FEBS Letters 587 (2013) 297–301

journal homepage: www.FEBSLetters .org

Calmodulin-induced structural changes in endothelial nitric oxidesynthase

0014-5793/$36.00 � 2012 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.febslet.2012.12.012

Abbreviations: NOS, nitric oxide synthase; eNOS, endothelial NOS; iNOS, nNOS,inducible and neuronal NOS; CaM, calmodulin⇑ Corresponding author.

E-mail address: persechinia@umkc.edu (A. Persechini).

Anthony Persechini ⇑, Quang-Kim Tran, D.J. Black, Edward P. GogolDivision of Molecular Biology and Biochemistry and Division of Cell Biology and Biophysics, University of Missouri at Kansas City, 5007 Rockhill Rd, Kansas City,MO 64110-2499, United States

a r t i c l e i n f o a b s t r a c t

Article history:Received 20 November 2012Revised 10 December 2012Accepted 11 December 2012Available online 22 December 2012

Edited by Peter Brzezinski

Keywords:Nitric oxide synthaseCalmodulinEnzyme regulationEnzyme structure

We have derived structures of intact calmodulin (CaM)-free and CaM-bound endothelial nitric oxidesynthase (eNOS) by reconstruction from cryo-electron micrographs. The CaM-free reconstruction iswell fitted by the oxygenase domain dimer, but the reductase domains are not visible, suggestingthey are mobile and thus delocalized. Additional protein is visible in the CaM-bound reconstruction,concentrated in volumes near two basic patches on each oxygenase domain. One of these corre-sponds with a presumptive docking site for the reductase domain FMN-binding module. The otheris proposed to correspond with a docking site for CaM. A model is suggested in which CaM bindingand docking position the reductase domains near the oxygenase domains and promote docking ofthe FMN-binding modules required for electron transfer.� 2012 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

2+ 2+

1. Introduction

The nitric oxide synthases (NOS) catalyze formation of NO andL-citrulline from L-arginine and oxygen, with NADPH as the electrondonor [1]. The three major mammalian forms of the enzyme arecommonly referred to as iNOS (inducible), nNOS (neuronal) andeNOS (endothelial) [2,3]. All of these are functional homodimersof 130–160 kDa monomers. Each monomer contains a reductaseand oxygenase domain joined by a �35 amino acid linker sequencecontaining a calmodulin (CaM)-binding domain [1]. The interfacebetween the two oxygenase domains appears to be responsiblefor maintenance of the enzyme dimer in solution [4–7]. Each reduc-tase domain contains an FMN-binding and FAD–NADPH-bindingmodule, joined by a �20 amino acid linker [8]. Electron transferduring catalysis occurs in trans, with electrons flowing from thereductase domain of one monomer to the heme reaction center inthe oxygenase domain of the other [9–11]. This appears to involvemovement of the FMN modules between their respective NADPH–FAD modules and docking sites on the oxygenase domains [12,13].

All three enzyme isoforms have negligible synthase activity inthe absence of CaM, which is bound with significant affinity toeNOS and nNOS only in its Ca2+-bound form, and to iNOS in both

its Ca -free and Ca -bound forms [14–17]. Various observationssuggest that CaM activates synthase activity both by increasingthe efficiency of electron transfer within the reductase domains,from NADPH to FMN via FAD, and by increasing the efficiency ofelectron transfer to the heme reaction centers via reduced FMN,in part by mobilizing the FMN modules [12,13,18–21]. Althoughcrystal structures have been determined for the dimeric eNOS oxy-genase domain [7], and for a dimeric form of the nNOS reductasedomain [8], no structures have been published thus far for an intactNOS isoform.

In this paper we present solution structures of full-length CaM-bound and CaM-free bovine eNOS at a nominal resolution of�25 Å,derived by reconstruction from cryo-electron micrographs of theenzyme in vitreous ice. These structures suggest significant newinsights to the structural relationship between the reductase andoxygenase domains, and how this relationship is affected by CaMto produce synthase activation.

