Perspectives of Structure-Sequence Dependent Stability of Collagen and Interaction of

Polyphenol Molecules with Collagen

V. Subramanian

Chemical Laboratory

Central Leather Research Institute Chennai

Introduction

Collagen is an extremely important protein, which provides mechanical strength and structural integrity to connective tissues

Nineteen different collagen types identified till date The identifying motif of the collagen is triple helix Prof Ramachandran and co-workers and Rich and

Crick The Gly-X-Y is the general repeating sequence of

the collagen (33% of Gly) Mutations in collagen chain can render the fibrils

unstable

Triple Helix

The collagen triple helix constitutes the major motif in fibril forming collagen and also occur as a domain in non-fibrillar collagens

Hydrogen bonding and presence of high content of imino acids provide stability to the three dimensional structure of collagen

The role of water mediated hydrogen bonding and hydration also play a crucial role in the stability of collagen

Recent experimental studies revealed that the presence of Arg in the Y position provides equal stability when compared to Gly-Pro-Hyp

The destabilizing nature of Asp in the Y position is also evident from the experimental studies

Therefore assessment of propensity of various amino acids to form collagen like peptides is an important area of research activity

Collagen Structure: An Indian Origin

Single Vs two hydrogen bond(s) If X and Y positions are imino acids, there is no

possibility of forming two hydrogen bonds The incorporation of other amino acids

provides a possibility of readdressing this question

StabilizingInter-strand

H-bonds

Collagen Triple Helix

Propensity of Various Amino Acids to Form Collagen Like Motif: Guest Host Approach

Propensity of various amino acids to form alpha helix and beta sheet have been addressed

Host-Guest peptide approach has been used to estimate the propensity

The presence of various amino acids not only influences the three dimensional structure but also the stability of collagen

The amino acid propensity-stability-function is an important area of research in molecular biophysics

Several experimental studies have been carried out on model collagen-like peptides to establish the propensity of various amino acids to form collagen

Circular dichrosim Spectroscopy has been used to develop triple helix propensity of various amino acids

The molar elipticity was monitored at 225nm while sample temperature was increased from 0 to 800 C

The melting curves were used to calculate fraction of folded states

Fraction folded has been used to compute vant Hoff enthalpies and free energy

These information provides experimental basis for predicting relative stabilities of various amino acids to form collagen like structure

Propensity scales will be used to compare the results obtained from modeling and simulations

Triple Helix Propensity Scale

)(

)(13)(

320

TF

TFcTK

4

3ln1exp)(

20

0 c

T

T

RT

HTK

m

Unraveling the Stability of Collagen: Experimental Studies by Brodsky and Co-workers

Host-Guest approach has been used to introduce new sequences in the general repeating Gly-Pro-Hyp sequences

Parameters such as melting temperature thermodynamics parameters from melting studies G of stabilization

of the host-guest collagen-like peptides have been studied to identify the influence of amino-acids towards the stability of collagen

Thermodynamics parameters for the Guest Host peptides

Biochemistry, 1996, 32, 10262.

Propensity data from Brodsky work

Biochemistry, 2000, 39, 14960.

Melting Temperature & Thermodynamic parameters for the Guest Host peptides in Y position

Melting Temperature & Thermodynamic parameters for the Guest Host peptides in X position

Collagen in Diseases

Mutation in collagen genes COL1A1 and COL1A2 leads to Osteogenesis Imperfecta (OI), a brittle bone disease

A point mutation in one of types collagen genes can cause disease

One of the main cause for OI is GlyAla mutation Glycine substitutions to another amino acid more severe

than mutations of X or Y in Gly - X - Y triplet. Understanding the stability of collagen upon mutation

becomes necessary Since collagen is a large protein, it is difficult to study the

influence of amino acids Various attempts have been made to probe the effect of

mutation in model collagen-like peptide sequences

Collagen in Diseases

Mutations in collagen leads to Osteogenesis Imperfecta (Type –I), Chondrodysplasis (type II), Ehlers-Danlos syndrome (type III), Alport syndrome (type IV), Bethlem myopathy (type VI) etc

Mutation in collagen genes COL1A1 and COL1A2 leads to Osteogenesis Imperfecta (OI), a brittle bone disease

