Ubiquitination of Neuronal Nitric-oxide Synthase in the Calmodulin-binding Site Triggers Proteasomal Degradation of the Protein*

Background: CHIP-dependent ubiquitination of NO synthase is an important regulatory mechanism. Results: A dozen ubiquitination sites on neuronal NO synthase were identified with 11 of the sites on either the oxygenase or calmodulin domain. Conclusion: Lysine residue 739 is the major site for poly-ubiquitination with other sites responsible for mono-ubiquitination of neuronal NO synthase. Significance: CHIP-dependent regulation of neuronal NO synthase turnover occurs primarily through lysine residue 739. Nitric-oxide synthase, a cytochrome P450-like hemoprotein enzyme, catalyzes the synthesis of nitric oxide, a critical signaling molecule in a variety of physiological processes. Our laboratory has discovered that certain drugs suicide-inactivate neuronal nitric-oxide synthase (nNOS) and lead to the preferential ubiquitination of the inactivated nNOS by an Hsp70- and CHIP (C terminus of Hsc70-interacting protein)-dependent process. To further understand the process by which altered nNOS is recognized, ubiquitinated, and proteasomally degraded, we examined the sites of ubiquitination on nNOS. We utilized an in vitro ubiquitination system containing purified E1, E2 (UbcH5a), Hsp70, and CHIP that recapitulates the ability of the cells to selectively recognize and ubiquitinate the altered forms of nNOS. LC-MS/MS analysis of the tryptic peptides obtained from the in vitro ubiquitinated nNOS identified 12 ubiquitination sites. Nine of the sites were within the oxygenase domain and two were in the calmodulin-binding site, which links the oxygenase and reductase domains, and one site was in the reductase domain. Mutational analysis of the lysines in the calmodulin-binding site revealed that Lys-739 is a major site for poly-ubiquitination of nNOS in vitro and regulates, in large part, the CHIP-dependent degradation of nNOS in HEK293 cells, as well as in in vitro studies with fraction II. Elucidating the exact site of ubiquitination is an important step in understanding how chaperones recognize and trigger degradation of nNOS.


Nitric-oxide synthase, a cytochrome P450-like hemoprotein enzyme, catalyzes the synthesis of nitric oxide, a critical signaling molecule in a variety of physiological processes. Our laboratory has discovered that certain drugs suicide-inactivate neuronal nitric-oxide synthase (nNOS) and lead to the preferential ubiquitination of the inactivated nNOS by an Hsp70-and CHIP (C terminus of Hsc70-interacting protein)-dependent process.
To further understand the process by which altered nNOS is recognized, ubiquitinated, and proteasomally degraded, we examined the sites of ubiquitination on nNOS. We utilized an in vitro ubiquitination system containing purified E1, E2 (UbcH5a), Hsp70, and CHIP that recapitulates the ability of the cells to selectively recognize and ubiquitinate the altered forms of nNOS. LC-MS/MS analysis of the tryptic peptides obtained from the in vitro ubiquitinated nNOS identified 12 ubiquitination sites. Nine of the sites were within the oxygenase domain and two were in the calmodulin-binding site, which links the oxygenase and reductase domains, and one site was in the reductase domain. Mutational analysis of the lysines in the calmodulin-binding site revealed that Lys-739 is a major site for polyubiquitination of nNOS in vitro and regulates, in large part, the CHIP-dependent degradation of nNOS in HEK293 cells, as well as in in vitro studies with fraction II. Elucidating the exact site of ubiquitination is an important step in understanding how chaperones recognize and trigger degradation of nNOS.
