(Received for publication, August 12, 1996, and in revised form, January 8, 1997)
From the Renal Division and Department of Medicine,
St. Michael's Hospital, University of Toronto, Toronto, Ontario M5S
1A8, Canada, the ¶ Department of Medicine, State University of New
York, Stony Brook, New York 11794-8152, and the ** Department of
Medicine, Division of Hematology, Emory University,
Atlanta, Georgia 30322
mRNA diversity represents a major theme of
neuronal nitric-oxide synthase (nNOS) gene expression in somatic
cells/tissues. Given that gonads often express unique and biologically
informative variants of complex genes, we determined whether unique
variants of nNOS are expressed in the testis. Analysis of cDNA
clones isolated from human testis identified a novel, testis-specific
nNOS (TnNOS) mRNA transcript. A predicted 3294-base pair open
reading frame encodes an NH2-terminal truncated
protein of 1098 amino acids. Measurement of calcium-activated
L-[14C]citrulline formation and nitric oxide
release in CHO-K1 cells stably transfected with the TnNOS cDNA
indicates that this protein is a calcium-dependent
nitric-oxide synthase with catalytic activity comparable to that of
full-length nNOS. TnNOS transcripts exhibit novel 5 mRNA sequences
encoded by two unique exons spliced to exon 4 of the full-length nNOS.
Characterization of the genomic structure indicates that exonic regions
used by the novel TnNOS are expressed from intron 3 of the
NOS1 gene. Although lacking canonical TATA and CAAT boxes,
the 5
-flanking region of the TnNOS exon 1 contains multiple putative
cis-regulatory elements including those implicated in
testis-specific gene expression. The downstream promoter of the human
nNOS gene, which directs testis-specific expression of a novel
NH2-terminal truncated nitric-oxide synthase, represents
the first reported example in the NOS gene family of transcriptional
diversity producing a variant NOS protein.
The nitric-oxide synthases (NOS)1
constitute a family with at least three distinct human isoforms:
neuronal (nNOS), inducible, and endothelial constitutive. These
apoenzymes are cytochrome P450-like hemeproteins that require
tetrahydrobiopterin, calmodulin, FMN, and FAD as cofactors. NOS enzymes
function as homodimers and catalyze the NADPH-dependent
five-electron oxidation of L-arginine to form
L-citrulline and NO (1-4). The three isoforms are
structurally similar, sharing in their primary amino acid sequences
binding sites for heme, calmodulin, FMN, FAD, and NADPH (1, 5, 6). However, nNOS differs from the other two isoforms in that it exhibits a
unique NH2-terminal extension containing a PDZ/GLGF motif
of 100 amino acids. PDZ/GLGF motifs participate in protein-protein interactions in a variety of proteins (7, 8). In fast twitch skeletal
muscle fibers, for example, nNOS is engaged to the sarcolemma dystrophin complex via this PDZ/GLGF motif (9). In a variety of
neurons, subcellular localization of nNOS in synapses involves PDZ/GLGF
anchoring to postsynaptic density 95 (PSD-95) and the related protein
PSD-93 (10). Furthermore, recent work has demonstrated that nNOS also
contains within the NH2-terminal region a unique binding
domain for PIN (protein inhibitor of nNOS) (11). Binding of PIN to this
domain results in significant inhibition of nNOS enzymatic
activity.
The biological roles for nNOS continue to expand. The enzyme is implicated in the regulation of neuronal cell biology and neurotransmission (1, 12, 13), as a major nonadrenergic-noncholinergic transmitter in enteric nerves (14, 15), in the neuroendocrine biology of the hypothalamus and pituitary (16-18), in modifying skeletal muscle contractile force and development (9, 19, 20), in the control of total body sodium content and body fluid homeostasis via its expression in the macula densa and distal nephron of the kidney (21-23), and in male sexual function (24), among other roles. Targeted disruption of the nNOS gene in mice by homologous recombination was recently reported (25). Studies using this genetic model have indicated involvement of the enzyme in neurotoxicity secondary to ischemia reperfusion injury in the central nervous system (26), in behavior (27), and in the development and normal function of the gastric pyloric sphincter (25). Significantly, a recent study in 27 families with inherited infantile pyloric stenosis has identified nNOS as a susceptibility gene in this human disorder (28).
