University of Southern California Dissertations and Theses

Studies On The Metabolism Of Iron

(USC Thesis Other)

Studies On The Metabolism Of Iron

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content STUDIES OK THE METABOLISM OF IROH
by
Philip James Charley
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Biochemistry and Nutrition)
June I960
72-604-5
CHARLEY, Philip James, 1921-
STUDIES ON THE METABOLISM' OF IRON.
University of Southern California, Ph.D.,
1960
Chemistry, biological
University Microfilms, A X ERO X Com pany, Ann Arbor, Michigan
THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED
UNIVERSITY O F SO U TH ER N CALIFORNIA
GRADUATE SCHO O L
UNIVERSITY PARK
LOS A NGELES 7, CALIFORNIA
p h . D fovo ( $ o Q , b
This dissertation, written by
Philip.. James„.Charley.......
under the direction of h%3....Dissertation Com
mittee, and approved by all its members, has
been presented to and accepted by the Graduate
School, in partial fulfillment of requirements
for the degree of
DOCTOR OF PHILOSOPHY
Dean
Date.......... ™ .O 9 6 0 .
DISSERTATION COMMITTEE
Chairman
L W JL
PLEASE NOTE:
Some Pages have i n d i s t i n c t
p r i n t . Filmed as re ceiv ed .
UNIVERSITY MICROFILMS
ACKNOWLEDGMENTS
It is with gratitude that I acknowledge the
counsel of a brilliant teacher and able research leader,
Dr. Paul Saltman. He stood ever ready to interrupt what
ever he was doing to assist in a problem, to offer sug
gestions, or sometimes to offer solace after an experi
mental setback.
Special thanks are also due to Dr. John Webb,
Dr. John Mehl, and Dr. Arthur Adamson for many helpful
discussions and criticisms. Their unique abilities and
interests in various areas of physical and biological
chemistry provided catalytic ideas to overcome theoretical
and experimental difficulties.
To Dr. Norman Kharasch I owe a special debt. It
was his inspirational teaching and guidance which con
vinced me to enter Graduate School in quest of a Doctorate.
I offer sincerest thanks to the members of my
Doctoral Committee: Dr. Donald Visser, Dr. Walter Marx,
and Dr. Jerry Donohue. They were always willing listeners
and sage counsellors with respect to experimental pro
cedures and new theories of iron metabolism.
Without the help of my friends and co-workers,
Clyde Stitt, Dr. George Kunitake, Ernest Shore, and
Morton Rosenstein, this research would have been
immeasurably longer, perhaps impossible; and definitely
much less enjoyable.
To my wife, Kay, and to Jim, Linda, and Bill, my
deepest thanks for sacrificing thousands of hours of
family camaraderie during my studies and research.
Finally, my thanks to the Truesdail Laboratories
and to Dr. Roger Truesdail, whose assistance and encour
agement made this work possible.
ii
TABLE OF CONTENTS
ACKNOWLEDGMENTS
Page
ii
LIST OF TABLES iv
LIST OF ILLUSTRATIONS v
Chapter
I. HISTORICAL INTRODUCTION 1
The Mucosal Block
The Role of Blood in Absorption and
Transport
The Binding of Iron to Apoferritin
Chelates in Iron Metabolism
Membranes as Ion Exchangers
II. OUTLINE OF PROBLEM AND METHOD OF ATTACK . . 31
III. EXPERIMENTAL AND RESULTS ................. 3*+
Chelates as Factors in the Movement of
Iron from Blood to Storage Tissue
The Binding of Iron by Ferritin
Some Properties of Sugar-Iron Chelates
Absorption of Iron from the Gut
IV. DISCUSSION IQh
V. SUMMARY 112
LITERATURE CITED 115
LIST OF TABLES
Table Page
1. The ability of EDTA and citrate to enhance
the uptake of Fe59 from transferrin by
rabbit liver slices...................... 4-5
2. The ability of various chelating agents to
remove Fe59 from rabbit transferrin by
dialysis................................ 4-8
3. The ability of various metabolites to re
move Fe59 from bovine transferrin by
dialysis................................ 50
4-. The uptake of radio iron by liver slices
incubated in serum..................... 55
5. Per cent of Fe59 activity transferred from
ferritin to transferrin across a
dialysis membrane in the presence of
various solutes at 0.01 M .............. 59
6. The effect of incubation time, reductants,
and sequestering agents on the formation
of ferritin from apoferritin and ferric
ammonium citrate....................... 64-
7. Elemental composition of ferric fructose
complexes ............................. 87
8. The regulation of iron absorption from the
gut by serum iron-binding capacity . . . 102
iv
LIST OF ILLUSTRATIONS
Figure Page
1. Diagrammatic representation of suggested
mechanism for the absorption of iron
into a cell of the intestinal mucosa,
the movement of the iron within the cell
and its delivery into the blood............ 5
2. Removal of ferritin under anaerobic
conditions............................... 23
3. The effect of heterologous transferrin and
ferric ammonium citrate on the uptake of
iron by liver slices of two species .... 38
*+. The site of radioactive iron incorporation
in serum . ............................. 39
5. The uptake of Fe^9 from transferrin by rat
liver slices as a function of citrate
concentration in the medium......... *+2
6. The influence of 0..01 M citrate on the rate
of uptake of Fe?9 from transferrin by
rat liver homogenates in Visking sacs . . . *+7
7. Titration curves of ferric ion alone and
in the presence of glucose......... 71
8. Zone of stable complex formation of ferric
iron with fructose................. 73
9. Difference spectrum of "fast" ferric fructose
compared to "slow" ferric fructose .... 76
10. Difference spectrum of ferric fructose
compared to ferric ion .......... 79
11. Electrophoresis of ferric chelates at two
p H ' s ................................ 83
12. Experimental conditions used to investigate
iron flux through the mucosal wall of
rabbits.............................. 92
v
Figure Page
13. Absorption of ferric sorbitol from rabbit
duodenum as measured by increase in
serum F e ? 9 ............................... 93
1^. Absorption of ferric fructose from rabbit
duodenum as measured by increase in
serum F e 5 9 ............................... 96
15. Absorption of ferrous ion from rabbit
duodenum as measured by increase in
serum F e 5 9 ............................... 97
16. Absorption of ferric ion and ferric sorbitol
from rabbit intestine as measured by
increase in serum F e 5 9 ................... 101
vi
CHAPTER I
HISTORICAL INTRODUCTION
The Mucosal Block
The most widely held theory of iron metabolism is
based upon the concept of the mucosal block. This hypoth
esis was proposed by Hahn and co-workers (1) in 19^3 >
based on their findings that prefeeding phlebotomized dogs
large amounts of iron (ca. 10 mg per kg) reduced the
absorption from subsequent dosages of this ion. They
considered that the gastro-intestinal mucosal epithelium
could accept or refuse iron depending on the body's need
for this element. These results were similar to the
original experiments of Fontes and Thivoile (2) who found
that repeated phlebotomies resulted in progressively
reduced quantities of iron in the feces of a dog fed a
constant daily amount of iron. Hahn and Whipple had
reported in 1937 (3) a series of experiments in which
anemic dogs were fed iron and the incorporation of iron
into hemoglobin was used as an index of absorption. These
authors concluded that iron entering the body was used in
toto for hemoglobin formation. The absence of new hemo
globin formation indicated a lack of absorption.
Later, (1) the measurement of storage iron present as
ferritin or hemosiderin was used together with hemoglobin
formation to follow more accurately iron uptake from the
diet.
Up to this time it had been tacitly assumed that
if excess dietary iron had been absorbed or was present
due to blood transfusions, the surplus could be excreted
via the feces. The work of McCance and Widdowson (*+, 5»
6) however, showed that negligible amounts of iron are
excreted via the urine or feces. Several observers (7j 8)
have confirmed these observations.
In 1937 Laufberger (9) isolated a crystalline
protein from horse spleen which contained up to 23$ dry
weight of iron. This protein was subsequently called
ferritin. Ferritin has since been found in many tissues,
including liver, testes, kidney, marrow, gastro-intestinal
mucosa, and cardiac muscle (10).
Ferritin is relatively resistant to heat de-
naturation, so that by simply heating a tissue homogenate
at 80° C and filtering it can be readily separated from
most other proteins (11). The crude ferritin thus formed
can be purified by the addition of cadmium sulfate,
whereupon octahedral brown crystals are formed (9). The
crystalline material isolated in this way was found to be
inhomogenous and varied in the amount of iron it contained
3
(12).
The iron was found to be present In the ferric
form as micelles and to have the composition (FeOOHjgPO^
(13, 1*+, 15). Electron microscopy has clearly shown the
ferric hydroxide micelles in crystalline ferritin (16).
“ While the iron in ferritin is quite stable in air,
it is converted to the ferrous form under reducing con
ditions and can be dialyzed away from the protein,
especially if a small ferrous chelate such as CL,a! dipyridyl
is present (13). This phenomenon has permitted the
preparation of apoferritin, a protein of uniform molecular
weight (17)» which crystallizes into colorless octahedra
under the same conditions as ferritin.
Granick (10) proposed that the principal role of
ferritin is to store iron. When iron is needed by the
organism, ferritin releases ionic iron into the system.
The presence of dietary iron stimulates the formation of
ferritin. Granick (10) found a marked increase in
crystallizable ferritin from guinea pig mucosa following
the feeding of high levels of ferrous iron. Tissue
ferritin is also increased in cases of hemolytic anemia
where transfusion therapy is used (18).
Iron was not absorbed unless it was "needed," and
could not be excreted when in excess. To explain these
experimental findings, the hypothesis of the mucosal block
If
was proposed. This theory has been championed by Granick
in his several reviews (10, 20, 21, 22, 23). Granick
explicitly states that, “the mucous membrane of the
intestine acts as if it knew when the body needed iron."
Granick felt that the primary regulator of iron absorption
was the ubiquitous protein ferritin. These ideas have
formed the basis for most research in iron metabolism.
A schematic presentation of the mucosal block is
shown in Figure 1.
Granick explains this mechanism as follows (19):
The movement of ferrous iron from the intestinal
tract into the cell and the movement of iron out of
the cell into the bloodstream may occur by the
establishment of concentration gradients, which are
made possible by the ease with which iron can be con
verted from the ferrous to the ferric state. The
assumption of an oxidation-reduction potential, or
rate of oxidation, that is greater in the front part
of the cell, and a reducing mechanism that is more
active in the rear part of the cell, might account
for a concentration gradient across the membrane of
the cell. On entering the mucosal cell, the ferrous
iron would be oxidized to the ferric state; the con
centration of ferrous iron in this portion of the cell
would remain low, establishing a concentration
gradient that would result in a continuous migration
of ferrous iron into the cell. At the other end of
the cell, the iron would be reduced from the ferric
to the ferrous state and would diffuse into the blood
where it would attach to the transferrin, thereby
maintaining a low concentration in the blood and
providing for perpetuation of transfer.
Apoferritin, a protein molecule, is continually
being synthesized and broken down in the cell.
Ferritin, the combination of apoferritin and iron
hydroxide, may accumulate in the cell if large amounts
of iron are fed. Iron so bound would eventually be
5
Lumen
of
gut
Synthesis Decomposition
^Apoferritinii
_ + +Ferrrtin
8FeOOH+l POJ
Mucosal
bloc |
Fe** - Fe** - F e**0, - Fe*
-M-
Oxid.» Red.
Red
Apoferritin
+
Fe**
Red
Fe**— -
Red > Oxid.
Blood
Fe**q-Fe***
Fe**+CO,+
3 phenolic groups
of transferrin—
Fe***transferrin
Figure 1.— Diagrammatic representation of a
suggested mechanism for the absorption of iron into
a cell of the intestinal mucosa, the movement of the
iron within the cell and its delivery into the blood
(19).
made available by some reducing mechanism in which
ferric iron became ferrous, was separated from the
apoferritin, and migrated into the blood stream to be
attached to transferrin.
It is implied that the lowered oxygen concentration of the
mucosal cells during conditions of anemia would permit the
iron to stay in the ferrous form and thus permit greater
absorption.
Granick has further proposed that perhaps apo
ferritin may be a structural part of the membrane and
contain "pores" equivalent in size to the micelles observed
in ferritin. Iron entering the cell must pass through
these pores; if too much iron is fed, the pores become
clogged with deposited ferric hydroxide. This in turn
prevents further absorption, until the reducing mechanisms
of the cell can clear the pores and permit additional
ferrous iron to enter.
Although Josephs (21 *) is in fundamental agreement
with the mucosal block concept, he points out that,
Actually the iron must pass across a cellular
membrane to unite with intracellular ferritin from
which it is "parcelled out" to the plasma across
another cellular membrane. In the cell, iron is
temporarily stored, and the rate at which it is freed
to unite with the transferrin of the plasma is
dependent on conditions about which little is known.
As recently as October, 1959j Bothwell (25)
states, "The absorptive process is an active one, and the
mucosal cells have the capacity of accepting or rejecting
iron according to the body's needs."
7
Recently, however, experimental evidence has been
obtained which is in conflict with the mucosal block
hypothesis. Keiderling and Wohler (26) demonstrated with
rats that in the presence of maximal amounts of ferritin
in the intestine, iron absorption continues. They showed
that when iron absorption was proceeding with maximum rate
that the greatest quantities of ferritin and hemosiderin
were present in the mucosal cells. Supporting these
findings, Heilmeyer (27) demonstrated that both the hemo
siderin and ferritin content of the gastrointestinal tract
were approximately proportional to the amount of iron
absorbed through the mucosa.
Conditions of simple iron deficiency as found in
hypochromic anemia do result in increased absorption of
dietary iron, and thus lend credence to the idea that
absorption is regulated by need. However, Dubach et al.
(28) have shown that iron is well absorbed by those
suffering from pernicious anemia where the direct need
for iron is not the fundamental deficiency.
The Bantus of South Africa have been found to
absorb large amounts of iron in a condition termed
nutritional siderosis (29, 30). The diet of the Bantus
consists of a high carbohydrate corn product cooked in
crude iron pots. Experimental siderosis was produced by
Kinney et al. (31) who found marked increases in iron
8
absorption in rats fed a corn grits diet supplemented
with ferric citrate.
In the heritable diseases, hemochromatosis, iron
is extensively deposited throughout the body. Dietary iron
continues to be absorbed despite the high concentrations
of ferritin isolated from the mucosa of those who were
afflicted with this condition (32) and the large amounts
of stored iron in other organs (28, 33, 3*+) • Evidently
the "mucosal block" is bypassed or is inoperative in this
disease.
The most striking evidence against the regulation
of iron absorption by the "mucosal block" is found in the
experiments of Reissman et al. (35) who administered large
quantities of ferrous iron to dogs and rabbits either by
duodenal tube or rectally. They observed dramatic in
creases in serum iron, the levels reaching as much as ten
times those of the total iron binding capacity of the
blood. It should be noted that the total iron absorbed
was only a very small percentage of the iron introduced
into the intestine. The mucosal wall was morphologically
intact, thus ruling out necrosis or rupture as possible
explanations for iron uptake.
Since it often has been observed that children can
be poisoned with excess ferrous sulfate (36), it must be
concluded that toxic amounts of iron can be absorbed if
9
the dose is high enough and if it is presented in an
appropriate form.
In both normal and anemic dogs fed non-toxic doses
of ferric ammonium citrate, Hahn et al. (1) found that the
iron was absorbed primarily in the stomach and duodenum.
Using fistulas on similar animals, they found that iron
could be absorbed equally well through the mucosa of the
lower small intestine. Granick (37) found, after feeding
guinea pigs ferrous ammonium sulfate dissolved in molasses,
that the amount of iron absorption decreased progressively
between stomach and large intestine. His semiquantitative
method involved the isolation and crystallization of
ferritin from the mucosal wall.
Most studies of iron absorption have utilized
ferrous iron (2^), and it has been concluded that iron
fed as ferrous ion is better utilized than iron fed as
ferric ion (38). It has been ascertained that ascorbic
acid (39) and sulfhydryl compounds (*+0) enhance iron
absorption, undoubtedly due to the fact that they are able
to keep the iron in the ferrous form. Groen et al. (*+1)
have found that several organic acids including glutamic,
aspartic, citric, and tartaric, as well as ascorbic,
increase iron uptake. They ascribe this enhancement to
increased acidity in the gut which would prevent the
precipitation of ferric hydroxide. Ferric hydroxide
10
precipitates unless the pH is substantially below 5, and
at pH 8, the maximum concentration of ferric ion possible
is 10"^® molar.
Divalent iron is far more soluble than trivalent
and this fact alone could explain its increased rate of
absorption. It is interesting to note, however, that in
cases of iron deficiency anemia, practically any inorganic
iron preparation can be used successfully (^2). These
observations, in general, are obtained from clinical
trials where an increase in the circulating hemoglobin is
correctly taken as evidence of satisfactory utilization
(^3, M+). However, there is no evidence that ferrous or
ferric iron is absorbed as such. Critical experimental
data have not been presented to differentiate between the
relative efficiency of the two soluble ionic forms of iron
and of their many possible organic and inorganic combina
tions.
The work of Herndon et al. (M-5) has shown that
sorbitol enhances ferric iron absorption. Sorbitol is not
completely metabolized and has no reducing effect on
trivalent iron. These authors postulate that carbohydrates
(e.g., sorbitol) either stimulate the absorption
mechanisms of the intestinal wall or protect various
substances in the intestinal lumen. They cite the work of
Fournier et al. (*+6) and Wasserman and co-workers (*+7) who
11
showed that several carbohydrates, as well as amino acids,
enhance calcium absorption from the gut of rats. In this
connection, it is of interest to examine the later work of
Lengemann (*+8), who has shown that preloading animals (or
their intestines) with carbohydrates does not stimulate
calcium uptake. His results point out that lactose or
glucose must be present with the calcium in the same