Nitric-oxide synthase (NOS) 2 is a cytochrome P450-like hemoprotein enzyme that catalyzes the conversion of L-argin-ine to L-citrulline and nitric oxide, an important cellular signaling molecule (1). NOS is a homodimer with each monomer having an oxygenase domain, which binds heme, tetrahydrobiopterin, and L-arginine, and a reductase domain, which binds FMN, FAD, and NADPH (2). A calmodulin-binding domain links the oxygenase and reductase domains, and the binding of calmodulin regulates the flow of electrons from the reductase to the oxygenase domain (2). Although a variety of regulatory mechanisms exist for the control of NOS, an important mechanism for the maintenance of NOS protein quality control is through ubiquitination by the C terminus of Hsc70-interacting protein (CHIP), an E3 ubiquitin ligase that binds both Hsp70 and Hsp90 through tetratricopeptide repeat domains (3,4). CHIP has been shown to be in native complexes with all three NOS isoforms as follows: neuronal NOS (nNOS), endothelial NOS, and inducible NOS (5)(6)(7)(8). Although CHIP is not the only ubiquitin ligase for NOS, as Parkin was found to ubiquitinate nNOS (9) and SPSB2 was found to direct the ubiquitination of inducible NOS independent of CHIP (10), the CHIP-dependent pathway appears to be the major one involved in the recognition and ubiquitination of altered dysfunctional forms of nNOS (11,12). The discovery that dysfunctional forms of nNOS were selectively culled came from in vitro and in vivo studies on the suicide inactivation of nNOS (13)(14)(15)(16). The suicide inactivation caused the covalent alteration of the heme or tetrahydrobiopterin of nNOS, and this has led to the concept that alteration of the structure of the active site cleft of nNOS triggers ubiquitination and degradation of the enzyme (5,17). Our laboratory has shown that the inactivated nNOS is preferentially ubiquitinated in an in vitro system containing purified Hsp70 and CHIP (6). More recently, greater levels of Hsp70 and CHIP were found associated to nNOS when ubiquitination was enhanced (18). Interestingly, N G -nitro-L-arginine, but not the D-isomer, protects against nNOS ubiquitination and decreases the levels of Hsp70 and CHIP found in complex with nNOS (18). Moreover, in studies where the individual domains of nNOS were expressed in HEK293T cells, Hsp90 and Hsp70 were found only in complexes with the oxygenase domain (18).
To further understand how the Hsp70/CHIP recognize and triage dysfunctional nNOS, in this study we examined where CHIP ubiquitinates nNOS. To our knowledge, the sites of nNOS ubiquitination are not known. We have identified a dozen sites on nNOS that are ubiquitinated by CHIP with the use of LC-MS/MS methods, and all but one of these sites was found on the oxygenase-calmodulin domain of nNOS. We show that lysine residues within the calmodulin-binding domain are important in the CHIP-mediated poly-ubiquitination of nNOS, and this site regulates degradation of nNOS in HEK293 cells. Lysine residue 739 was found to be the predominant site responsible for the ubiquitination and degradation of nNOS. These studies further support the notion that the Hsp70/CHIP interaction with the oxygenase domain is important in maintaining nNOS protein quality control.

EXPERIMENTAL PROCEDURES
Materials-(6R)-5,6,7,8-Tetrahydro-L-biopterin was purchased from the Schirck laboratory (Jona, Switzerland). Protein A-Sepharose, ubiquitin, ATP, creatine phosphokinase, L-arginine, N G -nitro-L-arginine, N G -nitro-D-arginine, and rabbit polyclonal anti-nNOS IgG were purchased from Sigma. Goat antirabbit IgG IRDye was purchased from Licor Biosciences (Lincoln, NE). His-tagged ubiquitin activating enzyme (E1) was from R&D Systems (Minneapolis, MN). Creatine phosphate was from Fluka (St. Louis, MO). Complete mini protease inhibitor mixture was from Roche Diagnostics. Nickel-nitrilotriacetic acid-agarose was from Qiagen Inc. (Valencia, CA). The cDNA for rat nNOS was kindly provided by Dr. Solomon Snyder (Johns Hopkins Medical School, Baltimore, MD). The cDNA for His-HA-tagged ubiquitin was from Dr. Yi Sun (University of Michigan). The cDNA for expressing the GST-tagged UbcH5a (E2, ubiquitin carrier protein) was kindly provided by C. M. Pickart (Johns Hopkins Medical School, Baltimore, MD). pET30a CHIP plasmid for expressing His-CHIP was kindly provided by C. Patterson (University of North Carolina, Chapel Hill). Nickel magnetic beads and Amicon ultracentrifugal spin filters were from Millipore (Billerica, MA). Untreated rabbit reticulocyte lysate was from Green Hectares (Oregon, WI). DE52 was purchased from Whatman.
