Luận án tiến sĩ: Physical and chemical analyses of melanins

Chapter One: Background Chapter Two: Insights to the Metal Binding to Melanin and its Effects on the Aerobic Reactivity of Melanin Chapter Three: Physiochemical Analysis of Melanosomes

PHYSICAL AND CHEMICAL ANALYSES

OF MELANINS

By Lian Hong Department of Chemistry Duke University

Dissertation submitted in partial fulfillment of

the requirements for the degree of Doctor

of Philosophy in the Department of Chemistry in the Graduate School

of Duke University

2006

UMI Number: 3244835

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy submitted Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted Also, if unauthorized copyright material had to be removed, a note will indicate the deletion

® UMI UMI Microform 3244835 Copyright 2007 by ProQuest Information and Learning Company

All rights reserved This microform edition is protected against unauthorized copying under Title 17, United States Code

ProQuest Information and Learning Company

300 North Zeeb Road P.O Box 1346 Ann Arbor, MI 48106-1346

An abstract of a dissertation submitted in partial

fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Chemistry in the Graduate School of

Duke University 2006

Abstract Melanin is a pigment that is ubiquitous in the biological world In addition to

its obvious function of adornment, melanin is believed to serve as a metal reservoir

and sunscreen This work examined the metal binding behavior, the chemical structures and the interaction with light of natural melanins isolated from several different sources

Sepia eumelanin was used as a model to study the metal binding behavior of eumelanins Infrared spectroscopic analysis suggests that Mg(II), Ca(II) and Zn(II) are

mainly bound to carboxyl groups and Cu(II) binds to hydroxyl groups at pH ~4 The effect of metal content on the aerobic reactivity was examined by analyzing the ability

of melanin to nick DNA Of the metals studied, Cu(II)- and Fe(III)-loaded melanin

showed the most damaging effects to DNA

The photoreactivity of melanin is most important in two parts of the body: the

skin and the eye We first studied melanosomes from hair, which correlate well with

the pigment in the skin Chemical analysis showed that eumelanin was the major pigment in the melanosomes in black hair, while the melanosomes from red hair contained both pheomelanin and eumelanin Photoelectron emission microscopic

(PEEM) analyses on the two melanosomes revealed that the photoionization threshold

of eumelanin was 4.4 eV and that pheomelanin was 3.8 eV, respectively The low threshold of pheomelanin may contribute to the high incidence of skin cancer in

redheads

11

We then examined the physical and chemical properties of bovine ocular melanosomes The collective data showed that iris melanosomes had a higher degree

of crosslinking and conjugation than choroid and RPE melanosomes of the same age

group The content of carboxyl groups and the degree of conjugation in melanosomes were smaller in older samples These tissue and age-related differences are likely to

result in variation of efficiency in melanin functions such as metal binding and light absorption

Finally, we examined the photoionization properties of human RPE

melanosomes and lipofuscin, another important pigment in the RPE RPE

melanosomes became more pro-oxidant with aging and lipofuscins were more pro- oxidant than melanosomes

iv

Chapter One: Background

Chapter Two: Insights to the Metal Binding to Melanin and its Effects on the

Aerobic Reactivity of Melanin

Chapter Three: Physiochemical Analysis of Melanosomes Extracted from

Human Black and Red Hair

Chapter Four: Physiochemical Analysis of Melanosomes Extracted from

Bovine Choroid, Iris and Retinal Pigment Epithelium

Chapter Five: Age-Dependent Ionization Potentials of Melanosomes and

Lipofuscin Isolated from Human Retinal Pigment Epithelium Cells

References

Biography

Page ili

vi

Vii

1X xii

The pH values of salt solutions after suspension of EDTA-washed

melanin

Summary of the changes in the IR spectra of melanin upon loading of

different metals

Chemical degradation analyses of human hair melanosomes

Metal ion contents of human hair melanosomes

Summary of the size analysis of the melanosomes isolated from

mature bovine eyes

Summary of the size analysis of the melanosomes isolated from

newborn bovine eyes

The contents of Na(I), Mg(II), K(I), Ca(ID and Zn(II) in bovine ocular

melanosomes determined by ICP-MS

Summary of amino acid analyses for bovine ocular melanosomes

Summary of chemical degradation analyses of bovine ocular

melanosomes

C, N and H mass percentages in melanosomes obtained by elemental

analysis are corrected according to their amino acid and metal

contents

C/N and O/N molar ratios on the surface of bovine ocular

melanosomes determined by XPS

Composition of carbon functional groups

Composition of oxygen functional groups

Composition of nitrogen functional groups

The peak positions of the aromatic rings and carbonyls in the IR

spectra of bovine ocular melanosomes

Summary of the size analysis of human RPE melanosomes from three

Proposed typical structures of eumelanin and pheomelanin based on

biosynthetic pathway and chemical degradation analysis of melanin

IR spectrum of Ca(II)-enriched melanin as a function of the solution

concentration of Ca(ID)

