Alternatively spliced neuronal nitric oxide synthasemediates penile erectionK. Joseph Hurt*, Sena F. Sezen†, Hunter C. Champion‡, Julie K. Crone†, Michael A. Palese†, Paul L. Huang§,Akira Sawa*¶�, Xiaojiang Luo*, Biljana Musicki†, Solomon H. Snyder*�**††, and Arthur L. Burnett†,††

Departments of †Urology and ‡Internal Medicine, Division of Cardiology, and Departments of ¶Cellular and Molecular Medicine, *Neuroscience,**Pharmacology and Molecular Sciences, and �Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine, Baltimore, MD 21205;and §Cardiovascular Research Center and Cardiology Division, Massachusetts General Hospital and Harvard Medical School, Charlestown, MA 02129

Contributed by Solomon H. Snyder, December 30, 2005

A key role for nitric oxide (NO) in penile erection is well established,but the relative roles of the neuronal NO synthase (nNOS) versusendothelial forms of NOS are not clear. nNOS- and endothelialNOS-deficient mice maintain erectile function and reproductivecapacity, questioning the importance of NO. Alternatively, residualNO produced by shorter transcripts in the nNOS�/� animals mightsuffice for normal physiologic function. We show that the � splicevariant of nNOS elicits normal erection despite a decrease instimulus-response characteristics and a 5-fold increased sensitivityto the NOS inhibitor, L-NAME. Residual nNOS� generates only 10%of the normal NO level in vitro but produces citrulline and diaph-orase staining reflecting in vivo NOS activity in pelvic ganglionnerves that is comparable to WT animals. Thus, alternativelyspliced forms of nNOS are major mediators of penile erection andso may be targets for therapeutic intervention.

penis � cavernous nerve � NADPH diaphorase

N itric oxide (NO) is synthesized by three major enzymesderived from distinct genes, inducible NO synthase (iNOS),

endothelial NOS (eNOS), and several variants of neuronal NOS(nNOS) (1). Penile erection is mediated by NO produced by bothnNOS and eNOS (2), the former initiating erection (3), and thelatter providing sustained maximal erection (4). The relativecontributions of different forms of NOS to penile erection areunclear, especially because erection is preserved in mice withtargeted deletion of nNOS� (nNOS�/�) as well as in eNOS�/�

animals (5–7). In the present study, using nNOS�/� (8),eNOS�/� (9), and doubly mutant mice (10) (dNOS�/�), we showthat alternatively spliced forms of nNOS mediate a major portionof penile erection.

Results and DiscussionWe evaluated erectile function by monitoring intracavernouspressure (ICP) in response to electrical stimulation of thecavernous nerve at different voltages (Fig. 1A). Erection ismaintained in all of the transgenic animals, and at the highestvoltage tested (6 V), there is no significant difference amonggroups. However, nNOS�/� and dNOS�/� animals are lessresponsive to low-voltage stimulation. At 1 V, nNOS�/� anddNOS�/� animals have only 4% and 26% of the stimulatedchange in ICP seen in WT animals. In contrast, eNOS�/�

animals appear supersensitive to electrical stimulation, with150–300% increased response at 1–4 V compared with WT.

We wondered whether the retained erection in mutant micereflects a residual NO-generating system or is unrelated to NO.We evaluated the influence of L-nitroarginine methylester (L-NAME), which inhibits all forms of NOS nonselectively (11).Strikingly, all three mutant mice are supersensitive to inhibitionof erection by L-NAME, with the doubly mutant mice 10 timesmore sensitive than WT and the nNOS�/� and eNOS�/� animals5 and 2.5 times more sensitive, respectively (Fig. 1B). Thissupersensitivity may reflect the greater ease of L-NAME toinhibit smaller amounts of nNOS in mutant animals. At higher

doses, L-NAME inhibition is similar among all groups. We alsonote striking differences in the patterns of stimulated erection intransgenic mice. dNOS�/� mice, in particular, demonstrate anirregular and dysfunctional priapic activity (Fig. 1C).