Expression and Purification of nNOS, nNOS 7R, nNOS 6R739K, nNOS 6R743K, Hsp70, Hsp40, GST-tagged UbcH5a, His-tagged Ubiquitin, and His-tagged CHIP-To make the 7R mutant nNOS, we mutated all seven lysine residues present in the calmodulin-binding region of nNOS (residues 725, 732, 733, 739, 743, 751, and 754) to arginine residues using the QuikChange multisite-directed mutagenesis kits (Stratagene, La Jolla, CA). The three 5Ј-phosphorylated primers used were 5Ј-CCCCCACGAGGCGGCGAGCTATCGGCTTTAGGAG-ATTGGCAGA-3Ј, 5Ј-GGCCGTCAGGTTCTCAGCCAGGC-TAATGGGACAG-3Ј, and 5Ј-GCCATGGCCAGGAGGG-TCAGGGCGACCATTCTCTAC-3Ј (codons for arginine are underlined). The template was pCWnNOS. The plasmid with the desired mutation was confirmed by sequencing and digested with PflM1 to liberate the 1863-bp fragment and then subcloned into PflM1 sites of the WT-pCWnNOS vector. The resultant construct was designated as pCW7R. We also created a mutant to restore individual native lysine residue 739 or lysine residue 743. This mutant was constructed with the use of a QuikChange II XL site-directed mutagenesis kit (Stratagene, La Jolla, CA). The template was pcDNA7R. The plasmid with the desired mutation was confirmed by sequencing. The plasmid was then digested with Sbf1 and Blp1 to liberate the 3271-bp fragment and subcloned into the WT-pCWnNOS vector. The resultant mutant was designated 6R739K and 6R743K.
The nNOS mutants (including a wild type as control) were bacterially expressed using the pCW vector and BL21 (DE3)competent cells (Stratagene), according to the manufacturer's recommendations. Cells from 1-liter cultures were harvested 48 h after induction with isopropyl 1-thio-␤-D-galactopyranoside and ruptured by French press at 1500 p.s.i. The lysates were processed and purified by 2Ј,5Ј-ADP-Sepharose and PD-10 gel filtration (GE Healthcare) chromatography as described previously (8). Heme was added as an albumin conjugate during the expression to convert all of the nNOS to the holo-nNOS dimer (19). The samples were concentrated with the use of a Centriplus YM-100 concentrator and stored at Ϫ80°C. The nNOS 7R and 6R739K mutants were found to be active (WT nNOS, 192 nmol of NO/min/mg of protein; nNOS 7R, 138 nmol of NO/min/mg of protein; nNOS 6R739K, 125 nmol of NO/ min/mg of protein).
His-CHIP, His-tagged human Hsp70, and His-ubiquitin were bacterially expressed and purified by nickel-nitrilotriacetic acid affinity chromatography as described previously (6). YDJ-1, the yeast ortholog of Hsp40, was expressed in bacteria and purified by sequential chromatography on DE52 and hydroxylapatite as described previously (20). GST-tagged UbcH5a was bacterially expressed and purified by GSH-Sepharose affinity chromatography as described (21).
In Vitro Ubiquitination of nNOS by Purified Ubiquitinating System-The nNOS was pretreated with Hsp70 and Hsp40 and then ubiquitinated. In these experiments, 60 M Hsp70 and 3 M Hsp40 were incubated for 30 min at 30°C with 4 M nNOS in a total volume of 20 l of 50 mM Hepes, pH 7.4, containing 100 mM KCl, and 5 mM dithiothreitol (DTT). To conjugate ubiquitin to nNOS, an aliquot (1 l) of this reaction mixture was incubated with a purified system containing an E1 ubiquitin-activating enzyme (0.1 M), an E2 GST-tagged UbcH5a (10 M), His-tagged CHIP (3 M), His-tagged ubiquitin (500 M), 1 mM DTT, 5 mM ATP in an ATP-regenerating system (50 mM ATP, 250 mM creatine phosphate, 20 mM magnesium acetate, and 100 units/ml creatine phosphokinase), expressed as final concentrations, for 1 h (unless otherwise indicated) at 30°C in a total volume of 40 l of 50 mM Hepes, pH 7.4, containing 100 mM KCl, and 5 mM DTT. After incubation, 80 l of sample buffer (5% (w/v) SDS, 20% (v/v) glycerol, 6 mg/ml DTT, and 0.02% (w/v) bromphenol blue in 125 mM Tris-HCl, pH 6.8) was added, and an aliquot (14 l) was loaded for Western blotting.
SDS-PAGE and Western Blotting-After boiling, samples were resolved on 5% (w/v) SDS-polyacrylamide gels and transferred to Immobilon FL membranes (Millipore, Billerica, MA) for 2 h at 100 V. The blot was probed with anti-nNOS (0.01%, w/v) and then a goat anti-rabbit IRDye secondary antibody (0.0067%, w/v), and the immunoreactive bands were visual-ized with the use of the LI-COR Odyssey Imaging System, and the bands were quantified using Image Studio software (LI-COR). The linearity of the detection system was validated with nNOS and ubiquitinated nNOS standards (0.03-0.75 g of nNOS/lane).