IR spectrum of Zn(I])-enriched melanin as a function of the solution

concentration of Zn(II) in the spectral region of a) 900-1900 cm” and

b) 2000-4000 cm”

IR spectrum of Cu(II)-enriched melanin as a function of the solution

concentration of Cu(II) in the spectral region of a) 900-1900 cm” and

b) 2000-4000 cm”

Image of a typical gel of pUC18 DNA with bands of two forms

Aerobic reactivity of melanins loaded with different metal ions in a)

dark and b) light reactions

The effects of [Fe(III] on the ability of melanin to break supercoiled

DNA in a dark reaction

AFM images of melanosomes isolated from a) black hair and b) red

hair

High magnification AFM images of black-hair melanosomes (a, b)

and red-hair melanosomes (c, d)

Integrated wavelength-dependent FEL-PEEM data for a) human black

hair and b) human red hair melanosomes

Time-dependent absorption of Fe(II)-Cyt ¢ following the mixing of

Fe(II)-Cyt c with equal masses of eumelanosomes and

SEM images of melanosomes isolated from a) mature bovine RPE, b)

newborn bovine RPE, c) mature bovine choroid, d) newborn bovine

choroid, e) mature bovine iris and f) newborn bovine iris

The distributions of the lengths of the long axis of a) newborn bovine

RPE melanosomes and b) mature bovine RPE melanosomes

The high-resolution X-ray photoelectron spectra of Cls in (a)

newborn choroid and (b) mature choroid melanosomes

The high-resolution X-ray photoelectron spectra of Ols in (a)

newborn choroid and (b) mature choroid melanosomes

IR spectra of a) newborn bovine choroid, iris and RPE melanosomes

b) mature bovine choroid, iris and RPE melanosomes in the range of

800-2000 cm `

The !C CP/MAS solid-state NMR spectra of a) mature choroid

melanosomes, b) EDTA-washed mature iris melanosomes, c) EDTA-

washed mature choroid melanosomes, d) EDTA-washed newborn iris

melanosomes and e) EDTA-washed newborn choroid melanosomes

SEM images of RPE melanosomes from patients a) 14, b) 59, and c)

76, years old, respectively

AFM micrographs of RPE melanosomes: a, b) a smooth granule and

c, d) a granule with a rough surface and e, f) a “transient” granule

SEM images of lipofuscin granules from donors a) 14, b) 59, and c)

76 years old

The emission spectra of RPE lipofuscin granules from donors a) 14, b)

59, and c) 76 years old

Integrated wavelength-dependent FEL-PEEM data for bovine ovoid

and rod-shaped melanosomes

Integrated wavelength-dependent FEL-PEEM data for melanosomes

isolated from human RPE cells from a) 14- and b) 76-year-old

specimens

Integrated wavelength-dependent FEL-PEEM data for lipofuscin

isolated from human RPE cells from a) 14- and b) 76-year-old

age-related macular degeneration chloroform-insoluble

chloroform-soluble

carboxyl

cross-polarization magic angle spinning

coefficient of variation Cytochrome

5,6-dihydroxylindole 5,6-dihydroxylindole-2-carboxylic acid dihydroxylphenylalanine

high performance liquid chromatography

inductively coupled plasma mass spectrometry iris pigment epithelium

infrared

1X

normal hydrogen electrode national institute of standards and technology nuclear magnetic resonance

neuromelanin

hydroxyl phosphate buffered saline

Dulbecco’s phosphate buffered saline without calcium and magnesium

Parkinson’s disease photoelectron emission microscopy

2,3,5-pyrroletricarboxylic acid

reactive oxygen specie retinal pigment epithelium relative sensitivity factor the length of the short axis scanning electron microscopy substantia nigra

transmission electron microscopy thiazole-4,5-dicarboxyl acid total melanin

X

TMS tetramethylsilane

XPS X-ray photoelectron spectroscopy

xi

Acknowledgements

First I would like to thank my thesis advisor Dr John D Simon for his guidance, support, encouragement and training throughout these past years His broad connections with research groups and scientists outside Duke are essential for the completion of my dissertation

I gratefully acknowledge our collaborators: Prof Kazumasa Wakamatsu, and Shosuke Ito for carrying out chemical degradation analysis; Bhavin B Adhyaru, Chi-