The preservation of electrically stimulated erectile function innNOS�/� animals suggests that the nerve fibers contain residualNOS. Accordingly, we compared total in vitro NOS activity inWT and transgenic penes (Fig. 2A). Total NOS activity isreduced 30–40% in eNOS�/� tissue and almost 90% in nNOS�/�

and double-knockout preparations. Residual activity is com-pletely inhibited by 500 �M L-NAME. Because iNOS has notbeen detected in normal WT rat and mouse penile tissues (5), wewondered whether alternatively spliced forms of nNOS mightmaintain erectile activity. In certain brain regions, alternativelyspliced and catalytically active nNOS� has been described (12),whereas another alternatively spliced form, nNOS�, also persistsin the nNOS�/� brain but lacks catalytic activity (13). Weconducted in situ hybridization for the predominant � form ofnNOS and the catalytically active � form (Fig. 2B). In WT mice,both NOS� and -� are prominent in the neuronal cell bodies ofthe pelvic ganglion, which give rise to the cavernous nerveinnervation of the penis. Similar localizations are observed witha common probe that recognizes both � and � forms. Stainingfor nNOS� mRNA is abolished in nNOS�/� animals, but noreduction of nNOS� is apparent, consistent with the design ofthe mutant mice in which exon-2 was deleted, leaving thoseforms that lack that portion of the N terminus of the proteinunaffected.

To investigate whether residual nNOS� produces NO in nervefibers, we stained penile tissue for citrulline, which reflects NOSactivity in intact animals (14) (Fig. 2C). We observe citrullinestaining in the cavernous and dorsal nerves of the penis as wellas endothelial tissue lining the cavernous sinuses and penileblood vessels. Citrulline staining is not reduced in nNOS�/�

penile sections. The retention of apparent NOS activity innNOS�/� intact penes may reflect a greater capability of residualNOS to generate citrulline (and presumably NO) in vivo but notin vitro or might reflect diminished degradation of citrulline inthe mutant animals. To distinguish these possibilities, we eval-uated neuronal production of NO in fresh tissue. First, wemeasured Ca2�-dependent NOS activity from purified cavern-osal nerve cells in transgenic animals transfected with a herpes-virus-lacZ marker (Fig. 3A), as described (15, 16). Although WTand eNOS�/� animals have similar neuronal NOS activity, asexpected, nNOS�/� and dNOS�/� animals have 60% and 57% ofthe WT activity. The activity is calcium-dependent and com-

Conflict of interest statement: No conflicts declared.

Abbreviations: NOS, NO synthase; nNOS, neuronal NOS; eNOS, endothelial NOS; iNOS,inducible NOS; dNOS, double (endothelial and neuronal) NOS; ICP, intracavernous pres-sure; L-NAME, L-nitroarginine methylester.

††To whom correspondence may be addressed. E-mail: ssnyder@jhmi.edu or aburnett@jhmi.edu.

© 2006 by The National Academy of Sciences of the USA

3440–3443 � PNAS � February 28, 2006 � vol. 103 � no. 9 www.pnas.org�cgi�doi�10.1073�pnas.0511326103

pletely abolished by 7-nitroindazole, a selective nNOS inhibitor.To further confirm the substantial nNOS activity in the trans-genic mice, we performed diaphorase histochemical staining ofisolated pelvic ganglia (17), which reflects NOS catalytic activity.This staining is similar in neuronal cell bodies for WT, eNOS�/�,nNOS�/�, and dNOS�/� mice (Fig. 3B).

In summary, our findings indicate that the alternatively splicednNOS� mediates a major portion of penile erection. AlthoughNOS catalytic activity monitored in penile tissue extracts isreduced 90% in nNOS mutant mice, no substantial reduction incitrulline or NADPH diaphorase staining or NOS activity ofcavernosal nerve lysates is evident in the mutants. nNOS�generation of NO�citrulline is also preserved in the brains ofnNOS�/� animals (14).