portion of isolated gut at the same time. Under these
circumstances, the absorption increased dramatically. The
effect was observed with all the alkaline earths tested.
Thus, the proposed mechanism by which the mucosal
block regulates and controls iron absorption appears to
have severe deficiencies. In the dosage range of
medicinal iron (ca. 1 mg per kg body weight), the percent
absorption is inversely related to the dosage level (2*0.
The presence of various foodstuffs, reducing agents, and
compounds which effectively precipitate iron (e.g., phos
phates) all appear to play a significant role in iron
♦
absorption.
The Role of Blood in Absorption
and Transport
It is widely accepted that the plasma is the chief
medium for iron transport (^9» 50). Peterson and Mann
(51), using rats with total lymph fistulas, showed only
insignificant amounts of iron were absorbed via the
lymphatics.
Present theory (*+0) holds that iron moves from the
mucosal cell as ferrous ion into the blood stream. The
high oxygen tension in this environment quickly oxidizes
the iron to the ferric form. As the trivalent ion, it is
tightly bound to a ( S -1 globulin which has been termed
‘ 'transferrin" or "siderophillin" (52). Wallenius (53) has
shown that all measurable serum iron is bound to this
globulin so long as the iron binding capacity of the
protein is not exceeded. Normal human serum contains ap
proximately 100 jigm of iron per 100 ml, with a saturation
value of approximately 300 jagm per 100 ml.
Cohn et al. (5^0 have reported the "affinity con-
stant" of ferric transferrin to be of the order of 10'
"at neutral reactions." This iron-protein complex is very
sensitive to pH. Schade et al. (55) were able to follow
the dissociation of the complex by measuring its charac
teristic absorption at ^60 mja. They found the highest
stability in the alkaline region above pH 7.5* As the
solution is made more acid, dissociation takes place until
at pH *+.8 only half the iron is bound. At pH ^.0 there
was no spectrophotometrie evidence of any complex present.
The average human contains M-.5 gm of iron, of
which about 73$ is in hemoglobin. About 23.5$ is found
in the form of ferritin and hemosiderin, stored primarily
13
in the liver and spleen. Important, though small,
fractions are found in muscle myoglobin, enzymes, and
other non-heme proteins (*f0). Although only about ^ to 5
mg of iron are normally bound to the circulating transfer
rin, this minute amount is still adequate to supply the
20 to 25 mg which is estimated to be the daily requirement
for new hemoglobin synthesis (56).
Transferrin concentration is regulated by unknown
factors. It has been shown that it increases during
enhanced requirement for iron as in pregnancy and iron-
deficiency anemia when the total serum iron is low. Con
versely, the concentration decreases during hemochroma
tosis and transfusion siderosis, when the total serum iron
is high. It is also regulated in part by factors control
ling other plasma proteins, such as albumin, during
infections and allied conditions (56).
Iron present in the transferrin molecule is used
preferentially for heme synthesis (57) and the rate of
radioactive iron removal from the plasma has been used as
an index of erythropoesis. There appears to be a
preferential utilization of iron released from destroyed
red blood cells for erythropoesis when compared with
storage iron. The condition of the body's iron stores
does not appear to affect the rate at which iron from
hemoglobin is re-used in new blood cells (58). These data
lb
indicate that the interchange of iron between transferrin
and reticulocytes occurs more readily than between the
blood protein and storage depots.
It is chiefly when the erythropoetic need for
iron diminishes while the supply is still high that the
transfer between transferrin and tissue storage depots is
highest. Such conditions are observed in transfusion
siderosis. Under circumstances where blood loss is high,
the degree of saturation of transferrin with iron
decreases and the erythropoetic centers utilize the iron
stored as ferritin and hemosiderin.
The data available are consistent with the view
that iron transfer to storage takes place only when
transferrin is nearly saturated with iron. Conversely,
the net flux in the reverse direction takes place when
the transferrin molecule has considerable capacity for
iron binding (59).
Jandl et al. (59) have shown that iron is
effectively transferred to reticulocytes and incorporated
into heme at lower iron saturation levels of transferrin
than could be demonstrated using a similar system where
rat liver slices were competing for the transferrin iron.
These authors have also shown that specific receptor sites
exist on or in the reticulocyte which bind the iron with
great avidity prior to incorporation into hemoglobin.
15
The uptake of iron onto these sites could be
inhibited by metabolic blocking agents. It was found that
transferrin or conalbumin, the iron binding protein of egg
white, would serve with approximately equal facility in
transferring radio-iron to immature red cells. This
situation contrasts with unbound ionic iron, which ap
peared to be bound non-specifically to mature red cells
as well as the reticulocytes. It is to be noted, however,
that the iron concentration used in these experiments
(30 mM at pH 7.1 *) cannot exist as free ionic iron and must
therefore have existed as an undetected precipitate of
ferric hydroxide. The composition and reactions of ferric
hydroxide change with respect to time and environmental
conditions (60).
It has been shown (59) that transferrin and
conalbumin can compete with reticulocytes for the avail
able soluble iron. Once incorporated into the heme
molecule, the iron is effectively removed from competition
with transferrin until the porphyrin ring is enzymatically
cleaved subsequent to the destruction of the aged red
blood cell. Some concept of the avidity of the heme
molecule for iron can be understood from the fact that the
complex has been found in oil deposits millions of years
old (61).
16
Laurell (62) has proposed that iron leaves
transferrin in the ionic form and is transferred to either
tissue or immature blood cells as such. Flexner et al.
(63) have postulated that the entire iron-transferrin
molecule passes through the membrane into the cell. The
work of Jandl and co-workers (59) does not settle the
question of the form of the mobilized ion, although they
were of the opinion that the transferrin molecule did not
enter the cell. They presumed that the protein molecule
impinged directly on the reticulocyte cell wall at the
time the iron was transferred (59)•
Laurell (6>+) has argued for the concept of ionic
iron as the intermediate on the basis that transferrin
entering the cell would result in iron turnover values
greater than that actually observed and that proteolysis
of the transferrin (e.g., to make hemoglobin) within the
cell would result in large amounts of free amino acid. It
is evident, however, that if the intact transferrin
molecule could enter a cell it could also, within reason,
withdraw in the same way.
Granick (21) has postulated that iron leaves the
bloodstream in the ferrous form to enter the storage tis
sues; principally the liver, spleen, and bone marrow. His
reasons for this are not clear, especially when he notes
in the same paper that "ferrous iron entering the
17
bloodstream is immediately autoxidized to the ferric form.”
He remarks that ferrous iron is more soluble than ferric
but gives no evidence for a change in oxidizing conditions
between blood supplying the gut and that of the liver. In
view of the close physical relationship between the in
testinal and liver circulation, a measurable change in the
oxygen pressure between these points would not be expected.
Even though the solubility of ferric iron is very
low, and the stability constant of transferrin-iron is
very high, there still exists an equilibrium of the ion
with the protein. Inman (65) has shown that the chelating
agent ethylenediamine tetraacetate can compete for
transferrin iron even though it is not as effective a
chelating agent as the protein molecule.
These data indicate that transferrin need not be
proteolysed in order for the bound iron to be released.
The transfer of iron from transferrin to cell may be
regulated by the dissociation of the transferrin-iron
complex, so long as the solubility of the ferric ion is
not exceeded. Laurell (66) has more recently proposed
that the distribution of iron is regulated by the laws of
mass transfer.
Granick (10), quoting Dubach et al. (67) and
Yuile et al. (68), states that "the degree of siderophil-
lin saturation with iron does not appear to affect
18
appreciably the rate of movement of iron from the mucosal
cell into the blood stream.*' A review of the first
reference (67) reveals the following statement:
In unpublished experiments, the authors have con
firmed Hahn's observations that anemia by itself does
not influence iron absorption. Uptake from the
intestinal tract has been shown to be independent of
plasma iron concentration. On the other hand, if
tissue iron reserves of subjects with hypochromic
anemia are partially restored by the parenteral
administration of large amounts of iron, uptake of the
metal from the alimentary tract becomes less complete.
The possibility has been explored that a factor may be
present in the blood of iron deficient subjects which
stimulates absorption, but in two unpublished experi
ments the infusion of large amounts of plasma from
iron deficient into normal dogs has failed to affect
the quantity absorbed.
The work of Hahn (1) to which Dubach refers is
that in which the net absorption of from 1 to 10 mg of
iron fed per kg of body weight was determined by the
response in circulating hemoglobin l1 * days later.
Yuile et al. (68), using three normal and four
anemic dogs, measured the amount of radio-iron absorbed
from the gut by assaying the hemoglobin several weeks
later. This was done under four experimental conditions:
with the animals in the normal state; with the animals in
the anemic state; and with or without the injection of an
amount of iron calculated to just saturate the transferrin
for several hours. In the case of the normal animals, the
maximum amount of the administered dose absorbed was 6%.
Saturation of the iron binding protein with injected
non-radioactive iron decreased the amount of plasma radio
iron. However, the total amount of Fe^9 absorbed was
approximately the same in both cases. In the anemic dogs,
the control absorption of the test dose was as high as 66$.
In two anemic animals, presaturating the transferrin
resulted in a marked decrease in absorption, while the
other two animals under the same experimental conditions
showed but moderate or slight decrease. From these data,
the authors conclude "that the plasma iron-binding protein
and its relative saturation play little role in iron
absorption."
A dog weighing 5 kg contains approximately *+00 ml
of blood of which, perhaps, 200 ml is serum. To saturate
the transferrin in the blood would require about 600 yxg of
iron. The normal animal would have approximately 200 )ig
of iron prior to saturation, and the anemic dog might have
a total serum iron on the order of 100 ^gm. Thus, the
available capacity of the blood to bind iron or transferrin
would be approximately *K)0 jugm or 500 pigm, respectively.
Huff and co-workers (69) have determined the plasma iron
turnover in five normal human subjects to be 0.35 mg per
kg per day. If this figure may be used for dogs, one may
calculate that a 5 kg animal would use about 75 Jigm of
iron from transferrin every hour. The amounts of iron fed
in Yuile’s work ranged from 7 mg (with maximal percent
20
absorption) to 50 mg (minimal percent absorption). In
considering the work of both Hahn (1) and Yuile et al.
(68), it should be recognized that the effects of the
degree of iron saturation of transferrin could be
completely obscured by the magnitude of the iron dose fed.
While the role of transferrin in regulating iron
absorption has not been unequivocally established, it has
been determined that its presence is not required for
movement of iron into or out of tissues in vitro.
Saltman and co-workers (70, 71» 72), working with
rat liver slices, have shown uptake of iron by the slices
against apparent concentration gradients. The iron,
offered to the cells as the ferric ammonium citrate
complex, moved at a rate controlled by diffusion. Brown
and Justus (73) > using everted gut loops, also showed that
the movement of iron across the mucosa did not depend on
active transport. In vitro, the iron initially diffusing
into liver cells concentrates in the nucleus (7*0. These
experiments were performed using isolated liver cells
incubated in medium containing approximately 30 jig of iron
(as ferric ammonium citrate) per ml. In time, the major
fraction of the iron accumulates in the cytoplasm.
21
The Binding of Iron to Apoferritin
Most of the iron diffusing into the liver cell is
bound as cytoplasmic ferritin, although hemosiderin is
also present, especially when body stores are high. The
mechanism by which iron is bound to ferritin has been in
vestigated by several workers. Michaelis (75) has stated
that while iron may be removed from ferritin without dif
ficulty, the reconstitution of ferritin from apoferritin
could not be accomplished by an "unobjectionable pro
cedure." Neither iron salts in the ferric or ferrous form,
nor colloidal ferric hydroxide were effective in recon
stituting ferritin. However, Bielig and Bayer (76) have
reported that they were able to obtain crystalline
ferritin by oxidizing ferrous ammonium sulfate in the
presence of the apoferritin. Loewus and Fineberg (77)
also state that they were able to make ferritin from apo
ferritin by incubation of the apoprotein with ferric
ammonium citrate and a reducing agent or with ferrous
ammonium sulfate in air. Their assay was based on pre
cipitation of the protein-iron complex with ammonium sul
fate followed by spectropho tome trie assay at ^-00 mja. In
this connection it is of interest to note that the spec
trum of ferritin in this region is practically Identical
with that of hydrated ferric iron.
22
It has been postulated (78) that the ferric iron
in ferritin is reduced to the ferrous form before leaving
the surface of the protein. Mazur (79) has outlined the
scheme shown in Figure 2. It is proposed that ferritin
has a preponderance of disulfide groups and has ferric
iron at the surface of the molecule. In the hypoxic
liver, ferritin is converted to the active form in which
free sulfhydryl groups chelate the ferrous iron. The
ferrous iron is now capable of dissociation, permitting it
to penetrate the cell wall. Green and Mazur (80) have
implicated uric acid as the compound responsible for the
reducing action.
Such conclusions are based on the rapid release
of iron by ferritin in the presence of reducing agents and
by the apparent inability of transferrin to bind, ferric
ion from solution. It should be recognized, however, that
it is not possible to have appreciable concentrations of
ferric ion at hydrogen ion concentrations which will hot
denature either protein. Most tests involving ferric
, , ion, , at neutral pH’s require the use of a chelating agent
to keep the iron in a soluble form. It is possible that
the stability constant and dissociation rate of these
ferric chelates have affected the observed results and
the conclusions that have been drawn from them.
23
Colloidal micelles of
(Fe OOH) e (Fe 0 0 P 0 sH2)
Cell wall
S
• • • •
• •
♦
LIVER
^ O T
s
Protein
ACTIVE INACTIVE
-Fe**^ ■ Fe*
Fe*
Figure 2.— Removal of ferritin iron under
anaerobic conditions (79)•
Chelates in Iron Metabolism
Chelates have long been recognized to play key
roles in the metabolism of many trace metals. The struc
tures of the important metallo porphyrins have been
reported (81). The stability constant of transferrin iron
is so high (82) that a chelate ring is indicated as
participating in its structure. Even the ferric hydroxide
micelles of ferritin are complexes, which are believed to
be composed of iron atoms linked to each other with oxygen
bridges (1^). Among other biological compounds which
form chelate rings are amino acids and proteins, pterins
and riboflavins, purines, nucleic acids, and ascorbic
acid (83, 8^).
The atoms involved in the formation of chelates
are those capable of donating electrons to metal ions—
principally oxygen, nitrogen, and sulfur. Many chelates
are formed by the displacement of a hydrogen ion from the
electro-negative element. While the stability constants
of two different ligands for the same ion may differ only
slightly at one pH, small changes in the hydrogen ion
concentration can exert marked changes in the avidity with
which the metal ion is bound. In mixtures of ligands, the
distribution of metal ions is in delicate balance; minimal
changes in pH as well as the concentrations of other metal
ions can effect transfer of an ion from one compound to
25
another (83). There are numerous possible ligands
distributed throughout biological systems. Because of the
relative insolubility of transition metal salts at the pH
prevalent under physiological conditions, the action of
chelates as key compounds in the metabolism of trace
minerals has been proposed by several authors. Pirzio-
Biroli (83) has investigated duodenal juice in an attempt
to find an iron chelate but obtained essentially negative
results. Laurell (66) feels that iron moves in body
fluids in ionic or "low molecular weight" form. Nielands
(85) has shown the presence of a strong iron chelate
"ferrichrome" is important in the metabolism of certain
fungi. Metzler, Ikawa, and Snell (86) have suggested that
pyridoxal functions in biological systems as a metal
chelate. Recently, Wacker and Vallee (8*+) have suggested
that metal ions incorporated in chelate rings may bear a
functional relationship to protein synthesis, and to the
transmission of genetic information. These authors found
that trace metals were strongly bound to both RNA and DNA
isolated from several sources.
Synthetic chelates are widely used as carriers of
iron in plant nutrition (87) to combat iron chlorosis, a
common deficiency condition in various soils of the world.
Chelates have been used to alleviate the toxicity
of heavy metal poisoning. A notable example in this field
26
was the early use of BAL (2,3 dimercaptopropanol) to
control the effects of arsenical war gas poisoning (88).
In addition, BAL has been successfully used for mercury
and bismuth poisoning. Kety (89) attempted to treat lead
poisoning with sodium citrate, since it was known that
lead formed a fairly strong complex with citrate. While
citrate has been used in an attempt to accelerate the
excretion of other metal ions (90), results have been dis
appointing since it is rapidly metabolized and does not
bind most ions with sufficient avidity. The polyamino
acids in general are extremely effective chelating agents
(91). Foreman and co-workers (92) have demonstrated a
tenfold increase in urinary iron excretion subsequent to
the intraperitoneal injection of calcium ethylenediamine-
tetraacetate (EDTA) in rats. This represented the first
removal of iron from the mammalian organism by artificial
means. The most successful compounds in clinical use for