A point mutation in one of type I collagen genes can cause disease

One of the main causes for OI is GlyAla mutation Glycine substitutions to another amino acid is more

severe than mutations of X or Y in Gly - X - Y triplet

Understanding the stability of collagen upon mutation becomes necessary

Collagen mimics and Biomaterial applications

Various physical and chemical properties make collagen as a versatile material for biomaterial applications

Studies on Collagen mimetics have been made to understand the strength of triple helix and for their application in biomaterials

In collagen mimetics, a variety of unnatural amino acids are incorporated in X and Y positions

K. N. Ganesh* and his coworkers have used 4 amino proline containing collagen like sequences

Murray Goodmann$ and his group made an attempt to template assembling of collagen like peptides using conformationally constrained organic molecule

*JACS, 1996, 118, 5156$JACS, 2001, 123, 2079

Frequency of Occurrence of Selected Triplets in Collagen

Propensity of Various Amino Acids to Form Collagen Like Motif

The propensity of various amino acids to form alpha helix and beta sheet have already been established

Host-Guest peptide approach has been used to estimate the propensity

The presence of various amino acids not only influences the three dimensional structure but also the stability of collagen

The amino acid propensity-stability-function is an important area of research in molecular biophysics

Several experimental studies have been carried out on model collagen-like peptides to establish the propensity of various amino acids to form collagen

Issues addressed

To determine the stability of collagen upon substitution of Gly-Pro-Hyp by other collagen-like triplets

To develop the propensity scale for various amino acids to form collagen-like peptides based on free energy of mutation

To probe the interaction between model collagen like peptides with polyphenols

Methodology

Ab initio and DFT calculations have been performed on collagen like triplets in both collagen and extended conformation

Free energy of solvation for these triplets have been computed for both conformations using Polarizable Continuum Method

Free energy of solvation has been used to compute the stability and amino acid propensity

The stability of these peptides have also been analysed by calculation of hardness

Free energy of various triplets have also been computed using classical molecular dynamics simulations

Using these values free energy change has been quantified

Model Collagen Triplets for Ab initio and DFT calculations

Gly-Pro-Hyp collagen-like conformation

Gly-Pro-Hyp Extended conformation

Superimposed structures of Gly-Pro-Hyp in both conformations

Relative Energy of Proline Containing Triplets

Relative Energy of Hyp Containing Triplets

Relative Energy of Triplets without Imino acids

Important Observations

The triplets containing proline or hydroxy proline are more stable in collagen-like conformation

Proline sterically restricts the N-C rotation and it has limited values of , – 63 ±15 degrees

Hence, proline can not be found in other known major protein motif

The dihedral angle corresponding to conformational energy minima for proline has been found to be –75 and 145o (, )

It can stabilize secondary structure of protein only when the allowed values of all other amino acids coincide with that of proline

Important Observations ………Contnd.

It is evident from the relative energy that Gly-Gly-Hyp does not stable in collagen like conformation

Recent experimental evidence confirms that glycine in the second position destabilizes the collagen triple helix

Solvation drastically alters the relative energy Proper ordering of the stability of various triplets needs

geometry optimization in solvent media

Free Energy Solvation

Important Observations

Solvation free energy of collagen-like sequences indicates that the triplets in collagen-like conformation can be hydrated better than its extended counterpart

The presence of polar and non-polar residues in the sequence drastically influences the solvation

Specifically Arg either in second or third position influences the solvation

Arg stabilizes the collagen-like sequence similar to stability provided by Hyp in the Y position

Free Energy Cycle

1

3 4

2

Assessment of Stability Using G

The propensity to form collagen-like sequence has been calculated using Gly-Pro-Hyp as reference

The calculated G ranges from 0.0 to 15.8 kcal/mol

The change in the free energy of Gly-Pro-Pro and Gly-Pro-Flp is close to Gly-Pro-Hyp

The most stable sequence is Gly-Pro-Hyp The general trend correlates well with the

experimental values derived from melting temperature studies on model systems

Triplets Involved in the Stability of Collagen: A Propensity Scale

-helix -turn

-sheet

A Propensity Scale Collagen

Chemical Hardness

Global hardness of various triplets in collagen-like and extended conformation has been calculated

It interesting to note that the chemical hardness values are more for the triplets in collagen-like conformation than extended conformation