The structural organization of the human nNOS gene has been described
(29). NOS1 is a complex locus consisting of 29 exons and 28 introns, localized to 12q24.2, spanning a region greater than 160 kb as
a single copy in the haploid human genome. The full-length open reading
frame of 4302 bp encodes a protein of 1434 amino acids. Translation
initiation and termination sites are located in exons 2 and 29, respectively. Structural and allelic mRNA diversity represents a
major theme in the regulated expression of this gene in somatic tissues
and cells, including cassette exon deletions (29), multiple examples of
exon 1 (30), and alternate polyadenylation signal usage (29). The
functional relevance of the diversity of mRNA transcripts remains
to be elucidated. Similar mRNA diversity has been observed in other
species. An alternative splicing event between exons 1 and 3 was
detected in wild type mice and mice in which exon 2 was disrupted by
homologous recombination, resulting in an NH2-terminal
truncated protein, nNOS (10). Furthermore, a cassette exon insertion
between exons 16 and 17 was shown to exist in the skeletal muscle of
rat (31).
Cardiovascular and neuroendocrine genes commonly exhibit expression
patterns in cells of germ-line lineage. In many cases, unique mRNA
transcripts are expressed. These variants are often informative with
respect to enzyme structure and function. Examples in which a novel
protein is expressed that differs from its somatic counterpart include
angiotensin-converting enzyme (32, 33), Ca2+/calmodulin-dependent protein kinase (34),
A-type lamin (35), glutamic acid decarboxylase (36),
proopiomelanocortin (37), and type platelet-derived growth factor
receptor (38), among others. Alternate promoter usage commonly
underlies this mRNA and protein diversity. In each of these cases,
characterization of the protein expressed in germ cells has led to a
better understanding of the biology and biochemistry of the protein
involved and provided important new insight into the regulation of the
gene. Furthermore, where 5
mRNA diversity exists in the 5
-UTR of
mRNA sequences of germ-cell transcripts, profound effects on
translational efficiency have been described, as in proenkephalin (39),
copper-zinc superoxide dismutase (40), GATA-1 (41), cytochrome
c (42),
-glutamyl transpeptidase (43), and
c-mos (44).
In this study, we report the cloning from human testis of a novel nNOS
cDNA (TnNOS) resulting from transcriptional activation of a
downstream promoter. This novel transcript encodes an
NH2-terminal truncated protein that is an analogue of the
mouse nNOS. The human nNOS
possesses NOS enzymatic activity
comparable to that of the full-length nNOS.
cDNA Cloning and Sequencing
Library ScreeningAn oligo(dT) + random primed gt10
human testis 5
stretch plus cDNA library and an oligo(dT) primed
gt11 human testis cDNA library (CLONTECH
Laboratories, Palo Alto, CA) were screened with restriction enzyme
fragments derived from a full-length human nNOS cDNA previously
described (45): 1344-bp BamHI fragment, exons 2-6; 1476-bp
BamHI fragment, exons 6-17; 867-bp
BamHI/XhoI fragment, exons 17-23; 2153-bp
XhoI/BamHI fragment, exons 23-29. Bacteriophage
were plated and transferred to nitrocellulose filters. Probes were
labeled to a specific activity of 8 × 108 cpm/µg
with [
-32P]dCTP (DuPont NEN; specific activity, 3000 Ci/mmol) using the random primer method. Hybridization and
post-hybridization washes were carried out at 65 °C.
Cross-hybridizing bacteriophage were isolated by plaque purification
prior to DNA sequence analysis as described (45).
Modifications of the RACE protocol were utilized to
isolate the 5-termini of mRNA transcripts for human nNOS in human
testis total cellular RNA (45). Briefly, total cellular RNA derived from normal human testis (1 µg) (CLONTECH) was
reverse transcribed with a gene-specific antisense primer (P1, exon 12, 5
-GCA CGA TCC ACA CCC AGT-3
) using SuperScript II reverse
transcriptase (Life Technologies, Inc.). RNase H (Pharmacia Biotech
Inc.) was used to remove template RNA, and first strand cDNA
product was tailed with dATP by terminal deoxynucleotide transferase
(Life Technologies, Inc.). First round PCR was performed for 35 cycles with primer annealing at 57 °C (Perkin-Elmer 480 thermocycler) (generic sense primers 5
-GAC TCG AGT CGA CGA ATT CAA
T(17)-3
, 2.5 pmol; 5
-GAC TCG AGT CGA CGA ATT CAA-3
, 25 pmol; gene-specific antisense primer P2, exon 8, 5
-CGT GCT TGC CGT CTG
TCC T-3
, 25 pmol). A second round of amplification was performed for
35 cycles with primer annealing at 57 °C (generic sense primer
5
-GAC TCG AGT CGA CGA ATT CAA-3
, 25 pmol; nested gene-specific
antisense primer P3, exon 7, 5
-CTT TGT TGG TGG CAT ACT TGA-3
, 25 pmol). PCR products were subcloned into pBluescript II SK(
) and
subjected to DNA sequence analysis.