promoting urinary excretion of undesirable elements are
salts of EDTA injected intravenously.
While the therapeutic use of chelates in metal
poisoning is proving successful, these chelates are no
more efficient than inorganic iron salts in effecting
absorption of iron through the intestinal wall. Larson
et al. (93) have shown that the addition of a 15*1 molar
ratio of EDTA, disodium salt to iron to the food of rats
27
decreases the net absorption of daily iron and increased
the excretion of iron in the urine, normally almost nil.
However, Mills (9^j 95) has noted that organic complexes
of copper present in herbage are better absorbed in the
rat than is the non-chelated ion.
From these conflicting data, it is evident that
the efficacy of a particular chelate is dependent on the
chemical nature of the complex as well as the organism
concerned.
The dependence of the stability constant of a
chelate on hydrogen ion concentration has been previously
noted. An additional requirement is a reasonable concen
tration of the ligand with respect to the metal ion and a
concomitant dearth of competing coordinating groups. If
satisfactory conditions are met and the pH is slowly
adjusted from an acidic to a basic condition, a chelate
will form. Indeed, this is one criterion for the demon
stration of the existence of a chelate (96). If the metal
ion forms an insoluble hydroxide, the rapid addition of
base, even in the presence of excess ligand, can cause a
precipitate. The presence of a sufficient amount of
ligand may then cause slow solubilization again.
In general, the metal chelates, which have been
used in most of the reported work, have been fed as
complexes containing the stoichiometrically correct ratio
of ion to ligand, i.e., without an excess of the coordina
tion agent. When this compound reaches the stomach, it is
subject to considerable dilution and to a hydrogen ion
concentration (pH 1 to 2) (97) which may be sufficient to
cause dissociation. As the fluid passes into the duo
denum, reformation of the complex may be blocked by the
presence of competing ions and by other ligands, which may,
as in the case of hydroxide and phosphate, effectively
precipitate the metal. Under circumstances such as
described above, there could be no conclusions regarding
the relative absorption efficacy of chelated vs. inorganic
ion.
In addition to the ability of organic ligands to
solubilize metal ions under basic conditions, the effective
charge in the vicinity of the ion is altered by the dis
placement of one or more hydrogen ions from the organic
molecule. For example, ferric citrate (98, 99) can have
a net charge of +3, +1, 0, -1, or -2, depending on the pH.
The stability constants of these various forms are vastly
different, and vary from less than 1 to 10^5.
Membranes as Ion Exchangers
It is well known that there is a potential
gradient across all membranes of living organisms (100).
While their exact structure is not known, these semi-
permeable barriers are believed to consist of a
29
lipid-protein complex (101). Sjostrand and Rhodin (102),
from electron microscope studies, concluded that the
o
plasma membrane was in the order of 200 A thick, consist-
o
ing of a lipid material about 50 A thick, sandwiched
o
between two protein layers about 70 A thick. Danielli
(103) bias envisioned a structure which is basically lipid,
but in addition contains a limited number of polar pores.
These polar pores are primarily protein in nature, and
must therefore carry a distribution of charges over their
surface.
Thus, biological membranes may be considered to be
of an ion exchange character (10*+). Ions are prevented by
electric repulsion from approaching the places on the
membrane where fixed charges of the same sign are located.
Such ions are virtually excluded from sufficiently narrow
pores. The greater the charge on the ion, the more
restricted is the movement. However, with increasing con
centration of electrolytes in the solution outside a
membrane, an increasing quantity of anions and cations
enter the pores and the ionic selectivity of the barrier
is effectively decreased (104-).
The selectivity of membranes to the passage of
certain electrolytes has been shown. Thus, the poor
absorption of sulfate ion by the intestinal tract is well
known (105)• Overton (106) found that the addition of
non-polar groups to a molecule Increased the permeability
through cell membranes, both animal and vegetable, while
the substitution of polar groups decreased the rate of
passage.
It is the purpose of the work reported in this
dissertation to propose methods whereby the known charac
teristics of membranes and chelates can influence ion flux
in living organisms.
CHAPTER II
OUTLINE OF PROBLEM AND METHOD OF ATTACK
While the literature is replete with clinical
studies on iron metabolism (2^,107), little is known of
the mechanisms by which iron is moved from the lumen of
the gut into the tissues and then to the various cellular
compartments and binding entities within the body. As can
be ascertained from the general lack of agreement among
the investigators in this field, the theories of this
mechanism are varied, ranging from pinocytosis (108) to
ionic diffusion (6*+). This dissertation presents a
possible mechanism by means of which the iron movement in
the body may be regulated. Experimental results will be
presented to support the conclusions drawn from this
hypothesis.
If cellular membranes exhibit some of the prop
erties of ion-exchange materials (10*+), it can be readily
concluded that uncharged or weakly charged solutes could
effect penetration of the membrane more easily than
highly charged particles. Furthermore, it is believed
(109, 110) that biological membranes contain pores through
which polar molecules are capable of diffusing. The
31
32
larger the diffusing solute, the slower may be the
expected passage through the pores (103).
It has been previously noted that the solubility
of ferric ion at physiological pH's is so low that it is
difficult to envision this form of the metal as that in
volved in iron flux. While the solubility of the ferrous
ion might be adequate for biological activities, it is
easily oxidized. It would be necessary to postulate both
oxidizing and reducing conditions at essentially the same
place and time.
It is therefore proposed that the metabolism of
iron occurs primarily through the mediation of small
soluble chelate molecules. It is suggested that the pas
sage of iron through the gut, its transport through the
blood and interstitial fluid, its storage in body tissues,
and its use in the erythropoetic and other centers are
functions of the following factors:
A. The stability constant, or measure of affinity of
the metal ion for each site where the ion is
located, under the pH and redox conditions
existing there, e.g., transferrin, hemoglobin,
ferritin, etc.
B. The total number of each kind of site available.
C. Competition by other ions for the sites.
Experiments both in vivo and in vitro have been
undertaken to study this problem; four principal aspects
of iron metabolism have been examined.
A. Chelates as factors in the movement of iron from
blood to storage tissue.
B. The formation of ferritin from apoferritin.
C. Studies on the properties of sugar-iron chelates.
D. The absorption of iron from the mammalian gut.
The techniques utilized and the results obtained
are presented in the following chapters.
CHAPTER III
EXPERIMENTAL AND RESULTS
Chelates as Factors in the Movement of
Iron from Blood to Storage Tissue
The initial work of Saltman et al. (70) had
involved the measurement of iron uptake by rat liver
slices from a Krebs-Ringer phosphate buffer solution.
The iron in this solution was added as ferric ammonium
citrate. This compound is a charged molecule whose struc
ture has not been elucidated. It is probable that the
iron is bound to the citric acid as a negative complex
(111) with the positively charged ammonium groups provid
ing electrical neutrality. Saltman et al. (70) showed
that under their experimental conditions, the incorporation
of tagged iron into liver slices as measured by radioactive
Fe?9
was directly proportional to the total iron uptake.
If the rate of uptake of iron by liver slices is a
function of the size and charge of the iron chelate
presented, the kinetics of accumulation would be consider
ably different if the iron were offered as the small
citrate chelate molecule rather than as the non-diffusable
transferrin molecule. The great affinity of transferrin
for iron would sequester a considerable amount of the
3^
35
metal in the medium surrounding the slice. It appears
that ferric citrate diffuses into the slice where it gives
up its iron to a chelate site of greater affinity.
Some authors (63) have postulated that the entire
transferrin molecule passes through the membrane into the
cell. It has been proposed that the transferrin might
require contact with specific cellular receptor groups
(59)* If these hypotheses are correct, some kinetic
effect might be observed if the iron were presented to the
slice in the transferrin of heterologous sera.
The following experiment was carried out to test
the effect of transferrin and of heterologous sera.
Seventy-five ml of fresh rabbit serum and a like amount of
chicken serum were obtained from a commercial source.'*'
The sera were centrifuged to remove a small amount of
cloudy precipitate. Krebs-Ringer phosphate buffer solution
was prepared (112) and the pH of sera and buffer were
adjusted to 7«1 +.
Ferric chloride containing Fe^9 of high specific
activity (18,000 millicuries per gm) was obtained in 1 N
HC1 solution from Oak Ridge National Laboratory. Ferric
ammonium citrate, containing 395 Jigm of iron per ml, was
prepared by adding stoichiometric amounts of citric acid
■^Mission Animal Supply, Los Angeles, California
36
p
to hydrated ferric chloride in distilled, deionized
water. The pH of this carrier iron solution was adjusted
to 7»*+ with 1 N I ' I H i +0H and it was then diluted to the
desired concentration. Radioactive iron, together with the
equivalent amount of citric acid, were added to an aliquot
of carrier iron solution, the pH again adjusted with
NHl^OH to 7.*+ and the solution further diluted to a final
concentration of 212 jigm of iron per ml.
One-half ml aliquots of the radioactive iron
solution were then added to 75 ml of each serum as well as
the Krebs-Ringer solution. The final concentration of
added iron was thus 1.*+ )xgm per ml. Three ml aliquots of
each of these solutions were pipetted into 20 ml beakers
and placed in a Dubnoff metabolic shaking incubator. The
temperature was held at 37° C. The rate of shaking was 80
cycles per minute. Three ml of each solution were counted
using a scintillation well counter equipped with a pulse
height analyzer peaked at the 1.098 Mev gamma emission
from the Fe^. Specific activities were as followsj
Rabbit serum .... *+5,100 cpm per 3 ml
Chicken serum .... *+8,700 cpm per 3 ml
Krebs-Ringer
solution *+6,700 cpm per 3 ml
2
All chemicals used in this and other experiments
were of A.C.S. reagent quality.
37
A 5 lb rabbit of unknown breed and a 3 lb Rhode
Island Red rooster were sacrificed by decapitation. Their
livers were perfused in situ with 100 ml of 0.9$ NaCl
solution via the portal vein. The livers were then excised
and rapidly sliced using a conventional Stadie-Riggs
microtome. The slices were blotted dry with filter paper
and weighed. Slices weighing between 60 and 100 mg wet
weight were incubated in the prepared solutions for pre
determined periods of time. Rabbit liver and chicken
liver slices were incubated in both sera. Only the rabbit
liver was incubated in the Krebs-Ringer solution. At the
conclusion of predetermined intervals, the slices were
removed, washed with three 10 ml portions of saline, and
placed in tubes for counting.
In order to ascertain the site of radioactive iron
incorporation in the proteins of the sera, portions of the
two sera were subjected to paper ionophoresis using a
barbiturate-acetate buffer (113)» pH 8.6 and ionic strength
0.1. Strips were developed at 1 ma per cm of width for a
period of 18 hours. The resolved proteins were stained
with bromphenol blue (ll^f) and the radioactivity scanned
using a gas flow Geiger tube linked to a ratemeter and
rectilinear recorder.
The experimental data are presented in Figures 3
and It is evident that there is no heterologous effect.
38
Oiooo
gj 600
Q.
0 4 0 0
rabbit liver in •
Krebs Ringer phosphate buffer
(ordinate is 1/5 scale)
□
☆ chicken liver, chicken serum
O chicken liver, rabbit serum
□ rabbit liver, chicken serum
★ rabbit liver, rabbit serum
30 4 0 50 60 70
Incubation time, minutes
80 90
Figure 3.— The effect of heterologous
transferrin and ferric ammonium citrate on the uptake
of iron by liver slices of two species.
39
distribution oi
radioactivity electrophoretic
pattern
vXvXvX
_
frfrOWQQfrfrWCWOQWgQqQaggOOOOOOOOO^
j 5j ea! ! M! LWSX0H0M9 S j i
Figure — The site of radioactive iron
incorporation in serum.
I
^•0
Transferrin from either species is equally effective for
presenting the iron to both species of liver. The Krebs-
Ringer solution is much more effective than either serum
in permitting uptake by the liver slices. These results
are compatible with the postulate that transferrin is not
a mandatory intermediate for iron uptake by liver. Since
the ferric ammonium citrate is many times more effective
than transferrin, one may conclude that the strong affinity
for iron possessed by the protein molecule prevents its
release and accumulation by the iron binding molecules
present in the cell. The results of the incubation of the
slices in Krebs-Ringer are identical with those obtained
by Saltman et al. (70).
Figure 4- illustrates that all the radioactivity
present in the serum is associated with the (3-1 globulin,
transferrin.
To verify the hypothesis that small chelate
molecules play an important role in iron transport, the
following experiment was carried out: 210 jig iron as
ferric ammonium citrate, containing 2.6 x 10^ counts per
min Fe^, was added to 110 ml of rabbit serum. This
amount of iron was calculated to saturate the binding
capacity of the transferrin. The serum was exhaustively
dialyzed against running distilled water to remove excess
citrate and other small molecules and ions. It was
necessary to centrifuge the serum before use to remove
precipitated euglobulins. The final specific activity of
the serum was 23,500 counts per min per ml. As in previous
experiments (see Figure *f), paper ionophoresis, followed
by radioactivity scanning, revealed that all the Fe^9 was
associated with the $-1 globulin. A rabbit weighing 3 to
5 pounds was sacrificed by cervical dislocation, and the
liver perfused with cold 0.9$ saline. The organ was then
excised and chilled in iced saline. Rabbit liver slices
were prepared with a Stadie-Riggs microtome, and samples
of tissue weighing 80 to 100 mg were added to 3*0 ml of the
dialyzed serum containing serial dilutions of citrate,
adjusted to pH 7.*+. Citrate was used as the chelating
agent for study in these experiments both for its favorable
stability constant and its occurrence in the in vivo system.
The slices were incubated with shaking for 1 hour at
37° C, removed, washed three times with saline, and radio
activity was measured as previously described. Incubation
for 1 hour or less permits the measurement of initial
rates of uptake of Fe^9. These measurements do not
reflect either the final concentration or capacity of the
tissue. Previous experiments have shown that the slices
maintain their cellular integrity under these conditions
(115). Figure 5 presents the data from a typical experi
ment. It is evident that an increased concentration of
Uptake, counts per minute per 1 0 0 mg.
k2
1600
8 0 0
005.01 .05
Citrate concentration(M)
Figure 5 .— The uptake of Fe^9 from transferrin
by rat liver slices as a function of citrate concentra
tion in the medium.
>+3
citrate enhances the initial rate of uptake of iron hy the
slices. When no citrate was added, there was a measurable
rate of accumulation probably due to the presence of
endogenous chelate molecules diffusing out of the tissue.
Several other chelates manifest similar activities,
including ethylenediamine tetraacetate (EDTA), which is
not a metabolic substrate. Table 1 demonstrates the
effect of EDTA compared with citrate on the uptake of
C jQ
transferrin-bound Fe" by liver slices.
By working with liver slices it is difficult to
determine if the iron dissociates from the transferrin, is
chelated, and then enters the cell; or if the metal
protein must impinge or complex directly in or on the cell.
Preliminary experiments indicated that transferrin-Fe^9
could not dialyze through Visking tubing. We, therefore,
perfused a freshly extirpated rat liver with 0 ,9 % saline,
minced the tissue, and then homogenized with 9 vol.
saline in a Potter-Elvehjem apparatus. One milliliter
aliquots of homogenate were tied into quarter-inch
diameter Visking sacs, and the exteriors washed carefully
several times with saline. Fe^-labeled serum was pre
pared and dialyzed as described above, and contained a
total activity of 15,090 counts per min per ml. One
series of sacs was incubated in the serum alone, the other
series in serum containing 0.1 M citrate at the same pH.
Filmed as received
without page(s)
UNIVERSITY MICROFILMS.
^5
TABLE 1*--The ability of EDTA and citrate to enhance the
uptake of Fe59 from transferrin by rabbit liver slices
Condition
Concentration
M
Uptake
Counts per min
per 100 mg liver
No chelate added
*+37
Citrate 0.01 l k -72
EDTA 0.01 l b 0 5
The specific activity of the serum was 2*+,890
counts per minute. Incubation was for 60 minutes.
h6
At the indicated time intervals, the sacs were removed and
washed with saline, and radioactivity was determined.
Figure 6 presents the data from a typical experiment. It
seems clear that citrate enhances the uptake of iron by
the homogenates via a mechanism which does not involve the
direct passage of the transferrin through the sac. The
most probable mechanism is the formation of a chelate which
can then pass through the semipermeable membrane. This
experiment indicates that the mechanism of chelation could
well play a similar role in intact cells.
That chelates can remove iron from transferrin was
directly demonstrated by the following experiment. One ml
aliquots of radioactive rabbit serum, prepared as described
above, containing 8300 counts per min per ml were dialyzed
against 5.0 ml of several different chelating agents. All
solutions were adjusted to pH 7«*+ before the dialysis,
and their pH was checked at the conclusion of the 2 b hour
period. A 2.0 ml aliquot of the external solution was
analyzed for radioactivity. Table 2 presents the data from
such an experiment. These results confirm that iron can
be removed from the transferrin in a molecular form which
is freely permeable through the Visking tubing. It should
be emphasized that the amount of iron removed is a direct
function of both the concentration and of the affinity of
the ligand for iron. The value given for the dissociation
>+7
CT>
!300
Q_
100
u>
Q .
ID
30 60 90
Minutes
Figure 6.— The influence of 0.01 M citrate on
the rate of uptake of Fe59 from transferrin by rat
liver homogenates in Visking sacs.
1+8
TABLE 2.— The ability of various chelating agents to remove
Fe59 from rabbit transferrin by dialysis