Yuan Cheng and Prof Clifford R Bowers for collecting '3C solid state NMR spectra; and Jacob Garguilo, Robert J Nemanich and Glenn S Edwards for performing

photoelectron emission microscopic analyses

I also appreciate the past and current members of the Simon Lab In particular,

I would like to thank Dr Yan liu for stimulating discussions, help of experiments and

sharing of her experience and knowledge with me I acknowledge her invaluable help

and her important inputs in chapter 2 and chapter 3 of this thesis I thank Dr

Alexander Samokhvalov for employing PEEM and analyzing black/red hair

melanosomes and Laura Anzaldi for help on the size analysis of human RPE

melanosomes I also thank Jenny Perry, Laura Lamb, Valerie Kempf, William Derek

Bush, Weslyn Ward and Jean Hatcher for various help

I owe special thanks to my husband Xiangqian for his patience, love, and

encouragement during these years of my studies Finally, I would like to thank my family and friends for their support

xii

Chapter One Background

Occurrence of Melanin and Melanogenesis Melanin is an omnibus term describing a large range of natural and synthetic phenolic-quinonoid pigments Since this current work is focusing on the characterization of natural melanins, only the properties of natural melanins will be discussed in this section Natural melanins are classified into

two groups according to their chemical structures and molecular precursors: the insoluble black to brown eumelanin derived from dihydroxyphenylalanine (DOPA),

and the alkaline-soluble yellow to red pheomelanin derived from DOPA and cysteine

(Crippa 1989; Ito 1993; Riley 1997; Ito 1998) Eumelanin is widespread in the biological world It occurs in the skin, hair, eye, brain and inner ear of mammals, the

feathers of birds, the ink sac of cuttlefish, and in some reptiles, amphibians and fungi

Pheomelanin, on the other hand, is mainly found in the feathers of birds and the hair

and skin of mammals

Melanin pigmentation occurs in a specialized organelle, the melanosome, in most pigment generating cells (Prota 1992) In addition to the melanin content, melanosomes contain structural proteins, enzymes, lipids, and metals (Prota 1992) Eumelanosomes, with eumelanin as the major pigment, are mostly integrated granules, uniformly and densely packed with pigments (Parakkal 1967) Pheomelanosomes, however, with pheomelanin as the major pigment, are unevenly packed with pigments and are loosely aggregated small granules (Parakkal 1967) Though eumelanin and

pheomelanin are chemically distinct, their melanogeneses are related and often occur

simultaneously and even in the same pigment generating cell, the melanocyte (Prota

2

1998) The formation of eumelanin or pheomelanin depends primarily on the access to

cysteine in the melanocytes (Prota 1998) The currently accepted melanogenesis pathway of melanin is shown in Figure 1-1 (Ito 2003)

The initial oxidation steps from tyrosine to dopaquinone are catalyzed by the

enzyme tyrosinase In the absence of cysteine, dopaquinone undergoes intramolecular cyclization, leading to the formation of 5,6-dihydroxyindole-2-carboxylic acid (DHICA) and 5,6-dihydroxyindole (DHI), which further polymerize and ultimately result in the formation of eumelanin (Ito 2003) In the presence of cysteine, dopaquinone will form a variety of cysteinyldopa adducts, which eventually form pheomelanin (Ito 2003) A variety of regulatory factors, including enzymes, hormones,

and metal ions, are present in the melanogenesis and mediate the synthesis of melanin

(Halaban 1990; Jackson 1992; Tsukamoto 1992; Prota 1998)

Structure of Melanin Melanin is insoluble, opaque and amorphous These characteristics obstruct the analysis of melanin by conventional analytical methods In order to study the chemical compositions of the pigment, a group of degradation methods have been developed to yield low molecular weight fragments (Benathan 1980; Ito 1985; Ito 1986; Crippa 1989) A significant amount of carboxyl-substituted pyrrole moieties (e.g., 2,3,5-pyrroletricarboxylic acid (PTCA)) and thiazole-4,5-

dicarboxyl acid (TDCA) with trace amounts of other thiazoles were identified from the permanganate oxidation products of eumelanin and pheomelanin, respectively

(Piattelli 1962; Fattorusso 1969; Benathan 1980; Ito 1985)

3

The hydrodic acid (HI) hydrolysis of pheomelanin forms aminohydroxyphenylalanine (AHP) (Minale 1967) PTCA was suggested as the

oxidation product of either DHICA or DHI with a connection at the 2-position

(Piattelli 1962; Ito 1986) TDCA and AHP were proposed as the products of

benzothiazines (Minale 1967; Fattorusso 1969)