A second 5-RACE analysis was performed to independently assess
putative start sites of transcription using an additional set of
gene-specific antisense primers and total cellular RNA derived from
multiple independent sources of normal human testis and brain
(CLONTECH). Reverse transcription was carried out
with a gene-specific antisense primer (P4, exon 5, 5
-CTT TGG CGA GAG GGA AGA G-3
). First round PCR amplification was performed with primer
annealing at 51 °C (gene-specific antisense primer P5a, exon 4, 5
-CAA TGT GCT CTT AAG GTG G-3
or P5b, exon 4, 5
-GAC CTT GAG GAA GCG
TGG AC-3
). Second round amplification was performed with primer
annealing at 49 °C (nested gene-specific antisense primer P6, testis
exon 2, 5
-TGG TTA TGT GAC CCT CGT TG-3
). PCR products were subcloned
into the pCRII vector (TA cloning kit, Invitrogen, San Diego CA) and
subjected to DNA sequencing analysis.
Characterization of Genomic Organization
A 15-kb EcoRI fragment was subcloned from a cosmid
genomic sublibrary of the human nNOS previously reported (29). Plasmid subclones were analyzed with restriction enzyme mapping and Southern blotting with oligonucleotides end-labeled with
[-32P]ATP (DuPont NEN; specific activity, 3000 Ci/mmol). All exonic sequences and exon-intron boundaries were
determined on both strands. Intron sizes were determined with
restriction enzyme mapping and PCR amplification with Elongase (Life
Technologies, Inc.) (46). 5
-Flanking regions of the two testis nNOS
exon 1 examples were sequenced on both strands and potential
cis-acting DNA sequences identified (Eukaryotic
Transcription Factor Data Base release 7.4, Genetics Computer Group
sequence analysis software package, Madison, WI).
RT-PCR
Semiquantitative RT-PCR was utilized to assess the tissue
specificity of novel nNOS mRNA transcripts. First strand cDNA
was synthesized with total cellular RNA (5 µg) derived from varied normal human tissues (CLONTECH) using random
primers and SuperScript II reverse transcriptase (Life Technologies,
Inc.). The following primers were used in PCR amplifications for testis
and full-length nNOS (see Fig. 3): testis exon 1 sense 5-TAG GTG GGG
GTT GAG AAA TG-3
; testis exon 1b sense 5
-AGA GCC TTG GTG GTA ACT-3
; exon 2 sense 5
-ACC AGA GTC AGC CTC CAA-3
; and exon 4 antisense 5
-CAA
TGT GCT CTT AAG GTG G-3
. PCR amplifications were performed under
well-characterized semiquantitative conditions in a total volume of 100 µl for 30 cycles. An 880-bp fragment of human
glyceraldehyde-3-phosphate dehydrogenase was amplified as a
quantitative control (sense primer, 5
-ACA TCG CTC AGA CAC GAT GG-3
;
antisense primer, 5
-GCT GTT GAA GTC AGA GGA GAC C-3
). PCR products
were size-fractionated by agarose gel electrophoresis and transferred
to GeneScreen Plus membranes (DuPont). Southern blots were hybridized
with [
-32P]ATP-labeled oligonucleotides positioned
internally to flanking PCR primers.
Heterologous Eukaryotic Expression of Testis nNOS
Eukaryotic expression vectors containing either the testis or
the full-length nNOS cDNAs were constructed as follows. For the
testis nNOS, the 5-end EcoRI/SphI fragment of
the 7.5-kb full-length human nNOS cDNA (45) was replaced with a
365-bp EcoRI/SphI fragment representing the major
5
-end of the testis nNOS cDNA. This 5247-bp testis nNOS cDNA
containing the full 5
-UTR, coding region, and 1615 bp of the 3
-UTR
was then inserted into the eukaryotic expression vector pcDNA3
(Invitrogen). For the full-length nNOS, an
XbaI/EcoRI fragment of the 7.5-kb human nNOS cDNA that contained 333 bp of the 5
-UTR, coding region, and 1615 bp of the 3
-UTR was subcloned into pcDNA3. CHO-K1 cells (ATCC, Rockville, MD) grown in Ham's F-12 medium supplemented with 10% fetal
calf serum (Life Technologies, Inc.) were transfected with TnNOS,
full-length nNOS, or empty vector using electroporation. Multiple
independent G418-resistant clones were selected 72 h later in 800 µg/ml G418 (Life Technologies, Inc.).
Western Blot Analysis
Stably transfected CHO-K1 cell monolayers were lysed with
boiling sample buffer (62.5 mM Tris-HCl (pH 7.4), 2% SDS,
10% glycerol, and 5% -mercaptoethanol). Cell lysates were boiled
for 5 min and centrifuged at 100,000 × g for 30 min.
