Applications of Density Functional Theory To Biological and Bioinorganic Chemistry 1st Edition Dennis R. Salahub All Chapter Instant Download
Applications of Density Functional Theory To Biological and Bioinorganic Chemistry 1st Edition Dennis R. Salahub All Chapter Instant Download
Applications of Density Functional Theory To Biological and Bioinorganic Chemistry 1st Edition Dennis R. Salahub All Chapter Instant Download
com
https://textbookfull.com/product/applications-of-
density-functional-theory-to-biological-and-
bioinorganic-chemistry-1st-edition-dennis-r-
salahub/
https://textbookfull.com/product/dielectrophoresis-theory-methodology-
and-biological-applications-1st-edition-ronald-r-pethig/
textbookfull.com
https://textbookfull.com/product/chemistry-of-metalloproteins-
problems-and-solutions-in-bioinorganic-chemistry-1st-edition-joseph-j-
stephanos/
textbookfull.com
https://textbookfull.com/product/molybdenum-and-tungsten-enzymes-
bioinorganic-chemistry-1st-edition-russ-hille/
textbookfull.com
https://textbookfull.com/product/developing-global-leaders-insights-
from-african-case-studies-eva-jordans/
textbookfull.com
Lonely Planet Vancouver Victoria 8th Edition Lonely Planet
https://textbookfull.com/product/lonely-planet-vancouver-victoria-8th-
edition-lonely-planet/
textbookfull.com
https://textbookfull.com/product/an-anthropology-of-marxism-cedric-j-
robinson/
textbookfull.com
https://textbookfull.com/product/laboratory-techniques-in-organic-
chemistry-4th-edition-jerry-r-mohrig/
textbookfull.com
https://textbookfull.com/product/i-me-mine-back-to-kant-and-back-
again-1st-edition-beatrice-longuenesse/
textbookfull.com
https://textbookfull.com/product/electing-the-pope-in-early-modern-
italy-1450-1700-first-edition-miles-pattenden/
textbookfull.com
Developing Resilience A Cognitive Behavioural Approach 2th
Ed 2nd Edition Michael Neenan
https://textbookfull.com/product/developing-resilience-a-cognitive-
behavioural-approach-2th-ed-2nd-edition-michael-neenan/
textbookfull.com
150
Structure and Bonding
Series Editor:
Editorial Board:
The series Structure and Bonding publishes critical reviews on topics of research
concerned with chemical structure and bonding. The scope of the series spans the
entire Periodic Table and addresses structure and bonding issues associated with all
of the elements. It also focuses attention on new and developing areas of modern
structural and theoretical chemistry such as nanostructures, molecular electronics,
designed molecular solids, surfaces, metal clusters and supramolecular structures.
Physical and spectroscopic techniques used to determine, examine and model
structures fall within the purview of Structure and Bonding to the extent that the
focus is on the scientific results obtained and not on specialist information
concerning the techniques themselves. Issues associated with the development of
bonding models and generalizations that illuminate the reactivity pathways and
rates of chemical processes are also relevant.
The individual volumes in the series are thematic. The goal of each volume is to
give the reader, whether at a university or in industry, a comprehensive overview of
an area where new insights are emerging that are of interest to a larger scientific
audience. Thus each review within the volume critically surveys one aspect of that
topic and places it within the context of the volume as a whole. The most significant
developments of the last 5 to 10 years should be presented using selected examples
to illustrate the principles discussed. A description of the physical basis of the
experimental techniques that have been used to provide the primary data may also
be appropriate, if it has not been covered in detail elsewhere. The coverage need not
be exhaustive in data, but should rather be conceptual, concentrating on the new
principles being developed that will allow the reader, who is not a specialist in the
area covered, to understand the data presented. Discussion of possible future
research directions in the area is welcomed.
Review articles for the individual volumes are invited by the volume editors.
Applications of Density
Functional Theory to
Biological and Bioinorganic
Chemistry
With contributions by
M. Causá • P.K. Chattaraj • A. Chakraborty • M. D’Amore •
A. Goursot • C. Garzillo • F. Gentile • E.S. Kryachko •
A. de la Lande • S. Pan • A.M. Putz • M.V. Putz •
R. Silaghi-Dumitrescu • D.R. Salahub • A. Savin •
R. Zhang • Y. Zhang
Editors
Mihai V. Putz D. Michael P. Mingos
Structural and Computational Inorganic Chemistry Laboratory
Physical-Chemistry Laboratory University of Oxford
West University of Timisoara Oxford
Timişoara United Kingdom
Romania
In the early twentieth century following the elucidation of the structure of atoms it
became evident that atoms and molecules with even numbers of electrons were far
more numerous than those with odd numbers of electrons. In 1916, G. N. Lewis
provided the first comprehensive description of ionic and covalent bonds, when he
postulated that atoms tend to hold an even number of electrons in their outer shells
and a special stability was associated with eight valence electrons, which he
speculated were arranged symmetrically at the eight corners of a cube. In 1919, I.