Amount Fe^9 Removed
Compound Concentration (Counts per min
W per 5.0 ml)
Water
5
EDTA 0.01 l*+32
o-Phenanthroline 0.01
551
Citrate 0.01 528
8-Hydroxyquino1ine 0.000*+ 268
Lactate 0.01 167
Lysine 0.01
85
Cysteine 0.01
113
k-9
constant of transferrin-iron of 10"? (5*+) seems highly
unlikely since the citrate-iron dissociation constant has
been estimated as 1 0 ~ ^ 5 (98).
It was desirable to ascertain which of the common
body metabolites were most effective in the removal of
iron from transferrin. Accordingly, 0.1 ml of radioactive
iron (as ferric ammonium citrate) was added to 100 ml of
fresh bovine plasma. The treated plasma was dialyzed for
1 day against running distilled water to remove unbound
Fe^9. Aliquots of 1.0 ml were placed in separate Visking
sacs. The outsides of the sacs were carefully washed and
each was counted in the scintillation well counter to
determine initial radioactivity. Solutions of various
metabolites were made up at a concentration of 0.01 M,
except in the case of tyrosine and cystine, where
saturated solutions were used. Each solution was adjusted
to pH 7»*+. The plasma in the sacs was then dialyzed
against 5.0 ml of each solution in large tubes. The necks
of the tubes were covered with aluminum foil, and the
tubes were placed in the dark, in a room held at 5° C.
The tubes were agitated at intervals for a period of 1
week. At the end of that time aliquots of the solutions
were withdrawn and counted. Table 3 presents the results
expressed as a percentage of the total original radio
activity in the plasma which diffused through the membrane.
50
TABLE 3«— The ability of various metabolites to
remove Fe59 from bovine transferrin by dialysis
Glutamate
1 .5%
Aspartate 5.2
Glycine
6.7
Citrate 65.1
Isocitrate
^8.7
OC -Ketoglutarate 15.6
Succinate 2.6
Oxaloacetate 1*8.1
Pyruvate 29.2
Lactate 12.2
Phosphate 36.2
Bicarbonate 3»0$
Results are expressed as a percentage
of the original transferrin iron which had dif
fused through dialysis tubing in the presence
of the metabolite in 1 week. For tyrosine and
cystine, saturated solutions were used. All
other solutions were 0.01 M.
51
Other amino acids, including tryptophane,
arginine, cystine, methionine, tyrosine, lysine, leucine,
phenylalanine, and proline, were ineffective— removing
less than 2% of the original activity. Kreb's cycle
intermediates, in general, are good chelating agents of
transition elements (116) and could serve as carriers
between non-diffusible proteins held within various body
compartments. Brown (117) has noted that the presence of
pyruvate and OC-ketoglutarate inhibits the in vitro
incorporation of iron into chicken erythrocytes. He
postulates that there is competition between two avid
sites for the iron.
It is well to note that a high stability constant
does not necessarily imply that a ligand will possess the
qualities required for efficient ion transport. A
rapidly diffusible chelate of low avidity which is able to
give up the metal to an intercellular site may prove more
effective than a poorly diffusible chelate of high
stability constant. The hypothesis that iron flux is
controlled by diffusion and by equilibrium binding of the
ion by proteins, nucleic acids, and small chelate mole
cules is in complete agreement with the prior work of
Saltman et al. (70). These workers demonstrated, with
liver slices, that iron transport in and out of cells was
not directly dependent on metabolic energy.
I
52
Radio iron added to transferrin as the citrate
complex is rapidly taken up by the protein and cannot be
removed by dialysis against distilled water. Preliminary
experiments indicated that apoferritin was also capable of
binding a large proportion of the iron from ferric citrate.
It is apparent from these observations that the binding
affinity of transferrin or ferritin for iron is sufficient
to effectively compete with citrate for iron. Based on
the experiments described above, it was postulated that
net iron uptake was controlled by three prinicpal factors:
(a) the total number of each kind of binding site (i.e.,
transferrin, ferritin, small chelate molecules, etc.);
(b) the stability constant of each site for iron, under
the pH and redox conditions existing there; and (c)
competition by other ions for the sites. To test this
hypothesis, the following experiments were performed:
Rabbits, *+ to 5 pounds, were bled via cardiac puncture
three times at weekly intervals. On each occasion 25 ml
of blood was removed. The repeated phlebotomies resulted
in depleted iron stores in the livers. A second group of
animals were injected on four occasions with a total of
57.5 mg Fe+++ as the iron dextran complex, Imferon^, to
supplement liver iron. Sixty hours before use some of the
^Lakeside Laboratories, Milwaukee, Wisconsin
animals from the first group were injected intramuscularly
with 25 mg of Imferon to provide an intermediate iron
level in the liver. These treatments yielded livers con
taining three levels of iron concentration. To prepare
serum containing two different concentrations of iron,
200 ml of rabbit serum was dialyzed against 1 liter of
distilled water for 2 b hours. This dialyzate was saved.
Insoluble euglobulins were removed by centrifugation, and
the clear serum was dialyzed against 5 liters of acetate
buffer, pH ^.2, to remove the remaining small molecules
and some of the bound iron. The serum was then made 0.1 M
in ascorbic acid to reduce the protein-bound Fe+++, and
dialyzed 2 b hours against running water. One aliquot of
this serum was treated with ferric ammonium citrate con-
taining Fe77 to provide a final concentration of 1.1 jag
iron per ml. The specific activity was 32,100 counts per
min per ml. This serum was classified as “depleted.1 1
Another aliquot of the acetate-dialyzed serum was treated
with ferric ammonium citrate containing Fe^9 to bring its
total iron concentration to 2 . b jig per ml, specific
activity 31j*+00 counts per min per ml. This serum was
classified as "supplemented." Both radioactive sera were
dialyzed against running distilled water for 2 b hours
prior to use. The small molecules and ions removed in the
original dialyzate were concentrated by evaporation and
returned to the serum at their original concentration to
assure the presence of the natural chelates and ions. In
the absence of these natural chelates, iron transport is
greatly inhibited. The pH of the sera was adjusted to 7A.
The treated rabbits were sacrificed, their livers removed,
perfused in saline, and sliced with a Stadie-Riggs micro
tome. The slices, 60 to 100 mg, were incubated with 3.0
ml aliquots of the two prepared sera for 15 and 60 minute
periods. The slices were removed, washed with saline, and
the radioactivity determined. Total iron in the slices
and serum was determined by a modification of the method
of Peters et al. (118). The modification consisted of
using an increased ratio of isopropyl to isoamyl alcohol,
from 3:1 to 6:1 for the solution of the ferrous iron
complexing agent, bathophenanthroline. More recently
Peters and co-workers (119) have made a similar modifica
tion. The liver slices were treated with HC1 and thio-
glycolate as used for the determination of serum iron.
Following this step, 0.25 gm of acid washed ground glass
was added to each slice, and it was thoroughly macerated
with a glass rod. Subsequent steps were as described
above (118). Table presents the data from a typical
experiment.
55
TABLE — Uptake of radioiron by liver slices incubated
in serum
Uptake
(ping Fe/100 mg liver)
Incub a*
tion
time
(min)
Supplemented
serum
2.1+ p g m Fe/ml
31,^00 counts/
min/ml
Depleted
serum
1.1 jugm Fe/ml
32,100 counts/
min/ml
Supplemented 15
liver
(7.*+ pgm Fe/ 60
100 mg)
^7
58
13
15
Intermediate 15
liver
(5*8 jagm Fe/ 60
100 mg)
65
88
18
21
Depleted 15
liver
(2.*+ jagm Fe/ 60
100 mg)
72
100
22
32
It is clear that the rate of iron movement from
the serum to the liver slices is directly influenced by
the relative availability of binding sites participating
in iron metabolism.
The Binding of Iron by Ferritin
Mazur and co-workers (78) have established that
iron can be transported between ferritin and transferrin
in the presence of agents able to reduce the ferric iron
bound to the ferritin by sulfur bridges. In an in vitro
system where ferritin and transferrin are prevented from
contact by a cellulose membrane, the presence of iron in
the ferrous state permits significant movement of the
metal across the barrier. The solubility of Fe++ at
physiological pH's is approximately 1.6 M (120). However,
it is mandatory that chelates be present to maintain
Fe+++ in a soluble and dialyzable form under similar
circumstances.
Mazur et al. (78) incubated rat liver slices with
a ferritin solution in the absence of added chelates,
under aerobic and anaerobic conditions. The solution was
later dialyzed against transferrin. They found a marked
increase in the movement of iron from ferritin to trans
ferrin under anaerobiosis which they attribute to the
reduction of ferritin iron, While the efficacy of
reducing agents in facilitating the removal of bound iron
from either ferritin or transferrin is not questioned, it
is well to point out that under conditions where oxygen is
limited, the concentration of respiratory intermediates,
especially organic acids, would increase. It has been
previously noted that these Kreb's cycle intermediates are
especially efficient for binding ferric iron. It seems
likely that the removal of iron from ferritin is not
solely a function of the reduction of the iron, but may be
caused by an increased pool of competing ligands. Anoxia
will produce conditions where both the redox environment
and the concentration of sequestering agents are such that
iron would leave body stores. Then, as Mazur (79) has
proposed, the increased iron available to the erythro-
poetic centers would stimulate heme synthesis. The
restoration of normal hemoglobin levels and the restora
tion of normal oxygen balance would regulate directly the
iron flux from the storage centers.
It was of interest to determine if reducing agents
were mandatory, of if chelating agents alone were capable
of removing iron from ferritin. Accordingly, the follow
ing experiment was carried out. A 5 pound rabbit was
injected with 25 mg of iron in the form of Imferon to
which Fe-^ had been added. At the end of b weeks, the
animal was sacrificed and the liver and spleen removed.
Crystalline ferritin was prepared as previously described
(11). The 5$ cadmium sulfate solution was made 0.01 M in
NaCl2 to increase the yield of crystals obtained (Dr. R. A.
Fineberg, personal communication).
Fifty ml of fresh rabbit serum was dialyzed
against 4- liters of 0.9$ saline for 2 b hours. Ten ml
aliquots of the prepared serum were then made 0.01 M in
each of the compounds listed in Table 5> and the pH's
adjusted to 7*0. The prepared ferritin was dissolved in
15 ml of 0.9$ saline and the pH adjusted to 7.0. Approxi
mately 1 ml aliquots of the ferritin solution, assaying
1^-00 counts per ml per min, were tied into Visking sacs
which were then thoroughly rinsed with saline. The sacs
were dialyzed against 5 ml the prepared sera in test
tubes and kept in a refrigerator; periodically the tubes
were withdrawn and shaken. At intervals of 2 and 6 days,
aliquots of the serum were removed, counted, and returned
to the tubes. Table 5 presents the results reported as
percentage of radioactivity originally present in the
ferritin which moved across the membrane. It is evident
from these results that the presence of reducing agents is
not required for the removal of iron from ferritin.
Bielig and Bayer (76), as well as Loewus and
Fineberg (77)» have both reported the preparation of
ferritin by the combination of ferrous iron and
59
TABLE 5.— Per cent of Fe-^ activity transferred
from ferritin to transferrin across a dialysis
membrane in the presence of various solutes at
0.01 M
After
2 days
After
6 days
Control 1 .0 %
0 .7 %
Dehydroascorbate 8.1 18.3
Ascorbate
3.7
7.8
Citrate 3.3
7.8
Thiosulfate 0 .6 % 0 .6 %
crystalline apoferritin. Bielig and Bayer noted that
red-brown crystals, characteristic in appearance to fer
ritin, were formed. Their measurement of the magnetic
susceptibility of a solution of this material (76) in
dicated that the iron present contained three unpaired
electrons. The same uncommon magnetic susceptibility was
found in natural ferritin by Michaelis et al. (1*+).
Recently, however, Shoden and Sturgeon (121) have demon
strated that ferric hydroxide gels can exhibit magnetic
susceptibilities identical with those found in ferritin,
depending on the rate of formation of the FeCOH)^ and con
centration of the ferric ion from which the precipitate is
made. It should also be recognized that the typical
crystal formation described by Beilig and Bayer (76) is a
property of the apoferritin protein (17, 20) and is not
influenced by the presence of ferric hydroxide other than
in its color. Loewus and Fineberg (77) used as their
criteria for the formation of ferritin an increase in
optical density of the solution at *+00 mja. It has been
previously noted that the spectrum of ferritin in this
region differs but slightly from that of hydrated ferric
ion. These authors incubated a solution of crystalline
apoferritin with ferric ammonium citrate in the presence
of liver slices, ascorbic acid or a variety of other
substances. Only with liver slices and with ascorbate,
61
however, could they detect significant uptake of the iron
by apoferritin. A negligible change was observed when the
apoferritin solution was incubated with ferric ammonium
citrate alone. They, therefore, postulated that ferrous
iron was necessary for ferritin formation and that ascorbic
acid was participating in the reaction primarily as a
reducing agent.
Insufficient attention has been given to the role
of ferric citrate as a powerful chelating agent. It should
be remembered that the addition of equimolar citrate to an
ionic iron solution results in the sequestering of
essentially all ionic iron at physiological pH's. Although
the complex is soluble and ionized, the iron therein is
tightly bound and is not readily available for reaction.
Ferric citrate is, however, in equilibrium with an
extremely minute amount of ferric ion. The presence of a
reducing agent, such as ascorbate, will shift the equilib
rium from the ferric citrate chelate to ferrous ion or to
a ferrous chelate of citric or ascorbic acid. In the
presence of light, even citric acid itself will reduce
ferric iron, which was demonstrated as follows:
A small quantity of 0(,0t/ dipyridyl was added to
a solution containing 395 ugm of ferric citrate per ml.
The ferrous dipyridyl complex is red; no interaction
between this ligand and ferric ion takes place. The
62
solution was divided into two equal volumes, one of which
was placed in the dark and one in daylight. Within 3
hours the solution in the light was pink, the intensity of
the red complex increasing with time. At the end of 1
week there was no observable change in the solution kept
in the dark, as determined by comparison with a fresh
preparation.
The procedure for crystallizing apoferritin
involves precipitation of the protein from solution by con'
centrated aqueous cadmium sulfate. These protein crystals
contain 6 to 7$ dry weight of cadmium (12). After washing
with concentrated KC1, Granick (12) found the cadmium
content could be reduced to 1.7$. It was not possible to
displace this small amount. It was thus possible that
cadmium may have played some role in the formation of fer
ritin as reported (76, 77).
It was desirable to test the hypothesis of Beilig
and Bayer (76) and Loewus and Fineberg (77) that reduced
iron was essential for the formation of ferritin from apo
ferritin. Crystalline apoferritin was prepared from
normal rabbit livers by the method of Granick (20). A
small amount was dissolved in 2% ammonium sulfate solution.
Ferric ammonium citrate was added so that the iron con
centration was 50 Jigm per ml of apoferritin solution. The
solution was divided and ascorbic acid was added to one
63
part so that there were 3 moles of ascorbate per gram atom
of iron present. Both solutions were adjusted to pH 7.0,
and were incubated in the dark for a period of 1 hour at
37°. After incubation the protein was precipitated and
washed with 50% (NHl^SOl^. solution and redissolved in 2%
ammonium sulfate. Absorbancy measurements were made at
^00 mji using a Beckman D. U. spectrophotometer, with the
original apoferritin solution without ferric ammonium
citrate as a blank. Under these conditions, the precipi
tate from the apoferritin plus ferric ammonium citrate
solution containing the ascorbate had an optical density
of 0.7j while that of control was 0.5* It is clear that
there was a direct incorporation or binding of ferric iron
by apoferritin.
To further study the phenomenon of ferritin
formation, a large amount of the apoprotein was prepared
from horse spleen. Diketogulonic acid, calcium salt, was
prepared by the method of Curtin and King (122). The
crystalline apoferritin from spleen was dissolved in 0.02 M
bicarbonate buffer, pH 7.^-. The effect of several com
pounds indicated in Table 6 on the binding of iron to apo
ferritin was carried out with 5 ml of the protein solution,
containing ferric ammonium citrate at a final concentration
of 79 jigm of iron per ml. The solutions were incubated in
a Dubnoff metabolic incubator. The protein was
6*+
TABLE 6.— The effect of incubation time, reductants, and
sequestering agents on the formation of ferritin from
apoferritin and ferric ammonium citrate
Added solute
Ratio of M
solute/gm
atom Fe*++
Incubation
time
(hrs)
Absorbancy
Control
---
1 0.261
Control
---
3
0.302
Ascorbate 1.5
1 0.2^9
Diketogulonate 1.5
1 0.221
Histidine 3.0 1
0.253
65
precipitated, washed, and redissolved as previously-
described. The absorbancies at ^-00 mji are tabulated in
Table 6. It is particularly noteworthy that diketogulonic
acid, the hydrolysed and oxidized form of ascorbic acid,
is equally as effective as ascorbic acid itself.
It is evident that ferritin, as measured by this
assay, can be formed from apoferritin without the inter
vention of ascorbate and without the necessity of the iron
being present as the ferrous ion. The formation of clear
red brown solutions was also observed upon the addition
of ferrous sulfate to crystalline beef serum albumin and
egg albumin, as well as apoferritin. It appears that this
assay is not specific for ferritin formation, but repre
sents a general phenomenon of stabilization of an other
wise insoluble metal hydroxide as a protein complex.
The possibility of competition between the protein
and small chelate molecule for the iron required investi
gation. It was felt that this could be studied with the
use of chelates of iron which were less stable than iron
citrate at pH 7.0-7.5» Traube and Kuhbier (123) had
noted that in the presence of the polyhydric alcohols
mannitol and sorbitol, sodium hydroxide did not precipi
tate trivalent iron. While 0.01 M citrate at pH 7*0 will
remove a portion of the iron bound to transferrin which is
retained in a Visking sac, an experiment showed a like
concentration of mannitol to be ineffective. Thus, the
iron mannitol complex appeared to have a stability con
stant less than the complexes of either citrate or ferritin.