Purification of nNOS-Ubiquitin Conjugates Using Nickel Magnetic Beads-For large scale preparation and purification of nNOS-ubiquitin conjugates, incubations were the same as above except that the concentration of nNOS was increased 5-fold to 20 M in a total volume of 800 l of the ubiquitination reaction mixture. After the 1-h incubation at 30°C, the sample was concentrated to 40 l in 100K spin filters (Amicon Ultra) to remove unconjugated ubiquitin, His-tagged CHIP, and other smaller proteins. This sample was then resuspended with 1 ml of equilibration buffer (50 mM sodium phosphate, 300 mM sodium chloride, 5 mM imidazole, pH 8.0) and then added to 400 l of nickel magnetic beads equilibrated with the same equilibration buffer. The sample was rotated with the beads for 30 min at 4°C. The supernatant was then removed, and the sample was washed three times with 1 ml of wash buffer (50 mM sodium phosphate, 300 mM sodium chloride, 10 mM imidazole, pH 8.0). After removal of the final wash, 70 l of SDS-sample buffer was added to the beads. The sample was then boiled, and 65-l aliquots were then loaded onto 5% SDS-polyacrylamide gels and then stained with Coomassie Blue dye. The appropriate nNOS-ubiquitin conjugate band (molecular mass range of 165-175 kDa) was excised from the gel and sent to MS Bioworks (Ann Arbor, MI) for analysis.
Identification of nNOS Ubiquitination Sites-Gel samples were washed with 25 mM ammonium bicarbonate followed by acetonitrile. Samples were then reduced with 10 mM DTT at 60°C followed by alkylation with 50 mM iodoacetamide at room temperature. Samples were then digested with trypsin (Promega) at 37°C for 4 h. Following digestion, samples were quenched with formic acid, and the supernatant was analyzed directly by nano-LC/MS/MS with a Waters NanoAcquity HPLC system interfaced to a ThermoFisher Orbitrap Velos Pro mass spectrometer. Peptides were loaded on a trapping column and eluted over a 75-m analytical column at 350 nl/min; both columns were packed with Jupiter Proteo resin (Phenomenex). The mass spectrometer was operated in data-dependent mode, with MS performed in the Orbitrap at 60,000 FWHM resolution and MS/MS performed in the LTQ. The 15 most abundant ions were selected for MS/MS. The MS/MS data were searched against the SwissProt database using Mascot search engine version 2.4 (Matrix Science, London, UK; version Mascot). Mascot was set up to search the IPI.RAT version 3.75 decoy database assuming the digestion enzyme trypsin with a maximum of two missed cleavages. The precursor and product ion tolerances were 10 ppm and 0.5 Da, respectively. Cyclization of N-terminal glutamine, deamidation of asparagine and glutamine, oxidation of methionine, acetylation of the N terminus, and ubiquitination (Gly-Gly) of lysine residues were specified in Mascot as variable modifications. Carbamidomethylation of cysteines was set as a fixed modification. Mascot DAT files were parsed into Scaffold (Proteome Software, Portland, OR) for visualization and validation using the Prophet tools (22,23). Minimum protein probability was 90%, and peptide probability was 50%; a minimum of two unique peptides was necessary for protein identification. Putative Gly-Gly-modified peptides were confirmed by manual inspection of the raw MS/MS spectra.
In Vitro Degradation of nNOS by Fraction II-DE52-retained fraction II was prepared from rabbit reticulocyte lysate as described previously (24). In studies where nNOS degradation was measured, purified nNOS (2 g) was incubated at 37°C in a total volume of 120 l of 50 mM Tris-HCl, pH 7.4, containing 2 mM DTT, 500 M His-ubiquitin, an ATP-regenerating system, and 2 mg/ml of fraction II. At indicated times, a 25-l aliquot of each sample was taken and quenched with 25 l of sample buffer. Samples were then submitted to SDS-PAGE and Western blotting.