Experimentally, Asp in the triplet does not favor collagen folding

Chemical hardness for Gly-Pro-Asp is observed to be less compared to the other sequences

Important Observations

B3LYP/6-31 G* level of theory predicted that collagen-like conformation of the Gly-Pro-Hyp is stable than the extended conformation by 0.46 kJ/mol

Hardness of triplets of sequences Gly-X-Y (without Hyp and Pro) is lower than the triplets containing Pro and Hyp residues

Emerging Roles of Computational Techniques in Tanning Theory

Computer model of bovine type I collagen has been simulated

Early report of molecular modeling of tanning processes has been made

Model peptides for collagen has been selected and interactions with various tanning materials simulated using force field as well as Density Functional Theoretical methods

Binding energies for various interactions of collagen like peptide with tannin molecules have been estimated

Computer Simulation of Collagen –like Peptide-Tannin Interaction

Collagen -Catechin Collagen -Epicatechin Collagen –Gallic Acid

Collagen -Quercetin

Interaction of gallic acid collagen like peptides

Gallic acid is a good anti-oxidant present in many plant sources

Gallic acid has been shown to selectively induce cell death in cancerous cell lines by binding to specific receptors or enzymes

Gallic acid finds major role in stabilization of collagen in tanning process of leather making

Collagen is an important and abundant connective tissue protein in animal kingdom

Gallic Acid

Collagen assemblies are stabilized by covalent and non-covalent interactions

A fundamental understanding on the interaction of gallic acid with collagen is important to unravel the nature of interactions that are required for the stabilization of collagen matrix

Theoretical calculations can be used for the determination and quantification of such interactions

In this view present investigation focuses on determining the interactions of different dipeptides with gallic acid

Such a study can not only be correlated to stabilization process involved in collagen but also will lead to the advancement on the knowledge of peptide-ligand interaction

Computational details

Three classical dipeptides of amino acids glutamic acid, lysine and serine chosen for the interaction studies with gallic acid

Dipeptides imposed with the and corresponding to the angles of collagen

Dipeptides and gallic acid built and energy minimized using modules available in Insight II(MSI, USA)

Four functional sites namely 3 OH groups and one COOH group present in the gallic acid identified to have the potential to interact with the side chain groups of the dipeptide

The geometry of the complexes optimized by a semi-empirical PM3 method using Gaussian98 suite of programs

Energy of the complex calculated using both Hartree Fock (HF) and DFT based B3LYP methods using 3-21G* & 6-31G* basis sets employing Gaussian 98w suite of programs

The interaction energy (VINT) calculated using supermolecule approach

VINT = TEcomplex – [ TEdipeptide + TEgallic acid]

Binding Energy (VBE) is, VBE = - VINT

Molecular electrostatic potentials (MESP) are useful in understanding the weak and non-covalent interactions. The electrostatic potential V(r) is defined as

ZA is the charge on nucleus A located at RA, and (r') is the electron density at a point r

MESP features of peptide-gallic acid complex have been studied by BLYP/DN using DMOL implemented in Cerius2

Molecular Dynamic (MD) calculations have been done for one of the complexes, to see the time evolution of the hydrogen-bonded complex. A time step of 1fs has been chosen and the MD simulations have been performed for 600 ps including an equilibration period of 100 ps

'

')'()(

rr

drr

Rr

ZrV

A

A

Discussion The functional groups para-OH, two meta-OH and COOH of

gallic acid have been assumed to act as a hydrogen bond donor/acceptor for different side chain groups of amino acids in dipeptide

Most of the complexes have exhibited hydrogen bonding in the complexes consisting of dipeptides-gallic acid

Complexes have exhibited binding energies in the range of 4 – 18 kcal/mol

Complexes of glutamic acid dipeptide with gallic acid have all exhibited hydrogen bonding with high binding energies

Some of the complexes of gallic acid with serine and lysine dipeptide have also exhibited hydrogen bonding

The interaction with COOH group of gallic and side chain COOH group of glutamic acid exhibited the maximum binding energy

All complexes calculated by HF methods predicted lower binding energies when compared to the binding energies predicted from DFT methods

Molecular electrostatic potential estimation of various complexes provided clues on the involvement of the electrostatics involved in the interaction process