While investigating the mannitol iron complex, it became
evident that many polyols, as well as many sugars, form
complexes of iron under the proper circumstances. Carbo
hydrates had been shown to produce remarkable, but unknown
effects in enhancing the absorption of inorganic ions,
notably iron and calcium, from the mammalian gut (*+5, *+7}
*+8). It appeared that these observations, together with
the relatively unrecognized chelating ability of the
sugars, warranted further investigation.
Some Properties of Sugar-Iron Chelates
Traube, Kuhbier, and Harting (12*0 investigated
complexes of sugars and metal ions. They reported this
preparation and isolation of barium-iron "salts" of
glucose, mannose, maltose, lactose, galactose, and
arabinose. They formulated the typical complex of glucose
They gave no evidence for such a structure other
than its elemental analysis. These workers prepared
sodium salts of iron complexes with mannitol and sorbitol
by mixing barium chloride and ferric chloride with the
sugar solution, followed by the addition of NaOH. Pre
cipitates were obtained which proved sparingly soluble in
water. Traube and co-workers were not able to prepare a
uniform product with fructose (12^-).
Lieser and Ebert (125) have reported complexes of
sugars with copper. Their analysis of the purified
complex showed 2 moles of copper present per mole of
hexose. Recently, Bourne et al. (126) have determined the
chelating power of polyhydroxy compounds, among them
glucose, fructose, mannitol, and sorbitol. These investi
gators added an excess of metal salt to a basic (pH 12)
solution of the polyol. The mixture was centrifuged and
the precipitated excess of metal hydroxide analyzed
quantitatively. The amount of metal remaining in solution
was found by difference. They concluded that polyhydroxy
compounds had the following relative ability to chelate
trivalent irons
D-sorbitol 10.75 gm atoms Fe per mole
ligand
D-mannitol 7*51 1 1 " M 1 1 "
D-fructose 0.80 " " " 1 1 "
D-glucose negligible
68
Compounds of iron and sucrose have been long
known. Saccharated oxide of iron has been used as an
intramuscular hematinic (127, 128). Due to occasional
toxic reactions (129), perhaps resulting from variations
in composition, it has not been widely accepted in
clinical practice. Bersin (130) gives the formula
(Fe(Fe(OH)2)^)i4 . for a compound of iron
with sucrose. The most useful agent for parenteral
therapy at present is the iron-dextran complex, "Imferon.1 1
Complexes of sugars with calcium are better known.
Calcium saccharate and calcium levulose have been uti
lized in the purification of sugar cane juice and fructose
respectively (131).
Chelates are usually formed between metal ions and
organic ligands which can act as acids or bases. The
formation of a chelate is accompanied by the displacement
of a proton from the acid or a decrease in the concentra
tion of the base. In this regard, it is of interest to
note that the existence of saccharate anions Ci2H21°ll~,
and C12H21O11— , has been demonstrated. The dissociation
constant of the sucrose molecule to give the anion and H+
was found to be l*f.6 x 10“^ . It was found that the
alkali metal saccharates were completely dissociated
salts up to a concentration of 0.1 N (132).
69
The preparation of iron-sugar chelates in this
research was carried out as follows: Solutions of either
ferric chloride or ferric nitrate were prepared at con
centrations of 0.1 M or 0.2 M with distilled water. It
was necessary to prepare fresh solutions immediately prior
to use; on long standing hydrolysis of the iron salt
resulted in the formation of a precipitate. This occurred
even though the pH's of the solutions were approximately
2. Where a standard iron preparation was required to
provide comparative data over a period of several days,
precipitation of ferric hydroxide was prevented by the
addition of concentrated HC1 to bring the pH below 1.
Solutions of sugar of the desired molarity were
prepared in distilled water. Since some of the sugars
have appreciable negative heats of solution, the mixture
was permitted to reach room temperature before use. The
sugar and iron solutions were then mixed and the pH
adjusted to the desired value with 1 N or 6 N NaOH. The
color of the iron solution, normally a pale yellow at acid
pH, darkened considerably upon the addition of sugar.
Such a change in spectral properties has been used to
establish the presence of a chelate (96). It was dis
covered that equimolar proportions of iron to monosac
charide were completely ineffectual in preventing
hydroxide precipitation, although the spectral change
was apparent.
Since the ionization constant of sucrose has been
determined, it was thought possible that the formation of
the glucose or fructose chelate resulted from the displace
ment of one or more hydrogen ions from the sugar molecule.
This displacement was demonstrated by the typical titra
tion curves shown in Figure 7. Upon the addition of the
sugar solution to the FeCNO^)^ solution, an immediate
slight increase in acidity was detected. The calculated
increase in hydrogen ion was 0.0012 M (Figure 7).
A titration curve of ferric fructose could not be
obtained since the solution continued to become more acid
as a function of time. This can be interpreted as either a
continued displacement of H+ from the sugar by the iron,
or a reduction of ferric iron to ferrous with the con
comitant formation of an acidic carbohydrate derivative,
such as glycolic and trihydroxy butyric acids.
The necessity for an excess of sugar over metal
ion, to form soluble chelates at basic pH, permitted an
evaluation of the relative sequestering ability of
several monosaccharides. Solutions, 0.1 M in ferric iron,
were prepared containing increasing concentrations of the
chelating agent being tested. These solutions were made
alkaline, as previously described. The minimum concen
tration of sugar adequate to prevent the precipitation of
71
0.014M glucose plus
_____i0.007M .Fe(N03):
3-
0 .0 0 7 M . Fe(NOJ.
X
Q.
ml. of 0.1 N NaOH added
Figure 7.— Titration curves of ferric ion alone
and in the presence of glucose. Fifty ml of solutions
used in each case.
72
a hydroxide was taken as an index of its sequestering
ability. The relative sequestering ability of polyols and
sugars for trivalent iron isi
fructose > sorbitol > glucose=galactose=maltose=
lactose > sucrose > ribose > erythrose
During the course of these experiments it became
apparent that not only was the ratio of iron to sugar of
critical importance for chelation, but also the absolute
concentrations of the two. An investigation into these
relationships resulted in the data presented in Figure 8.
Fructose can sequester iron at concentrations twofold
greater than that of the metal, but only when the absolute
concentration of the sugar is near saturation. The region
around pH 7, where precipitation occurs with low sugar
concentration, is of considerable interest. Evidence has
been presented that a complex is formed immediately upon
the addition of iron to the carbohydrate. As a solution
of such a complex is made basic, it will remain in solution
past the pH at which it will otherwise precipitate. If
there is an insufficient amount of carbohydrate present,
a precipitate will occur as the solution becomes more
alkaline. If additional base is added, at about pH 8, the
precipitate will re-dissolve and remain in solution up to
or past pH 1M-. The soluble complex of fructose with
ferric ion has been designated as "ferritose.'1
Fe***concentration,(M)
73
0.4
unstable chelate
formation
0.2
.06
.04
chelates stable
at all pH’s
0.02
0.01
4.0 20 0.4 0.2
Fructose concenfration/M)
0.02 0.04
Figure 8.— Zone of stable complex formation of
ferric iron with fructose.
7^
The coordination sphere of water molecules about
a metal ion is frequently influenced by the charge to the
extent that a hydrogen ion is displaced. Thus, a series
of "hydroxo complexes1 1 may arise in which the metal is
coordinated to hydroxyl groups. Two or more of these
simple complexes may combine by a process known as
olation, in which the metal ions are linked by hydroxyl
bridges. Under certain conditions, the olation process
may be accompanied by or followed by loss of water or the
simultaneous loss of hydrogen and hydroxyl ions to form an
"oxo compound."
The tendency of iron to form olation complexes
and, with time, oxo polymers, is well known (133)- The
addition of hydroxyl ion hastens the process of olation
(60). Anions which can enter the coordination sphere of
the metal effectively prevent olation. Even after olation
has occurred, certain anions can penetrate into the ferric
hydroxide polymer and effect changes in molecular size
and shape (133)•
It is probable that with the sugar chelates,
normal olation is inhibited by the presence of the sugar.
As the pH increases, concomitant olation occurs and a
compound forms which is essentially neutral. It is
postulated that the rate at which olation proceeds will
control the molecular size and therefore the presence or
75
i
absence of a precipitate. Thus, in a high concentration
of carbohydrate, the hindered rate of olation would retard
the formation of large insoluble polymers. A precipitate
may form in the presence of low concentrations of sugar.
Additional hydroxide may displace water from the coordina
tion sphere of the metal atoms and result in a negatively
charged complex which, due to electrostatic repulsion,
would tend to go back into solution. With a suitable
excess of sugar, and the rapid addition of base, a mole
cule of minimum size would be expected. To test this
concept the following experiment was performed: A solu
tion was prepared which was 0.1 M in FeCl^ and 1.6 M in
fructose. Six normal NaOH was slowly added to 100 ml of
the iron-fructose solution until the pH was 9*0. At no
time was a precipitate in evidence. The alkaline solution
was a dark red-brown color. An identical amount of base
was added rapidly with vigorous stirring to another 100 ml
of the iron-fructose solution. Under these conditions,
the solution resulting was a pale yellow brown. Similar
results are obtained over a range of concentrations. The
difference is due to greater absorption of the rapidly
neutralized complex at k-25 mji, as shown in Figure 9.
Approximately 2 ml aliquots of each basic solution
were placed in Visking sacs which were then tied off,
rinsed, and placed in a test tube containing 10 ml of
A bsorbance
76
•slow"
ferric fructose
dil. 1 / ioo
0.5
difference spectrum
"fasfvs'slow"
ferric fructose, p H u.o
0.025 M Fe***o.4M fructose
0.4
0.3
0 . 2
"fast‘f erric fructose
dil. i/ ioo
300 350 400 450 500 600 700
W avelength, m |jL
Figure 9»— Difference spectrum of "fast" ferric
fructose compared to “slow1 1 ferric fructose. All
solutions were adjusted to pH 11.0. Solutions of ferric
fructose used for spectral determinations against a
water blank were 0.00025 M in Fe+++.
distilled water. Tests for the presence of dialyzable
ferric iron in the water were made by withdrawing 0.5 ml
aliquots and adding 0.5 ml of a 2% potassium ferrocyanide
solution together with 1 drop of concentrated HC1. The
appearance of a deep blue precipitate of Prussian Blue
indicated the presence of triv^lent iron. It was found
that the complex prepared by rapid addition of base and
denoted by the name "fast ferritose" had dialyzed through
the bag in appreciable amounts in 15 minutes. Within 1
hour the color was uniformly distributed throughout the
test tube. The "slow ferritose" had not been able to
penetrate the cellulose sac.
After the slow ferritose had dialyzed for a day,
a small amount of pale blue color was observed upon ad
dition of acid ferrocyanide to a portion of the dialysate.
A test with ferrous sulfate and ferrocyanide confirmed the
suspicion that this color was due to the presence of
ferrous iron. Even after prolonged dialysis, no ferric
iron could be demonstrated outside the sac when slow
ferritose was used.
Complexes of ferrous iron with fructose were
prepared, but only at carbohydrate concentrations approxi
mately double those required for successful sequestering
of ferric ion. The characteristic color of the complexes
of both ferric and ferrous iron reflect the presence of
the hydrated ion. Slow ferric fructose is a deep red-
brown, while ferrous fructose is gray-green. Solutions of
ferric-fructose in excess fructose gradually turn the
identical gray-green color over a period of several weeks.
Tests of such solutions with CCjOC* dipyridyl revealed that
the ferric ion was gradually converted to the ferrous
form. The pink color of the dipyridyl complex developed
at a more rapid rate in acid solution. It is known that
oxidation of sugars by metal ions proceeds more rapidly in
alkaline solution (13*0. The previous observation thus
suggests that the iron is tightly bound to the fructose
under alkaline conditions and is not available for reduc
tion. Under acid conditions the complex dissociates more
readily and the d jt* ' dipyridyl removes ferrous ion from
solution, thus driving the redox reaction.
The color of the ferric fructose complex is
yellow under acid conditions, and red-brown under alkaline
conditions. The olated ferric ion absorption is very
strong in the lower wave lengths— so much so that the
presence of the darker complex, which is apparent
visually, cannot be demonstrated with the spectrophotom
eter. If, however, a difference spectrum is obtained using
a solution of ferric chloride as a blank, the increased
absorption of the complex at b 0 5 m p can be noted, Figure
10. These absorption curves were obtained with the
A bsorbance
79
i.5* io '^M.ferric
‘ructose jcM.
fructose excecc
. ^ pH 3.4
0.5
difference spectrum
c.025M.Fe
0.4
vs.
0.0 1 2 5 M . ferric fructose
0.3-
0.2-
600 700 450 500 400
300
W avelength, m ji
Figure 10.--Difference spectrum of ferric
fructose compared to ferric ion. All solutions were
adjusted to pH 3.h-.
80
Beckman model DK-2 spectrophotometer, using the tungsten
lamp.
Since efficient biological utilization of iron-
carbohydrate complexes may be directly dependent upon both
the size and the net charge, it was important to determine
the isoelectric region of the complex. The presence and
sign of the charge was determined by paper and starch
ionophoresis. Most of the weak acids and their salts
which are used as buffers in electrophoresis investiga
tions are unsuitable for use in chelate charge determina
tions since they may compete for the metal ion and change
the characteristics of the complex. This is especially
important when dealing with weak sequestering agents such
as the iron-carbohydrate complexes were thought to be.
Preliminary experiments were therefore conducted with
various buffer systems to determine if iron-buffer
chelates were formed. It was found that bicarbonate,
acetate, and formate did not complex with iron and were
therefore suitable for electrophoresis measurements.
Paper ionophoresis experiments were conducted in
0.1 M buffer at 2.5 ma per cm of strip width, using
■Whatman #3 filter paper. This current was used to promote
rapid movements and minimize diffusion. The time was
usually 16 hours, although it was sometimes possible to
obtain significant results in 8 hours.
I
81
The location of the iron chelate was ascertained
by spraying the air-dried paper strips with acid ferro
cyanide. The position of the fructose band was identified
with a resorcinol spray (135). Since there is buffer
movement by electro-osmosis, this fructose zone was used
as the origin to establish the distance which the chelates
had migrated.
The structure of ferric citrate has been suggested
to be that of a tridentate, hydrated, negatively charged
ion, as shown below (111).
COO
"00C CHp C ---0 — Pe OHp
I / \
CO 0 oh2
From the similarity of the possible coordinating groups,
it was expected that tartrate and malate might be strong
iron sequesterants. Chelates of tartrate, malate, and
citrate were made by adding the organic acid at a final
concentration of either 0.1 M or 0.2 M to 0.1 M solutions
of ferric chloride and adjusting to the desired pH with
6 N NaOH. From the distinct color differences which
developed in alkaline solution, it was clear that the
chelate species present were a function of the organic
acid concentration. Thus, at pH 7> the tartrate solution
containing equimolar proportions of iron and acid is dark
red. At a 2:1 ratio of acid to metal, the tartrate
solution is light yellow. The 1:1 malate solution was
unstable and a precipitate formed on adding base.
Aliquots of the organic acid chelates were adjusted to the
buffer pH, mixed with equal portions of the buffer and
approximately 0.01 ml applied to the paper. Ionophoresis
was carried out as above. Ferric fructose and ferric
sorbitol, both with a 32-fold excess of the polyol, were
also examined. Typical results are shown in Figure 11.
At pH 7 in bicarbonate buffer it was found that
all the chelates tested were anionic. The strong organic
acid complexes had approximately twice the effective
charge of either the sorbitol or fructose complexes. The
organic acid chelates had the same effective charge, even
though the species were different, as determined by the
color. By ionophoresis at various pH’s near the pKa's of
the three buffer systems, it was determined that the iso
electric region of ferric fructose lay between *+.5 and
b , 7 . Both ferric citrate and ferric tartrate are still
negatively charged at this pH. Difficulty was encountered
with the ferric sorbitol; a precipitate often appeared on
the addition of buffer. The iron may have been dissociated
from the sorbitol and thus the iron staining material
detected.may have been, in reality, ferric hydroxide.
83
ferric citrate llilfru c to se
ferric |
fructos^
ru cto se
O .l |i A cetate buffer, pH 4.5
+ ferric fructose
Ifru c to s e
ferric citrate fructose
0.1 |j l Bicarbonate buffer, pH 7 .0
Figure 11.— Electrophoresis of ferric chelates
at two pH's.
It was of interest to determine whether or not a
chelate of glucose and iron could be made in the absence
of hydroxyl ion. An excess of anhydrous glucose was added
to absolute ethyl alcohol, resulting in a saturated
solution (ca. 18 gm per liter) at room temperature. Two
hundred mg of anhydrous sublimed PeCl^ was dissolved in 10
ml of alcohol, which was then mixed with 100 ml of alco
holic glucose. There was no apparent color change when
compared with a control containing no glucose. On the
addition of a small amount of alcoholic NaOH, a brown
precipitate formed which rapidly dissolved on standing.
It was concluded that 0H“ is an integral part of the iron-
glucose complex.
Attempts were made to separate the iron glucose
and iron fructose complexes from the aqueous solution.
Ethanol proved an ideal agent. The iron glucose complex
is only sparingly soluble in 80$ alcohol, and thus could
be readily separated from excess glucose by precipitation.
Alcoholic solutions containing the iron glucose complex
were centrifuged, washed several times with 95$ ethanol,
and the complex recovered. These chelates proved to be
quite soluble in water after drying in vacuo for 1 week.
Elemental analysis revealed that one such preparation con
tained 1 mole of hexose to b gram atoms of iron.
85
Attempts to prepare a similar fructose iron
complex resulted only in recovery of a dark syrup. This
syrup, soluble in water, was also quite soluble in alcohol
and other polar solvents in which fructose would be