The cellular supernatant was collected and protein concentration
measured using the Bio-Rad protein assay kit II. Human testis protein
extracts were obtained from CLONTECH. Rat
cerebellar homogenates were prepared with a Polytron PT3000 (Brinkmann
Instruments). Protein samples were electrophoresed in 6%
polyacrylamide-SDS gels and transferred by electroblotting onto
nitrocellulose membranes. Blots were incubated with anti-human nNOS
monoclonal antibody (1 µg/ml) (Transduction Laboratories, Lexington,
KY) and subsequently with horseradish peroxidase-conjugated sheep
anti-mouse secondary antibody (1:20,000 dilution) (Amersham Corp.).
Signal detection was facilitated with ECL (Amersham Corp.).
Measurement of NOS Activity
Determination of L-[14C]Citrulline Formation (47)Stably transfected CHO-K1 cells were cultured in
six-well plates (Costar, Cambridge, MA) to confluency and allowed to
equilibrate in 1 ml of a physiological salt solution composed of (in
mM) 130 NaCl, 5 KCl, 1.5 CaCl2, 1 MgCl2, 25 HEPES (pH 7.4), and 10 D-glucose for
30 min at 37 °C. After being labeled with
L-[14C]arginine (5 × 105
cpm/ml) (Amersham Corp., specific activity, 300 mCi/mmol) for 30 min, cells were treated with either ionomycin (5 µM)
(Calbiochem) or vehicle for 20 min. The reaction was stopped by washing
the cell monolayers with 4 ml of ice-cold phosphate-buffered saline containing 5 mM EDTA. The monolayers were then extracted
with 1 ml of 0.3 M HClO4 for 20 min at 4 °C.
After neutralization, the extract was loaded onto a 1-ml wet bed volume
of Dowex AG 50WX-8 cation-exchange resin (Bio-Rad, Na+
form, 100-200 mesh) followed by 4 ml of water.
L-[14C]Citrulline in the 5-ml column effluent
was quantitated by scintillation counting. To validate the product of
cation-exchange chromatography, fractions of the column effluent were
resolved using TLC silica gel (Alltech Associates, Inc., Deerfield,
IL). Results indicated that L-[14C]citrulline
was the only detectable product.
Direct measurement of NO in cell culture was performed using an NO-sensitive electrode as described previously (48). Briefly, cells grown in 35-mm dishes (Costar) were preincubated in Krebs-Ringer buffer (pH 7.4) of the following composition (in mM): 136 NaCl, 5.4 KCl, 1.8 CaCl2, 0.8 MgSO4, 1 NaH2PO4, 10 HEPES, and 5.5 glucose, supplemented with 7.5 units/ml superoxide dismutase (Sigma). NO release was monitored with an NO-selective microprobe (Inter Medical Co., Nagoya, Japan). The working electrode made of platinum/iridium alloy was coated with a film containing KCl, NO-selective nitrocellulose resin (pyroxyline lacquer), and gas-permeable silicon membrane (49). A counter electrode was made of carbon fiber. The redox current was detected with a current-voltage converter circuit and continuously recorded. The probe tip (diameter, 25 µm) was enclosed in a Faraday's chamber and positioned 3-5 µm above the cell surface. Calibration of the electrochemical sensor was performed using varied concentrations of a nitrosothiol donor, S-nitroso-N-acetyl-DL-penicilamine (Sigma) (48).
Screening of approximately 2 × 106 plaques from two independent testis cDNA libraries yielded one positive clone. Restriction mapping and sequence analysis indicated that the 2.3-kb random primed cDNA insert represented exons 9-23 of human nNOS. No insertion or deletion of exons was detected.
5-RACE was utilized to define 5
-termini of nNOS transcripts in human
testis. This PCR-based approach allowed isolation of human nNOS
cDNA sequences upstream of exon 6. Primers were located in exons 12 (P1), 8 (P2), and 7 (P3) (Fig. 1A).
Thirty-three cDNA clones were sequenced. Nine of these clones were
identical in nucleotide composition to nNOS sequences defined in human
brain (45). Twelve independent clones featured a novel 5
-terminus. All
of these 12 clones exhibited 155 bp of new sequence upstream of nNOS
exon 4 (Fig. 1). This novel nNOS mRNA transcript, hereafter referred to as TnNOS, appeared to be expressed at a level comparable to
that of the full-length nNOS in the testis as judged by the number of
clones isolated in the 5
-RACE experiment and assessment of
semiquantitative RT-PCR signals (see below). Subsequent studies (Fig.
2, see below) indicated that this 5
-terminus was
encoded by two novel exons (Tex 1 and Tex 2) spliced to exon 4 of human nNOS. Characterization and analysis of the TnNOS mRNA transcript indicated an open reading frame of 3294 nt, which predicted a protein
of 1098 amino acids with a calculated molecular mass of 125,017 Da.