Dipeptide

Binding energies of gallic acid – dipeptide complex (kcal/mol)

m1 – OH (C1) p – OH (C2) m2 – OH (C3) COOH (C4)

3-21G*

6-31G*

3-21G*

6-31G*

3-21G*

6-31G*

3-21G*

6-31G*

Glutamic Acid

10.29 8.08 7.06 5.06 11.19 8.98 18.18 15.1

Lysine 5.51 3.75 11.65 9.51 8.97 6.53 14.51 12.18

Serine 7.7 6.32 10.96 10 6.23 5.1 12.92 10.19

Interaction energies of different sites of gallic acid with different dipeptides calculated using Density Functional Theory (B3LYP) with basis set 3-21G* and 6-31G*

Dipeptide

Binding energies of gallic acid dipeptide complex (kcal/mol)

m1 – OH (C1) p – OH (C2) m2 – OH (C3) COOH (C4)

3-21G*

6-31G*

3-21G*

6-31G*

3-21G*

6-31G*

3-21G*

6-31G*

Glutamic Acid

8.67 6.0 6.12 3.85 9.3 6.7 16.86 13.39

Lysine 3.55 –3.25 11.57 9.15 7.95 5.22 12.6 10.33

Serine 7.28 5.91 10.65 9.34 6.02 4.39 12.59 9.46

Interaction energies of different sites of gallic acid with different dipeptides calculated using Hartree Fock (HF) method with basis set 3-21G* and 6-31G*

Hydrogen Bonded Complexes of Glutamic acid Dipeptide and Gallic acid

C4C3C2

C4C2

Hydrogen Bonded Complex of Serine Dipeptide and Gallic acid

C2

Hydrogen Bonded Complexes of Lysine Dipeptide and Gallic acid

-ve MESP of serine gallic complex (C1)

-ve MESP of glutamic-gallic complex (C4) -ve MESP of lysine gallic complex (C2)

The functional groups para-OH, two meta-OH and COOH of gallic acid have been assumed to act as a hydrogen bond donor/acceptor for different side chain groups of amino acids in dipeptide

Most of the complexes have exhibited hydrogen bonding in the complexes consisting of dipeptides-gallic acid

Complexes have exhibited binding energies in the range of 4 – 18 kcal/mol

Complexes of glutamic acid dipeptide with gallic acid have all exhibited hydrogen bonding with high binding energies

Collagen-like Peptide Sequence

Difficult to handle large systems like collagen molecule for molecular simulation calculations

Interaction studies can be carried by building collagen like peptide sequence maintaining the uniqueness of collagen

A 9-mer sequence Ace-Gly-Pro-Hyp-Gly-Ala-Ser-Gly-Glu-Arg-Nme is built based on the repeatability of the sequences and on the presence of amino acids in the actual collagen molecule by imposing and constraints based on G N Ramachandran plot

Peptide sequence and polyphenolic molecules minimized using CVFF(Consistence Valence Force Field)

The polyphenolic molecules placed near the different sites of the peptide sequence and minimized

The binding energy of the molecules with the peptide sequence calculated based on the equation,

EB - Binding Energy (kcal/mol)

Epolyphenolics = Total energy of the minimized structure of polyphenolic molecules (kcal/mol)

Esequence = Total energy of the minimized structure of collagen-like peptide sequence (kcal/mol)

Interaction of Polyphenolics with Collagen-like Peptide Sequence

complexsequenceicspolyphenolB EEEE

Interaction of Polyphenolics with Collagen-like Peptide Sequence

Hydrogen bonded complex of Gallic acid with Serine residue of collagen like peptide (Binding Energy = 8 kcal/mol )

Hydrogen bonded Complex of Catechin with Peptide (Binding Energy = 18 kcal/mol )

Complex of Quercetin with collagen like peptide (Binding Energy = 12 kcal/mol )

Molecular Electrostatic Potential Surface (MESP) of Gallic acid–Collagen-like Peptide Complex

Positive electrostatic potential (0.7) surface of the complex of gallic acid with Glutamic acid residue of the peptide sequence

Negative electrostatic potential (-0.01) surface of the complex of gallic acid with Glutamic acid residue of the peptide sequence

Binding Energies of Polyphenol-Collagen-like Peptide Complexes

Complexes 

Binding Energies (kcal/mol)

Catechin Quercetin Gallic acid

1 19.10.2 13.70.2 8.20.1

2 16.40.1 16.20.2 7.10.1

3 15.60.2 12.20.1 6.10.2

[1]– Polyphenolic molecule interacted around the serine and glutamic acid residue of the collagen-like peptide sequence.[2]– Polyphenolic molecule interacted around the arginine residue of the model collagen-like peptide sequence.[3] –Polyphenolic molecule interacted around the hydroxyproline residue of the model collagen-like peptide sequence.

