Siderophore-mediated iron transport apparatus in Paracoccus denitrificans


Material Information

Siderophore-mediated iron transport apparatus in Paracoccus denitrificans
Physical Description:
Dionis, John B., 1961-
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 17945212
System ID:

Table of Contents
    Title Page
        Page i
        Page ii
        Page iii
    Table of Contents
        Page iv
        Page v
        Page vi
    Chapter 1. Introduction and background
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
    Chapter 2. Paracoccus denitrificans iron transport apparatus
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
    Chapter 3. Synthesis and characterization of a parabactin photoaffinity label
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
    Chapter 4. Extraction, assay and properties of a ferric parabactin outer membrane receptor in Paracoccus denitrificans
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
    Chapter 5. Conclusion
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
    Biographical sketch
        Page 148
        Page 149
        Page 150
        Page 151
Full Text

IN Paracoccus denitrificans






For their continual advice and support, this dissertation is
proudly dedicated to my mother and father. Their guidance over the

years has proven invaluable in reaching this point in my career.

I would like to express my sincere gratitude to my research
advisor, Dr. Raymond Bergeron, for his guidance during the past four
years. His genuine concern and advice during trying times have been

deeply appreciated.
I would like to express appreciation to the members of my super-
visory committee, Dr. Kenneth Sloan, Dr. Steven Schulman and Dr.
Charles Allen. I wish to extend a special word of thanks to Dr.
Richard Streiff, as a committee member, and for his helpful discussions

and advice.
I would like to thank all my friends in the lab, with a special
acknowledgment to Michael Ingeno and Dr. William Weimar for having made
significant contributions to this work. I would also like to extend a
special thanks to Irma Smith for her technical assistance in the prep-

aration of this dissertation.
I wish to thank my parents, brothers and sister for their love
and encouragement. I would also like to express special appreciation

to my grandparents for their love and inspiration.
Most importantly, I wish to express my deepest appreciation to my
loving wife, Jody, for always caring and displaying a genuine belief in
my abilities. Her ability to raise a family and complete her education
has been both an inspiration and driving force throughout my graduate
career. For making this experience the most memorable in my life, I
wish to express my love to my daughter Ariana. Just the sight of her
smiling face was enough to make even the most difficult days more




ACKNOWLEDGMENTS ....... ....................... .... iii

ABSTRACT ........ ............................... v


I. INTRODUCTION AND BACKGROUND ...... ............ 1

Microbial Iron Assimilation ..... .......... 1
Structural Organization of the Bacterial
Cell Envelope ........ ................ 12

II. Paracoccus denitrificans IRON TRANSPORT APPARATUS. 17

Introduction ..... ................. .... 17
Materials and Methods ...... ............. 18
Results ......... .................... 20
Discussion ...... .................. .... 28

PHOTOAFFINITY LABEL ....... ................ 41

Introduction ..... ................. .... 41
Experimental ........ ................. 43
Results ......... .................... 51
Discussion ...... .................. .... 83

Paracoccus denitrificans ... ............. .... 89

Introduction ..... ................. .... 89
Materials and Methods ................ .... 90
Results ...... .................... .... 98
Discussion ..... .................. .... 122

V. CONCLUSION ......... .................... 126

REFERENCES ......... .......................... ... 129

APPENDIX ........ ........................... .... 136

BIOGRAPHICAL SKETCH ...... ..................... ....148

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
IN Paracoccus denitrificans



August 1987

Chairman: Dr. Raymond J. Bergeron
Major Department: Medicinal Chemistry

The high affinity iron assimilation system in Paracoccus deni-

trificans involving its siderophore parabactin and the cognate membrane

receptor is examined. Parabactin was shown to deliver iron to the

microorganism via an "iron-taxi" mechanism in which the siderophore

iron complex binds to a membrane receptor where the iron is released

and transported into the cell. The deferrated ligand remains extracel-

lular and can be re-utilized in iron transport. The data also indicate

that Paracoccus denitrificans exhibits stereospecificity in its sidero-

phore requirement. The ferric L-parabactin chelate was most effective

in supplying iron to the microorganism. The ferric D-parabactin che-

late was unable to effectively supply iron to the microorganism in

various experimental protocols including growth and transport studies.

The synthesis of parabactin azide, the first catecholamide sider-

ophore photoaffinity label, is described. The photoaffinity label is

shown to have the same biological activity as parabactin in stimulating

the growth of Paracoccus denitrificans under low iron conditions. The

photoaffinity label was employed in the identification of the

siderophore membrane receptor. Additionally, the synthesis of amino

parabactin, which is subsequently attached to an activated sepharose

resin to produce the first polyamine siderophore affinity column, is


The role of isolated Paracoccus denitrificans membrane proteins

in siderophore-mediated iron transport is examined. High affinity,

stereospecific binding activity for ferric L-parabactin is shown to be

associated with the cell membrane and not with the cytoplasmic pro-

teins. An outer membrane preparation from cells grown in low iron med-

ium was found to retain ferric parabactin binding activity following

solubilization in a detergent containing buffer. Binding activity was

measured by means of a column-centrifugation technique which separated

free and receptor bound ferric parabactin. The presence of low-iron

inducible high molecular weight proteins as a major component in the

solubilized outer membrane preparation containing the highest binding

activity indicates that one or more of these proteins is involved in

the siderophore-mediated iron uptake system as a receptor for ferric



Microbial Iron Assimilation
Despite the fact that iron is one of the most abundant metals on

earth, nearly all life forms have had to develop sophisticated mechan-

isms for its access and assimilation. This is due, in part, to the

extreme insolubility of ferric ion at physiological pH (Ksp of

Fe(OH)3 = 10-38 M1)1 and the tendency of ferric hydroxide to polymer-

ize and precipitate as an oxyhydroxide polymer. It is also evident

that the uptake of this metal needs to be regulated at the membrane

level as a consequence of the role that iron plays in generating

hydroxy free radicals in the presence of hydrogen peroxide.2

Iron plays a crucial role in many biological redox systems.3-5
This is due to the generation of a large redox potential between the

interchange of the ferrous and ferric forms, an oxidation state which

is very sensitive to both pH and chelation. Iron plays a critical role

in the energy requirements of virtually all microorganisms. An excess

of the metal causes a fulminant growth of many microorganisms while

iron deprivation can substantially slow or even halt growth.6-8 The

respiratory chains of aerobic and facultative anaerobic bacteria have

in common the presence of cytochromes and non-heme enzymes, each with a

specific requirement for iron.9 In addition, there are other elec-

tron transfer proteins, such as the hydrogenases, iron-sulfur proteins

and dehydrogenases which have an absolute requirement for iron. As a


means to protect the cell, there are also microbial iron enzymes re-

sponsible for the metabolism of hydrogen peroxide and oxygen. These

include the catalases, peroxidases, oxygenases as well as certain forms

of superoxide dismutase. Other crucial roles which iron plays in bac-

terial metabolism include the nitrogenase enzyme involved in the fixa-

tion of dinitrogen, the aconitase enzyme of the citric-acid cycle and

ribonucleotide reductase which is required for DNA synthesis.9

The question arises as to why iron was incorporated into so many

important enzyme systems in virtually every type of living cell. Iron

is the fourth most abundant element on earth, surpassed by only oxygen,

silicon and aluminum. It is speculated that iron was firmly estab-

lished as a bioessential element during the anaerobic phase of life on

earth.10 During this stage of evolution, in which iron was in the

readily soluble ferrous form, there would be no need for specific iron-

binding compounds. However, when photosynthetic organisms began pro-

ducing oxygen, the soluble iron was transformed to its insoluble ferric

form. This imposed severe restrictions on iron availability and as a

result microorganisms developed the means for retrieving iron from

insoluble polynuclear complexes and for the delivery of iron to the

cell, where it can be processed for use in cellular metabolism. Micro-

organisms have evolved a group of low molecular weight, virtually fer-

ric ion specific ligands or siderophores to sequester exogenous ferric

ion and facilitate the transport of this metal into the cell.912

In general, siderophores are classed as either hydroxamates,13

such as desferrioxamine14 or as catecholamides,15 as exemplified by

enterobactin16 and the linear catecholamide parabactin17 (Fig. 1-1).

Both desferrioxamine and parabactin form extremely tight complexes




1 o
H _- /

0 0


H O O N \\ H NO H\

HO H --O

(o1 OH
FI I I t I I

(c) 0 S
HO 0
0C 0 0

HO OH 3 / OH


Figure 1-1. Structures of the siderophores. (a) Enterobactin,
(b) desferrioxamine and (c) parabactin.


with ferric ion, with formation constants, Kf, of 1031 M- and 1048

M-1, respectively.18'19 Enterobactin, which is a cyclic trimer of

2,3-dihydroxybenzoyl serine, is produced by most enteric species

including Escherichia coli and Salmonela typhimurium.16 The iron

(III)-enterobactin formation constant has been calculated to be 10

making this siderophore the strongest ferric ion chelating agent

known.20 Despite considerable structural variation between the classes

of siderophores, a common feature of these ligands is their ability to

bind ferric ion in a hexacoordinate, octahedral complex with formation

constants in the range of 1030_1050 M-1. Moreover, both the hydroxa-

mate and catecholamide siderophores utilize bidentate oxygen containing

moieties to form tight complexes with iron (III). The iron (III) being

a powerful lewis acid due to its high charge to size ratio forms very

stable bonds with the weakly polarizable oxygen atoms.21

It is generally found that fungi produce siderophores of the

hydroxamate variety, while bacteria frequently produce catecholamide

siderophores in addition to hydroxamates. For example certain strains

of E. coli produce the catecholamade siderophore enterobactin as well

as the hydroxamate siderophore aerobactin.22 There are also certain

siderophores namely mycobactins in which hydroxamate and phenolate

ligands are present in the same molecule.23 The gram-negative soil

bacterium Paracoccus denitrificans produces the catecholamide sidero-

phore parabactin. This was the first siderophore isolated which was

predicated on a polyamine backbone, in this case spermidine.24 To

each of the triamine nitrogens is affixed a bidentate ligand. The

terminal primary amino groups of the spermidine backbone have 2,3-dihy-

droxybenzoyl moieties attached while the secondary nitrogen has a 2-(2-

hydroxy phenyl)-4-carboxyl-5-methyl-2-oxazoline functionality fixed.

The original structure assigned by Tait was inaccurate in that the sug-

gested secondary N-functionality was given as N-(2-hydroxybenzoyl)-L-

threonine. It was only later that the functionality was demonstrated

to be an oxazoline ring.25 The soil bacterium Agrobacterium tumefac-

iens26 produces the catecholamide siderophore, agrobactin (Fig. 1-2).
This ligand is nearly identical to parabactin but it contains an addi-

tional hydroxyl on the central aromatic ring. Another hexacoordinate

polyamine catecholamide siderophore has been isolated from the cultures

of Vibrio cholerae.27 This ligand, known as vibriobactin (Fig. 1-2),

is predicated on the less frequently found symmetrical norspermidine

backbone and has in common with the previous two siderophores a phenyl

oxazoline ring system in which the oxazoline nitrogen is utilized in

iron chelation. It should be noted that the synthesis of the above

mentioned polyamine catecholamide siderophores has been accomplished in

our laboratories.28-30

Another interesting feature of the hydroxamate and catecholamide

ligands is that the siderophore iron complex can exist in two different

optical configurations arising from chelation of the metal in either a

left-handed A, or right-handed A, "coordination propeller" (Figs. 1-3

and 1-4).9 This can be visualized as one bidentate ligand being set

in a potentially preferred spatial orientation. The remaining two

bidentate ligands could then cluster around the iron in either the A or


(b) O N OH


H 0 N OH

CH 3 0 C 3 / L O

~ N

Figure 1-2. Structures of the siderophores. (a) Agrobactin and
(b) vibriobactin.

Figure 1-3. Representation of the A coordination isomer.

Figure 1-4. Representation of the A coordination isomer.

It is very interesting that the specific iron-siderophore trans-

port systems present in microbial species recognize only the sidero-

phore with the appropriate chirality about the iron center. The impli-

cations are that by producing siderophores from optically active com-

ponents, microorganisms have found a way to monopolize iron in a com-

peting environment. In other words, they can excrete siderophores

which are only recognizable to their own transport systems. The major

criteria for survival in this competing environment would then be the

ability to synthesize a chelator with a higher affinity for iron. It

would also be expected that those species which have evolved more than

one mechanism of iron-transport would be better suited for survival.

It should be pointed out that certain bacteria have evolved specific

transport systems for siderophores produced by other microorganisms.

For example, the enteric bacteria E. coli possesses a transport appara-

tus for the siderophore ferrichrome.31 This hydroxamate siderophore,

which E. coli does not synthesize, is produced by many species of

fungi and was first isolated by Neilands from the mycelium of the

smut fungus Ustilago sphaerogena.32

The concentration of iron required for maximal growth of most

gram-negative bacteria in culture ranges from 0.02 to 0.1 pg/mL.33

These iron requirements are influenced by a number of different

factors. Strongly aerobic organisms such as Pseudomonas, Azotobacter

and Mycobacterium have much higher iron requirements, approaching 5

ug/mL.33 A species of blue green algae, Nostoc muscorum, required

10 pg Fe/mL for maximal fixation of dinitrogen.34 It has been shown

that higher iron concentrations are required by microorganisms growing

in an environment containing metals such as A13+, Cr3+ or Cu2+.35,36

These metal ions can directly interfere with bacterial assimilation of

iron as they bind effectively to siderophores. The bacterial require-

ment for iron is also very sensitive to temperature. Studies indicate

that with increasing temperature there is a drastic reduction in sider-

ophore biosynthesis.37,38 This results in an increased iron require-

ment, as the microorganism can no longer transport this vital metal

using its high affinity system. In Salmonella typhimurium, a pathogen-

ic bacterium which also produces the catecholamide siderophore entero-

bactin, the concentration of iron required for maximal growth was

determined to be 0.1 pg/mL at various temperatures, ranging from 310 to

36.90C.37 However, when the incubation temperature was raised to

40.30C there occurred a concomitant loss of enterobactin synthesis and

an increase in the iron requirement for maximal growth by nearly 30

fold. As this critical temperature is within the range of most physio-

logical systems, it has been speculated that fever may play an impor-

tant role in the control of pathogenic microorganisms. By raising the

temperature to the critical point, the biosynthesis of iron transport

compounds could be turned off, resulting in the loss of the microorgan-

isms' ability to compete for iron present in the host.