Langmuir suggested that the structure of the periodic table could be rationalized
using an extension of Lewis’ postulates. In 1922, N. Bohr updated his model of the
atom by assuming that certain numbers of electrons (for example 2, 8, and 18)
corresponded to stable “closed shells.” In 1926, Schrödinger established a wave
mechanical description of the hydrogen atom which was subsequently extended to
polyelectron atoms. Pauli was the first to realize that the complicated numbers of
electrons in closed shells can be reduced to the simple rule of one per state, if the
electron states are defined using four quantum numbers. For this purpose he
introduced a new two-valued quantum number, identified by Goudsmit and
Uhlenbeck as electron spin. The resulting Pauli Exclusion Principle states that no
two electrons in a single atom can have the same four quantum numbers; if n, l, and
ml are the same, ms must be different such that the electrons have opposite spins.
The idea of shared electron pairs introduced by Lewis provided an effective
qualitative picture of covalent bonding and it still forms the basis of the universal
notation for chemical communication, but it was Heitler and London who in 1927
developed the first successful quantum mechanical expression for this bonding
model. Initially they provided a description of the bonding in molecular hydrogen,
but it was subsequently adapted to more complex molecules and its widespread
applications were articulated with great conviction by Linus Pauling. An alternative
molecular orbital description of chemical bonding originated from Burrau’s
description of the hydrogen molecule ion and this model was subsequently widely
developed by Mulliken and Lennard-Jones. The electrons occupy molecular
orbitals which are delocalized over the whole molecule and were filled according
to the Aufbau Principle and assigned quantum numbers according to the Pauli
v
vi Preface
ix
.
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
List of Contributors
xi
xii List of Contributors
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 DFT and ADFT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3 DFT-D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4 BODFT-MD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.2 Properties of Phosphatidyl Choline Lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3 Activation of Triplet Dioxygen by Bio-inspired Cuprous Complexes . . . . . . . . . . . . . . . 26
5 DFT/MM-MD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6 Constrained-DFT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
6.1 Methodological Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
6.2 cDFT and Population Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
6.3 Modeling Electron Transfer Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
6.4 Other Applications of cDFT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
7 Interpretational-DFT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
8 Conclusions and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
1 Introduction
Density Functional Theory (DFT) has become the workhorse of applied computa-
tional chemistry because of its particularly appealing combination of accuracy,
speed, and interpretability. Nearly five decades have transpired since the seminal
papers of Hohenberg and Kohn [1] and Kohn and Sham [2] and nearly nine since
the first use of a density functional in atomic theory by Thomas [3] and by Fermi
[4]. Each decade has seen steady, inexorable, progress—more accurate functionals
have been developed, better and faster algorithms have been implemented, and new
analytical concepts have been devised (see e.g., [5] for a review covering the
1964–2004 period). The new methodologies and techniques have allowed systems
of ever growing complexity to be addressed, to the point where, now, DFT is
starting to have a real impact on biological questions.
In this review, we will focus on biology and try to capture the state of the art in
studies that approach aspects of biological systems and processes from various
points of view, all of them involving DFT. We will exclude from our scope the
semiempirical Tight-Binding DFT (DFTB) and also Time-Dependent DFT
(TDDFT). Although these methods have been used to study systems of biological
interest, we choose to discuss here methods and applications that involve “real”
ground-state DFT, including hybrid functionals, at the core.
Recent Progress in Density Functional Methodology for Biomolecular Modeling 3
2.1 Methodology
with the operators in parentheses being, respectively, the kinetic energy of the
noninteracting Kohn–Sham reference system, the external potential (nuclear-
electron attraction), the classical coulomb potential felt at the point r due to the
electronic charge distribution (including the self-interaction), and the exchange-
correlation potential. ’i is a Kohn–Sham orbital and ei is the corresponding
Kohn–Sham orbital energy. The exchange-correlation potential is the functional
derivative of the Kohn–Sham exchange-correlation energy with respect to a varia-
tion of the density:
dExc ½rðrÞ
vxc ðrÞ ¼ : (2)
drðrÞ
The Kohn–Sham equations are exact but, of course, for practical calculations
approximations have to be made and these will determine the accuracy, the speed,
and the interpretability of approximate KS-DFT methods.