Translation initiation was predicted to occur within exon 5 at the
nucleotide sequence TCTGCG, which was consistent with a
Kozak consensus sequence for initiation of translation in eukaryotes
(50). 5
-UTR sequences of 311 nt showed stop codons in all three
translation reading frames. Although lacking 336 amino acids at the
NH2 terminus of the full-length nNOS protein, the novel
TnNOS protein contained all of the functional modular domains perceived
to be necessary for NOS enzymatic activity, specifically those for
binding of heme, tetrahydrobiopterin, calmodulin, FMN, FAD, and NADPH.
Further analysis indicated that this novel human nNOS protein is an
analogue of the mouse nNOS
. A third testis nNOS cDNA variant was
also isolated. This minor species (3 clones of the 33 isolated),
designated as TnNOSb, differed from TnNOS at sequences 5
of Tex 2, possessing a unique 95-bp exon 1 termed as Tex 1b (Fig. 1B).
The structure of this cDNA predicted a protein that is identical to
nNOS
. Nine of the 33 cDNA clones isolated from testis mRNA
sources revealed an interesting cassette insertion of Tex 2 between
exons 3 and 4 of the full-length nNOS (Fig. 3B). A 56-nt
insertion in the coding region of the full-length nNOS resulted in a
frameshift and introduced a stop codon (TGA) 371 nt downstream of the
splice junction within exon 6. The 1221-nt open reading frame encoded a
COOH-terminal truncated novel protein of 407 amino acids with a
calculated molecular mass of 43,788 Da (Fig. 1C).
Given the propensity for reverse transcriptase to pause at regions of
RNA secondary structure, a second 5-RACE protocol with more 5
primers
was performed to provide an independent assessment of RNA structure, to
reassess approximate transcription initiation sites, and also to assess
the tissue specificity of the TnNOS mRNA transcripts. Primers for
this additional 5
-RACE procedure (P4, P5, and P6) were located
upstream with respect to the previous 5
-RACE primers (P1, P2, and P3)
(Fig. 1A). Of 44 cDNA clones derived from two
independent samples of human testis RNA, no additional sequences
preceding Tex 1 were defined, confirming the structure of Tex 1. Of
note, no TnNOS cDNA sequences were isolated from human brain total
cellular RNA (see below).
Nucleotide
sequences reported in this study have been submitted to the
GenBankTM/EMBL Data Bank with accession numbers
U66360-U66362[GenBank][GenBank][GenBank]. DNA sequence analysis and restriction enzyme mapping of
genomic DNA fragments indicated that the novel 155-bp sequence at the 5-end of the TnNOS cDNAs was encoded by two unique exons (Tex 1, 99 nt; Tex 2, 56 nt) that were spliced to exon 4 of the full-length nNOS (Fig. 2A). The minor variant TnNOSb started with
another unique exon (Tex 1b, 95 nt) that was spliced to Tex 2. Tex 1, Tex 1b, Tex 2, and corresponding 5
-flanking regions were localized to
intron 3 of the full-length nNOS (Fig. 2A). Intron/exon
boundaries conformed to the GT/AG donor/acceptor site rule, maintaining
the consensus sequences described by Mount (51) (Table
I).
|
Analysis of 1800 nt of 5-flanking regions of Tex 1 revealed a GC
content of approximately 50%. The dinucleotide CpG was
underrepresented (45 observed, 109 predicted), consistent with the
balance of mammalian genomic DNA. Two independent rounds of 5
-RACE,
using different primers, indicated a common putative transcription
start site. We take these data to indicate that these genomic regions
represent sites for transcription initiation and that these 5
-flanking regions likely represent promoter regions active in testis tissues. Future analyses with primer extension, S1 nuclease, and/or RNase protection will be necessary to confirm this transcription initiation site. Although lacking a classical TATA box, sequence analysis of these
5
-flanking regions indicated the presence of numerous binding sites
for ubiquitous transcription factors (Fig. 2B) including an
inverted Sp1 site (CCCGCC) (52), two ATF/CRE-like elements (TGACGTCA)
(53), a high affinity site for NF1 (TGGCNNNNNNCCA) (54), a consensus
sequence for NF
B binding (GGGRHTYYHC) (55), an AP-1 site (TGASTCA)
(56), four AP-2 sites (GSSWGSCC)/(YCSCCMNSSS), and three inverted AP-2
sites (57, 58). In addition, a variety of cis-regulatory
elements implicated in testis-specific transcription were evident,
including three consensus GATA sites (WGATAR) (59) and five GATA-like
sites, a putative p53 half-element (RRRCWWGYYY) (60), an Ets family
protein binding site (SMGGAWGY) (61), three related Pu box sequences
(GAGGAA) (62) and three PEA3 sequences (AGGAAR) (61), an MEF-2 motif
(YTAWAAATAR) (61), and an insulin response element (IRE) site
(CCCGCCTC) (63) overlapping the Sp1 site, among others.