The virulence of many pathogens is dependent upon interactions

between host and pathogen. Indeed it is becoming more and more evident

that the virulence of many bacteria can be directly associated with the

production of siderophores thus implicating iron assimilation as a

critical component in the host-pathogen interchange.39'40 The avail-

ability of iron in the tissues of the mammalian host is severely

restricted.41 The majority of the total body iron is found intracel-

lularly as ferritin, hemosiderin and hemoglobin. The small quantity

which is extracellular and potentially available to invading pathogens

is tightly bound by high affinity iron-binding glycoproteins. Trans-

ferrin, the iron binding protein of serum, and lactoferrin, a similar

protein found in secretions, are normally unsaturated and limit the

availability of iron necessary for microbial growth to 10-12 M or

less.42 The ability of microbial pathogens to aquire iron in a

host is one factor determining the virulence of a microorganism and is

considered a prerequisite to infection.
In many cases the more virulent strains in a given species are

those which contain multiple pathways for the assimilation of iron.

The virulence factor in invasive strains of E. coli is not the expected

catecholamide siderophore enterobactin but the hydroxamate siderophore

aerobactin.43 At first glance this appears to be a contradiction

as the metal binding affinity of aerobactin is far inferior to entero-

bactin at physiological pH. An explanation implicating the role of

serum albumin on siderophore-mediated utilization of transferrin iron

has been advanced by Konopka and Neilands.44 Their initial results

demonstrated that the abilities of the siderophores enterobactin and

aerobactin to remove iron from transferrin in vitro were very dependent

on the composition of the medium. While the ability for enterobactin

to remove transferrin iron was greater in buffer solutions, the trans-

fer rate in serum was far superior for aerobactin.45 Further studies

demonstrated the effects of serum proteins on siderophore mediated

utilization of transferrin iron. Serum albumin, the most abundant

protein in human plasma accounting for 60% of the total, is known to

bind to many hydrophobic ligands.46 Serum albumin and human serum were

shown to be comparable in their selective ability to interfere with

the transfer of iron from ferric enterobactin to E. coli. Furthermore,

serum albumin was shown to bind enterobactin in a 1:1 stoichiometry

with a binding constant greater than 10 M.45 In contrast, aerobac-

tin had no detectable affinity for serum albumin and was able to trans-

fer iron to the cells in the presence of serum albumin. The explana-

tion for this behavior is probably best attributed to the different

chemical structures of the two siderophores. The catecholamide

siderophore enterobactin is a lipophilic molecule possessing 2,3-dihy-

droxybenzoic acid moieties. This aromatic character results in a sus-

ceptibility of the siderophore to bind to serum proteins. This tenden-

cy could also result in the formation of antibodies to the siderophore.

Evidence for the presence of enterobactin specific antibodies has been

reported.47 An immunoglobulin A antibody isolated from human serum

was shown to inhibit uptake of iron from enterobactin but not from fer-

richrome or citrate.48 These results suggest that the synthesis of

aerobactin is an important factor in the virulence of invasive strains

of E. coli. Aerobactin has the ability to remove transferrin bound

iron in the presence of serum proteins and to deliver the metal to the

invading pathogen.

Structural Organization of the Bacterial Cell Envelope

An examination of bacterial iron transport cannot be complete

without a general background into the structural organization of the

gram-negative bacterial cell envelope. This complex envelope consists

of various layers with distinct chemical compositions which impart

markedly different physical characteristics to each.49 These layers

are associated with one another through various points of attachment

and hence the function of each layer is modified by other cell wall

components. An interesting feature of the cell envelope is its contin-
ual synthesis in response to cell growth. In fact, the outer membrane

is a classic example of membrane biogenesis in which all major constit-

uents are synthesized elsewhere in the cell.50 Assembly of the

membrane requires translocation and insertion of each component into

the final membrane structure.

It is at the inner or cytoplasmic membrane that the structural

components of the cell wall are synthesized.51 These various com-

ponents include the basic chemical units of peptidoglycan, lipopolysac-

charide and phospholipid. The cytoplasmic membrane is a phospholipid

bilayer containing a large number of peripheral and integral pro-

teins.52 The cytoplasmic membrane also functions in the active
transport of substrates, including iron. Inner membrane vesicles from

cells of E. coli have been shown to transport [3H]-ferrichrome with a

K. of 0.2 UM.53 The transport mechanism indicated an energy dependence

as detected through the use of metabolic inhibitors as well as a need

for divalent cations.54

The peptidoglycan is an essential feature in all bacteria albeit

in highly variable amounts. This structure surrounds the bacterial

cytoplasmic membrane and prevents cell lysis in hypotonic environments

as well as contributes to cellular form and rigidity.55 It has also
been shown that the peptidoglycan is closely associated with the outer

membrane. A specific lipoprotein, covalently linked to the peptidogly-

can, extends outward towards the outer membrane and serves to anchor

the membrane by hydrophobic interactions with the phospholipids present

in the outer membrane.56 Due to a lack of extensive cross-linking,

the peptidoglycan of gram-negative bacteria does not constitute an

effective barrier to the passage of small molecules.

The area between the cytoplasmic membrane and the outer membrane

of gram-negative bacteria is known as the periplasmic space. This area

serves a similar function as the eukaryotic lysosome, which houses the

hydrolytic enzymes such as the proteases, lipases, phosphatases and

nucleases.57 Bacteria, however, do not contain lysosomes and in

order to prevent self-degradation have maintained these enzymes sepa-

rate from other cell components in the periplasmic space. This space,

which can comprise nearly 30% of the total cell mass, is also involved

in transport. Specific binding proteins associated with the outside of

the cytoplasmic membrane, transport amino acids and sugars through the

cytoplasmic membrane into the cell. Thus, periplasmic enzymes can act

on a wide variety of substrates which diffuse into this zone and con-

vert them to molecules which are transportable into the cell utilizing

the specific binding proteins and permeases.49 While there is no

direct evidence implicating periplasmic binding proteins in siderophore

mediated iron transport, their existence cannot be ruled out.

The layer which controls the passage of molecules into and out of

the periplasmic space is the "molecular sieve" layer better known as

the outer membrane. This layer, which is an asymmetric bilayer, is

composed mainly of lipopolysaccharide, phospholipid and protein.58

The lipopolysaccharide, comprising as much as 45% of the outer mem-

brane, is responsible for antigenicity, binding of specific enzymes and

contributes to the hydrophobicity of the outer membrane.59 It is the

lipopolysaccharide which to a major extent protects enteric bacteria

from lysis due to exposure to bile salts in the intestinal flora of


The outer membrane is made up of several major and minor pro-

teins. However, a minor protein can, under certain growth conditions,

be made in quantities as great as that of major proteins. Another

interesting property of the outer membrane is the nonspecific perme-

ability toward small, hydrophilic substances. The outer membrane can

thus provide specific and nonspecific channels for nutrients and ions

required for growth. Matrix proteins known as "porin" function to form

such passive diffusion pores throughout the outer membrane.60 The

porins render the outer membrane freely permeable towards hydrophilic

molecules (<650 Mw) such as sugars, amino acids and peptides.21

These major proteins are characterized by a tight, noncovalent associa-

tion with the peptidoglycan, presumably through electrostatic forces.

They are also unique in containing a high content of O-structure in

contrast to other intrinsic proteins. It is of interest that these

essential proteins located on the cell surface have become the target

for various phages.

It has been shown that in many bacterial species there are outer

membrane proteins which function as receptors in the uptake of ferric

ion siderophore complexes.52 These receptor proteins are produced

concomitantly with iron deprivation and bacterial siderophore produc-

tion. Characterization of these proteins by sodium dodecyl sulfate

polyacrylamide gel electrophoresis (SDS-PAGE) indicates a molecular

weight in the range of ca. 80,000 daltons for the ferrichrome and

enterobactin receptor proteins of E. coli.52 It is postulated that


these proteins may contact the cytoplasmic membrane at various adhesion

zones. Indeed, the ferrichrome receptor has been shown to be located

at such an area.52 The primary function of these receptor pro-

teins is the initial binding of the siderophore-metal complex. After

this step, there are a number of different scenarios which one could

envision. One such scenario would have the ferric siderophore complex

translocated into the cell, perhaps to the site of the cytoplasmic mem-

brane, facilitated via an adhesion zone. At this point, a specific

binding protein would transport the siderophore-chelate into the cyto-

plasm where the metal ion would undergo a reductive separation from the

ligand. Obviously, there can be a number of variations to this scen-

ario. The most striking being that release of iron from the sider-

ophore occurs at the membrane level. In this way, siderophore uptake

is not required for transport of the metal.

From the aforementioned discussion it is clear that iron

acquisition plays a critical role in the survival of microbial species.

The mechanism by which microorganisms compete for this vital metal has

been the subject of intensive research. It is hoped that the research

presented herein will contribute to the better understanding of microb-

ial iron transport.

Paracoccus denitrificans IRON TRANSPORT APPARATUS

Due to the physiological importance of iron coupled to its

extreme insolubility at physiological pH, living organisms have evolved

diverse pathways for iron assimilation. When confronted by iron-

deficient conditions, microorganisms synthesize low molecular weight

ferric ion chelating agents called siderophores. The siderophore and

the matching membrane-associated receptor which recognizes the ferric

chelate comprise the "high affinity" iron transport system in micro-

organisms.9 It is this high affinity iron transport pathway which has
received considerable attention. Less is known concerning the rela-

tively inefficient "low affinity" transport system of microbial iron

uptake. In this process it is believed that microorganisms have the

ability to utilize ferric ion without the necessity for synthesizing

specific siderophores and their corresponding membrane receptors for

the solubilization and transport of this metal. The concentration of

utilizable iron in the cell environment must be relatively high on the

order of 10 pM for maximal growth.9

In the present study we examine the high affinity system by

which Paracoccus denitrificans utilizes its siderophore parabactin in

its iron-transport apparatus. The effects of D-parabactin, L-para-

bactin, and citrate on the growth of Paracoccus denitrificans in iron

deficient media were determined. The results demonstrate the presence

of a stereospecific iron transport apparatus inducible by a low


concentration of utilizable ferric ion in the cell medium. Additional-

ly, the mechanism by which Paracoccus denitrificans utilizes its sider-

ophore to supply iron to the cell is examined.

Materials and Methods

Culture Media and Glassware

Water was purified by distillation in a Mega-pure still (Corning)

and passed through a deionizing filter (Sybron-Barnstead) prior to use.

All glassware was soaked in 3N hydrochloric acid for at least 1 h and

rinsed well with purified water.

Liquid culture medium contained per liter of water the following:

13.5 g of sodium succinate, 4.0 g of KH2PO4; 4.9 g of Na2HP04; 1.6 g of

NH4Cl; 0.4 g of MgS04/7H20; 1% v/v Tween 80 and 4.5 lmol MnS04. The pH

was adjusted to 7.0 with 5 M NaOH and filter sterilized through a 0.22 1

filter (Milipore). Atomic adsorption analysis indicated 2.0 PM iron

present in this culture medium.

Growth Studies

Iron-free ligands were added in methanol to empty, sterile Klett

flasks, evaporated to dryness under nitrogen passed through a 0.22 P

Millipore filter and redissolved in the sterile growth media. In addi-

tion ethylenediamine-di(O-hydroxyphenylacetic acid) (EDDA) and

Fe (NO3)3 were added to the liquid culture medium in each experiment as

required to maximize differential growth response to the various lig-

ands. Inoculations from a 24 h trypticase soy broth culture were made

into the liquid culture medium at 40 1L of broth per 10 mL of culture

medium and incubated with shaking at 300C. Growth rates were deter-

mined by monitoring Klett readings vs. time. Standard curves were gen-

erated correlating Klett units with colony forming units using serial

plating techniques. Growth studies were reported as CFU's vs. time.

Transport Assays

Individual colonies of Paracoccus denitrificans were inoculated

into 20 mL of trypticase soy broth and incubated for 24 h at 30'C with

rotary shaking. Inoculations were then made into liquid culture medium

at 20 UL of broth per 10 mL of culture media and incubated with rotary

shaking at 30*C. Cells were harvested after 48 to 72 h by centrifuga-

tion at 5000 g for 20 minutes at 4C, washed twice with cold culture

medium and resuspended in fresh culture medium to a Klett reading of

150. Stock solutions of [55Fe] and [3H] labeled chelates in Tris HCl

buffer pH 7.5 were added to the cell suspensions and incubated with

rotary shaking at 30'C. At various time points, aliquots were removed

(200 uL) and immediately diluted with cold culture medium (5 mL). This

mixture was then rapidly filtered through a Gelman GA-6 membrane fil-

ter. The filters had been presoaked for 24 h in a 1 mM unlabeled che-

late solution to decrease adsorption of the radiolabeled chelate to the

filters and were rinsed once with culture medium (5 mL) before filter-

ing the cell suspensions. After the cells had been collected, the fil-

ters were rinsed with cold culture medium (5 mL). Liquid scintilation

counting of the air-dried filters was then performed in 10 mL of

Biofluor cocktail. Control values of labeled chelates adsorbed to the

filters in the absence of cells were determined, then substracted from

the values obtained with cells present. Uptake of labeled chelates by

cell suspensions of Paracoccus denitrificans is presented as "percent

uptake," which indicates the percentage of the total amount of added

label that has been taken up by the cells.