The most important choice is that of the exchange-correlation functional. Here
there is a wide variety of options. Some standard choices are functionals of the
Generalized Gradient Approximation (GGA) (which we favor because of their
computational speed when used within the Auxiliary DFT approach and their
overall good level of accuracy—see below; PBE [6–8] is a common choice) and
the hybrid functionals that involve a component of Hartree–Fock exchange, the
B3LYP functional [9, 10] providing the prototypical example:
EB3LYP
xc ¼ EVWN
xc þ a0 ðEHF
x Ex
VWN
Þ þ ax ðEBecke
x EVWN
x Þ þ ac ðEVWN
c ELYP
c Þ;
(3)
Q-Chem [20], etc.) and also in our own code deMon2k [21]. We further outline the
use of fitting functions [22] and the Auxiliary-DFT (ADFT) [23] methodology
because it provides a real computational advantage provided (for the moment)
that pure, nonhybrid, functionals are employed. Of course one has to address the
accuracy question, both as concerns the inherent errors of approximate functionals
and the numerical errors associated with the choice of basis sets, auxiliary basis
sets, numerical integration grids, etc. These issues will be addressed in the discussion
of the applications below, but first we complete the overview of the methodology
with a discussion of density fitting and the ADFT methodology.
The use of Gaussian functions in DFT was pioneered by Dunlap, Connolly, and
Sabin [22] who, in 1982, formulated the LCGTO-Xa method that incorporated a
variational fit of the coulomb terms. With the usual LCAO approximation, and
using a general form for the exchange-correlation energy rather than Xa (an early
density functional that uses a local density approximation for exchange,
incorporating a parameter a that, arguably, accounts for correlation to a certain
extent), one can write the total energy in terms of the density matrix:
X 1 XX
E¼ Pmn Hmn þ Pmn Pst hmn k sti þ Exc ½rðrÞ: (4)
m;n
2 m;n s;t
Here Pmn and Pts denote elements of the density matrix and Hmn an element of the
core Hamiltonian (kinetic energy and nuclear attraction). The basis functions that
are used to expand the molecular orbitals are indicated by m; n; s and t. The (up to)
four-center electron repulsion integral is represented by:
ðð
mðr0 Þnðr0 Þ
hmnksti ¼ sðrÞtðrÞdr0 dr: (5)
jr r0 j
The key development came from the realization that if one could fit the density
using an auxiliary basis set then in practice one index could be saved, turning the
problem from a basically N4 scaling, where N is the size of the orbital basis set, to
N2M where M is the size of the auxiliary basis set. Writing
X
~ðrÞ ¼
r xk kðrÞ (6)
k
along with the expansion of the density using the density matrix:
X
rðrÞ ¼ Pmn mðrÞnðrÞ (7)
m;n
@e2 X X
¼ Pm;n hmnkki þ xl hlkki 08k: (9)
@xk m;n l
Solution of this equation is often called the Resolution of the Identity and is now
also common in wave function methods. A second auxiliary set was introduced for
the exchange-correlation potential; however, this used a least-squares procedure on
a numerical grid and it was not done variationally. This level of theory is identified
by the keyword BASIS in our deMon2k software since the density matrix is used
(within a min–max scheme).
Going one step further in terms of potential speed, Köster and coworkers [23]
studied the use of the auxiliary density throughout the variational process. The
corresponding auxiliary DFT (ADFT) energy expression is the following:
X X 1X
E¼ Pmn Hmn þ Pm;n hmnkkixk ~ðrÞ;
xk xl hkkli þ Exc ½r (10)
m;n m;n;k
2 k:l
where, in practice, as implemented in deMon2k, the orbital basis sets are contracted
Cartesian Gaussians while the auxiliary sets are primitive Hermite Gaussians. In
Eq. (10) the first term represents the core energy, the second the coulomb repulsion
energy of the electrons using the density matrix once and the fitted density once, the
third term complements the second in the variational fitting procedure and involves
only the fitted density and the final exchange-correlation term also involves only the
fitted density. The derivatives of this energy expression with respect to the density
matrix elements define the ADFT Kohn–Sham matrix elements:
@E X ~ðrÞ
@Exc ½r
Kmn ¼ ¼ Hmn þ hmnkki þ : (11)
@Pmn k
@P mn
After some algebra and using the variational property of the density fitting the
following equation is derived:
@E X
Kmn ¼ ¼ Hmn þ hmnkkiðxk þ zk Þ; (12)
@Pmn k
2.2 Applications
We start the discussion of biomodeling using straight DFT with a brief overview of
some recent work from Per Siegbahn and coworkers using the recent review of
Siegbahn and Himo as the leading reference [24]. Siegbahn has been a champion
for (properly converged) cluster models for a number of years and his well-chosen
applications have shed considerable light on classes of enzymes for which reliable
cluster models can be formulated. With recent increases in computer power and
program efficiency, cluster models with upwards of 150 atoms are now feasible; for
some, but not all, types of reactions, these are able to capture the essence of the
relevant free-energy profiles.