2. Materials and methods

A mutant of bovine eNOS containing a phosphomimetic S1179Dsubstitution was used for these investigations because its maximalCaM-dependent synthase activity is twice that of the native pro-tein, suggesting that in the presence of CaM more is in the fully ac-tive conformation [22]. This protein was expressed in Escherichiacoli and purified as described elsewhere [22,23]. The vertebrateCaM amino acid sequence, encoded by a rat cDNA, was expressed

298 A. Persechini et al. / FEBS Letters 587 (2013) 297–301

in E. coli and purified as described previously [24]. Immediatelyprior to preparation of samples for microscopy, 50 ll aliquots ofpurified eNOS were thawed and analyzed by size exclusion chro-matography on a Superdex 200 HR 10/30 column at 4 �C in a buffercontaining 25 mM Tris–HCl, pH 7.4, 100 mM KCl, 1 mM CaCl2 and1 mM dithiothreitol. Peak fractions previously shown to corre-spond with the intact dimeric enzyme were pooled, and the mono-mer concentration of eNOS was determined based on opticalabsorbance at 397 nm [23]. Prior to freezing on grids, the enzymewas diluted to a concentration of 30–150 nM in column buffer,with or without a 1.5-fold molar excess of (Ca2+)4–CaM. The appar-ent KD for the (Ca2+)4–CaM–eNOS complex is below 1 nM [25], sounder these conditions the enzyme should be saturated with CaM.

Fenestrated carbon films (Quantifoil Micro Tools GmbH) sub-jected to glow-discharge were used for application, blotting andfreezing of proteins in liquid ethane. The samples were stored in li-quid nitrogen until loading into a Gatan 626 holder and imagingwith a JEOL 1200 IIX electron microscope at 100 kV, using minimaldose protocols. Micrographs were recorded on Kodak SO163 film,using defocus values between 1.2 and 3 lm, and digitized usinga Hi-Scan drum scanner with a 5 Å pixel on the specimen. Individ-ual particles were selected from images wavelet-filtered to in-crease contrast, and the coordinates thus obtained were used toextract unfiltered particle images in 40 � 40 pixel (200 � 200 Å)boxes. The CaM-free and CaM-bound eNOS data sets each contain�25,000 images. Phase correction of the particle images was basedon defocus values estimated using the ACE software package [26].

Euler angles were assigned to each image based on projection-matching to de novo common-lines initial models. These were gen-erated from reference-free image averages of both data sets sortedinto classes by iterative multivariate statistical analysis with theEMAN software package [27]. An initial model derived in thismanner for each data set (±CaM) was used to initiate iterativeprojection-matching in 7� angular increments with twofold sym-metry imposed, using the EMAN software. A cutoff for correlationwith model projections eliminated approximately 35% of theparticles from the data. Convergence was reached within five toeight rounds of refinement based on round-to-round resolutioncalculations. To test model dependence, the two initial models(±CaM) were exchanged, and the final reconstructions of eachdataset were visually indistinguishable from those initiated withthe ‘‘correct’’ model. The amplitudes of the reconstructions werecorrected in defocus groups guided by a solution scatteringcurve of a similar-sized protein dimer, fatty acid synthase [28],

Fig. 1. Analysis of reconstructions. (A) Fourier shell correlation (FSC) plots for the CaM-frprojection classes. Nominal resolutions correspond to the reciprocal of the shell whreconstructions. (B & C) Characteristic projections (P) of the CaM-free (B) and CaM-bounddirectly from the data. Projections of the reconstructions before amplitude correctionReference-free class averages were derived from phase-corrected images as described in

at resolutions between 100 and 25 Å. The resolutions of thereconstructions were calculated by comparison of Fourier shellcoefficients (Fig. 1A), both yielding a limit of �25 Å at a 0.5 corre-lation value. Characteristic projections of both reconstructionscompare well with reference-free class averages of the phase-corrected data derived using the refine2d component of the EMANsoftware package (Fig. 1B and C).

Fitting and correlation of a simulated 25 Å resolution densitymap derived from the oxygenase domain dimer crystal structure(PDB ID = 1FOP) [7] was performed using the Chimera moleculargraphics package [29]. The DelPhi software suite [30] was usedto calculate the electrostatic potential surface for the crystal struc-ture displayed in Fig. 1C. A homology model for the eNOS FMNmodule was derived using standard methods from the nNOSreductase domain dimer crystal structure (PDB ID = 1TLL) [8].

3. Results

The final CaM-free and CaM-bound eNOS reconstructions aredisplayed in Fig. 2A and B as volumes enclosed at the level of steep-est density drop-off, which corresponds with the apparent surfaceof the protein. A simulated 25 Å density map (colored red) for theoxygenase domain dimer has been fitted to the CaM-free andCaM-bound reconstructions [7]. The reconstructions and simulatedoxygenase density are also represented in the figure as cross-sectional contour plots taken at the levels indicated in the recon-structions by lines 1–4. The red lines in the contour plots enclosethe simulated density for the fitted oxygenase domain dimer.