expected to be soluble. It became apparent that the rate
of formation of the complex, as well as the final pH, were
important factors in the type of chelate formed.
“Fast1 1 and "slow" ferric fructose were prepared as
previously described. The final pH was 9*0 in both cases.
As soon as the proper degree of alkalinity was achieved,
i
the solutions were immediately added to a 20-fold excess
of absolute ethanol. A yellow flocculent precipitate was
formed. The solution was centrifuged and the precipitate
washed and reprecipitated twice with absolute ethanol.
The solid buff-yellow powder remaining was dried 2 days
h.
in vacuo and analyzed with the following results:
Weight Percent
C H Asl1
"Fast" ferric fructose 27.38 5-01 + 29.37
"Slow" ferric fructose 22.93 *+.35 36.^5
Since the elemental compositions of these
substances were so different, it was necessary to deter
mine whether reproducible results could be obtained from
k
Truesdail Laboratories, Los Angeles, California
86
preparations made with different molar excesses of
fructose.
To 50 ml aliquots of 0.1 N FeCl^, a calculated
amount of fructose was added and the final volume brought
to 100 ml. The final solutions, all of which were .05 M
in Fe+++, contained 16, 32, and 6k- molar excesses of
fructose. Aliquots of each were removed and 6 N NaOH was
slowly added until the aliquot pH was 10.0. Based on this
initial titration, calculated volume's of NaOH were added
to each of the sugar-iron solutions and rapidly mixed.
The pH was immediately checked and the solution treated
with an excess of absolute ethanol. Both the 32x and 6kx
solutions contained a dark syrup which rapidly settled,
as well as a yellow precipitate which settled more slowly.
These were separated, washed, and dried in vacuo for 2
days. They were then subjected to elemental analysis with
the results indicated in Table 7»
It is evident that similar products result from
solutions containing different proportions of iron to
sugar. That the difference between a syrup product and a
flocculent precipitate depends upon the final pH of the
solution from which the solid is precipitated was demon
strated as follows:
"Fast" ferric fructose containing a 32x molar
excess of fructose was made as previously described and
87
TABLE 7‘^-Elemental composition of ferric fructose
complexes
Fructose iron
ratios in
aqueous solution
H Fe Na
l6x
32x
6l+x
32x (syrup)
64-x (syrup)
Weight per cent
33-07 5.04-
33.03 4-.94- 4-9.01 6.88 6 . 1 b
32.4-3 *+.82
34-. 50 5. l b
33.59 4-.98 4-7.28 6.83 7.32
I6x
32x
6 b x
32x (syrup)
64-x (syrup)
Atom per cent
2.75 5.04-
2.75 4-.94- 3.06 0.123 0.266
2.70 4-.82
2.86 5.14-
2.80 4-.98 2.95 0.122 0.318
88
the brown syrup was separated and dried. One gram of the
dry powder was dissolved in 25 ml of distilled water,
resulting in a solution of pH 11.*f. The solution was
divided into two equal portions, one of which was
acidified to pH 7»0. The ferric fructose from both was
then precipitated by an excess of absolute ethanol. The
precipitate obtained from the neutral solution was a buff
colored powder, while that from the alkaline solution was
a dark brown syrup, identical in appearance to that
originally recovered.
Absorption of Iron from the Gut
There is evidence (29, 30, 31) that the presence
of carbohydrates plays a significant role in the absorp
tion of dietary iron through the mucosa of the intestinal
wall. To pass from the lumen into the blood, iron must
either pass through the intercellular spaces or, what is
more likely, directly enter the cells. As previously
described, cellular membranes possess fixed charges within
them. Thus, it might be expected that iron-carbohydrate
chelates of low effective charge would more readily
penetrate charged membranes than either the metal ion
itself or some more highly charged small chelate molecule.
Any chelate passing through a cell or series of
cells enters into equilibrium with other ligands present
within the boundaries of the membrane and is directly
affected by the redox reactions of the cell. Nevertheless,
it was hoped that if the metabolic conditions within the
mucosal cells were approximately equal in experiments with
several animals, that the net ion flux through the mucosal
wall would be reflected in an increased iron content of
the blood. This was investigated in the following way;
Male rabbits of unknown breed, weighing between 5
and 7 pounds, were repeatedly phlebotomized via cardiac
puncture. A total of 75 ml of blood was removed at weekly
intervals. This was done to reduce the level of trans
ferrin-bound iron and to provide increased sensitivity in
observing iron uptake. This procedure did not render the
animals anemic since all had hematocrits between *+0 and
*+5» The animals were deprived of food for 2 b hours prior
to use.
The rabbits were anesthetized with 20% urethane in
saline by injection into the marginal ear vein. The
dosage used was 7*5 ml per kg. If additional anesthetic
were needed, 0.1 to 0.2 ml of nembutal (60 mg per ml) was
used in addition to the urethane. The areas of the neck
and abdomen were clipped free of fur and one carotid
artery and jugular vein dissected free of surrounding
tissue. The peritoneal cavity was entered and a 20 cm
section of the duodenum immediately distal to the pyloric
valve was located. This section was isolated from the
90
remainder of the gut with double ligatures, making sure
that the vascular bed was left intact.
Hypodermic needles were inserted in each end of
the isolated segment and the end of each needle was covered
with a short length of polyethylene tubing to prevent undue
damage to the mucosa.
A glass spiral was constructed which closely fit
the well of the scintillation counter previously described.
The total contained volume of the spiral was 1.75 ml.
This was connected to two lengths of polyethylene tubing,
I.67 mm inside diameter by 2.b 2 mm outside diameter. One
length of the tubing was equipped with a glass "T" and
syringe adaptor to permit the introduction or withdrawal
of fluids during the experiment. The total volume of the
tubing and spiral was 5*8 ml. One-half ml of heparin (50
mg per ml) was injected into the animal via the marginal
ear vein and 0.5 ml was injected subcutaneously to prevent
clotting. No difficulty with blood coagulation was ever
encountered when this procedure was followed.
The jugular vein and carotid artery were then
cannulated, in that order, using the free ends of the
polyethylene tubing. Blood from the rabbit's carotid
artery then flowed through the glass spiral inserted in
the scintillation well and returned via the jugular vein.
Total elapsed time in this circuit varied from 20 to H-0
91
seconds, depending on the animal's condition. The
experimental operations are illustrated in Figure 12.
Using the needles previously inserted in the gut
segment, the lumen was washed with 25 ml portions of
physiological saline at 37° C. Gentle pressure eliminated
excess saline from the gut after which the distal needle
was removed, and a ligature tied around the end. Four ml
of the iron solutions under investigation containing Fe^9
tracer were then injected into the proximal end of the gut
loop, which in turn was closed with thread. The needle
was left in position in order to prevent radioactive con
tamination of the blood. The activity of the blood was
monitored, using a pulse height analyzer set for the
1.098 Mev peak, linked to a rate meter and a rectilinear
recorder. To avoid excessive fluctuation, the rate meter
time constant was set for the lowest response— ^-0 seconds.
In all cases, the total amount of iron was 2 mg. Where
ferrous sulfate was used, it was prepared in oxygen-free
water, and all manipulations except the final injections
were carried out in an atmosphere of nitrogen gas. The
results of several experiments are presented in Figures
13, l b , 15, and 16.
Sorbitol has previously been cited (^5) as being
effective in promoting iron absorption. Figure 13
represents the results obtained from a b ml duodenal
Figure 12.— Experimental conditions used to
investigate iron flux through the mucosal wall of
rabbits.
B lood activity, counts/min.
93
150-
120
90-
inject 2 mg. Fe**'
as ferric sorbitol 60
30
10 20 30 40 0 50 60 70 80
Time, minutes
Figure 13.— Absorption of ferric sorbitol from
rabbit duodenum as measured by increase in serum
Fe59.
injection containing 2 mg of iron as ferric chloride.
Sorbitol^ was added to provide a 32 molar excess of the
polyol. The solution was brought to isotonicity by the
addition of sodium acetate, and the pH adjusted to k -,7
prior to injection. The acetate was 0.015 molar and pro
vided only a very slight buffering capacity. The total
radioactivity of Fe^ contained in the ml was 1,530,000
counts per min. By means of the "TM inserted in the
polyethylene tubing, samples of blood were withdrawn for
hematocrit and transferrin iron level determinations at
the beginning and end of the experiment. Occasionally,
throughout the course of the experiment, a small air
bubble was introduced into the tubing via the "T11; the
movement of the bubble could be followed through the
translucent tubing and permitted measurement of flow rate
in the circuit. A time interval of between 1 and 2
minutes is required before radioactivity, introduced into
the portal blood, can be detected by the counter. Figure
13 indicates that ferric sorbitol rapidly enters the
bloodstream with no detectible time lag other than that
imposed by the experimental limitations. It is to be
noted that blood radioactivity was still increasing when
the experiment was terminated. The contents of the gut
^Atlas Powder Co., supplied as a 70% solution.
95
were examined at the conclusion of the experiment. They
contained a viscous fluid of pH 8.0. A qualitative test
for ferrous iron with dipyridyl was negative.
A similar experiment with fructose, instead of
sorbitol, is illustrated by Figure 1^. In this case, a
total of 1,0*4-0,000 counts per min of Fe^9 was used as a
tracer; the concentration of total iron was 2 mg per ^ ml
of fructose solution. The rate of entry of ferric fruc
tose is significantly faster than ferric sorbitol. It is
seen that, after 1 hour, the level of radioactivity in the
blood reaches a plateau. This result was found in many
similar experiments where the initial rate of entry was
rapid and adequate time allowed before the animal was
sacrificed. The gut contents of this experimental animal
were similar to that observed with sorbitol. The gut
contained a mucoid solution of pH 7.8. In this instance,
traces of ferrous iron were detected in the solution.
To test the ability of non-chelated ferrous ion to
enter the blood, ferrous sulfate in 0 .9 % saline, was
prepared as previously described. Acetate, 0.015 M, was
also used to provide comparative data with other experi
ments. The total radioactivity of the sorbitol-iron
solution was 1,350,000 counts per min. Figure 15 presents
the results of this experiment. The initial slope of the
uptake curve is similar to that recorded for ferric
Blood activity, counts/m in.
96
300-
240
180
inject 2mg.Fe~*
as ferric fructose
120 -
60
80 70 60 50 30 40
Time, minutes
20
Figure 1^.— Absorption of ferric fructose from
rabbit duodenum as measured by increase in serum Fe?9.
Blood activity, counts/min.
97
300
sacrifice
240
180
inject 2 mg. Fe+ *
as FeS04
120
60
50 0 20 30 4 0 7 0 10 60 80
Time, minutes
Figure 15«— Absorption of ferrous ion from
rabbit duodenum as measured by increase in serum Fe79
fructose. Since the specific activity of the ferrous
sulfate was greater than that of ferric fructose, it can
be concluded that ferric fructose is assimilated more
rapidly. After about k-5 minutes the level of activity
remained constant until the animal was sacrificed.
The contents of the isolated section of gut were
drained into a beaker and the pH ascertained to be 7»5»
The mucoid solution, total volume 7.6 ml, was centrifuged
at 10,000 g for a period of 10 min. The clear supernatant
was decanted and an aliquot assayed for Fe^. The pre
cipitate was dissolved in concentrated HC1 and the
solution diluted to a known volume. An aliquot of this
solution was counted. The insoluble fraction contained
65$ of the total activity drained from the gut. The
original clear supernatant fraction containing 35% o f the
activity was divided into several aliquots. An aliquot
was tested for ferrous ion; no free ferrous ion was found.
Another aliquot was made acid with concentrated HC1 and
tested with a 2% potassium ferrocyanide solution. A pale
blue color, typical of ferrous iron, developed. A third
aliquot of the supernatant was made strongly basic, with
6 N NaOH. This aliquot was centrifuged; no precipitate
could be detected. As a control, 2 mg of iron in 7*6 ml
of water was brought to pH 7.5 with 0.1 N NaOH. A
copious dark green precipitate was immediately formed.
99
It is clear that the soluble ferrous iron present in the
intestinal contents was complexed with some endogenous
ligand.
The ability of ferric chloride to be absorbed by
the rabbit was tested. Ferric chloride was dissolved in
isotonic saline to bring the concentration of the iron to
2 mg per *+ ml. Fe^9 was then added to a total activity of
1,28^,000 counts per min. The pH was adjusted to 2.8,
which preliminary experimentation indicated was the
maximum degree of alkalinity at which no precipitate of
ferric hydroxide was visible. The solution was infused
into the isolated duodenal section as before. After 82
minutes there was essentially no increase in blood radio
activity above background.
Since it had been proposed that iron is absorbed
through the mucosa only in the ferrous state, it was
desirable to test this postulate in the same animal which
had shown negligible absorption of trivalent iron.
Accordingly, a solution of 4 ml of radioactive ferric
sorbitol was prepared as described previously. Sorbitol
was chosen since preliminary experiments had shown that
ferric iron was not reduced under these conditions. The
pH was adjusted to 2.8; the total radioactivity injected
was 1,0^-0,000 counts per minute. A section of the ileum
approximately 10 cm long, below the section of the gut
100
into which inorganic FeCl^ had been injected, was selected
and isolated by ligatures from the rest of the intestinal
tract. The gut loop was rinsed with warm saline and the
ferric sorbitol solution injected. Within- a few minutes,
as seen in Figure 16, the radioactive iron appeared in the
blood. At the end of 3 hours the animal was sacrificed.
The gut contents were removed from each loop and centri
fuged. The solution from the loop containing sorbitol
had a pH of 7«5} while that of the FeCl^ was pH 8.5. Wo
ferrous iron could be demonstrated in either solution.
Only 2,500 counts per minute of an injected total of
1,28*+, 000 counts per minute were present in a non
precipitated form in the ferric chloride treated section.
The radioactivity in the supernatant drained from the
ileac segment containing sorbitol was 183,000 counts per
minute, with the balance remaining in the precipitate.
It is evident from the preceding experiments that
the intestine of a rabbit is capable of secreting a sub
stance which will specifically complex with ferrous, but
not with ferric iron.
Serum was prepared from blood samples of all
experimental animals both before and at the conclusion of
the tests. In several experiments, iron-binding
capacities were determined. These are presented in Table
8. It can be seen, in cases where a plateau in the uptake
Blood activity, counts/min.
101
150
120
90-
inject 2mg.Fe4 4 4
as ferric sorbitol 60 - inject 2mg.Fe
as FeCls
30-
130 140 100 110
Time, minutes
120 90 80 70
Figure 16.— Absorption of ferric ion and ferric
sorbitol from rabbit intestine as measured by increase
in serum Fe59.
102
TABLE 8.— The regulation of iron absorption from the gut
by serum iron-binding capacity
Solution infused
into gut
Plateau in
uptake curve
reached
Transferrin binding
capacity available
Jig %
Before After
FeSOL,., pH k . 7 Yes
69
22
FeSOi^., pH k . 7 Yes 22*+ 21
FeSO^, pH b .7 Yes
293 13
FeSO^, pH b . 7 Yes
69
22
Ferric fructose,
pH b . 7 Yes Not done
31
Ferric fructose,
pH 8.0 No Not done 122
FeCl3, followed by
' ferric sorbitol No 161 150
103
curve was reached, that only a small amount of iron-
binding capacity remains in the blood. Such serum trans
ferrin saturation has invariably occurred when iron in
readily diffusible form was used. Under conditions where
a less rapidly assimilable chelate was used, iron absorp
tion continued until the time of sacrifice. It may be
concluded that under these circumstances, absorption
would also continue until the serum iron level approached
saturation.
CHAPTER IV
DISCUSSION
Evidence has been presented for the participation
of small chelate molecules in the transport of iron from
serum to tissue. It is unlikely that the chelate molecules
manifest activity as metabolic substrates in the regula
tion of iron flux. It has been demonstrated (7lj 73) by
inhibitor studies that iron uptake in liver and in everted
gut loops is a passive process not directly coupled to
respiratory mechanisms. Furthermore, EDTA, one of the
chelates used in this research, was effective in promoting
iron uptake by liver slices, but is not effectively
oxidized by the tissue. The chemical environment existing
in the blood precludes the existence of ferric ion in con
centrations in excess of 10"^ M; further, it is unlikely
that ferrous ion can exist in significant concentrations
at the oxygen tension present there.
Iron can exist as a chelate in a soluble and
readily mobilized form. Evidence for the participation of
chelates in iron metabolism was obtained by Reissman et al.
(35)• Large amounts of ferrous iron, inserted into the
duodenum or colon of dogs, resulted in a marked increase
10*+
105
in the concentration of citrate, lactate, and pyruvate in
the blood. These organic acids have been shown to bind
ferric ion with great avidity. Since blood and tissue
contain a number of such molecules known to bind metallic
ions, it is quite possible that they play £ fundamental
role in trace metal metabolism.
It is evident from the experiments presented in
the previous section that competition exists between
transferrin and liver tissue for metal ions. The movement
of iron between the various compartments of the mammalian
body has been shown to be greatly influenced by the con
centration and degree of saturation of the various
specific binding molecules.
The elegant experiments of Jandl and co-workers
(59) with reticulocytes demonstrate the influence of
binding capacities and affinities on iron transport. Iron
binding protein (i.e., transferrin) competed directly with
the reticulocytes for radioiron in the medium. The degree
of saturation of the transferrin governed the rate and
amount of Fe^ taken up by the reticulocytes, as well as
by rat liver slices. Whereas the membrane seems to play
a very decisive role in the regulation of iron uptake by
reticulocytes, the observations of Bass and Saltman (115)
clearly demonstrate that when ferric citrate is used, the
integrity of the membrane of the liver cell is not
106
intimately linked to iron transport in their system.
The molecular size and charge of complexes, as
well as their stability constants, markedly affect both
the rate of flux and final concentrations of metal ions.