Cell Culture and Transient Transfection-Human embryonic kidney (HEK) 293T cells were cultured in Dulbecco's minimum essential medium supplemented with 10% (v/v) bovine calf serum. Transient transfections of HEK293T cells were carried out with the use of a standard calcium phosphate method as described previously (25) in 6-well plates. The WT, 7R, 6R739K, and 6R743K rat nNOS cDNA was subcloned from PVL1393 (19) into the EcoRI and NotI sites of pcDNA3.1ϩ. Human CHIP cDNA was in pcDNA3 (8). The four types of nNOS cDNA (0.5 g) were cotransfected with CHIP cDNA (0, 0.1, 0.5, and 1 g) into 70 -80% confluent cells such that the total amount of cDNA was kept constant with vector plasmid. Cells were transfected for 48 h and, where indicated, were treated with lactacystin (10 M) for 20 h prior to harvesting. To harvest the cells, the media were remove, and cells were gently washed once with PBS. After removing the wash, 250 l of sample buffer was added to the wells, and the cells were agitated with a pipette tip. Samples were then boiled for 5 min and then subjected to SDS-PAGE and Western blotting. The data were normalized to the levels of ␤-tubulin, and the expression levels of the different forms of nNOS were confirmed to be similar (WT, 1.52 relative units; 7R, 1.49 relative units; 6R739K, 1.42 relative units; and 6R743K, 1.40 relative units). Fig. 1A, incubation of Hsp70-pretreated nNOS with a ubiquitination system containing purified E1, E2, and CHIP gives a time-dependent formation of higher molecular mass nNOS-bands, which we have previously established as nNOS-ubiquitin conjugates (15). Similar bands have been visualized by anti-ubiquitin IgG after immunoprecipitation with anti-nNOS IgG or with the use of 125 I-ubiquitin (15). Based on the estimated molecular mass calculated from the migration of the bands against molecular mass standards, we have labeled the first major band (170 Ϯ 2.3 kDa) above the native nNOS band (161 Ϯ 2.3 kDa) as the mono-ubiquitinated form of nNOS (mono-Ub). The higher mass bands, which appear after longer duration of incubation, are poly-ubiquitinated products (poly-Ub). Although in theory these bands may represent multiubiquitinated products (multiple mono-ubiquitinations), as will be shown below, they are predominantly poly-ubiquitin products. As shown in Fig. 1B, the time-dependent decrease in the nNOS band (closed circles) coincides with the time-dependent increase in the mono-Ub (closed squares) and poly-Ub (closed triangles) products. The sum of the nNOS and nNOS-ubiquitin bands at each time point (Fig. 1B, open squares) remains the same, suggesting that the majority of the nNOS protein is accounted for in each lane. It is important to note that unlike our previous studies where only a few percent of the total nNOS was ubiquitinated (6,11,16), the current system gives a much higher yield (ϳ50%) of ubiquitinated nNOS. As shown in Fig. 1C, the omission of ubiquitin (lane 2) or CHIP (lane 3) completely blocks the ubiquitination reaction. The use of a mutant ubiquitin (K0R), where the lysines used to form poly-ubiquitin chains have been mutated to arginine residues, was found to decrease the higher molecular bands, and we have labeled these as poly-ubiquitinated nNOS (Fig. 1C, lane 4). The mono-Ub band was unchanged as might be expected. The bands corresponding to nNOS and the formation of either mono-Ub or poly-Ub were quantified from the blots in Fig. 1C and is shown in Fig. 1D. The K0R mutant (condition 4, poly-Ub) decreases the poly-ubiquitin nNOS bands to the level seen with ubiquitin or CHIP omitted, clearly validating the nature of these bands. There are bands above the mono-ubiquitinated nNOS band and below the poly-ubiqutinated nNOS bands that are mixtures of poly-and multiubiquitinated products, which we have measured but are not shown here.

Ubiquitination by the Purified System Is Responsive to the State of the Heme/Substrate-binding Cleft of nNOS-Although
it was important to achieve a high yield of nNOS-ubiquitin conjugates for the mapping studies described below, it was equally important to assess the fidelity of the ubiquitination process. We have established that the substrate-binding cleft of nNOS controls the ubiquitination and degradation of the enzyme by Hsp70⅐CHIP (11,14). A diagnostic test of this mechanism is the decreased nNOS ubiquitination caused by L-isomer, but not the D-isomer, of N G -nitro-arginine (NNA) that can be demonstrated in in vitro studies with cell extracts, in intact cells, as well as in rats (11,14,17). As shown in Fig. 2A, ubiquitination by the purified protein system of nNOS pretreated with L-NNA (cf. lane 3 with lane 2) decreases the amount of higher molecular mass nNOS bands, corresponding to both poly-and mono-ubiquitin nNOS conjugates. The biologically inactive D-isomer of NNA has no effect ( Fig. 2A, lane 4), indicating the stereospecificity of the process. The quantification of the nNOS band clearly shows that L-NNA protects the native nNOS from ubiquitination (Fig. 2B, nNOS, cf. conditions 2 with  3), whereas the D-NNA has no significant protective effect (Fig.  2B, cf. conditions 2 with 4). Consistent with the effects on the nNOS, there is a significant decrease in the amount of both mono (Fig. 2B, mono-Ub)-and poly (Fig. 2B, poly-Ub)-ubiquiti- nated nNOS when L-NNA (condition 3), but not D-NNA (condition 4), is present. Thus, the stereospecific protection by the slowly reversible inhibitor, L-NNA, on nNOS ubiquitination in cells has been faithfully recapitulated by the purified ubiquitination system described here.