Siegbahn’s recent work focuses on the B3LYP hybrid functional for which he
claims “In spite of numerous attempts, it has been difficult to improve the accuracy
beyond that of this functional.” Of course, that does not relieve us from the
responsibility of addressing the question of accuracy for a given reaction. Siegbahn
cites three sources of inaccuracy in DFT methods (1) the self-interaction error,
(2) the inherent limitations of a single-determinant approach, and (3) the lack of van
der Waals interactions in the usual functionals, including B3LYP. According to
Siegbahn, errors (1) and (2) (using a spin-unrestricted formalism) tend to cancel and
this can at least partially explain the relatively good performance of B3LYP. He
also discusses the effects of varying the amount of exact exchange in the hybrid
functionals, leading to the rule of thumb that if the results do not change very much
when the amount of exact exchange is decreased from 20 % to 15 % the methodology
seems to be reliable. While this seems less than an ideal procedure from an “ab
initio” perspective, such procedures are necessary at the present stage of advance-
ment in the search for more accurate and generally applicable functionals. The body
of work using B3LYP indicates that it can provide results of useful accuracy if it is
applied with due caution. We will discuss the third source of error in the next
section on DFT-D methodologies. We only indicate here that the empirical
corrections for dispersion-like interactions can now be readily incorporated and
that they lead in many cases to significant improvement. They should become the
default option.
Two further aspects of Siegbahn’s cluster approach are (1) a coordinate-locking
scheme and (2) the use of a polarizable continuum method (PCM) to model the
electrostatic effects of the surrounding medium (protein and solvent). The coordinate-
locking scheme fixes the coordinates of key atoms on the periphery, hence preventing
large artificial movements of the active site groups. For very small cluster models
this approach can lead to artifacts, but as the cluster grows it behaves better and
better (and ultimately becomes unnecessary). The PCM approach assumes that the
surroundings can be represented as a homogeneous polarizable medium with a
dielectric constant that has to be chosen (often e ¼ 4 gives good results but in
some studies e is varied to gauge the sensitivity of the results to this parameter).
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
Recent Progress in Density Functional Methodology for Biomolecular Modeling 9
Fig. 1 Top: Reaction mechanism for the methylation of the Lysine side chain of histone by
S-adenosylmethionine using HKMT as the catalyst. Middle: The three cluster models used. Stars
indicate fixed atoms in the coordinate-locking scheme. Bottom: Potential energy profiles
(kcal/mol). Results for various values of the dielectric constant are shown in different colors.
Reproduced with permission from [24]
Clearly, the more protein residues and waters of solvation that are included in the
cluster model explicitly, the less critical will it be to resort to this somewhat doubtful
PCM approximation.
We turn now to histone lysine methyltransferase (HKMT) which catalyzes the
methylation of the N-terminal histone tail of chromatin using the S-adenosyl-
methionine (SAM) cofactor as the methylating agent (see Fig. 1).
Models of increasing size were used; Model I (46 atoms) contained only
truncated models of SAM and the substrate; Model II (72 atoms) also contained
two important tyrosine residues that form hydrogen bonds with the amino group of
10 D.R. Salahub et al.
the substrate and Model III (132 atoms, Middle part of Fig. 1) contains additional
groups that form a ring around the substrate and interact with the transferred methyl
group. The total overall charge on the models is +1. Considering only this particular
reaction one has a “quantum-chemistry friendly” situation, where all of the reaction
steps (reactants, transition state, and products) take place in a reasonably similar
environment (provided by the interacting residues of Model III, for example) and a
single positive charge is transferred from the cofactor to the substrate. Geometries
of the critical points were calculated at the B3LYP/6-31G(d,p) level with single-
point calculations of the energies using a larger 6-311 + G(2d,2p) basis. Solvation
effects were calculated as single points at the same level as the geometry
optimizations using a range of dielectric constants (2, 4, 8, 16, 80). As expected
the larger models showed little dependence on the value of e chosen (essentially
identical values for the entire range of e for the activation energy from the reactant
side and roughly a range of 4 kcal mol1 for the overall exothermicity of the
reaction) (see the bottom part of Fig. 1).