Combined with information on melanin’s biosynthetic pathway, it has been proposed and generally accepted that DHI and DHICA in varying oxidized forms are the major molecular constituents of eumelanin; and benzothiazine together with its

derivatives are the molecular constituents of pheomelanin (Ito 1998) Typical

structures of eumelanin and pheomelanin as proposed by Ito, are shown in Figure 1-2 (Ito 1998) The chemical composition of melanins may vary with species, tissue origins, age, and storage history of samples

Properties and Biological Functions of Melanin Melanin possesses an intriguing range of physicochemical properties: wide-range light absorption, ion-exchanging ability, photoreactivity, redox reactivity, and free radical scavenging capacity (Prota 1998; Sarna 1998) The biological functions of melanin greatly depend on these

physicochemical properties Here, the author will only briefly describe the properties

and functions related to this research, including the metal ion binding property and photoreactivity

HạN“ `OOH

Figure 1-2 Proposed typical structures of eumelanin and pheomelanin based on

biosynthetic pathway and chemical degradation analysis of melanin The positions of

(COOH) are connected with H or COOH The arrows show possible sites for attachments of other monomers (Ito 1998)

Melanin, especially eumelanin, is capable of binding a variety of metal ions, such as Na(), K(J), Mg(Il), Cad), Zn(ID, Fe(III), and Cu(II) (Cotzias 1964; Bruenger

1967; Potts 1976; Prota 1992; Sarna 1998; Liu 2003; Liu 2005; Liu 2005) It was suggested that ~ 20% of the monomeric units can bind to metals, assuming an average

monomer molecular weight of 200 (Potts 1976) The binding of metals involves

coordination to carboxyl groups, phenolic hydroxyl, and amine groups in melanin The

binding behavior depends on various reaction conditions, such as metal and melanin

concentrations, pH, ionic strength of reaction solutions, reaction time, and temperature (Froncisz 1980; Sarna 1980)

This metal binding capacity enables melanin to act as a reservoir of metal ions,

e.g., for Ca(II) (Boulton 1998) In addition, the binding of reactive heavy metal ions, e.g., Cu(J), Cu(ID, Fe(II) and Fe(III), by melanin has been proposed to function to avoid the production of reactive oxygen species (ROS’s) by reactions catalyzed by these metals (Sarna 1992; Zareba 1995) On the other hand, it has also been shown that when melanin is saturated with Fe(III), the integrity of the melanin could be impaired, resulting in the release of Fe(III), which can induce severe cellular damages (Zareba 1995)

Melanin absorbs light over a wide spectral range The absorption of melanin is featureless, increasing monotonically with decreasing wavelength in the UV-Vis range (Crippa 1978; Sarna 1984; Nofsinger 2002) Due to the particulate nature of melanin, the absorption spectra are likely the result of a combination of light absorption,

reflection and scattering The absorption depends greatly on the aggregation status of

melanin in solution (Huang 1989; Nofsinger 2002)

The spectral aspects of melanin (with significant absorption in the UV region) and the fact that the portion of the population with dark skin has a smaller chance to develop skin cancer have led to the belief that the melanin pigment in the skin should

be protective It has been suggested that melanin photoprotects by absorbing light and quenching free radicals or ROS’s (Boulton 1991; Sarna 1992; Bustamante 1993;

Dunford 1995; Zeise 1995; Rozanowska 1999; Schraermeyer 1999) Light energy

absorbed by melanin can be transformed into heat (Crippa 1978) However, a large body of evidence has shown that melanin could also photosensitize, undergo photoionization or photolysis, and generate ROS’s when illuminated with UV light

(Felix 1978; Crippa 1983; Kalyanaraman 1984; Korytowski 1987; Hill 1992;

Rozanowska 1997; Rozanowska 2002) These works showed that the photoreactivity

has a strong wavelength dependence, increasing rapidly with decreasing wavelength,

and that the efficiency of melanin to produce and to quench ROS’s varies with the melanin composition, age, and metal contents

Thus, a delicate balance may exist between the interrelated protective and deleterious characteristics of melanin Study of the metal binding behaviors, chemical properties and photoreactivity of melanins may provide a foundation for understanding the mechanisms of these protective and harmful functions

Contents of This Work In this dissertation, the author studied the physicochemical

properties of natural melanins isolated from several sources in an effort to provide fundamental information for understanding the functions of melanins from a physical

and chemical standpoint

Metal binding properties were examined using melanin isolated from Sepia officinalis This melanin is known to be mostly eumelanin with less protein content compared to other natural melanins and is easy to isolate and purify In addition, the

size of Sepia melanin granules, ~150 nm, smaller than common melanosomes, makes its suspension and reaction with metal ions efficient The functional groups responsible for the coordination of metal ions were studied by infrared spectroscopy