Identification of nNOS Ubiquitination Site(s) by LC-MS/MS
Analyses-The ubiquitination system described above was scaled up by a factor of 40, and the resulting mixture was filtered through a 100-kDa molecular sieve (Amicon Ultra) to remove unreacted His-ubiquitin and His-CHIP. The His-ubiquitin-nNOS conjugates present in the flow-through fraction were purified by magnetic nickel beads and submitted to SDS-PAGE. The band corresponding to the mono-ubiquitinated nNOS was excised and subjected to in-gel tryptic digestion. The tryptic peptides were analyzed by LC-MS/MS for the diagnostic GGassociated mass shift of 114.04293 Da. This GG addition verifies the ubiquitination as well as the specific lysine residue modified (26). The LC-MS/MS data were analyzed with the use of MASCOT algorithm, which was used to assess the match of the data to the rat protein sequence database. This method factors in the quality of the spectrum, how well the peptide matches the known sequence and accounts for a change in mass corresponding to the modification by GG. Putative Gly-Gly-modified sites were confirmed by manual inspection of the raw MS/MS spectra. As an example, one of the mass spectra that identify lysine residue 739 as a site of ubiquitination is shown in Fig. 3. The upper panel in Fig. 3 shows the theoretical b ions and y ions that should be produced from ubiquitin conjugation of residue 739. The b ions represent the sequence from the N terminus, whereas the y ions represent that from the C terminus. As shown in the lower panel in Fig. 3, the spectrum contains the indicated b 2 to b 9 ions. The sequence of amino acids corresponding to the fragment ions is also shown. The identified y 2 to y 9 ions along with the corresponding amino acid residues are also shown. The lysine residue at 739 has an additional mass of 114 Da, as indicated by the mass difference from y 4 and y 5 or from b 5 and b 6 , which corresponds to the addition of the signature ubiquitin remnant (GG) to the lysine residue. As shown in Table 1, a dozen ubiquitination sites on nNOS were identified by this approach. Nine of these sites are located in the oxygenase domain of nNOS (Table 1, numbers 1-9); two are located in the calmodulin-binding region (numbers 10 and 11), and one is located in the reductase domain (number 12).

Role of Calmodulin-binding Region in Ubiquitination of nNOS-
The lysine residues on NOS within the calmodulinbinding region have been of interest because it has been known that calmodulin binding decreases the in vitro ubiquitination of nNOS (16). Because of the known promiscuity of ubiquitination reactions (27), we constructed a mutant nNOS (7R) where all seven lysine residues found within the calmodulin-binding region were mutated to arginine residues (Fig. 4, 7R). We also constructed knock-in mutants for the two lysines found to be ubiquitinated by the LC-MS/MS analysis above (Fig. 4, 6R739K and 6R743K). We overexpressed and prepared purified 7R nNOS and 6R739K nNOS proteins for in vitro ubiquitination studies with purified proteins. We were unable to obtain purified 6R743K nNOS as overexpression in insect cells produced a clipped form of the protein during expression. Based on the extent of poly-ubiquitination obtained with purified 6R739K nNOS, we did not pursue the preparation of intact 6R743K nNOS further. We did, however, examine the 6R743K nNOS in HEK293T cells as described later. As shown in Fig. 5A, the ubiquitination of the nNOS 7R and 6R739K was compared with the wild type nNOS in the purified ubiquitination system. The polyubiquitination of nNOS 7R mutant is significantly decreased compared with the wild type nNOS (Fig. 5B, poly-Ub,  cf. solid bars with open bars), and this could not be explained by an overall decrease in the extent of ubiquitination as evident by the loss of the nNOS to the same levels (Fig. 5B, nNOS, cf. solid  bars with open bars). Interestingly, the amount of mono-ubiquitinated nNOS generated from nNOS 7R was unchanged (Fig.  5B, mono-Ub, cf. solid bars with open bars). The reintroduction of lysine 739 (nNOS 6R739K) recapitulated the poly-ubiquitination to the level found with the wild type enzyme (Fig. 5B,  poly-Ub, cf. gray bars with open bars). Role of Calmodulin-binding Region in the Degradation of nNOS-We next examined the proteasomal degradation of the wild type nNOS and nNOS 7R mutant in an in vitro system containing fraction II, which is the DE52-retained fraction of reticulocyte lysate that contains ubiquitinating enzymes and the proteasome (Fig. 5C, inset). We have previously shown that nNOS is proteasomally degraded in an ubiquitin-dependent manner in this system (15). The fraction II-dependent proteasomal degradation is also responsive to the nature of the nNOS heme cleft and dependent on ATP, CHIP, and Hsp70 (11). As shown in Fig. 5C, the nNOS 7R is not degraded after a 3-h incubation (squares), whereas ϳ40% of the wild type nNOS (circles) is degraded. The nNOS 6R739K (Fig. 5C, triangles) was proteasomally degraded by fraction II to a level intermediate between wild type and nNOS 7R.