It was found that the transition-state geometries for the three models were
very similar which is of practical importance, allowing the TS search to be
performed for small models and then only having to refine the structures for the
larger models. Moreover, the calculated energy differences are not very differ-
ent in the different models. In fact all of the barriers are close to the experi-
mental barrier of 20.9 kcal mol1. The solvation effects saturate quickly; 132
atoms in this case seem to be well converged, a fortunate circumstance for this
reaction and a good number of other reactions, because such cluster sizes are
well within the range of current possibility. It was emphasized that geometry
optimization is crucial; taking raw crystallographic geometries usually leads to
wrong energy profiles.
Some final comments on the entropy (free energy) are in order. Because of the
coordinate-locking scheme, there are a number of small imaginary frequencies
(<30i cm1) which, although they do not affect the energetics significantly, do
render the calculation of the harmonic frequencies and their associated entropy
contributions inaccurate. So the reported energies correspond to enthalpies and not
free energies. For the present HKMT case, Siegbahn refers to work by Hu and
Zhang [25] to show that the entropy effects are quite small. Hu and Zhang used a
QM/MM methodology (B3LYP with single-point MP2) with free-energy perturba-
tion theory (see below) for the MM contributions to the entropy and harmonic
frequencies for the QM part (now possible because there are no frozen atoms). The
QM part contained 66 atoms, SAM and the lysine side chain only. The free energy
barriers are found to be only about 1.1 kcal mol1 lower than the potential energy
barriers, confirming the validity of neglecting entropy for this particular reaction,
which is well contained within the QM part of the system. We will see below that
this is not a general result. For some reactions, such as those involving DNA or
RNA polymerase, entropy effects can be very large; they definitely cannot be
neglected.
For other examples of the valid use of the finite cluster model, we refer the reader
to [26–31].
Recent Progress in Density Functional Methodology for Biomolecular Modeling 11
In the central dogma of biology (“DNA to RNA to proteins”) the first step
transcribes the genetic code from DNA to messenger RNA. This task is accom-
plished by a marvelous nano-molecular machine, the RNA Polymerase (RNAP)
enzyme, which, once initiated, processes a DNA template strand, adding successive
matching nucleoside triphosphates to a growing chain of m-RNA. RNAP is a
complex multidomain protein containing about 3,500 residues and 28,000
nonhydrogen atoms [32]. One of the great triumphs of modern protein crystallog-
raphy is surely the elucidation of many aspects of the mechanism for transcription
for which Roger Kornberg was awarded the 2006 Nobel Prize in Chemistry. In
principle, the work of Kornberg and the others who have accomplished this
magnificent task is simple—isolate and crystallize proteins that are ligated by
various intermediates along the multistate reaction path and derive their structures
by X-ray analysis. In this way we get “flash-frozen” snapshots of the reaction
mechanism. This has been so successful that a movie of the process has been
made [33] which may be viewed at http://www.lmb.uni-muenchen.de/cramer/pr-
materials. We show some snapshots from the movie in Fig. 2.
This is clearly a more complex reaction than that of HKMT just described. Of the
seven steps shown in the movie all but one involve conformational changes of the
enzyme that are more in the domain of Molecular Mechanics than in that of
quantum chemistry (we are studying these conformational steps, but they are not
the topic of the present chapter). The sole exception is step 4, catalytic
incorporation, which actually involves a multistep mechanism of chemical
reactions. This is where theory and computation have to step in to help elucidate
the mechanism. The first steps of the process involve the construction of cluster
models for the calculation of relevant portions of the potential energy surface
corresponding to proposed reaction steps. Several key choices have to be made
for which reactions to consider.
The basic enzymatic function of RNAP is the transfer of the nucleotidyl motif
from the rNTP substrates to the hydroxyl at the 30 -end of the nascent RNA
transcript. The nucleotidyl transfer reaction can be simplified as
RNAP
RNAi þ rNTP ! RNAiþ1 þ HPPi : (13)
The catalytic center of RNAP includes the binding site for the 30 -end of RNA
and the insertion site for the incoming rNTP. In the nucleotidyl transfer reaction, the
30 -OH group in the sugar ring of the RNA primer reacts with the a-phosphorous
atom of a ribonucleoside triphosphate by nucleophilic attack, then the Pa–Oab bond
is broken and pyrophosphate (PPi) is released. Thus, a nucleotidyl addition to the
RNA primer is achieved. Structural and biochemical data have shown that the
active centers of all polymerases share certain common features: a pair of metal
ions (normally divalent magnesium ions Mg2+) and three universally conserved
carboxylates. The two-metal-ion mechanism for the nucleotidyl transfer reaction
12 D.R. Salahub et al.
Fig. 2 Snapshots from the NAC (Nucleotide Addition Complex) movie have been depicted that
correspond to different functional states of the EC. Reproduced with permission from [33]
was proposed by Steiz [34]. A schematic diagram of the ternary elongation complex
for yeast RNA polymerase is shown in Fig. 3.