Iron bound to transferrin, M. ¥. 88,000, diffuses very
slowly, if at all, into a liver slice. However, when
transferrin iron is incubated with liver slices in the
presence of chelates, the iron passes rapidly into the
slices. Sequesterants can also remove intercellular ions.
This was found to be the case in the intact rat by Fore
man et al. (92) who injected calcium EDTA and succeeded
in removing iron via the urine. Rubin and Princiotto
(136) injected the iron chelate of diethylenetriamine-
pentacetic acid (DTPA) into rats. The avidity of the
ligand for iron was so great that there was quantitative
excretion in the urine; no iron was incorporated into
tissue.
That the influence of charge, as well as size, is
important is illustrated by the data (93) which indicate
that some ferric chelates, such as EDTA, are not well
absorbed by the mammalian gut. Davis (137) has pointed
out that, within the cell, uncharged water-soluble
metabolites are found only as excretory and fermentation
products and occasionally as intermediates in purely
degradative reactions (e.g., acetaldehyde in alcoholic
107
fermentation). Ionized biosynthetic intermediates were
probably an evolutionary advantage to the cell, in that
they may have been retained more efficiently within the
confines of the cellular membrane. This concept is in
accord with observed ion exchange characteristics of bio
logical membranes. As has been previously pointed out,
if the presence of a chelate reduces the effective charge
on a metal ion, then it would be expected that the rate of
transport of the ion across a charged barrier would be
enhanced. It should also be noted that such facilitated
transport would prove of equal efficacy in providing for
the accumulation of otherwise charged ligands. Enhanced
diffusion of chelates might well play a part in the
active transport of metabolites into cells against con
centration gradients.
Once inside the cell, the concentration of free
ligand and the metal ion would not be expected to remain
the same as found outside. The competition of H*,
naturally occurring ligands and endogenous metallic ions
will drastically change the equilibrium concentrations.
The possibility that redox reactions are involved
in the control of iron metabolism should be emphasized.
It is well known that the dissociation constants of Fe++
XJ.J.
and Pe for various binding molecules are quite dif
ferent. The affinity of Fe+++ for most of the cellular
108
ligands which have been investigated is greater than that
of Fe++.
The rate at which ferritin can be formed from
apoferritin in the presence of ferrous ion and oxygen may
well be greater than that observed with ferric ion. This
finding, however, may be a function of the tremendously
increased solubility of the ferrous ion under physiological
conditions. The ability of Fe+++, if present in suitable
chelated form, to combine with apoferritin in the absence
of reducing substances has been demonstrated. It has
been shown that the existing data on the formation of
ferritin in vitro are subject to many experimental arti
facts and do not unequivocally implicate ferrous iron as
the unique reactive form of the metal.
Evidence has also been presented that the ferric
iron present in ferritin can be removed from the protein
molecule if a suitable chelating agent is available.
Contrary to published data (78, 79), it is not mandatory
that the solubilizing agent function in a reductive
capacity.
This research has demonstrated that ferric ion is
not transported across the intestinal mucosa in measurable
quantities. However, if the charge on the molecule is
reduced by the addition of a suitable ligand, even though
the molecular size is increased, the complex is readily
109
absorbed. Due to the presence of a mucosal secretion
which can complex with ferrous iron, but not ferric, the
rate at which the divalent metal alone is absorbed has not
been established. It will be of interest to elucidate the
nature of this natural ligand.
The role of carbohydrates in enhancing the
absorption of iron, as well as other trace elements, has
been studied. It has been shown that these substances,
under certain conditions, can complex with and solubilize
metal ions at neutral or alkaline pH, which otherwise
would cause precipitation of the ions as insoluble
hydroxides. The complexes of iron and carbohydrates are,
in general, negatively charged or neutral at physiological
pH's. They may be large polymers or small diffusible
molecules, depending not only on the relative concentra
tions of sugar and metal ion, but also the rate of forma
tion and the final pH.
The key role of serum transferrin in regulating
iron uptake from the lumen of the gut into the blood has
been illustrated. In all cases in which absorption had
ceased, as measured by blood radioactivity, it was found
that the serum iron-binding capacity was essentially
saturated. Where uptake was proceeding at the time the
animal was sacrificed, it was found that the transferrin
had not been saturated. The role of ferritin in this
control of iron transport across the gut seems of secondary
significance. The total amount of iron which can be
bound by circulating transferrin at any one time is quite
small. Transferrin is known to participate in iron move
ment between various cellular entities. The comparatively
small size of this compartment has probably been a primary
cause for the failure to recognize its fundamental role in
dietary iron absorption. It may well be that iron
absorption and incorporation into the mucosal cells may
continue even after the transferrin is saturated and iron
no longer moves into the blood. However, when mucosal
cells and transferrin are both saturated, no more iron
will be absorbed unless the permeability of the gut wall
is disrupted by excessive quantities of iron as is observed
in iron toxicity poisoning. Once the transferrin and
mucosal cells have been saturated, iron absorption appears
to be regulated by the rate at which transferrin-iron is
utilized by other tissues.
The essentiality of forming soluble iron chelates
must be emphasized. In the absence of carbohydrates, or
other suitable ligands, dietary iron is rapidly converted
into insoluble hydroxides which are unavailable for
absorption despite the presence of unbound iron-binding
capacity in transferrin. By understanding the requirement
for the chelated form of iron as well as the chemical
Ill
nature of the complexes of this ion with polyhydroxy
compounds and sugars, we can provide a rational explana
tion for the hitherto unknown effect of sorbitol 0+5) or
high carbohydrate diets (29, 30, 31) in causing enhanced
iron accumulation.
CHAPTER V
SUMMARY
Evidence is presented for the participation of low
molecular weight chelates in iron metabolism. It has been
shown that these chelates function in the flux of iron
from blood to tissue.
The role of carbohydrates in enhancing iron
absorption has been studied. Several sugars, including
glucose, fructose, ribose, and others, form soluble
complexes of iron. These chelates may be high molecular
weight polymers or small molecules of zero or negative
charge, depending on the rate of formation and final pH.
A minimum ratio of ■+ moles hexose to 1 gram atom of iron
is required to form complexes which are stable over a pH
range of 2 to 13. It has been demonstrated that these
soluble complexes of low net charge facilitate iron move
ment through the intestinal barrier.
The presence of ferrous iron is not mandatory for
the formation of ferritin from apoferritin, as had
previously been assumed. Ferric iron can b-? directly
incorporated if presented to apoferritin in suitable
chelate form. Similarly, ferritin will release its iron
-I
112
113
in the absence of reductants if suitable sequestering
agents are present.
It has been demonstrated that ferric iron is
easily absorbed by blood from the gut if it is presented
as an appropriate chelate of sorbitol or hexose. The
classical concepts of the "mucosal block" as the regulator
of iron metabolism must be re-examined in light of these
experiments. The blood, and not the intestinal mucosa,
exercises primary control of iron absorption.
LITERATURE CITED
LITERATURE CITED
1. Hahn, P. F., Bale, W. F., Ross, J. F., Balfour, W. M.,
and Whipple, G. H., J. Exp. Med., £8, 169 (l^S)*
2. Fontes, G., and Thivolle, L., Compt. rend. soc. biol.,
22, 685 (1925).
3. Hahn, P. F., and Whipple, G. H., Am. J. Med. Sci.,
121, 2 b (1937).
l+. McCance, R. A., and Widdowson, E. M., Lancett li,
680 (1937).
5. Widdowson, E. M., and McCance, R. A., Biochem. J.,
21, 2029 (1937).
6. McCance, R. A., and Widdowson, E. M., J. Physiol.
(London), 2z, 1^8 (1938).
7. Copp, D. H., and Greenberg, D. M., J. Biol. Chem.,
A . 385 (19I+6).
8. Karlefors, T., and Norden, A., Acta Med. Scand.
Suppl., 3^2 (1958).
9. Laufberger, V., Bull. Soc. Chim. Biol., 12, 1575
(1937).
10. Granick, S., Physiol. Rev., 2i, ^89 (1951).
11. Granick, S., J. Biol. Chem., 1^6, *+51 (19^2).
12. Maclnnes, D. A., Ann. N. Y. Acad. Sci., |+2, 175
(19^2).
13. Granick, S., and Michaelis, L., J. Biol. Chem., lj+Z,
91 (19^3).
1^-. Michaelis, L., Coryell, C. D., and Granick, S.,
J. Biol. Chem., 1^8. ^63 (19^3).
15. Granick, S., and Hahn, P. F., J. Biol. Chem., 155.
661 (19^f).
115
16.
17.
18.
19.
20.
21.
22.
23.
24-.
25.
26.
27.
28.
29.
30.
31.
32.
33.
116
Kuff, E. L., and Dalton, A. J., J. Ultrastructure
Res., 1, 62 (1957).
Rothen, A., J. Biol. Chem., Ig2, 679 (194*0.
Block, M., Bethard, ¥., and Jacobson, L. 0.,
J. Lab. and Clin. Med., 40, 781 (1952).
Granick, S., in“Metabolism and Function of Iron,"
Report of Nineteenth Ross Pediatric Research
Conference, p. 15) Columbus, Ohio (1956).
Granick, S., Chem. Rev., 385 379 (194-6).
Granick, S., Ann. N. Y. Acad. Sci., J+8, 657 (194-7).
Granick, S., Bull. N. Y. Acad. Med., 2£, 4-03 (194-9).
Granick, S., Bull. N. Y. Acad. Med., 32j 8l (1954-).
Josephs, H. W., Blood, 13, 1 (1958).
Bothwell, T. H., Medical Science, 6, 4-73 (1959).
Keiderling, W., and WiJhler, F., Arch. Exper. Path. u.
Pharmakol., 221. 4-18 (1954-).
Heilmeyer, L., in “Iron in Clinical Medicine,"
Wallerstein, R. 0., and Mettier, S. M. eds., p. 24-,
Univ. of Calif. Press, Berkeley (1958).
Dubach, R., Callender, S., and Moore, C. V.,
Blood, 3, 526 (194-8).
Gillman, J., Mandelstam, J., and Gillman, T.,
South African J. Med. Sc., 10, 109 (194-5).
Gillman, J., and Gillman, T., Arch. Path., J+0, 239
(194-5).
Kinney, T. D., Hegsted, D. M., Finch, C. A.
J. Exp. Med., 20, 137 (194-9).
Granick, S., and Michaelis, L., Proc. Soc. Exp.
Biol. Med., 66, 296 (194-7).
Bothwell, T. H., van Doorn-Wittkampf, H. vW.,
DuPreez, M. L., and Alper, T., J. Lab. and Clin.
Med., 4a, 836 (1953)*
117
3*+. Wallerstein, R, 0., and Robbins- S. L., Am. J. Med.,
lit, 256 (1953'T.
35. Reissmann, T. R., and Coleman, T. J., Blood, 10, 1+6
(1955).
36. Conley, B. E., J. Am. Med. Assoc., 170. 676 (1959).
37. Granick, S., J. Biol. Chem., 16J+, 737 (19“ +6).
38. Moore, C. V., Dubach, R., Minnich, V., and Roberts,
H. K., J. Clin. Invest., 2^., 755 (19^).
39. Greenberg, S. M., Tucker, R. G., Heming, A. E., and
Mathues, J. K., J. Nutrition, 6^, 19 (1957).
1+0. Gubler, C. J., Science, 123. 87 (1956).
l+l. Groen, J., Van Den Broek, W. A., and Veldman, H.,
Biochim et Biophys Acta, 1, 315 (19*+7).
1+2. Woodruff, C. E., Borden's Rev. Nutrition Research,
20, 61 (1959).
1+3. Kovacs, E., and Bernat, J. Therapia Hungarica, .2,
l (1957).
1+1+. Franklin, M., Rohse, W. G., de la Huerga, J., and
Kemp, C. R., J. Am. Med. Assoc., 166. 1685 (1958).
1+5. Herndon, J. F., Rice, E. G., Tucker, R. G.,
Van Loon, E. J., and Greenberg, S. M.,
J. Nutrition, 64, 615 (1958).
1+6. Fournier, P., Dupuis, Y., Susbielle, H., Alle, M.,
and Tardy, N., J. de Physiol., J+2., 339 (1955).
1+7. Wasserman, R. H., Comar, C. L., and Nold, M. M.,
J. Nutrition, 22.* 371 (1956).
1+8. Lengemann, F. W., J. Nutrition, 6£, 23 (1959).
1+9. Cartwright, E. G., Wintrobe, M. M., J. Clin. Invest.,
28, 86 (194-9).
50. Rath, C. E., and Finch, C. A., J. Clin. Invest.,
2&, 79 < i* 9 ) .
118
51. Peterson, R. E., and Mann, J. D., Am. J. Physiol.,
169, 763 (1952).
52. Eluenberg, A., and Laurell, C. B., Acta. Chem.
Scand., 2, 68 (1955).
53. Wallenius, G., Scand. J. Clin, and Lab. Invest.,
k , 2 b (1952).
5*+. Cohn, E. J., Surgenor, D. M., and Hunter, M. J. in
"Enzymes and Enzyme Systems," Edsall, J. T., ed.,
Harvard Univ. Press (1951).
55. Shade, A., Reinhart, R., and Levy, H., Arch. Biochem.,
20, 170 (19^9).
56. Laurell, C. B., Pharmacol. Rev., J+, 371 (1952).
57* Wasserman, L., Rashkoff, I., Leavitt, D., Mayer, J.,
and Port, S., J. Clin. Invest., 3-Lj 32 (1952).
58. Cruz, W., Hahn, P., and Bale, W., Am. J. Physiol.,
135, 595 (19^2).
59* Jandl, J. H., Inman, J. K., Simmons, R. L., and
Allen, D. ¥., J. Clin. Invest., 38, 161 (1959).
60. Bailar, J. C., Jr., "Chemistry of the Coordination
Compounds," p. V+8 et seq., Reinhold Publishing
Co. (1956).
61. Treibs, A., Angew. Chem., b ^ , 682 (1936).
62. Laurell, C. B., Acta Physiol. Scand., l}i> Suppl. *+6
(19^7).
63. Flexner, L. B., Vosburgh, J., and Cowie, D.,
Am. J. Physiol., 153, 503 (19*+8).
6M-. Laurell, C. B., Blood, 6, 183 (1951).
65. Inman, J. K., Thesis, Harvard Univ. (1956).
66. Laurell, C. B. in "Iron in Clinical Medicine,"
Wallerstein, R. 0., and Mettier, S. M., eds.,
p. 8, Univ. of Calif. Press, Berkeley (1958).
119
67. Dubach, R., Callender. S. T. E., and Moore, C. V.
Blood, 2, 526 (1948).
68. Yuile, C. L., Hayden, J. W., Bush, J. A., Tesluk, H.,
and Stewart, W. B., J. Exp. Med., 2£ j 367 (1950).
69. Huff, R. L., Hennessy, T. G., Austin, R. E.,
Garcia, J. F., Roberts, B. M., and Lawrence, J. H.,
J. Clin. Invest., 2£, 1041 (1950).
70. Saltman, P., Fiskin, R. D., and Bellinger, S. B.,
J. Biol. Chem., 220, 74-1 (1956).
71. Saltman, P., Fiskin, R. D., Bellinger, S. B., and
Alex, T., J. Biol. Chem., 220. 751 (1956).
72. Saltman, P., Frisch, H., Fiskin, R. D., and Alex, T.,
J. Biol. Chem., 221, 777 (1956).
73. Brown, E. B., Jr., and Justus, B. ¥., Am. J. Physiol.,
194-. 319 (1958).
74-. Bass, R., Bernick, S., and Saltman, P., Exptl. Cell
Research, 13. , 395 (1959)*
75. Michaelis, L., Adv. in Protein Chemistry, III. 53
(194-7).
76. Bielig, H., and Bayer, E., Naturwiss, 42, 125 (1955) •
77. Loewus, M., and Fineberg, R., Biochim. et Biophys.
Acta, 26, 44-1 (1952).
78. Mazur, A., Baez, S., and Shorr, E., J. Biol. Chem.,
2ia, 147 (1955).
79. Mazur, A., in "Metabolism and Function of Iron,"
Report of Nineteenth Ross Pediatric Research
Conference, p. 27, Columbus, Ohio (1956).
80. Green, S., and Mazur, A., J. Biol. Chem., 227. 653
(1957).
81. Fisher, H., and Orth, H., "Die Chemie des Pyrroles."
II Band I Halfte, Leipzig. Akademische
Verlagsgesellschaft (1937;•
82. Charley, P., Rosenstein, M., Shore, E., and
Saltman, P., Arch. Biochem. and Biophys. (in press X
83.
8* + .
85.
86.
87.
88.
89.
90.
91.
92.
93.
9*+.
95.
96.
97.
98.
99.
100.
120
Pirzio-Biroli, G., in "Metabolism and Function of
Iron," Report of the Nineteenth Ross Pediatric
Research Conference, p. 2*+, Columbus, Ohio (1956).
Wacker, W. E. C., and Vallee, B. L., J. Biol. Chem.,
23*+. 3257 (1959).
Nielands, J. B., Bact. Rev., 21, 101 (1957).
Metzler, D. E., Ikawa, M., and Snell, E. E.,
J. Am. Chem. Soc., £6? 6*+8 (195*+).
Chaberek, S., and Martell, A. E., "Organic Sequester
ing Agents," p. *+55j Wiley (1959).
Waters, L. L., and Stock, C., Science, 102. 601
(19*+5).
Kety, S. S., J. Biol. Chem., l]+2, l8l (19*+2).
Schubert, J., Science, 105. 389 (19*+7).
Schubert, J., Ann. Rev. Nuclear Science, 369
(1955).
Foreman, H., Huff, R. L., Oda, J. M., and Garcia, J.,
Proc. Soc. Exp. Biol, and Med., 2 2 i 520 (1952).
Larsen, B. A., Bidwell, R. G. S., and Hawkins, W. W.,
Can. J. Biochem. and Physiol., ^8, 51 (I960).
Mills, C. F., Biochem. J., 6^, 187 (1956).
Mills, C. F., Biochem. J., 63, 190 (1956).
Martell, A. E., and Calvin, M., "Chemistry of the
Metal Chelate Compounds," p. 19 et seq.,
Prentice Hall (1952).
Sunderman, F. W., and Boerner, F., "Normal Values in
Clinical Medicine," p. 230, W. B. Saunders (1953).
Warner, R., and Weber, I., J. Am. Chem. Soc., Z2i
5086 (1953).
Hamm, R. E., Shull, C. M., Jr., and Grant, D. M.,
J. Am. Chem. Soc., Z§.i 2111 (195*+) •
Rehm, W. S., Am. J. Physiol., 1*+*+. 115 (19*+5).
101.
102.
103.
10* + .
10?.
106.
107.
108.
109.
110.
111.
112.
113.
I l k - .
115.
121
Hober, B., "Physical Chemistry of Cells and Tissues,"
p. 272, Blakiston (1950).
Sjostrand, F. S., and Rhodin, J., Exp. Cell Res., *+,
*+26 (1953).
Danielli, J. F., in "Recent Developments in Cell
Physiology," p. 1, Academic Press (195*+).
Sollner, K., Am. N. Y. Acad. Sci., 177 (195*0.
Visscher, M. B., Fletcher, E. S., Jr., Carr, C. W.,
Gregor, H. P., Boshey, M. S., and Barker, D. E.,
Am. J. Physiol., l*+2. 550 (1§*+*+).
Overton, E., Pflugers Arch. ges. Physiol., £2, 115
(1902).
Wallerstein, R. 0., and Mettier, S. M., eds.,
"Iron in Clinical Medicine," University of
Calif. Press, Berkeley (1958).
Bessis, M. C., and Breton-Gorius, J., J. Biochem.
and Biophys. Cytology, 3, 503 (1957).
Collander, R., Trans. Faraday Soc., 33., 895 (1937).
Ussing, H. H., in "Recent Developments in Cell
Physiology," p. 33» Academic Press (195*0*
Charberek, S., and Martell, A. E., "Organic
Sequestering Agents," p. *+*+8, Wiley (1959).
Umbreit, W. W., Burris, R. H., and Stauffer, J. F.,
"Manometric Techniques and Tissue Metabolism,"
Minneapolis (19*+9) •
Roberts, S., and Kelley, M. B., J. Biol. Chem.,
222. 555 (1956).
Block, R. J., Durrum, E. L., and Zweig, G.,
"Paper Chromatography and Paper Electrophoresis,"
Academic Press, New York (1958).
Bass, R., and Saltman, P., Exp. Cell Res., 18, 560
(1959).
116.
117.
118.
119.
120.
121.
122.
123.
12* + .
125.
126.
127.
128.
129.
130.
131.
122
Bjerrum, J., Schwarzenbach, G., and Sillen, L. G.,
"Stability Constants, Part I," Chemical Society,
London (1957).
Brown, E. G., Nature, 182, 1091 (1958).
Peters, T., Giovanniello, T. J., Apt, L., and
Ross, J. F., J. Lab. and Clin. Med., J+8, 280
(1956).
Peters, T., and Giovanniello. T. J., J. Lab. and
Clin. Med., £L, b 9 k (1958).
"Handbook of Chemistry and Physics," *+0th Ed.,
Chemical Rubber Publishing Co., Cleveland (1959).
Shoden, A., and Sturgeon, P., Acta Haematol.,
(in press).
Curtin, C..0'H., and King, C. G., J. Biol. Chem.,
216. 539 (1955).
Traube, W., and Kuhbier, F., Ber., 6£, 187 (1932).
Traube, W.. Kuhbier, F., and Harting, H.,
Ber., 66, 15*+5 (1933).
Lieser, T., and Ebert, R., Ann., 532, 89 (1937 \
Bourne, E. J., Nery, R., and Weigel, H., Chemistry
and Industry, No. 30, 998 (1959).
Nissim, J. A., and Robson, J. M., Lancet i, 686
(19*+9) •
Slack, H. G. B., and Wilkinson, J. F., Lancet i,
11 (19*+9) •
Nissim, J. A., Brit. Med. J., 1, 352 (195*+) •
Bersin, T., Pharm. Acta Helv., 26, *+07 (1951) as
quoted in Honig, P., "Principals of Sugar
Technology," p. 6, Elsevier (1953).
Honig, P., "Principals of Sugar Technology," p. 5>
Elsevier (1953).
132.
133.
13*+.
135.
136.
137.
123
Aten, A. H. W., Van Ginneken, P. J. H., and
Engelhard, F. J. W., Rec. Trav. Chim. Pays-Bas,
itS, 753 (1926).
Mellor, J. W., “Comprehensive Treatise on Inorganic
Chemistry,” Vol. XIII, p. 831 et seq.,
Longmans (193*+) •
Wertheim, E., "Textbook of Organic Chemistry,"
Second edition, p. 265, Blakiston (19*+5) •
Aronoff, S., "Techniques of Radiobiochemistry,"
p. 102, Iowa State College Press (1956).
Rubin, M., and Princiotto, J. V., Conference of
Chelation Phenomena, New York, Dec. 17 (1959).
Davis, B. D., Arch. Biochem. and Biophys., £8j ^97
(1958).