We also examined the CHIP-dependent proteasomal degradation of nNOS in intact HEK293T cells (Fig. 6A). We have previously established that CHIP overexpression decreases the levels of nNOS in HEK293T cells (6). As shown in Fig. 6B, the quantification of nNOS clearly shows that the decrease in nNOS caused by CHIP (closed circles) is greatly attenuated by lactacystin (closed squares), indicating that CHIP causes the proteasomal degradation of nNOS. This is consistent with that found previously (6). As shown in Fig. 6C, we have examined the 7R, 6R739K, and 6R743K nNOS mutants under the same

TABLE 1 CHIP-catalyzed nNOS ubiquitination sites identified by LC-MS/MS analysis
The isolation of ubiquitinated nNOS and tryptic mapping of the ubiquitination sites by LC-MS/MS is described under "Experimental Procedures." Abbreviation is as follows: ox, oxidation.

No.
Site  conditions. Unlike in insect cells, the 6R743K nNOS was not appreciably clipped, and the intact form was blotted and quantified. As shown in Fig. 6D, the 7R nNOS (closed triangles) and 6R743K nNOS (open circles) were not significantly changed by CHIP overexpression. However, the 6R739K nNOS mutant (Fig. 6D, open squares) was significantly decreased by CHIP overexpression to a level that was similar to the wild type nNOS. The amount of 6R739K and wild type nNOS were not significantly different (p Ͼ 0.05) at any level of CHIP.

DISCUSSION
We show here that the CHIP-mediated ubiquitination of nNOS under in vitro conditions occurs primarily on the oxygenase domain of the enzyme. Of 12 identified ubiquitination sites, nine are located in the oxygenase domain, two are located in the calmodulin-binding region, and one is located in the reductase domain. Unfortunately, there is no known crystal structure of the full-length nNOS, but we have adapted a diagram of nNOS by Garcin et al. (28) to depict the location of the ubiquitination sites. As shown in Fig. 7, one monomer of nNOS is depicted in blue and the other in gray. In Fig. 7, the oxygenase domain (Oxy) is shown with an N terminus (N) placed near the dimer interface based on the location of the most N-terminal residue that has been defined in the oxygenase crystal structure (29). The C terminus of the oxygenase domain is connected to the calmodulin-binding domain (Fig. 7, CaM) that links the oxygenase domain to the reductase domain (Red). The ubiquitination sites appear to be primarily on the oxygenase and calmodulin domains. Based on the primary sequence, the ubiquitination sites in the calmodulin domain border the reductase domain. It is noteworthy that the Hsp70⅐CHIP complex has been shown to recognize primarily the oxygenase domain and that this binding is responsive to the conformation of the hemebinding cleft (18). The sites of ubiquitination are consistent with the notion that Hsp70 recognizes and binds the hemebinding cleft and directs the CHIP-dependent ubiquitination at nearby sites on nNOS (17).