Lessons from Molecular Modeling Studies

Molecular modeling studies have provided a basis to identify the interaction process involved in tanning

Catechin exhibited stronger binding, as compared to other polyphenolics chosen for the study

Many of the complexes exhibited hydrogen bonding and some exhibited electrostatic and weak interactions

MESP has revealed a lock and key type of electrostatic interactions involved in the stabilization of gallic acid and collagen-like peptide complex

Geometrical Issues in binding small molecules by collagen; A Prospective

Analysis

Computational Details

Four representative polyphenol molecules viz., gallic acid, catechin, epigallocatechingallate and pentagalloylglucose chosen for interaction studies

24-mer collagen triple helix corresponding to residues 193 to 216 (21 and 12 chains) of the native Type I collagen is constructed using the GENCOLLAGEN package

Following is the amino acid sequence of triple helix,

[Gly-Glu-Hyp-Gly-Pro-Hyp-Gly-Pro-Ala-Gly-Ala-Lys-Gly-Pro-Ala-Gly-Asn-Hyp-Gly-Ala-Asp-Gly-Gln-Hyp] 1

[Gly-Glu-Val-Gly-Leu-Hyp-Gly-Leu-Ser-Gly-Pro-Val-Gly-Pro-Hyp-Gly-Asn-Ala-Gly-Pro-Asn-Gly-Leu-Hyp] 2

The 24-mer triple helix and polyphenols are minimized using CVFF with a dielectric constant of 4.0

Collagen - an inside out proteinSide chain hydroxyl group of the amino

acids, serine and hydroxyproline, carboxyl group of aspartic acid, amino group of lysine and amide group of aspargine are potential interacting sites for formation of hydrogen bonds with polyphenols

Energy minimized structures of polyphenols

Vegetable Tannins

Catechin

Gallic acid

Epigallo Catechin Gallate

Penta galloylglucose

Energy minimized structure of 24-mer collagen triple helix

Complex between aspargine of T.Helix and gallic acid

Complex between aspartic acid of T.Helix and catechin

Complex between lysine of T.Helix and epigallocatechingallate

Complex between aspargine of T.Helix and pentagalloylglucose

Binding Sites in triple helix

Binding Energy (Kcal/mol)

Gallic acid (Gal)

Catechin (Cat)Epigallocatechi

ngallate (EGCG)

Pentagalloyl glucose (PGG)

9th residue Ser of C-chain (α2)

16.5 22.5 35.2 56.6

6th residue Hyp of A-chain (α1)

14.5 20.8 34.5 48.4

12th residue Lys of B-chain (α1)

19.2 23.8 37.9 41.1

21st residue Asp of A-chain (α1)

18.4 20.0 38.2 59.8

17th residue Asn of C-chain (α2)

14.1 23.7 34.3 52.8

Binding energies different complexes between polyphenols and triple helix

Interaction Site

Gallic acid (Gal) Pentagalloylglucose (PGG)

H-bondBond Dist Å

Bond Angle

H-bondBond Dist Å

Bond angle

9th residue Ser of C-chain (α2)

SerC9-(Cα)-C-O…H(3)O-Gal- 3.02 141

HypC15-(Cα)-O-H…O(19)-PGG

AsnA17-N-H…O(19)-PGGAlaA15C=O…H(20)-PGG

SerC9-(Cα)-C-O…H(10)O- PGG

2.843.172.762.93

177156156121

6th residue Hyp of A-chain (α1)

AspB21C=O…H(3)O-Gal 2.97 138

GluB2-(Cα)-C-O-H…O(15)-PGG

GluB2-(Cα)-C-O…H(24)-PGG

HypB3C=O…H(23)-PGGHypB3C=O…H(18)-PGGLeuC5-N-H…O(12)-PGG

3.042.882.963.083.17

163139131157149

12th residue Lys of B-chain (α1)