Three copies of the Alu family of short interspersed repetitive
sequences were found within 1.8 kb of 5-flanking region sequences (Fig. 2). Comparison with Alu consensus sequences revealed that these
three Alu elements belonged to two different Alu subfamilies. A
proximal element (
419 to
245 nt) on the bottom strand belonged to
the Alu-Sb subfamily (89% identity with the Alu-Sb consensus sequence), whereas the other two Alu elements (
1748 to
1579 nt and
1039 to
750 nt) on the top strand belonged to the Alu-Sx subfamily
(87 and 83% identity with the Alu-Sx consensus sequence, respectively). Alu repetitive sequence appears approximately every 4-8
kb in human genomic DNA (64). Therefore, the frequent appearance of
this repetitive element in the 5
-flanking region of human TnNOS is
atypical.
Testis exon 1b (Tex 1b) of TnNOSb and its 5-flanking region were
located farther 5
of Tex 1 (Fig. 2A). Putative
cis-regulatory elements are indicated in Fig. 2C.
An upstream CAAT box (CCAATC) was found at bp
330 (65). No TATA box
was found within this region. A consensus GATA site overlapped an
inverted c-fos/serum response element half site (GGACATC)
(66). An AP-1 site, three AP-2 sites, and two PEA3 sequences were found
within this region.
Northern blot analysis was performed with a 1.5-kb
EcoRI/NheI cDNA fragment corresponding to Tex
1 to exon 13. Total cellular RNA (20 µg) or poly(A) RNA (5 µg) from
normal adult human testis, brain, and skeletal muscle was used. A weak
hybridization signal for the 8-9-kb full-length nNOS was detected only
in RNA isolated from skeletal muscle, indicating that the full-length
nNOS and TnNOS were both uncommon transcripts under normal conditions. RT-PCR demonstrated that the TnNOS transcript was expressed exclusively in testis under normal conditions (Fig. 3A).
In contrast to the tissue-specific expression of TnNOS, full-length
nNOS mRNA transcript was expressed in a wide variety of human
tissues (Fig. 3B). The minor band present in Fig.
3Aii was not detected when the blot was hybridized with an
oligo probe located in Tex 2 (Fig. 3Aiii). This indicated
that Tex 2 was alternatively spliced in or out in TnNOS mRNA
transcripts. Such an event was not detected in 5-RACE clones, probably
due to its low frequency (
1:20 in comparison with the major band by
densitometric analysis). The biological significance of this cassette
deletion within the 5
-UTR is unclear, but the deletion may affect the
translational efficiency of TnNOS mRNA transcripts. Results from
5
-RACE indicated the potential insertion of Tex 2 between exons 3 and
4 of the full-length nNOS transcript. Consistent with these findings,
RT-PCR products corresponding to a cassette insertion of Tex 2 were
detected using an oligonucleotide probe located in Tex 2 (Fig.
3Biii). This amplicon was also detected as a minor band when
hybridized with an exon 3 oligonucleotide probe. When this same blot
was hybridized with oligonucleotide probes located in Tex 1 and 1b, no
signal was detected, indicating that these two exons did not exhibit
cryptic 3
-acceptor sites.
Similar experiments were performed with TnNOSb. Results showed that this transcript was expressed restrictively in the testis at an extremely low level under normal conditions. No alternative splicing events were detected (data not shown). These findings confirmed that this mRNA was a rare transcript under normal circumstances.
TnNOS mRNA Encodes a Functional NOSProtein extracts of
stable CHO-K1 transfectants containing full-length nNOS, TnNOS, or
pcDNA3 vector sequences were assessed with Western blot assay using
an anti-human nNOS monoclonal antibody directed toward the COOH
terminus. Results from multiple independent clones demonstrated that
the TnNOS cDNA encoded the truncated 125-kDa protein as predicted
from its nucleotide sequence (Fig. 4). In contrast, the
full-length nNOS cDNA encoded a 160-kDa protein similar to that
found in rat cerebellum tissue. Experiments using up to 150 µg per
lane of human testis total cellular protein from two independent
isolations failed to detect any specific signal for either full-length
nNOS or nNOS, indicating that this enzyme was expressed at
restricted levels in these tissues, perhaps in a cell-specific
fashion.