(IR) The aerobic reactivity of melanins with different metal contents was analyzed by

examining their ability to nick supercoiled DNA (Chapter 2)

The photoreactivity of melanins is especially important for two of the most exposed locations in the body: the skin and the eye Since it is difficult to obtain skin

samples and to isolate melanosomes from the skin, we instead studied melanosomes

from hair, which correlate well with the types and amount of melanin in the skin (Thody 1991) We isolated melanosomes from black hair and red hair, examined their

chemical, morphological, and photoreactivities with various methods, including

chemical degradation analysis, atomic force microscopy (AFM) and photoelectron

emission microscopy (PEEM) (Chapter 3)

As for ocular pigments, bovine ocular melanosomes were used as a model to

study the chemical and physical properties of ocular pigments due to the limited

9

amount of human eye samples Melanosomes were isolated from the iris, choroid and

retinal pigment epithelium (RPE) of cattle at two age groups and analyzed The tissue-

related and age-related differences among samples were examined (Chapter 4) For human ocular pigments, we focused on their photoionization properties In addition to melanosomes, another important pigment in the human RPE, lipofuscin, was also studied Samples from donors at 14, 59, and 76 years old were compared and age-

related photoionization changes were examined (Chapter 5)

10

Chapter Two

Insights into Metal Binding to Melanin and its Effects on

the Aerobic Reactivity of Melanin

This work was originally published as Hong, L.; Liu Y and Simon J.D

Photochemistry and Photobiology, 2004, 80(3), 477-481 and Hong L and Simon J.D

accepted by Photochemistry and Photobiology

11

INTRODUCTION

Natural eumelanins are associated with and have a considerable affinity for di-

and trivalent metal ions (Potts 1976; Sarna 1998; Liu 2005; Liu 2005; Liu 2005) It is believed that this pigment can serve as a reservoir of metal ions, e.g., Ca(II) (Boulton 1998), and as a trap of heavy metal ions, e.g., Cu(II) and Fe(III) (Sarna 1992; Zareba

1995) It has also been suggested that the integrity of the molecular structure of the

pigment, and thereby the eumelanosomes, could be impaired by high metal concentrations (Zareba 1995) Such a change could result in the release of heavy metal ions, for example Fe(IID, into the cytosol, which could induce cellular damage (Zareba 1995) The metal binding behavior, including binding capacity, affinity, and

sites of metals, in eumelanin and the effects of metal content on the aerobic reactivity

of eumelanin are important parameters for understanding the interaction and consequences of metal-melanin complexation

A variety of techniques, such as flame photometry, UV-Vis spectrometry, radiometry, and inductively coupled plasma mass spectrometry, have been employed

to study the metal binding affinity/capacity of melanin (Potts 1976; Liu 2004) In our

case, we have found that the affinity and capacity of metals for eumelanin can be determined using inductively-coupled plasma mass spectrometry by quantifying the amount of complexed and dissolved metal ions in solutions containing a fixed amount

of suspended divalent metal-free melanin and specified initial metal concentrations

12

(Liu 2004)

However, due to the opaqueness, insolubility, heterogeneity, and

amorphousness of eumelanins, it has proven difficult to elucidate molecular details of the metal binding sites The standard approaches of X-ray diffraction and nuclear magnetic resonance spectroscopy are not informative in this case

According to the proposed molecular structures for eumelanin, the pigment contains phenolic hydroxyl (OH), carboxyl (COOH) and amine groups (NH) as potential binding functional groups for metal ions (Riley 1997; Ito 1998) The binding

of metals will involve competing with H* for these functional groups As the pigment

exhibits infrared vibrational absorption bands associated with these functional groups,

IR absorption spectroscopy could be informative about the immediate binding

environment of metal ions in natural melanins

To this end, IR absorption spectroscopy has been employed to study the binding of Cu(II) and Fe(III) to several types of eumelanins (Bridelli 1999; Bilinska 2001; Stainsack 2003) Bilinska et al studied the interactions of Cu(II) with human hair eumelanosomes Stainsack et al examined the coordination of Cu(II) to synthetic eumelanin Their data suggested that carboxyl groups form the binding site (Bilinska 2001; Stainsack 2003) Bridelli et al studied the binding of Fe(III) by human

neuronmelanin Their work suggests that the hydroxyl group coordinates to the metal

ion (Bridelli 1999) This result is also consistent with the conclusions of our recent