Unique properties of nNOS� may explain its robust produc-tion of NO to preserve erectile function in nNOS�/� animals(18). nNOS� lacks the N-terminal PDZ domain that binds toPSD95, the postsynaptic density protein that anchors nNOS� toneuronal membranes and links nNOS to the NMDA glutamatereceptors (13). The lack of tethering to the plasma membrane for

Fig. 1. Transgenic mouse erectile responses to nerve stimulation. (A) Elec-trically stimulated erection as measured by increased ICP (22) is reduced at lowvoltages in nNOS�/� and dNOS�/� strains but becomes nearly the same as WTresponse at higher voltages. eNOS�/� animals are more sensitive than WT atall voltages tested. *, P � 0.01; **, P � 0.05. (B) L-NAME more potently inhibitserection mediated by 4-V stimulation in nNOS�- and dNOS-deficient mice,whereas eNOS�/� strains are more resistant to inhibition. L-NAME IC50 valuesare: WT, 4.2 mM; eNOS�/�, 1.6 mM; nNOS��/�, 0.95 mM; and dNOS�/�, 0.44mM. Data are mean � SEM for 3–15 animals per dose. (Inset) Magnification ofthe initial dose–response curves for the transgenic animals. (C) Representativetracings of 4-V stimulation of erection for 2 min (solid bar) followed byinjection of 15 mM L-NAME in the mouse penis (at hatchmarks) and secondstimulation at 4 V. The irregular unstable priapic response of the dNOS�/�

mouse is clearly shown. Peak change in ICP is measured as the highestcontinuous pressure peak maintained for at least 20 sec during the stimulusinterval.

Fig. 2. Presence of enzymatically active alternatively spliced nNOS in trans-genic mice. (A) Total NOS activity in penile homogenates measured in vitro bythe conversion of arginine to citrulline as described (23) and compared withWT activity. Total NOS activity in WT animals was 367 � 73 pmol�min per mgof protein. In controls, L-NAME added at 500 �M blocked all activity. Data aremean � SEM for six to seven animals. (B) In situ hybridization of WT andnNOS�/� mouse pelvic ganglia using sense and antisense probes to � or � nNOSisoform-specific RNA or to a sequence common to all nNOS isoforms. Solid andbroken arrows point to stained and unstained neuronal nuclei, respectively.(C) Citrulline staining in WT and nNOS�/� penile tissue. Bracket marks theendothelial surface of the penile artery. Arrow marks a distinctly stained nervefiber in the cavernous nerve. (Inset) Background staining in the absence ofprimary antibody.

Hurt et al. PNAS � February 28, 2006 � vol. 103 � no. 9 � 3441

PHYS

IOLO

GY

nNOS� may provide greater in vivo than in vitro activity. Also,the absence in nNOS� of the domain that binds to the proteininhibitor of NOS, an endogenous inhibitor of NOS (19), may alsopotentiate in vivo catalytic activity.

Because of its unique properties, nNOS� may be a noveltherapeutic target for treating erectile dysfunction, cardiovas-cular disease, and other conditions with NO deficiency. Acomplete nNOS-deficient mouse lacking nNOS�, -�, and -� isinfertile (20, 21). The nNOS��/�, eNOS�/�, and dNOS�/�

animals used in our studies are all fertile, indicating the crucialrole of functional nNOS in normal reproductive function.

MethodsPhysiologic Erection Studies and Inhibition. Age-matched adultmale C57BL6 (WT; The Jackson Laboratory), eNOS�/�,nNOS�/�, or dNOS�/� (from the laboratory of P.L.H.) micewere used in ICP studies (4). Mouse penis was denuded of skinand fascia and the corpus cavernosum was monitored with aneedle connected to a pressure transducer by using MATLABMathworks (Natick, MA) software. A bipolar electrode attachedto a Grass Instruments (Quincy, MA) S48 stimulator was placedaround the cavernous nerve and stimulation was conducted at 1,2, 4, and 6 V. For inhibition studies, stimulation was performedat 4 V for 2 min after a stable baseline was obtained. Then 10�l of the indicated concentrations of L-NAME was injecteddirectly into the penis followed by a 10-min stabilization phaseand then stimulated again for 2 min. The percent inhibition isexpressed as the ratio of �ICPL-NAME��ICPcontrol. IC50 valueswere computed by using dose–response curve-fitting softwarefrom GRAPHPAD PRISM (GraphPad Software, San Diego).