Various organic acids such as citric, succinic and malic, have

been suggested to play a role in iron metabolism. When the tricarbox-

ylate citrate was used in relatively high concentrations it can furnish

E. coli with iron in an apparent high affinity system.61 An 80,500

molecular weight outer membrane protein is expressed which has been

shown to be a specific ferric citrate receptor.61 It appears that

citrate may also act as a low affinity transport system in other

microorganisms as seen in the mycobacteria. While it was shown that

ferric citrate could promote the growth of the microorganism, there was

no evidence for the induction of specific membrane proteins.62

An experiment was performed to determine whether citrate was

capable of supplying iron to Paracoccus denitrificans. We compared the

effects of citrate and L-parabactin on the growth stimulation of

Paracoccus denitrificans in liquid culture media containing the

synthetic iron chelator EDDA. This chelator forms an iron complex

which the microorganism cannot utilize and hence results in a medium

which is severely lacking in any utilizable free iron. Figure 2-1

compares the ability of citrate and L-parabactin, present at 0.8 UM, in

promoting the growth of the microorganism. It can be seen that the

addition of L-parabactin produced a significant enhancement in growth

rate when compared to citrate or controls throughout the experimental

period. Citrate was incapable of facilitating stimulation of microbial

growth under these experimental conditions. The growth stimulation

from L-parabactin was expected, since this is the siderophore which




E 8.75




8.0 n
10 15 20 25 30 35 40 45 50
TIME (hours)
Figure 2-1. Growth rate of Paracoccus denitrificans in the
presence of L-parabactin (0) (0.8 uM), citrate (e) (0.8' uM), or
controls (1).


Paracoccus dinitrificans produces under low iron conditions. To deter-

mine whether Paracoccus denitrificans displays any stereospecificity in

its ferric parabactin-mediated iron-transport system, L-parabactin and

its enantiomer D-parabactin were compared for their ability to stimu-

late microbial growth. Both ligands would be expected to compete very

effectively for iron chelated to EDDA. In addition the iron (IIl) for-

mation constant for L-parabactin would be identical with that of the

enantiomeric form of parabactin. In this way any differences in growth

stimulation could be directly attributed to cell recognition of the

isomer with a preferred coordination about the metal center. High

field 1H NMR studies of the gallium (III) complex of L-parabactin and

the circular dichroism spectrum of the ferric siderophore indicate that

ferric L-parabactin exists in solution as the A-coordination isomer to

the exclusion of the A isomer.63,64 In contrast, by substituting

D-threonine into the synthetic scheme one produces the enantiomeric

form of parabactin which will chelate ferric ion to form the A coordin-

ation isomer to the exclusion of the A isomer.

Figure 2-2 demonstrates the effects of 20 pM L-parabactin and

D-parabactin on the growth of liquid cultures of Paracoccus denitrifi-

cans. Again, the addition of L-parabactin at the time of inoculation

resulted in a stimulation of growth throughout the experimental period.

In contrast, D-parabactin was much less efficient in stimulating the

growth of the microorganism compared to L-parabactin. However, there

was a slight growth stimulation in the presence of D-parabactin when

compared to controls. In this experiment the concentration of EDDA






8 I I I I I
8 12 16 20 24 28 32

TIME (hours)

Figure 2-2. Growth rate of Paracoccus denitrificans in culture
media containing ferric nitrate (0.05 VM) with L-parabactin (0) (2.0
UM), D-parabactin (A) (2.0 uM), or no carrier (e).

in the liquid culture medium was 0.55 mM while the concentration of
ferric nitrate was 0.05 vM. It appears likely that under these condi-
tions D-parabactin is operating via a mechanism other than the high
affinity iron uptake system. A possible explanation is that a slow

diffusion process is occurring whereby the hydrophilic D-parabactin
ferric chelate passes across the outer membrane of the microorganism
into the cell. This would not be expected to occur to any substantial
degree as the molecular weight of the ferric complex is very near the

exclusion limit of porin structures present in the outer membrane of
gram-negative bacteria.65 However, over the time course of the
experiment (ca. 32 h), small amounts amounts could conceivably pass
through the outer membrane and provide iron to the cell.
In a similar experiment, we compared the ability of L-parabactin

and D-parabactin to stimulate the growth of Paracoccus denitrificans
under severe iron deprivation conditions (Fig. 2-3). In this experi-
ment the concentration of EDDA present in the media was doubled and
ferric nitrate was omitted. The results from this experiment indicate
quite conclusively that Paracoccus denitrificans displays stereospec-
ificity in its siderophore-mediated iron transport apparatus. Only
L-parabactin was able to effectively stimulate growth throughout the
experimental period. Under these experimental conditions D-parabactin
was ineffective in facilitating growth stimulation as compared to con-
trols. This enantiomeric form of parabactin may be able to function

only as a low affinity iron transport system, and therefore be incap-
able of producing growth stimulation at the extremely low available

iron concentration present in this experiment. Only when there is
increased iron availability does one observe a growth stimulation

effect from D-parabactin.




9.0 .




5 10 15 20 25 30 35 40 45 50 55

Figure 2-3. Growth rate of Paracoccus denitrificans in the
presence of L-parabactin (0) (2.0 pM), D-parabactin (A) (2.0 PM), or
controls (.) without added ferric nitrate.

We have recently examined the mechanism by which Paracoccus
denitrificans utilizes its siderophore L-parabactin in its iron trans-

port apparatus.19,66 Siderophore-mediated iron transport was inves-
tigated by incubating cell suspensions with the ferric complex of

L-parabactin in which radiolabel was present either on the ligand or on

the metal. In this way it would be possible to differentiate between

the uptake of intact siderophore-metal complex "versus" uptake of iron

without corresponding accumulation of ligand. To monitor the fate of

radiolabeled metal, cell suspensions of Paracoccus denitrificans, grown

under iron deficient conditions, were presented with [55Fe]-ferric

L-parabactin at 1.0 vM and incubated with rotary shaking at 30C. Ali-

quots were removed at various time points, then filtered through a

0.45 V filter which was housed in a vacuum manifold system. After

rinsing, the filters were dried and the retained Fe was determined by

scintilation counting. The incorporation of radiolabeled metal contin-

ued throughout linearly for 1 h and then leveled off at approximately

50% of the total amount of radiolabel (Fig. 2-4). To determine whether

uptake of siderophore was occurring with uptake of iron, cells were

presented with the iron complex of the radiolabeled ligand [3HI

L-parabactin. In contrast to the rapid incorporation of radiolabeled

metal from [55Fe] ferric L-parabactin, there was no appreciable

uptake of labeled siderophore in cells incubated with E3HI ferric

L-parabactin (Fig. 2-4). In both instances a small percentage of

radiolabel (10%) was associated with the cells at the earliest time

points. However, only in the case where cells had been incubated with


20 40 60 90 120

Figure 2-4. Uptake of [55Fe] ferric parabactin (0), and [3H]
ferric parabactin (A) by Paracoccus denitrificans.


the [55Fe]-labeled siderophore was additional uptake of radiolabel
The previous data showed that Paracoccus denitrificans exhibits
stereospecificity in its siderophore requirement under low iron condi-
tions. To further demonstrate stereospecificity in its iron transport
apparatus [ 55Fe] ferric L-parabactin and its enantiomer [55Fe] ferric
D-parabactin were compared for their ability to donate iron to the
microorganism (Fig. 2-5). When cells were presented with E 55Fe ferric
L-parabactin there was, as expected, a rapid incorporation of radio-
labeled metal which had maximized after 60 min into the experiment. In
contrast, when cells were presented with [55Fe] ferric D-parabactin

there was no appreciable uptake of radiolabel in comparison to that
which was already cell-associated within the first minute of the assay.
It is thus apparent that enantioparabactin is incapable of supplying
iron to Paracoccus denitrificans in a high affinity transport system.
Considering the number and diversity of the iron specific sidero-
phores, it is not surprising that there are indeed a variety of mechan-
isms of microbial iron transport. Four such mechanisms have been elu-
cidated by which siderophores facilitate microbial iron assimilation.
In the "iron-taxi" delivery system67-69 the siderophore iron com-

plex binds to a specific membrane receptor protein in a temperature
independent step. The metal is then released from the ferric-sidero-
phore receptor complex to a presumed iron binding acceptor protein
which facilitates the transfer of iron into the cell. The deferrated
ligand, no longer having a high affinity for the specific binding pro-
tein, would then be displaced by another ferric siderophore molecule.



20 40 60 90 120
TIME (minutes)

Figure 2-5. Uptake of [55Fe] ferric parabactin (0), and [55Fe]
ferric enantioparabactin (A) by Paracoccus denitrificans.

The striking feature of the "iron-taxi" mechanism is that incorporation

of iron into the cell does not occur concomitant with siderophore

transport. In this way the siderophore remains extracellular at all

times allowing the siderophore to be reutilized in iron transport.

The more common mechanism of bacterial and fungal iron assimila-

tion is known as the "iron shuttle" delivery system.70'71 The notewor-

thy feature of this scheme is the transport of the intact siderophore-

iron complex, followed by intracellular dissociation of the metal-

ligand complex. The initial step in the "iron shuttle" mechanism, not

unlike the "iron-taxi," involves the temperature-independent binding of

the siderophore metal complex to a specific membrane receptor protein.

The next step, in contrast to the "iron-taxi," involves the transloca-

tion of the intact siderophore metal complex into the cell, presumably

to the site of the cytoplasmic membrane. Specific binding proteins,

located at the surface of the cytoplasmic membrane, then transport the

siderophore metal complex into the cytoplasm where the metal ligand

complex is dissociated. There have been two mechanisms advanced

describing the fate of the intracellularly released ligand. In the

"European mechanism," the subsequent step involves a translocation of

deferrated ligand from the cytoplasm to the extracellular environ-

ment.67 In this way the ligand can be reused to supply iron to the

cell. In the "American mechanism," after intracellular release of

siderophore bound iron, the ligand is degraded intracellularly.72

The ligand degradation products are then released into the extracellu-

lar medium.


A very pertinent example of iron transport mechanisms occurs in

the fungus Ustilago sphaerogena. This organism has evolved the means

for incorporation of iron by utilizing the hydroxamate siderophores

ferrichrome and ferrichrome A which it biosynthesizes. The trihydroxa-

mate siderophore ferrichrome donates iron to the cells in an apparent

"iron-shuttle" delivery system while the siderophore ferrichrome A

operates by the "iron-taxi" mechanism. 70,73 Before embarking on the

specifics of iron-transport in this system, it should be pointed out

that fungi are eukaryotic organisms and as such do not contain an

outer-membrane as found in the prokaryotic bacteria. Instead, they

possess a single cytoplasmic membrane which is surrounded by a hyphal

wall. In fungi, the receptors and iron-transport system are located at

the cytoplasmic membrane.74

In a double label experiment, Emery demonstrated that in the case

of ferrichrome, both ligand and metal were initially accumulated at the

same rate into the cell.69 When presented with [14C] and [59Fe]-

labeled ferrichrome, cell suspensions of Ustilago sphaerogena rapidly

incorporated label over the first 60 min of the experiment. At this

point in the experiment, the cells had taken up nearly 70% of the total

amount of 55Fe and 14C radiolabel present in the culture medium.

Interestingly, as the experiment progressed, there occurred a slow-

release of [ 14C]-labeled deferri-ferrichrome back into the culture

medium. This expulsion of deferrated ligand occurred until the percent

uptake had leveled off at approximately 40% of the total amount of

radiolabel present. In contrast, there was a continued uptake of metal

from [59Fe]-labeled ferrichrome which maximized near 100% uptake after
2 h.


To further investigate the mechanism of ferrichrome-mediated

iron-transport, the effects on cells incubated with the aluminum che-

late of [14C]-labeled deferri-ferrichrome were examined.69 Aluminum

forms a trivalent cation with a similar size radius as compared to that

for iron (III). As one might expect aluminum possesses a high affinity

for siderophores. Emery demonstrated that cell suspensions of Ustilago

sphaerogena rapidly incorporated the intact [14C]-ferrichrome Al (III)

chelate. Additionally, there was a marked absence of the release of

[14C]-ferrichrome back into the culture medium as was shown to occur
when the [14C]-ferrichrome iron complex was presented to the cells.

These results are in keeping with the proposed "iron-shuttle" mechanism

in which the metal and siderophore are taken up at identical rates.

After iron removal presumably by a reduction step, the free ligand is

expelled and reappears in the extracellular environment. In the case

of ferrichrome, the siderophore is released back into the culture medi-

um in unaltered form so that it may be reutilized in iron transport, as

in the "European mechanism." In the case of the aluminum chelate of

ferrichrome, the organism lacks a mechanism to remove the metal from

the stable aluminum complex so that release of free ligand back into

the culture medium does not occur.

Ecker amd Emery have also investigated the ferrichrome A mediated

iron transport system in Ustilago sphaerogena.70 By utilizing double
label transport assays with iron and gallium siderophore complexes, it

was determined that ferrichrome A was operating via an "iron-taxi"

mechanism. Whereas the [14C]-labeled ferrichrome iron-complex was

rapidly incorporated into the cells, there was no appreciable uptake
(<5%) of the [14C]-labeled ferrichrome A iron-complex. In contrast,

when [59Fe]-labeled ferrichrome A was presented to the cells there

was observed a near quantitative uptake of the metal after 30 min.