LysB12-(Cα )N-H…O(2)- Gal 3.28 126HypA18-(Cα)-O-H…O(3)-PGG

AsnB17-(Cα)-N-H…O(2)O-PGG3.093.12

122141

21st residue Asp of A-chain (α1)

AspB21-(Cα)-O-H…O(3)-Gal

HypB18-C=O…H(6)O-Gal2.892.91

128147

AspA21-N-H…O(9)-PGGGlyA19C=O…H(13)O-PGG

2.962.83

164174

17th residue Asn of C-chain (α2)

AsnA17-(Cα)-C=O…H(3)O-Gal 2.84 151AsnB17-(Cα)-C=O…H(8)O-PGG

HypA18C=O…H(4)O-PGGGlyA16N…H(3)O-PGG

3.272.9

3.42

159140161

Hydrogen bonds of complexes; their length and angle

Interaction Site

Catechin (Cat) Epigallocatechingallate (EGCG)

H-bondBonddist Å

BondAngle

H-bondBond dist Å

Bondangle

9th residue Ser of C-chain (α2)

SerC9-(Cα)-O-H…O(2)-Cat 3.04 161LysA12C=O…H(9)O- EGCGSerC9-C=O…H(3)O- EGCG

2.822.79

148132

6th residue Hyp of A-chain (α1)

HypA6-(Cα)-O-H…O(1)-Cat

AlaB9-N…H(12)O-Cat3.023.18

127137

ProB8-N… H(13)O- EGCG 3.25 142

12th residue Lys of B-chain (α1)

LysB12C=O…H(11)O-Cat 3.1 126 LysB12-(Cα)-N-H…O(2)- EGCG 3.41 150

21st residue Asp of A-chain (α1)

AspA21-(Cα)-C-O-H…O(4)-Cat

AlaB20-NH…O(2)-CatGlnA23-(Cα)-N-H…O(6)-Cat

3.083.223.24

150133164

AspA21-N-H…O(2)-EGCGGlnA23-(Cα)-N-H…O(6)-EGCG

3.33.26

147146

17th residue Asn of C-chain (α2)

GlyA16C=O…H(14)O-CatAsnA17-(Cα)-C=O…H(11)O-Cat

3.002.92

146151

GlyA16C=O…H(12)O- EGCGHypA18C=O…H(9)O- EGCGHypA18C=O…H(3)O- EGCG

AlaA20-N-H…O(2)- EGCG

2.992.822.9

3.13

140162156143

Total and contact surface areas of the collagen like triple helix and polyphenols in Å2

Collagen(24-mer)

Cat EGCG PGG Gal

CSA 1164 120 163 275 84

TSA 3825 268 382 688 160CSA – Contact surface

areaTSA – Total Surface Area

Gal Cat EGCG PGG

AT BT AT BT AT BT AT BT

Ser 92 61 110 78 219 124 462 205

Hyp 85 65 120 75 248 115 421 197

Lys 151 82 176 94 279 135 357 189

Asp 102 71 112 76 186 115 514 238

Asn 84 69 124 86 214 125 368 196AT – Solvent inaccessible Total Surface Area

BT – Solvent inaccessible Contact Surface Area

TSA of the complexes are in the range of 3840 – 4160CSA of the complexes are in the range of 1160 – 1250

Solvent inaccessible surface areas of the complexes in Å2

Plot of interfacial interacting volume Vs Binding energy of the complex

Interacting Interfacial Volume (Å3)

Plot of effective solvent inaccessible contact volume Vs Binding energy of the complex (inset): Plot of effective solvent inaccessible contact surface area Vs Binding energy of the complex

Plot of inverse of interacting interfacial volume (1/Int.Vol.) Vs inverse of binding energy(1/B.E) of the complexes

Ligation phenomena in collagen is being influenced by geometric parameters

Collagen complexation with small polyphenolic molecules, there may exist some minimum geometrical sizes and binding energies for influencing the long range ordering processes

Ability of polyphenol bearing flavanoid structure in management of arthritis and tanning may well result from their ability to reduce accessibility of solvent(water) to molecular surfaces of collagen

The present investigation offers the possibility to understand further recognition of phenomena associated with protein-protein and DNA-protein interactions in general, based on interfacial volume and surface areas