In Situ Hybridization. Digoxygenin in situ hybridization withunique N- or C-terminal probes was conducted as published (14)and according to the manufacturer’s instructions (Calbiochem)by using common or NOS�- or NOS�-specific probes corre-sponding to residues 100–786 of mouse nNOS sequence.

Citrulline Staining. Staining for glutaraldehyde-coupled citrullinewas performed as described (14) by using affinity-purifiedantibodies. Animals were killed and perfused with 5% glutar-aldehyde�0.5% paraformaldehyde, postfixed, and then cryopro-tected. Free-floating sections of mouse penis were reduced withNaBH4, blocked with 4% goat serum, and permeabilized in

0.1% Triton X-100 before incubating overnight with antiserumdiluted to 1:5,000 in PBS. Staining was developed with the ABCElite kit (Vector Laboratories).

nNOS Activity from Isolated Cavernosal Nerve. Mouse penes weretransfected with a herpesvirus (HSVlacZ) encoding V5 andlacZ (25 � 107 plaque-forming units) by direct injection of thecavernosum by using a 30-gauge needle and proximal occlusionof penile venous return by using an elastic band, as described(15). Two days later, penes were harvested and purified byusing anti-V5 antibody (Transduction Laboratories, Lexing-ton, KY) and miniMACS separation unit (Miltenyi Biotec).Confirmation of the cavernosal nerve cell population wasperformed by f low cytometry by using both V5 and lacZ.Constitutive enzyme activity was assayed in total penile ho-mogenates or isolated tissue, as described (22). Brief ly, sam-ples were sonicated in a solution of Tris�HCl (250 mmol/liter,pH 7.4�EDTA 10 nmol/liter�EGTA 10 mmol/liter) and cen-trifuged at 12,000 � g for 10 min at 14°C. The supernatant wasincubated in NADPH [10 mmol/liter)�[3H]L-arginine (1 mCi/ml)�CaCl (6 mmol/liter)�Tris�HCl (50 mmol/liter, pH 7.4)�tetrahydrobiopterin (6 �mol/liter)�f lavin adenine dinucle-otide (2 �mol/liter)�f lavin mononucleotide (2 �mol/liter)] for60 min at 24°C. The reaction was stopped with Hepes (50mmol/liter, pH 5.5) and EDTA (5 mmol/liter). The radioac-tivity eluted from a Dowex column was measured by liquidscintillation counting.

Diaphorase Histochemical Staining. NADPH diaphorase stainingwas performed on sections of mouse pelvic ganglia (17).Brief ly, 20- to 40-�m sections were fixed in 4% paraformal-dehyde and then developed in a reaction buffer containingNADPH and NBT. The sections were counterstained withNuclear Fast Red and fixed in EtOH, then mounted formicroscopy. Blanks were confirmed by omitting NADPH ornitroblue tetrazolium.

This work was supported by U.S. Public Health Service Grants DA00266and Research Scientist Award DA 00074 (to S.H.S.); by U.S. PublicHealth Service Grants DK64679 and DK67223 (to A.L.B.) and MH-069853; by foundation grants from National Alliance for Research onSchizophrenia and Depression and the Stanley Foundation (to A.S.); andby the American Heart Association, the W. W. Smith Charitable Trust,and the Bernard Family Foundation (to H.C.C.).

Fig. 3. Neuronal localization of residual NOS activity in transgenic mice. (A) Calcium-dependent and 100 �M 7-nitroindazole (7-NI) inhibited conversion ofarginine to citrulline in isolated cavernosal nerve tissue is maintained in eNOS�/� animals and reduced in nNOS�/� and dNOS�/� strains. Data are mean � SEMof three to five measurements. EGTA (5 mM) or L-NAME (100 �M) also completely inhibit NOS activity. (B) Histochemical localization of diaphorase activity tolarge neuronal cell bodies is maintained in pelvic ganglia from all transgenic strains. Data are representative of three separate staining procedures for gangliataken from two separate animals.

3442 � www.pnas.org�cgi�doi�10.1073�pnas.0511326103 Hurt et al.

1. Boehning, D. & Snyder, S. H. (2003) Annu. Rev. Neurosci. 26, 105–131.2. Burnett, A. L. (2004) Int. J. Impot. Res. 16 Suppl. 1, S15–S19.3. Burnett, A. L., Lowenstein, C. J., Bredt, D. S., Chang, T. S. & Snyder, S. H.