These results are in keeping with the proposed "iron-taxi" transport

mechanism in which only the metal is incorporated by the cells. In

this process the siderophore iron complex binds to a membrane receptor

where an enzymatic release of iron occurs. The deferrated ligand

remains extracellular and can be used again for iron transport.

In their continuing studies, Ecker and Emery witnessed an inter-

esting phenomena when the gallium complex of ferrichrome A was incubat-

ed with cells.70 While gallium (III) is very similar to iron (III) in

its coordination chemistry, it does not exist in a stable +2 oxida-

tion state.75 Substitution of iron by gallium would be expected to

result in a loss of metal incorporation if the siderophore were operat-

ing by an iron-taxi mechanism as enzymatic release of metal will not

occur. To the contrary, Emery observed a rapid and quantitative uptake

of metal when the [67Ga]-complex of ferrichrome A was presented to the

cells. When the [14C]-labeled ferrichrome gallium (III) complex was

incubated with the cells, there was also an incorporation of ligand

into the cells at a rate identical to the rate of metal uptake. An

explanation for these apparent contradictions in the iron-uptake mech-

anism mediated by ferrichrome A has been advanced. It would appear

that the mechanism of iron uptake from ferrichrome A depends upon the

ability of the cell to reduce the metal. Under iron-deficient or low

aeration conditions ferrichrome A operates by the iron-taxi mechanism.

However, if the cell cannot reduce the metal, as is observed under vig-

orous aeration, the substitution of iron by gallium or under iron-

replete conditions, the ligand is accumulated with the metal as an

intact complex.


The synthesis of ferrichrome in addition to ferrichrome A affords

a strong competitive advantage to Ustilago sphaerogena over other com-

peting organisms under conditions of iron deprivation. The finding

that only ferrichrome is synthesized under relatively high iron concen-

trations (50 vM) suggests a possible role in intracellular iron storage

or transport. Interestingly, at concentrations much higher than 10 UM

biosynthesis of ferrichrome A is completely shut down. Only under more

severe iron deficient conditions is the biosynthesis of this sidero-

phore up-regulated. It is not surprising that ferrichrome A possesses

a higher affinity for iron than ferrichrome. This is due to the con-

jugated nature of its hydroxmate functionalities in addition to the

tricarboxylic acid group which aids in lowering extracellular pH.70

Ferrichrome has been shown to be a potent growth factor in several fun-

gi and bacterial species.9 Indeed, E,. coli produces a specific

receptor for ferrichrome. Ferrichrome A, however, does not appear to

supply iron to other microorganisms and thus provides another means by

which Ustilago sphaerogena can monopolyze iron in a competing environ-


The high affinity iron transport system in Paracoccus denitrifi-

cans mediated by the siderophore parabactin appears to operate by the

"iron-taxi" mechanism. Cell suspensions grown under iron deficient

conditions rapidly incorporate iron when presented with [55Fe]-

labeled ferric L-parabactin at 30C. When cells were incubated with
[3Hj-ferric L-parabactin there was no appreciable uptake of label at

30GC. In experiments performed at OoC it was shown that uptake of

labeled metal had decreased dramatically, to a value nearly identical


to the [3HI-labeled ligand (Fig. 2-6). The first stage in the iron-

taxi mechanism involves binding of the siderophore iron complex to a

membrane receptor in a temperature independent step. It would appear

that the succeeding step involving iron release from the siderophore-

iron complex is impeded at OC so that iron transport into the cell

does not occur. The possibility still existed that parabactin was

operating by a mechanism other than the iron-taxi in delivering iron to

the cells. For instance, the parabactin iron complex could be entering

the cell intact followed by a rapid expulsion of deferrated ligand. In

this manner, the observable uptake of labeled ligand at any specific

time point would be negligible. To investigate such a possibility, it

was necessary to monitor the fate of the [3H]-labeled L-parabactin

gallium (IMl) chelate. As mentioned previously, gallium does not exist

in a stable +2 oxidation state. It has been shown that Paracoccus

denitrificans possesses a reductase capable of removing iron from para-

bactin. Thus, if parabactin were operating via an iron shuttle mechan-
ism one would expect to see transport of the intact parabactin gallium

complex inside the cell. Without a release mechanism for the metal

there would be an accumulation of radiolabeled ligand in the cell.

Contrary, when cells were incubated with the [3H]-labeled L-parabac-

tin gallium (III) complex, there was observed no appreciable uptake of

labeled ligand (Fig. 2-7). The small amount incorporated was identical

to that which was cell associated at the earliest time points. Thus it

appears very likely that parabactin is operating by an iron-taxi mech-

Additionally, the results indicate that the initial step in the

binding of the siderophore-iron complex to the cell membrane is a




20 40 60 90 120
TIME (minutes)

Figure 2-6. Uptake of [55Fe] ferric parabactin at 30% (0) and
40C 0); [3H] ferric parabactin at 30% (A) and 4C (V) by Paracoccus



20 40 60 90 120
TIME (minutes)

Figure 2-7. Uptake of [55Fe] ferric parabactin (0), and the
gallium (III) chelate of [3H] parabactin (A) by Paracoccus

specific process. Transport of radiolabeled iron was demonstrated only

for cells presented with [55Fe]-L-parabactin. Labeled iron pre-

sented in the form of the D-parabactin complex was not transported into

the cells. Similar results were obtained when measuring the ability of

L- or D-parabactin to stimulate the growth of liquid cultures of Para-

coccus denitrificans. Only in the presence of L-parabactin was there

considerable growth stimulation from the time of inoculation. Stimula-

tion from D-parabactin was observed only under less severe iron depriv-

ation conditions in an apparent low-affinity transport system. These

results suggest that Paracoccus denitrificans recognizes a specific

coordination isomer in which the chelating groups are in a left-handed

or A-coordination propeller about the metal, as in the case of the

ferric L-parabactin chelate and is able to discriminate between the

mirror image.

The stereospecific uptake of siderophore-iron complexes has also

been demonstrated in other microorganisms. There is now substantial

evidence that the chirality around the siderophore iron chelate is a

major factor that determines recognition of the complex by a specific

membrane receptor. Both Neurospora crassa and E. coli have receptor

systems which recognize the siderophore ferrichrome.76'77 This ligand

exists in the preferred A-coordination isomer. The synthetic enantio-

mer of ferrichrome which exists in the A-coordination isomer was inef-

fective in supplying iron to the microorganisms. It has also been dem-

onstrated that enantio-enterobactin was ineffective in donating iron to

E. coli as compared to the natural siderophore, enterobactin.76 Various

synthetic derivatives of enterobactin, including carbocyclic and

aromatic analogues, were tested to determine the structural boundary

conditions which the microorganism sets when utilizing the iron che-

late.78 Changes in the cyclized serine backbone of enterobactin

resulted in, surprisingly, only a slight decrease in its ability to

donate iron indicating that the chirality about the metal center is the

most important factor determining recognition by the bacteria.

Experiments with several parabactin analogues reveal that in Par-

acoccus denitrificans factors in addition to the chirality about the

metal center play a major role in iron transport and recognition.19

In fact relatively slight changes in the structure of the siderophore

were shown to drastically alter the iron-uptake characteristics. For

example, homoparabactin which contains a symmetrical 4,4 triamine back-

bone was investigated for its ability to stimulate growth of the micro-

organism in liquid culture medium. The results indicated that homopar-

abactin was far less efficient in promoting growth stimulation under

iron deprivation conditions as compared to parabactin. There was, how-

ever, an increase in growth as compared to controls in the presence of

homoparabactin. This sensitivity to variation of the chain length in

the polyamine backbone was further investigated by examining the abil-

ity of homoparabactin, norparabactin and parabactin to reverse EDDA

induced iron starvation on agar plates, seeded with Paracoccus deni-

trificans. At extremely low ligand concentration (0.1 nmoles) parabac-

tin and norparabactin were very efficient in their abilities to reverse

iron-starvation induced by EDDA. In fact, at 90 h the diameter of bac-

terial growth in the presence of norparabactin was just 15% less than

compared to parabactin. In contrast, there was no bacterial growth

detectable at 90 h in the presence of homoparabactin. As all three

ligands would be expected to exist in the A-coordination isomer, the

bacteria must use means in addition to chirality in the recognition of

the ferric-siderophore complex. It is evident that the disposition of

the spermidine backbone is one of those factors. While the symmetrical

3,3 backbone of norparabactin is recognized nearly as well as the

natural siderophore by the Paracoccus denitrificans iron transport

apparatus, the incorporation of one additional methylene unit into the

spermidine backbone is apparently a significant enough change to render

the complex unrecognizable to the membrane receptor or perhaps inter-

feres with the next step in the processing of the ferric siderophore

receptor complex.



Over the past few decades a great deal of research has been

amassed regarding microbial iron assimilation. In the 1950's came the

revelation that certain microbial natural products were functioning as

so called "iron carriers." Since that time a variety of these ferric

ion specific siderophores have been isolated and characterized. It is

now generally accepted that virtually all aerobic and facultative

anaerobic microbial species produce and/or utilize siderophores for the

sole purpose of solubilizing and sequestering iron to meet the metabol-

ic demands of the cell. While a number of siderophores have been

classified to date, it is safe to say that many more remain to be char-

acterized. Several intriguing mechanisms by which siderophores mediate

iron transport have been elucidated in both prokaryotic and eukaryotic

species. It is becoming more and more apparent that these iron uptake

mechanisms involve interactions between the siderophore iron complex

and specific surface receptors present in the cell. A logical progres-

sion which has followed over the past several years has been a shift to

research investigating the properties of the receptor proteins. A

recurring theme appears to be the induction of the membrane receptors

in conjunction with the synthesis of the siderophore in response to

iron deprivation.52 The demonstration of the induction of bacterial

outer membrane proteins in vivo during infection may hold much promise

in the eventual production of vaccines specific for these proteins.

While this potential avenue against infection is just now being recog-

nized, it is obvious that much needs to be learned regarding the nature

and composition of these proteins. Unfortunately, only in the case of

E. coli has a siderophore receptor been purified and its physical

characteristics elucidated.79

Our investigation into the mechanism of the parabactin mediated

iron transport apparatus in Paracoccus denitrificans strongly suggests

the presence of a membrane receptor responsible for the stereospecific

binding of the ferric L-parabactin complex. It is with these results

in mind that we have undertaken the isolation and characterization of

the membrane receptor protein responsible for siderophore binding and

iron removal in Paracoccus denitrificans. A potentially useful method

to explore the interactions between the siderophore iron complex and

the receptor protein is through the use of photoaffinity labeling. The

photoaffinity label contains a photoreactive functionality which upon

photolysis is capable of forming a covalent bond with the receptor pro-

tein. In this way a specific receptor protein can be selectively and

irreversibly labeled to facilitate its isolation and eventual structur-

al elucidation.

We have decided to examine the potential use of a photoaffinity

label to identify the membrane receptor protein responsible for binding

to the siderophore, parabactin. If successful, the properties of the

receptor will be investigated, e.g. molecular weight and chromatograph-

ic behavior. It would then be possible to isolate the unlabeled recep-

tor in order to characterize its physical and chemical properties.

The specific aims include the design and synthesis of a

photoaffinity label and the experiments evaluating its effectiveness at

substituting for parabactin at the receptor level. These include
1. A comparison of the conformation of the photoaffinity labels'

gallium chelate with that of parabactin's gallium chelate utilizing 300


2. The evaluation of the photoaffinity label's ability to pro-

mote the growth of Paracoccus denitrificans under limited iron condi-

tions; and

3. A comparison of the photoaffinity label's ability to promote

iron incorporation into cell suspensions of Paracoccus denitrificans

with that of parabactin utilizing both [55Fe] and [3H] labeled ligands.

Once having successfully shown that the photoaffinity label is

recognized by the parabactin receptor and that it promotes iron incor-

poration into the cell in the same manner as parabactin, it will be

possible to begin the actual labeling and isolation studies.



Reagents were purchased from Aldrich Chemical Co. and used with-

out further purification unless otherwise stated. The [55Fe] fer-

ric chloride, specific activity 43.6 Ci/g in 0.5 M hydrochloric acid

and the Biofluor scintillation cocktail were purchased from New England

Nuclear Corp. The [ 59Fe] ferric chloride specific activity 11.5 mCi/mg
in 1.0 N hydrochloric acid was purchased from ICN Biomedicals, Inc.

Activated CH-Sepharose 4B and Sephadex LH-20 were purchased from

Pharmacia Fine Chemicals. Gallium (III) nitrate nonahydrate was pur-

chased from Alfa. Filters used in the transport assays were Gelman

GA-6 membrane filters and were 0.45 pm pore diameter. Trypticase soy

agar and trypticase soy broth were purchased from BBL Microbiology Sys-

tems. Proton NMR spectra were recorded on a Nicolet NT-300, 300 MHz

instrument or a Varian EM-390 90 MHz instrument. IR spectra were

recorded on a Beckman Acculab 1 Spectrophotometer. Elemental analyses

were performed by Atlantic Microlabs, Atlanta, GA, or Galbraith Labora-

tories, Knoxville, TN. Melting points were taken on a Fisher-Johns

apparatus and are uncorrected. Photolysis experiments were performed

using a Rayonet type RS photochemical reactor. Gallium chelates used

for high field NMR studies were prepared by first solubilizing the fig-

and in an aqueous methanol solution at pH 8.5. A slight excess of gal-

lium III nitrate was then added and the pH adjusted to 7.5. The reac-

tion mixture was allowed to stir for 1 h at which time excess solvent

was removed in vacuo. Tritium parabactin was prepared and purified in

a manner identical to parabactin starting with L-[3H] threonine.

Final specific activity 1.05 x 107 DPM/mg.