The biological role of the multiple ubiquitination sites found from the in vitro studies is unclear. However, previous studies have demonstrated the ability of calmodulin, in the presence of Hsp90, to block nNOS ubiquitination (16). Moreover, Hsp90 is known to enhance the binding of calmodulin (30), and Hsp90 is known to protect nNOS from ubiquitination (9). Thus, our initial attempts focused on determining the functional significance of the ubiquitination sites within the calmodulin-binding domain. We employed a common strategy to account for the known promiscuity of ubiquitination by use of a mutant (nNOS 7R) where all seven lysines in the calmodulin-binding domain of nNOS were replaced with arginines, which are not able to conjugate to ubiquitin. The nNOS 7R was not poly-ubiquitinated nor degraded in vitro and was not degraded after overexpression of CHIP in intact cells. In contrast, knocking back in the lysine at residue 739 of the nNOS 7R recapitulated the polyubiquitination and degradation of the nNOS. Thus, we conclude that lysine residue 739 of nNOS is a major site of polyubiquitination and signals for proteasomal degradation of the enzyme. Interestingly, the in vitro mono-ubiquitination of the nNOS 7R mutant was similar to the wild type enzyme and indicates that sites apart from those in the calmodulin-binding site account for the mono-ubiquitination of nNOS. The monoubiquitination is a known signal for receptor internalization and endocytic degradation (27). The exact role of mono-ubiquitination of nNOS is not known, although this form is found in relatively high abundance in cells and in cytosolic preparations of rat brain (15).
Although many protein substrates for CHIP-dependent ubiquitination have been identified, there are only a few studies where the actual sites of ubiquitination have been characterized either by mutational analysis (31) or by LC-MS/MS methods (32)(33)(34), and to our knowledge there are none where both approaches were used, as in this study. Correia and co-workers (33,34) have shown that CHIP ubiquitinates CYP2E1 and CYP3A4 in vitro at multiple sites as determined by LC-MS/MS analyses, but no site was examined by mutational methods. Interestingly, the ubiquitination sites appeared to co-cluster with phosphorylation sites, and the notion of a phosphodegron for P450 degradation has been raised (33,34). We do not know if phosphorylation triggers the ubiquitination of nNOS. However, it is noteworthy that an affinity-purified antibody directed against a highly conserved domain comprising the lysine residue 317 to alanine residue 340 of CYP2E1 inhibited both catalytic activity and ubiquitination of the P450 (35). Thus, once again there is a link between the heme-binding cleft and ubiquitination. The authors speculated that this domain likely explains the substrate stabilization of CYP2E1 (35). The importance of the cleft is further supported by the discovery that CHIP mediates ubiquitination of CYP2E1 and that ubiquitination is enhanced after suicide inactivation of the P450 (36,37). Thus, it is highly likely that the heme cleft signals for ubiquitination similar to that found for nNOS. We do not rule out the possibility that cleft opening signals for phosphorylation of P450 enzymes, including nNOS.
In this study, we have characterized the sites of CHIP-dependent ubiquitination on nNOS and identified lysine 739 as the major site for poly-ubiquitination and signal for degradation in cells. The site of poly-ubiquitination is important in understanding how Hsp70⅐CHIP regulates nNOS degradation. Drugs and other xenobiotics greatly impact the degradation of nNOS either by enhancing turnover after suicide inactivation or by decreasing turnover when the natural substrate L-arginine or the slowly reversible inhibitor N G -nitro-L-arginine is present (14). D-Arginine or N G -nitro-D-arginine do not decrease turnover of nNOS indicating stereoselectivity of the process. The exact mechanism for these effects is not known, although they clearly involve interactions at the heme active site and Hsp70⅐CHIP (16). Chaperones regulate client proteins through interactions with ligand-binding clefts, and we have proposed that the conformation about the heme-binding cleft of nNOS is sensed by chaperones to selectively cull altered forms of nNOS for degradation (17,38). Consistent with this notion, N G -nitro-L-arginine decreases the amount of Hsp70 and CHIP bound to nNOS, whereas the N G -nitro-D-arginine has no effect (18). Accordingly, N G -nitro-L-arginine decreases ubiquitination of nNOS in a variety of in vitro and cell systems, decreases the  turnover of the protein in cells, and increases the amount of nNOS protein in vivo (11,14,16,18,39). Defining the sites of ubiquitination should ultimately provide a framework for understanding how the chaperones and CHIP recognize and bind to nNOS to ubiquitinate lysine residue 739 as a signal for degradation of nNOS to maintain protein quality control. It is intriguing to think that ligands that alter the conformation of the active site of nNOS could be developed to alter the amount of nNOS through alteration of chaperone-mediated degradation. The sites of mono-ubiquitination are also regulated by the active site conformation, but the functional role for these modifications is currently not known. Future studies will be directed at understanding the role of mono-ubiquitination in the regulation of nNOS.