We have considered several different detailed reaction mechanisms for the
nucleotide addition [36]. In all of them, the 30 OH group has to be deprotonated,
either by passing its proton to another residue of the enzyme or to a solvent water
Recent Progress in Density Functional Methodology for Biomolecular Modeling 13
base base
O sugar OH O sugar OH
deprotonation protonation
3'
O O H OH
RNA primer O2 rNTP
H P O
O
O O- P O
O
ASP485 Mg2+A O- P
O O-
O O-
Mg2+B
O
GLY484 O
ASP483 O
Fig. 3 The two-metal-ion mechanism and two-proton transfer model for the nucleotidyl reaction
of yeast RNA polymerase II [34, 35]. The ternary elongation complex consists of three parts: the
growing RNA transcript (blue), the incoming rNTP (red), yeast RNA polymerase active center
(black) that is mainly composed of two divalent magnesium ions, Mg2+ A and B, and three
conserved amino acid residues, aspartates. The two protons are directly or indirectly transferred at
two sites proposed according to deuterium isotope effect experiments. One site is the primer RNA
30 terminus (left square), where the RNA primer 30 OH must be deprotonated by the incoming
rNTP, a nearby residue or a mediated water molecule before nucleophilic attack. The other is the
a- and b-phosphate bridging oxygen atom (right square), where the pyrophosphate should be
protonated by a nearby residue or a mediated water molecule before it leaves. The three aspartate
residues ASP481, 483, 485 connected by PHE482 and GLY484 conjugate with two magnesium
ions to form the active center of yeast RNA polymerase
«Madame,
«Je ne sais pas si, comme vous le dites, vous vous êtes trompée
sur moi. Mais ce que je sais, c'est que je me suis trompé
cruellement sur vous. J'avais cru que vous étiez mes amies.
Vous le disiez, vous faisiez semblant de l'être, et je vous aimais
plus que ma vie. Je vois maintenant que tout cela est un
mensonge, et que votre affection pour moi n'était qu'une
duperie: vous vous serviez de moi, je vous amusais, je vous
distrayais, je vous faisais de la musique,—j'étais votre
domestique. Votre domestique, je ne le suis pas! Je ne suis celui
de personne!
«Vous m'avez fait durement sentir que je n'avais pas le droit
d'aimer votre fille. Rien au monde ne peut empêcher mon cœur
d'aimer ce qu'il aime; et si je ne suis pas de votre rang, je suis
aussi noble que vous. C'est le cœur qui ennoblit l'homme: si je
ne suis pas comte, j'ai peut-être plus d'honneur en moi que bien
des comtes. Valet ou comte, du moment qu'il m'insulte, je le
méprise. Je méprise comme la boue tout ce qui se prétend
noble, s'il n'a pas la noblesse de l'âme.
«Adieu! Vous m'avez méconnu. Vous m'avez trompé. Je vous
déteste.
«Celui qui aime, en dépit de vous, et qui aimera jusqu'à sa mort
mademoiselle Minna, parce qu'elle est à lui, et que rien ne peut
la lui reprendre.»
«Cher Monsieur,
«Puisque, à votre avis, il y a eu un malentendu entre nous, le
plus sage est sans doute de ne point le prolonger. Je me
reprocherais de vous imposer davantage des relations,
devenues pénibles pour vous. Vous trouverez donc naturel que
nous les interrompions. J'espère que vous ne manquerez pas,
dans la suite, d'autres amis, qui sauront vous apprécier, comme
vous désirez l'être. Je ne doute point de votre avenir, et suivrai
de loin, avec sympathie, vos progrès dans la carrière musicale.
Salutations.
«Josepha von Kerich»
Les plus amers reproches eussent été moins cruels. Christophe se vit
perdu. On peut répondre à qui vous accuse injustement. Mais que
faire contre le néant de cette indifférence polie? Il s'affola. Il pensa
qu'il ne verrait plus Minna, qu'il ne la verrait plus jamais; et il ne put
le supporter. Il sentit le peu que pèse tout l'orgueil du monde, au
prix d'un peu d'amour. Il oublia toute dignité, il devint lâche, il écrivit
de nouvelles lettres, où il suppliait qu'on lui pardonnât. Elles
n'étaient pas moins stupides que celle où il s'emportait. On ne lui
répondit rien.—Et tout fut dit.