13

study of the binding of Fe(III) to Sepia eumelanin using resonance Raman spectroscopy (Samokhvalov 2005)

Less attention, however, has been paid to the non-redox-active metal ions Mg(II) and Ca(II), yet both are abundant in human, bovine, and Sepia melanin

granules In addition, Zn(II) is found in all melanosomes and accumulates in RPE

melanosomes; the details of its binding are also largely unknown In this work, we studied the binding sites of Mg(ID, Ca(II), Zn(II) and Cu(II) to Sepia eumelanin by analyzing pH variations of metal solutions upon metal binding and IR spectroscopic

analysis

The aerobic reactivity of melanin is associated with its metal contents Melanin

has been proposed to act as a free radical scavenger (Boulton 1991; Sarna 1992;

Dunford 1995; Rozanowska 1999; Schraermeyer 1999) On the other hand, in addition

to its scavenging abilities, melanin is photosensitizing and produces reactive oxygen species (ROS’s), such as superoxide anion ( O%) and hydrogen peroxide (H2O2), under

UV irradiation (Hill 1992; Rozanowska 1997; Rozanowska 2002) The efficiency of

ROS’s generation is related to metal content in melanin; melanin saturated with Fe(III)

shows significant generation of ROS’s (Zareba 1995) Herein the effect of the metal binding on melanin’s aerobic reactivity was analyzed by examining the ability of melanins with different metal contents to nick supercoiled DNA

14

MATERIALS AND METHODS

Materials The ink sacs of Sepia officinalis were freshly dissected from wild cuttlefish (Richard K Stride, Dorset, UK) and shipped overnight Immediately upon arrival, the

melanin from the ink sacs was purified according to procedures previously published (Liu 2003) The metal chloride salts were > 99.99% purity (Sigma, St Louis, MO) pUC18 DNA, 2526 base pairs in length, was purchased from MBI Fermentas (Hanover, MD) Nanopure water (18.2 MQ) was used to prepare all samples KBr of

IR grade was purchased from Sigma

Sample Preparation Melanin with varying metal content was prepared as follows Ethylenediaminetetraacetic acid (EDTA)—washed melanin (with Na’ as the only metal

cation) was prepared by repeatedly suspending natural Sepia melanin into 40 mM

EDTA solution (pH=5.5, 16 hrs) for four times followed by washing with water The

final products were lyophilized Mg(II)-, Ca(II)-, Zn(II- and Cu(ID)-enriched melanins

were prepared by suspending EDTA-washed Sepia melanin (prepared at pH 5.5) in a metal salt solution (varying concentration up to 10 mM, pH ~4, with the ionic strength adjusted to 100 mM by NaCl) at room temperature for 16 hrs The metal-enriched

particles were purified by centrifugation, and then were lyophilized

pH Measurement The pH values of each salt solution were recorded before the addition of EDTA-washed melanin (1 mg/mL) and after 16 hours of the suspension of EDTA-washed melanin in the solution The pH measurement experiments were

15

performed twice, with a deviation less than 0.1

Infrared Spectroscopic Analysis The metal enriched melanin samples were kept in a vacuum desiccator in the presence of phosphorus pentoxide (P2Os) for more than 24

hours before examination by IR spectroscopy IR samples were prepared by mixing dried melanin granules with KBr in a mass ratio of ~ 1:1000, and the mixture were then pressed with an IR die resulting in transparent pellets IR spectra were collected

using a scanning FT-IR spectrometer (Bruker, Model IFS 66V/S) averaging 32 scans

at 4 cm” resolution

DNA Nicking Analysis The induction of DNA single-strand breakage by melanin

with different metal contents was examined as follows Supercoiled pUC18 DNA, was used either as provided or purified by ethanol precipitation The nicking reactions of different melanins with 14.3 nM of either the originally provided DNA or the purified

DNA were performed in phosphate buffer (5 mM, pH 7.4) or Tris-HCl buffer (10 mM,

pH 7.4) The concentration of melanin in solution was adjusted to obtain an

absorbance of 1.0 at 380 nm with a 1.0 cm path length, which yielded a concentration

of ~80 pg/mL The samples were irradiated for 10 min by 380-550 nm light with an average energy of 13 mW/cm’ from a filtered Hg—Xe lamp (1 kW, Spectral Energy, Chester, NY)

The nicking of supercoiled pUC18 DNA results in the conversion of DNA from the supercoiled form (Form I) to the relaxed form (linear or random coil form,