Growth Studies

The effects of parabactin and parabactin azide on the growth of

Paracoccus denitrificans in iron-deficient media were determined.

Iron-free ligands (2 pM) were added in methanol to sterile Klett

flasks, evaporated to dryness under nitrogen which had been passed

through a 0.22 p millipore filter and redissolved in sterile growth

media. The media was used as previously described.19 Growth rates

were determined by monitoring Klett readings versus time. Standard

curves were generated correlating Klett units with colony forming units

(CFU) using serial plating techniques. Growth studies were reported as

CFU's versus time.

HPLC was performed on a Rainin Rabbit 4P/4PX system utilizing a

C-18 reverse phase column. The mobile phase contained 80% acetoni-

trile/20% 0.01 M ammonium phosphate buffer, pH 3.0. For preparative

HPLC, the buffer was omitted from the aqueous phase.

Transport Assays

Individual colonies of Paracoccus denitrificans were inoculated

into 20 mL of trypticase soy broth and incubated for 24 h at 300C with

rotary shaking. Inoculations were then made into liquid culture medium

at 20 pL of broth per 10 mL of culture media and incubated with rotary

shaking at 300C. Cells were harvested after 48 to 72 h by centrifuga-

tion at 5000 g for 20 minutes at 4C, washed twice with cold culture

medium and resuspended in fresh culture medium to a Klett reading of

150. Stock solutions of chelates including 55Fe and 3H labeled ferric
parabactin and parabactin azide were added to the cell suspensions and

incubated with rotary shaking at 300C. At various time points, ali-

quots were removed (200 pL) and immediately diluted with cold culture

medium (5 mL). This mixture was then rapidly filtered through a Gelman

GA-6 membrane filter. The filters had been presoaked in a 1 mM unla-

beled chelate solution to decrease adsorption of the labeled chelate

to the filters and were rinsed once with culture medium (5 mL) before

filtering the cell suspensions. After the cells had been collected,

the filters were rinsed with cold culture medium (5 mL). Liquid scin-

tilation counting of the air-dried filters was then performed in 10 mL


of Biofluor cocktail. Control values of labeled chelates adsorbed to

the filters in the absence of cells were determined, then substracted

from the values obtained with cells present. Uptake of labeled

chelates by cell suspensions of Paracoccus denitrificans is presented

as "percent uptake," which indicates the percentage of the total amount

of added label that has been taken up by the cells.

Photolysis Experiments

Various concentrations of [ Fe] or [ H] labeled ferric parabac-

tin azide were incubated with Par~occus deRitrificans membrane prepar-

ations in the dark for 30 min at room temperature. Samples were then

placed in quartz tubes and photolyzed in the Rayonet photo-reactor for

5-10 min while rotating in a carousel. Protein was precipitated by

addition of ammonium sulfate to 80% saturation. Samples (50 Pg) were

then subjected to SDS-gel electrophoresis as described by Laemmli.

Autoradiography of the dried gel was performed by placing the gel

contact with Kodak X-Omat film and developed after 7 days. For detec-

tion of tritium label the gel was incubated with salicylate as

described by Chamberlain81 before drying.

2-Hydroxy-4-nitrobenzonitrile (1) was prepared from 2,4-

dinitrophenylacetic acid as previously described.82

4-Amino-2-hydroxybenzonitrile'HCl (2). Compound 1 (2.10 g, 12.80

mnmol) was dissolved in 200 mL of ethanol containing 5 mL of concentrat-

ed HCI. The reaction mixture was heated to reflux, at which point 3.0

g of iron powder was added in portions over a period of 3 h with con-

tinued heating. The reaction mixture was then cooled to room tempera-

ture. The ethanol was removed in vacuo and the residue was taken up in

water (150 mL) and extracted into ether (5 x 50 mL). The ether was
dried and concentrated to afford the crude product. Further purifica-

tion was effected by chromatography on silica gel, using 10% MeOH/CHCl3

as eluant to afford 1.10 g (50%) of product: mp 184-1850C to afford

1.10 g (50%) of product: mp 184-185*C (lit. 186C).83

4-Azido-2-hydroxybenzonitrile (3). Compound (2) (0.90 g, 5.28

mmol) was suspended in 10 mL of ice-cold concentrated HC1. To this

cooled solution NaNO2 (0.65 g, 9.4 mmol) in 5 mL of water was added

slowly with stirring over a period of 1 h. This was followed by

addition of NaN3 (0.60 g, 9.2 mmol) in 5 mL of water. The mixture
was stirred for an additional 1 h; then the cooled reaction mixture

was filtered and the residue washed with ice-cold water to yield 0.72

g of (3) in 85% yield. IH NMR (10:1 CDCl3/DMSO-d6)a 6.58 (m,2H),

7.44 (d, 1H); IR (KBr), 3200 cm"1 (br), 2240 (s), 2110 (s), 1600 (s),

1480 (s), 1285 (s). m.p. 148-1490C.

2-Hydroxy-5-nitrobenzonitrile (8) was prepared utilizing pub-

lished procedures.84
5-Amino-2-hydroxybenzonitrile'HCl (9) was prepared and purified

analogously as (2) mp 160-162C (lit. 162C).85

5-Azido-2-hydroxybenzonitrile (10) was prepared and purified as

described for compound (3): 82% yield; 1H NMR (10:1 CDCl3/DMSO-d6) a
7.03 (m, 2H), 7.11 (d, 1H); 1R (Kr), 3200 (br), 2230 (s), 2100 (s),

1510 (s), 1260 (s).
Ethyl 2-hydroxy-4-nitrobenzimidate'HCl (5). Vacuum-dried
2-hydroxy-4-nitrobenzonitrile (1) (0.71 g, 4.33 mmol) was suspended in

10 mL dry absolute ethanol and dry HCI was bubbled through the cooled

mixture for 1 h. The reaction mixture was allowed to sit for 48 hr at

room temperature at which point the product was recovered by filtration

and dried under vacuum to afford 0.90 g (84%) of product. 1H NMR
(DMSO-d6) 6 1.40-1.52 (t, 3H), 4.54-4.68 (q, 2H), 7.70-7.75 (d, 1H),

7.91-8.0 (d, 1H), 8.04-8.09 (s, 1H). m.p. 196-1980C Decomp.

Anal. Calcd for C9H11N204Cl: C, 43.83; H, 4.49; N, 11.36.

Found: C, 43.75; H, 4.52; N, 11.32.

Ethyl 4-amino-2-hydroxybenzimidate'HCl J6. Ten percent palladi-

um on carbon (0.99 g) was added to a solution of (5) (2.10 g, 8.51

mmol) in dry absolute ethanol (100 mL). The solution was stirred under

a hydrogen atmosphere for 7 h. The solution was then filtered through

a medium (10-15) frit and washed with ethanol. The solvent was evapor-

ated and the residue was purified by chromatography on silica gel, us-

ing 10% MeOH/CHC13 as eluent to afford the product (6) in near quanti-

tative yield. 1H NMR (7:3 CDCl3:DMSO-d6) 6 1.54 (t, 3H), 4.58 (q, 2H),

6.03 (br, 2H), 6.28 (d, 1H), 6.42 (s, 1H), 7.58 (d, 1H), 9.41-10.70 (m,

3H) m.p. 155-156C.
Anal. Calcd for C9H13N202C: C, 49.89; H, 6.05; N, 12.93.

Found: C, 49.82; H, 6.08; N, 12.90.

Ethyl 4-azido-2-hydroxybenzimidate'HCl (7). To a suspension of

(6) (0.96 g, 4.43 mmol) in 100 mL of ethyl acetate was added freshly

distilled isoamyl nitrite (714 pL, 5.32 mmol) and 75 UL of acetic

acid. The mixture was cooled and stirred for 3 h at which time (0.35

g, 5.38 mmol) of NaN3 was added. After an additional 3 h the reaction

mixture that resulted was filtered and washed with ethyl acetate. The

solvent was removed, and the residue was purified by chromatography on

silica gel using 10% MeOH/CHCI3, as the eluent to afford (0.25 g)
product. 1H NMR (DMSO-d6) 6 1.49 (t, 3H), 4.13 (q, 2H), 6.48 (d,
1H), 6.64 (s, 1H), 7.75 (d, 1H); IR (KBr), 3020 (br), 2090 (s), 1595
(br), 1510 (br), 1260 (br), 1085 (br).

Nitro parabactin (8). Vacuum dried ethyl 2-hydroxy-4-nitrobenz-

imidate (5) (0.25 9, 1.19 mmol) and bis N1,N8-2,3-dihydroxybenzoyl)-
N4-(L-threonyl)spermidine'HBr28 (1) (0.59 g, 0.98 mmol) were heated at

reflux in dry, degassed methanol (50 mL) for 20 h. The solution was

cooled and concentrated. The residue was dissolved in ethanol and dry
packed on Sephadex LH-20. Column chromatography of the residue on
LH-20, using 20% EtOH/benzene as the eluent, afforded 0.55 g (85%) of

product. 1H NMR (10:1 CDCI3/DMSO-d6) 61.35-1.51 (2d, 3H), 1.51-2.01
(m, 6H), 3.18-3.72 (m, 8H), 4.58-4.63 (2d, 1H), 5.31-5.48 (m, 1H),

6.50-6.65 (M, 2H); 6.81-6.91 (m, 2H), 7.01-7.15 (m, 2H), 7.58-7.77 (m,
3H), 7.81-8.15 (m, 2H); mass spectrum m/e 666 (M + 1).

Anal. Calcd. for C32H35N5011 0.5 H20: C, 56.97; H, 5.38; N, 10.38.
Found: C, 57.02; H, 5.38; N, 10.32
Amino parabactin (9). Ten percent palladium on carbon (0.24 g)
was added to a solution of (8) (0.54 g, 0.80 mmol) in dry absolute eth-
anol (75 mL). The solution was stirred under a hydrogen atmosphere for
6 h. The suspension was then filtered through a medium (10-15) frit

and the residue was washed with ethanol. Concentration of the filtrate
was followed by purification of the residue on Sephadex LH-20, using
20% EtOH in benzene as the eluent to afford 0.45 g (89%) of product.
1H NMR (10:1 CDCl3/DMSO-d6) 6 1.35-1.51 (2d, 3H), 1.55-2.10 (m, 6H),

3.20-3.71 (m, 8H), 4.47-4.66 (2d, 1H), 4.70-5.01 (m, 2H), 7.16-7.31 (m

2H), 7.33-7.46 (m, 1H), 8.05-8.41 (m, 4H); mass spectrum, m/e 636

(M + 1).
Anal. Calcd for C32H37N509: C, 60.46; H, 5.87; N, 11.02.

Found: C, 60.45; H, 5.83; N, 11.05.

Parabactin Azide (10J. Vacuum-dried ethyl 4-azido-2-hydroxy-
benzimidate (7) (0.13 g, 0.63 mmol) and bis N1,N8-(2.3-dihydroxy-
benzoyl)-N 4-(L-threonyl)spermidine'HBr (I) (0.30 g, 0.50 mmol) were

heated in dry degassed refluxing methanol (50 mL) for 20 h. The solu-

tion was concentrated and the residue was dissolved in ethanol then dry

packed on Sephadex LH-20. Column chromatography on LH-20, using 15%

EtOH in benzene as eluent afforded 0.25 g (76%) of product. 1H NMR

(10:1 CDC13/DMSO-d6) 6 1.41-1.56 (2d, 3H), 1.55-2.10 (m, 6H),3.21-3.71

(m, 8H), 4.60-4.71 (2d, IH), 5.35-5.49 (m, 1H), 6.52-6.73 (m, 4H), 6.96
(d, 2H), 7.17 (t, 2H), 7.63 (t, 1H), 8.02-8.26 (m, 2H); IR (KBr), 3350

(br), 2110 (s), 1640 (s), 1265 (s) cm-1; mass spectrum, m/e 661 (M + 1)
Parabactin Coupled Resin. Dry activated CH-Sepharose 4B resin
(Pharmacia Fine Chemicals, Uppsala, Sweden) was swelled in 10-3 M HCl

(50 mL) for 30 min. The swelled resin was rinsed in 0.1 M HEPES buffer

(N-2-Hydroxyethylpiperazine-ml'-2-ethanesulfonic acid) pH = 7.5.
A solution of amino parabactin (67 mg) dissolved in coupling
buffer containing methanol (4 mL) and 0.1 M HEPES buffer pH = 7.5 (2
mL) was added to 3 mL of the swelled CH-Sepharose 4B resin in a sealed

glass tube under an argon atmosphere. The tubes were rotated at 12 rpm

for 24 h. The resin was then filtered and exhaustively washed with

coupling buffer. Unreacted N-hydroxysuccinimide ester groups on the

resin were then deactivated by addition of 1 M ethanolamine HCl (0.1

mL/mL gel) pH = 8 for 30 min. The resin was washed with 3 cycles of

alternative 0.1 M TRIS-HCl buffer pH = 8 (50 mL) and 0.1 M Na acetate

buffer pH = 4 (50 mL). A control resin was subjected to identical

coupling conditions as above, however parabactin was added in place

of amino parabactin.
To a dilute solution of methanolic FeCl3 (1 mg/mL) was added

250 UL of 59FeC13 (100,000 counts/t00 UL). To 1.0 mL of resin was

added 500 UL of the radioactive iron solution and the mixture rotated

for 12 h. The purple resin that resulted was first washed with coup-

ling buffer (3 x 3 mL) then suspended in a 10 pM EDTA solution (3 mL)

and rotated for 30 min. The EDTA washings were continued until

counts on the control resin which had been reacted with parabactin

were negligible. In order to determine coupling efficiency, the

amount of iron bound to the amino parabactin coupled resin was meas-

ured. Sample radiation of the resin was counted with an automatic

gamma counter (LKB-Wilac Ria Gamma 1274, Wallac Oy, Finland) and

indicated 20% coupling efficiency.

