Generation of 3D

structure of protein
By: Bikash
Dutta
M.Pharm,
pharmaceutical
chemistry,
ripans
 Generation of 3D structure of protein is an important task to Understand the interaction
between drug(ligand) and protein molecule(receptor). It helps in the development of
structure based drug design by reveling active site, binding site and conformational
dynamic.
 The determination of three-dimensional protein structures at an atomic resolution is the
key to unlocking an understanding of biological functions and the molecular mechanisms
of diseases. Although the established experimental methods, such as X-ray
crystallography,Nuclear Magnetic Resonance (NMR) [8,9], and cryo-electron microscopy.
 Computational method such as homology modeling, AB initio modeling and alphaFold
use to generate the 3D structure of protein molecule.
Homology modeling:
Homology modeling is known as comparative modeling is an computer based approach to
predict the protein structure base on the sequence similarity to known structure or template.
It is Commonly used in structural biology and drug discovery when experimental techniques
like X ray crystallography and NMR spectroscopy are unavailable. The main idea is that
protein sequence with unknown structure is searched against the sequence database of
proteins where structures of all proteins are experimentally known and the unknown
structure is modeled from the evolutionarily closest or best matching protein from the
database.
Homology modeling starts with knowledge of the structure of a number of proteins and their
sequences which have been determined by experimental methods. The method uses this
previous knowledge to predict the structure of proteins for which we know the sequence, but
don’t yet know the 3D structure. For predicting the structure of the protein, we’ll first predict
the co-ordinates of N, Ca, Cb (backbone) and then the co-ordinates of the R-group (side-
chain) of each amino acid.
Steps involve in Homology modeling:
1. Template recognition and initial alignment. The best matching protein sequence from
the database, to our target is assumed to be the evolutionarily closest and its structure will
be used as a template to the model the structure of the target. The database search tool also
gives an alignment,i.e. Information of which regions of the target match which regions of the
template.
2. Alignment correction: Align the target sequence with template sequence to determine
conserved region. Multiple sequence alignment are useful for identifying regions that are highly
divergent, and hence better detecting the appropriate locations for insertions and deletions.
3. Backbone generation: After the target-template alignment is optimized, the protein
backbone structure (N-Ca-Cb) for the target is generated. This is achieved by simply copying the
coordinates of the template backbone to the target based on the alignment.
4. Loop modeling: There are two main approaches to model loops: knowledge based and
energy based. The former approach searches for conformations of loops that have similar
sequences and endpoints as the target in the database of known structures. The latter models
the loop conformation in an ab initio manner by predicting the loop structures with lowest
structural energies using force × eld functions and molecular dynamics. These methods provide
fairly accurate results for short loops with up to 5-8 residues.
5. Side chain modeling: Modeling the side chains involves predicting the value of Ca -Cb
torsion angle for each Rgroup attached to the backbone. The conformation of side-chains in the
structures, also called the rotamers, depends on the values on this torsion angle.
6. Model optimization: Now that all the aspects of protein structure are modeled for the
target, it is time make×ne changes in the structure to reduce the overall energy. This is
achieved in an iterative manner. In each iteration, the backbone conformations and rotamer
conformations are changed alternatively to lower the overall energy of the predicted structures.
Model optimization can also be performed by running a molecular dynamics simulation, which
starts with the current predicted structure, and makes small changes to the structure based on
the simulation.
Ab initio ( De Novo) modeling: Ab initio prediction is the challenging attempt to predict
protein structures based only on sequence information and without using templates. The
various another ab initio approach available are TOUCHSTONE-II, ROSETTA, and the most
widely used I-Tasser that is based on the Monte-Carlo algorithm. The ab initio approach
involves the basic protocol that starts with the searching of various conformations based
upon the basic amino acid sequences and leads to the generation of native folds. After the
recognization, as well as prediction, is done, the assessment of the model is done to analyze
the standards of the predicted structures.
Alpha fold: AlphaFold is an AI system developed by Google DeepMind that predicts a
protein’s 3D structure from its amino acid sequence. It regularly achieves accuracy
competitive with experiment.
X-ray crystallography:
X-ray protein crystallography is a technique by which it is possible to determine the three
dimensional positions of each atom in a protein. Now over 100 years old, x-ray
crystallography was first used to determine the three dimensional structures of inorganic
materials, then small organic molecules, and finally macromolecules like DNA and proteins.
To date, about 100,000 protein structures have been published in the Protein Data Bank,
with almost 10,000 added every year. To use this technique, the crystallographer obtains
protein crystals, records the diffraction pattern formed by x-rays passed through the
crystals, and then interprets the data using a computer. The result is a atomic-resolution
model of a protein.
Structure determination by NMR spectroscopy:
The structure determination by NMR spectroscopy depends critically on the
assignment of the NMR signals of the NMR active nuclei of the protein. NMR
spectroscopy has information about the structural geometry around every NMR
observable nuclei in the protein. This information may be read from the NMR
signals of the nuclei. It is therefore important that the signals in the NMR spectrum
have been assigned to the correct nuclei, those, which give raise to the signals.
From the NMR signals the distance to the near-by nuclei in the structure can be
read using NOE spectroscopy, the dihedral angles can be determined by coupling
constants and chemical shifts, the relative angles of a number of bond vectors can
be determined by residual dipolar coupling.
NMR spectroscopy has unfolded many mysteries of the molecule behaviors. Protein NMR
spectroscopy help to study the changes in protein conformation, denaturation, and internal
mobility, pH titration of individual ionizable amino acid side chains in enzyme active sites,
observation of hydrogen-bonded imino protons in tRNA, investigation of paramagnetic
centers in metalloproteins, etc[1]. This method measures the short distances and angles
between different protons and then computationally the protein structure is derived. The
modern spectroscopy of proteins requires multidimensional experiments that involve 1H, 13C,
and 15N nuclei in isotopically labeled proteins[1].
Cryo electro microscopy: Cryo EM is an advanced imaging technique used to elucidate
the three-dimensional (3D) structure of biological molecules and complexes at near-atomic
resolution. In this technique, the sample is rapidly frozen to temperatures below -150 °C,
trapping it in vitreous ice. The sample is then imaged from different angles using an
electron microscope, where a beam of electrons passes through the specimen, resulting in a
set of two-dimensional (2D) projections. Advanced computational algorithms are then used
to reconstruct a 3D model of the sample from these projections. The cryo EM era started in
the early 1980s with the development of a methodology that applied a flash freezing
approach that rapidly cools the samples at temperatures below the water glass-transition
temperature (around -140 °C) within milliseconds. This procedure avoided the nucleation of
crystals and generated amorphous ice in a process known as “vitrification”.
References:
1.Héctor Zamora Carreras, PhD,Cryo electron microscopy allows scientists to peer into the
microscopic world with unprecedented clarity and detail. Article, May 24, 2024.
2. Anastassios C Papageorgiou et al, Protein Structure Analysis and Validation with X-Ray
Crystallography
Methods, Mol Biol. 2021.
3.LibreTexts physics.
4. Tobias Schmidt, Andreas Bergne, Torsten Schwede , Modelling three-dimensional protein
structures for applications in drug design.
Thank you