Both the CaM-free and CaM-bound reconstructions accommo-date the oxygenase dimer crystal structure, but neither can accom-modate the reductase domains, which correspond to �60% of the260 kDa mass of the eNOS dimer. An excellent correspondencecan be obtained between the CaM-free reconstruction and the sim-ulated oxygenase domain density (correlation = 0.93). However,the reductase domains appear to have been lost by the image aver-aging inherent in the reconstruction process, indicating that theyare highly mobile with respect to the oxygenase domains. Thesmall amount of volume near the the ‘‘top’’ of the CaM-freereconstruction that is unaccounted for by the simulated oxygenasedensity may correspond with �65 amino acids missing from theN-termini of the crystal structure [7].

In the case of the CaM-bound eNOS, the highest-density portionsof the reconstruction also appear to match the simulated oxygenasedensity, but extensive peripheral density cannot be accounted for by

ee (s) and CaM-bound (d) reconstructions derived from odd and even images in theere the FSC value equals 0.5, which yields a limiting value of �25 Å for both(C) eNOS reconstructions compared with reference-free class averages (RF) derived

were generated after aligning them (corr = 0.96) using the Chimera software [29].Section 2.

Fig. 2. 3-D representations and cross-sectional contour plots of eNOS reconstruc-tions. (A and B) Surface representations of CaM-free (A) and CaM-bound eNOSreconstructions are shown in gray in the first row. A simulated density map for theoxygenase domain dimer fitted into the reconstructions is represented in red. Cross-sectional contour plots representing evenly spaced densities are also presented,taken at the levels indicated in by the numbered lines. The fitted simulatedoxygenase density is outlined in red. (C) The CaM-bound reconstruction displayedwith an electrostatic potential surface derived from the fitted oxygenase domaindimer illustrating the basic patches (colored blue) overlapped by volumes f and c.

A. Persechini et al. / FEBS Letters 587 (2013) 297–301 299

CaM alone, which at 16.8 kDa has only �10% of the mass of a syn-thase monomer. Most of this additional density is therefore likelyto correspond with the reductase domains, suggesting they are lessmobile in the presence of CaM. It appears to be concentrated in twovolumes on the surface of each oxygenase monomer, labeled f and cin Fig. 2B. As seen in Fig. 2C, volume f overlaps a basic patch on theoxygenase domain that has been proposed to participate in dockingof the FMN module during catalysis [12,13,21,31,32]. Volume c over-laps a second basic patch, which we propose corresponds with adocking site for bound CaM. In spite of these density concentrations,the diffuse nature of much of the additional density visible in theCaM-bound reconstruction suggests that significant portions ofthe reductase domains remain mobile.

4. Discussion

The low-resolution structures of intact eNOS presented heredemonstrate that the oxygenase domain portion of the eNOS dimercan be readily identified in solution by its correspondence to theX-ray crystal structure of the oxygenase domain dimer. However,the reductase domains, which comprise approximately 60% of theeNOS molecule, appear to be mobile with respect to the oxygenasedomains, especially in the absence of CaM (Fig. 2A). Indeed,mobility in the absence of CaM is sufficient to result in diffusionof almost all the reductase domain density by the reconstructionprocess. The apparent mobility of the reductase domains is presum-ably permitted by the �35 amino acid linker between reductaseand oxygenase domains, which could span an average distance ofas much as �50 Å [33]. A crystal structure for a dimeric form ofthe nNOS reductase domain suggested a model for the intactenzyme in which oxgenase and reductase domain dimers associatein a stacked arrangement [8]. Our results are not consistent withsuch a model for eNOS. They instead suggest that while theoxygenase domain dimer is maintained in solution, the reductasedomains are highly mobile with respect to it.

Additional protein visible in the CaM-bound reconstruction islikely to correspond to CaM and portions of the reductase domains.At the resolution of this analysis, we cannot definitively identifythe elements that comprise these structural additions. However,volume f is posited as the FMN module, due to its overlap with abasic patch on the oxygenase domain previously proposed to par-ticipate in CaM-dependent docking of this module (Fig. 2C) [21].Volume c overlaps an additional basic patch on the oxygenase do-main (Fig. 2C). A crystal structure determined for the complex be-tween CaM and an iNOS fragment containing the FMN module andCaM-binding domain indicates that bound CaM and FMN moduleare separated by a short �7 amino acid linker, a feature that ap-pears to be conserved in the synthases [31]. Thus, based on itsproximity to the presumptive FMN module, as well as its size, vol-ume c is consistent with bound and docked CaM. Docking of CaMin this manner would be expected to promote what otherwise ap-pears to be an intrinsically weak interaction between the FMNmodule and the adjacent oxygenase domain [12,13,34].

