From chemistry-request@server.ccl.net Sat Jan 29 00:14:59 2000 Received: from nucleus.harvard.edu (nucleus.harvard.edu [140.247.90.196]) by server.ccl.net (8.8.7/8.8.7) with ESMTP id AAA10436 for ; Sat, 29 Jan 2000 00:14:58 -0500 Received: from localhost (daizadeh@localhost) by nucleus.harvard.edu (8.9.3/8.9.3) with ESMTP id XAA26159 for ; Fri, 28 Jan 2000 23:10:07 -0500 (EST) Date: Fri, 28 Jan 2000 23:10:07 -0500 (EST) From: Iraj Daizadeh To: chemistry@ccl.net Subject: alpha-helix calculations (fwd) Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hello. I have recently asked the pdb list (pdb-l@rcsb.org) the same question concerning alpha-helix stability in water; the complete summary including their replies is shown below. Thanks again for the contributions and in general to those interested in this question. Iraj. xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Question: On Sat, 22 Jan 2000, Iraj Daizadeh wrote: > > Dear Members of the CCL: > > Here are some questions that I have been interested in for some time; your > thoughts would be appreciated. > > Here's the general problem: Given a alpha helix cut out from a PDB > protein, run md simulations to determine the folling: > > 1. How can we ensure from md simulations that the alpha helix is stable in > solution? We solvate the alpha helix the run md simulations for how long? > how much water [10 angstrom solvation shell?]? how many time steps? how > long will the calculation take? and on what machine-- Cray? > etc....details--details--details! > > 2. References showing that results of the above calculation have been > verified by experiment? References describing results of the above > calculations on such small subsets of PDB data? > > I ( as usual ) will post a summary of your responses. > > Thanks, Iraj. > xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Quite a feq MD simulations have been done on alpha-helices. I will point you to my own older work that has all the references: @Article{Spoel96a, author = {David van der Spoel and Bert L. de Groot and Steven Hayward and Herman J. C. Berendsen and Hans J. Vogel}, title = {Bending of the Calmodulin Central Helix: A Theoretical Study}, journal = "Prot. Sci.", year = 1996, volume = 5, pages = {2044-2053} } @Article{Spoel96b, author = {David van der Spoel and Hans J. Vogel and Herman J. C. Berendsen}, title = {Molecular Dynamics Simulations of {N}-terminal Peptides from a Nucleotide Binding Protein}, journal = "Proteins", year = 1996, volume = 24, pages = {450-466} } Especially this latter has a systematic search of peptides to test for stability. If you want to do calculations yourself: just get a couple of Pentium IIs, you can certainly do 1 or 2 ns a week per processor. You should not use a plastic bag (sphere) with water, since your helices will unfold and occupy a large volume. Use normal PBC, with a cubic or truncated octahedron box. Groeten, David. ________________________________________________________________________ Dr. David van der Spoel Biomedical center, Dept. of Biochemistry s-mail: Husargatan 3, Box 576, 75123 Uppsala, Sweden e-mail: spoel@xray.bmc.uu.se www: http://zorn.bmc.uu.se/~spoel phone: 46 18 471 4205 fax: 46 18 511 755 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx On Sat, 22 Jan 2000, Iraj Daizadeh wrote: > Here's the general problem: Given a alpha helix cut out from a PDB > protein, run md simulations to determine the folling: > > 1. How can we ensure from md simulations that the alpha helix is stable in > solution? I'd say you can't, not with 100% certainty; but with a good deal of work you might be able to develop a protocol with some degree of reliability, enough to be useful. One of the dilemmas in this area is evaluating the free energy of the folded state vs. the unfolded state, in large part because there are so many possible unfolded states. > We solvate the alpha helix the run md simulations for how long? Many nanoseconds, at least. You may need to determine that yourself, by testing against cases with known results, e.g. a couple cases where the helix is stable in water in expt, and a couple cases where it isn't. You should probably also deliberately refold the peptide into random coil configs (molten globule?) and subject it to the same protocol. > how much water [10 angstrom solvation shell?]? I recommend constant P,T methods in a periodic cell, with at least 12 A (typical nonbond cutoff) from the solute (protein) to the closest cell face. I wouldn't recommend using a sphere of water in general, and certainly not without employing e.g. a reaction field method. > how many time steps? Probably at least 10 ns per sim, but you should run sims of helices which are unstable when excised to determine that. > how long will the calculation take? A good Ph.D. project, or a very talented postdoc > and on what machine-- Cray? Whatever you can get access to-- considerable CPU time will be needed. Cray if you have the dollars or grant writing skills needed, but CHARMM on a Linux cluster would work for this as well. > 2. References showing that results of the above calculation have been > verified by experiment? References describing results of the above > calculations on such small subsets of PDB data? I know this has been done for paired helices, esp. the leucine zipper motif; the peptide GCN4-p1, 33 AA from the GCN4 protein, has been studied extensively by expt (X-ray crystal, NMR) and MD/MC techniques. > how long will the calculation take? A good Ph.D. project, or a very talented postdoc > and on what machine-- Cray? Whatever you can get access to-- considerable CPU time will be needed. Cray if you have the dollars or grant writing skills needed, but CHARMM on a Linux cluster would work for this as well. > 2. References showing that results of the above calculation have been > verified by experiment? References describing results of the above > calculations on such small subsets of PDB data? I know this has been done for paired helices, esp. the leucine zipper motif; the peptide GCN4-p1, 33 AA from the GCN4 protein, has been studied extensively by expt (X-ray crystal, NMR) and MD/MC techniques. Another area studied extensively by both expt and simulation techniques is helix bundles, including some with designed helices. I'm not as aware of the work on single, isolated helices; however, some of the dimer or bundle literature may have refs to single helix work. -- Rick Venable =====\ |=| "Eschew Obfuscation" FDA/CBER Biophysics Lab |____/ |=| Bethesda, MD U.S.A. | \ / |=| ( Not an official statement or Rick_Venable@nih.gov | \ / |=| position of the FDA; for that, http://www.erols.com/rvenable \/ |=| see http://www.fda.gov ) xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Dear Iraj, I cannot answer all your questions, but can give input on a couple. >1. How can we ensure from md simulations that the alpha helix is stable in >solution? > First of all, in general, I expect a free alpha helix NOT to be stable in aqueous solution - my argument is that the main-chain H-bonds should be able to form lower energy hydrogen bonds with water instead of its own chain elements if totally unconstrained. I suppose you could do a quick assessment by doing mechanics calc in water at various degrees of helix opening to assess energy changes to specifically address this. >how much water [10 angstrom solvation shell?]? Unfortunately, I think a stand alone solvation shell will only simulate a shell of water in a vacuum, which will be very different than in bulk water. I believe you need to use repeat boundary conditions with full hydration if you wish to simuate behavior in an aqueous environment. Best wishes, Bob Latour =========================================== Robert A. Latour Jr., Ph.D. Associate Professor of Bioengineering and Materials Science & Engineering 501 Rhodes Research Center Clemson University, Clemson, SC 29634 robert.latour@ces.clemson.edu tel: (864) 656 -5552 / fax: (864) 656-4466 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Dear Iraj, Yesterday I emailed you my opinion about your questions on alpha helix modeling and stability. The best reference that I have found on this is cited below - this paper does directly address alpha-helix stability and dynamics modeling in aqueous solution. M.J. Bodkin and J.M. Goodfellow, Competing interactions contributing to a-helical stability in aqueous solution, Protein Science, 4: 603-612 (1995). Regards, Bob Latour =========================================== Robert A. Latour Jr., Ph.D. Associate Professor of Bioengineering and Materials Science & Engineering 501 Rhodes Research Center Clemson University, Clemson, SC 29634 robert.latour@ces.clemson.edu tel: (864) 656 -5552 / fax: (864) 656-4466 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx >from pdb list: Dear Iraj Daizadzeh, You might try a medline search for the work of Louis Serrano and collaborators. They publish their opinion, that their AGADIR (Biopolymers 1997 Apr 15;41(5):495-509) formalism is quite well suited to estimate sequence-stability relationships in helices. Perhaps you do not really need an explicit simulation for your question. You should also be aware of the excellent work Barry Honig has contributed to the subject (e.g. J Mol Biol. 1995 Sep 22;252(3):351-65. ). For me, the acid test of whether a model or simulation is worth one's time is the question of whether the result can be experimentally verified. Very few explicit MD simulations score well on this criterion. Sincerely, Boris Steipe +---Dr. Boris Steipe---------------+ | Genzentrum | | Feodor-Lynen Str. 25 Tel +49 (0)89 2180-6987 | | 81377 Muenchen, Germany Fax -6999 | +------+ xxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxx Hi Iraj, I think your question is quite interesting. People have done MD simulations of alpha helices taking them out from proteins. For example, All the seven helices of bacteriorhodospin have been simulated. The simulations have been carried out on single helices. There are reports on molecular dynamics simulations on proline containing helices as well. I would probably search Medline on these to find protocols. Several helix bundles and leucine zippers have been simulated as well. I think quite a work has been done along these lines as it directly relates to secondary structure formation and protein folding. Warm regards, Sandeep Sandeep Kumar, Ph.D. Laboratory of Experimental and Computational Biology, NCI - FCRDC, Bldg. 469, Room 151, Frederick, MD 21702, USA. Phone: 301 846 6542 (Work) FAX : 301 846 5598 Email: kumarsan@ncifcrf.gov URL : http://www-lecb.ncifcrf.gov/~kumarsan xxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxx In addition to the references listed you may want to take a look at the work of Tirado-Rives and Jorgensen (Biochemistry late 80's or early 90's), Daggett and Levitt (JMB about the same time). C. Brooks also published on this a bit later in JMB. Regarding experimental systems, perhaps the most convenient is the "AQ" peptide studied by Baldwin and coworkers (JACS late 80's early 90's), this is a 13- or 14-mer with no formal charges. The same group studied a similar "AK" peptide (Marqusee & Baldwin PNAS 88 or 89), but this may be harder to study computationally. There is also some work by an Austrian or German group in JMB looking at the effect of cut-offs and electrostatics treatment in simulations of alpha-helices, worth tracking down. This paper anticipated some of the recent excitement on the use of rapid schemes for treatment of Ewald sums (Darden) for simulations of proteins and DNA. More recently, look at the work of S. Wang of Georgetown U. on several versions of simulations on the aforementioned "AQ" peptide (and early work by Wang and Sung). A 98 or 99 paper in J. Phys. Chem. used a modified version of CHARMM to study the peptide in a water box. Best regards, Max *********************************** Max Vasquez, PhD Protein Design Labs, Inc. 34801 Campus Drive Fremont, CA 94555 xxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxx From chemistry-request@server.ccl.net Sat Jan 29 15:07:28 2000 Received: from sunchem.chem.uga.edu (sunchem.chem.uga.edu [128.192.5.2]) by server.ccl.net (8.8.7/8.8.7) with ESMTP id PAA14709 for ; Sat, 29 Jan 2000 15:07:28 -0500 Received: (from kmarkey@localhost) by sunchem.chem.uga.edu (8.9.1/8.9.1) id OAA13777 for CHEMISTRY@ccl.net; Sat, 29 Jan 2000 14:00:13 -0500 (EST) From: Kristan Markey Message-Id: <200001291900.OAA13777@sunchem.chem.uga.edu> Subject: CCL:Excited States Methods To: CHEMISTRY@ccl.net Date: Sat, 29 Jan 100 20:00:12 +0100 (MEST) X-Mailer: ELM [version 2.4 PL21] Content-Type: text Hi, The Encyclopedia for Computational Chemistry, Paul v. Rague Schleyer, Norman Allinger, Tim Clark, Johann Gasteiger, Peter A. Kollman, Henry F. Schaefer, Peter R. Schreiner (Eds.), John Wiley & Sons, Ltd. Chichester 1998., provides in-depth articles on a variety of methods, including ones with which to calculate excited states. kristan -- ============================================================================== Kristan Markey Schreiner Group Tel: 706 583 0628 Center for Comp. Quant. Chem. FAX: 705 542 9454 University of Georgia kmarkey@gwdg.de Athens, GA 30602 http://www.gwdg.de/~kmarkey USA It is easier to be a "patriot" than to render your own country its proper due; it is easier to be a "civic leader" than to make your community a better place to live in; it is easier to be a "humanitarian" than to treat your own family with loving understanding; for the smaller the focus of attention, the harder the task. -- Sydney J. Harris =============================================================================