(1992) Science 257, 401–403.4. Hurt, K. J., Musicki, B., Palese, M. A., Crone, J. K., Becker, R. E., Moriarity,

J. L., Snyder, S. H. & Burnett, A. L. (2002) Proc. Natl. Acad. Sci. USA 99,4061–4066.

5. Burnett, A. L., Nelson, R. J., Calvin, D. C., Liu, J. X., Demas, G. E., Klein, S. L.,Kriegsfeld, L. J., Dawson, V. L., Dawson, T. M. & Snyder, S. H. (1996) Mol.Med. 2, 288–296.

6. Burnett, A. L., Chang, A. G., Crone, J. K., Huang, P. L. & Sezen, S. E. (2002)J. Androl. 23, 92–97.

7. Cashen, D. E., MacIntyre, D. E. & Martin, W. J. (2002) Br. J. Pharmacol. 136,693–700.

8. Huang, P. L., Dawson, T. M., Bredt, D. S., Snyder, S. H. & Fishman, M. C.(1993) Cell 75, 1273–1286.

9. Huang, P. L., Huang, Z., Mashimo, H., Bloch, K. D., Moskowitz, M. A., Bevan,J. A. & Fishman, M. C. (1995) Nature 377, 239–242.

10. Son, H., Hawkins, R. D., Martin, K., Kiebler, M., Huang, P. L., Fishman, M. C.& Kandel, E. R. (1996) Cell 87, 1015–1023.

11. Boer, R., Ulrich, W. R., Klein, T., Mirau, B., Haas, S. & Baur, I. (2000) Mol.Pharmacol. 58, 1026–1034.

12. Lee, M. A., Cai, L., Hubner, N., Lee, Y. A. & Lindpaintner, K. (1997) J. Clin.Invest. 100, 1507–1512.

13. Brenman, J. E., Chao, D. S., Gee, S. H., McGee, A. W., Craven, S. E.,Santillano, D. R., Wu, Z., Huang, F., Xia, H., Peters, M. F., et al. (1996) Cell84, 757–767.

14. Eliasson, M. J., Blackshaw, S., Schell, M. J. & Snyder, S. H. (1997) Proc. Natl.Acad. Sci. USA 94, 3396–3401.

15. Bivalacqua, T. J., Usta, M. F., Champion, H. C., Adams, D., Namara, D. B.,Abdel-Mageed, A. B., Kadowitz, P. J. & Hellstrom, W. J. (2003) J. Urol. 169,1911–1917.

16. Champion, H. C., Bivalacqua, T. J., Hyman, A. L., Ignarro, L. J., Hellstrom,W. J. & Kadowitz, P. J. (1999) Proc. Natl. Acad. Sci. USA 96, 11648–11652.

17. Burnett, A. L., Tillman, S. L., Chang, T. S., Epstein, J. I., Lowenstein, C. J.,Bredt, D. S., Snyder, S. H. & Walsh, P. C. (1993) J. Urol. 150, 73–76.

18. Gonzalez-Cadavid, N. F., Burnett, A. L., Magee, T. R., Zeller, C. B., Vernet,D., Smith, N., Gitter, J. & Rajfer, J. (2000) Biol. Reprod. 63, 704–714.

19. Jaffrey, S. R. & Snyder, S. H. (1996) Science 274, 774–777.20. Gyurko, R., Leupen, S. & Huang, P. L. (2002) Endocrinology 143, 2767–2774.21. Tranguch, S. & Huet-Hudson, Y. (2003) Mol. Reprod. Dev. 65, 175–179.22. Champion, H. C., Bivalacqua, T. J., Takimoto, E., Kass, D. A. & Burnett, A. L.

(2005) Proc. Natl. Acad. Sci. USA 102, 1661–1666.23. Bredt, D. S. & Snyder, S. H. (1989) Proc. Natl. Acad. Sci. USA 86, 9030–9033.

Hurt et al. PNAS � February 28, 2006 � vol. 103 � no. 9 � 3443

PHYS

IOLO

GY