From owner-chemistry-: at :-ccl.net Fri Nov 10 11:17:58 1995 Received: from bedrock.ccl.net for owner-chemistry |-at-| ccl.net by www.ccl.net (8.6.10/950822.1) id LAA18446; Fri, 10 Nov 1995 11:09:31 -0500 Received: from gtc05f for vis #at# gensia.com by bedrock.ccl.net (8.7.1/950822.1) id LAA28580; Fri, 10 Nov 1995 11:09:28 -0500 (EST) Received: from genny (genny.res.gensia.com) by gtc05f (5.x/SMI-SVR4) id AA12419; Fri, 10 Nov 1995 08:04:13 -0800 Received: by genny (940816.SGI.8.6.9/930416.SGI.AUTO) id LAA02691; Fri, 10 Nov 1995 11:01:08 GMT Date: Fri, 10 Nov 1995 03:01:07 +48000 From: Viswanadhan To: chemistry*- at -*ccl.net Subject: Re: Zinc parametrization Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi Netters: With regard to my question on parametrizing Zinc (vdW radius & charge) at the active site of an enzyme, I received a number of responses. Several refer to the papers of Merz & coworkers, but these authors do not use CHELPG and Gaussian to get the charges, and they constrain the residues. If the metal ion and the neighboring residues are constrained, it really does not matter that much what radius and charge are used in the simulation (to preserve the active site geometry though the final energies depend on these parameters). The charges on the interacting residues that coordinate with Zinc also need to be altered when deriving a model. I thank all of those who responded to my question. Here are the responses: ---------------------------------------------------------------------------------------------------- Vis: Goldschmidt (Naturwissenschaften 14:477, 1926 I think) gives the empirical radius of Zn2+ to be 0.83 ang while Pauling (JACS 49:763, 1927) computes 0.74 ang. For the CHELPG method, I don't believe that it hurts to choose a radius too large, but do not choose one that is too small! The important thing is the charge "felt" by neighboring atoms. So, choose a radius slightly smaller than the closest approach that you would expect from any other atoms. From the literature, the Zn-O distances in the first hydration shell are about 2.0 to 2.2 angstroms (see Chem Rev 93:1173, 1993), so I should think that your 1.39 angstrom value would work fine. Best of luck, Brian J. Teppen teppen*- at -*srel.edu Advanced Analytical Center for Environmental Sciences Savannah River Ecology Laboratory University of Georgia Drawer E Aiken, SC 29802 On Nov 6, 4:49am, Viswanadhan wrote: > Subject: CCL:G:Parametrization for Zinc / charges & vdW radius > I would like to use CHELPG charges which necessitates the use of right > vdW radii. For neutral Zinc, the Bondi radius (J. Phys. Chem. 68. 441) > is 1.39 A, which is clearly not appropriate. Can any one point to recent > references or suggest a method? I will be happy to summarize the > responses (with regard to vdW radii & modeling the charge in the active > site of an enzyme for the Zinc ion). Merz and coworkers have addressed the problem of zinc parameters for their study of human carbonic anhydrase II, JACS, 1991, 113, 8262-8270. Dave Huhta, now at Biosym/MSI, I believe, also worked on this issue with a different approach, JACS, 1990, 112, 4759-4767. Lee Bartolotti addressed the issue of the 6-12 potentials in J. Comp. Chem., 1991, 12(9), 1125-8. A formidable problem...best of luck. Regards, James jbrown ^%at%^ iris7.carb.nist.gov ------------------- Dr. Viswanadhan - While I offer no solutions, I do offer a caution: I would be careful when using CHELPG charges for the system you have described. In this system, it seems to me that the zinc ion would be fairly well buried. CHELPG produces charges which reproduce the electrostatic potential outside the molecule. Unfortunately, atoms that are buried inside the molecule do not affect the electrostatic potential much. The result is that CHELPG charges for buried atoms can vary over a wide range without changing the outcome much. This often leads to unreasonable charges on the buried atom. Dave Giesen Hi Vish, Sal Profeta here. Hope you and the group out there are doing well. I think the person who can best answer your query is Kenny Merz at Penn State. He's done a lot of work with zinc-containing systems within AMBER, and has published a few papers in JACS and JCC in the last 5 years on this. According to the Directory of Grad REsearch his eamil is: merz- at -retina.chem.psu.edu Good luck! Cheers, Sal ++++++De mortuis nil nisi bonum.++++++++++* Dr. Salvatore Profeta, Jr. Computational Polymorphism, Ltd. Research Computing & Consulting 104 Cottage Lane Durham, NC 27713-9387 (919) 544-2299/5304 (Fax) email: profeta~s \\at// glaxo.com (courtesy acct) ++++++Sic transit gloria mundi++++++++++++ Viswanadhan wrote: : : : I would like to use CHELPG charges which necessitates the use of right : vdW radii. For neutral Zinc, the Bondi radius (J. Phys. Chem. 68. 441) : is 1.39 A, which is clearly not appropriate. Can any one point to recent : references or suggest a method? I will be happy to summarize the : responses (with regard to vdW radii & modeling the charge in the active : site of an enzyme for the Zinc ion). : : Vellarkad N. Viswanadhan : Yes, what does constitute "good" vdW radii. In our work on Fe-S centers in proteins we've noticed that the actual charges calculated by CHELPG vary significantly depending on whether you use Breneman&Wiberg radii or Gavizotti&Spackman radii. Generally, Breneman & Wiberg based charges seem "better behaved" in our systems. Also, we've noticed that many groups use radii designed for other purposes and "plug" these into the CHELPG algorithm. What are the rationals people use in choosing their radii? -Brian -- ============================================================================= | .---------.| Brian W. Beck | E-mail Addresses: | |/\ | ||--------------------| brian;at;bert.chem.wsu.edu | || \\ WSU || Biochem/Biophysics | brian_beck \\at// wsu.edu | |\ - *|| WSU , 206 Fulmer | URL http://elmo.chem.wsu.edu/~brian | | | || Pullman, WA | VOICE (509) 335-4083 | | \___________|| 99164-4660 | FAX (509) 335-9688 | ============================================================================= Hi, You might want to look at the following reference in which charges and force field params for Zinc and for surrounding atoms are determined for the zinc finger in the DBD of the glucocorticoid receptor. M.A.L. Eriksson, T. Hard and L. Nilsson (1995) Biophys. J. vol68, pp 402-426. Molceular dynamics simulations of the glucocorticoid receptor DNA-Binding Domain in complex with DNA and free in solution. Charges were obtained from full geometry optimization of model system in the MOPAC/MNDO/ESP calcn. The Lennard-Jones params for Zn were taken from K. Merz (1991) J. Amer. Chem. Soc. 113:406-411. Hope this helps Laurent ********************************************************************** ** Laurent Chiche !! e-mail: ** ** !! chiche*- at -*cbs.univ-montp1.fr ** ** Centre de Biochimie Structurale !! ** ** Faculte de Pharmacie !! Tel: (33) 67.04.34.32 ** ** 15 Ave. Charles Flahault !! ** ** 34060 Montpellier - France !! Fax: (33) 67.52.96.23 ** **********************************************************************