Molecular modeling was performed to investigate the structuralfeasibility of our interpretations of volumes f and c in the CaM-bound reconstruction (Fig. 3). For the sake of clarity only one ofthe two symmetry-related pairs of f and c volumes is considered.Three orthogonal views of the model are presented alone (D–F)and superimposed on a surface representation of the CaM-boundeNOS reconstruction (A–C). The FMN module is positioned toachieve a reasonable correspondence with volume f on the surfaceof the green oxygenase monomer. It is depicted in magenta to indi-cate that it is continuous with the magenta oxygenase domain, andis therefore positioned in the model for the requisite transfer ofelectrons in trans to the heme in the adjacent oxygenase domain(colored green). The 18.9 Å distance between FMN and the reaction

Fig. 3. Molecular modeling of oxygenase and projections in CaM-bound eNOS. Orthogonal views of the fitted oxygenase domain dimer and modeled FMN module and CaMcomplexes are shown with (A–C) and without (D–F) display of the CaM-bound reconstruction. The oxygenase is depicted as a a-carbon tracing, with the two monomers(residues 67–482 and 69–482) colored green and magenta. The reaction center hemes are displayed as hard sphere models. The FMN module (residues 520–716) is depictedin magenta, to indicate that it belongs to the same monomer as the magenta oxygenase domain. Bound FMN is displayed as a hard sphere model. The CaM-binding domain(residues 494–512) is colored magenta, as it also belongs to the same monomer. Bound CaM is colored cyan, and bound Ca2+ ions are indicated by green spheres. The linker(GTLMAKR) between the FMN module and the CaM-binding domain is modeled as random coil. The linker between the magenta CaM-binding and oxygenase domains(KGSATKGAGIT) is not shown.

Fig. 4. A two-step model for CaM-dependent activation of the constitutive nitricoxide synthases. Aside from the two oxygenase domains (colored magenta andgreen), only one of two symmetry related sets of elements is depicted for the sake ofclarity. Reaction center hemes are represented by diamonds. Auto-inhibitorysequences in the reductase domain that may interact with the CaM-binding domainin the absence of CaM are not represented [12,18]. In step I, binding of Ca2+–CaM (c)activates the reductase domain and releases the FMN module (f) from the NADPH–FAD module (n). In step II, docking of bound CaM with the magenta oxygenasedomain positions the reductase domain as required to promote docking of the FMNmodule with the oxygenase domain in the other synthase monomer (green).

300 A. Persechini et al. / FEBS Letters 587 (2013) 297–301

center heme iron in this domain achieved in the model is similar tothe 18.8 Å distance estimated based on pulsed EPR data [32]. Thecomplex between (Ca2+)4–CaM and the eNOS CaM-binding domainis positioned adjacent to the magenta oxygenase domain in volumec, with the CaM-binding domain colored magenta to indicate itscontinuity with this domain. Bound CaM is colored cyan and boundCa2+ ions are indicated by green spheres. The short linker betweenthe FMN module and the C-terminus of the CaM-binding domain ismodeled as a random coil, demonstrating that it can span the dis-tance between volumes f and c, as required for the FMN module todock with the oxygenase domain in this model (Fig. 3).

Our results are consistent with a minimal two-step model forCaM-dependent synthase activation (Fig. 4). In the first step, CaMbinds a CaM-binding domain in the enzyme, activating the reduc-tase domain and mobilizing the FMN module. In the second step,CaM docks on the oxygenase domain that is continuous with theCaM-binding domain. This promotes docking of the FMN modulewith the adjacent oxygenase domain. The FAD–NADPH module isdepicted as remaining mobile with respect to the oxygenasedomain dimer, and this may account for some of the diffuse

A. Persechini et al. / FEBS Letters 587 (2013) 297–301 301

density observed in the presence of CaM. The idea that CaM acti-vates synthase activity through minimally a two-step process isnot new, although our results support a novel CaM-dockinghypothesis for the second step [12,18,34]. We have previouslydemonstrated that a similar binding and docking mechanism ap-plies to CaM-dependent activation of skeletal muscle myosin lightchain kinase, with the docking step apparently required to allowsubstrate access to the enzyme active site [35,36].