Il faillit mourir. Il pensa à se tuer. Il pensa à tuer. Il se figura du
moins qu'il le pensait. Il eut des désirs incendiaires. On ne se doute
pas du paroxysme d'amour et de haine qui dévorent certains cœurs
d'enfants. Ce fut la crise la plus terrible de son enfance. Elle mit fin à
son enfance. Elle trempa sa volonté. Mais elle fut bien près de la
briser pour toujours.
Il ne pouvait plus vivre. Accoudé sur sa fenêtre, pendant des heures,
et regardant le pavé de la cour, il songeait, comme quand il était
petit, qu'il y avait un moyen d'échapper à la torture de la vie. Le
remède était là, sous ses yeux, immédiat .. Immédiat? Qui le
savait?... Peut-être après des heures—des siècles—de souffrances
atroces!... Mais si profond était son désespoir d'enfant qu'il se
laissait glisser au vertige de ces pensées.
Louisa voyait qu'il souffrait. Elle ne pouvait se douter exactement de
ce qui se passait en lui; mais son instinct l'avertissait du danger. Elle
tâchait de se rapprocher de son fils, de connaître ses peines, afin de
le consoler. Mais la pauvre femme avait perdu l'habitude de causer
intimement avec Christophe; depuis bien des années, il renfermait
ses pensées en lui; et elle était trop absorbée par les soucis
matériels de la vie, pour avoir le temps de chercher à le deviner.
Maintenant qu'elle eût voulu lui venir en aide, elle ne savait que
faire. Elle rôdait autour de lui, comme une âme en peine; elle eût
souhaité de trouver les mots qui lui eussent fait du bien; et elle
n'osait parler, de crainte de l'irriter. Et malgré ses précautions, elle
l'irritait par tous ses gestes, par sa présence même; car elle n'était
pas très adroite, et il n'était pas très indulgent. Cependant il l'aimait,
ils s'aimaient. Mais il suffit de si peu pour séparer des êtres qui se
chérissent! Un parler trop fort, des gestes maladroits, un tic
inoffensif dans les yeux ou le nez, une façon de manger, de marcher
et de rire, une gêne physique qu'on ne peut analyser... On se dit que
ce n'est rien; et pourtant, c'est un monde. C'est assez, bien souvent,
pour qu'une mère et un fils, deux frères, deux amis, qui sont tout
près l'un de l'autre, restent éternellement étrangers l'un à l'autre.
Christophe ne trouvait donc pas auprès de sa mère un appui dans la
crise qu'il traversait. Et d'ailleurs, de quel prix est l'affection des
autres pour l'égoïsme de la passion, préoccupée d'elle seule?
Une nuit que les siens dormaient, et qu'assis dans sa chambre, sans
penser, sans bouger, il s'enlizait dans ses dangereuses idées, un bruit
de pas fit résonner la petite rue silencieuse, et un coup frappé à la
porte l'arracha à son engourdissement. On entendait un murmure de
voix indistinctes. Il se rappela que son père n'était pas rentré le soir,
et il pensa avec colère qu'on le ramenait encore ivre, comme l'autre
semaine, où on l'avait trouvé couché en travers de la rue. Car
Melchior n'observait plus aucune retenue; il se livrait à son vice, sans
que sa santé athlétique parût souffrir d'excès et d'imprudences, qui
eussent tué un autre homme. Il mangeait comme quatre, buvait à
tomber ivre-mort, passait des nuits dehors sous la pluie glacée, se
faisait assommer dans des rixes, et se retrouvait sur ses pieds, le
lendemain, avec sa bruyante gaieté, voulant que tout le monde fût
gai autour de lui.
Louisa, déjà levée, allait précipitamment ouvrir. Christophe, qui
n'avait pas bougé, se boucha les oreilles, pour ne pas entendre la
voix avinée de Melchior et les réflexions goguenardes des voisins...
Soudain, une angoisse inexplicable le saisit: il eut peur de ce qui
allait venir... Et aussitôt, un cri déchirant lui fit relever la tête. Il
bondit à la porte...
Au milieu d'un groupe d'hommes, qui parlaient à voix basse, dans le
corridor obscur, éclairé par la lueur tremblante d'une lanterne, sur
une civière était couché, comme autrefois grand-père, un corps
ruisselant d'eau, immobile. Louisa sanglotait à son cou. On avait
trouvé Melchior noyé dans le ru du moulin.