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Studies On The Enzymatic Formation Of Amino Acid - Ribonucleic Acid Compounds

PDF

Studies on the transport, metabolism and chemistry of iron-sugar chelates

PDF

Lipid Metabolism In Muscular Dystrophy

PDF

The Biosynthesis Of Ergothioneine

PDF

Effects of sex hormones on cholesterol metabolism

PDF

Studies On The Biosynthesis Of Collagen In The Rabbit Embryo Skin In Vitro

PDF

Galactose metabolism in human blood cells

PDF

Isolation And Characterization Of Hageman Factor And Plasma Thromboplastin Antecedent

PDF

Studies On The Effects Of The Plant Growth Substances On Algae

PDF

Pyrimidine Nucleoside Metabolism And Antimetabolites

PDF

Factors Affecting Cholesterol Absorption And Metabolism

PDF

Studies on acyl transfer reactions on n-acylimidazoles

PDF

Studies On A Modified Fat

PDF

A comparative study of the chemistry and biology of Heparin

PDF

The separation of complete and incomplete blood group antibodies

PDF

Studies of trypsin-binding ɑ₂ macroglobulin of human plasma

PDF

The ferric complexes of citrate, ETDA and NTA: their nature and the ligand exchange reaction with human transferrin

PDF

Structure and biochemical function of the photosynthetic apparatus

PDF

Effects Of 5-Aminodeoxyuridine And 5-Aminouridine On Metabolism Of Nucleic Acids

PDF

Microheterogeneity of fetuin

Asset Metadata

Creator Charley, Philip James (author)

Core Title Studies On The Metabolism Of Iron

Degree Doctor of Philosophy

Degree Program Biochemistry and Nutrition

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag chemistry, biochemistry,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Saltman, Paul (committee chair), Mehl, John W. (committee member), Webb, John L. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-548077

Unique identifier UC11363053

Identifier 7206045.pdf (filename),usctheses-c18-548077 (legacy record id)

Legacy Identifier 7206045

Dmrecord 548077

Document Type Dissertation

Rights Charley, Philip James

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA