From:	Bruce Allan Palfey <brupalf -AatT- umich.edu>
Date:	Fri, 15 May 1998 09:13:24 -0400 (EDT)
Subject:	LBHBs

Thanks to all of you who replied concerning my question on simulating
Low Barrier Hydrogen Bonds (LBHB) by MM methods.  It seems that the
majority of those responding don't expect good results, and that nothing
short of ab initio calculations will do the job.  Several people
expressed interest in references on LBHBs.  A heated debate has been in
progress in enzymology over the past several years.  A good, concise
overview of the field, with lots of leading references, can be found in
Pan & McAllister (1997) Journal of Organic Chemistry, 62, 8171-8176.

The inability of MM methods to handle LBHBs raises a (potentially)
bothersome point.  The refinement of macromolecular structures in X-ray
crystallography and NMR generally uses an MM force field, e.g., the
CHARMm force field in the ubiquitous X-plor.  Can structures of
molecules that contain LBHBs determined this way be considered accurate
if the force field used in solving the structure is not accurate?
Perhaps this only leads to errors localized to the region of the LBHB,
but since these generally occur at the enzyme active site (the region
where the highest accuracy is desirable), this situation may not be
satisfactory.  Any comments?

ciao,
Bruce

*****************************************************************************


My original post:

Over the past few years, low barrier hydrogen bonds (LBHB for short)
have received a fair amount of attention in enzymology.  They are
characterized by being short, having low isotopic fractionation factors,
and very large proton NMR chemical shifts.  Some say that they are
extremely favorable energetically and hence an important "tool" used by
enzymes in catalysis, while others disagree.  I was wondering if anyone
could tell me whether the common molecular mechanics force fields could
be expected to properly model LBHB.  Can one expect AMBER, CHARMm, etc.,
to accurately predict the LBHB lengths and frequencies?  Or are these
bonds unusual enough that new terms need to be included?

*****************************************************************************

From: "Samuel A. Abrash"  To: 'Bruce Allan
Palfey'
 Subject: RE: low barrier hydrogen bonds

Dear Dr. Palfey, I am very interested in the responses to your question,
because I was under the impression that MM models could not accurately
predict any bond lengths and frequencies involving weak intermolecular
forces.  I know that these can be obtained with reasonable accuracy
using medium sized gaussian basis sets and hybrid density functional
theories, but was not aware that any degree of accuracy could be
obtained with lower levels of theory.  However, my expertise tends
toward either ab initio work or experiment, so I may be sadly
misinformed about the state of the art in molecular mechanics
calculations. If I am correct, though, one thing that might interest you
is a new option within Gaussian 98 which allows you to use different
levels of theory on different portions of a molecule (ONIOM).  This
method was developed by the group of Keiji Morokuma at Emory, and would
allow you to use a relatively high level of theory to treat the low
barrier hydrogen bonds, while using successively lower levels of theory
to treat the rest of the molecule. Thank you for tolerating my ignorance
of MM methods.

Best regards, Sam Abrash sabrash { *at * } richmond.edu

*****************************************************************************

Date: Fri, 8 May 1998 19:04:21 +0200 (MET DST) From: "J.Lin"
 To: brupalf ^at^ umich.edu Subject: Re:
CCL:low barrier hydrogen bonds

Dear Bruce,

> Over the past few years, low barrier hydrogen bonds (LBHB for short)
have ^^^^^^^^^^^^^^^^^^^^^^^^^^ Could you give some references?

> received a fair amount of attention in enzymology.  They are
characterized > by being short, having low isotopic fractionation
factors, and very large ^^^^^ shorter than normal hydrogen bond?

> proton NMR chemical shifts.  Some say that they are extremely
favorable > energetically and hence an important "tool" used by enzymes
in catalysis, > while others disagree.  I was wondering if anyone could
tell me whether > the common molecular mechanics force fields could be
expected to properly > model LBHB.  Can one expect AMBER, CHARMm, etc.,
to accurately predict the > LBHB lengths and frequencies?  Or are these
bonds unusual enough that new > terms need to be included?

In AMBER '91 parameters set, there are still hydrogen bond parameters
with 12-10 functional form, but in new AMBER '94 and '96 parameters they
are already absorbed into electrostatic and Lennard-Jones (12-6)
interactions. It means, the direction of force field development of
AMBER is trying to avoid unnecessary functional terms and parameters.
Sorry that I'm not an expert on LBHB and could not tell you if AMBER can
do this job. In my personal speculation I would say if the LBHB effect
is not very quantum-mechanical, one can try to optimize the parameters
to describe his system.

regards,

Jung-hsin

--
--------------------------------------------------------------------------------
Jung-hsin Lin Forum Modellierung und Institut fuer Festkoerperforschung
Forschungszentrum Juelich GmbH. D-52425 Juelich. Germany Phone (Office):
+49-2461-61-2515; Fax: - 2983; email: j.lin $#at#$ fz-juelich.de Phone/Fax
(Home): +49-2461-54419; Addr:  Muenchener Str. 5, D-52428 Juelich,
Germany
------------------------------------------------------------------------------

*****************************************************************************

Date: Fri, 08 May 98 14:24:47 EDT From: Janet Del Bene
 To: brupalf \\at// UMICH.EDU Subject: Low-barrier
hydrogen bonds

Hi, Bruce, Realistic descriptions of "normal" hydrogen bonds are often
lacking using the empirical/semi-empirical levels of theory, and even
low-levels of ab initio theory.  So, I do not think that these methods
could give consistently reliable descriptions of the so-called "LBHB". I
am actually looking into this area, motivated primarily by one of the
biochemists at this University. Janet E. Del Bene

*****************************************************************************

Date: Fri, 8 May 1998 13:55:44 -0500 (CDT) From: Michael A McAllister
 To: Bruce Allan Palfey 
Subject:
Re: CCL:low barrier hydrogen bonds


Hello Brian, I was very interested to see your message on CCL.  We have
been interested in LBHBs from a computational standpoint for a few years
now, mostly at the non-empirical level.  There is no doubt that current
force fields would NOT be able to accurately model any kind of strong
H-bond interaction.  We have looked at some semi-empirical models and
even these are substantially lacking in accuracy.  One of our current
interests/projects is to actually incorporate some decent force field
parameters for "LBHB" type hydrogens into common FFs.  As far as I know,
this has not been done yet.

Mike M.

**************************************************************************
Michael A. McAllister Assistant Professor of Chemistry	voice:
940-565-4584 University of North Texas 	 	  fax: 940-565-4318 P.O.
Box 305070			email: McAllister { *at * } unt.edu Denton, TX 76203
www:
http://people.unt.edu/~mam0008
**************************************************************************

*****************************************************************************

Date: Sat, 9 May 1998 01:20:23 -0700 From: Jiangang Chen
 To: Bruce Allan Palfey  Cc:
Achim Kless  Subject: Re: CCL:low barrier hydrogen
bonds


Dear Bruce,

Common force field methods model the hydrogen bonding with classical
electrostatical method. However, Low barrier hydorgen bond has large
covalent character.  I doubt that calculation with existing force field
parameters could capture the correct energetics of low barrier hydorgen
bond.


Jiangang Chen UCLA

*****************************************************************************

Date: Mon, 11 May 1998 09:28:17 +0010 From: Andrey Bliznyuk
 To: brupalf - at - umich.edu Cc:
Andrey.Bliznyuk #*at*# anu.edu.au Subject: Re: CCL:low barrier hydrogen bonds

Hi,

> Over the past few years, low barrier hydrogen bonds (LBHB for short) >
have received a fair amount of attention in enzymology.  They are >
characterized by being short, having low isotopic fractionation >
factors, and very large proton NMR chemical shifts.  Some say that >
they are extremely favorable energetically and hence an important >
"tool" used by enzymes in catalysis, while others disagree.

The name low barrier hydrogen bonds appeared simply because some people
did not know what they are talking about and jumped to discover the
wonders of enzyme catalysis. The proper name for these bonds is charged
H-bonds, which are known to be strong and there have been plenty on
literature on this subject both experimental and theoretical (for
somewhat outdated review see, for example, A.A. Voityuk & A.A. Bliznyuk,
Zh. Struct. Khim, 1992, 33(6), pp. 157-183). The physical chemist
community was simply stunned by ignorance of published articles and did
not respond immediately. However, two paper seems to close the subject:
JACS, 1995, 117, pp.6970-6975 Chemistry and Biology, 1996, 3,
pp.163-170.

> I was > wondering if anyone could tell me whether the common molecular
> mechanics force fields could be expected to properly model LBHB. > Can
one expect AMBER, CHARMm, etc., to accurately predict the LBHB > lengths
and frequencies?  Or are these bonds unusual enough that new > terms
need to be included?

There is nothing unusual about these bonds from the computational point
of view, and all MM force fields would treat the energy of these bonds
more or less correctly. Of course, the charge transfer will not be taken
into account, and hence the results will not be exceptionally good, but
it is well known problem with MM approaches. And if the total energy or
geometry is what you are looking for, then it is not too bad.

Andrey

Dr. Andrey Bliznyuk ANU Supercomputer Facility IT services The
Australian National University Canberra, ACT 0200 Australia Email:
Andrey.Bliznyuk <-at-> anu.edu.au

*****************************************************************************

Date: Wed, 13 May 1998 10:19:33 -0700 (MST) From: Soaring Bear
 To: chemistry at.at www.ccl.net
Subject: CCL:CCL:low barrier hydrogen bonds

Considering how variable crystal hydrogen bonds are [see Gilli in "Fund
Princ of Molec Modeling" ed W Gans 1996 and GA Jeffrey & W Saenger
"Hydrogen Bonding in Biol. Structures" 1991] modeling forcefields can
only be expected to be an approximation.


^[[7m^[[5mSOARING BEAR^[[0m PhD     bear at.at pharmacy.arizona.edu
Computational Medicinal Chemistry; Molecular Modeling; Pharmacognosy
Protein & DNA Structural Biology; Cancer Biochemistry; Informatics; QSAR
http://ellington.pharmacy.arizona.edu/~bear/resume.html
currently
seeking employment





Similar Messages
05/08/1998:  low barrier hydrogen bonds
05/14/1998:  CCL:low barrier hydrogen bonds
05/19/1998:  Re: CCL:LBHBs
11/16/1993:  MM Parameters/Macrocycles
08/19/1994:       Intramolecular H-bonds: more refs
08/01/1996:  Re: CCL:M:Heat of formation calculation using MOPAC.
10/01/1993:  torsion of conjugated systems -- summary
02/01/1999:  Re: CCL:Hydrogen Bonds with DFT / GGA
06/08/1993:  undergrad computational chem
01/06/1994:  Summary of "Anomalous PM3 Hydrogen Interactions"


Raw Message Text