To determine whether the 125-kDa human nNOS variant was a functional
NOS, L-[14C]arginine to
L-[14C]citrulline conversion was measured in
stable transfectants. Results indicated that
L-[14C]citrulline formation increased
significantly following treatment with the calcium ionophore ionomycin
(5 µM, 20 min) (Fig. 5A). Production of L-[14C]citrulline was not
significantly different in stable CHO-K1 transfectants containing
full-length nNOS or TnNOS (Fig. 5A). Measurements of NO
release from cells were consistent with these findings. As shown in
Fig. 5B, stable TnNOS transfectants produced comparable
amounts of NO as full-length nNOS transfected cell lines following
treatment with calcium ionophore. Taken together, these data indicate
that the protein derived from the TnNOS mRNA transcript is a
calcium-dependent nitric-oxide synthase.
The human nNOS gene is a highly complex locus. In the current
study we have demonstrated that in the normal testis transcription can
initiate from downstream regions within intron 3. Depending on
alternative splicing, one or two novel exons are produced that in turn
are spliced to exon 4 of the full-length nNOS mRNA (Fig. 3A). Translation of this novel nNOS transcript initiates in
exon 5, resulting in a 125-kDa protein that is an analogue of the mouse nNOS. In comparison with the 160-kDa full-length nNOS, whose translation initiates in exon 2, this protein lacks the
NH2-terminal 336 amino acids but possesses NOS enzymatic
activity comparable to that of its full-length counterpart and is
therefore a functional nitric-oxide synthase. This novel nNOS
transcript is testis-specific and is designated as TnNOS in contrast to
the full-length nNOS.
Putative transcription initiation sites for TnNOS were mapped with
5-RACE methods. 5
-RACE has been used for the mapping of transcription
start sites of a number of other uncommon mRNA transcripts,
producing results consistent with those obtained with conventional
methods such as primer extension, RNase protection, and S1 nuclease
protection assays (45). 5
-RACE is robust compared with other cloning
approaches and is therefore a valuable alternative in structural
characterization of infrequent mRNA transcripts. TnNOS cDNA
clones isolated in the first 5
-RACE analysis were identical in 5
-UTR
length to results obtained from a further 5
-RACE analysis using a
different cloning strategy and multiple testis RNA sources. However,
further confirmation with other methods is warranted.
Translation initiation for this novel TnNOS transcript within exon 5 was confirmed by functional heterologous eukaryotic expression in
stable cell lines and assessment of protein size with immunoblot analysis (Fig. 4). Enzymatic activity was assessed in multiple independent stable transfectants using two independent methods. Measurements of whole cell calcium-activated
L-[14C]arginine to
L-[14C]citrulline conversion and NO release,
as assessed with an NO electrode (Fig. 5), indicate that this protein
is a functional NOS. A recent study has shown the existence of a
similar protein (nNOS) in wild type mice and mice in which the
full-length nNOS exon 2 was disrupted with targeted inactivation. In
these prior examples the underlying molecular mechanism involved
alternative splicing between nNOS exons 1 and 3 and subsequent
translation initiation in exon 5 (10). This is a separate and distinct
mechanism from what is reported here. Transient cDNA expression
demonstrated that mouse nNOS
possesses only minor NOS enzymatic
activity as assessed by in vitro methods. The discrepancy
between the results of this prior study and our results may reflect the
difference between transient and stable expression systems, the use of
different cell types, species differences, or translational repression
consequent to the complex 5
-UTR (523 nt) produced in the mouse
splicing variant (10).
The subcellular localization of the human nNOS in the testis may be
changed profoundly due to the lack of the NH2-terminal 336 amino acids. nNOS is mainly a membrane-bound enzyme (9, 67). It has
been demonstrated that in skeletal muscle nNOS is associated with the
sarcolemma through interaction of the NH2-terminal PDZ/GLGF
motif with the dystrophin complex (9). In neurons, synaptic association
of nNOS has been shown to be mediated by the binding of this
NH2-terminal PDZ/GLGF motif to proteins such as
postsynaptic density 95 (PSD-95) and the related novel protein PSD-93
(10). Without the PDZ/GLGF motif, nNOS
in nNOS mutant mice has been
reported to be restricted to soluble fractions of skeletal muscle (10).
It follows that the human protein derived from TnNOS would be similarly
distributed in testis. However, in stable CHO-K1 transfectants, Western
analysis using proteins derived from subcellular fractionization (67)
demonstrated that the human protein derived from TnNOS as well as the
full-length nNOS was distributed in both cytosol and particulate
fractions (data not shown). Because the full-length nNOS is not
restricted to membrane fractions in CHO-K1 transfectants, we were
unable to directly determine whether the truncated protein varies in subcellular localization in this specific cell type. It is likely that
unique cell-specific docking mechanisms exist in varied cell types.
The 5-flanking region of TnNOS may contain the promoter that directs
testis-specific expression of the nNOS gene. The presence within this
region of multiple binding sites for both ubiquitous and
testis-specific transcription factors is consistent with this proposal.