16

together referred as Form II) These two forms were separated by gel electrophoresis (1% agarose gel with a BioRad electrophoresis system, Hercules, CA) The gel was

stained by ethidium bromide (EB) and then imaged using a digital camera under

illumination of 302 nm light A correction coefficient of 1.4 was used to correct the

lower EB stainability of supercoiled DNA relative to relaxed DNA according to

Hermes-Lima et al (Hermes-Lima 1998)

RESULTS AND DISCUSSION

pH Measurements The natural pH of melanosomes is between 3 and 4 (Bhatnagar 1993) and we chose to perform experiments starting with an initial pH of 4 The pK,

of the three functional groups contained in the pigment are ~ 4.5 for COOH (the calculated value of COOH group of DHICA, SciFinder, 4790-08-3), > 10 for NH (considering the sp’ hybridization of the N atom), and ~9 and ~13 for the ortho- phenolic OH (pX, values for catechol (Szpoganicz 2002)) The EDTA-washed melanin was prepared at a pH of 5.5 and so the NH and OH are fully protonated and

the carboxylic acid sites are mostly deprotonated and coordinated to Na’ cations

When placed in the metal solutions at pH 4.0, the NH and OH remain fully protonated, but in the absence of coordination of the metal cations present in solution, a proportion

of the carboxylic acid sites become protonated because the pH < pK, (4.0 < 4.5)

17

A change in solution pH associated with the binding of a metal ion could

therefore provide insight into the function groups involved The final pH values following addition of melanin to the salt solutions are given in Table 2-1 The final

pH values decreased with increasing metal ion concentration And for the same metal

concentration, the pH values decreased in the order Mg(ID > Ca(ID > Zn(II) > Cu(II)

For the lowest metal concentrations, the final pH values were larger than those of the initial solutions for all metals For the highest concentrations of Zn(II) and Cu(II)

examined, the final pH values were lower than in the initial solution

The increase in pH observed upon addition of metal reflects the protonation of

COO’ groups in the pigment If the metal cation binds to the carboxylate group, then with increasing salt concentration the pH of the solution will approach that of the level

of the original solution, ~4 On the other hand, if the metal cation binds to either the hydroxyl or amino groups, coordination will result in release of protons (e.g., OH > OQ) and in this case, a significant proton concentration increase can be observed with

high enough metal concentration The data (Table 2-1) reveal that Cu(II) binding

induced a significant increase in H* concentration (~0.7 mM in response to a 10 mM

salt solution) compared to the other three metal ions (< 0.1 mM in response to 10 mM

salt solutions) EDTA-washed melanin can take up 1.3-1.5 mmol/g Mg(II), Ca(H) and Zn(II), and 1.1 mmol/g Cu(II) when the initial metal ion concentrations are 10 mM

18

19

(pH 4) Thus the difference in proton concentration change by the metal ions studied cannot be due to different extents of binding Cu(II) binding must involve

deprotonation of the functional group(s) OH and/or NH The pH data argue that

Ca(II), Mg(II), and Zn(II) are bound mainly to carboxylate groups We next further

validated these characteristics of the binding site with concentration-dependent IR

absorption data

IR Spectroscopic Analysis Binding of metal ions is expected to affect the transition

frequencies of the coordinated functional groups If metal binding involves the COO” group, then a concentration-dependent change in the infrared region characteristic of the C=O stretching frequency of the acid moiety is expected Likewise, if the metal

cation binds to OH or NH groups, changes in the infrared spectrum (loss of N-H and O-H intensity) reflecting deprotonation to accommodate binding of metal cations is anticipated

Published analyses of the infrared spectra of natural and synthetic melanins indicate the following peak assignments (Bridelli 1999; Bilinska 2001; Stainsack

2003): 3200-3500 em” for NH (~ 3200 cm") and OH stretching (~ 3400 cm”) in

indole or pyrrole systems; 1710 cm’! C=O stretching in COOH; 1610-1690 cm"! aromatic C=C and C=N bending and C=O stretching (non carboxylic acid); and 1250 cm”, phenol OH and the carboxyl OH

20

Table 2-2 summarizes the effect of added metals on the intensity of observed infrared transitions With increasing concentrations of bound Mg(II) and Ca(ID, the

intensity of the protonated COOH band decreases, while the NH and phenolic OH

(3400 cm’) band intensities are unaffected or slightly increased For the case of Zn(II) and Cu(II) binding, the change in intensities of the NH, OH and COOH bands depends

on the metal concentrations The reasons behind this are elaborated below

Ca(II) and Mg(II) binding Ca(II) and Mg(II) binding were manifested by a similar

concentration-dependent decrease in the 1710 cm peak (Figure 2-1 shows the

spectra for Ca(II)-loaded melanin) This spectral change reflects the concentration of