Acknowledgements

The authors would like to thank Benjamin Iwai and LeanneSzerszen for expert technical assistance. This work was supportedin part by NIH grant GM074887 to A.P. and by a University of Mis-souri Research Board grant to E.P.G.

References

[1] Masters, B.S.S., McMillan, K., Sheta, E.A., Nishimura, J.S., Roman, L.J. andMartasek, P. (1996) Neuronal nitric oxide synthase, a modular enzyme formedby convergent evolution – structure studies of a cysteine thiolate-ligandedheme protein that hydroxylates L-arginine to produce NO as a cellular signal.FASEB J. 10, 552–558.

[2] Silvagno, F., Xia, H. and Bredt, D.S. (1996) Neuronal nitric-oxide synthase-mu,an alternatively spliced isoform expressed in differentiated skeletal muscle. J.Biol. Chem. 271, 11204–11208.

[3] Sase, K. and Michel, T. (1997) Expression and regulation of endothelial nitricoxide synthase. Trends Cardiovasc. Med. 7, 28–37.

[4] Ravi, K., Brennan, L.A., Levic, S., Ross, P.A. and Black, S.M. (2004) S-nitrosylationof endothelial nitric oxide synthase is associated with monomerization anddecreased enzyme activity. Proc. Natl. Acad. Sci. USA 101, 2619–2624.

[5] Li, D., Hayden, E.Y., Panda, K., Stuehr, D.J., Deng, H.T., Rousseau, D.L. and Yeh,S.R. (2006) Regulation of the monomer–dimer equilibrium in inducible nitric-oxide synthase by nitric oxide. J. Biol. Chem. 281, 8197–8204.

[6] Stuehr, D.J. (1997) Structure–function aspects in the nitric oxide synthases.Annu. Rev. Pharmacol. Toxicol. 37, 339–359.

[7] Li, H.Y., Raman, C.S., Martasek, P., Masters, B.S.S. and Poulos, T.L. (2001)Crystallographic studies on endothelial nitric oxide synthase complexed withnitric oxide and mechanism-based inhibitors. Biochemistry 40, 5399–5406.

[8] Garcin, E.D. et al. (2004) Structural basis for isozyme-specific regulation ofelectron transfer in nitric-oxide synthase. J. Biol. Chem. 279, 37918–37927.

[9] Siddhanta, U., Wu, C.Q., Abusoud, H.M., Zhang, J.L., Ghosh, D.K. and Stuehr, D.J.(1996) Heme iron reduction and catalysis by a nitric oxide synthaseheterodimer containing one reductase and two oxygenase domains. J. Biol.Chem. 271, 7309–7312.

[10] Panda, K., Ghosh, S. and Stuehr, D.J. (2001) Calmodulin activates intersubunitelectron transfer in the neuronal nitric-oxide synthase dimer. J. Biol. Chem.276, 23349–23356.

[11] Sagami, I., Daff, S. and Shimizu, T. (2001) Intra-subunit and inter-subunitelectron transfer in neuronal nitric-oxide synthase – effect of calmodulin onheterodimer catalysis. J. Biol. Chem. 276, 30036–30042.

[12] Stuehr, D.J., Tejero, J. and Haque, M.M. (2009) Structural and mechanisticaspects of flavoproteins: electron transfer through the nitric oxide synthaseflavoprotein domain. FEBS J. 276, 3959–3974.

[13] Ilagan, R.P. et al. (2009) Regulation of FMN subdomain interactions andfunction in neuronal nitric oxide synthase. Biochem. 48, 3864–3876.

[14] Venema, R.C., Sayegh, H.S., Kent, J.D. and Harrison, D.G. (1996) Identification,characterization, and comparison of the calmodulin-binding domains of theendothelial and inducible nitric oxide synthases. J. Biol. Chem. 271, 6435–6440.

[15] Zhang, M. and Vogel, H.J. (1994) Characterization of the calmodulin-bindingdomain or rat cerebellar nitric oxide synthase. J. Biol. Chem. 269, 981–985.

[16] Anagli, J., Hofmann, F., Quadroni, M., Vorherr, T. and Carafoli, E. (1995) Thecalmodulin-binding domain of the inducible (macrophage) nitric oxidesynthase. Eur. J. Biochem. 233, 701–708.

[17] Spratt, D.E., Taiakina, V., Palmer, M. and Guillemette, J.G. (2007) Differentialbinding of calmodulin domains to constitutive and inducible nitric oxidesynthase enzymes. Biochemistry 46, 8288–8300.

[18] Daff, S. (2003) Calmodulin-dependent regulation of mammalian nitric oxidesynthase. Biochem. Soc. Trans. 31, 502–505.

[19] Craig, D.H., Chapman, S.K. and Daff, S. (2002) Calmodulin activates electrontransfer through neuronal nitric-oxide synthase reductase domain byreleasing an NADPH-dependent conformational lock. J. Biol. Chem. 277,33987–33994.

[20] Haque, M.M., Panda, K., Tejero, J., Aulak, K.S., Fadlalla, M.A., Mustovich, A.T.and Stuehr, D.J. (2007) A connecting hinge represses the activity of endothelialnitric oxide synthase. Proc. Natl. Acad. Sci. USA 104, 9254–9259.

[21] Panda, K., Haque, M.M., Garcin-Hosfield, E.D., Durra, D., Getzoff, E.D. andStuehr, D.J. (2006) Surface charge interactions of the FMN module governcatalysis by nitric-oxide synthase. J. Biol. Chem. 281, 36819–36827.

[22] Tran, Q.-K., Leonard, J., Black, D., Nadeau, O.W., Boulatnikov, I.G. andPersechini, A. (2009) Effects of combined phosphorylation at Ser-617 andSer-1179 in endothelial nitric oxide synthase on EC50(Ca2+) values forcalmodulin binding and enzyme activation. J. Biol. Chem. 284, 11892–11899.

[23] Tran, Q.K., Leonard, J., Black, D.J. and Persechini, A. (2008) Phosphorylationwithin an autoinhibitory domain in endothelial nitric oxide synthase reducesthe Ca(2+) concentrations required for calmodulin to bind and activate theenzyme. Biochemistry 47, 7557–7566.

[24] Persechini, A., Blumenthal, D.K., Jarrett, H.W., Klee, C.B., Hardy, D.O. andKretsinger, R.H. (1989) The effects of deletions in the central helix ofcalmodulin on enzyme activation and peptide binding. J. Biol. Chem. 264,8052–8058.

[25] Tran, Q.K., Black, D.J. and Persechini, A. (2005) Dominant affectors in thecalmodulin network shape the time courses of target responses in the cell. CellCalcium 37, 541–553.

[26] Mallick, S.P., Carragher, B., Potter, C.S. and Kriegman, D.J. (2005) ACE:automated CTF estimation. Ultramicroscopy 104, 8–29.

[27] Ludtke, S.J., Baldwin, P.R. and Chiu, W. (1999) EMAN: semiautomated softwarefor high-resolution single-particle reconstructions. J. Struct. Biol. 128, 82–97.

[28] Brink, J., Ludtke, S.J., Yang, C.Y., Gu, Z.W., Wakil, S.J. and Chiu, W. (2002)Quaternary structure of human fatty acid synthase by electroncryomicroscopy. Proc. Natl. Acad. Sci. USA 99, 138–143.

[29] Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S., Greenblatt, D.M., Meng,E.C. and Ferrin, T.E. (2004) UCSF Chimera–a visualization system forexploratory research and analysis. J. Comput. Chem. 25, 1605–1612.

[30] Li, L. et al. (2012) DelPhi: a comprehensive suite for DelPhi software andassociated resources. BMC Biophys. 5, 9.

[31] Xia, C., Misra, I., Iyanagi, T. and Kim, J.-J.P. (2009) Regulation of interdomaininteractions by calmodulin in inducible nitric-oxide synthase. J. Biol. Chem.284, 30708–30717.

[32] Astashkin, A.V., Elmore, B.O., Fan, W., Guillemette, J.G. and Feng, C. (2010)Pulsed EPR determination of the distance between heme iron and FMN centersin a human inducible nitric oxide synthase. J. Am. Chem. Soc. 132, 12059–12067.

[33] Fitzkee, N.C. and Rose, G.D. (2004) Reassessing random-coil statistics inunfolded proteins. Proc. Natl. Acad. Sci. USA 101, 12497–12502.

[34] Daff, S. (2010) NO synthase: structures and mechanisms. Nitric Oxide 23, 1–11.

[35] Krueger, J.K., Gallagher, S.C., Zhi, G., Geguchadze, R., Persechini, A., Stull, J.T.and Trewhella, J. (2001) Activation of myosin light chain kinase requirestranslocation of bound calmodulin. J. Biol. Chem. 276, 4535–4538.

[36] Persechini, A., Gansz, K.J. and Paresi, R.J. (1996) A role in enzyme activation forthe N-terminal leader sequence in calmodulin. J. Biol. Chem. 271, 19279–19282.