The proximal Sp1 site at
184 may represent a core element in the
transcriptional activation of this TATA-less transcription unit (68).
An unusual feature of this region is the presence of three clustered
Alu elements within 1.8 kb of the putative transcription initiation
site (Fig. 2, A and B). Originating from 7SL RNA,
Alu sequences are short, interspersed repetitive DNA elements,
appearing approximately every 4-8 kb in human genomic DNA and
representing about 5-6% of the total human genome (64). Most Alu
sequences have been considered to be functionally inert. However, when
located in promoter regions, Alu elements have been shown to exert
transcriptional repressor effects (69), transcriptional enhancer
effects (70, 71), or both (72). Alu sequences may be involved in
cell-specific expression of the erythropoietin receptor (73) and the
chain of Fc and T-cell receptors (72). Multiple Alu elements
clustering within the promoter/5
-flanking region have been reported in
a number of human genes including the poly(ADP-ribosyl) transferase
gene (74), lysozyme gene (75), prostatic acid phosphatase ACPP gene
(76), and erythropoietin receptor gene (73). An interesting functional
feature of clustered Alu elements in promoter regions is that those
located on opposite strands may be involved in the formation of DNA
tertiary structures implicated in juxtaposing distal
cis-regulatory elements to the proximal general
transcription factors (74). This feature may be of functional relevance
in the transcriptional regulation of TnNOS in that the most upstream
and the most downstream Alu elements of the 5
-flanking region are on
the opposite strands of the helix.
The functional relevance of this complex putative TnNOS promoter was not addressed in the current study. Future studies using promoter-reporter constructs in vitro and in vivo are required for the functional characterization of this transcriptional regulatory region. Based on the above findings, however, it is likely that multiple positive and negative cis-regulatory elements may cooperate in the tissue-/cell-specific and developmental regulation of TnNOS expression. Given its downstream localization, the TnNOS promoter is likely to be functional in the nNOS-deficient mice that were reported recently (25). Whether mRNA expression from this downstream promoter is modulated secondary to disruption of upstream regulatory regions remains to be assessed.
Recent studies suggest that NO/NOS play fundamental roles in maintaining reproductive endocrine function (24, 77-80). In particular, nNOS has been implicated in penile erection (24) and testicular function (78). However, the methods used in these prior studies did not allow a distinction to be made between the biological contribution of the proteins derived from full-length nNOS and TnNOS mRNA transcripts. We posit that TnNOS mRNA transcripts may exist in other sexually dimorphic organs of the male and female. The current study indicates that both mRNA transcripts are expressed in the testis in approximately equivalent amounts. Effort was made in this study to identify the cell types in testis that express the TnNOS mRNA transcripts using in situ hybridization. Although hybridization signals were detected in testis tissues of an adult human male and an adult baboon collected at necropsy, the low level of mRNA expression and/or detection precluded definition of cell-type specificity. The biological significance of this transcript warrants study. Jaffrey and Snyder (11) have recently cloned and characterized a full-length nNOS-inhibiting protein (PIN) that inhibits full-length nNOS enzymatic activity by protein-protein interaction with the NH2-terminal region (amino acids 163-245) of the full-length nNOS. The binding of PIN apparently prevents the dimerization of the enzyme by interfering with a unique dimerization domain located within the first 165 amino acids of the full-length rat nNOS. Interestingly, PIN mRNA transcript expression is most abundant in the testis (11). This suggests that PIN would inhibit full-length nNOS function in this organ. Given that the NH2-terminal truncated protein does not possess the PIN-binding domain and thus would not be inhibited by PIN, it is intriguing to postulate that this novel variant of nNOS derived from a unique downstream promoter sustains biological function of nNOS in the testis in the presence of PIN.
Cassette deletions of exons 9/10 or 10 in full-length human nNOS have been reported (29). An additional finding in the current study is that the testis exon 2 (Tex 2) originally isolated from testis was subsequently found to be alternatively inserted between exons 3 and 4 of the full-length nNOS in a variety of human tissues. Although occurring as minor events in comparison with the normal full-length nNOS transcript, the biological consequences of these alternative splicing events may be important. It is unclear what the determinant of these alternative splicing events is or how these events are regulated. Nor is it known whether these alternative mRNAs are translated in vivo and, if they are, what the biological significance is. Given that the NOS proteins all function as dimers in the enzymatic synthesis of NO, fruitful areas for future study would be whether a cell expresses both the full-length and shorter mRNAs, whether these shorter mRNAs are translated in vivo, and whether alternate proteins participate as heterodimers in the catalytic synthesis of NO.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U66360[GenBank], U66361[GenBank], U66362[GenBank].
We thank Dr. Yu Hu for his technical assistance in NO measurement and Eddy Wong for his assistance in DNA sequence analysis.