Christophe poussa un cri. Tout le reste du monde disparut, ses
autres peines furent balayées. Il se jeta sur le corps de son père, à
côté de Louisa, et ils pleurèrent ensemble.
Assis auprès du lit, veillant le dernier sommeil de Melchior, dont le
visage avait pris maintenant une expression sévère et solennelle, il
sentait la sombre tranquillité du mort entrer en lui. Sa passion
enfantine s'était dissipée, comme un accès de fièvre; le souffle
glacial de la tombe avait tout emporté. Minna, son orgueil, son
amour, hélas! quelle misère! Que tout était peu de chose auprès de
cette réalité, la seule réalité: la mort! Était-ce la peine de tant
souffrir, désirer, s'agiter, pour en arriver là!...
Il regardait son père endormi, et il était pénétré d'une pitié infinie. Il
se rappelait ses moindres actes de bonté et de tendresse. Car, avec
toutes ses tares, Melchior n'était pas méchant, il y avait beaucoup de
bon en lui. Il aimait les siens. Il était honnête. Il avait un peu de la
probité intransigeante des Krafft, qui, dans les questions de moralité
et d'honneur, ne souffrait pas de discussion et n'eût jamais admis
ces petites saletés morales, que tant de gens de la société ne
regardent pas tout à fait comme des fautes. Il était brave et, en
toute occasion dangereuse, s'exposait avec une sorte de jouissance.
S'il était dépensier pour lui-même, il l'était aussi pour les autres: il
ne pouvait supporter qu'on fût triste; et il faisait volontiers largesse
de ce qui lui appartenait,—et de ce qui ne lui appartenait pas,—aux
pauvres diables qu'il rencontrait sur son chemin. Toutes ses qualités
apparaissaient maintenant à Christophe: il les exagérait. Il lui
semblait qu'il avait méconnu son père. Il se reprochait de ne l'avoir
pas assez aimé. Il le voyait vaincu par la vie: il croyait entendre cette
malheureuse âme, entraînée à la dérive, trop faible pour lutter, et
gémissant de sa vie inutilement perdue. Il entendait cette
lamentable prière, dont l'accent l'avait déchiré naguère:
—Christophe! ne me méprise pas!
Et il était bouleversé de remords. Il se jetait sur le lit et baisait le
visage du mort, en pleurant. Il répétait, comme autrefois:
—Mon cher papa, je ne te méprise pas, je t'aime! Pardonne-moi!
Mais la plainte ne s'apaisait pas, et reprenait, angoissée:
—Ne me méprisez pas! Ne me méprisez pas!...
Et brusquement, Christophe se vit couché lui-même à la place du
mort; il entendait les terribles paroles sortir de sa propre bouche, il
sentait sur son cœur peser le désespoir d'une inutile vie,
irrémédiablement perdue. Et il pensait avec épouvante: «Toutes les
souffrances, toutes les misères du monde, plutôt que d'en arriver là!
»... Combien il en avait été près! N'avait-il pas failli céder à la
tentation de briser sa vie, pour échapper lâchement à sa peine?
Comme si toutes les peines, toutes les trahisons n'étaient pas des
chagrins d'enfant auprès de la torture et du crime suprêmes de se
trahir soi-même, de renier sa foi, de se mépriser dans la mort!
Il vit que la vie était une bataille sans trêve et sans merci, où qui
veut être un homme digne du nom d'homme doit lutter
constamment contre des armées d'ennemis invisibles: les forces
meurtrières de la nature, les désirs troubles, les obscures pensées,
qui poussent traîtreusement à s'avilir et à s'anéantir. Il vit qu'il avait
été sur le point de tomber dans le piège. Il vit que le bonheur et
l'amour étaient une duperie d'un moment, pour amener le cœur à
désarmer et à abdiquer. Et le petit puritain de quinze ans entendit la
voix de son Dieu:
—Va, va, sans jamais te reposer.
—Mais où irai-je, Seigneur? Quoique je fasse, où que j'aille, la fin
n'est-elle pas toujours la même, le terme n'est-il point là?
—Allez mourir, vous qui devez mourir! Allez souffrir, vous qui devez
souffrir! On ne vit pas pour être heureux. On vit pour accomplir ma
Loi. Souffre. Meurs. Mais sois ce que tu dois être:—un Homme.
L'ADOLESCENT
PREMIÈRE PARTIE
LA MAISON EULER