All spectral data are in the Appendix.


Design and Synthesis of a Parabactin
Photoaffinity Label
There are a number of ways one can consider activating parabac-

tin so that it will covalently and irreversibly bind to its bacterial

receptor, e.g. introduction of masked nitrenes, alkylating or acylat-

ing functionalities in the ligand. As we do not know anything about

the nature of the parabactin receptor, a fairly active nonfunc-

tionally selective device is required. The reagents selectivity must

be in its ability to bind largely to its own receptor and react with

it. In keeping with the synthetic problems and the required reactiv-

ity of the receptor label, we have chosen to consider parabactin

azides A + B (Fig. 3-1). The azide on photolysis will produce an

active nitrene capable of forming a covalent bond with the receptor

protein. Furthermore, the use of nitrenes generated photochemically

from azides to label isolated proteins or even cell associated

systems is well documented.86-88

The design and synthesis of the receptor label will of course

be predicated on parabactin. The molecule will ultimately contain a

tritium label. The boundary conditions we have set for this ligand

in terms of its required behavior are: (1) It must bind iron simi-

larly to parabactin. (2) It should promote uptake of iron by the

microorganism. (3) It must be accessible synthetically in moderate


The question of course arises as to precisely where to intro-

duce the azide functionality in the parabactin molecule. Clearly,

it must be introduced at a position which will not inhibit iron che-

lation, receptor association, or iron transfer into the cell. It is

clear from CPK models that we cannot put the azide next to the hy-

droxyl on the central aromatic ring because it is quite possible that

this will induce a different mode of ligand metal binding. This then

limits us to introducing the azide functionality para or meta to the

Figure 3-1. Structures of parabactin azide. (a) RI = N3,
R2 = H and (b) R1 = H, R2 = N3.

hydroxyl on parabactin's central aromatic ring (Figs. 3-2, 3-3). In

these systems the photoaffinity functionality should be far enough

removed from the ligand's chelating groups so as not to alter their

metal binding properties. The decision to place the azide function-

ality on the central aromatic ring as opposed to the terminal rings,

was based on the fact that the last step in the synthesis of parabac-

tin involves fixing of the central aromatic piece to N1,N -bis-(2,3-

dihydroxybenzoyl)-N4-threonyl spermidine (1).28 In this way the

azide functionality does not have to sustain any unwanted reaction

conditions involved in the synthesis of (I).

There is considerable precedent in the literature for the syn-

thesis of aryl azide imidates in high yield using the commercially

available 4-aminobenzonitrile. 89,90 In a model experiment designed

to evaluate the stability of an aromatic azide fixed to an oxazoline

ring, we synthesized ethyl 4-azidobenzimidate89 and successfully con-

densed it with the parabactin precursor (I) to yield the correspond-

ing parabactin analogue (Fig. 3-4) (mass spectrum m/e 646). The

compound proved to be stable in the dark and easily isolable by

chromatography on Sephadex LH-20. The stability of the azide imidate

to the condensation reaction conditions and its ease of synthesis did

not suggest there would be any problems in generating ethyl 4-azido-

2-hydroxybenzimidate. However, this was not the case.

As the last step in the synthesis of parabactin involves coup-

ling of ethyl 2-hydroxybenzimidate with (I) it is clear that this

imidate represents the optimum target for fixation of the azide

Figure 3-2. CPK model of the gallium (III) chelate of parabactin azide A.

Figure 3-3. CPK model of the gallium (III) chelate of parabactin azide B.




ClI 0 OH


f N

Figure 3-4. Synthesis of deoxyparabactin azide.


functionality. The standard protocol for the synthesis of aryl azide

imidates involves the initial conversion of a nitro aryl nitrile to

the corresponding amino aryl nitrile followed by conversion of the

amino function to the azide. The production of an azide generally

involves reaction of the amine in concentrated acid with aqueous sod-

ium nitrite followed by the addition of sodium azide. The intermedi-

ate diazonium compound is generally not isolated. This kind of syn-

thesis was attempted with 2-hydroxy-4-nitrobenzonitrile (1) (Fig.

3-5). The nitro group was first reduced to the corresponding amine

(2) utilizing iron and hydrochloric acid which proceeded in 50%

yield. Treatment of (2) with aqueous sodium nitrite in concentrated

hydrochloric acid followed by exposure of the diazonium compound to

sodium azide resulted in (3) in 80% yield. Unfortunately, we were

unable to convert (3) to the imidate (4) by treatment with anhydrous

ethanolic hydrochloric acid even in a variety of solvents. A similar

sequence was effected on 2-hydroxy-5-nitrobenzonitrile (8) (Fig.

3-6). Again the amino compound (9) was obtained in 50% yield and

conversion of this to the azide (10) proceeded in 80% yield. How-

ever, attempted conversion of this azide to the imidate (11) by

treatment with ethanolic hydrochloric acid, again in a variety of

solvents resulted in the amine (12). This collapse of the azide (10)

to the amine under acid conditions can probably be attributed to the

para relationship of the azide functionality to the phenolic hydrox-

yl. This instability suggests that ethyl 5-azido-2-hydroxybenzimi-

date might not be a suitable intermediate in the synthesis of




\b NO02

CL H2N5 C 2




(5) (6)

Figure 3-5. Synthesis of parabactin azide A precursor.


H2 CH3

02 N

pr OH
C =N






C =N

Figure 3-6. Synthesis of parabactin azide B precursor.

parabactin azide. Consequently, the alternative path (b) (Fig. 3-5)

was considered. The nitrile (1) was converted to the nitroimidate

(5) in 80% yield by treatment with anhydrous ethanolic hydrochloric

acid. We next determined that the nitro group in imidate (5) could

be reduced to the corresponding amine quantitatively by hydrogenation

over palladium. Fortunately the imidate was left intact and the

compound did not polymerize. Next the imidate (6) was taken to the

azide (7) in 30% yield by treatment with isoamyl nitrite and acetic

acid in ethyl acetate followed by exposure to sodium azide. The low

yield is probably associated with the hydrochloride's poor solubility

in ethyl acetate. Finally, ethyl 4-azido-2-hydroxy benzimidate (7)

(Fig. 3-5) was allowed to react with bis NI,N8(2,3-dihydroxybenz-

oyl)-N4-threonyl spermidine hydrobromide in refluxing methanol to

produce parabactin azide (Fig. 3-7) in 76% isolated yield.

All reactions after production of the amino compounds (2) and

(6) (Fig. 3-5) and (9) (Fig. 3-6) were carried out in minimal light.

The azide of interest, ethyl 2-hydroxy-4-azidobenzimidate when run on

TLC and exposed even briefly to a UV lamp for identification, quickly

turned yellow. Parabactin azide itself must also be handled

cautiously with regards to light exposure.

Proof of Structure

Several lines of evidence were utilized as proof of structure

for the azide. It is important to point out that because of the

compounds thermal and light sensitivity elemental analysis was






Figure 3-7. Synthesis of parabactin azide A.

unobtainable. The azide inevitably decomposed in the hands of the

analytical companies. A comparison of the 300 MHz 1H NMR of para-

bactin and parabactin azide reveals the two compounds to be essen-

tially identical, including duplicity of signals due to hindered

rotation about the central amide bond. However, as expected, the

aromatic region was different.

In order to assign the aromatic ring protons of the azide to

specific 1H NMR signals, we found it practical to compare the 300

MHz 1H NMR of the parabactin and parabactin azide gallium com-

plexes. In the spectra of the parabactin gallium (III) chelate, the

proton meta to the hydroxyl in the central aromatic ring is found to

be a triplet centered at 7.21 ppm (Fig. 3-8). However, in the spec-

tra of the parabactin azide gallium (III) complex, the signal at 7.21

ppm is absent (Fig. 3-9). These spectra indicate that the compound

at hand is a parabactin derivative substituted on the central aromat-

ic ring meta to the hydroxyl. The infrared spectrum of parabactin

azide run as a KBr pellet shows the characteristic absorption at 2110

cm-1 for N3. Unfortunately, the signal for the diazonium functional-

ity of the synthetic precursor can also occur at essentially the same


Additional proof of structure was obtained from the pyrolysis

and photolysis of parabactin azide (Fig. 3-10). The thermal decom-

position of parabactin azide was carried out in boiling tetralin at

2070. This resulted in a number of decomposition products including

amino parabactin (the synthesis of which is described below). This

arises from first loss of nitrogen leading to the unstable nitrene,

# H'_S



f--> doublet
e--> singlet
f--> doubletJ

i--> doublet
b--> doublet

Figure 3-8. Chemical shift and coupling data for the aromatic
ring protons of the parabactin gallium (III) chelate in d6-DMSO.

a coupled to f
b Coupled to e-f
C coupled to t
d coupled to i
e coupled to b
f coupled to oab
9 coupled to t

t coupled to cg.
I coupled to dt

a- 7.42 d
b- 7.21 t
C- 6.B4 d
d- 6.76 d
e- 6.66 d
f- 6.42 t
g- 6.35 d
t- 6.16 m
i- 6.07 t

7.q 7.2 7.0 6.8 6.6 6.4 6.2 6.0

7,q 7.2 7.0 6.8 6.6 6.4 6.2 6.0 PPM

Figure 3-9. 300-MHz 1H NMR spectrum of the aromatic ring
protons of the parabactin azide gallium (III) chelate in d6-DMSO.



HO CH 3 0 ON



Figure 3-10. Structures of the reaction products isolated from
the pyrolysis and photolysis of parabactin azide.

which then abstracts hydrogen from the solvent.91 The product

had an RF on silica gel identical with synthetic amino parabactin and

the same retention time on HPLC.

The photochemical decomposition of aromatic azides in the

presence of oxygen has recently been investigated.92,93 One of

the products, the corresponding nitro derivative was shown to be

formed by the trapping of triplet aryl nitrenes by triplet oxygen.

With the successful synthesis of nitro parabactin (see below), the

photochemistry of parabactin azide in the presence of oxygen could be

evaluated. The initial photolysis experiments were run using dilute

solutions (10-4 M) of parabactin azide in acetonitrile or benzene

under a steady stream of oxygen. Unfortunately, this resulted in the

formation of intractable tars and complete decomposition of the azide

as detected by HPLC. However, on a change of the solvent to acetone,

the azide was extremely stable and after photolysis for 24 h under a

stream of oxygen, several products were detected by HPLC. Indeed, a

major product had an identical retention time as synthetic nitro

parabactin, however, attempts at distinguishing this from parabactin

failed. The photodecomposition product was therefore isolated using

preparative HPLC. The isolated compound was then subjected to high

field NMR and the aromatic region compared to both synthetic nitro

parabactin and parabactin. Utilizing 1H NMR difference spectro-

scopy it was shown that the photodecomposition product was identical

to nitro parabactin indicating that indeed the nitrene was capable of

inserting into oxygen.

Finally, the most significant proof of structure was obtained

using fast atom bombardment (FAB) mass spectroscopy. A mass peak at

661 was observed corresponding to the intact azide along with a peak

at 635, arising from loss of molecular nitrogen and proton abstrac-

tion, corresponding to amino parabactin. It should be noted that

attempts at simple CI and El mass spectroscopy were unsuccessful in

observing a molecular ion. We regard the accumulated evidence as
proof of structure for parabactin azide. The synthesis has since

been repeated using tritiated threonine in the synthetic scheme thus

providing parabactin azide with a completely tritiated oxazoline

ring. The specific activity was determined to be 10.5x106 DPM/mg.

The ease of access and stability of the nitro imidate (5)

(Fig. 3-5) encouraged us to generate amino parabactin for use in the

production of a parabactin affinity chromatography column. The nitro

imidate (5) was condensed with the parabactin backbone (I) in reflux-

ing methanol to afford nitro parabactin in 85% yield. The nitro

functionality was then reduced to the corresponding amino compound by

hydrogenation over palladuim to afford amino parabactin (Fig. 3-11).

Amino parabactin was then reacted with activated CH-Sepharose 4B, in

aqueous methanol-HEPES buffer to produce the first polyamine sidero-

phore affinity column.

Activated CH-Sepharose 4B is produced by the esterification of

the carboxyl group of CH-Sepharose 4B using N-hydroxysuccinimide.

The N-hydroxysuccinimide active ester is highly reactive with primary

amino groups. Upon addition of ligand the N-hydroxysuccinimide group

is displaced and a stable amide bond is formed. Since activated



- "HO NH Br 0 OH
0 '- 0
0on, NNno




TO" CH3 /-0 O




NO1 0 N ON 0 OH



Figure 3-11. Synthesis of nitroparabactin and aminopara-

CH-Sepharose 4B is specific for primary amino groups other active

groups present in the ligand do not need to be protected prior to

coupling. The reaction was run for 24 h at room temperature and the

resulting matrix washed exhaustively with coupling buffer. The

washing was continued until the eluant was negative to ferric
radioactive assay employing 59FeC13 was used to
chloride. Araiatvasaeplyn Fe]wsuedo

determine the amount of ligand coupled to the resin. Gamma counting

of the 59Fe+3 chelated to the ligand, which binds in a 1:1 stoichi-

ometry, indicates a 20% coupling efficiency of amino parabactin with

the resin.