unionized COOH in the melanin, and therefore its decrease with increasing salt

concentration supports the conclusion that Ca(II]) and Mg(ID) bind to the ionized acid

group This conclusion is also supported by the pH data presented above Absorption

at the 3400 cm! and 3250 cm’ ‘peaks (OH and NH, respectively) was unaffected or slightly increased in intensity by metal binding, indicating that the OH and NH groups are not directly involved in the binding of Mg(ID) and Ca(II) The decreased intensity

of the 1250 cm’! peak (Figure 2-1) - overlapping absorption of the phenolic and carboxyl OH - also supports this conclusion The insensitivity of the 3400 cm™! peak

to salt concentration establishes that the phenolic groups are not affected by binding and so the decreased intensity in the 1250 cm” peak reflects the deprotonation of the

OH group of the carboxylic acid upon metal binding

21

Peak Mg(I) Ca(I)

*: with no obvious change except for the sample treated 10 mM metal solution,

which showed a significant decrease

> increase when [Cu(ID] in solution < 10 mM, but decrease for [Cuq])] = 10 mÀ⁄

22

10 mM solutions This data clearly shows the decrease of the 1710 cm" and 1250 cm"! peak intensities with increasing Ca(II) concentration

23

Zn(II) binding For [Zn(II] < 2 mM, the binding of Zn(II) showed a decrease in the

intensities of the transitions associated with COOH and NH, while the 3400 cm" peak for OH remained constant (or slightly increased) relative to that of either EDTA-

washed melanin or Mg(II), Ca(II)-loaded melanin (Figure 2-2.) This observation

suggests that Zn(II) binds to COOH and NH under these conditions However,

comparison of the changes in the COOH band for Zn(II) and Ca(ID/Mg(II) suggests

that this functional group accounts for most of the binding, with only a minor portion

arising from NH coordination

For larger concentrations of Zn(II), ¢.g., 10 mM Zn(ID), the intensity of the OH band at 3400 cm” decreased But considering that binding results in a release of 0.1

mM H" compared to a total binding of 1.5 mmol Zn(II) per gram of melanin, the involvement of OH as a coordinating ligand is a minor contribution, and likely reflects

a second, lower affinity binding site that becomes populated as the availability of carboxylate groups diminishes

Cu(II) binding In contrast to the observations for the three metals discussed above, the intensity of the COOH absorption band increased upon binding Cu(II) for solution concentrations < 2 mM (Figure 2-3a ) At the same time, the intensity of the OH band decreased (Figure 2-3b); the NH band was unaffected by Cu(II) binding These observations implicate the phenolic groups as binding sites Given the current

24

cm The spectra are normalized to the intensity of the 1620 cm” peak Since we

currently are not sure of the reason for the irregular IR spectrum of the 0.25 mM Zn(II)-enriched melanin, we compared the spectra of 0.5 mM with 2 mM (bottom grey curve) and with 10 m& (bottom black curve) This data clearly show the decrease of

the 1710 cm™ and 1250 cm” peak intensities with increasing Zn(II) concentration To facilitate the comparison with EDTA-washed melanin, the IR spectrum of EDTA- washed melanin is included (the thick grey line in b)

25

1710 and 3200 cm” peak intensities are both smaller

26

structural models for eumelanin, this would argue that catechols form the binding sites

of Cu(II)

For higher Cu(II) concentration the situation is more complex Comparison of the melanin IR spectra for exposure to 2 mM and 10 mM Cu(II) shows that the increase in salt concentration (and hence increase in the amount of bound metal) resulted in a decrease in the intensity of the COOH and NH bands This suggests that

at these higher concentrations, Cu(II) binds not only to the catechol groups, but to the

carboxylate and amino groups One possible interpretation of this data is that the

primary coordination site (catechols) are saturated for this amount of melanin in a ~2

mM solution Increased Cu(II) exposure then reveals a second site of lower affinity,

involving the carboxylic and amino groups

This observation, however, is not consistent with previous models for Cu(II)

binding, Froncisz et al and Sarna et al reported that the EPR spectrum of Cu(ID) in

melanin varies with pH (Froncisz 1980; Sarna 1980) Based on their data, they suggested that for pH < 5, Cu(II) binds to carboxylate groups; and for pH > 10, Cu(II)

binds to phenolic groups Our samples were examined for pH < 5 and the IR data

clearly showed effects of Cu(II on the phenolic absorption bands Their conclusions are based on the pKa values of the different functional groups (in the absence of metal) However, binding of Cu(II to these functional groups will change their

apparent pKa values It cannot be ruled out that in eumelanin, an o-OH (catechol)

27