neocortex, particulate NOS activity (indicative of NOS-Ill; ref 31) either was not detectable or amounted to less than 1% of the soluble NOS activity in cytosol and synaptosomes (indicative of NOS-I; refs 17, 22, 32). Thus, NOS-Ill represented only a negligible part of total brain detected
NOS activity (see also ref 23) and was not studied here further. The time course of NOS-I induction during development was biphasic, irrespective of whether Ca2+dependent soluble NOS activity or NOS-1155 immunoreactive protein was used to quantify levels of NOS-I. Expression of NOS was maximal with respect to total NOS activity at P9 in rat neocortex and P20 in mouse cerebellum; peak synaptosomal NOS expression was between P2 and P9 in rat neocortex and between P12 and P20 in mouse cerebellum (see Fig. 1 and Fig. 2). In a second phase, which began between P12 and P15 in rat neocortex and between P20 and P60 in mouse cerebellum, specific NOS activity decreased again-predominately in the synaptosomal fractions-whereas cytosolic activity values remained constant until P60 in rat neocottex. This time course of NOS expression corresponded well to the NOS1155 immunoreactivity as detected by Western blot analysis. Immunocytochemical staining of cerebella was used to verify the cellular localization of NOS-! during development. NOS-I expression was restricted to granule cells and basket/stellate neurons (Fig. 3; see also to ref 33). Endotheha of blood vessels that cover the surface of the cortex (upper margin of both panels in Fig. 3) and express NOS-IlI were immunonegative, demonstrating again the specificity of the antiserum for NOS-l. Early in cerebellar development (P8), NOS-I immunoreactivity was seen in granule cells in the internal granule cell layer and in the nascent molecular layer containing the axons (parallel fibers) of postmigratory granule cells. Granule cell precursors and premigratory granule cells in the external granule cell layer and the Purkinje cell layer were immunonegative. In the adult cerebellar cortex (P60), the molecular layer and the granule cell layer remained intensely immunostained. However, during the second postnatal week and continuing into adulthood distinct differences in the staining intensity of granule neurons became apparent (i.e., subsets of granule cells arranged in discrete clusters) and maintained a high level of NOS-I immunoreactivity, whereas other granule cells, again grouped together in clusters, no longer stained for NOS-I or
--,.I.
.
-‘
-
-:
____
Figure 3. Plasticity of NOS-l immunoreaction in the developing mouse cerebellar cortex. A) PB; B) P60. Bar = 100 JIm. Abbreviations: e, external granule cell layer; m, molecular layer; p, Purkinje cell layer; i, internal granule cell layer; g, granule cell layer; arrows, NOS-I-negative Purkinje cells enmeshed by a dense network of NOS-I immunopositive fibers.
802
Vol.
9
June 1995
The
FASEB Journal
OGILVIE
ET AL.
RESEARCH stained only very faintly (compare the patchy staining pattern of the granule cell layer in Fig. 38 to the rather homogenous staining of this layer in Fig. 3A; for further details, see ref 33). NOS.I
100
variants
80
In embryonal rat neocortex, a NOS-I-immunoreactive protein was expressed with an electrophoretic mobility (SDSPAGE) equivalant to an apparent molecular mass of 144 kDa (NOS-Iiw; see Fig. 1). On the other hand, NOS-1155 immunoreactivity or any other anti-NOS-I cross-reactive protein, in particular low molecular weight proteolytic fragments, were not detected. Tissue homogenates of the same age that clearly contained a NOS-Iiw immunostaining band, e.g., at E19, were devoid of NOS activity (see Fig. 1). We investigated whether the appearance of the NOS-Ii-immunoreactive protein correlated with the transcription of alternatively spliced NOS-I mRNA similar to mouse nNOS2 mRNA lacking a 315 bp in frame deletion. Rat neocortex total RNA was isolated at different developmental stages and analyzed by RT-PCR. Primers specific for rat NOS-I were designed to discriminate by size between nNOS-1 and the respective in frame-deleted nNOS-2 transcripts (11). However, the same amplification products (about 700 bp) were detected at all time points in embryonal and postnatal neocortex (Fig. 4). This value corresponded well to the theoretical amplification product of 692 bp calculated from the full-length rat NOS-I eDNA sequence (base 1315 to 2007). Thus, at all developmental stages nNOS-2 mRNA was below the detection limits of the RT-PCR method. Using additional primers spanning the entire NOS-I reading frame, all amplification products at all time points corre-
2036 1636
COMMUNICATION
0 0 1-
0
60
40
>
t U)
20
0
z 0
-9
-8
-7
-6
-5
-4
to NO2Arg
at
(NO2Arg], log M Figure
5.
Increased
sensilivity
of mouse
cerebellar
NOS-I
inhibition of cytosolic NOS-l activity at P7 (;co) and P20 (;cc) by NO2Arg; vertical bars indicate set (n = 3); **D