Growth Stimulation and Uptake Characteristics
of Parabactin Azide

The key issue in the design of a parabactin photoaffinity label

is that it be recognized by the Paracoccus denitrificans parabactin

receptor. The simplest way to ascertain whether or not the modified

parabactin is utilized by this receptor is simply to measure its

ability to stimulate the microorganism's growth in a low iron envi-

ronment. Consequently, both parabactin and parabactin azide were

compared in their ability to stimulate the growth of Paracoccus deni-

trificans in a minimal iron media (Fig. 3-12). As is clear from this

figure, both parabactin and parabactin azide stimulate microbial

growth to essentially the same degree. To further substantiate that

parabactin azide behaves biologically like parabactin, a closer

examination into its mechanism of iron-transport was required. It

has been demonstrated that parabactin supplies iron to Paracoccus









7 75
0 4 8 12 16 20


Figure 3-12. Growth rate of Paracoccus denitrificans in the
presence of parabactin (m) (2 pM), parabactin azide (A) (2 1M), or
controls (0).

denitrificans via an "iron-taxi" mechanism in which only the metal is

incorporated into the cells. We have compared the ability of para-

bactin azide to promote [55Fe] incorporation into the micro-

organism with that of parabactin.

Cell suspensions of Paracoccus denitrificans in late log phase,

grown under iron deficient conditions, were tested for their ability

to transport 1 pM [55Fe] ferric parabactin azide and [55Fe] ferric

parabactin. The data indicate that both parabactin and parabactin

azide promote iron uptake to essentially the same degree. There is a

fairly rapid incorporation of the [55Fe] label into the cells

during the first 10 min of the experiment, followed by a steady

increase until the iron uptake begins to level off at about 35% of

the total label taken up after 45 min (Fig. 3-13). In the proposed

"iron-taxi" mechanism of iron transport in Paracoccus denitrificans

there is no appreciable transport of tritium labeled ferric

L-parabactin. The ligand experiences only a transient association

with a membrane receptor and remains extracellular. In order to

monitor the fate of ligand in the ferric parabactin azide complex,

cell suspensions were presented with the [3H] ferric parabactin

azide complex. As demonstrated in Figure 3-14 there is no

appreciable transport of tritium labeled ligand (<10%) into the cell.

There was observed an initial 4% cell-associated label after the

first 5 min of the experiment which remained fairly constant over the

remainder of the experiment. As expected, there was a steady uptake

of the [55Fe] labeled parabactin azide over the course of this

experiment. Therefore it is apparent that in the dark parabactin








a 5 10 15 20 25 30 35 40 45

Time (minutes)

5 Figure 3-13. Uptake of [ 55Fe] ferric parabactin (a), and
5 Fe] ferric parabactin azide (0) by Paracoccus denitrificans.




0 25




5 10 15 20 25 30 35 40 45

Time minutes)

3 Figure 3-14. Uptake of [55Fe] ferric parabactin azide (), and
I3H] ferric parabactin azide (9) by Paracoccus denitrificans.

azide promotes iron incorporation into the cell in the same manner as
parabactin i.e., via the "iron-taxi" mechanism. It was also impor-

tant to demonstrate that parabactin azide was substituting for para-

bactin at the parabactin receptor site. An experiment designed to

evaluate the effectiveness of parabactin azide at competing with

parabactin for the receptor site was performed. Cells were presented

with 10 UM [55Fe]-ferric parabactin and various concentrations of

nonradiolabeled ferric parabactin azide. The cells were incubated in

the dark with rotary shaking at 30*C for 5 min, then assayed for

uptake of radiolabeled iron. The data (Fig. 3-15) fit the curve

predicted by simple competitive displacement of radiolabeled ferric

parabactin by nonlabeled ferric parabactin azide in a ligand-receptor

binding model with properties analogous to Michaelis-Menten enzyme

kinetics,94 with a KI = 2 VM. It should be mentioned that in this
experiment the percent uptake of radiolabel in the absence of

unlabeled parabactin azide was adjusted to 100% uptake. Additional

values in the presence of nonradiolabeled parabactin azide were then


Photolabeling Studies

As an initial study, the photoreactivity of parabactin azide

was determined utilizing UV spectroscopy. Spectra of a 10 UM sample

of parabactin azide in culture media were obtained after repeated

exposures to ultraviolet light. A Rayonet type RS photochemical

reactor fitted with six 350 nm UV lamps was used as a light source.

The spectra showed an absorption peak at 282 nm, corresponding to the

azide functionality. This peak was gradually reduced upon photolysis


Figure 3-15. Competitive inhibition of [55Fe] ferric parabac-
tin uptake (N) by nonradiolabeled ferric parabactin azide in the
dark. Incubation time: 5 min.

until its complete disappearance after illumination for a total of
four min.
The choice of the proper wavelength for irradiation is an
important consideration in photoaffinity labeling. Although irradia-
tion with 254 and 300 rum lamps was very efficient in decomposing the
azide functionality, irradiation at these shorter wavelengths was
found to be very deleterious to the bacteria as determined from
regrowth experiments. Therefore, we have chosen the longer wave-
length of 350 nm which will photolyze the azide but have less adverse
effects on the bacteria.
The following experiment was designed to evaluate the effects
of photolysis on ferric-parabactin azide mediated iron uptake. It
was first necessary to determine whether photolysis alone had any
adverse effects on iron-uptake in cell suspensions of Paracoccus
denitrificans. Cells, grown in iron deficient media, were harvested
in late log phase. The resuspended cells were then placed in quartz
tubes and photolyzed for 5 min while rotating in a carousel which
held them 8 cm from the lamps. Cells were then assayed for E55Fe]
ferric-parabactin uptake at 1.0 pM ferric chelate. A control experi-
ment without irradiation was run consecutively. The data indicate
that photolysis of cells alone resulted in approximatly 50% inhibi-
tion of iron uptake during the first 10 min of the experiment as com-
pared to cells with photolysis (Fig. 3-16). The photolyzed cells
were then able to gradually recover until the percent iron uptake was
nearly identical to unphotolyzed cells after 75 min (data not shown).
As a consequence, in an experiment involving photolysis of ferric-
parabactin azide, photolyzed cells in the presence of ferric parabac-
tin were used as controls.




W 10


0 2 4 6 8 10 12 14 16 18 20

Time (minutes)

Figure 3-16. Effects of photolysis on [55Fe] ferric parabac-
tin uptake by a Paracoccus denitrificans.
Uptake of [55Fe] ferric parabactin by untreated cells [a] or by
cells which had been photolyzed prior to addition of radiolabel [e].

Uptake of [55Fe] ferric parabactin by cells which had been
incubated with 30 pM nonradiolabeled ferric parabactin [M or ferric
parabactin azide [T] and photolyzed prior to addition of radiolabel.

Photolysis of cells in the presence of ferric parabactin azide

would be expected to irreversibly inhibit iron uptake by binding co-

valently at the ferric-parabactin receptor site and preventing fur-

ther iron uptake. In order to assay for irreversible inhibition,

cell suspensions were incubated with nonradiolabeled ferric-parabactin

azide at 30 UM and irradiated for 5 min. The cells were then centri-

fuged and washed twice with ice-cold culture media to remove any

unbound ligand. Resuspended cells, preincubated with shaking at 30,

were then assayed for iron-uptake with 1 VM [55Fe]-ferric para-

bactin. The data (Fig. 3-16) show that photolysis of cells in the

presence of ferric parabactin azide resulted in a slight inhibition

of iron-uptake as compared to photolysis in the presence of ferric

parabactin (10-15%).

The main objective in synthesizing the parabactin azide photo-

affinity label was to identify the receptor protein responsible for

siderophore binding in Paracoccus denitrificans in order to facili-

tate its eventual isolation. Experiments performed along these lines

have involved the photolysis of parabactin azide in the presence of

Paracoccus denitrificans membrane components. Samples were then sub-

jected to gel electrophoresis and incorporation of radiolabel asses-

sed by autoradiography of dried gels. Initial photolysis experiments

were performed utilizing [3H] labeled ferric parabactin azide.

When whole membranes were photolyzed in the presence of the tritium

labeled ligand there was no incorporation of radiolabel observed into

any of the protein bands as detected by autoradiography. It was

assumed that the low specific activity of the tritium labeled para-

bactin azide coupled to the inherent difficulties in detecting the


low energy beta particles were the causes. To circumvent these draw-

backs the [55Fe] ferric chelate of parabactin azide with a high

specific activity was employed. This has the added benefit of facil-

itating detection without the need for pretreatment of the gel with a

scintilation fluor. To demonstrate the ability of [55Fe] labeled

ferric parabactin azide to label the parabactin receptor site, photo-

lysis was performed in the presence of a high concentration of photo-

affinity label (20 uM). In this way the specific activity in the

reaction mixture (200 4l) was increased substantially to 4.0x107

DPM/100 ul. The autoradiograph (Fig. 3-17) demonstrates an undesir-

able degree of nonselectivity under these reaction conditions. Quan-

titation of labeled protein by fractionalization and solubilization

of individual gel slices followed by scintilation counting also

indicated a low efficiency of incorporated radiolabel. In order to

increase the photoaffinity labels ability to selectively label the

parabactin receptor protein the concentration of ligand was reduced

substantially. Figure 3-18 shows the results from one such experi-

ment where the concentration of parabactin azide was 0.5 uM (Lane 1)

and 2.0 pM (Lane 2). While in no way definitive, the autoradiograph

demonstrates an increase in selectivity when photolysis was performed

at low ligand concentrations. The association of radiolabel with the

low iron inducible high molecular weight proteins is evident. It is

noteworthy that these same protein bands appeared very intense when

photolysis was carried out at the higher ligand concentration.

Figure 3-17. Incorporation of radiolabel into Paracoccus deni-
trifican membrane preparations which had been incubated with
20 pM [55Fe] ferric parabactin azide and photolyzed for 5 min.

Figure 3-18. Incorporation of radiolabel into Paracoccus deni-
trificans' membrane preparations whicg5had been incubated with
0.5 pM (Lane 1) and 2.0 uM (Lane 2) [ Fe] ferric parabactin azide
and photolyzed for 5 min.


The use of photoaffinity labeling has many advantages over

traditional affinity labeling methods. Affinity labeling typically

involves modified ligands containing alkylating or acylating groups

which are capable of reacting with various nucleophilic species.

Thus the presence of amino acids at the binding site of the target

macromolecule is necessary for the potential formation of a covalent

bond with the affinity label. Photoaffinity analogues generate reac-

tive intermediates upon photolysis i.e., nitrenes or carbenes which

are capable of insertion into nucleophilic centers and of addition to

double bonds and aromatic systems present in the macromolecule. 95,96

As the photo-generated species has a short half life (10-3 sec),

label generated away from the active site would be expected to

react preferentially with solvent before migrating to a nonspecific

site. The photoaffinity label has the added advantage in that it is

chemically inert in the dark. In this way the photoaffinity label's

binding characteristics may be determined compared to the natural

substrate. An examination of these reversible binding constants is

extremely useful in designing photolabeling experiments. Nonspecific

photolabeling caused by nonspecific reversible binding of the photo-

affinity label is the most significant undesirable occurrence in

photolabeling studies.97 By determining the amount of reagent needed

to saturate the receptor sites one can optimize the efficiency of

site-specific photoaffinity labeling. Utilizing ligand concentra-

tions much above this concentration would very likely result in

increased nonspecific binding.


The results suggested that the parabactin photoaffinity label

would be an effective means by which to identify the membrane recep-

tor protein responsible for siderophore binding in Paracoccus deni-

trificans. It was demonstrated that parabactin azide formed a galli-

um (III) complex identical with the parabactin gallium (III) complex

as determined by 300 MHz 1H NMR. Furthermore, parabactin azide was

shown to facilitate the growth of the microorganism as effectively as

parabactin under the low-iron conditions present in the liquid cul-

ture medium. One might argue that parabactin azide was supplying

iron to the cells by some other mechanism. While this appeared

unlikely, it was nevertheless necessary to examine the mechanism of

iron-uptake mediated by parabactin azide. When cells were presented

with [55Fe] labeled ferric parabactin azide, there was observed a

rapid incorporation of radiolabel. The rate of uptake was identical

to the rate observed in the presence of [55Fe] labeled ferric

parabactin. To monitor the fate of ligand, tritium labeled ferric

parabactin azide was presented to the cells. As observed for para-

bactin there was no appreciable uptake of the tritium labeled ferric

parabactin azide. These results demonstrated that parabactin azide

was delivering iron to the microorganism in a manner identical to

parabactin via an "iron-taxi" delivery system. The evidence also

indicates that parabactin azide is able to compete effectively with

parabactin for the parabactin receptor site in the absence of light.

Increasing concentrations of nonradiolabeled ferric parabactin azide

resulted in a steady decrease in the percent uptake of radiolabel

from 10 PM [55Fe] ferric parabactin with an approximate inhibi-

tory concentration, KI, in the range of 2-5 UM.


Initial experiments demonstrated that parabactin azide was, as

expected, sensitive to irradiation. An absorption peak at 282 nm,

corresponding to the azide functionality, was shown to disappear upon

illumination. The photoreactive azide group upon photolysis decom-

poses to form a reactive but short-lived nitrene species. The aryl

nitrene generated can occur in two electronic states, the singlet

containing paired electrons or the triplet containing unpaired elec-

trons. Triplet nitrenes typically abstract hydrogen with radical

formation whereas singlet nitrenes are capable of inserting into the

active site of the receptor protein.96 The photolysis and pyrolysis

experiments involving parabactin azide show that the ligand is cap-

able of undergoing insertion reactions. If the nitrene generated by

the photolysis of parabactin azide was efficiently binding to the

parabactin receptor site one would expect to irreversibly inhibit

iron uptake by the microorganism. When cell suspensions of Paracoc-

cus denitrificans were photolyzed in the presence of ferric parabac-

tin azide there was observed a small degree of inhibition of iron-

uptake as compared to photolysis in the presence of ferric parabac-

tin. Due to variation in iron-uptake from one experiment to another

the significance of the datawas not determined. A closer examination

of the data in Figure 3-16 reveals that photolysis of cells in the

presence of parabactin or parabactin azide results in significant

inhibition of iron-uptake as compared to cells photolyzed in the

absence of ligand. After 20 min into the transport assay there is a

four-fold decrease in iron-uptake by cells which were photolyzed in

the presence of parabactin or parabactin azide. One explanation for

this observed behavior is that we simply have saturated the cells

with iron when photolysis was performed in the presence of 30 vM

unlabeled ferric siderophore. However, this requires that a regula-

tory mechanism exists by which the iron-uptake apparatus can be rap-

idly shut down at the first sign of high iron conditions. In this

assay, the cells are presented with the ferric siderophore for 5 min

during photolysis after which the cells are rapidly centrifuged and

washed twice with cold buffer to remove excess ligand. As the entire

washing procedure is complete in just a few minutes, we do not feel

that this brief exposure to a high concentration of ferric sidero-

phore can adequately account for the dramatic inhibition of iron-

uptake. Another plausible explanation which is consistent with the

results is that parabactin itself is acting as a photoaffinity label.

This would certainly explain why there was only a slight inhibition

of iron-uptake in the presence of parabactin azide as compared to

controls containing parabactin. In order to justify such a hypothe-

sis it would be necessary to demonstrate that parabactin undergoes

some structural changes as a result of U.V. irradiation. A solution

of parabactin (20 uM) in liquid culture media was photolyzed in a

quartz cuvette for several min utilizing six 3000 A U.V. lamps. The

U.V. spectrum of the photolysis mixture revealed significant pertur-

bations at 284 and 252 nm which increased with longer photolysis

time. While these results are preliminary they do suggest that the

observed inhibition of iron-uptake during photolysis in the presence

of parabactin may be a consequence of its photo-reactivity. This

will require further evidence such as the isolation and characteriza-

tion of a photo-decomposition product. Another experiment which

may be very informative would involve photolysis of membrane

preparations in the presence of [55Fe] ferric L-parabactin. An

autoradiographic analysis of the separated proteins demonstrating

incorporation of radiolabel into protein bands would be direct evi-

dence implicating parabactin as a photoaffinity probe.

Photolysis experiments designed to identify the parabactin

receptor protein have for the most part been inconclusive. The major

setback has been a lack of selectivity in the binding of the label to

membrane preparations. Figure 3-17 demonstrates the lack of selec-

tivity at high ligand concentration where it appears that increased

labeling is proportional to increased protein concentration. At low-

er ligand concentrations the selectivity was increased (Fig. 3-18).

A protein band in the 80,000 molecular weight range was more intense

than other labeled bands as detected by autoradiography. This high

molecular weight protein is not the most concentrated protein in the

membrane preparations. While a definitive statement cannot be made,

it is interesting to speculate on the role of these proteins in

siderophore-mediated iron-uptake. The possibility exists that these

proteins are subunits of the same receptor protein and that the

observed binding to each is representative of the physiological func-

tion of the receptor protein.
In other photoaffinity systems scavengers have been employed to

decrease nonspecific photolabeling.98 These protective compounds

prevent the photoaffinity label from binding nonspecifically without

decreasing the effectiveness of the label to bind at the receptor

site. We have utilized sodium salicylate for this purpose. Its use-
fulness in decreasing nonspecific binding of ferric D-parabactin to
membrane preparations will be demonstrated in the following section.
Unfortunately photolysis in the presence of salicylate resulted in a
similar degree of nonselectivity. Other modifications in the photo-
labeling experiments were made in an attempt to increase the selec-
tivity and efficiency of the photolabeling. These included photo-
lysis of membrane components prepared with or without detergent as
well as photolysis at different wavelengths. In all instances the
same degree of nonselectivity was observed. While the parabactin
photoaffinity label has not been as successful as hoped, the results
do direct attention to the high molecular weight proteins present in
membrane preparations of Paracoccus denitrificans. Their nature and

potential role as membrane receptors will be the topic of further in-
vestigation in the following section. Finally it should be noted
that a photoaffinity label for the siderophore-mediated iron trans-
port system in Neurospora crassa has recently been synthesized.99
The label is a derivative of the coprogen class of hydroxamate sider-
ophores. The synthesis was accomplished by reacting coprogen B,
which contains a primary amino group, with the commercially available
N-hydroxysuccinimide active ester of 4-azido-benzoic acid. In the

absence of light the photoaffinity label was shown to be taken up by
the iron transport system in Neurospora crassa similarly to the
natural product. It was also shown to be a competitive inhibitor of
coprogen uptake, KI = 5 pM, in the dark. The researchers have
proposed the use of the photoaffinity label in the identification of
the siderophore receptor protein in Neurospora crassa.

OUTER MEMBRANE RECEPTOR IN Paracoccus denitrificans

A great deal of interest has been raised as a result of recent

findings which demonstrated that iron-regulated outer membrane pro-

teins are expressed by pathogenic bacteria in vivo during infec-

tion. 100-104 Pseudomonas aeruqinosa isolated directly from the lungs

of a cystic fibrosis patient was shown to express iron-regulated

outer membrane proteins.I00 Bacterial isolates from the urine of

patients with urinary tract infection have also been shown to express

these proteins.102'104 Some of these low-iron induced proteins have

been shown to function as specific receptor proteins for ferric

siderophores. The finding that siderophore receptors are surface

exposed has led investigators to speculate on the feasibility of

vaccine production specific for the outer membrane proteins.105 If

successful these anti-siderophore receptor antibodies would be

expected to block siderophore-mediated iron uptake in the bacteria

resulting in a decreased metabolic state.

Preliminary investigations have demonstrated the presence of

naturally occurring antibodies in human sera which react with the

iron-regulated outer membrane proteins of E. coli.41 One of these

proteins is the receptor for the siderophore, ferric enterobactin.

Antiserum raised to this ferric siderophore receptor protein has also

been shown to react with iron-inducible proteins in Salmonella

typhimurium and Klebsiella pneumoniae, species capable of producing


enterobactin. This suggests that the antigenic properties and molec-

ular homology of the ferric enterobactin receptor protein are highly


Little is known regarding the physical characteristics of

ferric siderophore receptor proteins with the exception of the E.

coli ferric enterobactin receptor protein. Obviously, more research

is warranted regarding the physical characteristics of other bacter-

ial siderophore receptor proteins, beginning with their identifica-

tion and isolation. In the present study we furnish direct evidence

for the existence of a siderophore binding protein in the outer mem-

brane of Paracoccus denitrificans. High affinity, stereospecific

binding activity for ferric L-parabactin to isolated membrane compon-

ents is demonstrated. The physical characteristics of the ferric

siderophore binding protein has been investigated including an appar-

ent dissociation constant as well as pH and ionic strength effects on

binding affinity. Purification of the membrane receptor binding pro-

tein has been effected utilizing a detergent extraction procedure in

connection with anion-exchange chromatography.
Materials and Methods

55FeC13, specific activity 43.57 Ci/g, in 0.5 M HCl and Bio-

fluor Scintillation Cocktail were purchased from New England Nuclear.

Sephadex G 25-50, ethylene diamine di-o-hydroxyphenylacetic acid (EDDA),

deoxyribonuclease I (D-4527), and ribonuclease A (R-5125) were pur-

chased from Sigma Chemical Co. EDDA was deferrated and purified

according to the method described by Rogers.106 Triton X-100 was


obtained from Serva Biochemicals Inc. Acrylamide and ammonium sul-

fate were obtained from Schwarz/Mann Biotech. Gallium (III) nitrate

nonahydrate was purchased from Alfa. D-Parabactin was synthesized

and purified in a manner identical to L-parabactin but using D-BOC-

Threonine instead of L-BOC-Thr.28 [3H]-L-Parabactin was prepared

analogously, starting with [3H]-L-threonine. Final specific activity

was 10.5 x 103 dpm/Vg. Salicylic acid was twice recrystallized

from hot 0.01 N HCI and sublimed at 700C in high vacuum.

Bacterial Strain and Culture Conditions

Paracoccus denitrificans (ATCC Strain 19367) was maintained on

trypticase agar plates. Individual colonies were inoculated into 20

mL of trypticase soy broth and incubated with rotary shaking for 24 h

at 30'C. Inoculations were then made into a low-iron minimal salts

liquid media at 200 UL broth inoculum per 100 mL and incubated with

shaking at 30'C. The low-iron minimal salts media was prepared as

previously described,19 but with the addition of 0.2 mg/mL EDDA

and 0.5% Tween 80. Cells were harvested when turbidometric optical

densities were between 0.10 and 0.16 as determined by a Klett-

Summerson colorimeter (A.H. Thomas Co.).

Protein Determination

Protein concentration was estimated using bovine serum albumin

as the standard by the method of Lowry et al.107 with modifications

for membrane protein samples as described by Markwel et al.108

Isolation of Membranes

Cells of middle to late log phase (3-5 mg protein per mL) were

harvested by centrifiguation at 7000 x g for 25 min at 40C, and twice

washed in 50 mM Tris HCI, pH 7.5 containing 10 mM MgCl2, DNase (1

ig/mL), and RNase (1 Vg/mL) to give a thick suspension. The cells

were resuspended to a volume of 40 mL and homogenized by two passages

through a prechilled French pressure cell (18,000 psi). After sit-

ting for an additional 30 min at 40C, the crude homogenate was cen-

trifuged at 5000 x g for 30 min at 40C to remove unbroken cells and

large debris. The resulting supernatant was carefully removed and

ultracentrifuged at 105,000 x g for 60 min at 40C to isolate the cell

envelopes. The high speed supernatant was designated to be soluble
"cytosol" proteins. The pellet was resuspended in 50 mM Tris HCI, pH

7.5 to a protein concentration between 10 and 20 mg/mL to give a

yellow-green liquid which could be stored at 40C for at least two

weeks without formation of sediment or apparent loss of binding

activity. This preparation was designated to be whole cell "mem-


Extraction of Membrane Proteins
Solubilized protein extracts selectively enriched in inner

(cytoplasmic) or outer (cell wall) membrane components were prepared

based on a method described by Schnaitman. 109 The membranes were

pelleted at 105,000 x g, then resuspended and extracted with 100 mM

Tris HCI, pH 7.5 containing 2% Triton X-100 and 10 mM MgCl2. In

this and subsequent operations, volumes were adjusted to achieve

working protein concentrations of ca. 5 mg/mL of suspension. This

suspension was incubated 1 h at 40C and centrifuged at 105,000 x g,

for 60 min at 40C. The supernatant was designated the "Mg+2 TX-l00

extract." The Mg+2 TX-l0 extracted pellet was then resuspended and

extracted with 100 mM Tris HCI, pH 7.5 containing 2% Triton X-1O0 and

5 mM EDTA. After incubation at 4C for 1 h, the EDTA Triton X-1O0

suspension was centrifuged at 105,000 x g for 60 min at 4C. The

supernatant, designated the "EDTA TX-100 extract," contained solubil-

ized outer membrane proteins and displayed the highest specific

activity for binding ferric L-parabactin. Treatment of the "EDTA

TX-100 extract" with ammonium sulfate at 80% saturation at 40C for 2

h salted out the proteins which, after centrifugation, formed a mass

which floated on top of the solution. This mass was isolated and

redissolved in 50 mM Tris HCl, pH 7.5 containing 0.4% Triton X-100

but no EDTA. The resulting solution was adjusted to a protein con-

centration of 4-5 mg/mL and filtered through a 0.45 Um membrane to

give a crystal clear, almost completely colorless solution (in con-

trast to previous steps which had a yellow-green tinge). This solu-

tion was designated as "solubilized receptor" preparation and was

stored at 40C.

Preparation of Iron (III) Parabactin Chelate

In a typical preparation of 1 mL of 20 vM Fe (III) parabactin

chelate, 13 jig (21 nmoles) L- or D-parabactin in 26 jL methanol is

placed in a 1.5 mL polypropylene microfuge tube and blown dry under a

gentle stream of nitrogen. The residue is then dissolved in 69 PL

(690 nmoles) 0.01 N NaOH which had been thoroughly deoxygenated and


stored under argon. After the residue is completely dissolved, the

solution is diluted with 100 pL deoxygenated water and 1.117 jg (20

nequiv) iron (III) in 30 pL is added. Typically the iron solution is

prepared by diluting an atomic absorption standard (1000 Ug Fe/mL in

2% HN03; Aldrich) with water. Thus, the 30 pL of iron solution

contained ca. 350 nanoequivalents of H+. The solution should con-

tain excess base at this point [690-5(21)-350 = 235 nequiv OH-].

The purple chelate solution is diluted to 1 mL with 800 1L 100 mM

Tris Cl, pH 7.5. It is now important to first centrifuge the solu-

tion at 10,000 x g for 5 min, then filter the supernatant through a

0.2 jim membrane in order to remove any ferric hydroxide present. The

iron (III) parabactin in 100 mM Tris Cl, pH 7.5 gives a linear Beer's

plot over the concentration range 5-200 uM with a molar absorptivity

of 1.20 x 104 at 332 nm. The gallium (III) chelate of parabactin

may be prepared exactly as for that of iron (III) except that it is


Sephadex G-25 Column-Centrifiguation Binding Assay

A method modified from that described by Penefsky110 was

used to evaluate ferric parabactin binding activity. A 20 mg ( ca.

10%) plug of silylated glass wool was packed into the bottom of the

barrel of a 1 mL disposable tuberculin syringe. Acid-washed Pyrex

glass wool was silylated with 1% chlorotrimethylsilane in toluene (30

min), washed successively with toluene (2X), methanol (3X) and

acetone (2X), and dried at 100*C. This pretreatment of the glass

wool was essential for reproducible, quantitative recovery of protein

in the eluate. The plugged syringe barrel was then placed in a 15 mL